PE EIE[R-Rg RESEARCH ON - HJ Andrews Experimental Forest

PE EIE[R-Rg 

RESEARC H 

O N 

CONIFEROUS 

FOREST 

ECOSYSTEMS ° 

a symposiu m 

I77t 

THIS PUBLICATION CONSTITUTES A CONTRIBUTION TO TH E 

U.S./INTERNATIONAL BIOME PROGRAM

The editors : 

JERRY F. FRANKLIN, Principal Plant Ecologist, 

Pacific Northwest Forest 

and Range Experiment Statio n 

L . J. DEMPSTER, Consulting Editor , 

Corvallis 

RICHARD H. WARING, Associate Professor , 

Oregon State University , 

Corvallis 

Proprietary or brand names are used only to document actua l 

experience ; their use does not imply approval of the product to th e 

exclusion of others which may also be suitable .

Research on 

Coniferous Forest Ecosystems : 

First Year Progress in th e 

Coniferous Forest Biome, USu RP 

Proceedings of a Symposiu m 

held at 

NORTHWEST SCIENTIFIC ASSOCIATIO N 

Forty-fifth Annual Meeting 

Bellingham, Washingto n 

March 23-24, 197 2 

Edited by 

Jerry F. Franklin 

L . J. Dempster 

Richard H . Waring 

Published in 1972 by 

Pacific Northwest Forest and Range Experiment Statio n 

Forest Service, U .S. Department of Agriculture 

Portland, Orego n 

For sale by the Superintendent of Documents, U.S . Government Printing Offic e 

Washington, D .C., 20402 - Price $2 .50 Stock Number 0101-0233

Proceedings-Research on Coniferous Forest Ecosystems-A symposium . 

Bellingham, Washington-March 23-24, 197 2 

Foreword 

The research program organized under th e 

Coniferous Forest Biome is probably th e 

largest and most comprehensive single effor t 

at ecosystem analysis being carried out in th e 

Western United States . After a long period o f 

planning and discussion, it is finally in its firs t 

year of full-scale activity . Despite this youthful 

state there is great interest among ecologists 

both within and beyond the Biom e 

"boundaries" in the conceptual basis for th e 

Biome's program and present and planne d 

research . 

Consequently, at the invitation of th e 

Northwest Scientific Association, members o f 

the Coniferous Forest Biome organized th e 

symposium, "Research on Coniferous Fores t 

Ecosystems : First Year Progress in the Coniferous 

Forest Biome, US/IBP." The symposium 

was presented at the 45th Annual 

Meeting of the Northwest Scientific Association 

on March 23 and 24, 1972 . It highlighted 

the concepts and plans underlying major segments 

of the Biome program and the numerous 

new insights, techniques, and data which 

are resulting from the varied research activities . 

This volume contains the proceedings of 

that symposium. Publication was made 

possible only by the full cooperation and considerable 

contributions of the Pacific North - 

west Forest and Range Experiment Station, 

USDA Forest Service . Station staff not only 

assisted significantly in the editorial phases 

but handled the production in its entiret y 

from edited copy to final printing . Furthermore, 

this was done under severe deadlines in 

order that the volume be available for th e 

Fifth Assembly of the International Biologica l 

Program in August of 1972 . 

The editors would also like to acknowledg e 

the numerous contributions others made t o 

the production of this volume : To Drs . Dal e 

W. Cole and Frieda B . Taub for their significant 

part in organizing and conducting the 

symposium ; to the majority of authors an d 

referees who cooperated magnificently i n 

producing the necessary materials unde r 

severe time limitations ; to George M . Hansen , 

Betty J . Bell, Mildred I . Hoyt, and others i n 

Editorial Services of the Pacific Northwes t 

Forest and Range Experiment Station, wh o 

bore the burden of preparing the volume fo r 

printing ; and to Robert Romancier, Glend a 

Faxon, Virginia Hunt, James Overholser, Han s 

Riekerk, Charles Grier, and Laura Gregg fo r 

their part in obtaining the papers, typing, an d 

otherwise assisting in production of th e 

volume . 

Jerry F . Franklin 

L . J . Dempster 

Richard H . Waring

Contents 

SOME BROADER VIEWS OF BIOME ACTIVITIES 1 

Why a Coniferous Forest Biome ? 

Jerry F. Franklin 3 

Organization and research program of the Western Coniferous Forest Biome 

S . P. Gessel 7 

Findley Lake-the study of a terrestrial-aquatic interfac e 

P. R. Olson, D . W. Cole, and R. R. Whitney 1 5 

A comparative study of four lake s 

Frieda B . Taub, Robert L . Burgner, Eugene B . Welch, and Demetrios E . Spyridakis . . . . 2 1 

The modeling process relating to questions about coniferous lake ecosystem s 

Douglas M . Eggers and Larry M . Male 3 3 

Toward a general model structure for a forest ecosyste m 

W . Scott Overton 3 7 

Hydrologic modeling in the Coniferous Forest Biom e 

George W. Brown, Robert H . Burgy, R. Dennis Harr, and J. Paul Riley 4 9 

Preliminary considerations of the forest canopy consumer subsyste m 

M . A. Strand and W. P. Nagel 7 1 

An environmental grid for classifying coniferous forest ecosystems 

R. H. Waring, K . L. Reed, and W . H . Emmingham 7 9 

Pag e 

WATER AND NUTRIENT MOVEMENT THROUGH ECOSYSTEMS 9 3 

Modeling water movement within the upper rooting zone of a Cedar River soi l 

W. H. Hatheway, P . Machno, and E . Hamerly 9 5 

Elemental transport changes occurring during development 

of a second-growth Douglas-fir ecosyste m 

Charles C. Grier and Dale W. Cole 10 3 

Nutrient budget of a Douglas-fir forest on an experimental watershe d 

in western Orego n 

R. L. Fredriksen 11 5 

Nutrient cycling in throughfall and litterfall in 450-year-old Douglas-fir stand s 

Albert Abee and Denis Lavender 133

Page 

ESTIMATING BIOMASS AND OTHER STATE VARIABLES 14 5 

Direct, nondestructive measurement of biomass and structure i n 

living old-growth Douglas-fir 

William C . Denison, Diane M . Tracy, Frederick M . Rhoades, and Martha Sherwood . . . .14 7 

Estimation of biomass and transpiration in coniferous forests using tritiated wate r 

J. R. Kline, M . L . Stewart, and C. F. Jordan 15 9 

Theodolite surveying for nondestructive biomass samplin g 

Eugene E. Addor 16 7 

Estimates of biomass and fixed nitrogen of epiphytes from old-growth Douglas-fi r 

Lawrence H . Pike, Diane M . Tracy, Martha A. Sherwood, and Diane Nielsen 17 7 

Litter, foliage, branch, and stem production in contrasting lodgepole pin e 

habitats of the Colorado Front Range 

William H . Moir 18 9 

Small mammal and bird populations on Thompson Site, Cedar River : 

parameters for modelin g 

Sterling Miller, Curtis W . Erickson, Richard D . Taber, and Carl H . Nellis 19 9 

TERRESTRIAL PROCESS STUDIES 20 9 

Terrestrial process studies in conifers-a review 

R. B. Walker, D . R . M. Scott, D . J. Salo, and K . L. Reed 21 1 

Criteria for selecting an optimal model : terrestrial photosynthesi s 

Kenneth L. Reed and Warren L . Webb 22 7 

A model of light and temperature controlled net photosynthetic rate s 

for terrestrial plants 

Warren L. Webb 23 7 

Energy flux studies in a coniferous forest ecosyste m 

Lloyd W . Gay 24 3 

The lysimeter installation on the Cedar River Watershe d 

Leo J . Fritschen 25 5 

Initial steps in decomposition of Douglas-fir needles under forest conditions 

P. L. Minyard and C . H. Driver 26 1 

Seasonal and diurnal patterns of water status in Acer circinatu m 

James P. Lassoie and David R . M. Scott 26 5 

Development and testing of an inexpensive thermoelectrically cooled cuvett e 

David J . Salo, John A . Ringo, James H . Nishitani, and Richard B . Walker 273

Page 

AQUATIC PROCESS STUDIES 279 

Studying streams as a biological uni t 

James R . Sedell 28 1 

Exploring the aquatic carbon web 

Bruce Lighthart and Paul E . Tiegs 28 9 

Dynamics of nutrient supply and primary production i n 

Lake Sammamish, Washingto n 

Eugene B . Welch and Demetrios E. Spyridakis 30 1 

Hydroacoustic assessment of limnetic-feeding fishe s 

Richard E. Thorne 317

Some Broader View s 

of Biome Activities



Why a Coniferous Forest Biome? 

Jerry F . Frankli n 

Principal Plant Ecologis t 

Forestry Sciences Laboratory 

Pacific Northwest Forest an d 

Range Experiment Statio n 

USDA Forest Servic e 

Corvallis, Oregon, 

and Deputy Directo r 

Coniferous Forest Biom e 

US/I B P 

Abstract 

This introduction to the symposium outlines some of the reasons large integrated ecological researc h 

programs have been deemed necessary despite problems inherent in such "large science" programs . 

The Coniferous Forest Biome program is a n 

interdisciplinary research effort concerned 

with the structure and functioning of coniferous 

forest and associate aquatic ecosystems , 

particularly as they occur in western Nort h 

America. Initiated as a part of the U .S. International 

Biological Program it is one of si x 

biome programs organized in major biotic - 

environmental divisions or regions of th e 

United States-Deciduous Forest, Grassland , 

Desert, Tundra, Tropical, and Coniferous 

Forest. The first major funding of the Coniferous 

Biome occurred in September of 1970 , 

and there are now over 100 scientists from a 

total of 15 universities, national laboratories , 

and agencies involved in the program . 

Many aspects of this large and vital researc h 

program will be described in papers whic h 

follow. However, questions continue to arise , 

such as, Is this program necessary? What differentiates 

the Biome efforts from the numerous 

existing research programs on coniferou s 

forests, both large and small? Why has thi s 

"big science" effort been mounted ? 

So, before looking in detail at the Biome' s 

research activities, I'd like to share with yo u 

some of my views on the need for and impor - 

tance of the Biome program . These views are 

not all inclusive nor necessarily shared by al l 

involved in the program . 

To get close to home, let's consider firs t 

the increasingly complex nature of problem s 

facing forest land managers. Questions used to 

be relatively simple : What cutting system will 

regenerate a desired forest type? What effec t 

do different thinning or fertilizing regime s 

have on production of merchantable timber? 

What herbicides will most effectively permi t 

reforestation and discourage brush on cutover 

areas? Such questions are still interesting an d 

important . However, many of the more critical 

land management questions reach far beyond 

the relatively narrow confines of maximizing 

production of goods from forest lands : 

How do different cutting methods influence 

the flow and quality of water and, further , 

the characteristics of aquatic communities? 

What effect does fertilization have on th e 

nutrient content or fertility of the wate r 

draining from the treated forest land? What 

happens to various pesticides when applied to 

the land-how fast are they degraded, where 

do they accumulate, and what effect do the y 

have on nontarget organisms and the eco - 

3

system as a whole? And, if the biological an d 

physical questions were not sufficiently difficult 

(and they are), overriding economic an d 

social considerations further compound th e 

complexity of land management problems . 

The basic knowledge of natural science required 

for rational resolution of some of th e 

larger questions has thrust some new and 

difficult demarJds on the scientific community. 

Of coursel large amounts of information 

are required ; we often hear of an information 

explosion but, in fact, we are faced with a 

"demand" explosion-available data are to - 

tally inadequate to meet the information requirements 

of policymakers. More important , 

however, is the need for new kinds of information, 

particularly information on linkage s 

between the parts of forest ecosystems, suc h 

as land and water, and between economics , 

sociology, and natural sciences . In effect, 

information is required which involves th e 

links or relationships between the traditiona l 

units or disciplines of scientific attention an d 

organization . 

In addition to the study of linkages amon g 

different components of ecosystems, we ar e 

searching for properties that are unique to 

whole systems. We are asking how much tim e 

is required for an ecosystem to fully recover 

following disturbance, how many component s 

are redundant in the system, what are the 

major selection pressures to which ecosystems 

respond, and finally, are we forcing ecosystems 

to adapt to changes faster than is 

possible? 

It is also apparent that research program s 

must be planned so as to anticipate unforeseen 

questions and needs as well as answer 

immediate questions . In other words, scientists 

must work toward development o f 

general principles and models which will allo w 

us to predict from past research what wil l 

happen in a new situation which has not ye t 

been a subject of detailed study . Example s 

might be the use of basic models to anticipate 

the effect of a cutting system in a new forest 

environment or of a newly developed pesticide 

on the animal component of a forest 

stand. The importance of this capability for 

generalization is related to another recent 

phenomenon, the insistence that scientists 

provide immediate "best " answers to questions 

based on present available knowledge ; 

society is unwilling to wait for long periods 

for "final" solutions . 

These increased demands on the scientific 

community have to be met without the great 

increases in money and manpower experienced 

in the earlier postwar period . Since 

scientific resources are limited, they must be 

utilized more efficiently to meet society's 

needs for problem-solving information . Relevance 

and efficiency have become important 

considerations in planning and funding research 

programs . 

In many respects, the demands outlined 

above run counter to the traditional ways o f 

doing scientific research . Science has been 

strongly disciplinary in character with the 

greatest rewards going to the specialist who 

pursued his field in great depth . Even the applied 

scientist has tended to have a narro w 

focus looking, for example, at the effect of a 

cutting method on regeneration or thinning 

on wood yields, not at the overall effects of 

such treatments on ecosystems . 

From another viewpoint, science has been 

likened to an edifice of bricks gradually built 

up by the effort of many individual scientists . 

Unfortunately, there has been a strong tendency 

for each scientist to produce bricks of a 

dimension, shape, and material primarily of 

interest to him. Furthermore, at least in 

ecology, we have lacked overall blueprints for 

our edifices which would provide direction as 

to the kinds and number of bricks we need . 

There can be no questioning the need fo r 

strong basic research programs ; they are essential 

to the advancement of science and huma n 

knowledge. Traditional viewpoints and approaches 

do not adequately meet a great 

many of today's needs, however . . 

the scientific community is being 

forced to restructure a large part of its efforts , 

partially from a sense of its own responsibilities 

and partially in response to society ' s 

pressures. Nowhere is this more evident than 

in the fields of natural resources and ecology . 

New programs are interdisciplinary effort s 

which try to overcome inadequacies in pas t 

research efforts . Resources are concentrated 

on critical areas related, either directly or in - 

4

directly, to solution of major natural resourc e 

problems . 

The Coniferous Forest Biome program i s 

one of these integrated, interdisciplinary efforts. 

It has as its overall objective an under - 

standing of how materials, such as water an d 

nutrients, and energy enter, move through , 

and leave coniferous forest ecosystems, including 

both the terrestrial and aquatic components. 

The Biome program shares some 

common characteristics with many other larg e 

new ecological research programs : (1) It involves 

groups of scientists from many different 

fields working together toward some common 

goals. (2) It has a "systems" orientation ; 

it is concerned with all parts of the ecosystem, 

its total behavior, and the linkage s 

between the various components, not with a 

single piece . (3) Mathematical descriptions o r 

models of the various processes, subsystems , 

and total ecosystem are a key to organizatio n 

and synthesis of the effort . Conceptua l 

models provide the framework for structuring 

the research effort and determining dat a 

needs. (4) Continuous communication between 

the scientist participants, including 

sharing of data, is an essential feature . The 

efforts progress via the constant interchange 

between scientists, between modeling and th e 

field and laboratory research . 

In our symposium, we will try to introduc e 

you to the kinds of research being conducte d 

under the auspices of the Coniferous Forest 

Biome . The papers range widely in scope fro m 

general presentations on the conceptual basi s 

for major program segments to results of relatively 

narrow research projects . Unfortunately, 

it is too early to provide any major 

synthesis of activities ; this summer will be the 

first field season of essentially full funding . 

We have tried to emphasize the new concepts , 

techniques, and data which are emerging fro m 

the Biome's efforts . We hope that the 

"samples " of activities and philosophy which 

follow will make clearer what the Biome i s 

trying to do and how we are going about it . 

Acknowledgments 

The work reported in this paper was sup - 

ported in part by the U .S. Forest Service in 

cooperation with the Coniferous Fores t 

Biome, U .S. Analysis of Ecosystems, International 

Biological Program . This is Contribution 

No . 18 to the Coniferous Forest Biome . 

5



Organization and research program 

of the Western Coniferous Forest Biom e 

An Integrated Research Component of th e 

International Biological Program 

S. P . Gessel, Directo r 

Western Coniferous Forest Biome 

College of Forest Resources 

University of Washingto n 

Seattle, Washingto n 

Abstract 

The International Biological Program was initiated in 1964. The Coniferous Biome research of the Analysis 

of Ecosystems component did not begin until October 1970 . Other Biome programs, with the exception of th e 

Tropical, preceded the Coniferous with the first research in the Grasslands Biome beginning in 1968 . The IBP 

research, originally planned to terminate in 1972, has been extended in the United States through 1974 . It is 

likely that the Analysis of Ecosystems program will be extended beyond that date in order to give the integrate d 

research approach an adequate test. 

This introductory paper describes the administrative and research organization of the Coniferous Biome and 

presents the broad objectives. Principal research sites and the plans for extension to other areas of the Biome are 

also described. As research is in the very early stages, progress can be best judged by reviewing other papers in thi s 

Symposium. Aside from accumulation ofdata and development ofmodels, we have made significant contributions 

to the education ofboth undergraduate and graduate students by introducing new modeling concepts . 

Introduction 

The International Biological Program (IBP ) 

began in 1964 with the specific objective of developing 

a world study on the biological basi s 

of productivity and human welfare . The initiators 

were concerned with world-wide problems 

associated with increasing populations , 

world food supplies, changes in environmenta l 

degradation, and increasing resource use . 

It was originally conceived as an 8-year pro - 

gram to be completed by 1972 . However, be - 

cause many national programs started late , 

IBP has been extended through 1974. There 

are clear indications, however, that productiv e 

research programs such as the Biome effort s 

will continue beyond 1974 . 

The nature of IBP programs and rate of 

attainment of specific research objectives has 

been variable in the many participating countries. 

Rate of progress has been closely related 

to funds invested in the research . This discussion 

concerns the United States program and 

primarily those sections of the program under 

which Biome research falls. Our office will 

gladly supply interested scientists with references 

on the broader objectives and progra m 

of IBP . 

The United States IBP effort has evolved a s 

a series of large-scale, integrated research pro - 

grams coordinated by a national committee 

(sponsored by the National Academy of 

Sciences) and with the financial support of 

the National Science Foundation (NSF) . 

The Integrated Research Programs (IRP's ) 

are in two broad groups : 

A. Human Adaptability 

1. International study of Eskimo s 

2. Population genetics of American Indian s 

3. Biology of human population at high 

altitudes 

7

4. Nutritional adaptation to environmen t 

5. Ecology of migrant peopl e 

B. Environmental Managemen t 

1 . Operating programs 

a. Convergent and divergent evolutio n 

in the Americas 

b. Experimental biography of the sea 

c. Physiology of colonizing specie s 

d. Atmospheric dispersal (aeriobiology 

program ) 

e. Analysis of ecosystem s 

f. Conservation of ecosystems fo r 

scientific purposes 

g . 

Chemical and biological control o f 

organism s 

2 . Programs proposed but without funding 

a. Conservation of genetic material s 

b. Crop production under stress 

c. Productivity and conservation of 

marine mammals 

d. Nitrogen management 

e. Phenology 

U .S. IBP activities, particularly the Integrated 

Research Programs, provide a different 

focus for scientific inquiry ; i .e., integrated 

groups of scientists working toward solutio n 

of common problems instead of individual s 

working alone . The changes in focus have produced 

some trauma and a cadre of critics . 

Criticisms are still abundant but help to insur e 

better programs. Some critics say that integrated 

research efforts result in mediocre 

science, but program accomplishments mus t 

speak to this point now and in the future . 

Analysis of Ecosystems 

and th e 

Coniferous Forest Biome 

Having dealt with generalities of the total 

IBP, I now wish to explain briefly some specific 

parts of the total research effort. A s 

noted, the environmental management segment 

of U .S. IBP includes a subdivision calle d 

" Analysis of Ecosystems ." This is the largest 

and most active U .S . IBP program . It was originally 

conceived by Dr. Fred Smith of the 

University of Michigan, now at Harvard. The 

basic concepts involved are that world 

environments can be placed within broad 

units called Biomes and that integrated Biom e 

research efforts will provide the understanding 

for using and conserving the resources of 

that environment . 

Initially, the Analysis of Ecosystems pro - 

gram developed slowly . The Grassland Biome 

was organized first and has been a guide for 

other Biome programs . The total research 

effort will include the following Biomes : 

Grasslands Coniferous forest 

Desert 

Tundra 

Deciduous forest Tropical fores t 

Figure 1 illustrates the world environmenta l 

distribution of the Biomes . 

The first five Biomes are now functioning . 

The Tropical Forest Biome is still being organized; 

however, NSF is supporting tropical ecosystems 

research by such groups as the Organization 

for Tropical Studies . 

Figure 1 . Distribution of Biomes . 

The overall objectives of the Analysis of 

Ecosystems program are : 

1. To establish a scientific base for program s 

to maintain or improve environmenta l 

quality ; 

2. To derive broad principles of ecosyste m 

structure and function through an integration 

of the results of the six Biome studies ; 

3. To relate these principles to characteristic s 

of ecosystems such as persistence, stability , 

maturity, and diversity ; and 

4. To develop and refine a generalized adapt - 

able simulation model suitable for use in 

8

planning studies for new developmen t 

projects . 

Generalized objectives of Biome researc h 

include : 

1.To determine the driving forces, the processes 

causing transfers of matter and energy 

among components, the nonconcentratio n 

characteristics, and the controlling variable s 

in each Biome ; 

2. To determine the ecosystem response t o 

the natural and man-induced stresses appropriate 

to each Biome ; for example, large 

herbivores in grasslands, extreme and rar e 

weather patterns in deserts, periodic fluctuations 

of rodent populations in tundra , 

commercial use of timber in the coniferous 

forests, urbanization in the deciduou s 

forests, and nutrient retention in the wet 

tropical forests ; 

3. To understand the land-water interactio n 

characteristics of each Biome : Prairie pond s 

and reservoirs in the grasslands ; the abundance 

of shallow waters in wet tundra ; 

springs and temporary waters in deserts ; 

river and lake systems with anadromou s 

fish populations in coniferous forest ; pollution, 

and eutrophication in the deciduous 

forest; and large river systems in tropical 

forests; and 

4. To synthesize the results of these and previous 

studies into predictive models o f 

temporal and spatial variation, effects of 

pollutants and of exploitation, stability , 

and other ecosystem characteristics necessary 

for resource management in eac h 

Biome . 

The Coniferous Forest Biome 

The history and research progress of ou r 

own Coniferous Forest Biome brings this discussion 

somewhat closer to home . Coincident 

with the development of an Analysis of Ecosystems 

program and the initial organizatio n 

and funding of the Grasslands Biome, a large 

body of scientists in forested areas of the 

West began discussing a Coniferous Fores t 

Biome. The University of Washington hoste d 

the group's first formal meeting at Pack 

Forest in February 1968 which produced a 

plan for development of a Biome research 

organization . The University of Montana 

helped to continue the discussions with a 

symposium on "Coniferous Forests of the 

Northern Rocky Mountains " held at Missoula 

in September 1968 . 

In early 1969, I accepted an invitation to 

become the Director of the Coniferous Forest 

Biome . An organizational proposal for establishment 

of a skeletal office and development 

of a Biome research program was prepared 

and submitted to NSF . 

With these very limited funds and som e 

financial assistance from the Central Ecosystem 

office and the University of Washing - 

ton, and with a great deal of volunteer labor , 

a research proposal was prepared and submitted 

to NSF on December 31, 1969 . The 

initial $460,000 request was scaled to fund s 

NSF had available. The proposal was approved 

but not funded until September 15 , 

1970. We are, therefore, still in the very earl y 

stages of research . Even before the first proposal 

was funded, a second-year proposal wa s 

prepared and this, in turn, was funded at 1 . 2 

million dollars on January 1, 1972 . We are 

now preparing a proposal for 1973 and 197 4 

with a submission date of July 1 . 


Organization and Research 

The Central Biome office is at the University 

of Washington's College of Forest Re - 

sources . Intensive research sites have been 

established on the H. J. Andrews Experimental 

Forest of the U .S. Forest Service in 

the Oregon Cascade Range and the Cedar 

River watershed of the City of Seattle in th e 

Washington Cascade Range . Although many 

other areas had been nominated as potentia l 

research sites, these two were chosen because 

of their concentrations of available scientists , 

developed facilities, history of research, an d 

ability to provide the types of research area s 

needed to accomplish the research goals of 

the Biome . 

The total Biome research program originally 

called for additional coordinating site s 

9

where validation studies could be carried out . 

Funds initially restricted development outside 

the intensive sites ; however, plans for coordinating 

programs, sites, and activities are no w 

being developed by the Extrapolation an d 

Application Committee under the direction o f 

Dr. William Laycock of the U .S. Forest Service. 

It appears that a diverse group of Biome - 

wide or coordinating site activities including : 

(1) major efforts to coordinate with other on - 

going ecosystem research programs in th e 

Biome; (2) cooperative studies of ponderos a 

pine ecosystems with the Grasslands Biome ; 

(3) examinations of fire and disease influences 

on coniferous forests ; and (4) validation o f 

some initial modeling efforts will develop . 

The Biome 's organizational structure use d 

in developing, organizing, and administerin g 

the research is shown in figures 2 and 3 . The 

overall Biome research plan is developed b y 

the Scientific Directorate, but the individua l 

research proposals are generated by the re - 

search committees . Yearly programs develop 

about as follows . First, research tasks or 

objectives are defined. Funding guidelines fo r 

each of these are also outlined based on total 

funding expected from NSF . Research committees 

then define specific subprogra m 

objectives and after agreement by the Scientific 

Directorate, the chairman of a researc h 

committee calls for proposals from the scientific 

community. Proposals are reviewed b y 

research committees and by the Scientifi c 

Directorate for task relevance and adequacy . 

Following revisions, deletions, and additions 

the proposals are incorporated into the Biom e 

proposal. A group of scientists from othe r 

Biomes review the draft of the Biome proposal 

before its final preparation and sub - 

mission to NSF . 

Objectives of Coniferou s 

Forest Biome Research 

I will not attempt to detail the research i n 

this brief report, but I should comment o n 

the main objectives of all Biome research, i .e . , 

development and validation of models of ecosystems 

function. The purpose of producing 

these models is to provide understanding o f 

coniferous ecosystems useful for management 

BIOME ADVISORY BOARDI DIRECTOR] -- ANALYSIS OF ECOSYSTEM S 

EXECUTIVE BOAR D 

{ CENTRAL OFFICE 

] 

[- TECHNICAL COMMITTEES -1 

Chemical Analyse s 

Remote Sensin g 

Genetic s 

Phenolog y 

Data Managemen t 

Internationa l 

DEPUTY DIRECTORS 1- EXECUTIVE DIRECTORATE 

SITE 

DIRECTOR S 

Terrestrial Division 

Aquatic Divisio n 

Ecosystem Integratio n 

Division 

SCIENTIFI C 

DIRECTORATE 

H . J . Andrew s 

Lake Washington Drainag e 

Validation Site s 

TRRESTRIAL DIVISIO N 

- ECOSYSTEM INTEGRATION DIVISIO N 

AQUATIC DIVISIO N 

Primary Production Modeling Management Lakes 

Food Chain Processes Interface Streams 

Biogeochemical Processes 

Extrapolation and Application 

Physical Processes 

Figure 2. Coniferous Forest Biome organizational chart . 

10

DIVISION 

COMMITTEES AND SUBCOMMITTEE S 

PRIMARY PRODUCTION FOOD CHAIN PROCESSES BIOGEOCHEMICAL CYCLING PHYSICAL PROCESSE S 

TERRESTRIAL 

1 . 

2 . 

Processe s 

Biomass an d 

Structure 

1 . 

•2, 

Biologica l 

Geochemical 

1 . 

2 . 

Hydrolog y 

Meteorolog y 

LAKES 

STREAMS 

.- 

AQUATIC 

1 . 

2 . 

3 . 

Water Column Processe s 

Bottom Related Processe s 

Higher Consumer Processe s 

MODELING MANAGEMENT INTERFACES EXTRAPOLATION AND APPLICATIO N 

ECOSYSTE M 

INTEGRATION 

Figure 3 . Research committees detail . 

and protection purposes . The following statement 

by Dr . Chapman, Deputy Director for 

Biome modeling activity, explains Coniferou s 

Biome modeling philosophy . 

The primary aim of this ongoing re - 

search is to increase our understanding 

of whole ecosystems within the Wester n 

Coniferous Forest Biome ; and to this 

aim an area of high priority is the development 

of models . . . with special emphasis 

on modeling of the aquatic-terrestrial 

interface . 

The analysis of an ecosystem shoul d 

begin by constructing models from existing 

data . The second step should be collection 

and analysis of data for testing an d 

refining the first models . This require s 

participating scientists to spend consider - 

able time thinking how their part of th e 

system relates to the rest and about general 

modeling processes. As the kinds of 

additional data that are needed are identified, 

scientists will be able to conduc t 

specific field studies designed to furthe r 

improve the models . This approach establishes 

a continual feedback loop . 

It is important, then, that participatin g 

scientists consider the understanding of 

systems modeling to be a major responsibility 

. The necessary modeling cannot be 

done by, say, a systems engineer, biometrician, 

or biomathematician . It will not 

be necessary for each participant to be - 

come a full-fledged system engineer or 

biomathematician (although it would be 

highly desirable if a few did so), but it will 

be essential that each participant re-orient 

his approach and philosophy somewhat , 

for all should become systems modelers , 

in one capacity or another . 

These points, made in our first proposal, can be 

reemphasized in other ways . In the first place , 

every scientific study beyond simple observation 

involves some type of model-verbal , 

graphical, or mathematical . The development 

of subsystem models is the responsibility of th e 

scientist doing the substantive research an d 

constitutes an essential part of their annual 

reports to the Biome . The modeling group provides 

support for those in need of modeling 

assistance and in fact has such a supportiv e 

function as one of its primary missions . 

The other primary mission of the modeling 

group is the development of the overall ecosystem 

model . This is one ultimate aim of th e 

Biome study, and while still at an early stage 

it will be possible to construct a total system 

computer model of low resolution . A better 

model will be achieved when reasonably 

sophisticated functional subsystem model s 

1 1

have been developed . Prior to their development 

the overall ecosystem model will be for 

the most part a translation of simple graphica l 

and verbal models into mathematical language 

with many aspects of the system interpolate d 

on the basis of empirical submodels or eve n 

crude estimates . 

Ultimately both total systems models an d 

submodels will require testing. This testing 

will indicate the further research needed to 

answer new questions or refine parts of existing 

models . The modeling team plays a role in 

assisting in the design of such new research. I n 

this way the model building plays an essentia l 

feedback role in the program . 

In addition, both subsystem models an d 

larger ecosystem models require estimates of 

parameters and available data often appear s 

inadequate. In these circumstances, it is neces - 

sary to test the sensitivity of the model to th e 

errors of estimation. Does it make a great deal 

of difference if the estimates are crude? If 

not, the model may proceed . If it does make a 

significant difference, additional observation s 

are required and further research in such area s 

is necessary. Obviously, such sensitivity tests 

cannot be made until models are fairly wel l 

developed ; some models will be at this stage 

this year. 

Specific Goals 

The goal of our Coniferous Biome research 

is development of a basic understanding of 

coniferous forest ecosystems, including both 

terrestrial and aquatic components, so tha t 

ecological constraints on and opportunitie s 

for increased production of fiber, food, water , 

and wildlife can be recognized . The overall 

strategy includes identification of the majo r 

components and processes, both physical an d 

organic, within the ecosystem, and definition 

of their interrelationships . The definition of 

interrelationships will be accomplished 

through a systems analysis and modeling pro - 

gram . As an overall guide for Biome research , 

we recognize the following general objectives : 

1. To determine the major factors, both components 

and processes, that control the 

productivity and distribution of organisms 

in coniferous forest ecosystems, includin g 

(a) an analysis of the structure and distribution 

of the principal resources, (b) definition 

of the functional relationships between 

biotic, decomposer, consumer, and 

producer components of the systems, an d 

(c) analysis of the forms and degrees of 

stability in these systems . 

2. To examine the linkage of terrestrial an d 

aquatic components in coniferous fores t 

ecosystems, including (a) water, energy , 

and transport of chemicals (including pesticides), 

(b) direct transport of terrestrial 

products into the aquatic system, e .g. , 

through litter fall and surface erosion, an d 

(c) return of organic and inorganic materia l 

from the aquatic to the terrestrial environment 

through movements of fish, birds , 

and insects . 

3. To determine how various manipulation s 

influence the structure and function of 

coniferous forest ecosystems using bot h 

unit watersheds and plot studies . Special 

attention is directed to the influences of 

manipulations on (a) stability and productivity 

of these systems and (b) the linkage s 

between terrestrial and aquatic components 

of the systems. 

4. To understand population dynamics of 

those major components of each trophic 

level which appear to influence significantly 

the sustained productivity and stability 

of various coniferous forest ecosystems 

within the Biome . 

5. To produce models of temporal and spatia l 

variations in coniferous forest ecosystem s 

or system components. These models will 

include factors affecting productivity an d 

stability of the systems and the linkages between 

terrestrial and aquatic environments , 

forecasting the behavior of these systems 

and their relationships to human manipulation 

. 

6. To apply to specific models in the solutio n 

of major use problems in the Biome are a 

and assist other groups or agencies to d o 

likewise . 

Since these long-term objectives provid e 

only general guidance, considerable time an d 

effort is spent in defining annual researc h 

tasks in order to focus the research an d 

12

modeling activities and provide a basis fo r 

phasing of the program . 

The specific research tasks planned for 

1973 and 1974 are to : 

1. Complete initial programs in terrestrial production 

process measurement and modeling, 

using both physiological and meteorological 

techniques, and link results to behavior 

(net productivity, transpiration , 

etc.) of actual stands . Begin developing 

data for the production process models fo r 

other species and environments and for 

different age classes important in th e 

Biome, including linkages of individual tre e 

processes to behavior of stands . 

2. Develop elemental cycling models of forest 

stand level ecosystems based upon proces s 

and transfer functions such as decomposition, 

ionic leaching, weathering, organism 

uptake, storage, and return, and meteorological 

and biological additions . This will 

include stressed systems. This fine resolutio n 

work will serve as a basis for development o f 

unit watershed programs . 

3. Begin describing paths of energy and material 

flow through consumers (food webs) in 

coniferous ecosystems under study, wit h 

particular attention to movement and accumulation 

of toxic materials and consequen t 

effects of diversity, flow paths, and primar y 

producers . 

4. Complete initial nutrient, water, and energ y 

flow models for unit watersheds and begi n 

their refinement with particular attention to 

incorporation of process models . For selected 

compartments and transfers and i n 

simplified form, begin examining the abilit y 

of the watershed models to describe behavior 

of (a) larger drainages and (b) strongl y 

contrasting coniferous ecosystems across th e 

Biome, and to predict (c) system response s 

under various stresses and over long tim e 

spans (in connection with manipulation an d 

successional research) . 

5. Quantify and model paths and rates of material 

and energy transfers across terrestrial - 

aquatic interfaces, specifically includin g 

quantification of transfers in and comparisons 

between (a) undisturbed small stream s 

and small lakes, (b) lakes of various topolog y 

and climate, and (c) undisturbed and clear - 

cut watersheds with stream systems . In the 

comparisons of interface transfers in undisturbed 

and altered systems, study consequences 

of terrestrial disturbances on aquat - 

ic productivity and stability with initial 

emphasis on stream systems . 

6 . Evaluate and model the trophic dynamic s 

of four lakes of significantly different 

nutritional status, topology, and climate . 

Evaluation and modeling will be by way of 

three subdivisions of the aquatic system : 

(a) the water column, including the rapidly 

changing community of phytoplankton , 

microbes, and zooplankton, (b) the 

benthos, including the moderately slowly 

changing community of larger plants, sediments 

and longer life span invertebrates , 

and (c) higher consumers-long life spa n 

invertebrates and fish . Integrate these submodels 

into a total system productivity 

model that allows determination of importance 

of various internal and externa l 

nutrient sources to lake productivity an d 

suggests methods of management of productivity. 

Finally, explore opportunitie s 

for validation elsewhere . 

7. Examine responses of coniferous ecosystems 

or subsystems to selected stresse s 

(manipulations) in terms of material and 

energy flows and of productivity . This tas k 

is largely structured as validation of model s 

and will utilize existing as well as newl y 

acquired data. Manipulations may includ e 

(a) clearcutting, (b) fertilization, (c) addition 

of toxic materials (pesticides an d 

heavy metals), (d) addition to or remova l 

of nutrients from aquatic systems, and 

(e) burning . 

8. Examine long-term (successional) behavio r 

of coniferous ecosystems specifically including 

: (a) development of a successiona l 

model for prediction of forest compositio n 

and biomass changes, primarily by cooperation 

with the Deciduous Forest Biome an d 

acquisition of necessary data base for 

validation of the Oak Ridge model(s) ; 

(b) descriptions of selected subsyste m 

structures and processes across different 

forest age classes to test successional hypotheses 

regarding : (1) complexity and redundancy 

of detritus subsystems, (2) ratio 

13

of gross production and respiration, and 

(3) conservation of nutrients within the 

system ; (c) examination of eutrophication 

processes in lake systems using a series o f 

lakes of varying "ecological stage" (nutritional 

status) . 

Each of these tasks involves an integratio n 

of efforts by several discipline groups . Expanded 

and more rapid interchange betwee n 

discipline groups, specialists, and modelers , 

and between interdisciplinary units workin g 

on different levels of resolution or systems, i s 

essential and will be pursued by all feasibl e 

methods. Expanded contact, including personnel 

exchange will be developed with othe r 

Biomes . Specific cooperative programs wit h 

the Deciduous Biome on succession and wit h 

the Grassland Biome on ponderosa pine - 

grassland systems are planned . Extension of 

program activities to a Biome-wide basis i s 

essential to the 1973-74 program . 

Progress 

The papers presented in this symposium 

will indicate to some extent the progress in 

Biome research . Something which perhap s 

will not be as apparent in the presentations i s 

our accomplishments in working together so 

that the total of our group research effort is 

greater than the sum of the individual parts . I 

believe some of our best progress has been i n 

getting terrestrial researchers to consider th e 

relationship of their work to aquatic components 

and how these systems interface. Th e 

same kind of crosslinks have been establishe d 

among different disciplines on each campu s 

involved in the study and perhaps even mor e 

important across state and county boundarie s 

between different institutions . The working 

relationships established between scientists at 

Oregon State University, the University o f 

Washington, and the U .S. Forest Service have 

been particularly outstanding. U .S. Forest 

Service scientists, especially those in the 

Pacific Northwest Forest and Range Experiment 

Station, have made a major contributio n 

to the success of the program . 

For organizational and budget reasons w e 

have concentrated our research effort at tw o 

major sites, the Lake Washington-Cedar Rive r 

system and the H. J. Andrews Experimenta l 

Forest. We must depend upon research at 

these areas to provide the initial theory an d 

models. We do recognize the broad researc h 

responsibility embodied in the term Coniferous 

Forest Biome and that we must expan d 

the base by carefully choosing additional site s 

for study and data accumulation in order to 

test models and theory . We also hope that 

researchers will use the principal research site s 

to validate models they may have develope d 

elsewhere . 

We are currently engaged in a study of th e 

entire Biome area for the purpose of selecting 

areas and projects which will extend Biom e 

research in an orderly and efficient manner . 

We are also developing inter-Biome research 

efforts which will enable us to exchange 

models and validation studies with all othe r 

Biomes . 

These remarks should serve to illustrate 

that the Coniferous Biome is in the early 

stages of a research program . Much of our effort 

so far has been related to organizing 

working groups, setting up general and specific 

research goals and then writing the necessary 

proposals . We hope the next 2 years can 

be devoted to the actual research and that in a 

similar meeting in 1974 we can demonstrate 

the success of the program . 



ported in part by National Science Foundatio n 

Grant No . GB-20963 to the Coniferous Fores t 




14



Findley Lake the study of a 

terrestrial-aquatic interface 

P. R . Olson and D . W. Col e 


an d 

R . R . Whitney 

College of Fisheries 



Abstract 

The linkage of the aquatic properties of a small lake to the terrestrial landscape is under examination in th e 

Findley Lake Basin of the Cedar River watershed. This pristine, 10 ha, oligotrophic lake is situated at 1,128 m 

elevation in a 162 ha watershed. During the initial year of investigation, this research program has focuse d 

primarily on a description of the ecosystem components. This initial phase of the program and the long-term 

objectives are discussed . 

Introduction 

Findley Lake is located in the upper portion 

of the Cedar River watershed in Kin g 

County, State of Washington . Cedar River 

watershed is the municipal watershed of th e 

City of Seattle and is one segment of the Lak e 

Washington watershed which is one of the intensive 

study sites of the Coniferous Fores t 

Biome, a portion of the United States Inter - 

national Biological Program (Gessel 1972) . 

A mission of the IBP is to demonstrat e 

how sound biological information can be 

synthesized and used in assessing alternatives . 

More specifically the Coniferous Fores t 

Biome addresses its attention to specific 

resource problems within the Biome . In general 

terms the Findley Lake interface study is 

directed to: better understanding of productivity 

and its maintenance on forest lands an d 

adjacent surface waters ; to determine the role 

of land and vegetation management in water 

quality and quantity ; and to measure the 

interaction of animal populations upon these 

areas . 

The City of Seattle Water Department ha s 

agreed to dedicate this entire drainage basin a s 

a research area for the IBP, Coniferous Forest 

Biome program. This cooperation includes 

securing the timber rights of the property and 

developing access into the area which at th e 

present time is by trail . The intent at Findley 

Basin is to maintain the pristine environmen t 

and minimize human influence while conducting 

the research . 

Watershed Description 

Findley Lake is located at 1,128 m elevation 

in a cirque with a maximum elevation of 

1,450 m and a total acreage of 162 ha . The 

aquatic portion of the Findley Lake basi n 

consists of Findley Lake proper with an are a 

of 10 ha and two smaller ponds of abou t 

0.4 ha each. Figure 1 defines the drainage 

basins of each of the ponds and lakes . Findley 

Lake proper intercepts 60 percent of th e 

watershed, and the one shallow pond intercepts 

14 percent. The lower, deep pond inter - 

15

Figure 1 . Aerial photo of Findley Lake basin indicating 

drainage basin of upper lake and two smalle r 

ponds. 

cepts each of the other two basins as well a s 

26 percent of the watershed directly . 

The soils of the watershed have bee n 

mapped by Bockheim and Ugolini (n .d .) . 

Many of the slopes range from 30 to 40 percent 

; widespread talus accounts for 16 .2 percent 

of the total basin. Soils of mixed materials 

are divided as forested, semiforested, an d 

nonforested ; they account for 56 .2, 4 .3, an d 

1.6 percent, respectively, of the total basin . 

The residual soils on the ridges are 17 .5 percent 

forested and 4 .2 percent nonforested . 

Del Moral (1972) has completed a vegetation 

survey of the watershed . Seven vegetation 

types were distinguished and their distributions 

mapped . Most of the area is timbere d 

by relatively homogenous old-growth Pacifi c 

silver fir (Abies amabilis) . 

Taber' has conducted a preliminary survey 

of the terrestrial vertebrate activity in the 

1 P. R. Olson, J . S . Bockheim, F . C. Ugolini, an d 

others . A terrestrial-lake interface program, Findle y 

Lake watershed. Coniferous Forest Biome Interna l 

Report No . 25 . (In press .) 

watershed. The elk (Germs canadensis) and 

pikas (Ochotona princeps) have a noticeable 

impact on the vegetation of the basin . Migratory 

,vertebrates such as deer (Odocoileus 

hemionus), elk, and birds could influence the 

movement of nutrients in or out of the basin . 

In addition, the elk activity in some of the 

meadow areas has obviously disturbed th e 

soil . 

Snow accumulations of 4 m often occur i n 

the Findley Lake basin . The open surface 

period on Findley Lake is approximately 4 

months. The lake frequently is not free of 

snow cover until late July and usually start s 

to close about mid or late November. A thin 

layer of ice will often form in early November 

and soon thereafter, snowfall will cover th e 

surface. The snowfall accumulates rapidly and 

insulates the surface from additional freezing . 

The latent heat of the lake water thaws th e 

ice layer originally formed and the snowpac k 

effectively supports itself . As snow accumulates, 

the light penetration is greatly reduced . 

In February of 1972 there were 2 .9 m (9 . 5 

feet) of snow ; the light readings beneath th e 

snowpack were approximately 1 .4 lux-0 .0 3 

percent of the surface illumination . Th e 

oxygen concentration of the water colum n 

was close to saturation except within 2 m o f 

the bottom where the oxygen decrease wa s 

quite abrupt. Temperatures at this tim e 

ranged from 1 0 C at the surface to 3° C a t 

25 m . 

Limnological surveys were performed b y 

Welch (see footnote 1) on a monthly interval 

during the open summer period . Findley Lake 

is extremely clear with visibility extending to 

depths of 15 m . Since the lake is relativel y 

deep (about 28 m) and unproductive, th e 

hypolimnion during the summer period remained 

well oxygenated . Maximum water 

temperature noted at the lake surface durin g 

the surveys was 18° C . The dominant zoo - 

plankton found was Diaptomus shoshon e 

with a modest population of Holopedium 

gibberum present early in the season. Th e 

Diap torn us population persisted after 3 

months of snow cover . 

The analysis of water samples b y 

Spyridakis (see footnote 1) is covered mor e 

extensively by Taub et aL(1972). Mean values 

16

of the most pertinent variables were : total 

phosphorus 5 µg/2, orthophosphorus 1 µg/2 , 

nitrate-nitrogen 3 pg/2, silica 76 pg/Q, calciu m 

1 mg/2, magnesium 0 .6 mg/2, sodium 1 mg/2 , 

and potassium 25 µg/2 . Chlorophyll determinations 

were 0.3 pg/2 and primary production 

was 370 mg C/m 2 /day . 

Other surveys which are relatively complet e 

at this time are the paleoecological analysis by 

Tsukada (see footnote 1) of numerous 1 m 

long sediment cores . These cores are bein g 

analyzed for concentration of pollen , 

diatoms, Cladocera, inorganic elements, an d 

organic chemicals including chlorophyll . Th e 

1 m cores were not sufficiently deep to give a 

complete record ; however, they did disclos e 

the fact that the surrounding environmen t 

had been disturbed at least twice in the recen t 

past. There was evidence of a fire with subsequent 

pollen change and, in a deeper portion , 

a 500 percent rise in sedimentary chlorophyll . 

Longer cores will be necessary to explain thi s 

phenomenon . Soil profiles disclosed evidence 

of two relatively recent fires as well as three 

volcanic ash depositions . 

A macrobenthic invertebrate survey wa s 

taken by Paulson (see footnote 1) of the lake , 

the ponds, and the interconnecting streams . 

May flies and caddis flies appear to be th e 

dominant consumers in the lake . May flies 

were represented by Baetidae and caddis flie s 

by two species of Limnephilidae . Midge larvae 

were present, two species of stone fly and als o 

Simuliidae in the creek . In addition a sphaeri d 

clam and a planorbid snail were found in fai r 

abundance . Three species of amphibians wer e 

abundant in the lake and pond : rough-skinne d 

newt (Taricha granulosa), western toad (Bufo 

boreas), and Cascades frog (Rana cascadae) . 

Less common were the northwestern salamander 

(Ambystonza gracile) and the Pacifi c 

treefrog (Hyla regilla) . All evidence to dat e 

indicates there are no fish present in the lake . 

Present and Future 

Program s 

Year one study on Findley Lake was primarily 

a survey year . The effort at present is 

directed at a continuation of the previou s 

research with a higher resolution to supply 

the necessary input to modelers . Studies will 

be intensified on nutrient availability and 

quantification of primary, secondary, an d 

tertiary production in both the terrestrial and 

aquatic environments . 

The objective of the proposed pedological 

study by Ugolini is to understand soil - 

weathering processes and their relations to th e 

b i o geochemical cycle. Gravitational water 

interconnects the atmospheric, biological , 

lithological, and hydrological components of 

the biosphere, bringing the reactants together 

and transporting the end products to ne w 

sites. The very essence of soil formation is th e 

migration of ions and particulate matter . Prior 

to establishing the kinds and rates of migration 

it is necessary to acquire a knowledge o f 

the soil system in both its chemical an d 

mineralogical composition . Thus estimations 

of nutrient capital in soils and forest floor o f 

the Findley Basin will be determined . 

The common denominator throughout the 

entire ecosystem is water . Water input as 

precipitation will be determined and followed 

through the aerial portion of the system . Thus 

crown wash and stemflow analyses, both 

qualitative and quantitative, are essential as 

well as determination of litterfall rates . The 

elements of prime concern will be initially N , 

Ca, Mg, K, and P . This aspect will be investigated 

by Cole and Gessel who will follo w 

water through the soil system collectin g 

leachates beneath the forest floor, at th e 

boundary between A and B horizons, beneat h 

the rooting zone in the C horizon and withi n 

the ground water table. Techniques used here 

will be similar to those previously reporte d 

for the Thompson Site (Cole and Gessel 

1968) . 

The surface water flow within the basi n 

will be surveyed and chemically analyzed b y 

Spyridakis. Many of the surface flows ar e 

temporary in nature associated with snow - 

melt; hence, the quantity of surface flow i n 

the upper system will not be critically deter - 

mined, but Wooldridge will quantify the 

hydrologic flow from the entire basin. Th e 

nutrient accumulation within the snowpack 

will also be determined . 

17

Nutrient levels in stream and lake water a s 

a function of seasonal variations and samplin g 

location will be determined by Spyridakis . 

Determinations will include temperature, conductivity, 

suspended and dissolved solids , 

light penetration, turbidity, color, dissolve d 

O 2 , pH, soluble and particulate C, chemical 

oxygen demand, dissolved SiO 2 , Na, K, Ca , 

Mg, Fe, Al, Mn, chloride, sulfate, bicarbonate , 

and carbonate . In addition, the lake sediment s 

will be characterized and their contribution t o 

the overlying waters will be determined . 

Welch will determine the annual and seasonal 

primary production in the lake an d 

relate it to the available nutrients . Special 

attention will be directed to regulation an d 

control dictated by the elements which may 

be limiting. Attention will be directed to th e 

phytoplankton cell size and their utilizatio n 

by zooplankton . The zooplankton composition 

and biomass will be determined on a 

regular sampling schedule. The predatio n 

impact upon zooplankton will probably b e 

quite minimal, because of the lack of fish i n 

the system ; however, observations will b e 

made to determine the potential consumptio n 

by insects and amphibians . A consideration 

for the fourth year of study may be the introduction 

of cutthroat trout into the lowes t 

pond in the lake basin to determine th e 

impact of predation upon the secondary production, 

after a comparison of the pond t o 

the lake for a period of time . 

The contrast existing between the tw o 

ponds and the lake presents an interestin g 

study by itself. Both ponds are 0.4 ha in area 

but have different depth profiles . The upper 

pond is only 1 .5 m deep and with the loss of 

the snowpack in midsummer will cease to 

overflow and decrease to 1 .0 m in depth . The 

lowest pond has not been accurately sounde d 

but appears to be at least 10 m deep and continually 

receives the overflow from the upper 

lake. The shallow pond opens earlier in th e 

summer, has a higher heat budget, and freezes 

earlier in the fall. This contrast will provide an 

interesting comparison in studies of decomposition, 

lake bottom water interface, an d 

growth and reproduction of amphibians an d 

invertebrates . 

One of the concerns in the aquatic area is 

the self-sustaining capabilities within that are a 

and the rapidity of recharge when the syste m 

is limited in some important essential . Questions 

that have been asked repeatedly in th e 

past deal with the role or importance of 

decomposition in the aquatic system. The 

Findley Lake basin presents a unique opportunity 

to measure the dynamics of decomposition 

when light is abundant and when ligh t 

is greatly reduced or eliminated for a 6 month 

period by snow cover. Findley Lake will be 

used primarily as a comparison to the othe r 

three lakes in the system as discussed by Tau b 

et al. (1972) . The flux and pool sizes of 

carbon in the water column will be deter - 

mined by Lighthart from a compartment 

analysis of (1) dissolved inorganic carbon , 

(2) phytoplankton, (3) zooplankton , 

(4) dissolved organic carbon, (5) detritus, and 

(6) heterotrophic bacteria using a C 14 technique. 

This technique requires the measurement 

of primary productivity as the initial 

step in measuring exchanges between compartments. 

The role of the aerobic, facultative 

anaerobes, and anaerobic bacteria will be 

determined by Matches . 

The oxidation of organic matter in th e 

water column to be done by Packard will be 

estimated by means of an oxygen uptak e 

method measuring enzyme activity (Packard 

1969). By this method he will compare th e 

seasonal changes in respiratory activity o f 

phytoplankton and zooplankton . Pamatmat 

will determine the annual deposition of 

organic matter on the lake basin and its rat e 

of decomposition . This will be determined by 

placing in the sediment a bell jar equippe d 

with oxygen probe, thermistor probe, and a 

stirring mechanism (Pamatmat and Bans e 

1969). The oxygen loss will be monitore d 

with regard to time so that consumption rate s 

may be determined. This will be related to th e 

oxygen budget of the overlying water column . 

Estimates of the contribution of detrital 

biomass to the aquatic food chain will b e 

determined by Taub . This will involve th e 

biomass estimate, and its organic carbon , 

organic nitrogen, and caloric contribution t o 

zooplankton, benthic invertebrates, an d 

amphibians . In addition, Taub will investigat e 

the biological aspects of nitrogen transforma - 

18

tions (Klucas 1969) . An overall nitrogen 

balance estimate will be determined from estimates 

of nitrogen fixation and denitrification . 

Attempts will also be made to determine th e 

transfer from detrital material such as chiti n 

and protein to the dissolved plant nutrien t 

ammonia . The interchange between the 

various dissolved states of nitrogen, ammonia , 

nitrate, and nitrite will also be investigated . 

Discussion 

The Findley Watershed relates to the overall 

objectives of the Coniferous Biome i n 

three specific ways : 

1. It provides an excellent opportunity to 

study linkage between terrestrial processes 

and the limnological properties of a lake 

system. The research is designed to assign 

values to the rates and quantities of elemental 

and organic exchange between these 

systems . 

2. Findley Lake represents an extreme i n 

oligotrophic conditions in lake systems and 

should provide a very interesting genera l 

lake model. It will serve as an excellen t 

comparison with the lower three lakes o f 

the Lake Washington drainage basin in 

studies of nutrient income vs . productivity 

relationships. Consequently, it is a critical 

component of the four lake program of th e 

Coniferous Biome . 

3. Because of the similarity in terrestrial community 

structure between Findley Lake 

and some of the reference stands at th e 

Biome's H. J: Andrews Intensive Sit e 

(Gessel 1972), it provides a linkage between 

the two biome study sites . Th e 

effort at the Andrews Site will concentrat e 

on unit watershed models, which will include 

transpiration, terrestrial primary production, 

and stream hydrologic models, al l 

of which should have excellent applicatio n 

in the Findley watershed . 

To study the interface phenomenon , 

Findley Lake has been diagrammatically stratified 

into a series of discrete ecosystem components 

connected by transfer functions . 

Figure 2 represents a first approximation of 

how these components are interconnected 

Figure 2 . Terrestrial-lake interface diagram with transfer 

functions numerically listed : (1) atmospheri c 

exchange, heat, light, water and gas; (2) nutrient 

assimilation; (3) nutrient leaching ; (4) hydrologica l 

output ; (5) consumption; (6) detrital contribution ; 

(7) decomposition ; (8) return from decomposition ; 

(9) movement across interface; (10) migration i n 

and out of watershed . 

and the nature of the linkage between th e 

terrestrial and aquatic systems . Each of the 

component boxes will necessitate a quantitative 

estimate of the biomass or effective 

reservoir involved . The pathways between 

components will not only involve a quantitative 

estimate but also a measure of rate o f 

movement between the components . As the 

research progresses, the level of resolution wil l 

become extremely critical before a total systems 

analysis is possible . A few important 

variables which will be kept in mind are : 

(1) spatial dimensions of the system under 

study, (2) time base of the study, and (3) th e 

nutrient elements and/or caloric equivalent s 

of the components . It is very likely that th e 

modeling effort will help prescribe a desirabl e 

time base . 

19



ported by National Science Foundation Grant 

No. GB-20963 to the Coniferous Fores t 

Biome, U .S. Analysis of Ecosystems, Inter - 

national Biological Program . This is Contribution 

No . 20 to the Coniferous Forest Biome , 

IBP . 

Literature Cited 

Bockheim, J . G., and F . Ugolini .[n .d.] Soil s 

and parent materials of Findley Lake , 

Snoqualmie National Forest, Washington . 

Northwest Sci . : (In press . ) 

Cole, D . W., and S. P . Gessel. 1968. Cedar 

River research-a program for studyin g 

pathways, rates and processes of elemental 

cycling in a forest ecosystem. 53 p . Forest 

Resour. Monogr ., Univ . Wash. Contrib . No . 

4. Seattle . 

Del Moral, R . 1972. Vegetation survey o f 

Findley Lake Basin . Am. Midland Nat . : (In 

press .) 

Gessel, S . P. 1972. Organization and researc h 

program of the Western Coniferous Fores t 

Biome . In Jerry F. Franklin, L . J . 

Dempster, and Richard H . Waring (eds .) , 

Proceedings-research on coniferous forest 

ecosystems-a symposium, p . 7-14, illus . 

Pac. Northwest Forest & Range Exp . Stn . , 

Portland, Oreg . 

Klucas, R. V. 1969. Nitrogen fixation assessment 

by acetylene reduction . Eutrophicat 

i o n -B i o s timulation Assessment Work - 

shop. Proc. 1969 : 109-116 . Univ. Calif . , 

Berkeley . 

Packard, T. T. 1969. The estimation of th e 

oxygen utilization rate in sea water fro m 

the activity of the respiratory electro n 

transport system in plankton . 115 p . Ph .D . 

thesis on file, Univ . Wash . 

Pamatmat, M. M., and K. Banse . 1969 . 

Oxygen consumption by the seabed . II. In 

situ measurements to 180 m depth . 

Limnol. Oceanogr. 14 : 250-259 . 

Taub, Frieda B., Robert L . Burgner, Eugene 

B. Welch, and Demetrios E . Spyridakis . 

1972. A comparative study of four lakes . 

In Jerry F . Franklin, L . J. Dempster, an d 

Richard H . Waring (eds .), Proceedingsresearch 

on coniferous forest ecosystems- a 

symposium, p . 21-32, illus. Pac. Northwest 

Forest & Range Exp . Stn ., Portland, Oreg . 

20



A comparative study 

of four lakes 

Frieda B . Taub , 

Robert L . Burgner, 

Eugene B . Welch, 

an d 

Demetrios E . Spyridakis 



Abstract 

It is increasingly evident from many unfortunate case histories that lake communities are sensitive in a 

variety of inadequately understood ways to watershed management practices and abuses . If we are to conserv e 

or manage our array of lakes of the Coniferous Biome successfully, we must gain a sufficient understanding t o 

predict the impact of various perturbations on the general lake community structure and production . Initial 

effort in the Coniferous Biome is being directed toward an intensive comparative study of an array of four lake s 

in a single watershed that provide a spectrum of human disturbance and manipulation . These are lakes 

Washington, Sammamish, Chester Morse, and Findley in the Lake Washington-Cedar River Drainage. A study of 

Cedar River, the main drainage system and a link between two of the lakes, contributes to both the lakes' stud y 

and the smaller stream investigations at Andrews Forest . 

The Place of Lake Studies 

in the Biome 

It may well be asked why there should be 

an intensive study of lakes within a CONIF- 

EROUS BIOME . Certainly, the biome is 

named for the most conspicuous organisms , 

the terrestrial conifers . However, it can be 

shown that aquatic studies and the terrestrial 

studies each have much to gain by being 

combined . 

Unlike the European IBP, the US IBP chose 

to combine the initial working groups for 

Freshwater Productivity and Terrestrial Productivity. 

It was anticipated that this approach 

would best encourage truly biome - 

wide studies, not artificially divided b y 

discipline boundaries, such as the Findle y 

Lake study which is discussed in the previou s 

paper. This Biome also includes studies of th e 

marine contribution to freshwater and terrestrial 

communities by anadromous fish . 

Limnologists have long recognized that th e 

terrestrial influences on lakes are of utmos t 

importance. The physical characteristics of a 

lake are determined largely by its topology 

and climate which will determine the degree 

and frequency of mixing. It had long been 

thought that lakes progressed from oligotrophic 

to eutrophic as an inevitable 

sequence . Margalef (1968) and Odum (1969 ) 

have suggested that lakes will progress towar d 

increased oligotrophy if the surrounding lan d 

does not contribute nutrients . Thus the majo r 

successional pattern of a lake may be determined 

by the terrestrial input . 

The nutrient inputs to a lake are very largely 

determined by the nature of the soil, the 

land use, and the amounts and patterns of 

water flow within the drainage basin . Since 

the major nutrient inputs may be made by the 

streams, rather than by shore drainage, events 

quite some distance from the lake can exert 

major effects. The Biome study provides 

greater opportunity for evaluating terrestria l 

21

inputs into the lakes . Recent studies have 

shown this kind of information to be necessary 

in considering effects of land use . 

The terrestrial scientists should not be unmindful 

of the effect that aquatic bodie s 

exert on the surrounding land mass . Man 's 

needs for water have resulted in virtually all 

civilizations developing in proximity to rive r 

drainage basins . Locally, the impact of water 

on man's land use is also obvious . Seattle 

developed into a major city because of it s 

location and port facilities on Puget Sound . 

Lake Washington has had an impact o n 

Seattle by preventing growth eastward in a 

contiguous fashion . A great desire for water - 

front property has prompted the developmen t 

of homes and parks all along its shore front , 

and therefore brought about the necessity o f 

bridges and major road systems to permi t 

residents of the east shore to have convenien t 

access to Seattle . Similarly, the existence o f 

Lake Sammamish and its recreational an d 

esthetic assets have created the developmen t 

of parks, communities, and therefore, roads , 

along its shore . Because Chester Morse Lak e 

serves as a water supply for Seattle, the entir e 

watershed of this lake has been fenced off an d 

been inaccessible to the public for 60 years . 

Land use has been restricted to controlle d 

logging designed so as to protect wate r 

quality. Within the Chester Morse watershed , 

Findley Lake has been protected incidentally . 

Aquatic bodies also influence terrestrial 

communities by the frequency and extent o f 

floods. The temptation to build on level land 

has resulted in construction on flood plain s 

and the subsequent need for dams and othe r 

flood control measures . These in turn have 

altered the terrestrial environment and th e 

potential uses of the land . 

The distribution of terrestrial wildlife is 

also influenced to a very great extent by th e 

availability of the water supply . A host of 

water birds collect their food from the water 

but otherwise nest and conduct the nonfeeding 

parts of their lives on the land . In addition, 

there are animals such as otters (Lutra 

canadensis), racoons (Procyon lotor), and 

bears (Ursus spp.), which take a rather substantial 

amount of their food from aquatic 

bodies, especially after the salmon spawning . 

Over geological timespans, water has a 

major impact on the land . The streams an d 

rivers are major sculptors of the terrestrial 

topology. Lakes are not merely a sink for up - 

land nutrients but may become the source o f 

future land . Much of our prime agricultura l 

land is lake fill . The occurrence of peat in th e 

local drainage system gives further evidenc e 

that much of our present land was previousl y 

occupied by an aquatic community . 

Characteristics of the 

Four Lakes 

The four lakes in the Coniferous Forest 

Biome 's Cedar River intensive site-Washington, 

Sammamish, Chester Morse, and 

Findley-provide contrasting conditions (fig . 

1) . The physical parameters of the lakes are 

shown in table 1 . 

1. Lake Washington (fig . 2), the lowermost 

lake, has a documented history of eutrophication 

and recent sewage diversion . 

The dominant fish is the anadromous 

sockeye salmon (Oncorhynchus nerka) 

which was introduced about 30 years 

ago . Other anadromous and resident fis h 

populations also occur here. The lake is 

used intensely for recreation, including 

sports fishing . The shores of the lake are 

largely urbanized . 

2. Lake Sammamish (fig . 3) represents an 

intermediate condition, as it was oligotrophic 

until very recently, underwent 

limited eutrophication as the forest wa s 

replaced by an increasing urban community, 

and is currently undergoing 

sewage diversion. The response of this 

lake to sewage diversion has been substantially 

less dramatic than that of Lake 

Washington . Surprisingly, the hypolimnion 

is anerobic much of the summer . 

Some of the same fish species, includin g 

sockeye salmon, occur here as in Lak e 

Washington, but in smaller numbers . 

This lake drains into Lake Washington 

from the north via Sammamish Slough . 

3. Chester Morse Lake (fig . 4) is the upper - 

most of the two large lakes in the Ceda r 

River watershed . Although the lake level 

22

Figure 1 . Map of Lake Washington drainage. 

Table 1.--Physical characteristics of lakes in Lake Washington drainage basi n 

Lake 

Maximum 

depth 

Mean 

depth 

Water 

Area Volume Length Elevation exchange 

tim e 

m m km 2 km 3 km m years 

Findley 30' 10 1 0.09 0.0009' 1,13 1 

Chester Morse 35 13 .0 6.54 .085 8 .1 473 0 . 3 

Sammamish 31 17 .7 19 .8 .350 12 .9 12 2. 2 

Washington 64 33.0 87 .6 2 .884 31 .5 8 .6 3 .0 

' Tentative values . 

2 3

was slightly raised in 1902, the lake in 

other respects is the least disturbe d 

major lake in the state . Access to th e 

lake has been rigidly controlled sinc e 

1911 and no fish have been planted o r 

harvested since that time. It contains no 

anadromous fish, but has resident populations 

of rainbow (Salmo gairdneri) , 

Dolly Varden (Salvelinus malma), and 

whitefish (Prosopium sp .) . The shore is 

completely undeveloped . 

4. Findley Lake (fig . 5) is among several 

small lakes in a totally unmanipulate d 

watershed . It and its watershed are being 

used in an integrated terrestrial-aquati c 

study . The land area consists of undisturbed 

mature coniferous forest. Th e 

lake has no resident fish population bu t 

has zooplankton and salamanders i n 

moderate abundance . 

Specific tasks of the lake-stream study in 

the Lake Washington-Cedar River watershe d 

include assessment of (1) terrestrial influences, 

including forest and urban land use , 

(2) nutrient budgets, (3) sources of energy t o 

support the food chain, (4) community structure 

and metabolism, (5) effects of food avail - 

ability on successive trophic levels, (6) effects 

of fish on lower trophic levels, (7) effects o f 

anadromous fish on other fish populations , 

and (8) outputs to the marine community b y 

anadromous fish . 

Research Program 

In 1971, fieldwork was initiated on chemical 

budgets, primary and secondary productivity, 

and fish populations, in coordinatio n 

with other concurrent studies. A literature 

compilation of past aquatic studies in th e 

watershed was completed and the transfer o f 

pertinent data to the IBP data bank initiated . 

A major hydrological study of the lake-river 

system was initiated by another agency 

(R I B C O, River Basin Coordinating Committee) 

. Decomposer studies were organized 

in conjunction with the terrestrial program. A 

preliminary interface study on Findley Lake 

watershed was accomplished . A generalized 

trophic level model was developed and its 

shortcomings were analyzed . Modeling efforts 

on existing data from other lakes wer e 

initiated by the modeling group . 

In 1972 the fieldwork mentioned above ha s 

been expanded and fieldwork on decomposition 

and nutrient regeneration processes has 

been initiated . Field effort and modeling 

structures are reorganized into subsystems o f 

bottom related processes, water colum n 

processes, and higher consumer dynamics . 

Models of optimum strategies and density 

dependent relationships are being developed . 

Detailed descriptive models of the lakes wil l 

be completed . Data sets at potential extrapolation 

sites are being identified and a 

coordination framework was established . 

In 1973 the full field program will continu e 

with phasing and evaluation of specific 

studies, including elimination of those whic h 

have either met their objectives or which are 

not profitable lines of pursuit . Greater 

emphasis will be placed on using hydrologica l 

data and interfacing with terrestrial aspects of 

the program as well as on perfecting sub - 

system models and linking them together int o 

a more complete model . Coordination with 

extrapolation site programs will be increased , 

including data set analysis . 

In 1974, successful and fruitful program s 

will be continued as necessary, with phasin g 

of research in many instances to accomplish 

tasks deferred for orderly sequential progression 

under limited funding . Major efforts wil l 

be directed to integration of all work and to 

use of extrapolation data for model validatio n 

and generalization of intensive site results . 

Focus of Research Program 

The stream-lake study will address a number 

of questions) which have relevance to 

both science and analysis of environmenta l 

impacts. These questions are concerned with : 

(1) the mechanisms by which purely physica l 

constraints, such as structure of lake basin, 

These questions and the subsystem models have 

been developed with the help of the modeling group . 

Mr. Larry Male of the Center for Quantitative 

Science, University of Washington, is particularl y 

acknowledged for his contributions to these ideas . 

24

OUTLE T 

PINS 

LANE 

LYIMIIMP 

MICOGS 

LANE 

LAKE IAMMAMIS H 

DEPTH IN EETEA 1 

LAKE WASHINGTO N 

DEPTH IN METERS 

o 2 .. 

0 I E km Figure 3 . Map of Lake Sammamish . 

Ip►OIA N 

ONEER 

Figure 2 . Map of Lake Washington . 

Figure 4. Map of Chester Morse Lake. Figure 5 . Map of Findley Lake . 

25

climate, surrounding terrestrial environment , 

prevailing winds, lake circulation dynamics , 

rates of water and inorganic nutrient inflow 

and outflow, and chemical precipitation of 

essential nutrients control the productivity , 

water quality, and community structure of 

aquatic ecosystems ; (2) the dynamics of nutrient 

and energy cycling through both biological 

and physical processes ; (3) the 

mechanisms by which community structure 

may inhibit or enhance overall productivity 

and water quality ; (4) the response of complex 

aquatic ecosystems to a variety of 

natural and artificially created disturbances ; 

and (5) the mechanisms these systems have 

evolved for coping with these disturbances . 

The nature of these questions justifies th e 

coordinated program which is planned. Some 

of the questions this program will bear upo n 

and how they intend to be tackled will now 

be discussed . 

The phytoplankton and bacterial communities 

in lakes not only provide the basis fo r 

complex food webs, they essentially contro l 

the overall water quality and the apparent 

state of eutrophication . The sustained abundance 

and species composition of the bacterial 

and phytoplankton communities depend 

to a large extent upon the rate at whic h 

they are supplied with essential inorganic 

nutrients and energy sources. Most of these 

compounds are introduced from the surrounding 

terrestrial environment and the 

atmosphere through biological fixatio n 

processes . The rate at which a lake is supplied 

with allochthonous materials may determin e 

its productivity structure . Several mechanisms 

may be responsible for determining a lake' s 

response to allochthonous nutrient input . 

These mechanisms relate both to the suppl y 

of nutrients to the phytoplankton and bacteria 

communities and to the physical an d 

biological forces which influence the production 

dynamics of these communities . 

1 . Dynamics of release of inorganic nutrients 

from allochthonous organic material 

of varying nutrient richness : The first indication 

is that nutrients in nutrient 

poor organic material, e .g., wood, are released 

slowly. Since the four lakes in thi s 

study can be characterized by the nature 

and extent of allochthonous input, thi s 

mechanism can be studied . 

2. Nutrient availability as a function o f 

thermal distribution, circulation dynamics, 

and cycling dynamics of other chemical 

compounds : Vertical density gradients 

(determined by vertical therma l 

distribution gradients) can control th e 

availability of essential nutrients to autotrophic 

organisms . The density gradient s 

function as a barrier to the mixing o f 

nutrients from deeper waters into th e 

productive surface waters . Since the 

density barrier also inhibits the passag e 

of oxygen from the surface into th e 

deeper waters the deeper waters some - 

times become depleted of oxygen . Thi s 

anerobic condition enhances the solubility 

of compounds which bind phosphorus 

in forms unavailable to phytoplankton 

. Thus, the deep waters may be 

rich in nutrients which are relatively unavailable 

to autotrophic organisms . Th e 

four lakes in this study vary greatly i n 

the formation of their density gradient s 

and thus their capacity to trap nutrients . 

This theory is only partially applicabl e 

to any particular lake since there exis t 

other mechanisms for distributing nutrients. 

The shape of the lake basin, direction 

and strength of prevailing winds , 

internal circulation currents, disturbanc e 

of the bottom sediments by mechanica l 

mixing, and dynamics of nutrient trans - 

port through the reduced layer of th e 

sediment are all mechanisms which ar e 

being studied to bear upon the proble m 

of nutrient cycling . 

3.Biological cycling of nutrients : Since the 

communities of phytoplankton and bacteria 

in lakes are often limited by the 

availability of essential nutrients they 

have tended to evolve elaborate mechanisms 

for the conservation and recyclin g 

of nutrients. In the absence of these 

mechanisms, production would mos t 

likely be nil. Nutrients cycle through 

communities of phytoplankton, zoo - 

plankton, bacteria, protozoans, and 

littoral aquatic plants. The nature of 

these cycles and how they operate with - 

26

in the restrictions of the physical environment 

are being tackled . Since the 

four lakes in this study have evolved to 

their present structure in the presence of 

different nutrient stresses we might 

expect them to exhibit various capacitie s 

to conserve and cycle nutrients . 

4. Effect of community structure on productivity 

and distribution of energy : Th e 

production of herbivores (zooplankton ) 

cannot be explained entirely by production 

of primary producers (phytoplankton). 

The physical form of the production 

(i.e., size distribution or colonia l 

form of algal cells) may preclude consumption 

by zooplankton . However , 

zooplankton may in fact contribute to 

changes in the physical form of primar y 

production by being size-selective 

feeders. Since changes in the form an d 

species composition of phytoplankto n 

communities are indices of pollution an d 

eutrophying environments this process i s 

being investigated . 

In the same way that changes in th e 

physical form of phytoplankton production 

may control the production of zoo - 

plankton, the community structure, siz e 

distribution, and spatial distribution o f 

zooplankton control the feeding behavior 

and subsequent production o f 

predators . 

The terminal link in the food web of three 

of the study lakes is the community of fis h 

species . Much literature has been devoted to 

models which try to explain the populatio n 

dynamics of these organisms without incorporating 

their interaction with the environment 

. The communities of fish can influenc e 

the nature of production by altering th e 

community structure of other vertebrate an d 

invertebrate communities upon which the y 

feed and indeed may alter the entire community 

(Hurlbert et al . 1972). Their own production 

dynamics are related to the structure 

and abundance of the prey communities an d 

also to mortality dynamics which are deter - 

mined by the state of the physical environment 

at critical life stages . Three projects wil l 

bear upon the questions related to the fis h 

community . 

Presently, a three submodel system (figs . 6 

and 7) is being used to develop the propose d 

field studies ; these are organized around water 

column, bottom related, and higher consumer 

processes . These subsystems were chose n 

because the couplings within each are mor e 

tightly linked and the span-of-time resolution s 

OVERALL SYSTE M 

Figure 6 . Overall system categories. 

LAKE SUBSYSTEM (C) 

NIGHE R 

CONSUME R 

DYNAMIC S 

(C? , 

Figure 7. Lake subsystem categories. 

27

more similar than between the different subsystems 

. For example, the processes in the 

water column involving exchanges betwee n 

nutrients, algae, shortlived zooplankton, and 

bacteria can better be handled by a singl e 

team than by several teams, each interested i n 

a specific trophic level. Further, this information 

may have to be handled on a daily o r 

hourly basis, whereas the higher consumer 

production, e .g., fish, will be handled on a 

seasonal basis . 

Studies Planned for 1973-74 

The individual studies tentatively planne d 

for 1973 and 1974 for the four lakes are 

shown in table 2 . These are described briefl y 

here. The interface study specific to th e 

Findley Lake watershed is described in th e 

previous paper . 

Water Column Process Studie s 

The categories and relationships are show n 

in figure 8. Much of the past work in th e 

Cedar River drainage lakes defined the annua l 

nutrient supply to Lake Sammamish an d 

Figure 8 . Categories and relationships of sub-syste m 

(C)a ; water column processes . 

documented its rate of recovery compared t o 

that of Lake Washington as a result of nutrient 

diversion. Recently, limnological conditions 

have been monitored in Chester Mors e 

and Findley Lakes to compare their trophi c 

character with that of Lakes Sammamish an d 

Washington . Ultimate objective is to develop a 

model with enough generality to encompas s 

the range in trophic character now observabl e 

among the lakes . 

Lake Washington responded rapidly to a 

diversion of over one-half its annual supply o f 

phosphorous . Edmondson (1970) has documented 

its recovery to a trophic state simila r 

to that recorded over 20 years ago in a matte r 

of only 3 years after the completion and 7 

years after the beginning of sewage diversion . 

Lake Sammamish, a mesotrophic lake wit h 

similar flushing time, located only 10 miles t o 

the east of Lake Washington, has no t 

responded noticeably in 3 years following 

nutrient diversion . Possible reasons for this 

difference in rate of response involve th e 

morphometry of the two lakes . 

The nutrient input to the lake and the 

levels of dissolved inorganic nutrients will b e 

related to the growth and production of 

phytoplankton and its consumption by zoo - 

plankton . These studies will be in sufficient 

detail to examine changes in species composition 

and the importance of size categories a s 

they may relate to optimum feeding strategies. 

More detailed, but less frequent estimates 

of the flow of carbon through all 

components of the food web will be made by 

the techniques of Saunders (1969) usin g 

radioactive bicarbonate, radioactive detritus , 

and radioactive dissolved organics . Both of 

these field-oriented studies will be correlate d 

with the modeling on the dynamics of nutrient 

distribution and flow through aquatic ecosystems. 

A deferred study concerns the rol e 

of fungal parasites on algae to explore the 

hypothesis that algal blooms may cease 

through disease processes, rather than throug h 

the limitation of nutrients . 

The mathematical models which have been 

developed to explain the production dynamic s 

of phytoplankton production admit that the 

concentration of phytoplankton may vary continuously 

from the surface to the lake bottom . 

28

Table 2.-Proposed lake studies (by investigator and abbreviated title ) 

Aquatic Director: R. L . Burgner 

Lake Studies : F. B. Taub, Chairman 

Department University Study 

Water Colum n 

(E. B. Welch, Chairman ) 

1. E. B. Welch Civil Engineering Univ. Washington 

P. R. Olson Forest Resources Univ. Washington 

Zooplankton-productio n 

2 . E. B . Welch Civil Engineering Univ. Washington Phytoplankton-production 

3. B. Lighthart Inst. Freshwater Studies Western Washington Carbon web 

4. B. Lighthart Inst. Freshwater Studies Western Washington Bacteria 

5. T. Packard Oceanography Univ. Washington Oxidation-respiratio n 

6. F . B. Taub Fisheries Univ. Washington 

J. T. Staley Microbiology Univ. Washington 

Nitrogen transformation 

7.*D. E. Spyridakis Civil Engineering Univ. Washington 

R. F. Christman Civil Engineering Univ. Washington 

Nutrient budgets 

Bottom-Related Processes 

(D. E. Spyridakis, Chairman) 

8. D. E. Spyridakis Civil Engineering Univ. Washington 

R. F. Christman Civil Engineering Univ. Washington 

Nutrient budget s 

9. F . B. Taub Fisheries Univ. Washington Detrital inpu t 

10. J. R. Matches Fisheries Univ. Washington Bacterial decomposition 

11. M. M. Pamatmat Oceanography Univ. Washington Oxidatio n 

Higher Consumers 

(R. R. Whitney, Chairman ) 

12. R. L. Burgner Fisheries Univ. Washington 

Limnetic fis h 

R. E. Thorne Fisheries Univ. Washington 

13. A. C. DeLacy Fisheries Univ. Washington Limnetic fish feed 

14. R. S. Wydoski Fisheries Univ. Washington 

Benthic-littoral fis h 

R. R. Whitney Fisheries Univ. Washington 

*See also Bottom-Related Processes. 

29

The net rate of primary production at any 

particular depth depends upon the temperature, 

light intensity, essential nutrient concentration, 

and zooplankton grazing pressure a t 

that depth. The vertical distribution an d 

abundance of phytoplankton is thus a functio n 

of varying net production, sinking rates of algae 

and mixing dynamics of the water column . Th e 

total production per unit area of lake surfac e 

may be obtained by integrating the concentration 

of phytoplankton from top to bottom . 

The next step is to incorporate the size selective 

feeding of zooplankton and determine it s 

effect upon production (since algal cells o f 

different sizes possess varying capacities t o 

absorb nutrients) . 

The effective grazing rate by zooplankto n 

is modeled as a function of the phytoplankton 

density (at high densities the zooplanktons 

feeding structures become saturated) . 

The efficiency of assimilation of phytoplankton 

by zooplankton is also allowed to depen d 

upon algae abundance (efficiency decrease s 

with increasing abundance) . 

The overall model relating the phytoplankton, 

zooplankton, and nutrients is a system o f 

partial differential equations (in time and 

depth) . 

The submodel describing the physical 

cycling of nutrients (developed in the botto m 

related processes group) is an integral part of 

the phytoplankton, zooplankton submodels . 

The number and kinds of heterotrophi c 

bacteria will be assessed to give some information 

on the constancy of community structure 

of the decomposers . The utilization an d 

ultimate destruction of fixed organic material 

for the entire system will be estimated by 

respiration studies and verified by substrat e 

disappearance . The respiration an d 

heterotroph abundance values will also be correlated 

with the regeneration of nutrients . 

The nutrient budget and effective concentrations 

will be assessed . All of the above studie s 

will be carried out in units of biomass or wil l 

be convertible to biomass and units of carbo n 

by calculation of size distribution. Determinations 

of the magnitude of nitrogen transformation 

within each of the above compartments, 

as well as the search for evidence nitrogen 

fixation and its opposite process, denitri - 

fication, will be made . 

The detailed carbon cycle through th e 

phytoplankton, zooplankton, bacteria and 

pools of dissolved inorganic carbon, dissolved 

organic carbon, and detritus are bein g 

modeled initially with a varying rate compartment 

system. The rates of flow between th e 

compartments are to be functions of light , 

temperature, and P0 4 and NO 3 concentrations. 

Later the rates will also be modeled a s 

functions of 0 2 and Si concentration. Thi s 

model will allow an analysis of what paths o f 

the cycling process limit the rate of 

production . 

Bottom Related Process Studie s 

The compartments are shown in figure 9 . 

The necessary chemical analyses to assess thi s 

aspect of the nutrient budget will also be 

done. Inputs of detrital materials to the sediments 

and their potential incorporation int o 

higher food chains via bottom invertebrates t o 

the fish will be assessed . The oxidation of 

organic material will be assessed by respiration 

rates of the sediments. The disappearanc e 

of substrates such as cellulose and chitin an d 

information on the bacteria associated wit h 


(C)b ; bottom related processes. 

30

these processes will be measured and correlated 

with nutrient regeneration and respiration 

rates. The production of littoral material s 

will be assessed and compared with limneti c 

production and terrestrial inputs . The combined 

information on limnetic, benthic, an d 

littoral photosynthesis and respiration wil l 

permit comparisons of P/R ratios for th e 

various lakes within this study as well as fo r 

wider comparisons . 

A sophisticated theoretical model describing 

the complex physical and biological cycle s 

of inorganic nutrients has been proposed . Thi s 

model incorporates the features of densit y 

gradients as a barrier to mixing, circulatio n 

currents, chemical precipitation of essentia l 

nutrients, transport and release of nutrient s 

from the bottom sediments as a function o f 

0 2 concentration, pH, and bacterial decomposition, 

nutrient input in organic and in - 

organic form from allochthonous sources, an d 

atmospheric input through biological fixatio n 

processes . This theoretical model will be mad e 

more precise and useful by formulating it as a 

collection of well-defined assumptions an d 

mathematical equations. This model wil l 

allow an analytic comparison of the cyclin g 

processes in the four lakes . 

Higher Consumer Process Studie s 

Compartments are shown in figure 10 . Th e 

measurements of population parameters o f 

young sockeye and other limnetic feeding fis h 

will provide comparative data on seasona l 

changes in numbers, biomass, growth, an d 

mortality rates relative to possible controllin g 

factors, including recruitment, physical an d 

chemical environment, food availability an d 

characteristics, behavior and competition , 

predation and other removal . Characteristics , 

habits, and interactions of the benthic an d 

littoral fish will be studied . Study of th e 

benthic food supply will be temporarily deferred. 

The interaction between the benthi c 

and littoral fish and the limnetic feeding fis h 

is receiving emphasis . In all cases, there will b e 

a search for understanding of feeding strategies 

and growth dynamics as they influenc e 

community structure of the zooplankton an d 

fish . 


(C)c; higher consumer dynamics . 

The vast food webs, typical of lake ecosystems, 

seem to be so complicated as to defy 

description . The higher consumers not only 

respond physiologically to their surroundin g 

physical environment, they exhibit complex 

behavioral mechanisms to cope with changin g 

food supplies and varying physical conditions . 

A conceivable model for this system is a varying 

rate compartmental model where the rate s 

of transfer among various species depend o n 

the physical environment . Another approach 

which is finding favor among theoretical 

ecologists is the theory of optimal strategies . 

The essences of this theory is that selective 

processes and adaptive mechanisms tend t o 

produce populations of organisms which 

strive to maximize their energy intake subjec t 

to the physical restrictions imposed by the 

environment and their own physiological an d 

morphological limitations . If the community 

structure of the prey organisms available to a 

predator is known then a mathematical representation 

of an optimal strategy model wil l 

predict the amounts and kinds of prey a 

predator will ingest. Coupled with models 

which describe the mortality structure of 

populations the optimal strategy model will 

effectively handle the questions concernin g 

31

community structure and energy flo w 

dynamics . 

It is obvious the subsystems above are artificial 

and that coordination of techniques an d 

measurements is necessary. For example, th e 

nutrient budget of the lake cannot be developed 

without the water column and th e 

bottom processes information, nor without 

the hydrological studies being coordinated 

with RIBCO and assessment of nutrien t 

sources . 

The present (1972) studies are necessarily 

descriptive, to supply the material for the firs t 

rough models . As the inadequacies of the 

crude models become apparent during the 

course of 1972, there will be increasing inter - 

play between the modeling and the field 

investigations and programs correspondingl y 

modified . 



ported in part by National Science Foundation 

Grant No . GB-20963 to the Coniferous Fores t 





Edmondson, W. T. 1970. Phosphorous nitrogen 

and algae in Lake Washington after 

diversion of sewage . Science 169 : 690-691 . 

Hurlbert, Stuart H., Joy Zedler, and Deborah 

Fairbanks . 1972. Ecosystem alteration b y 

mosquitofish (Gambusia affinis) predation . 

Science 175: 639-641 . 

Margalef, Ramon. 1968. Perspectives i n 

ecological theory. 111 p., illus. Chicago , 

London: Univ . Chicago Press . 

Odum, Eugene P . 1969. The strategy of ecosystem 

development . Science 164 : 

262-270 . 

Saunders, George W . 1969. Some aspects of 

feeding in zooplankton . In Eutrophication , 

p. 556-573. Nat. Acad. Sci., Washington, 

D.C . 

32



The modeling process relating to 

questions about coniferous 

lake ecosystems 

A bstract 

Douglas M . Eggers 

Fisheries Research Institute 


an d 

Larry M . Mal e 

Center for Quantitative Scienc e 


The role of sound conceptualization and meaningful questions in the modeling process is discussed . Th e 

salient features of lake communities are reviewed. Included are factors which must be considered in answering 

questions involving the higher consumers of lake ecosystems . 

Before attempting an analytical model on e 

must conceptualize the system being modeled . 

If one 's conceptualization is good, then one is 

able to ask relevant and meaningful questions . 

Since ecosystems are infinitely complex, on e 

must make assumptions and simplifications in 

constructing analytical models which are with - 

in our capabilities . Because of these assumptions 

and simplifications it is impossible t o 

construct general ecological models-a general 

model being one which will answer or anticipate 

all questions. Therefore models must be 

constructed relevant to specific questions . 

A set of meaningful questions should be th e 

foremost component in the strategy which a 

worker adopts in the construction of a model . 

These questions allow the modeler to make set s 

of assumptions and simplifications which mak e 

a model both meaningful and workable . 

Considering a coniferous lake ecosystem, if 

one is interested in questions of seasonal succession 

of phytoplankton or perhaps, more 

long-term successional changes along a gradient 

lake nutrient level, then one must construct 

models which differentiate the algae with 

respect to species . This model must incorporate 

algal adaptive mechanisms to critical resources . 

On the other hand, if the question is one of 

primary production, then it may be worthwhile 

to construct models where the algae are lumpe d 

together into a uniform mass of quasi - 

organisms . This is a fundamental assumption of 

the Riley et al. (1949) model . But without 

species differentiation a means of handling the 

interesting questions in ecology of competition, 

predation, selection, and adoption i s 

destroyed . 

The following are some of the factors w e 

have considered in developing our conceptua l 

framework of a coniferous lake ecosystem . 

The structure and dynamics of the salient 

features of aquatic communities are reviewed . 

Attention is given to processes which hav e 

traditionally been neglected in quantitative 

models . 

Generation times in the algae are short . 

Thus the selection process may be rapid . This 

can give rise to rapid shifts in the species composition 

of phytoplankton as the environment 

changes (Hutchinson 1967) . Because of seasonal, 

diurnal, temporal and other spatial 

patterns of environmental change, selectio n 

processes are in a continual state of flux . The 

patchiness in spatial distribution of the organisms 

plus interruption in selection processes 

by continually changing environmental 

conditions, allows many species to coexist i n 

the lake ecosystem . The state of being th e 

33

est adapted organism vacillates betwee n 

species with time as the environment changes . 

The level of nutrients in a lake ultimatel y 

determines the level of phytoplankton whic h 

the lake can support . Nutrients are utilized b y 

plants as components of structural molecules , 

such as DNA, RNA and enzymes . If nutrien t 

levels are in excess of structural demands an d 

other conditions are favorable for photosynthesis, 

then reproduction can occur . Nutrient s 

are rare elements in lake ecosystems, very 

eutrophicated lakes being exceptions . Aquati c 

communities have evolved many mechanisms 

fOr their conservation (Pomeroy 1970) . Algae 

and littoral plants can absorb nutrients almos t 

instantaneously from the water . Aquatic plants 

can take up nutrients as they are available and 

store them until conditions become more suit - 

able for growth when the nutrients are utilized . 

Light and temperature conditions necessary for 

photosynthesis are much more predictable , 

with definite seasonal and diurnal patterns , 

than nutrient levels, which the algae cells en - 

counter more or less randomly . Then, from the 

viewpoint of algae, a good strategy to evolve 

would be a means of acquiring nutrient s 

whenever they occur and wa iting to make use 

of them until sufficient light and temperature 

conditions (these are predictable) prevail . 

In addition to being able to utilize nutrient s 

as they occur, the phytoplankton-zooplankto n 

community recycles nutrients rapidly . As th e 

algae die and sink to the bottom, nutrients are 

released in two ways . Autolytical release occurs 

by simple diffusion. Mechanical release occurs 

when the cell membrane ruptures . Betwee n 

25 and 70 percent of the nutrients containe d 

in a sinking, dead organism are released i n 

these ways (Johannes 1968) . 

Zooplankton eat substantial proportions o f 

the algae . Herbivores are large in relationship 

to their food supply . Since zooplankton are 

eating a rich source of essential nutrients , 

they consume many more nutrients than they 

need for structural components. The excess is 

simply excreted into the water . 

The classical role of bacteria, as the principal 

agents of nutrient regeneration is questioned 

by Johannes (1968) and Pomeroy 

(1970). Bacteria accumulate nutrients ove r 

the level required in growth, as algae do . 

Johannes (1968) believes that protozoa, ofte n 

neglected as important organisms in lak e 

systems, are responsible for regeneratin g 

nutrients by eating bacteria and excreting th e 

excess nutrients . 

In addition to biological means of nutrient 

cycling, many purely physical processes affec t 

nutrient distribution in lake systems . Nutrients 

enter the lake through the inflow (both surface 

flow and seepage) and precipitation . Nutrients 

leave the lake through outflow (both surface 

flow and seepage) . Sediments are very important 

in the nutrient budgets of lakes . They may 

act as a trap for essential nutrients, such a s 

phosphorous. The phosphorous is tied up b y 

iron and precipitated out at high redox potentials. 

Decreasing redox potential usually ac - 

companies decreasing oxygen concentration . 

When the redox potential falls below a certain 

level the phosphorous goes into solution, there - 

by becoming available to photosynthesizers i n 

the waters above . In oligotropic lakes th e 

redox potential is always high and phosphorous 

bound in falling detritus is lost to th e 

system when it reaches the sediments . Th e 

cycling of nitrogen is more complicated an d 

even more intimately tied in with the redo x 

potential . The rate at which essential nutrients 

are recycled through the community, interchanged 

with the sediments, and lost and 

replenished through inflow and outflow deter - 

mines the productivity of the lake . The 

response of this system to perturbation, measured 

not only by the level of productivity but 

by changes in community structure, is th e 

central theme of our research effort . 

Production is usually defined as the total 

elaboration of organic matter by photosynthetic 

organisms in a specified time period , 

while productivity is the production per unit 

time. Photosynthetic rate depends principall y 

upon light intensity, temperature, and essential 

nutrient concentration (Riley 1946 , 

Rhyther 1956, Tailing 1961) . Photosynthetic 

rate is different for different species of alga e 

(Talling 1955) . The production under a uni t 

area of lake surface is the sum over species o f 

the integral with respect to time and depth o f 

the algae densities times their respectiv e 

photosynthetic rates . The above generalization 

may appear overly simple ; however, the 

34

complexity of lake productivity will become 

apparent. 

The exponential decay of light intensity is 

a function of depth . Temperature varies with 

depth . Although factors leading to lake stratification 

are well understood, the construction 

of predictive models for temperature distributions 

in lakes has only recently begun . Light 

intensity varies with time, because of seasonal 

and diurnal patterns plus changing weathe r 

conditions . There are vertical currents due t o 

advection and eddy diffusion in lakes (Rile y 

et al . 1949) . These currents cause temperature, 

nutrient concentrations, and the population 

densities to vary with time . We must als o 

consider feedback mechanisms such as selfshading 

(Tailing 1960), and local depletion of 

nutrients by algae . 

Another interesting feedback mechanis m 

involves the relation between phytoplankto n 

and zooplankton densities . The rate of grazing 

on phytoplankton is determined by zooplankton 

density and its relative grazing rate (Riley 

1947) . On the other hand, the relative grazing 

rate of zooplankton is a function of phytoplankton 

density ; a saturated grazing rate is 

reached at high algal population densities 

(Holling 1959, Riley 1947) . Zooplankton also 

influence nutrient availability by excretin g 

soluble nutrients into the water . 

In addition to the limnetic plankton communities, 

there are, in lake ecosystems, littoral 

communities and, in areas deeper tha n 

maximal light penetration, benthic communities. 

The processes that occur in these communities 

are somewhat different than in th e 

limnetic community . Littoral communities 

are similar to terrestrial communities in many 

respects . Emergent and floating forms have 

access to atmospheric carbon dioxide . Attached 

forms have direct access to nutrients 

held in the sediments ; whereas, limnetic algae 

are free floating and must depend upon larg e 

ratios of surface area to volume and upo n 

passive sinking for efficient nutrient absorption. 

In littoral areas, herbivores are small , 

relative to plant size . They consume little of 

the plant biomass. Littoral plants show seasonal 

die back, with much material being de - 

composed . As in terrestrial systems, bacteri a 

and detritus feeders are the primary agents of 

nutrient regeneration . 

Hutchinson and Bowen (1950), in a radiophosphorous 

study, showed that littoral 

plants can soak up and hold a large quantity 

of nutrients . Thus limnetic and littoral plant s 

do compete for nutrients . The littoral plants 

are less efficient than algae in absorbing nutrients 

because of their smaller ratio of surfac e 

area to volume . The littoral system's ability to 

hold nutrients (because of slow nutrient re - 

generation by bacterial decomposition an d 

relatively small grazing pressure) offsets inefficient 

nutrient absorption. Phytoplankton , 

because of their short life span and hig h 

predation rate, cannot hold nutrients as long 

as littoral plants, but they are very efficient in 

absorbing nutrients . 

Deep benthic communities are made up o f 

detritus feeders that feed upon seston rainin g 

down from the productive epilimnion . They 

may be important in nutrient regeneration 

(Johannes 1968) . These animals constitute 

important food sources for higher orde r 

consumers . 

A promising theory which may be of great 

help in generating and answering question s 

about higher consumers, is that of feeding 

strategies. For a review of a substantial literature 

on the subject, see Schoener (1971) . The 

essence of the theory of feeding strategies i s 

that selective processes tend to produce populations 

of organisms which strive to maximiz e 

their energy (essential nutriment) intake . 

Predators may accomplish this end in severa l 

ways; however, it appears reasonable that i t 

would pay to eat the largest prey or the easiest 

to capture whenever they are encountered , 

chasing relatively small difficult-to-catch prey 

only as a last resort . It is currently being 

demonstrated that this theory applies to a 

large variety of animals. The consequence is 

that one is able not only to predict the foo d 

habits of individual species but also to estimate 

their growth dynamics . 


The work reported in this paper was supported 

by National Science Foundation Grant 

GB-20963 to the Coniferous Forest Biome , 

35

U .S . Analysis of Ecosystems, International 

Biological Program . This is Contribution No . 

22 to the Coniferous Forest Biome, IBP . 


Holling, C. S. 1959. Some characteristics of 

simple types of predation and parasitism . 

Can. Entomol . 91 : 385-398 . 

Hutchinson, G. E. 1967. A treatise on limnology. 

Vol. II. Introduction to Lake 

Biology and the limnoplankton . 1115 p . , 

illus. New York : John Wiley & Sons . 

and V. T. Bowen. 1950 . Limnological 

studies in Connecticut . IX . A 

quantitative radiochemical study of th e 

phosphorous cycle in Linsley Pond . Ecology 

81 : 194-203 . 

Johannes, R . E . 1968 . Nutrient regeneration 

in lakes and oceans . In M . R . Droop and E . 

J. F. Wood (eds .), Advances in microbiology 

of the sea, Vol. I, p . 203-213 . New 

York : Academic Press . 

Pomeroy, L. R. 1970 . The strategy of mineral 

cycling . In Annu. rev. of ecol. and syst. , 

Vol. I, p. 171-190. Palo Alto : Annu. Rev . , 

Inc . 

Rhyther, J . H. 1956 . Photosynthesis in the 

ocean as a function of light intensity . 

Limnol. & Oceanogr . 1 : 61-70 . 

Riley, G . A. 1946. Factors controlling phytoplankton 

populations on Georges Bank . J . 

Mar . Res . 6 : 54-72 . 

1947 . A theoretical analysis o f 

the zooplankton population on George s 

Bank . J. Mar. Res . 6: 104-113 . 

, H. Stommel, and D . F. Bumpus . 

1949. Quantitative ecology of the plankto n 

of the western North Atlantic . Bull . 

Bingham Oceanogr. Coll. 12 : 1-169 . 

Schoener, T. W. 1971. Theory of feeding 

strategies . In Annu. rev. of ecol. and syst . , 

Vol. II, p. 369-404 . Palo Alto : Annu. Rev . , 

Inc . 

Talling, J. F. 1955. The relative growth rates 

of three plankton diatoms in relation t o 

underwater radiation and temperature . 

Ann. Bot. 14 : 329-341 . 

1960 . Self-shading effects in 

natural populations of a planktonic diatom . 

Wett. Leben 12 : 235-242 . 

1961 . Photosynthesis under 

natural conditions . Ann. Rev. Plant 

Physiol. 12 : 133-154 . 

36



Toward a general model structur e 

for a forest ecosystem 

W. Scott Overto n 

Professor, Departments of 

Statistics an d 

Forest Management 

Oregon State University 

A primary objective of the Coniferous Forest 

Biome is the development of a family o f 

models for a forest ecosystem as a whole . Thi s 

paper will discuss the general structures an d 

forms which represented our views toward th e 

end of 1971 . It is emphasized that these ar e 

evolutionary ; at least part of the value of thi s 

paper is as a record of the path we took i n 

model development . The value of these structures 

in furthering ecosystem theory is als o 

argued . 

Thus, a primary focus of this paper is wit h 

respect to ecosystem research at the "tota l 

system" level, that is, concerning the ecosystem 

as an object . It is another goal of ecosystem 

research to study the relations amon g 

component parts and further to study th e 

system properties of those parts . Populations , 

communities, assemblages, food chains, productivity 

processes, and many other sub - 

system structures are of interest, in their ow n 

right, as are the functional and relationa l 

aspects of these structures . In fact, one of th e 

great difficulties in modeling the ecosystem as 

a (single) object is the strong tendency to 

perceive the ecosystem as an assemblage o f 

poorly defined subsystems such that th e 

couplings are vague, at best, and often entirely 

undefined . Model structures are developed 

which, it is hoped, will contribute t o 

elaboration of these subsystem concepts, a s 

well as the concept of the total system as 

object. 

A brief statement of my view of models , 

and of their role in science seems desirable . 

Any particular model for a real world system 

(or object) can be decomposed into a set o f 

statements regarding that object . Some of 

these statements will be definitive, some wil l 

represent firm knowledge about the object , 

and some will be in the form of infirm knowledge 

or simplifications (which statements w e 

usually call assumptions) . If we consider th e 

collection of all possible models for an object , 

then the set of all statements of definition 

and all statements of firm knowledge represent 

all that is known about the object, s o 

that at this level the model, the canonical 

model, is identical to the state of knowledge 

about the object . 

Working models usually emphasize som e 

elements of knowledge and neglect others, s o 

that in practice the models we use are some - 

thing less than "complete ." However, most 

working models represent, in a very real way , 

the knowledge and theory of a real world 

system . Some subset of the total knowledge i s 

coupled with some set of simplifying assumptions 

and the collection given a form an d 

structure by a mathematical or relational 

algorithm or convention . The process puts th e 

present knowledge into a new form . If th e 

new form serves to increase the understandin g 

of the system, then the modeling operation i s 

37

successful . 

Thus we can see modeling as the impositio n 

of structure on existing knowledge . As such , 

it is a joint activity between those familia r 

with forms and structures and those familiar 

with a body of knowledge . In time, the structure 

imposed becomes integrated into th e 

scientific paradigm of the object, providin g 

new insight and a base for extending concepts 

. In this view, modeling is an integra l 

part of the theoretical advance of a subject 

science, not an external, nor peripheral , 

activity engaged in by mathematicians o r 

other specialists . 

Now I don 't think that I am the only on e 

with this view, but there are sufficiently man y 

opposing views among modelers and sufficiently 

many misconceptions among subjec t 

specialists that verbalization of this perspective 

is essential in an introduction of th e 

present topic . I view our efforts to develop a 

total ecosystem model as a contribution to 

the scientific paradigm of ecosystems . Ou r 

primary activity as modelers is the conceptualization 

of theoretical ecosystem structure s 

and behavior. The present paper deals wit h 

the structural aspects of a developing family 

of models for the forest ecosystem at a date 

late in 1971, and with the strategy we have 

adopted to pursue our objectives . 

Approaches to System 

Theory and Definition 

Klir (1969, 1972) discusses in some detail a 

variety of definition forms and approaches t o 

developing a general theory of systems . This 

seems a productive direction and the following 

section is my own interpretation and 

elaboration of these passages in Klir . There 

may be inadvertent deviations from Klir 's 

intent . 

The State Variable (or State Space) approach 

seems to describe the model structur e 

currently most popular in several fields, including 

ecology . The essence of this approac h 

is that the system is defined as a set of variables, 

the state variables, and a set of variable s 

representing the environment, in a particular 

temporal resolution . A popular version de - 

fines compartments, with the state variable s 

describing the contents of the compartments . 

The definition is completed by specificatio n 

of algorithms for change of the state variables 

in time . 

Klir has proposed a general systems theory 

which, although not distinct from the stat e 

variable approach, has certain features whic h 

apply nicely to study of ecosystems . Specifically, 

he identifies five ways in which a 

system may be defined : 

1. By the set of external quantities and th e 

resolution level 

2. By the given activit y 

3. By the permanent behavio r 

4. By the Universe-Coupling structure 

5. By the State-Transition structure 

The models of the system follow definitio n 

forms 3, 4 and 5 . Definition form 1 is used in 

the planning stages of a modeling or data 

collection activity, and a collection of data i s 

a realization of definition form 2. It follows 

that definition form 1 is implied by form 2 

and by forms 3, 4, and 5. When applied to th e 

same system, the forms must be mutually 

consistent . In the Coniferous Biome, we hav e 

adopted a particular combination of the thre e 

model forms for our model of the forest ecosystem. 

We view each system (or subsystem ) 

as modeled at two levels : 

1. Holistically, according to its Behavior (o r 

State-Transition structure) . 

2. Mechanistically, according to its Universe-Coupling 

structure . 

That is, we view Klir ' s Behavior and S-T 

structure as useful forms for characterizing 

the holistic behavior of a system "as a n 

object," and consider that such holistic characterization 

is necessary for each define d 

system or subsystem . Further, we consider 

that in most cases we will also wish to model 

the system according to the U-C Structure , 

that is, as a collection of subsystems, eac h 

modeled according to its Behavior and with 

the collection coupled in a manner appropriate 

to the behavioral forms used . 

Some further elaboration of these term s 

seems necessary, but a formal definition (as i n 

Klir 1969 and Orchard 1972) is inappropriate . 

An attempt to informally define some of th e 

38

terms follows. The external quantities (or 

variables') are those relevant quantities associated 

with the attributes of the object. Part 

of them are inputs (produced by the environment) 

and part are outputs (produced by th e 

object). Specification of these quantitie s 

constitutes definition form 1 . By definitio n 

form 2, an activity is a particular set of value s 

representing the system over a particular se t 

of time instants . 

The permanent behavior is a time invarian t 

relation between the outputs of interest and 

other quantities. In order to develop this in 

general, it is necessary to augment the specified 

external quantities by past values of th e 

external quantities (i .e., by a memory), the 

augmented set of quantities constituting the 

principal quantities . The outputs of interest 

are then the dependent quantities (a subset of 

the principal quantities), and the relation i s 

defined between these and the rest of the 

principal quantities . 

It is necessary to emphasize that this definition 

specifies that the principal quantities ar e 

generated from the external quantities alone . 

If internal quantities (i .e ., quantities defined 

on the subsystems but not on the whole) were 

allowed, then the state variable approac h 

would be a special case of definition form 3 . 

As it is, the two approaches (S-V and Behavior) 

have special cases in common . 

The State-Transition Structure is defined 

by recognition of the State as the instantaneous 

value of the external quantities, of the 

stimulus as the instantaneous value of th e 

input quantities and by a time invariant relation 

which sends the system from a particular 

state and stimulus at time t into the set of 

possible states at time t + h . The S-T structure 

is a useful special case of the permanent 

behavior, and in subsequent treatment wil l 

not be explicitly identified . 

One last point regarding behavior . In either 

form (of behavior) the time invariant relatio n 

can be either deterministic (i .e ., the relation is 

a mapping) or probabilistic (i .e., the relatio n 

is one to many) and in the latter case eithe r 

definition (Behavior or S-T structure) must be 

' The quantities are defined on the object . Variables 

are model counterparts . 

augmented by an appropriate probability 

function . 

Our notation, then, for the holistic definition 

of the system will be : 

S={Z,M,Y ;R} (1 ) 

where Z represents instantaneous input 

quantities (variables ) 

Y represents instantaneous output 

quantitie s 

M represents memory quantities (past 

values of Z or Y quantities ) 

and R is the appropriate time invariant relation 

between Y, on the one hand 

and Z and M on the other . 

It is the specification of the relation, R , 

which is the goal of ecosystem research at thi s 

level. R is the time invariant relation whic h 

permits description of the behavior of th e 

system (definition form 3) in place of th e 

simple "data record" provided by definitio n 

form 2 . In fact, it is a fair statement that a 

field data collection relates to system definition 

by the second form and that it is the rol e 

of data analysis and theoretical synthesis t o 

progress to forms 3 and 4 . 

The behaviorial representation of the system 

S as an object, as by (1), is followed by 

its perception as a coupled collection of sub - 

systems (sub-objects ) 

S = {So, S I , Sk ; C } (2) 

where C is the specification of the coupling s 

among the subsystems S i , . .., Sk. Typically , 

the coupling between two subsystems, say S i 

and Sj will be the coupling variables defined 

by ZjnYj and ZjnYi . That is, some of the 

outputs of subsystem Sj become inputs o f 

subsystem Si and vice versa, and these coupling 

variables define the couplings among th e 

subsystems . 

In Klir's theory, it seems to be assumed 

that all of the external quantities of the system 

S appear explicitly as external quantities 

of the subsystems (along with many ne w 

quantities which are external to Si, i = 1, . . ., k , 

but internal to S) . However, it appears to me 

that it is the nature of different levels o f 

organization that the external quantities 

differ not only in resolution and detail but 

possibly also qualitatively and that definition 

39

of the external quantities of S would be impossible 

in the context of isolated subsystems . 

What is needed is some translation of sub - 

system external variables into system externa l 

variables. Clymer and Bledsoe (1970) have 

addressed a similar question in consideratio n 

of "interfacing" two subsystems at the sam e 

level and use the term "slave model" to designate 

the set of equations for changing resolution. 

This term seems inappropriate in the 

context of composition of subsystem proper - 

ties into properties of the whole ; such a 

process is more master than slave . I have 

chosen the term "Ghost System i2 and notationally 

designated So as the integrating sub - 

system which derives the properties of th e 

whole from the properties of the parts . In its 

simplest form, So is a resolution expander of 

inputs and reducer of outputs . Whether it will 

need to assume a more complex form is yet t o 

be seen . 

In this manner we have, by statements (1 ) 

and (2), defined the ecosystem at two organizational 

levels, and may elaborate lower level s 

by the device of decomposing subsystems i n 

the same manner . That is, if S may be defined 

at two levels, so may the elements of {Si}, an d 

the subsystems so defined, to any desire d 

degree of fineness . Ultimately, it is necessary 

to end with a behavioral model for all sub - 

systems, because it is this model that yield s 

explicit form to the relational expressions . 

Several points are of interest here . First, i t 

is not clear that a hierarchical decompositio n 

is always desirable . That is, a finer resolutio n 

may not best take the form of a decomposition 

of next most coarse form of interest . 

Hierarchical forms, however, have man y 

advantages and will be used wherever possible . 

Second, it is quite clear that a uniform resolution 

is unnecessary . That is, a very coarse 

resolution model of one subsystem may b e 

coupled with a fine resolution model of 

another, and such an arrangement will have 

many advantages . Not the least of these is th e 

advantage of manageability . If we want to 

take a closeup view of some parts of the system, 

there is no reason to, and many reason s 

not to, look at the rest of the system at th e 

2 After Koestler's "The Ghost in the Machine ." 

same degree of magnification . A hierarchical 

subsystem structure will be most conducive t o 

a variable resolution, but again, it is no t 

necessary . 

Application of the 

General Theory to the 

Forest Ecosystem 

With this introduction to the theoretica l 

background, we can now take a look at th e 

coarser ecosystem structures as we now perceive 

them . A diagrammatic convention i s 

established which will be recognized as simila r 

to Forrester's, but which differs in some 

important respects . Let boxes designate compartments 

(to be conceptualized as storage 

containers), let ovals or circles designate systems 

(or subsystems), and let diamonds designate 

a set of variables . Solid lines designate 

coupled flow variables and dotted lines represent 

control variables. As a convention , 

diamonds are eliminated from the flow couplings, 

as these variables are apparent fro m 

the context, but identification of contro l 

variables is important . Occasionally these are 

also eliminated from a figure to reduc e 

clutter, and in the illustrations given here , 

most control paths are eliminated for th e 

same reason . 

In figure 1 is depicted a representation of a 

system as a whole, with inputs and outputs . At 

this level, the entire system (the forest ecosystem) 

is defined according to (1) . In figure 

2, the same system is elaborated by identification 

of major subsystems, the terrestrial biota , 

the aquatic biota and the hydrologic system . 

This involves definition of the system according 

to (2) and it should be emphasized at thi s 

point that there are very many ways in whic h 

this could be done . 

Implementation of this form (fig. 2) requires 

decisions regarding boundaries an d 

specifications of the nature of the coupling s 

between subsystems . An example will illustrate 

both points. Consider the uptake of 

water by higher plants . Question : Do we wish 

to consider this water as having transferred 

40

function, but this appears awkward . 

Once all the couplings are identified and 

specified, then it is no longer necessary to consider 

one subsystem in elaborating the interna l 

structure of another. The process of couplin g 

specification effectively uncouples the syste m 

into subsystems, each of which can be developed 

independently, provided that the identified 

external variables of the subsystems ar e 

maintained . 

I will use the word integrity in this context . 

Elaboration of the system at one level o f 

organization is unrestricted so long as th e 

integrity of the next higher level of organization 

is maintained. This does not appear at 

this writing to be a trivial problem, but th e 

problem exists under any formulation o f 

model structure and is well identified unde r 

the present structure, hence potentially 

manageable . 

Figure 1 . The system viewed as an object receivin g 

inputs from the environment and producin g 

outputs . 

from the hydrologic system to the terrestrial 

biotic system? Or, will we let stored water 

remain in the hydrologic system? The answer 

to this is arbitrary and should be made on th e 

basis of ultimate convenience . If the specification 

of couplings is simpler for one boundary 

definition than for another, then that boundary 

is preferred . 

The nature of the couplings (or couplin g 

variables) is also important . Suppose we le t 

water enter the biotic system at the momen t 

of uptake. Then the transfer of water fro m 

hydrologic to biotic depends on (1) the stat e 

of the hydrologic system (specifically, th e 

amount and distribution of water available i n 

the soil) and on (2) the biotic demand fo r 

water. These two variables must be identifie d 

as output variables of the respective system s 

in order to elaborate the transfer . It then i s 

immaterial whether the calculation of transfer 

takes place in the hydrologic or the bioti c 

system, but it seems appropriate that it b e 

included in the latter as an uptake function . 

In fact, one can even leave the function o f 

transfer outside both systems as a coupling 

Decomposition of the 

Terrestrial Biotic System 

With this background, I can develop further 

details of our present conceptualization . 

These results are presented in much greate r 

detail in a project document titled "Modeling , 

Round One," prepared jointly by myself an d 

a number of others in our program . This 

document summarizes a series of seminar - 

workshops in which we examined the definition 

of subsystems and the nature of thei r 

couplings . Reports of details of several sub - 

systems are reported by others in this symposium. 

I will use several subsystems for illustration 

which are not treated elsewhere . 

In figure 3 is seen a specification of primary 

subsystems of the terrestrial biotic 

system. These are not even, in the sense o f 

evenness of importance, but rather are chose n 

to emphasize the relationships we wish t o 

emphasize at this level . Only two compartments 

are specifically identified, the plan t 

biomass (B. & S.) and the detritus (D .) compartments, 

with all others embedded in the 

respective systems . Identification of the biomass 

compartment forces explicit identification 

of a minor subsystem, litterfall. Th e 

41

Figure 2 . The major subsystems of a forest ecosystem . Specification of the coupling variables among the majo r 

subsystems allows independent development of their internal structure . 

42

Figure 3. The December 1971 version of the subsystems of the Terrestrial Biotic System . The external variable s 

of the whole system match those illustrated for this subsystem in figure 2 . 

traditional decomposition and fixation sub - 

system has been uncoupled into the three sub - 

systems represented as Decomposition, Fixation 

and Weathering, and Interchange and 

Uptake (I-U) . 

The value of this formulation is several 

fold. First, we can recognize that this is a 

formal expression of generally held ideas, s o 

that only the formality, and perhaps the 

identification of the I-U process, is in an y 

degree innovative. The formality gives sub - 

stance to the concepts, and the formal U-C 

structure identifies the tasks facing us as w e 

elaborate this structure . These tasks are : 

(1) specification of the coupling variables , 

(2) elaboration of internal structure of the 

subsystems, and (3) analysis and synthesis o f 

behavior and properties of the systems and 

4 3 

t+

subsystems as objects . In addition, the formal 

structure effectively partitions our effort and 

ensures that all . bases are covered . Boundaries 

between subsystems and the appropriat e 

coupling variables must be identified jointl y 

by those responsible for the subsystems involved. 

After boundaries and coupling variables 

are agreed upon, then the activities o f 

elaboration of internal structure are independent 

among subsystems . This is another 

expression of the earlier observation tha t 

identification of the couplings effectivel y 

uncouples the subsystems . 

These values of the model formulation can 

be expressed as explication, conceptualization, 

and organization . Existing concepts are 

given explicit expression, new concepts ar e 

developed and research efforts are given structure 

and organization, all by the formal ecosystem 

model . The fact that the model structure 

is arbitrary in no , way obviates thes e 

values . On the contrary, the arbitrariness o f 

the model structure enhances these values , 

because as we proceed with the process of 

elaboration, and bring existing knowledg e 

into sharper focus by the process of modeling , 

we are constantly forced into adjustments, re - 

orientation, and reorganization . 

Structures, like figure 3, are not meant t o 

be permanent . They live and die like generations 

of insects . In fact, the insect analogy is 

quite appropriate for figure 3 . This structure 

was conceptualized in the spring and earl y 

summer of 1971, during Round One . I t 

developed underground, so to speak, all during 

the fall, undergoing several transitions, t o 

emerge in December in the form presented . 

By mid-January 1972 this form had lived ou t 

its life and given birth to a new form which is 

yet in the larval stages and not ready for th e 

light of day . 

At the next lower organizational level , 

progress has been made in elaboration of the 

internal structure of the food chain subsyste m 

and the I-U subsystem . The first is reported 

here by Strand and Nagel, but the second i s 

not represented in this symposium and I 

would like to include a brief description of 

that subsystem as additional illustration o f 

the way in which we are using the Universe - 

Coupling structure . 

Figure 4 represents a model form date d 

back in 1971, which form has been replace d 

by a new but incompletely developed form, in 

accordance with the changes being made a t 

the next lower resolution (fig . 3). However , 

figure 4 serves to illustrate the concept of th e 

I-U process and, again, the point is made tha t 

this is developed more fully in the document s 

of Round One . 

The I-U subsystem is postulated on the following 

statements : 

1. If nutrients remain in soil solution, they 

must be lost to the system by soil and 

groundwater transport . 

2. Nutrients must be in solution in order t o 

be available to uptake . 

3. Some microorganisms and some highe r 

plants are "leaky ." 

4. It is concluded that a successful syste m 

must have a tightly coupled, highly interactive 

and buffered subsystem of nutrient 

interchange and uptake . 

The modeling contribution here is th e 

recognition (4) that the three stated feature s 

of the traditional processes of nutrient inter - 

change are such that a successful terrestria l 

ecosystem (i .e., one which does not lose it s 

nutrients downstream) must have a tightly 

coupled I-U subsystem . That is, the traditional 

study by individual process canno t 

possibly answer the questions we want to ask , 

unless we explicitly define the couplings . 

Coupling definition is difficult in a tightl y 

coupled system, and appears exceedingl y 

difficult in this one . The identified task is th e 

conceptualization of properties and behavior 

of the I-U subsystem as a whole . 

It is my very strong conclusion here that 

the Biome research effort should be re - 

oriented to accommodate the concept of th e 

I-U subsystem. Present conceptualization an d 

specification of that subsystem is yet very 

primitive and will receive considerable attention 

in the next year . 

From the model structure point of view , 

the primary productivity subsystem is th e 

least well defined in our model . This is, in 

part, because most of our research effort ha s 

been at lower organizational levels than i s 

necessary for ecosystem models . It is a giant 

step from tree to community . The primary 

44

I 

I 

I 

t 

/ INTERCHANG E 

a UPTAKE BY SOI L 

HIGHER 

MICROORGANIS M 

PLANT S 

ION I C 

BONDI NG, 

HUMU S 

1 

SOI L 

SOLUTION ----~-- 

1 

Figure 4 . The December 1971 version of the subsystems of the Interchange and Uptake System. The externa l 

variables of the whole system match the couplings of the I-U subsystem of figure 3 . 

effort has been at the tree level (and below ) 

and the model structures for the ecosystem s 

are communities. The necessary effort here is 

apparent . 

Representation of Spatial 

and Environmental Variatio n 

The Universe Coupling structure is als o 

appropriate to the problem of modeling a 

heterogenous ecosystem . If the system i s 

stratified into homogenous subsystems , 

identified by spatial or environmental criteria , 

then the models appropriate for each of the 

strata can be coupled together to form a 

model for the whole . This is the approach we 

have taken in our watersheds . A watershed is 

sufficiently variable, with regard to physical 

and biological processes, that we feel it i s 

necessary to stratify so as to provide essential 

uniformity of processes within strata. Th e 

process of composition, of coupling th e 

4 5

several models together to form the model for 

the watershed as a whole, is the reverse of the 

decomposition process previously discussed. 

Now we are faced with the question of defining 

appropriate external variables of th e 

watershed as an object after having initiall y 

defined the watershed as a collection of sub - 

objects . 

These questions are obviously appropriate 

to consideration of a general model for, say , 

the entire H . J. Andrews Experimental Forest , 

or for the Coniferous Forest Biome, sob that 

we consider them central to our overall 

objectives. By concentrating on the proble m 

of stratifying a watershed, and on the problems 

of modeling the strata and the collective , 

we are attempting to devise strategies which 

will be useful in later extension of the modeling 

process . 

It turns out that if we view the construction 

of strata in essentially the same perspective 

as the construction of subsystems in th e 

U-C structure, the same general criteria hold . 

Generally speaking, subsystems (strata ) 

should be constructed in such a manner tha n 

couplings are minimized (or simplified), an d 

such that important processes are containe d 

within the subsystems . This criterion is no t 

always compatible with the second one, tha t 

strata should be environmentally homogenous, 

within, and occasional conflicts arise . 

In any event, we are attempting hierarchica l 

stratification of watersheds, with topographic 

features defining the primary strata and vegetation 

types the secondary, the two levels giving 

uncoupling and homogeneity, respectively . 

Summary 

Some of the ways in which the Universe- 

Coupling structure is being used in building a 

hierarchical, modular system of models for 

the Coniferous Biome have been discussed . 

These models are hierarchical by virtue of th e 

identification of systems at one level o f 

organization as subsystems of the system s 

defined at the next higher level. They ar e 

modular by virtue of the uncoupled nature ; 

identification of the coupling variables at an y 

level allows complete flexibility of subsystem 

representation, provided that the integrity of 

the couplings is maintained . 

A general criterion for constructing sub - 

systems is provision of simple coupling s 

among subsystems with tight relations contained 

within . This principle is violated quite 

badly when trophic levels are used for sub - 

system definition . An alternate structure for 

consumers is being investigated in our program, 

as reported in the paper by Strand and 

Nagel . 

In applying the U-C structure to spatial 

units, another criterion is employed . Here it i s 

desirable to provide homogeneity of environment 

and process within strata, with variatio n 

among strata of no concern . This criterion i s 

sometimes in conflict with the first one o f 

minimal couplings, so that compromises are 

necessary . 

The paper has dealt primarily with aspect s 

of structure and criteria for application of th e 

structural form . Emphasis has been on th e 

first two of the three tasks which the U- C 

structure defines, to wit : (1) specification of 

the coupling variables, (2) elaboration of 

internal structure of subsystems . The third 

task, analysis and synthesis of behavior of the 

subsystems as objects, is given little attention , 

but this probably is the most important, fro m 

the point of view of ecosystem theory . As 

mentioned earlier, the conceptualization of 

holistic system properties and behavior i s 

poorly advanced. The U-C structure provides 

an excellent basis for the attempt to develop 

this concept . 

One last point is made regarding the differences 

between the Universe-Couplin g 

approach and the State Variable approach . 

Since the State Variable model can als o 

assume hierarchical form (Goguen 1970), an d 

under common explicit models of subsyste m 

behavior the U-C model will reduce to a stat e 

variable model, one might question the practical 

validity of the distinction . Essentially, th e 

differences are that one (U-C) is structur e 

oriented and the other variable oriented an d 

that one (U-C) is oriented simultaneously to 

holistic and mechanistic representation an d 

the other solely to mechanism . In the application 

of ecosystem models so far produced, th e 

state variable approach seems adequate , 

46

except for the retention of fine detail . Th e 

tendency in constructing models on this base 

has been to develop fine resolution mechanistic 

models in which the variables defined a t 

the finest resolution are in direct relation t o 

all the other variables in the system . Variabl e 

orientation seems to lead to variable sanctity . 

One might make a case for the position that 

the state variable approach was developed b y 

engineers who have no conceptualizatio n 

problem . They know what their variables are , 

and the relations among them, and they ar e 

satisfied with mechanistic models . 

The U-C structure is seen as a means o f 

introducing holism, in addition to mechanism, 

into the model conceptualization process . By 

focusing on the subsystem structures, and b y 

attempting to describe the behavior of the 

system and subsystems at all levels in terms of 

their behavior as objects, each system an d 

subsystem is modeled at two levels, (1) holistically 

and (2) mechanistically in terms of th e 

holistic behavior of its subsystems . This 

approach is held to be much more promising , 

from the point of view of development of 

ecosystem theory, than the currently popula r 

approach . 



in part by National Science Foundatio n 

Grant No. GB-20963 to the Coniferous Forest 

Biome, U .S . Analysis of Ecosystems, International 

Biological Program. This is Contribution 



Clymer, A. B., and L . J. Bledsoe. 1970 . A 

guide to the mathematical modeling of a n 

ecosystem. In R. G . Wright and G . M. Van 

Dyne (eds.), Simulation and analysis of 

dynamics of a semi-desert grassland, p . 

75-99 . Colo . State Univ . Range Sci . Dep . 

Sci. Ser. No . 6 . 

Forrester, J . W . 1971 . Counterintuitive behavior 

of social systems . Technol. Rev . 73 : 

53-69 . 

Goguen, J. A . 1970 . Mathematical representation 

of hierarchically organized systems . In 

E . O . Attinger (ed .), Global systems 

dynamics, p . 112-129 . New York : Wiley . 

Klir, G. J. 1972 . The polyphonic general 

system theory . In G . J . Klir (ed .), Trends in 

general systems theory, p . 1-18 . New York : 

Wiley-Interscience . 

Klir, George J . 1969 . An approach to general 

systems theory. 323 p. New York: Van 

Nostrand Reinhold Co . 

Orchard, R . A .. 1972 . On an approach to general 

systems theory . In G. J . Klir (ed .) , 

Trends in general systems theory, p . 

205-250. New York : Wiley-Interscience . 

Strand, M . A., and W. P. Nagel. 1972 . Preliminary 

considerations of the forest canop y 

consumer subsystem . In Jerry F. Franklin , 

L. J. Dempster, and Richard H . Warin g 

(eds.), Proceedings-research on coniferou s 

forest ecosystems-a symposium, p . 71-77 , 

illus. Pac. Northwest Forest & Range Exp . 

Stn., Portland, Oreg . 

47

Proceedings-Research on Coniferous Forest Ecosystems-A symposium. 


Hydrologic modeling in the 


George W . Brow n 

Chairman, Hydrology Subcommittee 


Robert H . Burgy 

Hydrology Project Leade r 

University of California, Davi s 

R . Dennis Har r 

Hydrology Project Leade r 


J . Paul Riley 

Hydrology Project Leader 

Utah State University 

Abstract 

The objective of the hydrology program is to prepare a model which will provide predictions of the hydrologi c 

state of a coniferous watershed at any desired time and in any desired place, where state is defined by the inpu t 

needs of the other submodels or systems, particularly the producer and biogeochemical processes. Subsurface flo w 

is the dominant runoff mechanism in coniferous watersheds and one of the least understood processes in 

hydrology. Research projects in hydrology seek to understand this process using three different techniques . One 

project relates subsurface flow to soil properties using direct measurement techniques. Another project approaches 

the problem using simulation techniques. The third project utilizes systems analysis and statistical decompositio n 

of runoff events to make inferences about subsurface flow. These studies of hydrologic processes will b e 

incorporated into a hydrologic model and linked to studies of other systems. The first step in linking our mode l 

with those of other groups is watershed stratification, a problem now solved by our modeling efforts. 

Introduction 

Water is an essential component of any ecosystem. 

In the Pacific Northwest, water is a 

dominant element. The coniferous forests o f 

this region are noted as some of the best - 

watered terrestrial ecosystems in the United 

States; water is the major linkage which tie s 

the terrestrial portion of the coniferous ecosystem 

to the aquatic portion . 

Water performs several functions which 

foster this linkage . Water must be viewed as a 

carrier. It carries organic and inorganic nutrients 

between the several compartments of the 

terrestrial portion of the ecosystem and fro m 

the terrestrial to the aquatic portion . Wate r 

also carries sediment from the terrestrial t o 

the aquatic portion of the system . 

Water must also be viewed as a nutrient 

itself. It is an essential component of most 

biologic processes . The availability of water i n 

the soil governs both the initiation and termi - 

49

nation of any process as well as the rate a t 

which it proceeds . 

Objective of the Hydrology Progra m 

The objective of our program is to prepare 

a model which will provide predictions of th e 

hydrologic state of a coniferous watershed at 

any desired time and in any desired place , 

where state is defined by the input needs of 

the other submodels or systems, particularl y 

the producer and biogeochemical processes . 

Structure of the Hydrology Modeling Effort 

Our modeling effort is organized to pa y 

particular attention to the many functions o f 

water in the forest ecosystem . A generalized 

model for water flow through a forest syste m 

was conceptualized long ago . This model is 

often called the hydrologic cycle . Rothacher 

et al. (1967) showed for the H . J. Andrews 

Experimental Forest that water movement i n 

the forest soil and evapotranspiration are th e 

most significant processes governing water 

flow . 

Our studies focus upon subsurface wate r 

movement. One project measures subsurface 

flow directly in the study watershed . Another 

study is a computer simulation of a water - 

shed. This approach will provide yet anothe r 

avenue for assessing soil moisture and the sub - 

surface flow of water. The simulation model 

will be calibrated using 14 years of record o n 

watershed 2 . Then the model will be verifie d 

with the data available on watershed 10 . A 

third project seeks to develop techniques for 

predicting the subsurface flow componen t 

using a systems analysis technique for statistical 

decomposition of the hydrograph into its 

components . Other studies will contribut e 

submodels to the simulation of the hydrologi c 

system. The work of the Primary Producer 

group on a transpiration model as well as th e 

work of the Meteorology group on an evapotranspiration 

model will aid our modeling 

effort significantly . 

We are concurrently working toward a 

more spatially refined model, the character of 

which is determined as much by biologic a s 

hydrologic constraints . Our latest efforts hav e 

focused upon devising a system for stratification 

of watershed 10 which is amenable t o 

both hydrologic and biologic models . 

Other hydrology projects are included i n 

the biome effort. We hope to provide hydro - 

logic measurements as a part of the work a t 

Findley Lake and the evapotranspiratio n 

study at the Thompson site . We shall soo n 

begin to solicit and organize available dat a 

from coniferous watersheds throughout th e 

West in anticipation of extrapolating our 

model to other ecosystems . 

This has provided a broad overall view of 

the Coniferous Biome Hydrology Program-its 

structure to achieve a better understanding o f 

water flow through the ecosystem and its 

interaction with other system components . A 

more detailed description of the hydrolog y 

efforts follows . Each project focuses upon 

evaluating water flow in a forested watershed . 

Analytical techniques vary between projects , 

but the ultimate goal of modeling subsurfac e 

flow mechanisms and watershed respons e 

links all projects together . 

Subsurface Movement 

of Water on 

Steep, Forested Slopes l 

With the exception of stream channel interception, 

the hydrographs of watersheds in the 

forested, steep topography of western Oregon 

reflect overall subsurface movement of water . 

Watershed response is rapid but without 

surface runoff (Barnett 1963, Rothacher et al . 

1967) . Although subsurface flow is by far th e 

major component of the hydrograph, virtuall y 

nothing is known about the process on stee p 

slopes. The objective in our study is to characterize 

the subsurface movement of water i n 

steeply forested topography . 

1 Authored by R. Dennis Harr, Assistant Professor, 

Oregon State University, Corvallis . 

50

The study areas are located at several of th e 

lower watersheds in the H . J. Andrews Experimental 

Forest near Blue River, Oregon . Vegetation 

is typical of the low-elevation Douglasfir 

forest. Study slopes average about 75 per - 

cent. Soil depth is variable with maximu m 

depths in excess of 5 meters . Because of high 

porosities (70-80 percent) and large proportions 

of macropores, these soils drain rapidly . 

Permeabilities of 5,000 and 900 mm per hou r 

have been noted on nearby watersheds fo r 

surface soil and subsoil, respectively (Dyrnes s 

1969) . 

Methods 

Initial investigations are being directe d 

toward describing the physical properties of 

the porous medium through which wate r 

moves on its way to a stream . These field and 

laboratory investigations will indicate where 

water movement most likely occurs . 

Drilling with a portable power drill wil l 

follow a grid pattern over a small stream-toridge 

portion of slope . At each grid point soi l 

depth, depth and thickness of saprolite, an d 

depth to unweathered bedrock will be deter - 

mined. Additional drilling between initial gri d 

points will indicate in more detail the surfac e 

contour of the impermeable parent material . 

Aluminum tubing placed in each hole wil l 

provide access for measurement of ground - 

water level or soil moisture content . 

In the laboratory, undisturbed soil core s 

taken from various depths in soil pits locate d 

over the study area are being analyzed . Such 

properties as porosity, pore-size distribution , 

stone content, permeability, and moisture retention 

characteristics are being evaluated . 

The type and amount of measurements to 

be made during and following winter stor m 

events in 1972-73 will depend on the in - 

formation gathered during initial field an d 

laboratory investigations now underway . 

Anticipated measurements include soil moisture 

content, vertical and lateral extent o f 

saturated flow, soil moisture tension, precipitation, 

and water outflow from the base of 

the slope . Drilling and tracer studies will 

attempt to define the source area for this 

water . 

Preliminary Results 

Although the study has just recently begun , 

certain observations have provided qualitative 

information concerning the subsurface flow 

process on steep slopes. Precipitation moves 

downward under the influence of gravity until 

this movement is obstructed . In some parts of 

the study area this obstruction may be cause d 

by rock fragments which cause shallow , 

localized saturation as evidenced in several 

soil pits during a period of heavy rain . Where 

rock fragments are not present, downward 

movement of water continues until the relatively 

impermeable parent material is reached . 

Here saturation occurs, flow acquires a horizontal 

component, and water begins movin g 

toward the stream. 

At some point on the slope this saturated 

flow is concentrated into pipelike subsurfac e 

channels. The cause of this concentration i s 

unknown but could conceivably result fro m 

the microrelief of the impermeable material , 

from bedrock fractures, or from decayed root 

channels. At the toe of the slope the channels 

are spaced about 1-6 meters apart . They lie on 

the bedrock surface and appear associated 

with surface micro-relief. Shapes of their cross 

sections range from circular to flat rectangular. 

Width is also variable, ranging from 1 

centimeter to about a meter. Where these 

channels discharge into the stream channel , 

they are separated by soil which may contai n 

a shallow saturated lower layer from whic h 

seepage occurs. Water velocity of the seepage 

appears to be several orders of magnitud e 

lower than that of the subsurface channels . 

The latter accounts for the greatest portion o f 

stormflow . 

The subsurface channels evident at the to e 

of the study slopes may be outlets of a sub - 

surface drainage system much like that de - 

scribed for other humid areas (Jones 1971) . 

Water can be observed discharging from suc h 

channels in roadcuts and recent soil slumps at 

various slope positions in the vicinity of th e 

study area. Such a subsurface drainage syste m 

could account for the rapid hydrologic 

response of these steep slopes . 

51

Computer Simulation 

of Fores t 

Watershed Hydrology 2 

A hydrologist is often faced with the need 

to predict system responses under variou s 

possible management alternatives . One approach 

to this problem is to apply the technique 

of computer simulation, whereby a 

quantitative mathematical model is developed 

for investigating and predicting the behavio r 

of the system. In this study, a computer 

model is being developed to simulate th e 

hydrologic responses of a forest watershed , 

emphasizing the measurable variables related 

to the plant communities and soil types of th e 

watershed . The model represents the inter - 

related processes of the system by function s 

which describe the different components o f 

physical and biological phenomena in a 

watershed . 

Scope and Objectives of th e 

Simulation Stud y 

In the first phase of the study, the scope is 

being limited to the formulation of a fundamental 

model of watershed hydrology whic h 

takes precipitation as the basic input and 

evapotranspiration and streamflow as output s 

of the system . The various component processes 

within the system are linked by the conservation 

of mass principle. Depending upon 

energy levels, water can vary among its solid , 

liquid, and vapor forms ; hence, the energy 

budget is used as an auxiliary tool for maintaining 

the water balance . That aspect of th e 

system involving water as a carrier of nutrients 

and sediments will be examined in a 

subsequent phase of the study . Under this 

next phase a water quality submodel will b e 

formulated and added to the quantity mode l 

now being developed . 

2 Authored by J . Paul Riley, Professor, Utah Wate r 

Research Laboratory, Utah State University, Logan ; 

George B . Shih, Research Engineer, Utah Water 

Research Laboratory, Utah State University, Logan ; 

and George E . Hart, Associate Professor . Fores t 

Science, Utah State University, Logan, Utah . 

The specific objectives of the current phase 

of the study are stated as follows : 

1. To develop and verify (calibrate an d 

test) a hydrologic simulation model for a 

small forested subwatershed on the H . J . 

Andrews Experimental Forest . 

2. To estimate through model sensitivit y 

studies the relative importance of variou s 

processes within the hydrologic syste m 

of the model, with particular emphasi s 

on evaluating the soil moisture and inter - 

flow components . 

Hydrologic System Models 

Several hydrologic simulation models ar e 

currently available . Examples which might be 

cited include Crawford and Linsley (1964) , 

Sittner et al . (1969), and Riley et al.(1966) . 

However, in order to meet the needs of thi s 

study all existing models require some modifications 

and further development . Therefore , 

on the basis of previous work at Utah State 

University a computer model is being developed 

to simulate the hydrologic behavior of 

forest watersheds . The model will be applicable 

to a wide variety of geographical area s 

and management problems . In this study, data 

from watershed 2 on the H . J. Andrew s 

Experimental Forest will be used to demonstrate 

the utility of the model . Figure 1 summarizes 

the geophysical features of the study 

area (Rothacher et al . 1967). Data requirements 

include air temperature, precipitation , 

runoff hydrographs, and characteristics of th e 

watershed (average slope, degree and aspect , 

vegetative cover, density of vegetative cover , 

soil moisture holding characteristics, an d 

drainage density). Other observed records , 

such as snow depth, soil moisture content will 

be used to check the performance of th e 

model in simulating various component processes 

of the system . 

Figure 2 illustrates the various componen t 

processes represented in this model, with th e 

boxes representing storage locations and th e 

lines transfer functions . Under this study, th e 

hydrology of the drainage area will be synthesized 

first as a lumped parameter model in 

which the entire watershed area is considere d 

as a single space unit. On this basis, a dis - 

52

K m 

Figure 1 . Watershed 2, H . J. Andrews Experimental Forest, Oregon . Area: 60.3 hectares ; aspect: NW ; 

average slope (percent) : 61 .1 ; elevation min .: 526 m; elevation max .: 1,078 m ; main channel length : 

1,108 m; drainage density: 4.3 km/km 2 ; precipitation (1952-62) : 2,400 mm/yr; runoff (1953-62) : 

1,560 mm/yr; average evapotranspiration (1959-62) : 540 mm/yr . 

tributed parameter model will be developed i n 

which the watershed will be divided into four 

space units, roughly corresponding to subwatersheds 

within the area . The model will 

compute continuous daily streamflow for 

each subarea and route the contribution of 

each down the streams to the gaging station , 

where the computed and observed discharges 

will be compared. Other important outpu t 

functions from the model will include soi l 

moisture, actual evapotranspiration, and sno w 

depth. Several of the component processe s 

which are illustrated by figure 2 are discussed 

in the following sections . 

Interceptio n 

Interception is the part of precipitation 

that is caught temporarily by forest canopie s 

and then redistributed either to the atmosphere 

by evaporation or sublimation or to the 

forest floor. The amount of interceptio n 

depends upon storm size and intensity, an d 

canopy type and density. A report by 

Rothacher (1963) showed that throughfall in 

the study area was related to storm size b y 

the equation : 

Throughfall = 0 .8311 x (gross precipitation) - 0 .117 (1 ) 

In equation 1 throughfall is, of course , 

bounded by the condition that it must be 

greater than or equal to zero . Although larger 

amounts of snow may be temporarily intercepted 

than rain, there is strong evidence that 

most intercepted snow ultimately falls and be - 

comes part of the snowpack . 

53

CLOUD S 

I 

EVAPORATION 

SUBLIMATION 

INTERCEPTIO N 

0 

SNOW SNOW SNO W 

z 

STORAG E 

H 

z 

0 

H 

EVAPOTRANSPIRATION 

WATER STORED ON 

LAND SURFACE 

SNOWMEL T 

H 

EVAPOTRANSPIRATION 

0 . 

H 

H 

w 

H 

SOIL MOISTURE 

OVERLAND FLO W 

INTERFLOW 

STREAM 

CHANNEL 

STORAG E 

SURFAC E 

OUTFLOW 

GROUNDWATER 

INFLOW 

GROUNDWATER 

STORAGE 

EFFLUENT FLOW -Aim. 

INFLUENT FLOW E- 

GROUNDWATER OUTFLO W 

Figure 2 . A flow diagram of the hydrologic system within a typical watershed area. 

54

An alternative way of considering interception 

quantities in a model is to express 

interception rate as a decaying function of 

time limited by an average interception storage 

capacity for the watershed canopy . This 

approach was incorporated into a watershed 

simulation model by Riley et al . (1966) . 

Forms of Precipitation 

Snow Storage and Melt 

Only two forms of precipitation, rain an d 

snow, are considered in this study, with a 

surface air temperature criterion being applie d 

to establish the occurrence of these two 

forms. Figure 3 (U .S. Army Corps of Engineers 

1956) shows that at a temperature of 

1.5°C there is a 50-percent chance that th e 

precipitation will be in the form of snow . A 

straight-line fit to figure 3 is used to deter - 

mine the portion of rain in a given day ac - 

cording to the following equation . 

Ta - Ts 

R=P 

(2 ) 

T r- Ts 

in which 

R = estimated portion of total dail y 

precipitation occurring as rai n 

P = total daily precipitation 

Ta 

T s 

Tr 

= mean daily surface air temperatur e 

= mean daily air temperature below 

which all precipitation is assumed 

to occur as snow 

= mean daily air temperature above 

which all precipitation is assumed 

to occur as rain 

Precipitation falling as snow will be accumulated 

on the watershed until air temperatures 

rise sufficiently above the freezing poin t 

to initiate snowmelt . 

10 0 

1 0 

9 0 

2 0 

8 0 

3 0 

7 0 

4 0 

5 0 

6 0 

7 0 

8 0 

9 0 

100 

-1 .5 -1 .0 -0 .5Ts 0 0 .5 1 .0 2 .0 3 .0 4 .0 Tr 

SURFACE AIR TEMPERATURE, ° C 

Figure 3 . Frequency distribution of precipitation in rain and snow forms . 

0 

55

LIQUID WATER-HOLDIN G 

CAPACITY OF SNOWPAC K 

Figure 4 . A flow chart of the snow accumulation and ablation processes. 

56

Snowmelt 

A flow chart of the snow accumulation an d 

ablation processes is shown by figure 4 . Rate 

of snowmelt depends primarily upon the rate 

of energy input to the snowpack . However, 

both the complex nature of snowmelt an d 

data limitations prevent a strictly analytica l 

approach to the simulation of this process , 

and air temperatures are frequently applied a s 

an index of available energy . Examples of 

researchers who have used this approach ar e 

Pysklywec et al. (1968), Anderson and Craw - 

ford (1964), Amorocho and Espildor a 

(1966), and Eggleston et al. (1971). Becaus e 

temperature data are the only indicators o f 

energy levels available on watershed 2, a 

degree-day approach based upon the work o f 

Eggleston et al. (1971) will be used in th e 

model of this study to represent the snowmelt 

process at the surface of the snowpack . This 

component submodel includes mathematical 

relationships for various phenomena involve d 

in the snowmelt process . The submodel is 

applicable to any geographic location by determining 

appropriate constants for certain re - 

lationships through a verification procedure . 

The relationship for surface melt rate i s 

expressed as follows : 

RIs 

Prg 

Mrs =kmkvRIh Ta( 1-A ) +Ta 80 (3) 

in whic h 

km = a constant of proportionality 

kv 

RI h 

RI s 

= vegetation transmission coefficient 

for radiatio n 

radiation index for a horizontal surface 

at the same latitude as th e 

particular watershed or zone under 

study 

radiation index for a particula r 

watershed zone possessing a known 

degree and aspect of slop e 

T a = surface air temperature in °C 

A = albedo, or reflectivity, of the snowpack 

surfac e 

Prg 

precipitation reaching the snow surface 

in the form of rain, in 

centimeters 

Infiltration 

Rates of water supply on the ground surface, 

whether in the form of rainfall minus 

interception or snowmelt, must exceed infiltration 

rates before any surface runoff occurs . 

The infiltration rate depends on the physical 

and moisture characteristics of the soil, as 

well as the surface organic conditions, and it 

is often expressed in the form of Horton's 

exponential equation . However, the soils of 

watershed 2 are very porous and no overlan d 

flow has been observed . Thus, all precipitation 

reaching the ground surface is assumed to 

infiltrate into the soil and to move to th e 

stream channels as subsurface flow . 

Soil Moisture 

Soils on the study watershed are relatively 

deep and have a high porosity . Data on the 

physical properties of the watershed soils ar e 

available (Rothacher et al . 1967), and this 

information will be used to determine the soi l 

moisture holding characteristics . 

The computer model allows infiltrating 

water to satisfy first the available moistur e 

holding capacity of the soil within the roo t 

zone of the forest canopy . When the available 

soil moisture holding capacity is reached , 

additional infiltration is assumed to percolate 

by gravitation either somewhat laterally with - 

in the root zone or downward to deeper soi l 

zones. Water which moves laterally usuall y 

reaches a surface channel within a relativel y 

short period of time, whereas deep percolation 

moves from the watershed more slowl y 

and sustains streamflow during dry seasons . 

From preliminary studies (Rothacher et al . 

1967), approximately 87 percent of the tota l 

annual precipitation reaches the ground surface, 

and about 75 percent of this quantity 

becomes surface runoff . From an analysis of 

streamflow hydrographs it is estimated tha t 

about 10 percent of the runoff comes fro m 

baseflow, which is contributed from dee p 

percolation . Thus, of the average annual 

precipitation of 2,400 mm which falls on th e 

watershed approximately 2,100 mm enter th e 

soil, 440 mm are abstracted by evapotranspiration, 

and the remaining 1,660 mm leav e 

57

the watershed as surface runoff, with 170 m m 

of this quantity occurring as baseflow . Th e 

soil moisture content computed by the mode l 

will be checked with observed data . 

Evapotranspiratio n 

Factors affecting evapotranspiration includ e 

temperature, solar radiation, wind, humidity , 

and consumptive use by plants . However , 

only temperature and humidity data are 

available for the watershed . Among the commonly 

used evapotranspiration equations 

(Veihmeyer 1964), the Penman equation i s 

perhaps the most rational, but the data requirements 

are extensive . For this reason, the 

modified Hargreaves (Veihmeyer 1964), will 

be used in this study . The equation is stated 

as follows . 

U =E Kd(0 .38-0.0038h)T a (4) 

in which 

U = the daily potential evapotranspiration 

in centimeters 

d = the daily daytime coefficient de - 

pendent upon latitud e 

h = the mean daily relative humidity at 

noon 

Ta = the mean daily surface air temperature 

in ° c 

K = a monthly consumptive use coefficient 

which is dependent upo n 

plant related characteristics, such a s 

species, growth stage, and densit y 

on the watershe d 

The influence of soil water on evapotranspiration 

has been the subject of much re - 

search and discussion . It is now generall y 

recognized that there is some reduction i n 

evapotranspiration rate as the quantity o f 

water within the root zone decreases . In this 

study it will be assumed that evapotranspiration 

occurs at the potential rate through a 

certain range of the available soil moisture . A 

critical moisture level is then reached at whic h 

actual transpiration begins to lag behind th e 

potential rate . Within this range of the avail - 

able soil moisture the relationship betwee n 

available water content and transpiration rat e 

will be assumed to be virtually linear . Thus, 

E = U, [Mes < Ms(t) < M cs] (5 ) 

in which 

E = daily evapotranspiration adjuste d 

for the influence of soil moistur e 

levels 

M s quantity of water stored within th e 

root zone and available for plan t 

use at any time, t 

Mes limiting root zone available moisture 

content below which soil moisture 

tensions reduce evapotranspira - 

tion rates 

M cs root zone storage capacity of water 

available to plants 

M s (t) 

E = U [Mes > M s (t) >_ O] ( 6 ) 

Mes 

Considering the pressure effect, the total rate 

of gravity water storage depletion through 

both interflow and deep percolation is 

assumed to be directly proportional to th e 

quantity of water in this form of storage remaining 

in the soil profile at any particula r 

time. The interflow portion of this depletion , 

Nr, will be expressed as follows : 

Nr ( t) = Ki 

ddGs = Ki (Kg Gs (t) ) 

in which 

K i = interflow depletion coefficien t 

Kg = gravity water depletion coefficien t 

-K t 

That is, Nr = KiKgGs(0)e g , in which gravity 

storage at time, t = 0 is represented b y 

Gs(0) and no input to Gs is assumed to occur 

between t = 0 and any other time, t . It is estimated 

that on watershed 2 about 90 percen t 

of the gravity water storage leaves the area as 

interflow, in which case KiKg = 0 .9 . 

Groundwater 

(7 ) 

Water enters groundwater storage as dee p 

percolation from the overlying plant root 

zone. The rate of deep percolation, Gr, is 

numerically equal to the total rate of gravity 

water depletion within the root zone less th e 

interflow rate . Thus, 

58

d Gs 

Gr = (1 - Ki) d t 

By integrating equation 9 over a specific tim e 

period the accumulated inflow to the ground - 

water basin, Gam,, is estimated for this tim e 

period . 

Gw = f 

o 

G 

(1 - Ki) d dt (10) 

dt 

If the groundwater basin is considered as a 

linear reservoir, the outflow rate is given by 

the expression 

Qrg = K b Gw ( 11 ) 

in whic h 

K b = a coefficient which is estimated 

from dry season streamflow hydrograph 

s 

Qrg = the outflow rate from the groundwater 

reservoir 

By combining equations 9 and 11, the net 

rate of storage change within the groundwater 

basin is derived a s 

Gr - Qrg d t 

By substituting equation 11 into equation 1 2 

and rearranging terms, the following relation - 

ship is obtained . 

d dtg = Kb Mr (t) - Qrg (t)l (13) 

The rate of discharge from the groundwate r 

basin as baseflow is obtained by solving equation 

13 for Qrg. 

Runoff 

d G am, 

(9) 

(12) 

The possible sources of streamflow at any 

reach within a channel are overland flow (surface 

runoff), interflow, groundwater, and up - 

stream input. Manning's equation is usuall y 

applied to compute overland and channel flow 

rates at any point . Under conditions on watershed 

2, however, surface runoff does not occur , 

and channel routing on a daily time increment 

is not significant . Therefore, runoff rates at th e 

stream gage are given by summing the interflow 

and groundwater discharge rates . 

Model Verification 

Model verification includes calibration of 

the model parameters to a particular area , 

testing the sufficiency of processes defined i n 

the model, and examining the prediction performance 

of the model . A self-calibration subroutine 

will be included in the model whereb y 

the program will search for optimal model 

parameter values . Under this procedure each 

water year is used as a unit for optimizatio n 

and the objective function is to minimize th e 

variance between observed and compute d 

streamflow (Shih 1971). The sufficiency of 

processes defined in the model is reflected in 

the dispersion of parameter values resultin g 

from each year of calibration . After th e 

model is calibrated, those years of data whic h 

were not used ' for calibration are used to 

examine the confidence level of prediction s 

by the model . A flow diagram of the mode l 

verification procedure is shown by figure 5 . 

Model Parameters 

Model parameters are the coefficients use d 

in defining the processes which have not bee n 

accurately measured or which cannot b e 

directly measured. By establishing the value s 

of these coefficients the general model i s 

fitted to the hydrologic system of a specifi c 

watershed . Depending upon the resolution of 

the model and the availability of data, the 

number of parameters to be calibrated may 

vary. In order to avoid using a large numbe r 

of degrees of freedom in the calibration 

process and to save computation time, the 

number of model parameters should be kept 

as few as possible. In this study, the preliminary 

model parameters to be calibrated are 

interception storage capacity (SI), snowmelt 

coefficient (Ks), soil moisture retentio n 

capacity (Mcs), gravity water depletion coefficient 

(Kg), and groundwater recession coefficient 

(Kb) . 

59

(AVAILABLE DAT A 

WATER YEA R 

HYDROLOGI C 

DATA 

CALCULATE WATERSHED COEFFICIENT S 

ESTIMATE THE INITIAL VALUE AN D 

LIMIT OF MODEL PARAMETERS SE T 

UP PARAMETER VALUE CONSTRAINT S 

ENTIRE WATERSHED AS SINGLE UNIT 

WATERSHED HYDROLOGI C 

SIMULATION MODE L 

1 

CALIBRATE THE 

MODEL PARAMETERS 

CHECK ACCURACY AN D 

PARAMETER VALUE DISPERSION 

IMPROV E 

MODE L 

DAILY TIME INCREMENT 

CONVERT MODEL COEFFICIENTS 

AND PARAMETERS FOR DAIL Y 

TIME INCREMEN T 

CALCULATE AND ESTIMAT E 

SUBWATERSHED COEFFICIENT S 

AND PARAMETERS 

CALIBRATION, RESULTS STUD Y 

AND COMPARISON WITH 

MONTHLY MODE L 

NO Ị 

IMPROV E 

DEFINITION 

OF PROCESSE S 

SUBWATERSHE D 

MODEL 

NO 

SENSITIVIT Y 

ANALYSIS 

APPLICATION 

Figure 5 . Flow diagram for verification procedure . 

I 

6 0

Calibration of Parameter s 

In general, it is anticipated that realisti c 

parameter values are established through th e 

calibration procedure . However, when streamflow 

is the only available component for 

checking the model, it is possible that several 

combinations of parameter values will yiel d 

satisfactory agreement between observed an d 

computed outflow hydrographs . The proble m 

of establishing unique parameter values i s 

approached on the basis of hydrologic judgment 

and by using "interior" observations , 

such as snow depth and soil moisture, a s 

check points on model performance . Other 

ways of testing the model include the tim e 

distribution of output quantities, such a s 

stream flow, and known (or estimated ) 

monthly or annual quantities . For example , 

for watershed 2 it is estimated that interception 

storage is about 0 .5 cm, and total interception 

amounts to approximately 17 percen t 

of the annual precipitation . 

Under the self-calibration technique model 

parameter values are altered or purturbed in a 

random sequence and the resulting changes in 

the objective function are examined (Shi h 

1971). A computer flow chart for the calibration 

subroutine is shown by figure 6 . The 

entire program model, including the calibration 

subroutine, will be synthesized on a 

hybrid computer. 

Sensitivity and Management Studies 

Sensitivity 

A sensitivity analysis is performed b y 

changing one system variable while holdin g 

the remaining variables constant and notin g 

the changes in the model output functions . If 

small changes in a particular system parameter 

induce large changes in the output or respons e 

function, the system is said to be sensitive t o 

that parameter . Thus, through sensitivity 

analyses it is possible to establish the relative 

importance with respect to system respons e 

of various system processes and input functions. 

This kind of information is useful fro m 

the standpoint of system management, syste m 

modeling, and the assignment of priorities i n 

the collection of field data. Under this study, 

the verified model will be used to perfor m 

various sensitivity analyses for the hydrologi c 

system of watershed 2 . 

Management 

Opportunities for management of a fores t 

watershed are widely varied, and range fro m 

changes in logging practices to forms of soil 

treatment. Actual implementation of a 

management scheme depends upon benefit s 

gained as compared with possible disbenefits . 

The simulation model developed under thi s 

study will not make direct comparisons of 

benefits and disadvantages, but will predict 

changes in the system output associated wit h 

given management alternatives . Under this 

study the capability of the model will be 

demonstrated for rapidly testing many 

possible management alternatives . 

Much of the work discussed by this paper i s 

based upon past developments in watershed 

simulation at Utah State University . Hydro - 

logic modeling of the H . J. Andrews Experimental 

Forest for Coniferous Forest Biom e 

has just begun, and the preceding discussio n 

has been influenced by a consideration o f 

particular conditions in the study area . How - 

ever, the model will be fundamental in 

concept, and therefore generally applicable in 

a geographic sense . Whenever feasible, th e 

model will use basic equations which are valid 

for short time intervals to define the variou s 

processes in the model . The output will the n 

be summed for application to daily or longe r 

time increments . For example, Horton 's infiltration 

equation is applicable in minute units , 

but by summing these quantities, the model i s 

capable of calculating equivalent daily infiltration 

rates . For the area under this study, how - 

ever, it is assumed that water supply rates a t 

the ground surface do not exceed infiltratio n 

capacities, so that surface runoff does not 

occur, thus considerably simplifying th e 

model calibration process . 

System functions which are important t o 

forest management and other aspects of th e 

total project include evapotranspiration and 

soil moisture . These functions, along wit h 

streamflow, will be estimated by the mode l 

for use in other parts of the total system 

61

SET UP PARAMETER OPTIMIZIN G 

ORDER IN RANDOM SEQUENC E 

CHANGE THE PERTURBATION DIRECTION 

WITHIN GIVEN PARAMETER LIMI T 

ASSIGN OPPOSITE DIRECTIO N 

OF PERTURBATION I N 

NEXT CYCLE 

PARAMETERS TESTE D 

WITHOUT SIGNIFICANT 

IMPROVEMENT 

L = L + 1 

RESULT S 

FROM THIS S E 

OF PARAMETER VALU E 

SATISFY TH E 

CONSTRAINTS 

i RETURN 

Figure 6 . Flow chart for the model calibration . 

I 

E2

model being developed under the Coniferou s 

Biome Program. The important underlying 

feature throughout the entire study will b e 

that all of the separately described hydrologi c 

processes and phenomena are interlinked into 

a total system . Thus, from the model, hope - 

fully, it will be possible to evaluate the relative 

importance of the various items, explore 

critical areas where data and perhaps theor y 

are lacking, and finally establish guidelines fo r 

the improved management of forest water - 

sheds . 

INPU T 

(lOVO 

(A) 

SYSTEM 

(KNOWN ) 

SYSTEM SYNTHESI S 

OUTPUT 

^-TI €N NNowN 1 

Hydrologic Systems Analysis 3 

INPUT 

(KNOWN) 

SYSTE M 

(UNKNOWN) 

OUTPU 

T (KNOWN ) 

The purpose of this research is to devise a 

technique for statistical decomposition of a 

hydrologic event such that system processe s 

such as precipitation, subsurface flow, an d 

evapotranspiration, which contribute to th e 

observed streamflow can be separated and 

described. This technique will therefore pro - 

vide one more avenue for determination of 

the subsurface flow process on forest soils . 

The technique chosen for this research is a 

form of systems analysis . 

Systems, Definitions and Basic Principles 

System may be defined as an aggregate of 

physical parts that do not change with time , 

operating on an input to produce an output , 

both being functions of time . The simplified 

representation of a watershed, given in figure 

2, can be considered as a system whose input 

is precipitation and runoff its output . Syste m 

"synthesis" is a technique employed when th e 

system is known in terms of a mathematical 

equation; the objective is to determine th e 

nature of the output for any class of inpu t 

(fig. 7) . In system "analysis" a syste m 

response function or kernel which best de - 

scribes a given input-output pair is derived 

(fig. 7) . The term "best" implies that the 

derived kernels are not unique . Combination s 

of both techniques can be used for the solution 

of hydrologic problems . A system can 

3 Authored by Z. G. Papazafiriou, Research 

Associate, and R . H . Burgy, Professor, University o f 

California, Davis . 

(C) 

(B) 

SYSTEM ANALYSI S 

MULTI-INPUT/OUTPUT SYSTEM 

Figure 7 . Illustration of systems . 

have one input and one output or many in - 

puts and outputs (fig. 7). A system i s 

"lumped parameter" if input and output ar e 

functions of a single variable . Otherwise, th e 

system is of the "distributed parameter" type . 

If the system response at any time, due to a 

given input, is uniquely determined, th e 

system is said to be "deterministic." If the 

system response is subject to uncertain influences, 

the system is "stochastic o r 

probabilistic . " 

A quantity z is a "functional" for the function 

x(t) in the interval (a,b), if it depend s 

upon all values taken by x(t), when t varies i n 

the interval (a,b) . An illustration of a functional 

is given in figure 8 . The output of a 

system is a functional of the input and, fo r 

the same reason, runoff is a functional o f 

precipitation. A system is "time invariant " if 

it does not change with time. Such system s 

can be represented by functionals . "Physically 

realizable" is a system whose output at time t 

depends only upon past values of the input . 

63

can be almost linear, but there is no linea r 

system which can be almost nonlinear . In 

general, linearity is a limiting case of non - 

linearity. Therefore, any theory or technique 

adequate for a general nonlinear system i s 

equally adequate for linear systems . 

Deterministic Linear Hydrologic Systems 

The theory behind most linear methods ca n 

be generalized in the following manner . Suppose 

that s and a are continuous variable s 

representing position in space, and t and r 

define position in time . Consider the linear 

P. D. E . of the general form 

L[g(s,t)] = f(s,t) (17) 

Figure 8 . Demonstration of a functional . 

Hydrologic systems are physically realizabl e 

where L is linear P . D. operator of arbitrary 

since their outputs (runoff) at time t depen d 

order, and g(s,t) some function which satisfie s 

only on the past values of their inputs (precipitation) 

. The "memory" of a system is th e 

equation 17 within a certain region R . Given 

the appropriate homogeneous boundary conditions 

along R, the solution of equation 1 7 

time period between some past time and th e 

present for which the output depends only 

can be written according to Hildebran d 

upon the input . If the output depends only 

(1958) as 

on the present value of the input, the syste m 

is said to be a "no-memory" system . If the 

g(s,t) =f f G(s,t ; a,r) f (a,r) da dr (18) 

output of a time invariant system is analyti c 

R 

about zero input at some time to, the syste m If 

is "analytic". Analyticity is very important , 

f(s,t) = fs (s) ft (t) (19) 

since if a system is analytic, its output can b e 

equation 18 can be written in the for m 

expanded in Volterra series . Hydrologic 

systems are assumed to be analytic. 

g(s,t) =f f(r) [f G(s,t ; a,r) do] dr (20) 

A deterministic system H is said to be 

s 

"linear," if given the inputs X 1 (t) and X 2 (t) 

such that 

If fs (s) is spatially invariant, we may write 

y 1 (t) = H [AX 1(t)] (1 4) 

f 

ti ti 

g(s,t) = G (s,t ; r) f (r) dr (21) 

y 2 (t) = H [BX 2 (t)] (15) 

implies that 

Y1 (t) + Y2 (t) = Y[ X (t)] ( 16) 

= H [AX 1 (t) + BX 2 (t) ] 

= AH [X 1 (t)] + BH [X 2 (t)] 

that is, in a linear system, each member of a 

sequence of input values influences the out - 

put independently of every other . This is the 

well known principle of superposition. If a 

system does not satisfy the above condition i t 

is said to be "nonlinear . " A nonlinear system 

where 

ti ti 

f (r) = f(s,t), and G(s,t ; r) (22) 

= fG(s,t ; a,r) da 

s 

We can write equation 21 in differentia l 

equation form as 

d 

An(s,t) 

dtn't) + An 1(s,t) dn-lg(s,t) + 

dt n- 1 

+ A 0(s,t) g(s,t) = f(s,t) (23) 

64

Usually, we are interested in the output variable 

at some particular point in space ( a 

particular gaging station), in which case equation 

23 becomes 

An(t) dng(t) 

+ An-1(t) 

do-1g(t) 

+ 

dtn dt n-1 

▪ + A0(t) g(t) = f(t) 

(24) 

which describes a spatially lumped parameter , 

time-varying linear system. If we assume that 

the parameters in equation 24 are time in - 

variant, we obtai n 

An ddgnt) + An-1 do-lg(t) + 

▪ 

Aog(t) = f(t) 

dtn- 1 

(25) 

which describes a time invariant, lumpe d 

parameter linear system . If we assume that 

the system is completely at rest at t=0, we can 

write equation 25 in the form of the con - 

volution equation 

t 

g(t) = f h(r) f(t-r) dr (26 ) 

0 

which is the basis of the unit hydrograp h 

theory and many other hydrologic techniques. 

In equation 26, g(t) represents runoff, 

f(t) rainfall, and h(t) is the kernel, or in thi s 

case, the unit hydrograph . 

In summary, application of equation 26 

implies that the watershed behaves as a linear 

system, it is time invariant, the rainfall is 

uniformly distributed over the watershed 

area, and the watershed is completely at res t 

at the beginning of the rain . The "effective" 

precipitation is used as an input to th e 

system. This implies that we know some 

method for the separation of the runof f 

hydrograph into base flow and direct runoff 

(fig. 9) . This approach in fact is a combination, 

using system synthesis for the estimation 

of the effective precipitation, and syste m 

analysis for the estimation of runoff. Methods 

using this technique have been developed b y 

Snyder (1955), Eagleson et al. (1966), Nash 

(1957, 1960), O'Donnell (1960), Dooge 

(1965), and others . 

Deterministic Nonlinear Hydrologic System s 

As it was noted in a previous section, a 

time invariant analytic system can be expanded 

in Volterra series . Such an expansion 

can be written in the form 

y(t)=ho+ f h 1( T1) x ( t-r1) dr 1 

+ f f h 2 (T1 ,T2) X (t-T1) X(t-T 2 )dr 1 dr 2 

+f f hn(T 1 Tn) x(t-T 1 ) ... 

x(t-Tn) dr 1 

. . . drn 

+ (27) 

EFFECTIVE RAINFALL = DIRECT RUNOF F 

xUI 

Y(T) 

BASE FLOW 

m.'UtAut%A%atU 

Figure 9 . Estimation of effective rainfall through hydrograph separation . 

65

where hi are the kernels of the system an d 

ho=0 unless a source or sink is present. If th e 

system is physically realizable, and has finite 

memory (as it happens with hydrologi c 

systems), equation 27 can be written in th e 

form u 

r y ( t) = J h 1 (r 1 ) x(t-r1) dr 1 

U 

0 

u 

+ f ,/ 3 r h 2 ( T 1 ,T2) X4-T 1 ) X(t-T2 )dr1 dr2 

0 0 

+ . . . . . . . . . . . . . . 

+ 

u 

f . . . 

u 

f hn (T 1 . . .Tn)x(t-T1) . . . 

0 0 

x(t-T n )dT i 

. . . dTn 

subject to the condition 

h i(t) = 0 for all r < 0 

(28) 

and where u is the length of the memory . If , 

instead of continuous functions, discrete set s 

of data are used, equation 28 can be written 

in the form 

U 

y(T) = H 1(S1) X(T-S 1 ) 

S 1 = 0 

U U 

+ E E H 2 (S1,S2) X(T-S1 ) 

S 1 =0 S 2 =0 X(T-S 2 ) 

U U 

+ E . . . E H n(S I . .. S n ) X(T-S 1 ) 

Si =0 Sn = O X(T-S n ) (29) 

The first term in equation 28 or 29 represents 

a linear system, that is an ordinary con - 

volution integral. The other terms are a 

generalization of the convolution integral . In 

general, the i th term represents a pure subsystem 

of order i. Therefore our system y(t) i s 

composed of the summation of a linear and a 

series of nonlinear subsystems . If we represent 

the successive terms of the system by H i (t) , 

H 2 (t), . . ., Hn(t), . . ., respectively, the system 

can be written in the for m 

y(t) = H1(t) + H 2 (t) + . . .. + Hn(t) + . . . . (30) 

An illustration of a nonlinear system is give n 

in figure 10 . 

A system is identified whenever its kernel s 

hn(t) are calculated. This evaluation is base d 

on the past behavior of the system. Theoretically, 

after knowing all the kernels, th e 

response of the system can be calculated fo r 

any given set of input values . 

Equation 28 (in the form given) is very 

generalized and its usage is limited and tim e 

consuming . It is desirable to reduce the operations 

involved . For example, given a sequence 

of inputs, sometimes sequential values ar e 

highly correlated, and the system itself may 

smooth out rapid fluctuations . Hydrologic 

systems demonstrate both of these characteristics 

. Thus, we may look for ways whic h 

can describe the sequence by a smalle r 

number of parameters . If the sequence x 1 . . . . 

Xi . . . . Xn represents observations at times 

t1 . . ..tj . . . .tn, respectively, they can be approximated 

by a polynomial of the for m 

k 

X k(t) = aa +alt+ . ..+aktk= a m tm (31) 

m= 0 

where k

LINEAR 

Y~(T) 

2nd ORDER 

Y2 (T) 

3rd ORDER 

Y 3 (T) 

INPUT 

X (T) 

OUTPUT 

Y (T ) 

} 

nth ORDER 

Yi (T ) 

Figure 10 . Illustration of nonlinear system representation by Volterra Series. 

the memory, an exact match could be made 

to each of the input values . However, n o 

economy in the description would have bee n 

affected . Instead, fewer polynomials can b e 

used. This introduces an error, but due to th e 

nature of the operation, it will be the leas t 

error possible . 

Methods using the above procedure hav e 

been introduced by Jacobi (1966), Harder 

and Zand (1969), and Brandstetter an d 

Amorocho (1970) . It is this method which is 

being considered for use in this study . There 

are certain distinct advantages associated with 

the process . The method is quite general and 

can handle a variety of problems such as runoff, 

chemical quality of runoff waters, and 

suspended sediment predictions, when inpu t 

values are properly weighted by functions 

describing the physical processes involved in 

each case. Once the response functions or 

kernels of the system are evaluated, they may 

be used for predictions given any sequence o f 

inputs. It requires a minimum amount of 

data, possibly 1 to 2 years of good records . 

Systems can be used in cascade, lik e 

predict predict 

precip.-runoff quality or sediment 

It can be used for quantitative evaluations o f 

the changes created by any type of watershe d 

management procedure by evaluating th e 

kernels of the original and the manage d 

hydrologic system. Emphasis will be given to 

establishing weighted functions for the bes t 

description of the physical process, and o n 

deriving tools for the greatest economizatio n 

of the procedure . 

67

Watershed Stratification : 

A Problem on Watershed 10 4 

W e have also begun to structure the 

hydrologic model for watershed 10 concurrently 

with the simulation study underwa y 

for watershed 2 and the watershed system s 

analysis. Our objective here is to prepare a 

fine-resolution hydrologic model which in - 

corporates the transpiration model from th e 

Primary Producers, the evapotranspiratio n 

model from Meteorology, and provides soi l 

moisture and subsurface water flow for th e 

Primary Producer and Bio-Geochemical Processes 

groups . The first step in structuring suc h 

a model is system stratification . 

One key to the analysis of complex system s 

is the compartmentalization of the syste m 

into homogeneous subsystems which can the n 

4 Authored by George W . Brown, Associate Professor, 

Oregon State University, Corvallis . 

be isolated for study . This should be done in 

such a way that the linkage between compartments 

is simple and direct . 

Another key is placement of compartmen t 

boundaries in such a way that the cells ar e 

easily uncoupled, or are coupled as a simpl e 

linear cascade . To do otherwise would complicate 

the modeling considerably. A hydrologic 

model that consists of a series of compartments 

arranged as a branching cascade is 

extremely difficult to manage . Water flow 

from an upper compartment must be some - 

how divided between lower compartments in 

the cascade. The basis for such division i s 

generally obscure and usually arbitrary . 

In our attempt to structure the hydrologi c 

model for watershed 10 at the next level o f 

resolution, we began by setting compartmen t 

boundaries along stream courses . This automatically 

decoupled the compartments, sinc e 

water does not cross the channel (fig . 11) . 

Next, it was necessary to consider the arrangement 

of the plant communities within 

Figure 11. Initial and secondary stratification of watershed 10, H . J. Andrews Experimental Forest, Oregon . 

68

the watershed. The hydrologic model and the 

primary producer model are obviously linked . 

The principal coupling variable betwee n 

models is the soil moisture profile . Soil moisture 

is that portion of the "hydrologic state " 

of the watershed which closely regulates plant 

growth. Plants, in turn, influence soil moisture 

by transpiration . Thus, superimpositio n 

of the primary producer 's vegetative structure 

upon the structure of the hydrologic system i s 

essential . This structure was combined wit h 

the initial hydrologic stratification and de - 

fined the subcompartment boundaries . Sub - 

compartment boundaries approximate th e 

vegetative type-map boundaries and are arranged 

into riparian, midslope and ridgeto p 

zones. These zones undoubtedly reflect th e 

changes in soil moisture regime within th e 

watershed . Also, this stratification allows u s 

to consider flow between subcompartments a s 

simple linear cascades . 

This final stratification for watershed 1 0 

will provide the basis for sampling schemes t o 

characterize soil moisture, water flow and 

other hydrologic and biologic processes necessary 

for the next round of model construction. 

It is essential to note that this stratification 

is compatible for modeling hydrologi c 

processes and is also compatible for linking 

the hydrologic model with that of the primary 

producers . It is the major achievement 

of our initial modeling effort and sets the 

stage for continued progress . 



in part by National Science Foundation 

Grant No . GB-20963 to the Coniferou s 

Forest Biome, U .S . Analysis of Ecosystems , 

International Biological Program . This is Contribution 

No. 24 to the Coniferous Forest 

Biome . 


Amorocho, J., and B . Espildora . 1966 . Mathematical 

simulation of the snow melting 

processes. Univ. Calif. Water Sci. & Eng . 

Pap. No. 3001, 156 p . Davis . 

Anderson, E. A., and N. H . Crawford . 1964 . 

The synthesis of continuous snowmelt run - 

off hydrographs on a digital computer . 

Stanford Univ ., Dep. Civil Eng. Tech. Rep . 

No. 36, 103 p . Palo Alto, Calif . 

Barnett, L. O . 1963. Storm runoff characteristics 

of three small watersheds in wester n 

Oregon. 84 p. M .S. thesis, on file at Colo . 

State Univ., Fort Collins . 

Brandstetter, A ., and J. Amorocho. 1970 . 

Generalized analysis of small watershed 

responses. Univ. Calif. Dep. Water Sci . & 

Eng . Pap. No. 1035, 204 p . Davis . 

Crawford, N. H., and R. K. Linsely . 1964 . 

Digital simulation in hydrology : Stanfor d 

Watershed Model IV . Stanford Univ . Dep . 

Civil Eng . Tech. Rep. No. 39, 210 p . Palo 

Alto, Calif . 

Dooge, J . C . I. 1965. Analysis of linear systems 

by means of Laguerre functions . J . 

Soc . Ind . Appl. Math. Cont., A. 2 : 

396-408 . 

Dyrness, C . T. 1969 . Hydrologic properties o f 

soils on three small watersheds in th e 

western Cascades of Oregon . USDA Forest 

Serv. Res. Note PNW-111, 17 p . Pac . 

Northwest Forest & Range Exp . Stn., Portland, 

Oreg . 

Eagleson, P . S., R. Mejia, and F . March . 1966 . 

The computation of optimum realizabl e 

unit hydrographs. Water Resour . Res. 2(4) : 

755-765 . 

Eggleston, Keith O ., Eugene K . Israelsen, and 

J . Paul Riley . 1971 . Hybrid computer simulation 

of the accumulation and melt processes 

in a snowpack . Utah State Univ . Coll . 

Eng., Utah Water Res. Lab., Pap. No . 

PRWG 65-1, 77 p . Logan, Utah . 

Forsythe, G . E . 1957 . Generation and use of 

orthogonal polynomials for data-fittin g 

with a digital computer . J . Soc. Ind. Appl . 

Math. 5(9) : 74 . 

Harder, J . A., and S . Zand . 1969 . The identification 

of nonlinear hydrologic systems . 

Hydraul . Eng. Lab. Tech . Rep . HEL8-2, 8 3 

p. Univ. Calif., Berkeley . 

Hildebrand, F . B. 1958. Methods of applie d 

mathematics . 523 p. New York : Prentice - 

Hall . 

69

Jacobi, S . L. S. 1966. A mathematical mode l 

for nonlinear hydrologic systems . J. Geol . 

Res. 71 : 4811 . 

Jones, A. 1971 . Soil piping and stream channel 

initiation. Water Resour . Res. 7(3) : 

602-610 . 

Nash, J. E . 1957 . The form of the instantaneous 

unit hydrograph. Int. Assoc. Sci . 

Hydrol . Gen. Assem. Toronto, Proc ., 3 : 

114-121 . 

. 1960. A unit hydrograph study 

with particular reference to British catchments 

. Inst . Civil. Eng. (Proc.) 17 : 

249-282 . 

O ' Donnell, T . 1960. Instantaneous unit 

hydrograph derivation by harmonic analysis. 

Assoc . Sci. Hydrol. Publ. No. 51, p . 

546-557 . 

Pysklywec, D . W., K. S. Davar, D . I. Bray . 

1968 . Snowmelt at an index plot. Water 

Resour. Res . 45 : 937-946 . 

Riley, J. P ., D . G . Chadwick, and J . M . 

Bagley. 1966. Application of electroni c 

analog computer to solution of hydrologi c 

and river-basin-planning problem : Utah 

Simulation Model II . Utah State Univ . , 

Utah Water Res . Lab. PRWG 32-1, 121 p . 

Logan, Utah . 

Rothacher, Jack. 1963 . Net precipitation 

under a Douglas-fir (Pseudotsuga menziesii) 

forest. Forest Sci . 9 : 4 . 

, C . T. Dyrness, and Richard L . 

Fredriksen . 1967. Hydrologic and related 

characteristics of three small watersheds in 

the Oregon Cascades . USDA Forest Serv . 


54 p. Portland, Oreg. 

Shih, C . C. 1971. A simulation model for predicting 

the effects of weather modificatio n 

on runoff characteristics . 116 p . Ph.D . 

thesis, on file at Utah State Univ ., Logan . 

Sittner, W. T., C. E. Schauss, and J . C . Monro . 

1969. Continuous hydrograph synthesi s 

with an API-type hydrologic model . Water 

Resour . Res. 5(10) : 1007-1022 . 

Snyder, W. M. 1955. Hydrograph analysis b y 

the method of least squares . Am . Soc . Civil 

Eng. Proc., J. Hydraul . Div. 81(793), 25 p . 

U .S. Army Corps of Engineers . 1956 . Summary 

report of snow investigations, sno w 

hydrology . 437 p . Portland, Oreg. 

U .S . Department of Agriculture . 1964 . Irrigation 

water requirements . Soil Conserv . 

Serv. Eng . Div . Tech. Release No . 21 . Washington, 

D .C . 

Veihmeyer, F . J. 1964 . Evapotranspiration . In 

Ven to Chow (ed.), Handbook of applie d 

hydrology, p . 11-1 to 11-23 . New York : 

McGraw-Hill . 

70


Bellingham, Washington-March 23-24, 1972 . 

Preliminary consideration s 

of the forest canopy 

consumer subsystem 

M. A. Strand and W . P . Nagel 


Corvallis, Orego n 

Abstract 

A food chain beginning with herbaceous materials produced in the forest canopy is delineated as a 

subsystem. A conceptual model of energy flow through this subsystem includes assumptions and definitions 

relating to components of the canopy food chain, processes by which energy is transferred, and the energ y 

pathways. From this conceptual model, it is observed that the canopy food chain is directly coupled to othe r 

consumer subsystems through common predators and through the detritus component . Besides the direct 

relationship through the food base to primary production, this subsystem may also influence the plant syste m 

by its effects on the sites ofplant hormone production and on the fate of mobile nutrients in the ecosystem . 

Introduction 

In the past, modeling efforts concernin g 

consumers have been centered on populatio n 

dynamics, while more recently energetics hav e 

been receiving attention . Our consideration s 

of consumers will also concern energy transfer. 

We will focus on the couplings betwee n 

consumers and other parts of the ecosystem . 

These couplings may be direct (i .e., involve 

energy transfer to or from a consumer) o r 

indirect (i .e., involve the influence of on e 

component on energy transfer between othe r 

components of the ecosystem) . The majo r 

subdivisions of the ecosystem are discussed b y 

Overton (1972) . 

To facilitate our understanding of th e 

consumer system, it would be convenient t o 

subdivide it into a hierarchical set of units an d 

observe the couplings between them and their 

couplings with other ecosystem components . 

If our units are to be subsystems, then th e 

most logical divisions would occur aroun d 

groups of components with high degrees o f 

interconnections. The majority of linkage s 

with a subsystem's components should occu r 

within the subsystem ; couplings to components 

of other subsystems should be comparatively 

rare. The classic way to defin e 

consumer subsystems has been by trophi c 

levels; however, for our purpose this divisio n 

may not be the most meaningful one . By 

basing the division of the consumer system o n 

trophic level designations, the direct linkage s 

with regard to energy flow occur between th e 

subsystems and only indirect linkages, such a s 

competition, occur within them . We hypothesize 

that a better subdivision of the consumer 

system could be made by defining th e 

subsystems as food chains on the basis of a 

stratification of their primary food base . 

The first division in our hierarchy will b e 

into grazing and detritus food chains . Th e 

grazing food chain includes primary consumers 

feeding on plant material and a serie s 

of secondary consumers that prey on the 

herbivores and on each other . To further 

subdivide this food chain, the food base i s 

stratified into forest canopy herbaceous material, 

woody material, forest floor herbaceou s 

material, and roots . Food chains are define d 

by the food material eaten by the primar y 

consumers in the chain . Likewise, the detritu s 

food chain which recycles detritus to lowe r 

71

grades of detritus is divided according t o 

particle size. The two resulting food chain s 

use fine or coarse detritus as their primar y 

food sources. By uncoupling these six foo d 

chains (four grazing and two detritus) with 

regard to their food bases, the first orde r 

consumers are reasonably unique to a particular 

food chain subsystem, while commo n 

predators couple the subsystems together . 

The purposes of this paper will be to defin e 

the forest canopy food chain and to develop a 

conceptual model of it as an example of th e 

type of preliminary consideration which is a 

prerequisite to mathematical modeling. 

The forest canopy food chain forms a 

subsystem that begins with herbaceous material 

produced by trees and follows it throug h 

its many transfers to detritus . The food bas e 

is primarily photosynthetic tissue, needles an d 

leaves of trees ; however, buds, young twigs , 

immature cones, and seeds will also be included, 

since consumption of these foods i s 

an interrelated process . For present considerations, 

we will confine our concern to an 

old-growth, coniferous forest canopy . Successional 

changes, epidemic outbreaks, an d 

environmental gradients will not be include d 

in our discussions . It is assumed that variations 

in energy flow from year to year are no t 

appreciable, so there are periods of short-ter m 

stability . The conceptual model we will pro - 

pose is a description of this type of stabl e 

condition . 

Our discussion will consist of four parts : 

the components of the subsystem, the processes 

involved in energy transfer, energy path - 

ways, and the interconnections between thi s 

and other food chain subsystems as well a s 

the subsystems defined by Overton (1972) i n 

this symposium . In each section we wil l 

present a set of assumptions and definition s 

that represent our notion of this subsystem . 

The set will form our conceptual model . 

Components 

The components of a subsystem are the 

locations of energy storage and vehicles o f 

energy transfer along the food chain . Components 

are defined by ecological function 

rather than taxonomic criteria . However great 

species diversity might be in a forest canopy , 

functional roles are common to many species , 

enabling them to be regarded as a few larg e 

populations. To avoid the pitfall of over - 

generalization about the superpopulations, a 

brief discussion of the variations of behavio r 

relevent to energy transfer of our components 

is included . 

Primary consumers are the key links in th e 

food chains, diverting energy produced by th e 

plants into the animal system . In the fores t 

canopy most consumers are insect grazers 

which feed mainly on new tissue from expanding 

needles or leaves (Keen 1952) . The 

feeding stage either hatches or emerges near a 

feeding site, so that food finding initially i s 

not a problem . Many of them feed gregariously 

when larvae are small, but disperse t o 

"feed singly later. The main disseminatio n 

phase, however, is the adult where female s 

seek out new feeding sites for oviposition . 

The feeding period occurs primarily in th e 

spring and early summer with resting phase s 

(pupae, eggs, or overwintering larvae) beginning 

in late August or early September . A 

second group of primary consumers include s 

insects with sucking mouth parts . They act as 

physiological sinks and remove dissolve d 

nutrients from the xylem sap (Way an d 

Cammell 1970). Only one significant vertebrate 

foliage feeder has been reported in th e 

forest canopy-the red tree mouse (Phenacomys 

longicaudus) (Maser 1966). It eats 

needles of Douglas-fir (Pseudotsuga menziesii ) 

leaving the central xylem strand . Smal l 

branches are cut and carried to the nest where 

they are consumed . 

Plant reproductive tissues are also consumed. 

Immature cones are attached mainl y 

by cone and seed feeding insects ; some fee d 

exclusively on the seeds, while others feed o n 

the cone preventing seed development . Mature 

seed may be consumed prior to dissemination 

by certain birds and squirrels . Thes e 

omnivorous birds use seeds as a primary foo d 

source during the winter ; they remove undisseminated 

seeds while the cones are stil l 

attached to the tree (Isaac 1943) . Squirrel s 

cut cones and extract the seeds on th e 

ground ; the subsequent fate of the seeds i s 

72

considered as part of another food chai n 

subsystem which utilizes forest floor herbaceous 

material . 

Predators and parasites constitute the nex t 

links in the food chain . Small invertebrate s 

are consumed generally by other invertebrates 

. Parasitic forms may feed on larvae , 

pupae, or eggs of the host . Often several 

generations of parasites are produced for 

every single host generation ; hyperparasitis m 

of several degrees is possible. Predaceous 

invertebrates include both spiders and insects . 

Some, as the orb-spinning spiders, wait fo r 

mobile forms to be trapped, while other s 

move about actively in search of food an d 

feed on nearly any invertebrate they ca n 

capture (Graham 1952) . 

During the mating and nesting season , 

omnivorous birds feed on insects with eg g 

hatch often being correlated with peak insec t 

abundance (Lack 1954) . Very small larvae are 

usually not eaten since they are inconspicuou s 

and many would be required to satisfy a 

bird 's energy needs . Flycatchers, which ar e 

not abundant in the winter, breed in the 

forest canopy and feed mainly on adult 

insects . Other insectivorous birds search th e 

foliage for the larger larvae . Predation of adult 

birds is rare, but nest predators may be foun d 

consuming both eggs and young nestlings . 

Other predatory birds are also rare components 

of the forest canopy ; however, th e 

spotted owl (Strix occidentalis) is an important 

predator of vertebrate foliage feeder s 

(Nussbaum 1972 1 ) . 

The components of the forest canopy food 

chain are restricted to those animals that ca n 

complete at least the feeding stages of their 

life cycles above the shrub stratum. In our 

considerations of the forest canopy, we hav e 

restricted ourselves to endemic population 

conditions and have not considered the larg e 

numbers of defoliators found in some areas . 

The actual population densities found under 

"normal" circumstances is not known for th e 

old-growth stands, and estimates of thes e 

densities will be a subject of future research . 

We will consider herbivores to be predator 

limited and predators to be food limited . All 

' Personal communication . 

components will be assumed to be balanced 

with respect to immigration and emigration 

except for seasonally migrating birds . 

In summary, we have designated nine 

superpopulations or components in the forest 

canopy food chain subsystem : grazing insects , 

sucking insects, vertebrate foliage feeders , 

seed and cone insects, omnivorous birds , 

parasitic invertebrates, predaceous invertebrates, 

nest predators, and other predator y 

birds. Energy flow will be followed between 

these components ; possible transfers within 

them will not be considered explicitly . 

Processes 

Certain processes are essential for the transfer 

of energy, and we have defined five main 

processes for the food chain subsystems : 

consumption, elimination, respiration, assimilation, 

and death . The definitions we use may 

not be the same as those found in ecologica l 

literature, but our definitions are consistent 

with the purpose of defining direct couplings 

within and between the food chains and other 

subsystems of the ecosystem . 

The concept of consumption is often 

equated with ingestion ; however, the importance 

of consumers may not be measured by 

what they ingest alone . Shelter building an d 

wasteful feeding behavior may result in fa r 

greater death of the food resource than th e 

amount ingested. Hence consumption wil l 

include ingestion as well as other activities o f 

the animals that result in the loss of life o f 

their food. All forms of waste production b y 

animals after the food is ingested is include d 

in the process of elimination . Thus, energ y 

lost as a result of indigestion, metaboli c 

wastes, or losses of integument will be transferred 

to detritus via elimination. Assimilatio n 

is defined as individual secondary production . 

It is the process by which energy from on e 

trophic level is incorporated into the tissue of 

the next. The process of respiration include s 

the energy loss to the atmospheric sink as th e 

result of maintenance metabolism and o f 

work. Death includes mortality losses as a 

result of factors other than consumption b y 

another component of a food chain. Both the 

73

processes of death and elimination result i n 

the production of detritus, but they must b e 

considered separately since they differ in their 

consequences to the food chain components . 

Elimination does not affect the potentia l 

activity of a component, whereas, death may . 

The purpose of defining these five processes 

is to identify the means by whic h 

components are directly coupled with eac h 

other or with other subsystems of the ecosystem. 

Each one is influenced by indirec t 

couplings . For example, feeding behavior , 

population age structure, and energy demand s 

are indirect couplings between the feeding 

population and the process of consumption . 

Nutritional quality of the food may influenc e 

the processes of assimilation and elimination . 

These indirect couplings may be thought of as 

informational transfers and may be as impor - 

tant to the system as the direct couplings . 

Energy Pathways 

The general energy pathway is diagramme d 

in figure 1 . The boxes are used to show sites 

of energy storage . They are associated with a 

set of variables which are used to describe the 

site. The consumed tissue compartment is a 

hypothetical site but is included to separate 

the processes of consumption and assimilation 

. Circles are used to indicate the processe s 

which are employed to transfer energy between 

sites . Solid arrows mean true flows of 

materials or energy (direct couplings), whil e 

dotted arrows and diamonds indicate lines o f 

influence (indirect couplings) from an energ y 

storage site to a process . 

= Consumption 

WO 

0 

0 

= Waste (Subproces s 

of consumption 

= Assimilatio n 

Respiration 

= Death 

= Elimination 

4 

Food quality an d 

availabilit y 

Population ag e 

structure 

Food quality 

Food qualit y 

Population feeding an d 

sheltering behavio r 

Abiotic environmen t 

Figure 1 . Basic energy pathway through food chain subsystem . See text for explanation of symbols . 

74

PRIMARY PRRODUCTIO N 

■ 

FOLIAGE 

CONES AN D 

SEEDS 

r 

GRAZIN G 

VERTEBRATE S 

GRAZIN G 

INSECTS 

SUCKIN G 

INSECTS 

SEED AN D 

CONE INSECTS 

f 

V 

PREDACEOU S 

BIRDS 

PARASITI C 

INVERTEBRATES 

PREDACEOUS 

INVERTEBRATE S 

Figure 2 . The major energy pathways between components of the forest canopy food chain . 

The paths of energy (fig. 2) may be 

followed beginning in the spring when th e 

buds of the conifers swell and insect herbivores 

emerge and feed . While they are small , 

the grazing and sucking insects are preyed 

upon by other invertebrates . Later they are 

preyed upon by birds that switch from 

conifer seeds to insect larvae during thei r 

breeding season . Energy leaves the system in 

the form of food for predators in other foo d 

chains, migrating birds, detritus, and heat . 

The amounts of energy following these routes 

is not known yet . Consumption rates, relative 

densities of consumers, food quality variability, 

assimilation efficiencies, an d 

migration habits are also not known for mos t 

of the consumers . These problems will require 

further research before the model may b e 

quantified. The two diagrams and the definitions 

and assumptions that were outlined i n 

the preceding sections represent our curren t 

conceptual model of the canopy food chain . 

Interconnections with 

Other Subsystem s 

The conceptual model presented here of 

the forest canopy consumer subsystem ma y 

be used to describe possible interrelation s 

between it and other subsystems. By definition, 

this subsystem contains consumers tha t 

feed primarily in the forest canopy ; therefore , 

we would not expect it to be tightly coupled 

to other food chains . However, since the food 

base defines the subsystem and, in the case of 

the canopy food chain, herbivores are mor e 

specific in food habits than the carnivores, it 

is evident that direct couplings will probabl y 

result from the feeding behavior of th e 

secondary consumers . For example, adul t 

insects from other subsystems may serve as 

food for flycatcher birds which live primaril y 

in the canopy . Predaceous invertebrates from 

the forest floor feed upon canopy specie s 

75

when they move to the floor for overwintering 

or pupation. The food chain with primary 

consumers that feed on fine detritus i s 

coupled to this subsystem through the detritus 

output, and the canopy food chain ma y 

influence the energy flow rates within th e 

detritus food chain by the quality of its 

detrital outputs . 

Besides these connections with other foo d 

chain subsystems, the canopy consumers ar e 

closely linked to the primary productio n 

system . The consumers are influenced by th e 

quantity and quality of herbaceous materia l 

produced by the trees . However, the influence 

of the canopy food chain on primary production 

is more subtle. Although consumption o f 

foliage or cones is not considered to be at a 

rate to affect production in the old-growt h 

canopy of the model, the preferential consumption 

of new needles and expanding cone s 

may affect the nutrient capital of the trees . 

These consumption sites are physiologica l 

sinks, and mobile nutrients are actively moved 

to them from senescent foliage and storag e 

sites (Sweet and Wareing 1966) . The loss o f 

these sinks may mean the return of certain 

nutrients to the soil solution rather than thei r 

retention by the trees (Rafes 1970) . Also , 

consumption of bud tips reduces the sites o f 

growth hormone production ; this hormone i s 

necessary for cambial division, and loca l 

deficiencies may result from the gregariou s 

feeding habits of some herbivores (Kozlowsk i 

1969). These indirect couplings may prove t o 

be the most significant role of the food chai n 

subsystems in the old-growth forest . 

Summary 

The consumer system is composed of a 

series of grazing and detritus food chains . Th e 

primary consumers are thought to be reason - 

ably unique to each food chain, and direc t 

energy transfers between the food chain s 

occur via common predators . The processes of 

energy transfer that link the components of 

the food chains with each other and the rest 

of the ecosystem are : consumption, assimilation, 

elimination, respiration, and death. Indirect 

links are formed by informational transfers 

which influence energy flows . 

The forest canopy food chain includes the 

primary consumers that feed on herbaceous 

material produced in the canopy and the 

series of related predators . This food chain is 

directly linked to primary production by 

consumption of herbaceous materials, to th e 

other food chains by common predators, and 

to the detritus component through production 

of detrital material . Primary productio n 

influences the food chain processes by variations 

in nutritional quality, spatial arrangement, 

and quantity of material produced . Th e 

influences of the food chain on primar y 

production are more subtle . Preferential feeding 

habits of many canopy grazers affect th e 

number of sites of plant hormone productio n 

and the fate of mobile nutrients in th e 

ecosystem . 


The work reported in this paper wa s 

supported by National Science Foundatio n 

Grant No. GB-20963 to the Coniferous Fores t 


Biological Program . This is Contributio n 

No . 25 from the Coniferous Forest Biome . 

76

Graham, S. A. 1952. Forest entomology . 35 1 

p., illus. New York : McGraw-Hill Book Co . , 

Inc . 

Isaac, L. A. 1943. Reproductive habits of 

Douglas-fir . 107 p ., illus. Washington, D .C . : 

Charles Lathrop Pack For . Found . 

Keen, F . P. 1952. Insect enemies of western 

forests. U .S. Dep. Agric. Misc. Publ . 273 : 

75-126 . 

Kozlowski, T . T. 1969. Tree physiology ' an d 

forest pests . J. For. 67 : 118-122 . 

Lack, David . 1954. The natural regulation of 

animal numbers. 343 p., illus. Oxford , 

England: Clarendon Press . 

Maser, C . O. 1966 . Life histories and ecology 

of Phenacomys albipes, Phenacomys longicaudus, 

Phenacomys siluicola . M .S. thesis 

on file, Oreg . State Univ., Corvallis . 

Overton, W . S. 1972 . Toward a general model 

structure for a forest ecosystem . In Jerry F . 


Franklin, L. J. Dempster, and Richard H . 

Waring (eds .), Proceedings-research on 

coniferous forest ecosystems-a symposium, 

p . 37-47, illus . Pac . Northwest Forest 

& Range Exp . Stn., Portland, Oreg . 

Rafes, P. M. 1970. Estimation of the effect s 

of phytophagous insects on forest production 

. In D. E. Reichle (ed .), Analysis of 

temperate forest ecosystems, p . 100-106 . 

Berlin : Springer-Verlag . 

Sweet, G . B., and P. F. Wareing. 1966. Role 

of plant growth in regulating photosynthesis. 

Nature 210 : 77-79 . 

Way, M . J., and M. Cammell . 1970. Aggregation 

behavior in relation to food utilizatio n 

by aphids . In A . Watson (ed.), Animal 

populations in relation to their food re - 

sources, p. 229-247. Oxford, England : 

Blackwell Sci . Pub . 

77



An environmental grid fo r 

classifying coniferous 

forest ecosystems 

R . H . Waring, 

K . L . Reed, 

W. H . Emmingham 

School of Forestry 

Oregon State Universit y 


Abstract 

To develop models which will predict primary production and forest composition across the Biome, we 

suggest that the environment be defined operationally-as it affects certain plant responses . Models can be 

developed which indicate how water, nutrients, light, temperature, and mechanical factors interact through th e 

plant to control primary production and plant composition. Plant responses including water stress, stomatal 

behavior, foliar nutrition, and phenology, among others, were coupled to environmental variables to create plan t 

response indices for soil moisture, temperature, transpiration, and soil fertility . By correlating indicator species 

with plant response indices, an ecosystem can be defined environmentally without direct measurements . Other 

processes including biogeochemical cycling, consumer population dynamics, and hydrologic functions may als o 

be related to the environmental grid. 

Introduction 

Diversity in environment and vegetation i s 

characteristic of the Coniferous Forest Biome . 

Such diversity is esthetically pleasing but presents 

formidable problems of classification . 

Classification is necessary, however, if we ar e 

to understand the processes which affect the 

composition and allow for continued productivity 

of the forests and related water, wild - 

life, and recreational resources . 

If we could understand how the environment 

influences forest growth and composition, 

we would be well on the way toward 

reassessing our past management successe s 

and failures . We could also make our detailed 

ecosystem studies more widely applicable in 

both time and space, thus providing a frame - 

work for intelligent land management. Th e 

problem lies (a) in identifying which proper - 

ties of the environment should be measured , 

(b) coupling these to important biological 

processes, and (c) finally in predicting certai n 

key properties of ecosystems . 

In this paper we attempt to establish a conceptual 

basis for evaluating environment, then 

test the general approach, and finally suggest 

necessary modifications to permit extensio n 

across the Coniferous Forest Biome . 

General Theory 

The Operational Environment 

Excellent descriptions of vegetation and 

regional physiography exist for many region s 

of the world . In the Pacific Northwest the 

most thorough treatments have been provide d 

by Daubenmire (1952), Krajina (1965), and 

Franklin and Dyrness (1969) . Detailed 

physiographic classification such as that developed 

by Hills (1959) has proven most use- 

79

ful in regions with relatively uniform climate . 

Probably the most helpful classifications 

from a standpoint of predicting the nature o f 

forest ecosystems, however, have been thos e 

related to environmental gradients (Warming 

1909, Sukachev 1928, Pogrebnjak 1929 , 

Bakuzis 1961, Ellenberg 1950, 1956, Row e 

1956, Whittaker 1956, 1960, Loucks 1962 , 

Waring and Major 1964) . Unfortunately, al 

previously defined environmental gradient s 

cannot be directly applied outside the particular 

region where they were developed . We feel 

environmental gradients can be more widel y 

utilized only when the environment i s 

coupled to basic processes controlling plan t 

growth and composition . We believe a key t o 

expanding the gradient analysis approach is t o 

focus more closely upon the basic physiological 

behavior of plants . 

Fortunately a foundation for a process - 

oriented approach has already been laid . 

Mason and Langenheim (1957) defined th e 

idea of an "operational environment" as on e 

which directly affects an organism. Implicit i n 

their concept of environment is the recognition 

of specific plant responses to environmental 

stimuli. In fact, these authors felt tha t 

"environment" could not exist independently 

in an ecological sense . It must have an effec t 

upon an organism . Although they did not de - 

fine measurable environmental stimuli, thei r 

very definition of an "operational environment" 

focuses attention upon the influenc e 

rather than on the origin of the stimulus . 

Thus, in an interpretive ecological sense, i t 

is more important to assess the availability o f 

water to plant roots than the origin of tha t 

water. This distinction greatly reduces the 

number of factors requiring consideration, for 

although altitude, slope, and other physiographic 

features are correlated with vegetation, 

such indirectly operating factors may be 

ignored if the mode of action can be identified 

and measured . Ellenberg (1956), Bakuzis 

(1961), Waring and Major (1964), and other s 

have suggested that vegetation responds t o 

changes in water, temperature, light, and 

chemical and mechanical factors. Although 

these factors do not operate independently , 

they cannot completely substitute for one 

another . 

General Approac h 

We still have to translate the concept of an 

"operational environment" into a design adequate 

for research . As a first step we can diagram 

the flow of material through a plant into 

various compartments and identify the controls 

on the rate of flow imposed by the 

environment, the organism, or the amount o f 

material in a given compartment. Figure 1 

Figure 1 . General model of primary production . 

Material inputs of H 2 0, CO 2 . and energy flow int o 

the system to form carbohydrates . The rate of incorporation 

(photosynthesis) is a function of 

various plant responses. These responses are 

triggered by a host of environmental stimuli : 

temperature, light, humidity, soil water potential , 

soil fertility, and mechanical stress. The plant 

responses interact to provide two functions: (1) a 

control on carbohydrate storage and (2) a contro l 

on growth. Losses from the system are throug h 

respiration (R), death of roots and other organ s 

and in litter (L), and consumption by animals (C) . 

presents such a diagramatic model . Th e 

rectangles are the compartments, and the flo w 

of materials is indicated by solid lines, whil e 

dotted lines indicate the transfer of information 

through valves controlling the rate of 

material flow from one compartment to an - 

80

other. The plant integrates environmenta l 

stimuli. Arrows indicate the direction of flow , 

and circles are special functions regulated b y 

information from the plant integrator . 

We interpret environmental variables as 

being external to the system and impingin g 

upon it . The system, in this case the organism , 

responds only in accordance to the externa l 

stimuli. Modifications in the environmen t 

brought about by the organism itself are no t 

distinguished from those originating throug h 

other means . 

The amount of material transfer from environment 

to plant is determined by the kin d 

and magnitude of the response ; i .e., the 

amount of CO2 converted into carbohydrate 

is determined by the photosynthetic rate . The 

amount of carbohydrate utilized in th e 

growth of foliage, stem, and roots is also 

affected by the stage of plant development 

and other plant responses . As long as the carbohydrate 

pool does not become exhausted , 

or one of the plant responses is not death, 

biomass can accumulate . 

When the environment initiates a letha l 

plant response, the transfer of carbohydrate s 

is stopped and all of the biomass compartments 

empty . Biomass loss also occur s 

through respiration, consumption by animals, 

the normal fall of litter, and death of roots 

and twigs . 

Unfortunately, we do not, at present, have 

adequate knowledge for complete evaluatio n 

of such an energy flow model, even for a 

single species . We hope to acquire muc h 

needed data through the Biome program. Still 

it is safe to say that we will lack complete 

data on most species to predict accurately th e 

system 's response to external variables . Our 

approach, therefore, is to study physiologica l 

responses in relation to the environment , 

develop functional models, and from thes e 

create plant response indices as a means of 

interpreting environmental gradients . Further , 

we propose to correlate ecosystem propertie s 

with these plant response indices . Figure 2 

ENVIRONMENTA L 

STIMUL I 

Absolute Humidit y 

Radiation 

Air & Soil Temperature 

Soil Wate r 

Soil Fertility 

Wind and Snow 

PLANT RESPONSE 

INDICES 

ECOSYSTE M 

CHARACTERISTIC S 

Plant Moisture Stress 

Productivity 

Transpiratio n 

Compositio n 

PHYSIOLOGICAL 

CALIBRATION 

j---* Photosynthesi s _ 

Hydrologic Stat e 

PROCESS MODELS 

VALIDATION 

Nutritional 

Decomposition 

4 

PREDICTION 

Temperature Growth Consumer Population s 

Mechanical Stress 

/ 

/ PHYSIOLOGICA L 

/ BEHAVIO R 

Stomatal Respons e 

Cambial Activity 

Photosynthesi s 

Transpiration 

Plant Moisture Stres s 

Phenology 

Foliar Nutritio n 

Figure 2 . Physiological process models couple environmental stimuli to the physiological behavior of selecte d 

conifers . From an understanding of the process models, plant response indices are derived which reflec t 

major environmental gradients . These indices locate ecosystems within an environmental grid and throug h 

correlation permit certain ecosystem characteristics to be predicted . 

8 1

epresents the basic linkages between environmental 

stimuli, physiological behavior, an d 

the generation of plant response indices whic h 

are to be correlated with ecosyste m 

characteristics . 

The Organis m 

If we strive for both a predictive model and 

an understanding of the general processes con - 

trolling plant composition, we must compromise 

by linking observations of plant distribution 

and growth for many species to th e 

physiological behavior of certain trees . For i n 

this Biome, trees represent the dominant for m 

in which energy accumulates in forest ecosystems 

. 

The size of reference plants must also b e 

standardized . Although establishment is critical 

for determining the composition of an y 

plant community, young seedlings represen t 

both adapted and ill-adapted plants . Mature 

plants, on the other hand, are usually so well 

established that their responses only hint at 

the critical selective factors operating at th e 

time of establishment and early growth . As a 

compromise, in this Biome we study recently 

established conifers between 1-2 m tall . Few 

such trees will die during the course of measurements 

; yet they are small enough to 

reflect the major environmental forces actin g 

at the critical time of establishment . 

Having decided upon the important environmental 

factors and the size and kind of 

reference plant, we are still faced wit h 

coupling the environment to plant responses . 

The most commonly selected response has 

been growth in some form or another, bu t 

growth reflects the total integration of al l 

environmental influences and a host of different 

environmental combinations can resul t 

in similar productivity . We choose response s 

related to the entire plant, but those whic h 

may be more closely identified with specifi c 

environmental stimuli . 

Physiological Behavior 

Because plants are a complex system, no 

response is completely unrelated to another . 

Therefore most responses cannot be inter - 

preted without knowledge of others, and 

cause and effect relationships can rarely b e 

implied. The most important behavior 

includes : 

1. Phenological development . This reflects 

past environment and serves to define 

periods during which the sensitivity of 

the plant is similar. Lack of attention to 

phenological development is a majo r 

reason why otherwise good environmental 

data are often difficult to interpret 

ecologically (Azzi 1955) . 

2. Carbon dioxide exchange provides a 

ready means of quantifying the effect s 

of light and temperature upon photo - 

synthesis . 

3. Plant moisture stress (plant water potential) 

before dawn is an excellent measur e 

of how a reference plant responds to soil 

water status. 

4. Stomatal resistance reflects the influenc e 

of evaporative stress and soil drought, a s 

well as certain other environmental variables 

. 

5. Foliar nutrition indicates the interactio n 

between supply of nutrients in the soil 

and the demand for nutrients . 

Results of Past Studie s 

To help clarify the above discussion, we 

draw upon our experience during the last 8 

years in an environmentally and floristicall y 

diverse region of southwestern Oregon (Whit - 

taker 1960, 1961 ; Waring 1969) . 

At all or some of the 25 forest stand s 

described briefly in table 1, environmental 

data on air and soil temperature, radiation, 

humidity, soil moisture, and soil fertility were 

recorded . Physiological response studies o n 

two widely distributed conifers, Douglas-fi r 

(Pseudotsuga menziesii) and Shasta red fir 

(Abies magnifica var . shastensis), included observations 

of phenology, plant water potential, 

stomatal resistance, and foliar nutrition . 

Productivity was estimated by measuring terminal 

elongation or total height of dominant 

old-growth trees. Details of this work hav e 

been published elsewhere (Waring and Clear y 

1967, Waring 1969, Atzet and Waring 1970 , 

82

Table 1.-Description of stands 

Stand Elevation 1 Slope Aspect Parent material Dominant vegetation 

Meters 

Percent 

3 780 45 N Granite Douglas-fir, black oak, 

ponderosa pine 

21 550 75 N Metavolcanic Douglas-fir, black oak, 

Oregon white oa k 

8 1,280 40 SW Granite Ponderosa pine, Douglas-fi r 

12 1,585 70 WSW Green schist Ponderosa pine, Douglas-fir 

1 1,490 25 W Granite White fir, ponderosa pine , 

Douglas-fir 

11 1,370 35 SW Granite Ponderosa pine, sugar pine , 

white fir, Douglas-fi r 

13 1,340 20 W Green schist Douglas-fir, ponderosa pine , 

sugar pine, white fir 

19 1,250 70 W Mica schist Douglas-fir, sugar pine, 

white fir 

2 1,675 60 WNW Granite White fir, Douglas-fir 

9 1,550 55 NNW Metavolcanic White fir, sugar pine , 

Shasta red fir 

14 1,585 45 E Green schist Douglas-fir, white fir 

22 1,460 50 N Metasedimentary Douglas-fir, white fir 

7 1,920 20 N Granite Shasta red fi r 

16 1,890 40 SW Mica schist Shasta red fi r 

17 1,830 10 E Metavolcanic Shasta red fi r 

6 2,040 35 NNE Granite Mountain hemlock, Shasta red fi r 

15 2,010 10 NE Mica schist Mountain hemloc k 

18 2,135 30 NE Granite Mountain hemlock 

24 2,040 45 NNE Metavolcanic Mountain hemlock 

4 1,920 65 SE Ultrabasic Jeffrey pine, incense-cedar , 

western white pin e 

5 1,710 65 SE Ultrabasic Jeffrey pine, incense-ceda r 

25 1,740 5 SE Ultrabasic Jeffrey pine, white fir, 

incense-cedar, Douglas-fi r 

20 760 70 NNW Mica schist Douglas-fir, Pacific ye w 

23 1,400 10 N Granite Engelmann spruce, Douglas-fir , 

white fi r 

10 1,740 55 N Metavolcanic Brewer spruce, Shasta red fir, 

mountain hemlock 

8 3

Cleary and Waring 1969, Reed 1971, Emmingham 

1971, Waring and Youngberg 1972) . 

In this article we will restrict discussion to 

interpretation of results, concentrating upo n 

the operational effects of water, temperature , 

and soil fertility . Mechanical stress from sno w 

creep or ice storms was important as was win d 

(Tranquillini 1970), but such adverse mechanical 

forces are closely associated wit h 

temperature gradient. Light, too, was important 

in understanding succession an d 

accounted for more than 75 percent of th e 

variation in terminal elongation of Abies 

when other environmental factors were 

restricted to narrow ranges (Emmingham 

1971) . In our stands, which were mature 

forests, maximum height of mature trees wa s 

used as an index to productivity . In such a 

situation, earlier reductions in growth due to 

shading or mechanical damage could be ignored 

. 

Measurement and Interpretation of 

Plant Water Stress 

During the growing season, plant moistur e 

stress measured by the Scholander pressur e 

chamber (Scholander et al . 1965, Waring an d 

Cleary 1967, Boyer 1967) provided a goo d 

estimate of plant water potential (Waggone r 

and Turner 1971) . We took measurements a t 

monthly intervals, with more frequenc y 

during September when the vegetation wa s 

under the greatest stress . 

Plant moisure stress usually increases fro m 

a predawn minimum to some maximum leve l 

in the afternoon . Predawn stress represent s 

the nearest equilibrium between soil and plant 

water potential . Not only the diurnal patter n 

but also the seasonal changes in plant wate r 

stress are important . The first reflects an imbalance 

between transpiration and water 

uptake ; the latter can be related to the avail - 

ability of soil water over a season . In figure 3 , 

three contrasting forest types are shown t o 

have different seasonal water stress curves . 

The type dominated by Engelmann spruce 

(Picea engelmannii) was restricted to mois t 

sites, while oak (Quercus kelloggii) grew 

where stress was sufficient to bring about 

cessation of cambial activity in the reference 

Figure 3. Seasonal changes in minimum plant moisture 

stress associated with different forest ecosystems 

(Waring 19701 . All data are from 1- t o 

2-m-tall Douglas-fir. For this particular set of data , 

the end of season Plant Moisture Stress Inde x 

would be 30 for the Oak Type, 18 for the Pin e 

Type, and 7 for the Spruce Type . 

plants. Where no rainfall occurs during the 

summer, the minimum night moisture stres s 

at the end of the growing season is an index 

to the plant moisture conditions throughout 

the entire season . From figure 3, we see the 

oak type had a plant moisture stress index o f 

30 ; the pine, an index of around 18 ; and the 

spruce, 7 atmospheres . Where summer precipitation 

is important, a mathematical description 

of the seasonal trend is desirable (Ree d 

1971) . Winter drought due to cold or frozen 

soils should also be similarly evaluated . 

Measurement and Interpretatio n 

of Temperatur e 

At each of the 25 forest stands, air temperature 

and soil temperature were recorded 

on 30-day thermographs because both root 

and shoot temperatures are important . At the 

time, we did not have photosyntheti c 

response information to help interpret th e 

effect of various combinations of temperatur e 

and light. Because winter temperatures are 

often below freezing, winter photosyntheti c 

activity is probably less in the study regio n 

than in the milder climate along the Pacifi c 

84

coast. Without the benefit of good photo - 

synthetic data, we used laboratory studie s 

conducted by our colleague, D . P. Lavender . 

Lavender found that the dry weight increas e 

of Douglas-fir seedlings was closely related to 

both air and soil temperatures . With these 

data a temperature-plant index was develope d 

by summing the potential growth possible for 

each day during the growing season (Cleary 

and Waring 1969) . Other variables were 

assumed nonlimiting. In the field, the Temperature 

Growth Index ranged from 30 near 

timberline to nearly 100 on oak and pin e 

forests. Douglas-fir was not found where th e 

index was below 40 . 

To visualize this information more effectively, 

the distributions of selected conifer s 

are presented in relation to these two rathe r 

simple plant response indices (fig. 4). The distributional 

patterns closely reflect the adapta - 

I . I . I I 

10 20 30 

PLANT MOISTURE STRESS, ATM . 

Figure 4 . Distribution of natural regeneration in relation 

to gradients of moisture and temperature de - 

fined by plant response indices (Waring 1970) . 

Validation stands, symbolized by 0, had vegetation 

predicted by the intercept of their plant 

response indices . 

tion of the various conifers . In a local region , 

the variation in distribution of different 

species provides a means of predicting the environment 

through association with measurements 

taken on reference plants . The plan t 

response indices were correctly predicte d 

from knowledge of plant distributions fo r 

three validation stands, indicated as divide d 

circles in figure 4 . It is significant that suc h 

predictions are possible without physiological 

observations on species other than the reference 

plants and without special attention t o 

events controlling establishment . Once such 

relationships are established, the vegetatio n 

can provide understanding of the operational 

environment without requiring additional 

measurements of any kind . We shall expand 

this idea later . 

Measurement and Interpretation of 

Stomatal Respons e 

Conifer stomata are most difficult to observe, 

usually being sunken and occluded b y 

wax . The resistance which they offer to th e 

transfer of water vapor from the interior of 

the needles can be assessed by determinin g 

the rate of water vapor movement with a 

diffusion porometer (Waggoner and Turne r 

1971). The aperture of stomata may be estimated 

by observing the pressure necessary t o 

force a 50-percent ethanol solution through 

the pores (Fry and Walker 1967). Th e 

diffusion resistance is then determined by 

calibrating these pressures with reduction i n 

transpiration under known vapor pressure 

gradients (Reed 1971) . The latter procedure 

was followed in our past fieldwork . We found 

that stomatal resistance increased as the soi l 

moisture became less available during th e 

growing season . Further increase in stomatal 

resistance was possible during the day i f 

evaporative stress was high . These relationships 

were quantified and developed into a 

simulation model by Reed (1971) . Wit h 

knowledge of temperature, humidity, and 

nocturnal plant water stress, the model predicted 

on a daily basis, both potential transpiration 

(PT) and transpiration (T) wit h 

stomatal control . These values were summed 

for the entire season and confirmed the obser - 

85

vation that species such as Brewer spruc e 

(Picea breweriana), Port-Orford-cedar 

(Chamaecyparis lawsoniana), vine maple 

(Ater circinatum), and rhododendron 

(Rhododendron macrophyllum) are restricte d 

to areas with low potential transpiration . Th e 

most valuable index, however, was the rati o 

of actual to potential transpiration (T/PT) 

Which integrates for water, the demand, th e 

supply, and the control b the plant . Where 

sot moisture was never inadequate and th e 

evaporative stress remained low, the ratio wa s 

1 .0. Where water became limiting and hig h 

evaporative stress was common, ratios of 0 . 3 

were calculated . Under these conditions, both 

forest composition and growth were dramatically 

affected . 

Measurement and Interpretation of Soi l 

Fertility and Plant Nutritio n 

Soil profiles from each stand were reconstructed 

to a depth of 60 cm and taken into 

controlled environment chambers where seedlings 

of Douglas-fir and Shasta red fir were 

grown for a period of 5 months (Waring an d 

Youngberg 1972) . The dry weight yields at 

the end of the experiment were used as a bio - 

assay of soil fertility, representing the supply 

of nutrients available to conifers . Foliar analyses 

were made on reference trees, first durin g 

the time of maximum demand when ne w 

foliage was being produced and again after al l 

shoot and diameter growth had ceased in th e 

fall. In the first period, 1-year-old foliage wa s 

assessed because it represents a major sourc e 

of mobile nitrogen, phosphorus, and potassium. 

Three categories of nutrient availability 

can be identified : (1) where no nutritional 

stress occurs during the year ; (2) where nutrition 

is adequate only when shoot and 

diameter growth has ceased ; and (3) where 

nutrition is inadequate year around (Warin g 

and Youngberg 1972) . These three categorie s 

may be further refined and are of consider - 

able value in reaching decisions concernin g 

fertilization. The composition of vegetatio n 

is, however, insensitive to this classification of 

nutrient availability . 

Only on soils where an imbalance of nutrients 

or toxic amounts of certain heavy metals 

were present did special plant communitie s 

develop. In this paper, therefore, we hav e 

assigned plants to one of three categories of 

tolerance to infertile soils developed from 

ultrabasic parent materials : In class 1 ar e 

those which are tolerant ; class 2 is made up of 

those species that are intolerant ; and in class 3 

are those plants both tolerant and competitively 

restricted to ultrabasic soils . 

Vegetation as an Index to Environmen t 

Can we use plants to define their environment, 

or more precisely, the environment expressed 

through reference plants? To test thi s 

idea, we selected 47 species from more than 

600 in the local flora and recorded thei r 

distributional limits in relation to various 

environmental plant indices (table 2) . For 

some plants we knew only the approximat e 

ranges and occasionally we had insufficien t 

data to set particular index limits . 

In forest ecosystems where plant composition 

is known, the range of environmenta l 

indices may be predicted from information i n 

table 2 . In table 3, the plant response indice s 

were thus predicted for the 25 stand s 

described in table 1 . Therefore for the blac k 

oak forest, stand 3, where Rhus diversiloba , 

Arbutus menziesii, Lonicera hispidula, and 

Quercus chrysolepis grew, one can assess fro m 

table 2 that the Temperature Growth Inde x 

lies between 98 and 96, the Plant Moistur e 

Stress Index at 25 .4, and the ratio of transpiration 

to potential transpiration as 0 .29 . 

Plants that were exclusively adapted t o 

ultrabasic soils, such as Jeffrey pine, isolated 

those ecosystems (stands 4, 5, and 25) . Those 

forests with oak (stands 3, 8, and 21 ) 

exhibited the greatest water stress, th e 

warmest environments, and the greatest control 

of transpiration . At the other extreme , 

mountain hemlock forests (stands 6, 18, 15 , 

and 24) had the coolest environments and 

lowest potential transpiration . 

In figure 5, the midpoint of the predicted 

Temperature Growth Index from table 3 is 

plotted against the Temperature Growt h 

Index calculated from temperature records . 

The regression has an r2 of 0 .93, and is highl y 

86

Generic name 

Table 2.-Ecological distribution of selected species in relation t o 

plant response indices' 

PT , PT , T/PT , T/PT , PMS , PMS , TGI, TGI, Soil 

maxi- mini - maxi- mini- maxi- mini- maxi- mini- tolermum 

mum mum mum mum mum mum mum anc e 

ANACARDIACEAE 

Rhus diversiloba 30.0 17 .7 0 .42 0 .29 25 .4 20 .3 100 80 1 

BERBERIDACEA E 

Achlys triphylla 16.8 1 .00 .57 16 .2 80 60 2 

CAPRIFOLIACEAE 

Lonicera hispidula 

COMPOSITA E 

30.0 17 .7 .29 30.0 25 .4 100 95 1 

Arnica latifolia 13.5 7 .5 1 .00 .46 19 .1 5 .2 75 35 2 

CUPRESSACEA E 

Libocedrus decurrens 

ERICACEAE 

19.5 12.2 1 .00 .40 30 .0 5 .2 85 52 1 

Arbutus menziesii 30.0 - 1 .00 .29 30 .0 5 .2 98 68 1 

Arctostaphylos uiscida 30.0 17 .7 .42 .29 25 .4 20 .3 100 90 1 

Rhododendron macrophyllum 12.5 - 1 .00 .57 16 .2 - 100 2 

Vaccinium membranaceum 12.5 .63 .46 19 .1 12 .8 60 45 2 

Vaccinium scoparium 10.3 1 .00 .46 19 .1 - 50 2 

FAGACEA E 

Castanopsis chrysophylla 19.5 - 1.00 .51 15 .3 - 85 40 1 

Quercus chrysolepis 30.0 12 .0 1.00 .63 30 .0 8 .0 100 78 1 

Quercus kelloggii 30.0 17 .6 .42 .29 30 .0 15 .0 98 80 1 

Quercus sdeeriana 16.8 - 1 .00 .46 19 .1 - 70 50 1 

Quercus uaccinifolia 30.0 .40 - 30.0 100 - 1 

GARRYACEA E 

Garrya fremontii 30.0 12 .7 .62 .40 30 .0 12 .6 70 50 1 

LABIATA E 

Monardella odoratissima 21 .4 .62 .40 20 .3 12 .6 100 35 1 

LEG UMIN OS A E 

Lathyrus polyphyllus 13.1 - 1 .00 .63 12 .8 5 .2 85 45 2 

Lupinus leucophyllus 30.0 12 .7 .86 - 30.0 8 .4 70 50 3 

LILIACEAE 

Clintonia uniflora 13 .1 7 .5 1.00 .51 16 .2 5 .2 85 35 2 

Disporum hooheri 16 .8 12 .2 1.00 .51 16 .2 5 .2 85 45 2 

Xerophylium tenax 19 .5 - 1 .00 .40 19 .1 8 .4 80 40 1 

PINACEA E 

Abies concolor 19.5 - 1 .00 .46 19 .1 5 .2 84 47 1 

Abies magnifica 12.5 7 .5 1 .00 .46 19 .1 5 .2 59 34 2 

var . shastensis 

Picea breweriana 10.3 1 .00 .46 19 .1 52 1 

Picea engelmannii 11 .0 - 1.00 .63 12 .8 - 47 - 2 

Pinus jeffreyi 25.0 12 .7 .61 - 25 .0 8 .4 100 52 3 

Pinus lambertiana 21 .4 11 .0 1 .00 .29 25 .4 5 .2 96 45 1 

Pin us ponderosa 21 .4 13 .5 1 .00 .29 25 .4 5 .2 98 68 2 

Pinus monticola 19.5 7 .8 1 .00 .46 19 .1 5 .2 70 44 1 

Pseudotsuga menziesii 21 .4 - 1 .00 .29 25 .4 5 .2 98 47 1 

Tsuga mertensiana 11 .0 7 .5 1 .00 .46 19 .1 5 .2 59 34 2 

POLEMONIACEA E 

Phlox adsurgens 16.8 10 .3 1.00 .46 19 .1 5 .2 85 52 2 

Polemonium californicum 7 .8 - 1 .00 .98 5 .2 - 50 30 2 

POLYGONACEAE 

Polygonum dauisiae 7 .8 1 .00 .98 5 .2 40 30 2 

POLYPODIACEAE 

Onychium densum 

RANUNCULACEAE 

30.0 12 .7 .62 18 .1 5 .2 80 50 3 

Anemone deltoidea 19 .5 7 .8 1 .00 .51 19 .1 5 .2 85 45 1 

Anemone lyallii 19 .5 7 .8 1 .00 .51 15 .3 5 .2 75 45 2 

ROSACEAE 

Amelanchier pallida 16.8 .62 .46 19 .1 75 50 1 

Rubus lasiococcus 12.5 1 .00 .46 19 .1 - 60 - 2 

Rubus paruiflorus 

SAXIFRAGACEA E 

16 .8 1 .00 .51 16 .2 5 .2 85 50 2 

Ribes viscosissimum 13 .1 7 .8 1 .00 .51 16 .2 5 .2 85 40 2 

Tiarella unifoliata 13 .1 - 1 .00 .63 12 .0 - 85 47 2 

SCROPHULARIACEAE 

Pedicularis racemosa 

VALERIANACEAE 

12.5 7 .5 1 .00 .51 16 .2 5 .2 65 35 2 

Valeriana sitchensis 

VIOLACEA E 

7 .8 1 .00 .98 5 .2 45 30 2 

Viola glabella 

Viola sempereirens 

13 .1 

13 .1 

7 .5 

- 

1 .00 

1 .00 

.63 

.57 

12 .8 

16 .2 

5 .2 

5 .2 

95 

85 

35 

45 

2 

2 

' 

P otential transpiration (PT) is calculated for April through September based on a minimum stomatal resistance of 4 se c 

cm Transpiration is expressed in g cm - 2 

The ratio of simulated transpiration (T) to potential (PT) reflects the degree of stomatal control exhibited by a referenc e 

conifer. 

Plant Moisture Stress (PMS) is expressed in atm and represents predawn measurements on reference conifers near the end 

of the summer dry season (Sept .) . 

A Temperature Growth Index (TGI) reflects the potential for Douglas-fir seedling growth as a function of air and soil 

t emp erature for the entire growing season. 

The soil tolerance index indicates whether the species is exclusively restricted (class 3), tolerant (class 1), or excluded (class 

2) from infertile ultrabasic soils . 

Values of 30 in PT and PMS columns indicate species are known to occupy environments more extreme than possible fo r 

coniferous forest . 

8 7

Table 3.-Predicted plant response indices for 25 forest ecosystems based on th e 

inclusive limits of species present with known ecological distributions ' 

Forest type 2 

Stand 

PT , PT, T/PT , T/PT , PMS , PMS, TGI , TGI , Soi l 

maxi - mini- maxi - mini- maxi - mini- maxi- mini - tolermum 

mum mum mum mum mum mum mum anc e 

Black oak 3 21 .4 17 .7 0 .29 0 .29 25 .4 25 .4 98 95 1 

Black oak 21 3 21 .4 17 .7 .29 .29 25 .4 25 .4 96 95 1 

Ponderosa pine 8 21 .4 17 .7 .42 .40 20 .3 20 .3 96 90 1 

Ponderosa pine 12 3 16 .8 13 .5 .62 .46 19 .1 12 .6 70 68 1 

Mixed conifer 1 13 .5 13 .5 .62 .51 15 .3 5 .2 75 68 2 

Mixed conifer 11 16 .8 13 .5 1 .00 .57 15 .3 5 .2 70 68 1 

Mixed conifer 13 13 .1 12 .2 1 .00 .63 12 .8 8 .4 80 68 1 

Mixed conifer 19 13 .1 12 .2 1 .00 .63 12 .8 5 .2 84 68 2 

White fir 2 12 .5 12 .2 .62 .51 15 .3 12 .6 65 50 1 

White fir 9 12 .5 12 .2 .63 .57 16 .2 12 .8 59 60 2 

White fir 14 3 13 .1 11 .0 1 .00 .63 12 .8 8 .4 70 52 2 

White fir 22 3 13 .1 10 .3 .62 .63 12 .8 5 .2 70 52 2 

Shasta red fir 7 3 7 .8 7 .8 1 .00 .98 5 .2 5 .2 45 45 2 

Shasta red fir 16 12.5 7 .5 1 .00 .51 15 .3 5 .2 59 40 1 

Shasta red fir 17 7 .8 7 .8 1 .00 .98 5 .2 5 .2 45 45 2 

Mountain hemlock 6 7 .8 7 .5 1 .00 .98 5 .2 5 .2 45 35 2 

Mountain hemlock 15 3 11 .0 7 .5 1 .00 .46 19 .1 5 .2 59 34 2 

Mountain hemlock 18 3 7 .8 7 .5 .62 .46 19 .1 12 .6 40 35 2 

Mountain hemlock 24 3 7 .8 7 .8 1 .00 .98 5 .2 5 .2 45 44 2 

Jeffrey pine 4 16 .8 12 .7 .61 .51 15 .3 12 .6 70 52 3 

Jeffrey pine 5 19 .5 12 .7 .40 .40 18 .1 12 .6 70 52 3 

Jeffrey pine 25 3 19.5 12 .7 .61 .51 15 .3 8 .4 70 52 3 

Yew 20 3 17 .7 16 .8 .51 .42 16 .2 15 .0 85 80 2 

Engelmann spruce 23 11 .0 11 .0 .63 .63 12 .8 12 .8 47 47 2 

Brewer spruce 10 10.3 7 .8 .62 .46 19 .1 12 .8 50 50 2 

' Potential transpiration (PT) is calculated for April through September based on a minimum stomatal resistance of 4 sec 

cm^1 . Transpiration expressed in cm- 2 

The ratio or simulated transpiration (T) to potential (PT) reflects the degree of stomata] control exhibited by a referenc e 

conifer. 

Plant Moisture Stress (PMS) is expressed in atm and represents predawn measurements on reference conifers near the en d 

of the summer dry season (Sept .) . 

A Temperature Growth Index (TGI) reflects the potential for Douglas-fir seedling growth as a function of air and soil 

temperature for the entire growing season . 

The soil tolerance index indicates whether the species is exclusively restricted (class 3), tolerant (class 1), or excluded (clas s 

2) from infertile ultrabasic soils . 

2 From R . H . Waring, Forest plants of the eastern Siskiyous : their environmental and vegetational distribution. Northwes t 

Sci . 43 : 1-17, 1969 . 

'Simulation of PT and T/PT was not possible because of inadequate data . The predicted values appear reasonable whe n 

compared with other stands with similar vegetation, temperature, and plant moisture stress indices . 

88

x 

-0 o 90 

LL 

L 

70 

L 

a) 

a 

E 

50 

a ) 

t 

-o a ) 

L 

a- 30 

30 50 70 90 

Observed Temperature Growth Inde x 

Figure 5 . Relationship between the temperature index 

derived from the growth response of Douglas - 

fir to temperature and the midpoint values predicted 

from the presence of indicator species in a 

given ecosystem (table 3) . Validation stands, 0 , 

with one exception fell within the 95-percent 

confidence limits . r2 = 0.93 . 

30 

significant . Similarly, figure 6 presents the regression 

for plant moisture stress where the r 2 

was 0.77, also highly significant. The symbol 

8 represents validation stands which generall y 

fell within the 95-percent confidence intervals 

. 

From the comparisons made earlier, we 

showed that environmental plant indices ca n 

predict the forest composition, and conversely 

from the above relationships, we not e 

that environmental indices may be predicted 

from knowledge of the distribution of specifi c 

plants . 

Environmental Plant Indices as Predictor s 

of Forest Productivity 

In this study the only index to productivit y 

was maximum tree height . Particularly on th e 

infertile ultrabasic soils, this index overestimates 

productivity, for although individua l 

trees reach considerable height, the density o f 

trees remains low. Reed and Waring l reporte d 

that the ratio of actual to potential transpiration 

(T/PT) combined with the Temperatur e 

Growth Index (TGI) accounted for 96 per - 

cent of the variation observed in the maxi - 

mum height of Douglas-fir . For all dominant 

trees, these and related indices accounted fo r 

86 percent of the variation in height . The 

model for analytic solution is : 

Maximum height (in meters) = 

- 102.3 + 0.334 PT + 316.4 T/PT + 0.405 TGI 

- 186.8 (T/PT) 2 - 3.25 (T/PT)PT + 0.631 (T/PT)TGI 

This is encouraging, considering the limited 

sampling of tree heights and the expecte d 

differences among species . The model was not 

improved by including the index to soil fertility, 

again suggesting that total height may 

be an inadequate measure of productivity . 

Figure 6. Relationship between the moisture stress 

index during the period when soil water may b e 

limiting to reference conifers and the midpoin t 

values predicted from the presence of indicato r 

species in a given ecosystem (table 3) . Validation 

stands, 0, fell within the 95-percent confidence 

limits . r2 = 0.77 . 

30 

Data Requirement s 

If the environmental grid is to be applie d 

across the Coniferous Forest Biome, then a 

coordinated effort must be made to test an d 

improve the approach . Present efforts by 

1 Unpublished data . 

89

Walker, Reed, Scott, and Webb are channele d 

toward providing additional physiological 

information on Douglas-fir, one of four pro - 

posed reference plants . These investigators are 

concerned with photosynthesis, water uptake , 

transpiration, and translocation . Lavender an d 

Hermann are responsible for providing functional 

relationships between foliage, stem, an d 

roots. Through stem and twig analysis and b y 

aging foliage, we hope to gain measures of net 

primary production . 

Other reference plants include Pinus 

ponderosa, Tsuga heterophylla, and Abies 

lasiocarpa ; additional studies on these specie s 

are necessary . 

To extend and validate the genera l 

approach described in this paper, we will re - 

quire certain kinds of information at selecte d 

forest environments across the Biome . Forest 

stands will be selected which represent equilibrium 

stages ; i.e., approach the maximu m 

accumulation of biomass, on homogeneou s 

areas at least 30 m 2 . 

Within these stands, bud break will b e 

recorded on 1- to 2-m reference conifers b y 

observing five lateral branches on each of a t 

least five trees. Cambial activity will b e 

assessed by using the pin-into-cambium technique 

which identifies the termination of cel l 

divisions (Wolter 1968) . Stand compositio n 

and structure will be estimated from basa l 

area and height estimates of trees and under - 

story sampling of shrub and herbaceous cover . 

Nocturnal moisture stress will be measure d 

on three to five reference trees at 2-week intervals 

throughout the growing season and at 

less frequent intervals during the dorman t 

season, unless winter dessication is likely . 

Temperature under the stand near referenc e 

trees will be recorded at 1 m above and 20 c m 

below ground level. Latitude location o f 

stands will allow for calculation of day length 

for different seasons . Nutrition will b e 

assessed from a composite sample of 1-yearold 

foliage collected from the lateral branches 

of at least five trees at the time of new shoot 

elongation and during the late fall . 

Short wave radiation will be measured in a 

nearby open area or above the stand. Data 

will be expressed in cal cm -2 day -1 . Absolut e 

humidity will be measured daily at solar noon 

during the growing season and if possible continuously 

year around. Precipitation will include 

both monthly values and average maxi - 

mum snow depth . Estimates of wind velocity 

are desired where conditions appear to warrant 

measurements . 

These data will not only be valuable to hel p 

construct the environmental grid but wil l 

serve to interpret process studies in decomposition, 

consumer population dynamics, and 

other ecosystem properties . 


All field investigations reported in thi s 

paper were generously funded under an extended 

McIntire-Stennis Federal Grant . Th e 

simulation of transpiration, regression equations 

for productivity, and a computer program 

for defining plant response indices o n 

the basis of overlapping plant distribution s 

were developed under support by Nationa l 

Science Foundation Grant No . GB-20963 t o 

the Coniferous Forest Biome, U .S. Analysis o f 

Ecosystems, International Biological Program . 

This is Contribution No . 26 from the Coniferous 

Forest Biome . 


Atzet, T., and R. H . Waring. 1970. Selectiv e 

filtering of light by coniferous forests an d 

minimum light energy requirements fo r 

regeneration . Can. J. Bot. 48 : 2163-2167 . 

Azzi, G . 1955. Agricultural ecology . 424 p. 

London : Constable and Co . 

Bakuzis, E. V. 1961 . Synecological coordinates 

and investigations of forest ecosystems 

. 13th Congr. Int. Union For. Res . 

Org. (Vienna) Proc . 2, Sec . 21½ . 

Boyer, J . S . 1967 . Leaf water potentials measured 

with a pressure chamber. Plant 

Physiol. 42: 133-137 . 

Cleary, B . D., and R. H. Waring. 1969. Temperature 

: collection of data and its analysis 

for the interpretation of plant growth an d 

distribution . Can . J. Bot . 47 : 167-173 . 

Daubenmire, R . 1952. Forest vegetation of 

northern Idaho and adjacent Washington , 

90

and its bearing on concepts of vegetatio n 

classification . Ecol . Monogr . 22 : 301-330 . 

E l lenberg, H. 1950. Landwirtschaftliche 

Pflanzensoziologie . Bd . I . Unkrautgemeinschaften 

als Zeiger fur Klima and 

Boden . 141 p . Ulmer, Stuttgart . 

1956 . Aufgaben and Methoden 

der Vegetationskunde . In H . Walter (ed .) , 

Einfuhrung in die Phytologie. Bd. IV . 

Grundlagen der Vegetationsgliederung . Tel 

1 . 136 p . Ulmer, Stuttgart . 

Emmingham, W . H . 1971 . Conifer growth and 

plant distribution under different light environments 

in the Siskiyou Mountains of 

southwestern Oregon . 50 p. M .S. thesis, on 

file at Oreg . State Univ ., Corvallis . 

Franklin, J. F., and C. T. Dymess. 1969 . 

Vegetation of Oregon and Washington . 

USDA Forest Serv . Res . Pap . PNW-80, 21 6 

p., illus. Pac. Northwest Forest & Range 

Exp . Stn., Portland, Oreg . 

Fry, K. E., and R . B . Walker. 1967 . A pressure 

infiltration method for estimatin g 

stomatal opening in conifers . Ecology 48 : 

155-157 . 

Hills, G. A. 1959. A ready reference to th e 

description of the land of Ontario and it s 

productivity. 142 p. Ont . Dep. Lands & 

Forests . 

Krajina, V . J. 1965. Biogeoclimatic zones and 

classification of British Columbia . In V . J . 

Krajina (ed .), Ecology of western North 

America, vol . 1, p. 1-17 . Univ . Brit . Columbia, 

Vancouver . 

Loucks, O. L. 1962. Ordinating forest communities 

by means of environmental scalars 

and p h y t o sociological indices. Ecol . 

Monogr . 32 : 137-166 . 

Mason, H. L., and J . H. Langenheim . 1957 . 

Language analysis and the concept of environment. 

Ecology 38 : 325-339 . 

Progrebnjak, P . S. 1929 . Uber die Methodik 

der Standortsuntersuchungen in Verbindung 

mit den Waldtypen. Int. Congr . 

Forest Exp. Stn. (Stockholm) Proc., p . 

455-471 . 

Reed, K. L. 1971. A computer simulation 

model of seasonal transpiration in Douglasfir 

based on a model of stomatal resistance . 

132 p . Ph .D . thesis, on file at Oreg . State 

Univ., Corvallis . 

Rowe, J . S . 1956. Use of undergrowth plan t 

species in forestry . Ecology 37 : 461-472 . 

Scholander, P . F., H . T. Hammel, D . Broad - 

street, and E . A. Hemmingsen. 1965. Sap 

pressure in vascular plants . Science 148 : 

339-346 . 

Sukachev, V. N . 1928. Principles of classification 

of the spruce communities of European 

Russia. J. Ecol . 16 : 1-18 . 

Tranquillini, W ., 1970 . Einfluss des Winde s 

auf den Gaswechsel der Pflanzen Umsha u 

in Wissenschaft and Technik 26: 860-861 . 

Waggoner, P . E ., and N. C . Turner . 1971 . 

Transpiration and its control by stomata in 

a pine forest . Conn. Agric. Exp. Stn . Bull . 

726, 87 p. New Haven . 

Waring, R. H. 1969 . Forest plants of the east - 

ern Siskiyous : their environmental and 

vegetational distribution. Northwest Sci . 

43 : 1-17 . 

1970. Matching species to site . 

In R. K. Hermann (ed .), Regeneration o f 

ponderosa pine, p . 54-61 . Oreg. State 

Univ., Corvallis . 

and B. D. Cleary . 1967. Plan t 

moisture stress : evaluation by pressure 

bomb. Science 155 : 1248-1254 . 

and J. Major . 1964. Some vegetation 

of the California coastal redwoo d 

region in relation to gradients of moisture , 

nutrients, light, and temperature . Ecol . 

Monogr . 34: 167-215 . 

and C. T. Youngberg . 1972 . 

Evaluating forest sites for potential growt h 

response of trees to fertilizer. Northwest 

Sci . 46 : 67-75 . 

Warming, E . 1909. Oecology of plants . 422 p. 

London : Oxford Univ . Press . 

Whittaker, R . H. 1956. Vegetation of th e 

Great Smoky Mountains . Ecol. Monogr . 

26 : 1-80 . 

1960. Vegetation of the Siskiyou 

Mountains, Oregon and California . 

Ecol . Monogr . 30 : 279-338 . 

1961 . Vegetation history of th e 

Pacific Coast States and the "central" significance 

of the Klamath region . Madron o 

16 : 5-23 . 

Wolter, K . E . 1968. A new method for marking 

xylem growth . Forest Sci . 14 : 102-104 . 

91

Water and Nutrien t 

Movement Through Ecosystems 

93



Modeling water movement within 

the upper rooting zone of a 

Cedar River soil 

W. H . Hathewa y 

P. Machn o 

an d 

E . Hamerl y 


Seattle, Washington 9819 5 

Abstract 

The Richards equation for unsaturated soil water flow is used to represent flow in a natural forest soil . Th e 

differential equation and the corresponding finite difference technique used to obtain an approximate solutio n 

are discussed. Independent estimates of soil moisture conductivity and of initial soil water content at severa l 

depths are used as inputs to the finite difference equation. Conductivity was estimated by laboratory techniques, 

and initial conditions in the field were measured by tensiometer. A tension lysimeter system provided 

estimates of soil water flow. Predicted values were compared with values measured in the field . Results sugges t 

that the model gives a satisfactory representation of actual soil water flow despite considerable variability in 

forest soil properties. 

Introduction 

In the Coniferous Forest Biome, we hav e 

been interested in models which represent th e 

movement of water through unsaturated soil s 

for use as possible components of a larger 

computer model which describes water relations 

within a local soil-plant-atmospher e 

system. Other components of the larger 

model include uptake of water by the roots of 

trees, conduction of water through the vascular 

system to the leaves, and transpiratio n 

from the leaves to the surrounding atmosphere 

. When we find that existing submodel s 

can be used-possibly with slight modifications-in 

our larger soil-plant-atmospher e 

model, we prefer to use them rather than t o 

invent them anew . Where existing models ar e 

not well suited to our objectives, we prefer t o 

develop new ones . For example, we are currently 

constructing a model which represent s 

the flow of water through the vascular syste m 

of a tree. We evaluate existing or new sub - 

models on the basis of (1) their behavior in 

isolation and (2) their compatibility with 

other components of the system . In the 

present paper we report our experiences with 

the Richards equation, a model which de - 

scribes the flow of water in an unsaturate d 

soil . 

The Model 

The Richards equation (Richards 1931) is 

essentially a modification of the well-known 

Darcy relationship for saturated flow i n 

porous media. Because the Richards model 

allows for variation in hydraulic conductivity 

with changes in soil water content, it is wel l 

suited for the description of water flow in 

unsaturated media . Darcy 's law with variabl e 

conductivity and negative hydraulic head can 

be written (Rose 1966 ) 

v=K aia 6 ae aZ 

K (1) 

95

Here v is the volume of water crossing a uni t 

area per unit time (cm/min) , 

K is the hydraulic conductivity (cm/min) , 

is the soil water pressure or suctio n 

(hydraulic head) (cm) , 

0 is the soil water content (volume of 

water per unit volume of soil), an d 

z is the height above datum (referenc e 

level) of the point under consideratio n 

(cm) . 

By introducing the concept of soil-wate r 

diffusivity, defined b y 

purposes of calculations . Soil moisture specified 

as a function of time is required at th e 

upper and lower boundaries of the soil column 

. Their solution permits the simultaneous 

calculation o f 

(Dj + 

Dj+) ) 

2 (oJ+1 zj+1 - 01 zj 

This expression approximates D a8 and hence 

v, the volume flux of water, apart from th e 

constant K (cf. equation 2) . 

D = -K a alP d , 

where D is measured in c m 2 /min, Childs and 

Collis-George (1950) cast the Darcy equatio n 

in a form similar to Fick 's law of diffusion : 

v = -Da 8 -K (2 ) 

Applying the equation of continuity , 

ad + av 

at az 

= 0 (which implies that there are n o 

sources or sinks of water in the system), one 

obtains the full Richards soil flow equation : 

30 __ a ad aK 

at az (D az ) + a z 

(3 ) 

Soil moisture is obtained as a function o f 

depth (z) and time (t) by integrating (3) . 

Accordingly the Richards model is compatibl e 

with our larger soil-plant-atmosphere model , 

which requires soil moisture at various depth s 

and times as an input to a submodel whic h 

represents the uptake of water by the roots o f 

a tree. Our current uptake model is that of 

Gardner (1960), a diffusion-type flow equation 

in which time rate of change of soil wate r 

concentration (0) at any point (z) in the soi l 

is related to the moisture gradient betwee n 

that point and an absorbing root . 

Since D and K depend on 0, (3) is non - 

linear in 0, and a solution is possible only i f 

numerical approximations are used . Remso n 

et al . (1965) developed a computer program 

which gives approximate results when initial 

moisture content is specified at each of n+ 1 

equally spaced points or nodes into which a 

soil column zn centimeters thick is divided for 

Available 

Observational Data 

At the Allan Thompson Research Center , 

Cedar River Watershed, the downward flow o f 

water through soil profiles has been measure d 

for several years by a tension lysimeter syste m 

(Gessel and Cole 1965) . A prescribed suctio n 

on the porous lysimeter plate causes drainage 

through the plate to approximate the drainag e 

in the adjacent soil . Results are recorded i n 

permanent form on paper tape by an automatic 

data recording system . Our data com e 

from a series of controlled experiments, carried 

out in 1970, which were designed t o 

study the effects of rainfall on wetting front s 

in the rooting zone of a Douglas-fir stand . 

Lysimeter plates were installed at depths o f 

11 and 25 cm and rainfall was simulated by a 

below-canopy sprinkler system . Soil moisture 

was measured by tensiometer at a point mid - 

way between the soil surface and the uppe r 

lysimeter plate, and water flux through th e 

lysimeter was recorded by a flow meter connected 

to the data recorder . 

For the present study only the data for th e 

11-cm plate were used. Since an abrup t 

change in the physical characteristics of th e 

soil occurs between the 11- and 25-cm levels , 

a different set of estimates of the conductivity 

and diffusivity parameters is require d 

for the lower layer . A model coupling th e 

flow of water through the two layers has bee n 

developed but has not yet been tested . 

96

Parameter Estimation 

In order to estimate conductivity (K) and 

diffusivity (D), both of which are functions of 

soil water content (0), we used standar d 

laboratory and computational procedures 

(Richards 1948, Green and Corey 1971) . The 

so-called moisture release curve, a plot of ,1i 

against 0, was also obtained in the laboratory . 

Curves for K, D, and as functions of 0 are 

shown in figure 1 . 

Our Cedar River core samples of the 

Everett gravelly sandy loam were biased to an 

unknown degree by the removal of the large 

stone fraction before the laboratory analysis . 

These stones constitute perhaps up to 20 per - 

cent of the volume of the Everett soil series . 

Because these stones tend to impede the flo w 

of water in unsaturated soils-the path aroun d 

them is longer than it would be for smaller 

obstructions-we believe our estimates of K 

may be somewhat too large . This decrease in 

conductivity associated with the presence o f 

large stones is especially noteworthy if th e 

soil is highly unsaturated because water 

strongly held by matric forces flows only 

along capillary paths around the stones an d 

not directly through the interstices of aggregations 

of large stones and cobbles . 

Since conductivity of water in unsaturate d 

soils is extremely difficult to measure directly, 

it has become common practice to estimate 

this parameter from pore-size distribution 

data, or equivalently, from the soil 

moisture release curve. Green and Corey 

(1971) have reviewed three methods fo r 

calculating K based on pore-size distribution 

and found that all give good results whe n 

compared with measured data . Our curve for 

K as a function of 0 is based on laboratory 

measurements of K for the saturated Everett 

soil and adjusted for unsaturated soils by us e 

of the Marshall pore-interaction relationship 

with a matching factor (Marshall 1958) . 

The shape of the moisture release curve i s 

also affected by sampling procedures. Removal 

of large stones clearly biases the result s 

in the direction of a soil with smaller pores . 

Accordingly, the correct relationship is some - 

what to the left of the one determined in the 

laboratory . This implies that the slope aO i s 

different from that indicated by our laboratory 

data and that our estimate of the dif - 

fusivity D = -K is biased . 

ao 

Computational Procedures 

We used the Remson computer program , 

which provides an approximation to the solution 

of (3). As we have already noted, it is 

possible to calculate at the same time a n 

approximation to the volume flux of water , 

v = -D ae, at any depth in the soil. For thi s 

calculation initial moisture content must b e 

specified at a number of equally space d 

points, called nodes, along the vertical soil 

profile. Our nodes were set at depths of 1, 3 , 

5, 7, 9, and 11 cm. Initial moisture content 

was measured at approximately 5 .5 cm by 

tensiometer. The lysimeter plate at 11 cm 

depth provided another point of known 

moisture content . Because suction at the 

lysimeter plate was maintained at a constant 

level, the moisture content at that point coul d 

be obtained directly from the empirically 

determined moisture release curve. Estimate s 

of initial values at other nodes were obtaine d 

by linear extrapolation . Moisture content at 

the upper boundary (the soil surface) varie d 

with incoming precipitation . 

Results and Discussion 

Observed flow of water through the lysimeter 

plate is compared to that predicted by 

the flow equation in figures 2, 3, and 4 which 

summarize the outcomes of three field experiments, 

and the corresponding computer runs . 

The results follow a consistent pattern . Predicted 

and observed flows begin, peak, an d 

subside at about the same time . Maximu m 

predicted flow is about 25 percent greate r 

than observed. At lower levels of flow th e 

correspondence is closer . 

We conclude that the Richards flow equa - 

97

2 

10 7100 .160 .220 .280 .340 .400 .460 .520 .580 •640 -70 0 

2 

10 100 .160 .220 .280 .340 .400 .460 .520 .580 .640 .700 

100 0 

90 0 

t.u 

I- 60 0 

70 0 

0 60 0 

50 0 

U 

40 0 

30 0 

0 20 0 

100 

0 

-10 IQQ .200 .300 400 .500 .600 .70 0 

WATER FILLED POROSITY - THETA . (CU . CM . WATER)/(CU . CM . SOIL ) 

Figure 1 . Soil water diffusivity (top), conductivity (middle), and suction (bottom) a s 

functions of soil moisture content (0) . Everett gravelly sandy loam . M indicates mea n 

values used in model . Three samples were used to obtain the moisture-release curve . 

98

F , 

20 .0 40 .0 60 .0 80 .0 100 .0 120 .0 140 .0 160 .0 180 .0 200 . 0 

ELAPSED TIME . (MIN .) 

Figure 2 . Precipitation (P), observed soil water flow (F), and modeled 

soil water flow (M) at 11 cm. Lysimeter suction was 50 cm of water . 

.0600 

120 .0 140 . 0 

Figure 3 . Precipitation (P), observed soil water flow (F), and modeled 

soil water flow (M) at 11 cm. Lysimeter suction was 88 cm of water . 

99

.0500 

.040 0 

z .030 0 

E 

L.) 

0 .020 0 

.0140 

I 

4 0 20 .0 40 .0 60 .0 80 . 0 100 . 0 120 . 0 146 . 0 

ELAPSED TIME . (MIN . ] 

Figure 4 . Precipitation (PI, observed soil water flow (F), and modele d 

soil water flow (M) at 11 cm . Lysimeter suction was 146 cm of water . 

tion provides a satisfactory description of th e 

flow of water in the upper rooting zone o f 

our Douglas-fir stand . Since the Everett 

gravelly sandy loam is extremely variable, our 

results probably should not be extrapolate d 

much beyond the local area in which the y 

were obtained . 

We have not been able to account for th e 

25 percent discrepancy between predicted 

and observed peak flows . A number of possible 

explanations have occurred to us . It can 

be argued that the lysimeter system is not a 

perfect device for measuring water flow in a 

vertical column of soil . For example, in the 

early stages of a flow experiment, acceleratio n 

of flow due to lysimeter suction may be 

greater than that due to gravity. Perhaps more 

significant may be variations in flow associated 

with the lack of homogeneity of an i n 

situ forest soil. We suspect that stones or 

roots in soil above the lysimeter plate diverte d 

water toward or away from the plate . Perhaps 

more important, our estimates of initial value s 

of 0 at several nodes are not accurate. 

It is probably more important to emphasiz e 

that the field experiments described here wer e 

conducted for purposes rather far removed 

from our present one of evaluating th e 

Richards equation as a possible component of 

a soil-plant-atmosphere model . We also wish 

to point out that the Richards equation has 

not received adequate evaluation in field situations, 

perhaps because the stringent under - 

lying assumptions are rarely satisfied . Th e 

rather close correspondence between model 

and observation that we have obtained suggests 

that the model may be more robust tha n 

has been generally believed . 



by National Science Foundation Grant 

No. GB-20963 to the Coniferous Forest Biome , 

U .S. Analysis of Ecosystems, Internationa l 


27 to the Coniferous Forest Biome . 

100

Childs, E. C., and N . Collis-George . 1950 . The 

permeability of porous materials . Roy . Soc. 

London Proc. A 201 : 392-405 . 

Gardner, W. R. 1960. Dynamic aspects of 

water availability to plants . Soil Sci . 85 : 

63-72 . 

Gessel, S . P., and D . W. Cole. 1965 . Influence 

of removal of forest cover on movement o f 

water and associated elements through soil . 

J . Ann . Water Works Assoc . 57 : 

1301-1310 . 

Green, R . E., and J . C . Corey . 1971 . Calculation 

of hydraulic conductivity : a further 

evaluation of some predictive methods . Soil 

Sci. Am. Proc . 35 : 3-8 . 


Marshall, T. J. 1958 . A relation between 

permeability and size distribution of pores . 

J. Soil Sci . 9 : 1-8 . 

Remson, I ., R. L. Drake, S . S. McNeary, and 

E. M. Wallo . 1965. Vertical drainage of an 

unsaturated soil . J. Hydraul . Div., Am . Soc . 

Civil Eng., Proc . 91 : 55-74 . 

Richards, L . A. 1931. Capillary conductio n 

through porous mediums . Physics 1 : 

318-333 . 

1948 . Porous plate apparatus fo r 

measuring moisture retention and trans - 

mission by soil . Soil Sci . 66 : 105-110. 

Rose, C . W. 1966 . Agricultural physics . 226 p . 

London: Pergamon Press . 

101



Elemental transport changes occurring 

during development of a second-growth 

Douglas-fir ecosystem 

Charles C . Grie r 

Research Associat e 

an d 

Dale W . Cole 

Associate Professo r 



Seattle, Washington 98195 

Abstract 

Mineral cycling processes in a second-growth Douglas-fir ecosystem have been monitored for nearly 10 year s 

at the A. E. Thompson Research Center in western Washington. In this interval, substantial year-to-yea r 

differences in quantities of elements transferred have been observed. For example, a comparison of transfers 

during the 1964-65 and 1970-71 measurement years showed that input of calcium by precipitation increase d 

roughly 300 percent while calcium transfer by throughfall, stem flow, and leaching from the forest floo r 

increased roughly 600 percent, 350 percent and 220 percent, respectively . In the same interval, the mass an d 

elemental content of the standing crop increased roughly 15 percent while forest floor mass and elementa l 

content remained constant. A portion of this difference is related to the increased mass accumulated in th e 

standing crop. However, the major factor causing this large difference is climatic variation . These differences in 

elemental transfer indicate the need for continuous observation of transfer processes to resolve the effects of 

ecosystem development on mineral cycling against this background of climatic variation . 

Introduction 

The growth and development of a coniferous 

forest ecosystem does not normally occu r 

in a steady linear fashion . Rather, intervals of 

rapid stand growth may be interrupted b y 

periods of slower growth reflecting changes in 

growing conditions . In turn, the cycling of 

elements within a forest ecosystem can be expected 

to reflect these changes . 

Intensive study of the mineral cycling processes 

in second-growth Douglas-fir stands ha s 

been in progress at the A. E. Thompson Re - 

search Center since 1961 . During this time, a 

substantial body of information has been collected 

concerning the amounts, pathways , 

rates of transfer, and mechanisms controlling 

transfer of elements by the mineral cyclin g 

processes of this ecosystem. Between 196 1 

and 1966, research was concentrated largely 

on detailed descriptions of ecosystem components 

(Cole, Gessel, and Dice 1967) . 

Research at the Thompson site since 196 6 

has had two major objectives : (1) to provide 

information describing the changes in mineral 

cycling processes as the stand grows an d 

(2) to examine the year-to-year variation i n 

mineral cycling processes . More recently, a 

third objective has been to examine the relationships 

between climatic factors and th e 

various mechanisms controlling rates of transfer 

along the different transfer pathways . 

During the course of research at th e 

Thompson site, a number of changes in th e 

103

patterns of elemental cycling have been observed 

. Some of these changes are related t o 

changes in the structure or mass of the ecosystem; 

other changes are related to short - 

term variations in climatic factors . The 

purpose of this paper is to examine several of 

the changes in mineral cycling processes a s 

they have been observed during the 10 years 

of research at the Thompson Research Center . 

Experimental Area 

These studies were conducted at the A . E . 

Thompson Research Center, an area specifically 

developed for the study of elemental 

cycling in second-growth Douglas-fir stands . It 

is located about 64 km southeast of Seattle , 

Washington, at an elevation of 215 m in the 

foothills of the Washington Cascades . A full 

description of the geology, soils, vegetation , 

and climate is given by Cole and Gessel 

(1968) . 

The study site is located on a glacial out - 

wash terrace along the Cedar River. This outwash 

terrace was formed during the recessio n 

of the Puget lobe of the Fraser ice sheet abou t 

12,000 years ago . 

The soil underlying the research plots is 

classified as a typic haplorthod (U .S. Department 

of Agriculture 1960) and is mapped a s 

Everett gravelly, sandy loam. This soil contains 

less than 5 percent silt plus clay an d 

normally contains gravel amounting to 50-8 0 

percent of the soil volume. The forest floor is 

classified as a duff-mull (Hoover and Lun t 

1952) and ranges from 1 cm to 3 cm thick . 

This forest floor represents the accumulatio n 

since 1931 when the present stand was established 

following logging (around 1915) an d 

repeated fires . 

The present overstory vegetation is a 

planted stand of Douglas-fir (Pseudotsuga 

menziesii (Mirb.) Franco) which was established 

about 1931 . Currently, the trees average 

about 19 m high with a crown density o f 

about 85 percent . 

The principal understory species are salal 

(Gaultheria shallon Pursh), Oregon grap e 

(Berberis nervosa (Pursh) Nutt .), red huckleberry 

(Vaccinium parvifolium Smith), and 

twinflower (Linnaea borealis L . ssp . americana 

(Forbes) Rehder) . Various mosses are 

the principal understory vegetation beneath 

the denser portions of the canopy . 

The climate is typical of foothill condition s 

in the Puget Sound Basin . Temperatures have 

ranged from -18° C to 38° C, but these extremes 

are seldom reached . The average temperature 

for July is 16 .7° C and for January is 

2.8° C. The average annual precipitation i s 

136 cm, almost all falling as rain . Precipitation 

rates are generally less than 0 .25 cm/hour 

and over 70 percent falls between Octobe r 

and March . 

Field Plots 

The facilities at the research site are designed 

to provide precision measurement of 

the flux of elements within this forest ecosystem. 

Some transfers are monitored continuously; 

others, with less potential for rapid 

change, are measured at regular intervals . A 

detailed description of the collection facilitie s 

at this site is given by Cole and Gessel (1968) . 

The following section will briefly describe 

these facilities with emphasis on recen t 

changes in the system . 

Sampling 

Precipitation collections are made at two 

locations in the stand . One collector is at the 

top of a 30-m tower and is fitted with a flowmeter; 

the other is an automatic collecto r 

(Wong Laboratories) which opens at the onse t 

of precipitation and is mounted on a 10- m 

tower in a clearcut adjacent to the plots . 

Canopy drip (foliar leaching) is collected i n 

six randomly located screened funnel s 

mounted in the necks of collection flasks . 

Stemflow is diverted from the stem of the six 

sampled trees at breast height (1 .35 m) by 

soft rubber collars and is collected in covered 

160-liter polyethylene trash cans . Litter is collected 

on eight randomly located 0 .21 m 2 

plastic litter screens. Leaching through the 

forest floor and soil is measured using the 

tension lysimeter system developed by Col e 

(1968). This system permits direct measure - 

104

ment of the flux of ions through the soil . I n 

this plot, 3 lysimeters per soil horizon are located 

in the soil at the base of the forest 

floor, at the base of the A l horizon (4 cm) 

and B 2 horizon (30 cm) and at 100 cm in th e 

C horizon . 

Instrumentatio n 

Meteorological measurements including ai r 

and soil temperature profiles, are made on a 

continuous basis utilizing a data logging system. 

Details of this instrumentation are give n 

by Cole and Gessel (1968) . 

The data logging system also permits th e 

continuous monitoring of the pH and conductivity 

of water moving through this ecosystem. 

These facilities are described by Col e 

(1968). Volumes of water flow between different 

components of the system are measured 

using the resistance flowmeter describe d 

by Cole (1968) . 

Chemical Analysis 

Analytical methods used on samples fro m 

this site have changed over the years. For 

methods used in earlier studies, the reade r 

may consult the following reports : Rahma n 

(1964); Cole and Gessel (1965) ; and Cole , 

Gessel, and Dice (1967) . The methods currently 

in use will be briefly described in the 

following paragraphs . 

Water Analysis 

Determinations of calcium, magnesium , 

potassium, and sodium are normally made 

directly on water samples using an Instrumentation 

Laboratories-353 atomic-absorptio n 

spectrophotometer . Calcium is determined in 

a nitrous oxide-acetylene flame ; and magnesium, 

potassium, and sodium are deter - 

mined in an air-acetylene flame . Nitrogen an d 

phosphorus are determined from a hydroge n 

peroxide-sulfuric acid digest (Linder and Harley 

1942) of a 10-fold concentration of th e 

solutions. Nitrogen is determined using a 

micro-Kjeldahl distillation (Jackson 1958) . 

Phosphorus is determined using the chlorostannous-reduced 

molybdophosphoric blue 

color method of Jackson (1958). Total io n 

concentration is estimated from specific conductance 

using an empirical method (Logan 

1961) and pH is determined instrumentall y 

using established methods . Further details of 

water analysis are given by Grier (1972) . 

Soil Analysis 

Total soil nitrogen is determined using th e 

standard Kjeldahl digest and distillation o f 

Jackson (1958) . Cation exchange capacity i s 

determined using the neutral ammonium acetate 

leaching method (Jackson 1958) . Ex - 

changeable cations are determined spectrophotometrically 

from the ammonium acetat e 

leachate . Soil carbon is determined instrumentally 

using a Leco carbon analyzer . Phosphorus 

is extracted from the soil by digestio n 

in 36 N sulfuric acid and 30 percent hydroge n 

peroxide ; concentrations are then determine d 

following the same procedures as for water 

samples . 

Tissue and Forest Floor Analysi s 

Elemental assays of forest floor and plant 

tissue are done using methods described by 

Grier and McColl (1971) . 


Table 1 shows how organic matter, calcium, 

potassium, nitrogen, and phosphoru s 

were distributed among different component s 

of the ecosystem at the Thompson site at th e 

end of 1965 (Cole, Gessel and Dice 1967) . 

Since the time of this determination, transfer s 

between these various components have bee n 

monitored to assess the changes and rates of 

change of elemental distribution in the stand . 

In the interval between 1966 and the present, 

the volume of woody tissue in the stan d 

has increased by an average of 11 m 3 /ha/yf ' . 

This represents approximately a 15-percent 

increase in volume since 1966 . Increases in 

mass and elemental content of most components 

of the standing crop are estimatedusing 

estimation procedures developed fo r 

105

these stands by Dice (1970)-to be proportional 

to the increase in woody tissue mass . 

However, the foliar mass has probably remained 

about constant in this interval . Thus 

at the present time, the quantities of elements 

and organic matter in the standing crop at this 

site are about 15 percent greater than show n 

in table 1 . 

In the same interval, other components of 

the ecosystem have shown relatively little 

change. For example, the forest floor was intensively 

sampled in 1961, 1 1965 (Cole, Gessel, 

and Dice 1967), and 1969 (Grier and Mc - 

Coll 1971) . Results of these studies are summarized 

in table 2 . With the exception o f 

magnesium content, the forest floor of thi s 

D . W . Cole . Unpublished data on file at the Colleg e 

of Forest Resources, University of Washington, Seattle . 

Table 1 .-Distribution of N, P, K, Ca, and organic matter (g/m 2 ) in a 35-year-old second-growt h 

Douglas-fir ecosystem at the Thompson Research Center (Cole, Gessel, and Dice 1967 ) 

Component N P K Ca 

Organic 

matte r 

TREE 

Foliage current 2.4 0 .5 1 .6 0.7 199 . 0 

older 7.8 2 .4 4.6 6.6 710 . 7 

Branches current .4 .1 .3 .2 51 . 3 

older 4.0 .9 3.2 6.5 1,337 . 3 

dead 1.7 .2 .3 3.9 814 . 5 

Wood current 1.0 .2 1 .0 .4 748 . 5 

older 6.7 .7 4.2 4.3 11,420 . 2 

Bark 4.8 1.0 4.4 7.0 1,872 . 8 

Roots 3.2 .6 2.4 3.7 3,298 . 6 

Total tree 32 .0 6.6 22 .0 33 .3 20,452 . 9 

SUBORDINAT E 

VEGETATION .6 .1 .7 .9 101 . 0 

FOREST FLOO R 

Branches .5 .1 .4 .8 142 . 3 

Needles 3.5 .4 .5 2.7 300 . 5 

Wood 1.4 .2 .8 1 .7 634 . 5 

Humus 12 .1 1.9 1 .5 8.5 1,199 . 9 

Total forest floor 17 .5 2.6 3.2 13 .7 2,277 . 2 

SOIL 

0-15 cm 

80 .9 116 .7 7 .9 31 .3 3,837 . 2 

15-30 cm 86 .8 119 .5 6 .6 19 .6 3,693 . 5 

30-45 cm 76 .1 98 .0 5 .2 15 .2 2,829 . 0 

45-60 cm 37 .1 53 .6 3 .7 8.0 795 . 5 

Total soil 280 .9 387 .8 23 .4 74 .1 11,155 . 2 

TOTAL ECOSYSTEM 331 .0 397 .1 49 .3 122 .0 33,986 .3 

106

Table 2 .-Forest floor composition changes during an 8-year period in 

a Douglas-fir ecosystem at the Thompson Research Center 

Year 

sampled 

Dry 

weight 

N 

g/m 2 

P Ca Mg K 

1961 1,542 15 .4 2.2 11 .2 3 .8 2 . 4 

1964 1,500 16 .1 2.4 12 .0 3.7 2 . 4 

1969 1,430 14 .0 12 .3 1.9 2 .7 

ecosystem has changed only slightly since 196 1 

indicating that litterfall is balanced by decomposition. 

Similarly, there has been no significant 

change in elemental capital in the soi l 

(table 3) due both to the large reserves of mos t 

nutrients in this soil and to efficient cyclin g 

processes within the ecosystem . The decreas e 

in magnesium capital in the forest floor between 

1965 and 1969 may indicate increased 

rates of internal redistribution of magnesiu m 

within the standing crop . However, this i s 

doubtful in view of the small change in magnesium 

capital in the soil (table 3) . 

The quantities of cations transferred between 

components of the ecosystem generall y 

increased between 1964-65 and 1970-71 (table 

4). Comparison of some transfers during th e 

1970-71 measurement year with those of th e 

1964-65 measurement year (table 4) reveal s 

some rather substantial differences in both th e 

overall amounts transferred and in the relativ e 

importance of the various pathways . 

For example, input of calcium and potassium 

by precipitation is roughly four-fol d 

greater now than in 1964-65. This increas e 

may reflect a general increase in atmospheri c 

pollution in the Puget Sound basin . The quantities 

of calcium and potassium returned t o 

the soil from the aboveground vegetation hav e 

also increased since the 1964-65 measurement. 

As an example, return of calcium an d 

potassium by stemflow has increased 25- and 

35-fold, respectively, since measurement during 

1964-65 . Increases in return by foliar 

leaching also were observed (table 4) bu t 

these were not of the same magnitude as th e 

Increased return by stemflow. No explanation 

has been found for the differential increases 

in return by stemflow and foliar leaching . 

Return of elements by litterfall also in - 

creased between 1964-65 and 1970-71 (table 

4) . These differences are probably due both 

to normal year-to-year fluctuation of litterfal l 

and also to the increasing proportion o f 

branch material observed as a component o f 

the litter . 

Increases were also observed in leaching of 

elements through the soil profile . Leaching of 

calcium and potassium from the forest floo r 

roughly doubled between 1964-65 an d 

1970-71 (table 4) . A portion of this increase 

may result from increased rates of litter de - 

composition as well as the increased input in 

precipitation . 

As previously noted, the forest floor mas s 

appears to have approached a steady state 

condition (table 2) ; the present mass representing 

a quasi-equilibrium state, achieve d 

since the stand was established in 1931, afte r 

repeated fires through the area . On the other 

hand, litterfall quantity has increased durin g 

the period of observation of this stand. In 

table 5, monthly litterfall measured at thi s 

site in 1962-63 (Rahman 1964) is compared 

with measurements made during 1970-71 ; the 

annual mass of litterfall has apparently increased 

roughly 50 percent over this 8-yea r 

period. Thus the increased quantities of ele- 

10 7

Table 3.-Elemental capital in the soil of a second-growth Douglas-fir ecosystem 

at the Thompson Research Center as determined in 1965 and 197 1 

Depth 

g/m 2 

N P K Ca Mg 

A) 1965 : 

0-15 81 117 7.9 31 3 . 7 

15-30 87 120 6.6 20 3 . 5 

30-45 76 98 5.2 15 2 . 5 

45-60 37 54 3.7 8 1 . 2 

Total soil 281 389 23 .4 74 10 . 9 

B) 1971 : 

0-15 86 123 8.3 40 3 . 2 

15-30 82 111 6.6 21 3 . 3 

30-45 71 104 5.3 10 2 . 7 

45-60 40 52 4.2 11 1 . 5 

Total soil 279 390 24 .4 82 10 .7 

Table 4. Transfers of elements between components of a Douglas-fir ecosyste m 

at the Thompson Research Center during two stages of development ' 

Item 

Calcium (g/m 2 ) Magnesium (g/m 2 ) Potassium (g/m 2 ) Sodium (g/m 2 ) 

1 2 1 2 1 2 1 2 

Precipitation 0.28 0.93 ND 0.22 0.08 0.47 ND 1 .6 8 

Throughfall .35 2.05 ND .49 1 .07 2 .83 ND 2 .5 9 

Stemflow .11 3.85 ND .60 .16 4.00 ND 2 .3 0 

Litterfall 1 .11 1.78 ND .20 .27 .47 ND .0 7 

Leaching fro m 

forest floor 1 .74 3.87 ND 1 .10 1 .05 2.60 ND 2 .1 8 

1 1=transfers measured during 1964-65 measurement year . Reported by Cole, Gessel and Dice (1967) . 

2=transfers measured during 1970-71 measurement year . 

108

Table 5 .-Monthly amounts of litterfall during two intervals in th e 

development of a second-growth Douglas-fir ecosyste m 

Month 

Quantity (g/m 2 ) 

1962-63 1970-7 1 

October 73 .0 89 .0 

November 17 .9 44 . 5 

December 9.2 14 . 2 

January 6.5 10 . 0 

February 15 .9 3 . 8 

March 6 .2 13 . 8 

April 7 .4 5 . 8 

May 12 .8 ( I ) 

June 11 .4 ( 1 ) 

July 5 .2 39 .1 1 

August 5 .5 21 . 0 

September 12 .3 41 .6 

' Litterfall amounts for May, June, and July 1971 are combined in July collection . 

ments leached from the forest floor are a t 

least partially due to increased litter decomposition 

rates . 

Many of the changes in amount and relative 

importance of the transfer pathways sinc e 

1965, are probably due to changes in the mass 

and structure of this forest ecosystem . How - 

ever, many of the changes are completely ou t 

of proportion to estimates of increased mas s 

of the standing crop . 

Observations made since 1966 at th e 

Thompson site indicate that the return of elements 

to the soil and their subsequent distribution 

through the soil profile is sensitive t o 

certain climatic factors ; primarily precipitation 

and temperature . Variations in these 

climatic factors have been observed to cause 

substantial variation in quantities transferred 

along most pathways . As an example, the re - 

turn of elements to the soil by stemflow an d 

foliar leaching were found to be partially 

regulated by the distribution of precipitation ; 

especially precipitation occurring during th e 

summer. Figure 1 shows the monthly return 

of potassium by stemflow from one tree fo r 

1970 and 1971 ; the total amounts transferred 

being 0 .50 gm during 1970, and 0 .75 gm during 

1971 . Much of the greater amount re - 

turned during 1971 was returned during a fe w 

summer rainstorms . Foliar leaching, although 

not shown, exhibits a similar pattern of behavior. 

Apparently the extent of removal of 

potassium and other elements by leachin g 

from Douglas-fir foliage is related to its 

phenological stage . This suggests that precipitation 

occurring in certain critical parts of the 

year may remove elements that would other - 

wise remain in the plant . 

Temperature, the other major factor causing 

variation in year-to-year transfer rates , 

exerts its control on transfers primaril y 

through its effect on organisms active in de - 

composition . Decomposer activity regulate s 

the availability of nutrient elements in tw o 

ways ; first, by regulating the rate of mineralization 

of elements and second, by its influence 

on levels of the mobile bicarbonate 

anion in the soil solution . 

10 9

(A . ) 

f} Q 

2 0 

Figure 1 . 

A. Volume of stemflow from a 15-cm diameter Douglas-fir during 2 successive years . 

B. Quantity of potassium transferred to soil from a single 15-cm diameter Douglas-fir during 2 successive years . 

.2 

i 

N 

U 

N 

O 

U 

0 

20 

15 

4 JUNE 

6 - 7 JUN E 

1 k i 

1 1 k l 1 

2400 

O6DO 1200 1800 

TB[ OF DAY 

Figure 2 . Temperature effects on CO2 evolution from forest floor at Thompson Research Center (Ballard 1968) . 

110

Figure 2 illustrates how CO 2 production 

from the forest floor is related to temperatur e 

in this ecosystem (Ballard 1968) . Temperature 

effects on mineralization rates are probably 

directly proportional to the temperatur e 

effects on CO 2 production rates . 

The bicarbonate ion is the major anion i n 

soil solutions at the Thompson site. Moreover , 

the concentration of this ion in the soil solution 

has been shown to be the major facto r 

controlling rates of cation leaching in these 

soils (McColl 1969) . Temperature effects o n 

the levels of the bicarbonate ion are related t o 

decomposer activity, since the equilibrium between 

CO2 in the soil atmosphere and HCO3 

in the soil solution is related to CO 2 concentration 

. Consequently, leaching rates in th e 

soil of the Thompson site are strongly relate d 

to temperature (McColl 1969) . 

Figure 3A illustrates the relationship between 

estimated ion concentration in fores t 

floor leachates and air temperature . As can be 

seen from these data, the ion concentratio n 

closely follows temperature during the year . 

Monthly variations in mean ion concentration 

of forest floor leachates for 3 successive years 

(figure 3B) are primarily the result of differences 

in mean monthly temperature . 

Total transfer by leaching is regulated both 

by temperature and the amount and date of 

-1C 

~'' 0''404%4P` EC I P . CONDUCTANC E 

ONTHLY PRECIP . 

~. _ _ -,'• 

■ 

JAti FEB MAR APR NAY JUNE JULY AUG SEPT OCT NOV DE C 

0 

0 

JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DE C 

Figure 3 . 

A. Mean monthly precipitation, air temperature, and specific conductance of precipitation and forest floo r 

leachates for the Thompson site during 1970 . 

B. Specific conductance of forest floor leachates for 3 consecutive years at the Thompson site . 

11L

Figure 4 . Rates of calcium leaching through soil profile and rate of precipitation during 1970-71 growing season . 

occurrence of precipitation . The largest quantities 

of elements are transferred during 

periods of relatively high temperatures and 

ample precipitation . Figure 4 illustrates the 

relationship between leaching of calcium from 

the forest floor and A horizon of the minera l 

soil in 1970-71 and precipitation in tha t 

interval . As can be seen from these data, the 

greatest quantities of calcium were transferre d 

during the spring and autumn when decomposer 

activity was relatively high and sufficient 

precipitation fell to flush the decomposition 

products into the soil . 

Conclusions 

Mineral cycling processes within a forest 

ecosystem are responsive to changes in th e 

mass and structure of the standing crop an d 

also to the shorter term effects of climatic 

variation . Of the two, the effects of climatic 

variation are the most readily seen . For this 

reason, any efforts to describe mineral cycling 

processes within a forest ecosystem must include 

sufficient data that effects of changes in 

mass and structure of the ecosystem can be 

resolved and separated from the large year-toyear 

effects of climatic variability . 

Present and future mineral cycling research 

at the Thompson site is directed towards 

separating and describing the possibly inter - 

dependent effects of growth and climate on 

mineral cycling processes in second-growt h 

Douglas-fir . 




Grant No . GB-20963 to the Coniferous 

Forest Biome, U .S. Analysis of Ecosystems , 

International Biological Program. This i s 

Contribution No. 28 to the Coniferous Forest 

Biome, IBP . 

112

Ballard, T. M. 1968 . Carbon dioxide production 

and diffusion in forest floor materiala 

study of gas exchange in biologically 

active porous media. 120 p . Ph .D . thesis on 

file, Univ . Wash., Seattle . 

Cole, D . W . 1968 . A system for measuring 

conductivity, acidity, and rate of flow in a 

forest soil . Water Resour. Res. 4 : 

1127-1136 . 

and S . P . Gessel . 1965 . Movement 

of elements through a forest soil a s 

influenced by tree removal and fertilizer 

additions . In C. T . Youngberg (Ed .) Forestsoil 

relationships in North America, p . 

95-104. Corvallis, Oreg. : Oreg. State Univ . 

Press . 

and S. P . Gessel . 1968 . Cedar 


pathways, rates and processes of elemental 

cycling in a forest ecosystem. For. Resour . 

Monogr . Contrib . No . 4, 53 p . Univ. Wash . , 

Seattle . 

S. P. Gessel, and S . F. Dice . 

1967 . Distribution and cycling of nitrogen , 

phosphorus, potassium, and calcium in a 

second-growth Douglas-fir ecosystem . In 

Primary productivity and mineral cycling in 

natural ecosystems-symposium, p . 

193-197. Assoc . Advan. Sci. 13th Annu . 

Meet. Orono : Univ . Maine Press . 

Dice, S . F. 1970 . The biomass and nutrien t 

flux in a second-growth Douglas-fir ecosystem 

(a study in quantitative ecology) . 

165 p . Ph .D . thesis on file, Univ . Wash . , 

Seattle . 


Grier, C . C. 1972 . Effects of fire on the movement 

and distribution of elements within a 

forest ecosystem. 167 p. Ph.D. thesis on 

file, Univ . Wash ., Seattle . 

and J. G . McColl . 1971 . Forest 

floor characteristics within a small plot i n 

Douglas-fir in western Washington . Soil Sci . 

Soc. Am. Proc . 35 : 988-991 . 

Hoover, M . D., and A . H . Lunt . 1952 . A key 

for the classification of forest humus types . 

Soil Sci . Soc. Am. Proc . 16:368-370 . 

Jackson, M. L. 1958 . Soil chemical analysis . 

498 p. Englewood Cliffs, N .J . : Prentice- 

Hall, Inc . 

Linder, R . C., and C . P . Harley . 1942 . A rapid 

method for the determination of nitroge n 

in plant tissue . Science 96 : 565-566 . 

Logan, J. 1961 . Estimation of electrical conductivity 

from chemical analysis of natural 

waters. J . Geophys . Res . 66: 2479-2483 . 

McColl, J . G . 1969. Ion transport in a forest 

soil : models and mechanisms . 214 p . Ph.D . 

thesis on file, Univ . Wash., Seattle . 

and D. W. Cole. 1968 . A mechanism 

of cation transport in a forest soil . 

Northwest Sci . 42 : 134-140 . 

Rahman, A . H. M . M. 1964. A study of the 

movement of elements from tree crowns b y 

natural litterfall, stemflow and leaf wash . 

117 p. M.F. thesis on file, Univ . Wash . , 

Seattle . 

U .S. Department of Agriculture. 1960. Soil 

classification-a comprehensive system-7th 

approximation. 265 p. Washington, D .C . : 

U .S. Govt. Print. Office . 

113



Nutrient budget of a Douglasfir 

forest on an experimenta l 

watershed in western Oregon 

R . L . Fredrikse n 

Soil Scientist, Forestry Sciences Laborator y 

Pacific Northwest Forest and Range Experiment Statio n 

Forest Service, U .S . Department of Agricultur e 


Abstract 

Annual loss of nitrogen, phosphorus, silica, and the cations sodium, potassium, calcium, and magnesiu m 

followed the same pattern as annual runoff which is heavily dominated by winter rainstorms arising from th e 

Pacific Ocean. Even though 170 and 135 cm of water passed through this Douglas-fir ecosystem for the 2 year s 

reported here, this ecosystem conserved nitrogen effectively as indicated by an average annual dissolved nitroge n 

outflow of 0.5 kg/ha from an annual average input of 1 .0 kg/ha in precipitation . There was a small annual ne t 

loss of phosphorus (0.25 kg/ha) . Cation input in precipitation was less than 10 percent of sources from minera l 

weathering-indicating that mineral weathering was the principal source of cations to the system. Average 

annual net losses of calcium, sodium, magnesium, and potassium were : 47, 28, 11, and 1 .5 kg/ha, respectively. 

Silica loss of 99 kg/ha-yr was the largest of all constituents and came entirely from within the forest system . 

Although loss of sediment was low during the period of study, loss of nutrients by soil erosion may become of 

major importance over a longer time scale due to widely spaced unsampled catastrophic erosion . 

Introduction 

The purpose of this study was to measur e 

the inputs, losses, and retentions of specifi c 

plant nutrients . 

Forested watersheds have been used fo r 

decades to study the hydrologic cycle . 

Bormann and Likens (1967) indicated tha t 

small watersheds are suitable for studies of 

chemical cycles if the ecosystem is located o n 

a watershed with relatively impermeable bed - 

rock . Then, all annual chemical fluxes int o 

and out of the forest system can be attribute d 

to a definable forest ecosystem, and error s 

due to deep seepage can be minimized . Small 

watersheds containing soil-plant systems hav e 

been generally accepted as useful units fo r 

ecological research . 

Nutrient cycles begin with the establishment 

of vegetation, and the quantity o f 

nutrients cycled undoubtedly increases as the 

forest passes through successional phases . At 

all points, along this developmental sequence , 

the supply of nutrients in the soil-plan t 

system is regulated by the balance betwee n 

(1) the inputs from the atmosphere an d 

mineral weathering and (2) the outflow b y 

soil erosion or chemicals dissolved in stream 

water. A continued level of fertility of th e 

system therefore depends upon the retentio n 

of nutrients in the cycle from loss by leachin g 

and soil erosion . 

Nutrient budget studies were begun on two 

small watersheds at the H . J. Andrews Experimental 

Forest in 1967 . Only the initial results 

from one of these and the chemicals in solution 

will be reported here . The followin g 

questions were asked : 

1 . What quantities of nitrogen, phosphorus , 

the cations (sodium, potassium, calcium , 

magnesium), and silica reach the site i n 

precipitation? 

115

2. How effectively are these chemicals retained 

by the forest ? 

3. What hypotheses can be made about th e 

processes that regulate the gain or loss o f 

these chemicals from the forest ? 

The 10.1-hectare study watershed, designated 

number 10 in Forest Service studies , 

occupies strongly dissected topograph y 

characteristic of the west side of the Cascade 

Range. It rises from 430 m elevation at the 

stream gaging station to 670 m at the highes t 

point on the back ridge. The overall slope is 

44 percent, but side slopes and the headwal l 

range up to 90 percent due to deep incision of 

the basin into the main ridge . Soils, fro m 

weathered tuff and breccia materials, hav e 

weakly developed profiles that often overlie 

saprolite up to 6 m deep . Tree roots are mos t 

abundant in the surface 0 .6 m and probably 

do not penetrate much deeper than 2 .4 m . 

The saprolite is porous and therefore can 

transmit water . The average annual precipitation 

is 230 cm . 

Forest vegetation on the watershed consist s 

of remnants of a 450-year-old Douglas-fi r 

(Pseudotsuga menziesii) stand and islands of 

various younger age classes which originate d 

when the old-growth stand was broken up by 

windthrow, disease, or fire . Because the basin 

is oriented toward the southwest, the sid e 

slopes within it face south, west, and north . A 

considerable diversity in vegetation on thes e 

slopes has recently been classified for th e 

Experimental Forest by Dyrness, Franklin , 

and Moir (personal communication) . Douglasfir 

climax forests occupy ridgetop and upper 

slope south aspect positions on the watershed . 

A dense evergreen shrub layer, compose d 

mainly of golden chinkapin (Castanopsis 

chrysophylla), Pacific rhododendron (Rhododendron 

macrophyllum), and salal (Gaultheria 

shallon) codominates the site beneath a spars e 

stand of old-growth trees. Western hemloc k 

climax forests make up the remainder of th e 

vegetation on the watershed . On these habitats, 

the dominant overstory is mainl y 

Douglas-fir and western hemlock (Tsuga 

heterophylla) . Proceeding from west to north 

aspects along a gradient of increasing moisture, 

the shrub cover shifts from vine maple 

(Acer circinatum)-salal, to rhododendron - 

Oregon grape (Mahonia nervosa) and finally 

to vine maple-western sword fer n 

(Polystichum munitum) on moist, northfacing 

middle and lower slopes . 

Methods 

Precipitation and streamflow are sample d 

by the following techniques . Precipitation at 

the top and bottom of the watershed i s 

measured by standard rain gages ; samples for 

chemical analysis are obtained from a specia l 

collector mounted on a tower 18 m tall in th e 

center of a nearby clearcut . At the outlet o f 

the watershed, an H-flume gages streamflo w 

and a proportional water sampler, capable of 

sampling both dissolved and suspended materials 

from the stream, composites discret e 

water samples in a polyethylene carbo y 

(Fredriksen 1969) . The samples are remove d 

at approximately 3-week intervals throughou t 

the year. In 1970 and 1971, grab sample s 

were taken from low summer flows ; and during 

the winter of 1971-72, a series of gra b 

samples were collected from streamflow during 

storm runoff events . In the laboratory , 

solids are removed by filtration through a fine 

glass fiber filter (Whatman GF/C) from bot h 

precipitation and streamflow samples and 

chemical analyses were performed on bot h 

fractions. The fraction that passes the filter i s 

considered to be dissolved although we recognize 

that the filtered water may contain som e 

particles smaller than 2 microns in diameter . 

Since the concentration of each chemica l 

constituent is an estimate of the mean for th e 

time during which the sample was collected , 

the input or loss for each sampling period is 

the product of precipitation or streamflow, 

respectively, and the concentration of eac h 

constituent . 

Chemical analyses were done in duplicat e 

as follows : Ammonium and dissolved organi c 

nitrogen by distillation and digestion, respectively, 

on 1/2-liter samples and detection b y 

Nesslerization; nitrite by the sulphanilamid e 

method ; nitrate by reduction and detection a s 

nitrite ; orthophosphorus by the molybdat e 

blue method ; total phosphorus by the 

molybdate blue method following a per - 

116

sulfate-sulfuric acid digestion in the autoclave ; 

reactive silica by the molybdate yello w 

method ; cations sodium and potassium by 

flame emission ; and calcium and magnesium 

by atomic absorption following addition of 

lanthanum as a masking agent . 

Detection limits (mg/1) for the following 

chemicals were : organic nitrogen, 0 .005 ; 

nitrate and nitrite, 0 .0005 ; phosphorus , 

0.005; sodium, 0.05 ; potassium, 0 .1 ; calcium , 

0.4; magnesium, 0.05 ; and silica, 0 .1 . 

Results 

Nutrient Input in Precipitatio n 

Total amounts of nutrients dissolved i n 

precipitation and the annual rainfall are give n 

in table 1 . Nitrogen enters the forest syste m 

mainly in organic substances ; nitrate average d 

only 13 percent of all nitrogen forms for th e 

2-year period. In that period, only trac e 

amounts of ammonia nitrogen were detected , 

and a small quantity of nitrite in 1970. The 

phosphorus came mainly in organic form in 

two collections in the fall of 1969 . Of the 

cations, input of calcium was the greates t 

followed by sodium, magnesium, and potassium 

in descending order . Traces of silica were 

found in only five of the 14 collections in 

1970 . 

There was also an input of particulate 

matter from the atmosphere that measure d 

upwards of 3 to 5 kg/ha-yr . Although the 

average dustfall for 3 years was nearly equally 

divided between the dry season (April - 

October) and the wet season (November- 

March), the proportion in the wet season 

varied from 33 to 65 percent of the total fo r 

individual years . 

Nutrient Loss in Streamflo w 

Nitrogen is transported in the stream al - 

most entirely in the form of organic matter 

(table 2). Though water samples from thi s 

watershed were contaminated with nitrat e 

from the exhausts of propane heaters, the 

nitrate nitrogen outflow from another un- 

Table 1.-Nutrient input in precipitation for th e 

period October 1 through September 3 0 

Constituent in 

rainwater 

1969 1970 

-------------------- hg/ha-yr 

NO 3 -N 0 .06 0 .2 0 

N0 2 -N 0 .0 1 

Organic N 1 .02 .69 

All N 1 .08 .9 0 

Ortho P ( 1 ) .0 1 

Total P ( 1 ) .2 7 

Na 1 .17 2 .3 4 

K .27 .1 1 

Ca 7 .65 2 .3 3 

Mg .72 1 .3 2 

Si (' ) ( 2 ) 

Precipitation (cm) 251 215 

1 Missing data. 

2 Trace . 

11 7

Table 2 .-Nutrient budget for dissolved constituents of precipitation an d 

runoff for the period October 1 through September 3 0 

Element 

1969 1970 

Input Outflow Net Input Outflow Net 

N L08 +0 .50 0 .90 0 .38 +0 .5 2 

Total P (') (') 1 (') .27 .52 -.2 5 

Na 1 .17 33 .66 -32.49 2 .34 25 .72 -23 .3 8 

K .27 1 .24 -.97 .11 2 .25 -2 .1 4 

Ca 7 .65 53 .65 -46 .00 2 .33 50 .32 -47 .9 9 

Mg .72 12 .70 -11 .98 1 .32 12 .44 -11 .1 2 

Si (') (') (') (2 ) 99 .3 -99 . 3 

H2 0-cm 251 170 81 215 135 80 

' Missing data . 

2 Trace . 

disturbed watershed nearby was 0 .00 4 

kg/ha-yr . A similar nitrogen outflow can be 

expected for the study watershed . Ammonium 

and nitrite do not occur in measurable 

concentrations . 

Total phosphorus outflow, both organi c 

and inorganic, was about in the same order o f 

magnitude as outflow for nitrogen. The 

organic form predominates in spring, summer, 

and fall seasons but reaches minimum value s 

in midwinter when the forest and soils are 

repeatedly flushed by rainwater . Orthophosphorus 

concentration remained nearl y 

constant throughout the year, varying fro m 

0.01 to 0 .02 mg/l. 

Cation and silica outflows were much large r 

than those of nitrogen and phosphorus . Potassium 

export was much smaller than that of 

other cations . Silica outflow was the largest of 

all the dissolved constituents . Transport of 

suspended sediment was 67 and 37 kg/ha-y r 

for 1969 and 1970, respectively . 

Net Loss from the Watershed 

The nutrient budget in table 2 results from 

the subtraction of the nutrient outflow in the 

stream and the input in precipitation . Net 

loss or gain is indicated by positive and negative 

values, respectively . 

The small annual nitrogen gain of about 0 . 5 

kg/ha was nearly the same for the 2 years . For 

all other mineral constituents of runoff, more 

was lost than was gained in precipitation . 

Phosphorus losses are relatively small and in 

the same order of magnitude as nitroge n 

losses. The cation and silica losses were fro m 

one to two orders of magnitude larger than 

those of nitrogen and phosphorus . Of th e 

cations, calcium losses were greatest followe d 

in descending order by sodium, magnesium , 

and potassium. Of all constituents, silic a 

losses were greatest . Though precipitation an d 

runoff were greater in 1969, the use of water 

by the forest was nearly equal for the 2 years . 

118

Time Trends of Outflow and Concentratio n 

of Selected Chemicals 

Movement of water in the forest system i s 

obviously the dominant factor regulating th e 

flow of chemicals . However, moderation o f 

the annual chemical outflow pattern can b e 

expected from processes such as organi c 

decomposition, uptake into plants, an d 

mineral weathering . These processes may b e 

regulated by the hydrologic and thermal state 

of the forest system. 

The time trends of soil temperature an d 

water content are indicated on figure 1 . Be - 

cause precipitation occurs mainly during th e 

fall and winter months with less than 2 percent 

of the annual total in July and August, 

substantial withdrawals of water from the soi l 

are made by the forest during the summer . 

Water content of the top 30 cm of soi l 

reaches minima, near the wilting point , 

usually in late August or early September . 

Fall rainstorms normally recharge the soi l 

mantle to field capacity by the end o f 

December, although recharge of the surfac e 

soil may occur from 1 to 2 months earlier . 

The soil temperature cycle is the inverse o f 

the soil water cycle . Maxima near 14° C occur 

in August and minima from 0° to 2° C are 

common from December through March . 

Precipitation and runoff for 1970 are illustrated 

in figure 2 for the periods that composite 

streamflow and precipitation samples were 

taken. Precipitation exceeded runoff by a 

wide margin in the fall of 1969 until soils 

reached field capacity sometime in late 

December or early January . The soil remained 

near field capacity until precipitation fell to 

low levels in early May . Wetting fronts from 

precipitation and snowmelt can pass through 

the soil mantle of the watershed during this 

period. Runoff declined rapidly with the in - 

crease of evapotranspiration and reduction of 

precipitation in the spring and summe r 

months . 

Organic nitrogen, calcium, and silica wer e 

chosen to illustrate the differences that exis t 

in annual concentration and outflow patterns 

(figs . 3, 4, and 5) . The influence of water ca n 

be seen immediately from the general correspondence 

of the stream runoff and chemical 

outflow cycles . Minimum chemical outflo w 

occurs in summer and fall while streamflow i s 

low; maximum chemical outflow comes with 

high runoff in the winter months. However , 

the concentration patterns vary considerabl y 

between the three, and there are notabl e 

differences in outflow between organic nitrogen 

and the other two (calcium and silica) 

during the winter season . 

Organic Nitroge n 

The high concentration of organic nitrogen 

in October 1969 (fig. 3) came at a time when 

soils were dry (fig. 1) and could store all 

precipitation except that which falls directly 

into the stream or runs off rock and shallo w 

soil areas adjacent to the stream . After a 

decline in early November (fig . 3), the concentration 

and outflow rose to a peak in 

December when soils probably had reached 

field capacity (figs . 1 and 2) and wetting 

fronts from precipitation could pass through 

the soil mantle . Organic nitrogen outflo w 

remained high and reached a secondary peak 

in late January at peak runoff (fig . 3). As 

precipitation and runoff declined in February , 

concentration and outflow reached low value s 

(figs. 2 and 3). The concentration and outflow 

rose again in late winter and early sprin g 

as precipitation increased (figs . 2 and 3) . Outflow 

and concentration declined in April an d 

May as precipitation declined and soil wate r 

storage dropped below field capacity (figs . 1 , 

2, and 3) . Increased concentration in summe r 

had little effect on outflow due to very low 

runoff (fig . 3) . 

Though the forest system retains mor e 

organic nitrogen than it loses (table 2), th e 

gain is not uniform throughout the year . 

From mid-December to mid-February, th e 

forest system lost more in outflow than i t 

gained in precipitation; for the remaining 1 0 

months, there was a net retention of organi c 

nitrogen (fig . 6). For the 2-month midwinte r 

period, the outflow in runoff (0 .28 kg/ha) 

exceeded the input in precipitation 

(0.16 kg/ha) by 0 .12 kg/ha. There was an in - 

creasing margin during the 2 months . 

Contributions to the stream from the fores t 

119

Figure 1 . Water content of the surface 30 cm and temperature at 15 cm in the soil of an old-growth Douglas-fi r 

forest. 

120

30_ 

20_ 

E 

U 

10- 

0 

Figure 2 . Precipitation and runoff for an undisturbed Douglas-fir forest for each sampling period, 1969-70 . 

121

0.12_ _30 

rn 

z 

0 

Q 

z 

w 

0 

z 

0 

0 

z 

Q 

c 

o.10_ 

0.08_ 

0.06_ 

rn 0.04_, 

- 

.-.Outflow 

Concentratio n 

___ Runoff 

_20 

Jo 

E 

U 

0 

u_ 0.02_ 

I 

D 

0 

0.00 

2 r 0 

Figure 3 . Concentration and outflow of organic nitrogen dissolved in stream water and runoff per samplin g 

period from an undisturbed Douglas-fir forest, 1969-70. 

122

_20 

E 

U 

_10 

0 

10' '12 ' 

4- 1969-* 

' 2 1 1 4 1 

1970 

1 6 ' 1 8 

0 

Figure 4 . Concentration and outflow of calcium dissolved in stream water and runoff per sampling period fro m 

an undisturbed Douglas-fir forest, 1969-70 . 

123

_3 0 

. Outflow 

Concentration 

_ __ Runoff 

E 

U 

0 

Figure 5 . Concentration and loss of silica dissolved in stream water and runoff per sampling period from a n 

undisturbed Douglas-fir forest, 1969-70 . 

124

Figure 6. Input in precipitation and loss in runoff of dissolved organic nitrogen for each sampling period , 

1969-70 . 

125

0.10_ 

0.08 _ 

Pea k 

0.04_ 

Concentration in precipitation 

0.02 

I I I I I I 

0.2 0.4 0.6 0.8 1 .0 1 .2 

STR EAM FLOW (m 3/ km 2 -sec ) 

Figure 7 . Concentration of organic nitrogen dissolved in streamfiow during a storm runoff event Novembe r 

25-29, 1971 . 

126

and precipitation can be illustrated from a se t 

of samples taken during a storm runoff even t 

in late November 1971 (fig. 7). The rapi d 

initial rise of organic nitrogen concentratio n 

can be attributed to flushing of availabl e 

sources within the vegetation, soil, and atmosphere 

. The concentration declined as source s 

within the system were depleted and stream - 

flow rose by about one order of magnitude t o 

peak flow. The concentration rapidly declined 

on the recession curve to values slightly larger 

than the mean concentration in precipitation . 

Calcium 

Calcium outflow closely followed the run - 

off cycle (fig. 4) . The twofold to threefold 

increase in concentration from winter to the 

warm seasons had only minimal control o n 

calcium outflow compared with the hundredfold 

variation in annual runoff. The abrupt 

concentration drop in late November and th e 

rise in May corresponds approximately to th e 

time that the soil reaches field capacity in th e 

winter and falls below field capacity in the 

spring (fig . 1) . 

The calcium and bicarbonate carbon con - 

tent of the streamflow follow a similar pat - 

tern (fig. 8). Winter season values were nearl y 

constant whereas warm season stream con - 

tents rose sharply for streamflow values les s 

than 0.1 m 3 /km 2 -sec. The change in slop e 

(fig. 8) corresponds approximately to th e 

change in streamflow regimes (fig . 4) and th e 

time that soil water content rises to fiel d 

capacity in the winter and departs from it i n 

summer (fig . 1) . 

Silica 

The occurrence of minimum silica concentration 

in the dry and warm seasons of th e 

year and maxima in the winter and earl y 

spring (fig. 5) strongly contrasts calcium concentration-time 

trends (fig . 4). The silic a 

concentration increase in December an d 

January corresponds to the time of expecte d 

recharge of the soil to field capacity . Th e 

continued elevation of silica concentratio n 

through the winter season suggests tha t 

soluble silica is either released by weathering 

processes at that time or that it accumulated 

during the warm season of the year and i s 

removed when wetting fronts flush the soil 

mantle in midwinter . The reduction of concentration 

in late April corresponds to th e 

time that subsoils drain as the forest responds 

to evaporative demand and precipitation 

lessens . 

Discussion 

Organic nitrogen has seldom been identified 

in precipitation . Tarrant et al. (1968) 

found values of nitrogen input (1 .49 kg/ha-yr ) 

similar to those reported here (table 1) and in 

the same proportions of organic to nitrat e 

nitrogen (87 and 13 percent, respectively) . 

Organic nitrogen from 0 .02 to 0.20 mg/1 was 

also found in New Zealand snows by Wilson 

(1959). Only trace amounts of ammoniu m 

were found in the present study and in tha t 

by Tarrant et al . (1968). However, Moodie 

(1964) reported from 1 to 6 kg/ha-yr of 

ammonium nitrogen at several agricultural 

sites in western Washington. Values of input 

for cations, sodium, potassium, calcium, an d 

magnesium (table 1) were generally similar to 

those reported by Moodie . 

Atmospheric particulate matter may con - 

tribute substantially to nitrogen and phosphorus 

input . A large proportion of this i s 

undoubtedly of local terrestrial origin durin g 

the dry season and arises from road dust , 

pollen, and smoke particles . From November 

through March, when the local environment i s 

regularly dampened by precipitation, particulate 

matter is thought to be carried from th e 

Pacific Ocean in prevailing westerly winds . 

Presumably, the finest particles pass the filter s 

we use to separate solids from water and con - 

tribute to the "dissolved nutrients ." Nearly al l 

the phosphorus and 72 percent of the nitrogen 

resulted from a digestion of substances in 

rainwater collected during the dry seasons 

(table 1) . 

This forest system retains nitrogen ver y 

effectively . The net gain of 0 .5 kg/ha (table 2 ) 

is nearly equal to that reported by Cole et al . 

(1967). The fact that this result was duplicated 

on two Douglas-fir stands, very divers e 

127

8J 

Calciu m 

0 

I I I I 

0.25 0.50 0.75 1.00 

STREAM FLOW (m 3/ K km 2 -sec ) 

Figure 8 . Concentration of calcium and alkalinity as bicarbonate carbon in streamflow, 1970-72 . 

1 28

in terms of soil and stand structure, suggests 

that this retention property may be genera l 

for Douglas-fir forests west of the Cascade 

Range in the Northwestern United States . 

Although an explanation for this retentio n 

must await the completion of research now in 

progress, it appears from the small organic 

nitrogen outflow that most of the solubl e 

organic substances released from decomposition 

of detritus and organism excretions are 

rapidly incorporated by the organisms of th e 

forest. Nitrate nitrogen outflow is also regulated 

to levels much less than the input i n 

precipitation. Either nitrification rates are 

very low in the soils or the nitrate is retaine d 

and rapidly incorporated by the fores t 

system. Nitrification is active in the acid soil s 

of conifer forests on the Oregon coast (Bolle n 

and Lu 1968), but this information is lacking 

for soils of the H . J . Andrews Experimental 

Forest . 

The status of nitrogen capital of a Douglas - 

fir forest is the result of a number of processes 

including input in precipitation, fixation 

by free living organisms such as blue-green 

algae and soil bacteria, and fixation by micro - 

organisms living in symbiotic association wit h 

a host plant. If the forest gains 0 .5 kg/ha 

annually, as is indicated by this short perio d 

of record, 8,000 years would be required t o 

accumulate the 4,000 kg/ha of organic nitrogen 

found in the more fertile soils of the 

study watershed. Additions of nitroge n 

symbiotically fixed by Ceanothus velutinu s 

may be somewhat larger . If Ceanothu s 

velutinus stands persist for 20 years and fi x 

20 kg/ha annually as Zavitkovski and Newto n 

(1968) suggest, and if generations of thes e 

brush stands are induced by fire at intervals o f 

200 years, then 16,000 kg/ha of nitrogen 

would accumulate in 8,000 years. From this, 

deductions must be made for losses by soi l 

erosion and microbial and fire volatilization . 

In any case, inputs of nitrogen in precipitation 

must be considered of importance compared 

with other mechanisms of gain and los s 

by the forest . 

Phosphorus outflow from the forest wa s 

similar to that of nitrogen (table 2) . Loss 

occurs as both organic phosphorus and orthophosphorus 

. The organic form predominates 

in the warm seasons of the year when there i s 

a rapid biological turnover of phosphorus b y 

the forest. Concentration of the ortho form re - 

mains fairly constant throughout the year and 

is thought to arise from mineral weathering . 

The losses of the cations sodium, potassium, 

calcium, and magnesium were large 

compared with those of nitrogen and phosphorus 

and represent excess amounts of thes e 

cations to the annual requirements of th e 

forest vegetation (table 2) . These losses are 

approximately four times those reported fo r 

the Hubbard Brook ecosystem in New Hampshire 

except for the potassium values whic h 

were nearly equal to those reported her e 

(Likens et al. 1970). Cole et al. (1967) also 

reported lower losses of potassium and calcium. 

The cation and silica losses (table 2) are 

estimates of the annual release of these chemicals 

by mineral weathering . In this regard, it 

is interesting to note that the ratio of 

Si :Ca:Na:Mg weathering rates for 1970 ar e 

8.9 :4.3 :2.1 :1 . 

Another mechanism for nutrient loss is th e 

particulate matter that is transported by the 

stream. This material may arise from soil 

erosion or from organic detritus from the 

forest. Losses of suspended sediment fro m 

this study (67 and 37 kg/ha-yr) were much 

lower than the 12-year annual average of 13 3 

kg/ha from another undisturbed watershed in 

the Experimental Forest (Fredriksen 1970) . 

Although we have not quantified the chemical 

loss by this means in this study, previousl y 

published annual losses of 0.16 kg/ha for 

organic nitrogen represent less than half of 

the dissolved component loss (table 2), still a n 

important part of the total nitrogen loss 

(Fredriksen 1971) . Cation loss was much les s 

than 1 percent of the loss in dissolved form . 

However, since erosion losses are very sporadic 

and dependent on extremes of storm 

runoff events, and the results (Fredrikse n 

1971) are indicative of the chemical loss during 

a relatively quiescent period, they shoul d 

be taken as minimum estimates of nutrien t 

loss by soil erosion processes . 

Biological and chemical processes of th e 

forest that control nutrient mobilization an d 

retention are undoubtedly closely linked to 

environmental factors. The movement of 

129

water is dominant because flowing water i s 

required to transport materials to the sites of 

physical-chemical and biochemical activity 

and to remove byproducts that appear in the 

outflow of experimental watersheds . 

Definite concentration changes in th e 

stream occur at the transition periods fro m 

drying to wetting of the forest system (figs . 3 , 

4, and 5) . A key feature is either retention o f 

precipitation (and reaction byproducts) in soil 

storage or the passage of wetting front s 

through the soil mantle when the soils are a t 

field capacity . It is not surprising that, of th e 

annual totals, 77 percent of the calcium, 8 0 

percent of the organic nitrogen, and 83 per - 

cent of the silica were exported from th e 

forest in the 5 months from December 196 9 

through April 1970 . During this period, th e 

soil mantle was at field capacity and wettin g 

fronts from winter rainstorms regularl y 

flushed the entire forest system . 

For organic nitrogen (fig . 3), the period of 

maximum outflow corresponds to the tim e 

when wetting fronts begin to pass through the 

soil mantle in early December . The widening 

margin between input and outflow (fig. 6 ) 

indicates an increasing contribution to outflow 

of organic nitrogen from the forest 

system. Although we do not know how muc h 

organic nitrogen from precipitation directl y 

supplements streamflow, the increased concentration 

and content in precipitation an d 

streamflow in late January suggest that 

sources of organic nitrogen in precipitatio n 

may directly augment the outflow from this 

experimental catchment . 

There is a definite change in the calcium 

content of streamflow as the hydrologic state 

of the forest system changes from winter to 

summer (fig. 8). The calcium outflow was 

closely paralleled by the content of carbonate 

and bicarbonate-the predominant anions of 

the stream . McColl and Cole (1968) demonstrated 

that cation mobilization is effectively 

controlled by these mobile anions that are 

formed by the carbonation of water with 

carbon dioxide released by respiring organisms 

of the forest . As yet, we cannot explain 

the processes that maintain a nearly constant 

concentration of calcium and these mobil e 

anions over the range of winter streamflow . 

Silica is nearly absent in precipitation an d 

is therefore entirely generated within the soil - 

forest system (fig. 5). Although the paren t 

materials of soil are the primary source o f 

silica, silica is cycled in generous amounts b y 

forests as indicated by litterfall studie s 

(Remezov et al. 1955) and the occurrence o f 

plant opal crystals in forest soils (Paeth 

1970). Silica in streamflow may originate 

from primary mineral weathering, secondary 

mineral dissolution, or release from decomposition 

of detritus . Silica may be taken out o f 

solution by formation of secondary mineral s 

within the soil . The occurrence of minimu m 

silica concentration in the dry and war m 

seasons of the year and maxima in the winte r 

and early spring strongly contrasts calciu m 

concentration-time trends (fig . 4) . 



ported by the U .S. Forest Service in cooperation 

with the Coniferous Forest Biome, U.S . 

Analysis of Ecosystems, International Biological 

Program. This is Contribution No . 29 to th e 

Coniferous Forest Biome . 


Bollen, W. B., and K . C. Lu. 1968. Nitroge n 

transformations in soils beneath red alde r 

and conifers . In J. M. Trappe, J . F. Franklin, 

R . F . Tarrant, and G . M. Hansen (eds .) , 

Biology of alder, p. 141-148 . Pac. North - 

west Forest & Range Exp . Stn., Portland , 

Oreg . 

Bormann, F. H., and G. E. Likens . 1967 . 

Nutrient cycling . Science 155 : 424-429 . 

Cole, D . W., S. P. Gessel, and S . F . Dice . 

1967 . Distribution and cycling of nitrogen , 

phosphorus, potassium and calcium in a 

second-growth Douglas-fir ecosystem . In 

Primary production and mineral cycling in 

natural ecosystems, p. 197-232. Univ. 

Maine Press . 

Fredriksen, R. L. 1969. A battery powered 

proportional streamwater sampler . Wate r 

Resour. Res . 5: 1410-1413 . 

130

1970. Erosion and sedimentation 

following road construction and 

timber harvest on unstable soils in three 

small western Oregon watersheds. USDA 

Forest Serv. Res. Pap. PNW-104, 15 p . , 

illus. Pac. Northwest Forest & Range Exp . 

Stn., Portland, Oreg. 

1971 . Comparative water quality-natural 

and disturbed streams followin g 

logging and slash burning . In Forest lan d 

uses and stream environment, p . 125-137 . 

Oreg. State Univ., Corvallis . 

Likens, G . E., F . H. Bormann, N . M. Johnson , 

and others . 1970. Effects of forest cuttin g 

and herbicide treatment on nutrien t 

budgets in the Hubbard Brook Watershedecosystem 

. Ecol. Monogr . 40: 23-47 . 

McColl, J. G., and D . W . Cole. 1968 . A mechanism 

of cation transport in a forest soil . 

Northwest Sci . 42 : 134-140 . 

Moodie, C . D. 1964 . Nutrient inputs in rain - 

fall at nine selected sites in Washington . 

Wash . Agric. Exp. Stn., Wash. State Univ. 

Paeth, D. C. 1970. Genetic and stability relationships 

of four western Cascade soils . 12 6 

p. Ph .D . thesis, on file at Oreg . State Univ. , 

Corvallis. 

Remezov, N. P., L. N. Bykova, and K. M. 

Smirnova. 1955. Nitrogen and mineral 

cycles in forests . Akidemiya Nank SSSR. 

Trudy Instituta Lesa ., p. 167-194 . 

Tarrant, R . F ., K. C. Lu, C . S. Chen, and W. 

B. Bollen. 1968. Nitrogen content of precipitation 

in a coastal Oregon forest opening. 

Tellus XX : 554-556 . 

Wilson, A . T. 1959. Organic nitrogen in New 

Zealand snows . Nature 183 : 318-319 . 

Zavitkovski, J ., and M . Newton . 1968. Ecological 

importance of snowbrush , 

Ceanothus velutinus, in the Oregon Cascades 

. Ecology 49 : 1134-1145 . 

131



Nutrient cycling in throughf all 

and litterfall in 450-year-old 

Douglasfir stands 

Albert Abee 

an d 

Denis Lavender 

School of Forestry 



Abstract 

Comparisons of nutrient concentrations (N, P, K+, Ca ++ , Mg++) found in canopy throughfall and litterfall 

were made on the H. J. Andrews Experimental Forest. Six old-growth Douglas-fir (Pseudotsuga menziesii ) 

stands were studied which represented six forest communities common to the western Cascades ofOregon. These 

community types span a large portion of the temperature and moisture gradients present in the area. The 

preliminary data indicate that nutrient concentration in throughfall was highest during the summer and fall, and 

lowest during the winter. Nutrient input through throughfall generally followed the same trends . Nutrient 

return through litterfall was greatest in the needles. More amounts of N, P, and Ca++ were transferred to the soil 

through litterfall than through throughfall, while more K + and Mg++ were added to the soil through throughfall. 

Litterfall was maximum during the winter. Future studies will correlate the results from the nutrient analysis to 

the moisture and temperature gradients . 

Introduction 

The worldwide interest of scientists in litterfall 

production during the past century, has 

been shown by Bray and Gorham (1964) i n 

their review of litter production in the forest s 

of the world. Methodology reports ranged 

from utilization of randomly located collection 

devices of varied design, separation, oven - 

drying, and chemical analysis of several litter 

components, to merely raking up and air drying 

the litter on a unit area basis. In spite of 

the large number of papers cited in the abov e 

review, data of litter production from natural , 

old-growth ecosystems are meager. Even less 

is known about litterfall in old-growth 

Douglas-fir (Pseudotsuga menziesii) forest 

types. The examination of seasonal fluctuations, 

nutrient concentration changes associated 

with defoliation, and nutrient composition 

of various litterfall categories are scarce 

(Kira and Shidei 1967) . 

The first published report of an investigation 

of litterfall in coniferous forests of the 

Pacific Northwest is that of Tarrant, Isaac , 

and Chandler (1951). These workers collected 

the litter of several species for 1 year and estimated 

nutrient movement by multiplyin g 

litter weight by the percent elemental conten t 

of foliage collected from trees, an inexact procedure 

. More detailed measurements of th e 

nutrient cycle in Douglas-fir forests have bee n 

published for stands in New Zealand (Wil l 

1959) and the United States (Dimock 1958) . 

In addition, workers at both the University o f 

Washington (Rahman 1964) and Oregon Stat e 

University 1 have collected substantial dat a 

describing litterfall in both managed an d 

natural Douglas-fir stands . Riekerk and Gessel 

(1965) and Cole and Gessel (1968) summariz e 

1 D . P . Lavender, unpublished data . 

133

a number of very sophisticated studies o f 

nutrient movement through Douglas-fir ecosystems 

in Washington . 

All of the above studies, save that of Tar - 

rant et al ., however, were concerned with 

litterfall and nutrient movement through relatively 

young stands . 

Several studies (LeClerc and Breazeal e 

1908, Mes 1954, Tukey and Amling 1958 , 

and Tukey et al . 1958) have demonstrated 

that rainfall may remove substantial quantities 

of nutrient elements from the foliage of 

horticultural plants . Similarly, studies of th e 

elemental content of precipitation under 

forest stands (Tamm 1951, Madgwick an d 

Ovington 1959, Will 1959, and Voigt 1960 ) 

have demonstrated that rainwater which ha s 

passed through tree crowns ( "throughfall") 

contains significantly higher quantities o f 

many nutrient elements than rainfall collected 

in adjacent openings . 

In the Pacific Northwest, studies reporte d 

by Rahman (1964), Tarrant et al . (1968) and 

Cole and Gessel (1968) have yielded data 

which describe the movement of nutrient s 

from the atmosphere and tree crowns to th e 

forest floor by precipitation . Finally, unpublished 

data by Lavender describe the 

movement of nutrients from the crowns of 

both fertilized and control second-growth 

Douglas-fir stands to the forest floor by 

precipitation . 

The purpose of the present study was to 

measure the movement of nutrients in canop y 

throughfall and litterfall in several associatio n 

types of old-growth Douglas-fir stands . These 

community types were selected to represent 

the range of environments occurring on the H . 

J. Andrews Experimental Forest and are also 

indigenous to the Pacific Northwest. This 

paper will report on the results of our efforts 

to date . 

Study Area 

The H . J. Andrews Experimental Forest en - 

compasses 15,000 acres and is characterize d 

by steep topography with approximately one - 

fifth of its land area in gentle slopes o r 

benches. Elevations within the forest vary 

from 457 m to more than 1,523 m . Precipitation 

is heavy, varying from 226 cm per year a t 

lower elevations to as much as 356 cm per year 

along the highest ridges. A considerable snowpack 

develops on the higher slopes while rai n 

predominates at the lower elevations. Mean 

temperatures within the forest range fro m 

35°F in January to 65°F in midsummer 

(Berntsen and Rothacher 1959) . 

Methods 

Within the Experimental Forest, six communities 

were chosen (table 1), each name d 

for characteristic plants in both the overstory 

and understory . The six old-growth communities 

are presented in order of increasing 

elevation. Each of the six plots are 0 .2024 

hectare in size and are equipped with eight 

litter traps ; each is 2,601 cm 2 in area, located 

on a random basis in each plot . Litter was 

collected every 4 to 6 weeks during the snow - 

free months of 1970-71 . Heavy snow pack 

prevented litter collection during much of th e 

winter of 1970 . Therefore, data describing 

nutrient movement in the litter for this perio d 

are weak because : (1) the necessarily infrequent 

collections do not permit accurat e 

assessment of the rate of litterfall, and (2) lit - 

ter which remained in the traps for lon g 

periods was subjected to leaching . The following 

fall, three litter traps on each plot wer e 

equipped with a 113-liter reservoir to collec t 

the precipitation which passed over the litter . 

Analysis of this water will provide a measure 

of the nutrients leached from the litter . 

Crown and stem maps were made for each 

plot to aid in evaluating the variation of litter - 

fall between traps . 

After collection the litter for each trap wa s 

dried at 70°C, separated into classes (needles , 

cones, twigs, branches, hardwoods, bark, 

lichens and mosses), and weighed. Prior to 

chemical analyses, litter from the eight litter 

traps per plot was composited into tw o 

samples, each representing four traps . In addition, 

some consolidation of litter collected on 

different dates was necessary . Each sample 

was analyzed to determine the levels of nitrogen, 

phosphorus, potassium, calcium an d 

134

Table 1.-Characteristics of study plot s 

Plot 

Elevation 

PPT/ 

year 

Percent species 

composition ) 

Diameter 

range 

Average 

d .b.h . 

Basa l 

area 2 

Stems / 

hectare 

meters centimeters - - - centimeters- - - 

1 . Pseudotsuga- 457 211 .8 Psme 89 .3 11 - 163 46 .7 83 .06 27 7 

Holodiscus Tsme 5 . 4 

Tabr 3 . 6 

Acci 1 . 7 

2 . Tsuga- 488 228 .3 Psme 17 .0 8 - 139 37 .8 97 .55 494 

Rhododendron- Tsme 69 . 0 

Berberis Tabr 8 . 0 

Thpl 5 . 0 

Conu 1 . 0 

3 . Tsuga- 762 230 .6 Psme 26 .5 10 - 213 68 .1 120 .87 168 

Polystichum Tsme 61 . 8 

Tabr 8 . 8 

Thpl 2 . 9 

4 . Tsuga - 610 229 .6 Psme 52 .8 8 - 157 34 .0 69 .77 45 0 

Rhododendron - Tsme 23 . 1 

Gaultheria Tabr 1 . 1 

Thpl 19 . 8 

Cach 2 . 2 

Conu 1 . 1 

5 . Tsuga-Abies- 975 Psme 16 .1 8 - 173 66 .3 129 .60 27 7 

Linneae Tsme 46 . 4 

Thpl 33 . 9 

Tabr 3 . 6 

6 . Abies- 1,311 Psme 46 .3 8 - 117 55 .4 109 .63 33 1 

Tiarella Tsme 16 . 4 

Abam 

Abpr 

37 .3 

1 Psme Pseudo tsuga menziesi i 

Tsme Tsuga mertensiana 

Tabr Taxus brevifolia 

Acci Acer circina turn 

Thpl Thuja plicat a 

Conu Cornus nutalli i 

Cach Castanopsis chrysophylla 

Abam Abies amabilis 

Abpr Abies procera 

2 Square meters/hectare . 

135

magnesium. 

In addition to the litter traps, each plot was 

equipped with- four 20-inch-high rain gages . 

Each gage was assigned to one of 20 rando m 

locations within the plot after each collection , 

in accordance with a method described b y 

Wilm (1943). Higher elevation plots als o 

contain a rain gage on a platform 10 feet fro m 

ground level to provide a water sample durin g 

months of heavy snow cover . Water was collected 

and the volume measured at approximately 

2-week intervals . The samples wer e 

returned to the laboratory on the day of 

collection, filtered, and stored at -12°C unti l 

thawed for analysis. For analysis, the fou r 

samples per plot were combined into tw o 

samples and analyzed for total potassium , 

calcium, magnesium, orthophosphate, an d 

total phosphorus. Nitrogen in the form of 

ammonium, nitrate, nitrite, and organic nitrogen 

was also determined . 

Precautions were taken to keep contamination 

of water samples to a minimum . Funnel s 

with glass wool stoppers were provided for 

each rain gage to keep organic matter fro m 

contaminating water samples . Mercuric chloride 

was added to the rain gages in the summer 

and fall to keep microorganism activity 

to a minimum . The cold temperature helpe d 

to reduce microorganism and insect activit y 

during the winter . 

Several techniques were investigated in an 

effort to arrive at a measure of crown density . 

Basal area and volume poorly describe intercepting 

crown cover in old-growth defectiv e 

stands as canopy development tends to re - 

main constant after trees reach maturity ; 

hence, direct estimates were used . Photo - 

graphs which were taken above each samplin g 

point with a 35-mm camera were shown over 

a spherical dot grid to give a means of comparing 

crown densities . 

(fig. 1). Water sample concentrations were 

highest during the summer when precipitatio n 

was minimal (table 2) . Concentrations were 

lowest during the winter when precipitatio n 

was highest. As precipitation decreased fro m 

winter to spring, concentration of throughfal l 

samples for each element increased . Throughfall 

concentrations were also high during th e 

fall when precipitation first starts . 

Unlike N, the average total input of P , 

Mg+, Ca++, and K+ generally follows the 

same trend as did the concentration curves 

(fig. 2). The greatest amount of each element 

was leached out during the fall when precipitation 

first washes the canopy . Decreasin g 

amounts were leached out with increasin g 

precipitation . Potassium input reached a lo w 

point during the winter, at which time 68 per - 

cent of the total precipitation had fallen, an d 

z 

0 

I- 

2 

CC 

I- 

z 

w 

U 

z 

0 

U 

0 

100 

80 

60 

40 

20 


Throughfall Results 

Nutrient concentration in throughfall 

samples for plots 1 through 4 for all elements 

appeared to be the same during each season 

FALL 

Figure 1 . Average concentration of plots 1 through 4 

for each element by season . 

0 

136

Z 

w 

2 

w 

J 

w 

3 

0 

Table 2. Average total precipitation increased sharply from winter to spring , 

across all plots by season slightly decreasing from spring to summer . 

Season 

Precipitation 

Calcium and P reached low points during th e 

spring, at which time 92 percent of the tota l 

pre pitation had allen, and increased fro m 

spring to summer. Magnesium input was great - 

centimeters 

est from fall to winter and remained approximately 

the same from winter to summer. 

Fall 

70.89 

Nitrogen input slightly increased from fall to 

winter, decreasing from winter to sprin g 

Winter 

92 .0 5 

reaching a low point during the summer . 

There appears to be no difference in term s 

Spring 

57.9 6 

of net kg per hectare per year between plot s 

1 to 4 for each element with the exception o f 

Summer 

18 .0 8 

plot 3 (table 3) . Plot 3 had more K + and les s 

Ca++ than plots 1, 2, and 4 . 

Total 238 .98 

100 

8 0 

20 

Figure 2. Total average kg per hectare of plots 1 

through 4 for each element by season . 

Throughfall Discussion 

In general, the total nutrient input an d 

throughfall concentrations were highest in th e 

summer and fall and lowest during the winte r 

and spring months. This seems to indicate 

that each tree or canopy has a constant fraction 

of elements which can be removed fro m 

the foliage through leaching elements . Once 

the rains start in the fall, the majority of eac h 

nutrient is leached out . As the rains increase 

in quantity and duration during the winter 

• and spring months, the available fraction of 

nutrients is further depleted. Decreasing 

precipitation from spring to the end of summer 

allows the nutrient fraction to increase 

• 

again until the total fraction of leachabl e 

nutrients is reached . 

Variations found between plot 3 and plot s 

1, 2, and 4 with respect to K + and Ca++ coul d 

be due to differences in soil types (data no t 

available yet) . If soil types are different with 

respect to nutrient availability, the difference s 

between plots could be explained by luxur y 

consumption . 

Another possible source of the nitroge n 

found in the throughfall samples is nitrogen - 

fixing bacteria . Jones (1970) in his study o f 

nitrogen fixation by bacteria in the phyllo - 

sphere of Douglas-fir (Pseudotsuga douglasii) 

in England isolated bacteria from the leaf 

surfaces of Douglas-fir. He found that th e 

bacteria could fix atmospheric nitrogen whe n 

provided with a carbohydrate source . The fat e 

13 7

Table 3 .-Net kg per hectare per year of nutrients collected in throughfall gages 

++ Mg++ 

Item N P K + Ca 

Input from atmosphere ) 1 .298 0 .232 0 .106 2 .085 1 .27 3 

Throughfall input : 

Plot 1 

3 .999 2 .308 17 .416 5 .983 2.608 

2 2 .979 2 .398 15 .749 4 .438 2 .08 6 

3 3 .729 2 .970 30 .350 2 .134 1 .45 6 

4 2 .710 3.283 23 .379 5 .104 2 .34 3 

Average 3 .354 2 .740 21 .724 4 .416 2 .123 

1 Fredriksen unpublished data-data collected from open area on the H . J . Andrews Experimental Fores t 

at 610 meters . 

Table 4 .-Distribution of metric tons/hectare between litter components by plo t 

Plot 1 2 3 4 5a 5b 1 6 Average 2 

Needles 2 .002 2 .246 2 .950 3 .200 2 .533 2 .533 3 .741 2 .77 7 

Percent of total 32 .84 35 .56 46 .44 62 .30 15 .17 55 .91 54 .10 47 .1 5 

Reproductiv e 

structures .834 1 .141 1 .284 .742 .536 .536 .518 .74 3 

Percent of total 13 .68 18 .06 20.22 14 .46 3 .20 11 .83 7 .49 14 .3 1 

Wood material 2 .280 2 .760 1 .876 1 .009 13 .479 1 .267 2 .524 1 .95 3 

Percent of total 37 .40 43 .68 29 .53 19 .66 80 .74 27 .96 36 .50 33.1 4 

Hardwoods and 

mosses 1 .022 .175 .401 .197 .195 .195 .206 .36 5 

Percent of total 16 .77 2 .77 6 .32 3 .84 1 .17 4 .30 2 .98 6 .2 0 

Total 6 .138 6 .317 6 .512 5 .131 16 .694 4.530 6 .916 5 .89 1 

Note : In plot 5, an extremely large slab of bark from a nearby snag fell into a trap causing high values for tota l 

tons/hectare . Over a longer period of time, this type of variation between litter components can be expected to 

occur randomly throughout each plot . However, due to the limited sampling time thus far recorded, the one 

extreme value will be temporarily ignored . 

1 Excluding extreme bark sample . 

2 Excluding 5a . 

138

of the nitrogen was not determined. However , 

Jones suggested that it could be washed to th e 

ground . 

Litterfall Dat a 

Despite the differences in stand characteristics 

shown in table 1, little variation in total 

litter production was found among stands fo r 

the year 1970-71 (table 4) . Average yearl y 

litterfall production for all plots was 5 .8 9 

metric tons/hectare . This is approximately 1 % 

times the average 3 .5 metric tons per hectare 

reported by Bray and Gorham (1964) fo r 

cool, temperate forests, but closer to the yiel d 

they reported for a latitude comparable t o 

their study area (fig . 1). From worldwid e 

data, these authors reported that nonlea f 

litter averaged from 27 percent to 31 percent 

of total litter production . The stands reported 

here averaged 47 percent nonleaf (woody ) 

litter for the 1-year period . 

In terms of total kg/hectare of litter, the 

vast majority fell during the winter (fig . 3) . 

This is the period when snowfall is greatest , 

consequently much litter breaks under th e 

1600 

1400 

1200 

weight of the snow . Needle cast was greatest 

in the fall, decreasing during the winter, an d 

gradually increasing during spring . Hardwoo d 

and moss litter was greatest in the fall, de - 

creasing throughout the rest of the year . 

Woody material and cone litterfall was greatest 

during the winter. 

Nutrient concentration of litterfall components 

varied considerably among plots an d 

seasons (table 5). Plots 1 and 3 were chosen 

to represent the range of values that can be 

found in nutrient return through litterfall . 

There appears to be no consistent trend b y 

season or plot for the litter component concentrations. 

However, average yearly concentrations 

of each nutrient for each litter clas s 

are comparable between plots 1 and 3 . There 

is a substantial difference between tota l 

kg/hectare for N, P, K +, and Ca++ betwee n 

plots (table 6). Plot 1 had more kg of Ca ++ 

per hectare than did plot 3 . Plot 3 had greate r 

amounts of N, P, and K + than did plot 1 . 

Little difference occurred among plots fo r 

Mg++ 

The greatest portion of nutrient inpu t 

through litterfall came in the needles (tabl e 

7). Needle litterfall contributed about 54 per - 

cent of the total nutrient input . Cone litterfall 

accounted for 13 percent while twig litterfall 

accounted for 11 percent of the total . Together, 

the needle, cone, and twig litterfall 

account for 78 percent of the total nutrien t 

input through litterfall . 

Litterfall Discussio n 

40 0 

200 

Figure 3 . Average kg per hectare by season and litter 

component . 

Variations in nutrient concentration foun d 

between litterfall components among plot s 

and season can be expected if foliage characteristics 

such as age and species are no t 

constant . We also observed that the age of th e 

tissue, and when it falls, varies throughout th e 

year for each plot. This is primarily due to 

environmental parameters such as win d 

action, rainstorms, and snowfall . Difference s 

in soil types could also have affected concentrations 

. 

Differences between total nutrient inpu t 

through litterfall (table 6) are affected by th e 

distribution of litter components within th e 

total. Where two plots seem to produce corn - 

139

Table 5.-Average percent of N, P, K +, Ca ++, and Mg++ by season, plot, and litter componen t 

Litte r 

component 

and season 

Plot 1 Plot 3 

N P K + Ca ++ Mg++ N P K+ Ca ++ Mg+ + 

Needles : 

Fall 0 .368 0 .087 0 .109 1 .901 0 .017 0 .434 0 .120 0 .161 1 .815 0 .02 0 

Winter .691 .108 .210 1 .305 .018 .763 .124 .280 1 .103 .08 1 

Spring .462 .124 .159 1 .241 .028 .717 .119 .103 1 .112 .01 7 

Summer .467 .143 .200 1 .740 .034 .474 .105 .145 1 .286 .02 5 

Average .497 .115 .169 1 .546 .024 .597 .117 .172 1 .329 .02 0 

Cones : 

Fall .530 .084 .122 .397 .045 .518 .063. .099 .149 .01 5 

Winter .294 .029 .046 .146 .016 .387 .033 .051 .148 .01 1 

Spring .488 .049 .133 .214 .041 .544 .064 .102 .178 .01 1 

Summer .487 .068 .151 .369 .027 .561 .063 .159 .206 .01 6 

Average .449 .057 .105 .281 .032 .502 .055 .103 .170 .01 3 

Twigs : 

Fall .340 .033 .051 .852 .009 .434 .051 .124 1 .107 .01 6 

Winter .375 .040 .057 1 .261 .011 .398 .032 .075 1 .110 .01 2 

Spring .424 .072 .076 .999 .081 .358 .047 .046 .823 .00 8 

Summer _ .408 .043 .105 1 .034 .013 .363 .055 .102 1 .054 .01 5 

Average .386 .047 .072 1,036 .013 .388 .046 .086 1 .023 .01 3 

Branches : 

Fall .218 .017 .030 .713 .008 .207 .013 .025 .522 .00 6 

Winter .028 .017 .080 .598 .006 .296 .039 .082 .843 .01 0 

Sprin g 

Summer .102 .008 .030 .338 .00 5 

Average .213 .017 .055 .655 .007 .201 .020 .046 .567 .00 7 

Bark : 

Fall .332 .047 .076 .566 .014 .417 .038 .074 .517 .01 0 

Winter .404 .033 .050 .963 .010 .431 .035 .057 .413 .009 

Spring - _ .476 .080 .055 .300 .00 9 

Summer .320 .030 .051 .637 .011 .541 .056 .163 .474 .01 2 

Average .352 .036 .059 .722 .011 .466 .052 .087 .426 .01 0 

Hardwoods : 

Fall .630 .095 .167 2 .395 .037 .591 .124 .475 2 .428 .06 1 

Winte r 

Sprin g 

Summer .546 .092 .302 2 .127 .07 0 

Average .588 .094 .234 2 .261 .053 .591 .124 .475 2 .428 .06 1 

Mosses and lichens : 

Fall .417 .053 .117 .408 .016 .659 .126 .229 .413 .01 9 

Winter 1 .264 .100 .132 .567 .015 1 .234 .122 .350 .329 .01 7 

Spring - - -- _ - - - - 

Summer 1 .313 .100 .246 .376 .020 1 .427 .130 .284 .364 .01 8 

Average .998 .084 .165 .450 .017 1 .106 .126 .287 .368 .018 

140

Table 6. Total kg per hectare per year for plots 1 and 3 for each element by litter class 

Litter class 

Plot 1 Plot 3 

N P K+ Ca ++ Mg++ N P K + Ca ++ Mg+ + 

Needles 7 .93 2 .27 3 .16 35 .89 0 .47 15 .61 3 .82 5 .89 44 .08 0 .6 5 

Percent of total 36.12 58 .85 49 .54 50 .22 44 .61 47 .76 68 .20 60 .25 69 .90 59 .7 9 

Cones 3 .46 .45 .83 2 .06 2 .13 6 .60 .69 1 .33 2 .05 .1 7 

Percent of total 15 .74 11 .59 13 .03 2 .89 20 .31 20 .23 12 .40 13 .63 3 .24 15 .4 6 

Branches 1 .97 .16 .74 5 .09 .06 .82 .12 .29 2 .15 .0 2 


Twigs 3 .52 .38 .66 12 .01 .10 3 .54 .34 .75 9 .56 .1 0 

Percent of total 16.07 9.86 10 .33 16 .79 9 .50 10 .82 6 .20 7 .67 15 .15 9 .2 7 

Bark 1 .62 .11 .07 3 .56 .30 3 .00 .28 .49 3 .18 .0 6 

Percent of total 7 .39 3 .17 3.08 4 .97 2 .92 9 .18 5 .00 5 .04 5 .03 5 .1 5 

Hardwoods 2 .33 .39 .67 12 .48 .17 .38 .06 .11 1 .51 .0 5 


Moss and lichens 1 .12 .08 .12 .39 .01 2 .73 .28 .90 .54 .0 3 

Percent of total 5 .10 2 .11 1 .98 .54 1 .32 8 .35 5 .00 9 .16 .85 3 .0 9 

Total 21 .95 3 .86 6 .39 71 .49 1 .07 32 .68 5 .59 9 .77 63 .06 1 .09 

Table 7.-Average percent of litterfall totals for plots 1 and 3 for each element by litter class 

Litter class N P K + Ca ++ Mg++ Average 

Needles 41 .9 63.5 54 .9 60 .1 52 .2 54 .5 2 

Cones 18 .0 12 .0 13 .3 3 .1 17 .9 12.8 6 

Branches 5 .7 3 .0 7 .3 5 .3 3 .7 5 .0 0 

Twigs 13 .4 8 .0 9 .0 16 .0 9 .4 11 .1 6 

Bark 8 .3 4 .1 4 .1 5 .0 4 .0 5 .1 0 

Hardwoods 5 .9 5 .1 5 .8 9 .9 10 .1 7 .3 6 

Moss and lichens 6 .7 3 .6 5 .6 .7 2 .2 3 .7 6 

r 

14 1

parable total quantities of litter on a yearly 

basis, amounts of the various litter components 

are important . The amounts of each 

litter component are important becaus e 

concentrations of nutrient elements vary fo r 

each litter component (see table 5) . Table 4 

shows that the greatest differences betwee n 

plots 1 and 3 in terms of kg/hectare of litter - 

fall occur between needles and woody material. 

Plot 3 produced 947 kg/hectare more of 

needles than plot 1, while plot 1 produced 

403 kg/hectare more of woody material tha n 

plot 3. However, table 5 shows that th e 

average concentration of nitrogen, fo r 

example, is much higher in needles than it is 

in woody material ; consequently, variations 

such as found in table 6 are brought about . 

Nutrient input in both throughfall and litterfall 

appears to have the same trend for K + , 

P, and Ca++ when comparing plots 1 and 3 . 

The litterfall analysis shows plot 3 havin g 

more K +, P, and less Ca ++ than plot 1 . Th e 

throughfall data for plots 1 and 3 show th e 

same results (table 3) . However, differences 

between N and Mg++ input through throughfall 

and litterfall for each plot do not agree . 

Table 3 shows little difference in N input fo r 

each plot, while table 6 shows N being highe r 

in plot 3 . A possible explanation for this ha s 

been given in the preceding paragraph . Tabl e 

3 shows that plot 1 had more Mg ++ input 

than plot 3 . Litterfall input for Mg ++ was 

about the same between plot 1 and 3 . A 

possible explanation for the difference between 

plots 1 and 3 for Mg++ in throughfall 

might be that the hardwood litterfall ac - 

counted for 10.1 percent of the total Mg ++ 

input (table 7) . Table 1 shows that 1 .7 per - 

cent of the species composition in plot 1 wa s 

hardwoods, while plot 3 shows no hardwoods . 

More amounts of N, P, and Ca ++ were 

transferred to the soil through litterfall tha n 

through throughfall ; while more K + and Mg ++ 

were added to the soil through throughfal l 

(table 8) . 

There appeared to be no relationship between 

nutrient concentration and elevation . 

Rather, concentration on each plot seemed t o 

be correlated with crown density . Basal are a 

and crown density in old-growth Douglas-fi r 

stands seem to be distributed randomly ove r 

the sites on the H . J. Andrews Experimental 

Forest. Consequently, no real correlatio n 

could be seen between basal area and crow n 

density with moisture and elevation . Thi s 

could also be due to the fact that the range o f 

environments sampled on the H . J. Andrews 

Experimental Forest was not great enough , 

indicated by total precipitation . Table 1 

shows that there was no real difference i n 

total precipitation between plots 1 through 4 . 

Summary 

The preliminary data indicate that nutrien t 

concentration in throughfall varied with sea - 

son . Highest concentrations were found in th e 

summer and lowest concentrations during th e 

winter. Nutrient input through throughfal l 

generally followed the same trends as di d 

nutrient concentrations . Nutrient return 

through litterfall was greatest in the needles . 

Together, the needles, twigs, and cones ac - 

counted for 78 percent of the nutrient inpu t 

Table 8.-Total nutrient input in kg per hectare per yea r 

Input N P K + Ca ++ Mg++ 

Average throughfall input 3 .3544 2 .740 21 .7230 4 .416 2.12 3 

Average litterfall input 27 .3235 4 .725 8 .0784 67 .2738 1 .08 0 

Total 30 .6779 7 .465 29 .8014 71 .6888 3 .203 

142

through litterfall . More amounts of N, P, and 

Ca++ were transferred to the soil throug h 

litterfall than through throughfall, while mor e 

K+ and Mg++ were added to the soil through 

throughfall. Litterfall was maximum during 

the winter . 



in part by National Science Foundation 




No. 30 to the Coniferous Fores t 

Biome . 


Berntsen, C. M., and J . Rothacher . 1959 . A 

guide to the H . J. Andrews Experimenta l 

Forest. USDA Forest Serv . Pac. Northwes t 

Forest & Range Exp . Stn ., 21 p . Portland , 

Oreg . 

Bray, J. Roger, and Eville Gorham . 1964 . 

Litter production in forests of the world . 

Advan. Ecol. Res . 2: 101-157 . 

Cole, Dale W ., and Stanley P. Gessel. 1968 . 

Cedar River Research . A program for 

studying the pathways, rates, and processes 

of elemental cycling in a forest ecosystem . 

Univ . Wash., Coll . Forest Res ., Inst. Forest 

Prod. Contrib. No . 4, 54 p . 

Dimock, E . J. 1958 . Litter fall in a young 

stand of Douglas-fir. Northwest Sci . 32 : 

19-29 . 

Jones, K. 1970. Nitrogen fixation in th e 

phyllosphere of the Douglas-fir, (Pseudotsuga 

douglasii) . Ann. Bot. (London) 34 : 

239-244 . 

Kira, T ., and T. Shidei . 1967 . Primary production 

and turnover of organic matter in different 

forest ecosystems of the wester n 

Pacific. Jap . J. Ecol. 17(2) : 70-87 . 

Le Clerc, J. A., and J. F. Breazeale . 1908 . 

Plant food removed from growing plants by 

rain or dew . In J . A. Arnold (ed.), U .S . 

Department of Agriculture yearbook, p . 

389-402 . 

Madgwick, H. A. I., and J. D. Ovington . 1959 . 

The chemical composition of precipitatio n 

in adjacent forest and open plots . Forestry 

32(1) : 14-22 . 

Mes, Margaretha G . 1954. Excretion (recretion) 

of phosphorus and other mineral elements 

by leaves under the influence of rain . 

S. Afr . J . Sci. 50(7) : 167-172 . 

Rahman, Abu Hamed Mohammed Mojibur . 

1964 . A study of the movement of elements 

from leaf crowns by natural litter - 

fall, stemflow and leaf wash . 119 p . M.F . 

thesis on file, Univ . Wash ., Seattle . 

Riekerk, Hans, and Stanley P . Gessel. 1965 . 

Mineral cycling in a Douglas-fir fores t 

stand . Health Phys . 11 : 1363-1369 . 

Rothacher, Jack. 1963 . Net precipitation 

under a Douglas-fir forest . Forest Sci . 9(4) : 

423-429 . 

Tamm, Carl O. 1951. Removal of plant nutrients 

from tree crowns by rain . Physiol . 

Plant. (4): 184-188 . 

Tarrant, R . F., Leo A. Isaac, and Robert F . 

Chandler, Jr . 1951. Observations on litte r 

fall and foliage nutrient content of som e 

Pacific Northwest tree species . J. For . 

49(12) : 914-915 . 

, K . C. Lu, C . S. Chen, and W . B . 

Bollen. 1968. Nitrogen content of precipitation 

in a coastal Oregon forest opening . 

Tellus 20(3) : 554-556 . 

Tukey, H . B., Jr., and H. J. Amling . 1958 . 

Leaching of foliage by rain and dew as an 

explanation of differences in the nutrien t 

composition of greenhouse and field-grown 

plants. Mich. Quart. Bull. 40(4) : 876-881 . 

, H. B. Tukey, and S. H. Wittwer. 

1958. Loss of nutrients by folia r 

leaching as determined by radioisotopes . 

Proc. Am. Soc. Hortic. Sci. 71 : 496-506 . 

Voigt, G. K. 1960. Alternation of the composition 

of rainwater by trees. Am . Midland 

Nat . 63(2) : 321-326 . 

Will, G . M. 1959. Nutrient return in litter an d 

rainfall under some exotic conifer stands i n 

New Zealand . New Zealand J. Agric. Res. 

2 : 719-734 . 

Wilm, H . G. 1943. Determining net rainfal l 

under a conifer forest . J. Agric. Res. 67 : 

501-512 . 

143

Estimating Biomass 

and Other State Variables 

145


Bellingham, Washington-March 23-24, 1972 . 

Direct, nondestructive measurement of 

biomass and structure in living, 

old-growth Douglas-fir 

William C . Denison, 

Diane M . Tracy , 

Frederick M . Rhoade s 

Department of Botan y 

and Plant Pathology 



Martha Sherwood 

Department of Biolog y 

University of Orego n 

Eugene, Orego n 

Abstract 

Previous studies of biomass and structure of Douglas-fir have examined trees less than 100 years old. This 

paper describes methods for measuring older trees, illustrated by data from a tree 60 m tall and 450 years old. 

Rock climbing techniques, modified for use on trees, are employed to climb the main trunk . A movable spa r 

provides access to lateral branches. The trunk is measured, the position of each branch system is located on it , 

and the branch systems are scored for 10 variables related to biomass and structure. An importance value, 

calculated for each branch system, is used in selecting a set of branch systems for detailed measurement. The 

data permit diagramatic reconstruction of the tree, or estimates of the distribution or total amount of component 

parts . 

Introduction 

The techniques described in this paper are 

of two kinds ; technical climbing methods an d 

methods of tree description and measuremen t 

which depend upon technical climbing to pro - 

vide access to the top of the tree . In combination, 

these techniques are designed to provid e 

quantitative descriptions of the biomass, surface 

area, and spatial distribution of th e 

aboveground parts of individual trees . 

The climbing techniques developed from a 

National Science Foundation-Undergraduate 

Research Participation study of epiphytes o n 

old-growth Douglas-fir (Pseudotsuga 

menziesii) . Previous studies (Coleman, Muenscher, 

and Charles 1956; Hoffman and Kazmierski 

1969) were limited to epiphytes 

within 2 m of the ground, but this study wa s 

designed to study the canopy populations a s 

well. An initial attempt to use felled trees wa s 

unsuccessful because the surface of the trun k 

which hit the ground was destroyed and th e 

branch systems with their epiphytes wer e 

shattered and scattered . The climbing techniques 

were developed as an alternative an d 

proved to be practical, effective, and 

economical . 

Five old-growth Douglas-firs have bee n 

rigged and climbed to date, but only one has 

been subjected to the measurement and analysis 

described herein. Previous studies of biomass 

in this species (Burger 1935 ; Newbol d 

1967 ; Reukema 1961) have been limited t o 

trees less than 100 years old, but the tree s 

used in our study are 450 years old and large ; 

147

60-80 m tall, 1-1 .5-m dbh, with their lowest 

branches 12-25 m above the ground . 

Climbing trees of this height is dangerous . 

A fall would probably be fatal . But probably 

there is greater danger of injury to personnel 

under the tree in the event equipment or part s 

of branch systems are accidentally dropped 

upon them . We have worked through two sea - 

sons without injury . 

The climbing techniques are described in 

detail to permit others to adopt them . Th e 

rock-climbing techniques on which they are 

based are described in books on mountaineering 

(Blackshaw 1970 ; Manning 1967) but may 

be unfamiliar to biologists . 

The access technique, the methods o f 

description and measurement, and one application 

of the data, a diagramatic reconstruction 

of a tree, are treated in this paper . 

Another paper in this symposium (Pike et al . 

1972) describes the estimation of tree surfac e 

area and epiphyte biomass . 

Riggin g 

Rigging involves an initial ascent and preparation 

of the tree for subsequent climbing . 

The ascent is slow, and requires exceptiona l 

agility and endurance . A team of three experienced 

climbers should rig the tallest trees i n 

a day and a half . 

The procedure used in ascending the tre e 

and placing climbing and belay ropes is outlined 

in figures 1-4 . 

1 

Figure 1 . A lag screw is used to fasten a steel hanger 

to the tree . 

The Access Techniques 

Access to the tree involves three steps : rigging, 

climbing, and use of the spar, eac h 

described below. Rigging and climbing techniques 

are modified from direct-aid rock - 

climbing techniques . Basic safety procedures , 

equipment, terminology, and philosophy hav e 

been adopted from mountaineering . Anyone 

without prior mountaineering experience who 

expects to adopt these techniques should consult 

one of the standard texts (Blackshaw 

1970; Manning 1967) . The specialized rope 

and hardware may be obtained from either 

mountaineering or yachting suppliers . 

Basic equipment for all climbers includes : 

hard hat, heavy climbing shoes, and a harnes s 

of nylon webbing to which a belay rope i s 

attached. The climber is always belayed-tha t 

is, she is protected by a safety rope held b y 

another experienced climber . Climbing an d 

belay ropes are 11 mm nylon "goldline" an d 

the webbing used in making slings, stirrups , 

etc . is of nylon and 40-50 mm wide . 

Figure 2 . A carabiner is clipped into the hanger . 

148

Figure 3. Two additional carabiners are used to faste n 

two climbing stirrups to the first carabiner. (Only 

the upper ends of the webbing stirrups are shown . ) 

The belay rope is clipped through the carabiner . 

5 

Figure 4 . The climber ascends the climbing stirrup s 

until the first carabiner is at her waist . Tension on 

the belay rope holds her in position while sh e 

drives the next lag screw . 

6 

Figure 5 . The climbing rope is fastened to the tree b y 

four lag screws, hangers, and carabiners . A loop, Figure 6 . The belay rope runs through a block hun g 

tied near the end of the rope, is connected to the by heavy chain with welded links. The chain i s 

carabiners by two loops of webbing. The free end fastened around the tree above a limb and secure d 

of the rope is tied around the tree using a bowline . by two 1-cm bolts with washers and lock nuts . 

149

Lag screws, 8 by 150 mm, are used t o 

attach equipment. They do not sag or "work 

out" of the thick, soft bark of Douglas-fir as 

nails do . The screws are driven with a hammer 

but may be tightened or removed with a 

wrench . 

A climbing rope is attached near the top of 

the tree and subsequent ascents are made on 

this fixed rope . Figure 5 illustrates the usual 

method of attachment. 

The block for the belay line is attached 

below the point at which the climb path i s 

obstructed by branches. In climbing abov e 

this point the belay rope is carried up throug h 

carabiners placed during rigging . Below th e 

block, on the open trunk, the climber is belayed 

from above through the block . Figure 6 

illustrates the method of attachment for th e 

block . 

Climbing 

Climbing requires mastery of unfamiliar 

skills, but no greater strength or agility tha n 

climbing a ladder, a 70-m ladder . An experienced 

climber, in good physical condition , 

can climb a 75-m tree in 20 minutes or less , 

depending upon the distribution of branche s 

and the lean of the trunk. A clumsy, middleaged 

man can clamber up with equal safety , 

but it takes longer . 

The techniques involved in climbing th e 

fixed rope are illustrated in figures 7-13 . 

7 

8 9 

Figure 7 . A band of rubber cut from a motorcycle 

innertube is fastened around the sling and over the 

toe of the boot to prevent the climber from accidentally 

stepping out of the sling . 

Figure 8 . When weight is placed on the jumar, it s 

toothed gate grips the rope so that the jumar ca n 

move neither up nor down . 

Figure 9 . When there is no weight on the jumar, it s 

gate may be lowered and the jumar moved up or 

down the rope . 

Figure 10. The jumar cannot be removed from th e 

rope unless the safety catch is depressed to permit 

the gate to open fully . 

150

1 1 

Figure 11 . The climber places her weight on on e 

foot, thus locking the jumar on that side, while sh e 

moves the other jumar up or down . She ascends or 

descends by shifting her weight from one foot t o 

the other and moving the opposing jumars. Sh e 

stands in webbing slings which hang from th e 

jumars . The jumars are connected to each other by 

a short length of rope which passes through a carabiner 

clipped to her belay harness . 

Figures 12 and 13. At those points at which th e 

climbing rope is fastened by a carabiner, th e 

jumars must be disconnected from the rope, one at 

a time, and moved around the carabiner . 

Most of the carabiners used in rigging ar e 

removed once the tree is rigged, but a few ar e 

left in place to prevent the climbing rop e 

from hanging away from the trunk . 

Shouted instructions from climber to belayer 

are normally adequate during climbing , 

but once the climber enters the canopy conversation 

is severely impeded by distance an d 

intervening foliage . A pair of small radi o 

transceivers facilitates longer communication s 

and permits the belayer to record data dictated 

by the climber . 

The Spar 

Use of a special boom, hereafter called "the 

spar," permits the climber to reach any point 

within 4 m of the trunk . 

The spar is specially constructed of three 

lengths of mast-grade spruce laminated in a 

"tee" cross section. It is 4 m long and together 

with its hardware and rigging it weighs 

15 kg. It is awkward to maneuver into place , 

especially where there are branches close 

above the one to be sampled. Under normal 

151

14 

Figure 14 . The inboard end of the spar. The hinge is 

of steel, 8 mm thick except for the central block , 

which is 36 by 36 mm . The top of the spar is a 

single piece of spruce, 22 mm high, 75 mm wide , 

and 4 m long . The bottom is of two pieces, eac h 

22 mm thick, 100 mm high, and 4 m long, glued 

face to face . 

Figure 15. The outboard end of the spar. Note that 

the supporting ropes are not fastened to the out - 

board end . They pass through blocks and guides 

on opposite sides of the spar and are secured at th e 

hinge end . 

circumstances it should be possible to raise 

the spar from the ground and have it ready 

for use in 2 hours or less ; in difficult circumstances 

it may take a full day . 

Figures 14-17 illustrate the constructio n 

and operation of the spar . 

The ropes used to support the outboard 

end of the spar are of nylon, but, unlike thos e 

used in climbing, are of a special braid whic h 

minimizes stretching (Samson Yachtbraid) . 

This prevents the spar from sagging out of 

position as the weight of the climber moves 

away from the tree . 

Note that the ropes supporting the spar ar e 

fastened at the inboard end of the spar, not a t 

the outboard end . The position at which th e 

spar rests depends upon the relative tensio n 

on these ropes . Thus a climber on the spar 

may maneuver herself through 180° of arc , 

without returning to the trunk, by pulling on 

one or the other of the ropes running along 

the spar . 

The climber sits in a "swing seat" suspended 

by loops of webbing . A climbing stirrup 

permits her to move along the spar b y 

alternately sitting in the seat while she move s 

the stirrup and standing in the stirrup whil e 

she moves the seat . 

In moving the spar, the ropes are coiled an d 

the hinge is folded back along the spar an d 

fastened . In placing it on the tree, it is hung , 

hinge downward, and maneuvered until th e 

hinge is in place . The hinge is fastened to th e 

tree with lag screws . Then the outboard end i s 

lowered into place. In removing the spar the 

process is reversed . 

The use of these three techniques in combination 

enables the climber to work with 

comparative freedom and safety for hours at a 

time if need be . 

152

Figure 16. The spar in use. The suspending ropes pass 

upward through carabiners on opposite sides of 

the tree and are tied off near the inboard end of 

the spar . The spar is shown in the open but woul d 

normally be placed adjacent to a branch system . 

Figure 17 . Raising the spar . A "trolley" of light 

nylon line is stretched from a clear area on th e 

trunk to an open area on the ground . The spar is 

packaged, with its ropes coiled and hinge folded , 

and clipped to the "trolley" with carabiners . A 

handline and pulley aid the climber in pulling th e 

spar up into the tree . 

Tree Description 

and Measurement 

Description and measurement of a tree proceeds 

in two steps : an initial survey, in whic h 

the trunk and all of the branches are de - 

scribed, followed by detailed measurement o f 

a sample set of branch systems . The tree i s 

arbitrarily subdivided into components : th e 

main trunk plus a number of branch systems . 

A branch system is defined as one or mor e 

branches, living or dead, arising from the same 

point on the trunk. Several branch system s 

may arise at the same height, but at differen t 

points around the circumference of the trunk . 

For examples of kinds of data recorded and 

calculated, the reader should refer to table 1 

while reading the following sections . 

Survey 

As the tree is climbed, the trunk is marked 

off vertically in meters, by use of tape. At 

convenient intervals, 5 to 10 m, a benchmar k 

is established and its height verified by transit 

from the ground. The diameter and inclination 

of the trunk are recorded at the benchmarks. 

It is not safe to climb above the poin t 

where the trunk is 10 cm or less in diameter , 

so we have treated the portion above tha t 

point as an exceptional branch system . When , 

through damage to the original leader, one or 

more secondary leaders have developed, the y 

are treated in the same fashion as the main 

trunk . 

153

Each branch system is numbered. The position 

of each branch system on the trunk i s 

recorded by height and compass quadrant . 

Each branch system is scored for 10 variables. 

Most of these variables relate to the 

structure and biomass of the tree, but sinc e 

this study began as a survey of epiphyte distribution, 

three variables relate to epiphyte load . 

1. Number of Main Axes 

A "main axis" is defined as a branch more 

than 4 cm in diameter, originating within 1 

dm of the trunk. Six classes are recognized : 

no axes (i .e., no branch more than 4 cm i n 

diameter), one axis, two axes, three axes, fou r 

axes, and five or more axes . 

2. Extension from Trun k 

The extension is measured as the horizonta l 

distance from the trunk to the furthest tip o f 

any part of the branch system . Five classes are 

recognized : 0-1 m, 1-3 m, 3-5 m, 5-10 m, and 

more than 10 m . 

3. Total Length of Living Axe s 

An "axis" is any branch more than 4 cm i n 

diameter. The total estimated here is of th e 

length of all living axes within the branc h 

system . Six classes are recognized : 0-1 m, 1-5 

m, 5-10 m, 10-15 m, and more than 20 m . 

4. Total Length of Dead Axe s 

The definition and classes are as in 3 above , 

except that here the branches are dead . 

5. Area of Branchlets and Foliag e 

An estimate is made of the area of a n 

imaginary figure formed by connecting th e 

outer tips of all branchlets with the point o f 

attachment to the trunk and projecting th e 

outline on a horizontal plane . Six classes ar e 

recognized : 0-1 m 2 , 1-2 m 2 , 2-5 m2 , 5-10 m 2 , 

10-20 m 2 , and more than 20 m 2 . 

6. Density of Branchlets and Foliage 

Within the area delimited in 5 above, a n 

estimate is made of the area occupied b y 

branchlets and foliage . Six classes, expressed 

as percents of the total area, are recognized : 0 

percent (no foliage), 0-20 percent, 20-40 percent, 

40-60 percent, 60-80 percent, an d 

80-100 percent . 

7. Maximum Diameter of Axi s 

The maximum diameter of the largest axi s 

is measured to the nearest centimeter . 

8. Lichen Cover : All Specie s 

The lichen cover is estimated for all axes in 

the branch system . Only foliose and fruticos e 

species are counted . The cover is expressed a s 

a percent of the total surface area on all axes . 

Six classes are recognized : 0 percent (n o 

lichens), 0-20 percent, 20-40 percent, 40-60 

percent, 60-80 percent, and 80-100 percent . 

9. Lichen Cover : Lobaria species 

The cover is estimated and expressed for 

Lobaria species alone, using the same classes 

listed in 8 above . 

10. Bryophyte Cover : All Specie s 

The cover is estimated and expressed for al l 

bryophytes using the same classes listed in 8 

above . 

The data from this survey are tabulated , 

with the branch systems arranged by number , 

in the order in which they occur on the tree . 

A sample group of branch systems is the n 

selected for detailed measurement . 

Sample Selectio n 

The number of branch systems selected fo r 

detailed measurement (n) depends upon th e 

complexity of the tree : two branch system s 

are selected in a small, uniform tree ; three, i n 

a normal tree ; four or five, in a tree with more 

than one leader . In our example, five branc h 

systems were selected and measured . 

An importance value (v) is calculated fo r 

each branch system according to the following 

formula : 

where , 

v = b 2 +b 1 2 + ad + 1 + m 

b = class number (1-6) of total lengt h 

of living axes 

b 1 class number (1-6) of total lengt h 

of dead axe s 

a = class number (1-6) of area of 

branchlets and foliage 

d = class number (1-6) of density o f 

branchlets and foliage 

l = class number (1-6) of lichen cover , 

all specie s 

m = class number (1-6) of bryophyte cove r 

The importance value is listed for each 

branch system, and both running totals and a 

grand total (V) are calculated . The grand total 

(2,111 in our example) is set equal to the 

number of branch systems to be selected (five 

154

in our example) and the running totals recalculated 

. For example, branch number 87 has 

an uncorrected running total of 28 . The corrected 

total is: 28(5/2,111)=0 .066. After thi s 

correction, each branch system is represented 

by an interval, proportional to its importance 

value, which is the difference between 

its corrected running total and that of 

the preceding branch system . 

A random number is drawn : a three-place 

fraction of 1 .000 . In our example, the number 

drawn was 0 .870. Those branch systems 

are selected which have intervals which include 

: the random number, 1 + the random 

number, 2 + the random number, etc . In ou r 

example, the numbers 0 .87, 1.870, 2 .870 , 

3 .870, and 4 .870 corresponded to branc h 

system numbers 7, 33, 61, 93, and 116 . 

The branch systems selected in this way ar e 

measured in detail as described below . 

Branch System Measuremen t 

The measurement of each selected branc h 

system is carried out in 10 stages . The first 

four stages occur in the field and require the 

use of the spar ; the remaining stages occur i n 

the laboratory . 

1. Each axis is marked off in decimeter 

lengths and the total length of the main axe s 

is recorded . 

2. The diameter is recorded every 4 dm . 

3. The transition points (4-cm diam) a t 

which an axis becomes a branchlet are 

identified . 

4. At one-fourth of the transition point s 

the distal portion, consisting of branchlets 

and attached foliage, is cut off and lowered t o 

the ground for further study . 

5. The annual rings are counted at the cu t 

end of the removed branch and the age of the 

sample is recorded . 

6. The foliage-bearing portion of each 

sample is cut into lengths bearing a singl e 

year's needles . 

7. A subsample of 20 of each year 's 

needles is measured (length and width) while 

fresh . 

8. The foliage and attached fragments of 

branchlet are dried at 85° and weighed . Th e 

weight of each year's foliage is recorded, as is 

the weight of its associated segments o f 

branchlet . 

9. The older branchlets, those without 

foliage, are cut into lengths representin g 

approximately 1 year 's growth and their 

lengths and diameters are recorded . 

10. The segments of older branchlet are 

dried at 85° and weighed . 

Analysi s 

Several types of information can be summarized 

from data accumulated by the techniques 

described . Two of these, surface are a 

of the trunk and limbs and epiphyte biomass , 

are described in a paper by Pike et al .(1972) . 

Analyses of wood and foliage biomass ar e 

being developed, but at present our results ar e 

incomplete since they are based on a singl e 

tree . 

Another kind of analysis results in a diagram 

of the tree (fig . 18) ; a graphic summarization, 

to scale, of the structure of an individual 

tree . This diagram is an east-wes t 

vertical section . It serves both as a map fo r 

those working on the tree and as a chart of th e 

distribution of branch systems and foliage . 

Reconstruction of the trunk is based upon 

benchmark measurements of diameter, height , 

and inclination. A second leader, originatin g 

on the north side of the main trunk at th e 

40-m level, is drawn to the same scale but 

displaced to the right . Only those branc h 

systems are shown which project to the eas t 

or to the west. From existing data one could 

draw a comparable north-south section displaying 

the remaining branch systems . Eac h 

branch system is shown in its correct positio n 

on the trunk and the length of the longes t 

horizontal line indicates, to scale, the distanc e 

the entire branch system projected away fro m 

the trunk . However, the arrangement of axes 

and foliage within a branch system is symbolic 

rather than pictorial . 

Living branch systems are represented by a 

parallelogram, representing foliage, usuall y 

accompanied by one or more solid lines , 

representing axes . Parallelograms without 

solid lines are branch systems in which n o 

branch was more than 4 cm in diameter . 

155

156 

Figure 18 . Semidiagrammatic reconstruction of Tree Number 1, Watershed 10 . 

Only those branch systems extending to the east or west are shown . Th e 

second trunk is drawn to scale at the right height, but is displaced to the right . 

Solid horizontal lines represent living branches (axes) or branch systems 

more than 4 cm in diameter . The length of the line indicates, to scale, how fa r 

the whole branch system extends from the trunk . Dead branches are represented 

by dashed lines . The parallelograms represent branchlets and foliage . 

The length of a parallelogram represents the horizontal area covered by th e 

branch system, and the height represents the density of branchlets an d 

foliage within that area . Parallelograms without horizontal lines are branc h 

systems with no branches more than 4 cm in diameter . Numbered branc h 

systems are those selected for detailed measurement .

Dashed lines are dead branches . The area of a 

foliage parallelogram is a measure of the 

amount of foliage within the branch system . 

The length of a parallelogram is a function of 

the area of the branch system ; the height of a 

parallelogram is a function of the density of 

foliage within the branch system . 

Individual branch systems are easily identified. 

For example, those branch systems 

which were selected for detailed measuremen t 

(numbers 7, 33, 61, 93, and 116) are identified 

by number on the diagram . 

Discussion 

Although the access technique is still being 

improved, it is now very nearly routine fo r 

trees with a growth pattern similar t o 

Douglas-fir . It could undoubtedly be modified 

for use in measuring large trees with different 

growth patterns . It would be particularly 

interesting to attempt to apply the method to 

trees of the upper canopy in lowland tropica l 

wet forests . 

We have not had adequate opportunity t o 

evaluate the accuracy of our methods of 

description and measurement . The basic pat - 

tern seems satisfactory, but there will b e 

changes made before we extend it to additional 

trees . We will increase the precision o f 

initial estimates of lengths of axes . We need 

better methods for estimating foliage biomass . 

The importance value used in selecting branc h 

systems combined values for foliage biomas s 

with values for epiphyte load . This resulted i n 

selection of a subset that was a poor compromise 

between the conflicting requirements fo r 

measurement of tree biomass and measurement 

of epiphyte biomass . In the future , 

separate importance values will be calculate d 

for tree biomass and epiphyte biomass ; an d 

separate sets of branch systems will be selected 

and measured . 


The authors are grateful to many people 

who contributed to the development of thi s 

project . Don Kirkpatrick taught us the ascent 

technique and made the first ascent . Othe r 

climbers (Diane Nielsen, Tom Denison, Jan e 

McCauley, and Karen Berliner), working wit h 

the authors, have helped in data taking an d 

have made improvements in climbin g 

methods. Dr. Jack Culver (Benton Boa t 

Works) suggested improvements in the desig n 

of the spar, and James Ewanowski (Arborea l 

Constructions) built it. Sue Carpenter helpe d 

with the drawings . Dr. Scott Overton suggested 

the basic format of the sampling pattern 

and has guided its development . 

This project began (July-August 1970) as a 

student project under the National Scienc e 

Foundation Undergraduate Research Participation 

Program (Grant Number GY-7641) in 

the Department of Botany and Plant Pathology, 

Oregon State University . Subsequent 

work has been supported by National Science 

Foundation Grant Number GB-20963 to the 

Coniferous Forest Biome, U .S . Analysis o f 


This is Contribution No . 31 from the Coniferous 



Blackshaw, A . 1970 . Mountaineering . 552 p . 

Hammondsworth : Penguin Books . 

Burger, H . 1935 . Holz, Blattmenge and 

Zuwachs. II. Die Douglasie. Mitt. Schweiz . 

Anst. forstl . Vers. 19 : 21-72 . 

Coleman, B . B., W. C. Muenscher, and D . R . 

Charles . 1956 . A distributional study of th e 

epiphytic plants of the Olympic Peninsula , 

Washington . Am. Midland Nat . 56 : 54-87 . 

Hartley, H. O. 1966. Systematic samplin g 

with unequal probability and without re - 

placement. J. Am. Statist . Assoc . 61 : 

739-748 . 

Hoffman, G. R., and R . G. Kazmierski . 1969 . 

An ecological study of the epiphytic bryophytes 

on Pseudotsuga menziesii on the 

Olympic Peninsula, Washington. I . A 

description of the vegetation . Bryologist 

72 : 1-18 . 

Manning, H. (ed.), 1967. Mountaineering the 

freedom of the hills . 2d ed . 485 p . Seattle : 

The Mountaineers . 

157

Newbold, P . J. 1967 . Methods for estimatin g 

the primary production of forests . I .P.P . 

Handbook #2 . 62 p. Oxford, England : 

Blackwell Sci . Publ . 

Pike, Lawrence H., Diane M . Tracy, Martha A . 

Sherwood, and Diane Nielsen . 1972. Estimates 

of biomass and fixed nitrogen o f 

epiphytes from old-growth Douglas-fir . In 

Jerry F. Franklin, L . J . Dempster, an d 

Richard Waring (eds .), Proceedings-researc h 

on coniferous forest ecosystems-a symposium, 

p. 177-187, illus . Pac . Northwest 

Forest & Range Exp . Stn ., Portland, Oreg . 

Reukema, D . L. 1961 . Crown developmen t 

and its effect on stem growth of six 

Douglas-firs . J. For. 59 : 370-371 . 

Table 1 .-Exemplary data for branch systems on tree #1-a Douglas-fi r 

on Watershed 10, H . J. Andrews Experimental Fores t 

c 

w 

N 

so 

S 

O0 

m 

m o 

EO sO+ ., 

V Q 

.w 

G 

o ~ 

•d S+ 

C F 

Total Lengt h 

of Branches 

E al 

E ~ 

Lichen Cove r 

All 

Specie s 

-% - 

Lobari a 

-% - 

w E 

x 

o 

il w 

-m- 

x m 

Liv e 

-m - 

Dea d 

-m - 

N• m 

-cm- 

-% - 

w 

-c LL H 

O 0 

0 0 

WI V 

-%- 

88 59 .1 TOP 

87 59 .1 S 

86 58 .8 W 

85 58 .5 E 

84 58 .2 W 

83 58 .2 N 

82 58 .2 S 

81 57 .6 S 

22 41 .1 W 

21 40 .8 W 

20 40 .2 W 

19 39 .3 W 

18 38 .4 W 

17 38 .1 W 

2 20 .4 N 

1 15 .9 N 

128 49 .8 W 

127 49 .5 N 

93 40 .5 E 

92 40 .5 N 

91 40 .2 N 

90 40 .2 N 

89 39 .9 N 

1 1-3 1-5 30 2-5 10-20 40-60 40-60 15 1 5 

1 1-3 1-5 5 1-2 20-40 0-20 0-20 13 2 8 

1 1-3 1-5 6 1-2 0-20 0-20 0-20 11 3 9 

1 0-1 0-1 5 0-1 0-20 0-20 5 44 

1 1-3 1-5 5 1-2 020 0-20 0-20 11 5 5 

1 1-3 1-5 8 0-1 0-20 0-20 0-20 9 6 4 

1 0--1 0-1 5 0-1 0-20 0-20 0-20 6 7 0 

2 1-3 1-5 6 2-5 20-40 0-20 0-20 0-20 17 8 7 

1 1-3 1-5 9 0-1 40-60 0-20 0-20 11 94 9 

1 0-1 0-1 6 0-1 0-20 5 95 4 

1 1-3 1- 5 

8 0- 1 

0-20 12 96 6 

At this point the second trunk leaves the tree, 40m, N Compas s 

2 3-5 5-10 7 5-10 60-80 0-20 0-20 0-20 33 99 9 

1 5-10 5-10 8 2-5 20-40 20-40 0-20 22 102 1 

1 0-1 0-1 3 0-1 80-100 9 103 0 

5 3-5 10-15 8 10-20 40-60 0-20 0-20 20-40 41 137 8 

4 3-5 10-15 5-10 10 10-20 60-80 0-20 20-40 55 143 3 

Data for trunk 2 which left the tree between branch 19 & 2 0 

1 3-5 1-5 6 1-2 0-20 0-20 0-20 0-20 12 1445 

1 3-5 1-5 6 1-2 0-20 0-20 0-20 11 L45 6 

1 3-5 1-5 12 5-10 20-40 0-20 0-20 20-40 21 206 2 

1 1-3 1-5 9 0-1 20-40 0-20 20-40 11 207 3 

1 0-1 0-1 9 0-1 0-20 0-20 0-20 6 207 9 

1 3-5 5-10 11 2-5 20-40 0-20 0-20 20-40 23 210 2 

1 1-3 1-5 7 0-1 0-20 0-20 0-20 9 211 1 

158


Bellingham, Washington-March 23-24, 1972 

Estimation of biomass and transpiratio n 

in coniferous forests using 

tritiated water 

Abstract 

J . R . Kline, 

M . L . Stewart, 

C . F . Jorda n 

Radiological Physics Divisio n 

Argonne National Laboratory 

Argonne, Illinois 60439 

Nondestructive measurement of biomass and transpiration rates in trees is dependent on a new application of 

the established theory of tracer dynamics in steady state systems. The method utilizes HTO as a tracer for H2 0 

in the plant. Both transpiration and biomass measurements require experimental determination of tritiu m 

activity in the tree as a function of time. Biomass measurement requires one additional parameter-the mean 

residence time of water in the plant. In this paper we examine various theoretical and experimental alternative s 

for determining mean residence times. Data supporting one alternative is presented. The measurement of 

conducting tissue biomass is also discussed . 

Introduction 

Kinetic theory and tracer distributio n 

methods have been used extensively in measurement 

of the properties of flowing system s 

both in biology and engineering . The theoretical 

aspects and the underlying assumption s 

governing the practical application of tracer 

dynamics have been thoroughly discussed by 

authors in many fields. Reports by Zierler 

(1964), Bergner (1961, 1964a, 1964b, 1965 , 

1966), and Ljunggren (1967) have describe d 

the use of tracer dynamics and kinetic theory 

in the nondestructive measurement of flo w 

rates, mean residence times, and compartmental 

volumes of steady state biological an d 

engineering systems . 

It is the purpose of this paper to presen t 

the state of the art in the application of th e 

theory and tracer methods to the nondestructive 

measurement of tree biomass and transpiration, 

and to examine the experimenta l 

parameters that can be measured as well a s 

the underlying theoretical assumptions on 

which they are based . 

The nondestructive measurement of transpiration 

rates and biomass in trees utilize s 

tritiated water (HTO) as a tracer for water . 

Tritiated water is added to the tree water pool 

by injection into the trunk near ground level . 

The fate of the HTO tracer-labeled water in a 

tree is followed by monitoring tritium activity, 

as a function of time, in foliage and smal l 

branches. Both transpiration and biomas s 

measurements require experimental determination 

of tritium concentration-tim e 

curves. Biomass requires additionally th e 

mean residence time of water in the tree and 

the mean moisture content of the tree . 

Theoretical Discussion 

Transpiration Measurement 

Application of the theory of tracer dynamics 

to the problem of measuring transpira- 

159

tion rates in plants has been previously 

demonstrated' 2 (Kline et al . 1970) . The 

theory itself has been discussed extensively b y 

Bergner (1961, 1964a, 1964b, 1965, 1966) , 

Zierler (1964), Ljunggren (1967), and Orr and 

Gillespie (1964) . Transpiration measurement s 

depend upon use of the Stewart-Hamilto n 

equation shown by equation 1 : 

M = F f 0 - f(t)dt , (1 ) 

where M = total activity of tritium initially 

injected (disintegrations pe r 

minute, DPM) , 

F = the flow rate of water throug h 

the tree (ml/hour x tree) , 

f(t) = activity distribution of tritiu m 

at the points of exit from th e 

system (DPM/ml), and 

t = time (hr) . 

Equation 1 states simply that the product 

of the flow rate (F) and the total integral of 

the curve of activity versus time (fig. 1) is 

equal to the total activity of the tracer whic h 

was originally injected . In practice the 

activity-time curve is measured experimentally 

and the total activity injected is 

fixed by the experimenter . The flow rate (F ) 

is the only unknown quantity in equation 1 

and is solved algebraically . The value of F is 

the daily flow rate which is averaged over day - 

time and nighttime flows. The average i s 

taken over the full time interval of residenc e 

of tritium in the tree . Shorter term resolution 

of transpiration is not possible with thi s 

method . Examples of transpiration rate s 

which have been obtained using equation 1 

with tritiated water as the tracer are given in 

table 1 for field-grown coniferous trees . 

1 J. R. Kline, C. F. Jordan, and R . C . Rose . Transpiration 

measurements in pines using tritiated wate r 

as a tracer . In D. J. Nelson (ed .), Third Nationa l 

Symposium on Radioecology, Proc ., May 10-12 , 

1971, Oak Ridge, Tenn . (In press . ) 

'J . R . Kline, M. L. Stewart, C . F . Jordan, an d 

Patricia Kovac_ Use of tritiated water for determination 

of plant transpiration and biomass under fiel d 

conditions . In Symposium on the Use of Isotopes an d 

Radiation in Soil-Plant Research Including Applications 

in Forestry, Proc . Int . At . Energy Agency Conf . 

SM-151, December 13-17, 1971 .Vienna, , Austria . (I n 

press .) 

50 30 0 

Figure 1 . Typical activity-time response curve obtained 

by injecting a jack pine (Pins banksiana) 

tree with tritiated water and sampling twigs as a 

function of time for tritium content . Peak arriva l 

time is indicated by Tp . 

Biomass Measuremen t 

The measurement of tree biomass is base d 

on that part of the theory of tracer dynamic s 

which permits calculation of the pool size o f 

the compartment through which flow take s 

place . In trees the pool refers simply to th e 

average total water content of the tree . Whe n 

pool size has been computed, it is a simpl e 

matter to convert to biomass using the aver - 

age moisture percentage of the wood . 

Derivation of an expression which permit s 

computation of compartmental pool size i s 

given by Zierler (1964) . Zierler's expression i s 

given by equation 2 : 

C =F Tm , (2) 

where C = compartment volume (ml) , 

F = flow rate through the compartment 

(ml/hr), and 

Tm = mean residence time of th e 

flowing substance (hr) . 

Equation 2 states simply that the compartment 

pool size is given by the product of th e 

160

Table 1.-Transpiration rates, mean residence times, computed biomasses, and observe d 

biomasses for field-grown red and jack pine tree s 

Red pine (Pinus resinosa) (1970 ) 

Tree 

number 

Dbh 

(CM) 

Transpiration 

rate (F ) 

(ml/hr) 

resi ence 

time (Tp ) 

(hr) 

Computed 

dry biomas s 

(kg) 

Observed 

dry biomass 

(kg) 

3 14 .1 1170 50 42 .4 42 . 6 

4 9.5 222 110 17 .7 12 . 2 

5 12 .1 935 42 28 .4 26 . 3 

6 13 .3 1280 42 38 .9 35 . 1 

7 12 .3 725 49 25 .7 26 . 7 

8 8 .2 413 49 14 .6 12 . 1 

10 12 .3 254 98 18 .0 20.6 

Mean tree weight ± SE 26 .5 ± 4 .1 25 .1 ± 4 . 3 

Mean forest biomass ± SE (kg/ha) 7.7 x 10 4 ± 1 .2 7.3 x 10 4 ± 1 . 2 

Jack pine (Pinus banksiana) (1971 ) 

Tree 

number 

Dbh 

(CM) 

Transpiration 

rate (F) 

( 

/hr) 

re isdeece 

time (T ) 

(hr) p 

Computed 

dry biomass 

(kg ) 

Observe d 

dry biomass 

( kg) 

1 10 .5 807 49 38.5 19 . 9 

2 12 .0 833 52 33.5 26 .7 

3 10 .0 540 48 21 .5 19 . 9 

4 10.0 941 27 25.6 32 . 0 

5 8.9 570 45 25 .4 18 . 1 

6 11 .0 833 25 19 .2 26 . 0 

7 11 .0 609 23 17 .0 32 . 3 

8 9.5 506 46 24 .6 24 . 6 

9 10 .4 492 27 18 .3 20 . 9 

11 11 .4 1083 26 28 .0 28 . 5 

12 9.4 903 24 27 .2 20 . 0 

Mean tree weight ± SE 23.3 ± 2 .0 24 .4 ± 1 . 5 

Mean forest biomass ± SE (kg/ha) 6.3 x 10 4 ± 0 .5 6.1 x 104 ± 0 .4 

16 1

flow rate and the mean residence time of th e 

system . 

In plants the pool size (C) is also given b y 

the difference between wet and dry weigh t 

(W-D) of the plant . Moisture fraction is conventionally 

calculated by equation 3 : 

W-D 

W 

= f , 

where W = wet weight of sample (gm) , 

(3) 

D = dry weight of sample (gm), an d 

f = fractional moisture content . 

Equation 3 holds equally well for the cas e 

where the determination is done on the entire 

plant or for a representative subsample of the 

plant. In the case where the entire plant is the 

sample, W-D can be substituted for C in equa - 

tion 1 resulting in equation 4, which is an 

expression for the moist weight of the plant : 

W=1F•Tm . (4) 

f 

Moist weight can be converted to dry weigh t 

using equation 3 . This results in equation 5 

which is an expression for dry biomass of th e 

plant : 

D = 11-f FTm . (5) 

f 

Equation 5 requires the experimental determination 

of f, F, and Tm for its solution . In 

the absence of a feasible method for deter - 

mining f on the entire tree, it is necessary t o 

measure it on subsamples. Ideally the subsamples 

should be weighted for different tre e 

parts such as trunk, branches, and leaves . 

Since there is usually no method available for 

measuring weighting factors, we have followed 

the practice of estimating moistur e 

content of the tree trunk since this represent s 

the greater portion of the biomass of the tree . 

In our experience, moisture content of plan t 

parts has not been greatly different from one 

another and no serious errors are introduced 

by following this procedure . If plants are 

found where unweighted estimates of f diffe r 

from the true weighted value, then this coul d 

be a significant source of error in the estimate 

of biomass . 

The value of F in equation 5 is the mean 

flow rate which has previously been calculated 

using equation 1 . This means that th e 

reliability of the estimate of biomass can b e 

no better in general than the reliability o f 

flow or transpiration rate . Biomass estimate s 

are normally expected to have lower statistical 

precision than transpiration estimate s 

since they require the use of additional measured 

parameters . 

The most difficult parameter to measure i n 

equation 5 is Tm, the mean residence time . 

There appear to be at least three possibl e 

approaches to obtain this quantity : (1) measure 

the slope of the declining branch of th e 

activity-time curve ; (2) compute the firs t 

moment of the curve ; or (3) measure th e 

transit time between the point of injection 

and the point of exit of the tracer from the 

system (Donato et al . 1964) . 

The slope method for mean residence tim e 

would be valid in the case where the plant is 

labeled to equilibrium with the tracer . If all of 

the water molecules of the trees were labeled 

equally with HTO, then the rate at which tritium 

activity declines in the tree would b e 

proportional to the amount of tritiu m 

present . Such systems are described by an 

equation of the form A = AO e- Xt where AO i s 

initial activity, A is activity at time t and X i s 

the rate constant of loss . The term X is th e 

slope when data described by this relationshi p 

are plotted on semilogarithmic coordinates . 

The mean residence time for such a relation - 

ship is simply the reciprocal of X (Tm = 

This is a frequently used relationship for obtaining 

Tm although in practice it is some - 

times used without verification of the assumptions. 

In some systems it is possible to label t o 

equilibrium by injecting the tracer into the 

system continuously ; however, in large trees 

this would entail larger than desirable releases 

of radioactivity to the environment . Therefore, 

it is preferable to label by the instantaneous 

pulse method . When the pulse method 

of labeling is used, the injected material 

moves upward in the tree while retaining its 

pulse shape. The pulse may undergq consider - 

able broadening but the system cannot be 

assumed to have achieved uniform labeling . In 

162

this case the slopes of activity-time curve s 

have two components, one reflecting the turn - 

over rate of water and the other reflecting the 

pulse shape. There is no method currently 

available for resolving these components in a n 

activity-time curve, and therefore the slope of 

the curve cannot validly be used to compute 

Tm . 

Ljunggren (1967) has described the compu - 

tation of mean residence times for flowin g 

systems using the first moment of th e 

activity-time curve. The first moment is th e 

centroidal axis of the activity-time distribution-that 

vertical axis which divides th e 

distribution into two parts having equal areas . 

Equation 6 indicates the method : 

f0`°t f(t) dt 

Tm* _ (6) 

f ff(t) dt 

In general, Tm* will always be the tru e 

mean residence time of the tracer in th e 

system under study . The tracer mean residence 

time of the system will satisfy the relationship 

of equation 2, however, only under 

the conditions of identical behavior of trace r 

and substance traced . In trees, the nominal 

mean residence time for water, Tm, is no t 

equal to the tracer mean residence time Tm * 

since the tracer apparently undergoes interactions 

with the conducting vessels of the 

tree . The relationship is given by the 

expression 

Tm *=f 1 Tm +f2 TH 

where TH is the residence time of the fractio n 

of tritium which has undergone some interaction 

with the wood and fl and f2 are fractions 

of the total tritium which pass throug h 

the tree without interaction and with inter - 

action, respectively . Possible interactions include 

isotopic exchange of tritium wit h 

hydrogen of the wood or diffusion of tritiu m 

into non flowing compartments of plant 

water. Experimentally Tm*>Tm has been 

found for all trees which have been examine d 

to date, indicating that the term f2 TH has a 

nonzero value in trees . This phenomenon has 

been termed "holdback" by others who have 

examined tracer dynamics in flowing vessels 

(Ljunggren 1967) . Because of "holdback" the 

value of Tm* cannot be used without correction 

to solve equation 2 . The problem of finding 

an appropriate value for f2 TH is presentl y 

unsolved . 

In general, the mean residence time for a 

flowing system can always be obtained by 

measuring the activity distribution of a trace r 

in the system at two points along the flow 

pathway (Ljunggren 1967) . If Ti is the time 

of passage of the tracer at point 1 and T2 the 

time of passage at point 2 further down - 

stream, then Tm = T2 - Ti where Tm is the 

mean residence time between the two sampling 

points. Since the tracer normally undergoes 

peak broadening, Ti and T2 are taken as 

the times when the peak of the distributio n 

passes the sampling points . In trees we fix th e 

initial position of the tracer by the injectio n 

at time Ti = O . In this special case Tm = T2 . 

The value of T2 could be measured at an y 

point in the tree downstream to the injection 

point. In the special case where the downstream 

sampling point is tree foliage, the n 

Tm = Tp where Tp is simply the time of pea k 

arrival in the foliage . 

As a first approximation it can be assumed 

that Tp is not affected by "holdback" as was 

Tm* because tritium is probably remove d 

from free flowing forms equally over the en - 

tire activity distribution . That is, interactio n 

of tritium with conducting vessels could a s 

well occur with the isotope in the leadin g 

edge of the distribution or the trailing edge . 

The peak position would, therefore, not b e 

affected by these interactions . This assumption 

requires experimental verification whic h 

is given in the results section . 

Correction for Nonconducting Tissue 

The foregoing suggests that tritium trace r 

experiments can only be used to measur e 

actually conducting biomass in trees . Such tissues 

as bark, flowers, fruit, and nonconducting 

heartwood will not be included in the estimate. 

Roots are not included in the estimate 

since the tracer injection is normally done i n 

tree trunks above the roots . In practical biomass 

measurements for forestry purposes, th e 

163

most serious problem is the omission of non - 

conducting heartwood from the direct measurement. 

The theory of tracer dynamics cannot 

be used for direct measurement of thi s 

quantity, and it is, therefore, necessary t o 

make an approximation . Equation 7 expresse s 

the relationship between total biomass an d 

the biomass of heartwood and sapwood : 

where VT 

VTPT = VHPH + V SPS , (7) 

total volume of tree tissu e 

( cm3 ) , 

weighted mean wood 

density (cm3 ) , 

VH ; VS= volume of heartwood an d 

sapwood (cm 3 ), an d 

PH ;PS 

density of heartwood an d 

sapwood ( ) 

cm 3 

Substituting the relationship s 

PH/PS = K 

and VH/V S = A 

into equation 7 results in equation 8 : 

VTPT = V SpS ( AK + 1) . ( 8 ) 

Assuming that the volume of heartwoo d 

and that of the total tree can be approximated 

by a right circular cone, an expression 

for A is derived as follows : 

r2 

where r 

PT 

X = Ar (2r + Ar ) 

= mean radius of heartwood at 

base (cm), and 

Ar = mean thickness of sapwoo d 

at base (cm) . 

Upon substituting for A in equation 8, a fina l 

expression for tree biomass is obtained : 

r 2 K 

VTpT = D = V S p S ( Or(2r + Dr) + 1) 

. (9) 

The numerical group V.T.I. is the total plan t 

biomass (D) and the group VsPs is the conducting 

or sapwood biomass as measured b y 

the tritium method. Equation 9 is not sensitive 

to the assumption that wood volumes ar e 

approximated by right circular cones. Th e 

same result is obtained for a right circular 

cylinder or for intermediate figures . 

Equation 9 has not yet been evaluated experimentally 

. It is proposed here to suggest 

some of the anticipated lines of research to b e 

undertaken in the Coniferous Biome. A principal 

problem for the solution of equation 9 

lies in accurate measurement of the quantitie s 

r and Ar . These quantities fundamentally refer 

to the radii of nonconducting and conductin g 

wood, respectively . In the simplest case they 

may be coincident with heartwood and sapwood 

as observed visually. Their evaluatio n 

could then be done by straightforward measurement 

of tree cores . 

An accurate solution also depends on th e 

nature of the transition zone between con - 

ducting and nonconducting tissue . If this is 

sharply defined, then r and Ar will be wel l 

defined and an accurate solution to equation 

9 can be obtained . If the conduction under - 

goes a gradual transition across the radius o f 

the tree, there may be no practical method 

for assigning values to r and An It is possibl e 

that reasonable values can be obtained b y 

studying tritium distribution along wood 

cores which have been taken from tritium - 

labeled trees. These considerations apply t o 

large trees which have appreciable volumes o f 

heartwood . In the trees for which we hav e 

experimental data, heartwood was a mino r 

part of the total tree volume and was no t 

considered . 

Results 

The mean residence time Tm is the most 

difficult parameter of equation 5 to obtain , 

principally because it is not generally known a 

priori which of several possible means of computing 

it is the correct one . The desired value 

is the nominal mean residence time Tm 

(equation 2); however, in the usual non - 

destructive experiment this is not directly obtainable 

. Where tritium-injected trees have 

been harvested, however, the water pool siz e 

(C) is directly obtainable from biomass an d 

moisture measurements, and it is possible to 

164

solve equation 2 for Tm since the flow rat e 

(F) is known . The values of Tm obtained b y 

this direct method were compared to th e 

values obtained by the three indirect method s 

which were previously discussed . 

The nominal mean residence time (Tm) ha s 

been computed for several harvested field - 

grown coniferous trees and the values compared 

with those obtained from curve slopes , 

from first moment calculations and from peak 

arrival times . The results show in general, that 

Tm > TS for slope calculations and Tm < 

Tm * for first moment calculations . 

The nominal mean residence time agree s 

most closely with Tp which was obtained b y 

the midpeak or peak arrival time method . 

2 0 

I 0 

o ► 1 I I_ I 1 

0 10 20 30 40 50 6000 110 120 13 0 

Tm(hr ) 

Figure 2 . Relationship between nominal mean residence 

time (Tm) of plant water and mean residence 

time obtained by peak arrival (Tp) fro m 

activity-time curves . 

Figure 2 shows the relationship between T m 

and Tp for a group of harvested trees . Th e 

slope of the linear least squares relationshi p 

between the two estimates is 0 .916, and the 

intercept is 7 .05 as compared with the expected 

slope of one and expected intercept of 

zero (1 :1 line, fig. 2). The coefficient o f 

determination (r2 ) is 0.70 for the relationship. 

The results suggest that Tp is an unbiased 

estimator of Tm . This confirms that Tp 

is the best estimate available for Tm since 

other known possibilities have been examined 

and found to be biased either above or belo w 

the true values . 

The scatter of data points on figure 2 indicates 

that forest biomass determinations must 

C 

at present be approached statistically . Any 

particular determination of mean residenc e 

time of water may be substantially in error ; 

however, a group of determinations appear s 

to converge on the true Tm values of th e 

group. Experimental error reduction is on e 

objective of continuing studies in the Coniferous 

Biome program . 

Table 1 shows a comparison between tre e 

biomass as computed by equation 2 an d 

actual harvested weights. These data were obtained 

in a uniform age plantation of red pin e 

(Pinus resinosa) in 1970, and a similar stan d 

of jack pine (Pinus banksiana) in 1971. Th e 

results show individual examples of substantial 

error; however, the estimates of mean tre e 

biomass and mean forest biomass as measured 

by the tritium method agree well with thos e 

estimated by direct harvest . There is, of 

course, no necessity for determining biomas s 

of a group of trees on all individuals simultaneously 

. Biomass can be determined at an y 

time during which transpiration flow is takin g 

place. The results suggest that the biomass of 

these forests could have been reliably determined 

by the tritium method alone . 

Conclusions 

Experiments designed to measure transpiration 

rates in field-grown trees may also b e 

used with little extra data collection for non - 

destructively measuring tree biomass . The tritium 

method can, in general, be used only for 

determination of biomass which is actually 

transmitting water . Flowers, fruits, bark, and 

nonconducting heartwood are not included i n 

the estimate. We have proposed a method for 

including heartwood in the estimate but have 

not yet proved it experimentally . 

The most difficult parameter to obtain for 

calculation of biomass is the nominal mea n 

residence time of water in the plant . After 

examining several alternatives for obtainin g 

this quantity, we conclude from theory and 

experiment that the time of arrival of pea k 

tritium activity in tree foliage is the most reliable 

estimate of nominal mean residence time . 

Estimates of biomass of field-grown coniferous 

trees show considerable statistical varia - 

165

tion; however, the mean of a group of such 

determinations was a reliable and unbiase d 

estimate of the mean obtained by direct harvest. 

With additional research on the problem 

it may be possible to reduce the experimental 

error of the method and to improve the 

reliability of individual estimates . 


The work reported in this paper was performed 

under the auspices of the U .S . Atomic 

Energy Commission and supported in part b y 

National Science Foundation Grant No . 


U .S . Analysis of Ecosystems, Internationa l 




Bergner, P .-E . E. 1961 . Tracer dynamics: I . 

A tentative approach and definition of 

fundamental concepts . J. Theor. Biol. 2 : 

120-140 . 

. 1964a . Tracer dynamics and th e 

determination of pool sizes and turnove r 

factors in metabolic systems . J. Theor . 

Biol . 6 : 137-158 . 

. 1964b. Kinetic theory : Som e 

aspects on the study of metabolic processes 

. In R. M. Kniseley and W . N. Taux e 

(eds.), Dynamic clinical studies with radioisotopes, 

p. 1-18 . U .S. At. Energy Comm . 

TID-7678. Germantown, Md . 

. 1965 . Exchangeable mass : 

Determination without assumption of isotopic 

equilibrium . Science 150 : 1048-1050 . 

r . 1966. Tracer theory: A review . 

Isotope & Radiat . Tech . 3 : 245-262 . 

Donato, L ., C. Giuntini, R . Bianchi, and A . 

Maseri . 1964 . Quantitative radiocardiography 

for the measurement of pulmonary 

blood volume . In R. M. Kniseley 

and W. N. Tauxe (eds .), Dynamic clinical 

studies with radioisotopes, p . 267-283 . U .S . 

At. Energy Comm . TID-7678 . German - 

town, Md . 

Kline, J . R ., J. R. Martin, C . F . Jordan, and J . 

J. Koranda. 1970 . Measurement of transpiration 

in tropical trees using tritiate d 

water. Ecology 5(6) : 1068-1073 . 

Ljunggren, K . 1967. A review of the use of 

radioisotope tracers for evaluating parameters 

pertaining to the flow of materials in 

plant and natural systems. Isotope & 

Radiat. Tech. 5(1) : 3-24 . 

Orr, J. S., and F . C. Gillespie . 1968. Occupancy 

principle for radioactive tracers i n 

steady state biological systems . Scienc e 

162 : 138-139 . 

Zierler, K. L. 1964. Basic aspects of kineti c 

theory as applied to tracer distributio n 

studies . In R. M. Kniseley and W . N. Taux e 

(eds.), Dynamic clinical studies with radioisotopes, 

p . 55-79 . U .S . At. Energy Comm . 

TID-7678. Germantown, Md . 

166



Theodolite surveying fo r 

nondestructive biomas s 

sampling 

Eugene E . Addor, Botanist 

U .S . Army Corps of Engineers 

Waterways Experiment Statio n 

Vicksburg, Mississippi 93180 

A bstract 

By theodolite surveying, the relative location of points in space may be calculated by triangulation . With th e 

aid of computers, data gathered by theodolite surveying may provide a dimensional analysis of individual trees. 

Because the system is nondestructive, the rates and patterns of change in the spatial structure of trees and stand s 

may be monitored by repetitive surveying. This paper presents a preliminary test of the approach upon trees in a 

40-year-old Douglas-fir (Pseudotsuga menziesii) plantation in western Washington. From experience gained in 

the initial experiment, recommendations are made to increase the precision of repetitive measurements . 

Introduction 

The theodolite is an instrument used in precise 

surveying to locate points in space by triangulation 

. The use of high speed computers 

for converting angle measurements to poin t 

locations allows theodolite surveying techniques 

to be used for describing the physical 

structure of vegetation assemblages in considerable 

detail with relative ease . Such a surveying 

procedure has been used to construc t 

simulation models for studying the effects of 

vegetation on engineering activities (West e t 

al . 1971) . Since the method is nondestructive , 

it offers the possibility of repetitive samplin g 

with high inherent precision. Exploratory 

surveys were made recently to test the applicability 

of the system for this purpose (West 

and Allen 1971) . This paper examines som e 

data from a Douglas-fir stand at the Coniferous 

Biome intensive site in Washington . 

Description of the 

Surveying System 

The procedure requires two theodolite s 

placed at an arbitrary distance apart and located 

conveniently to the subject trees (fig . 

1) . The vertical and horizontal angles fro m 

each instrument to every point located in th e 

sample space are measured with respect to a 

base line. The instruments can be move d 

about to obtain clear lines-of-site to desire d 

points in the sample space and every instrument 

location (turning point) is referenced t o 

the base line by conventional traverse survey - 

Figure 1 . Instrument set-up for surveying spatia l 

structures of trees. Two theodolites are in use ; th e 

instrument in the middle is a spotting laser . 

167

ieoo -- 

~e0o - 

I/00 - 

1200 - 

10o0 - 

u 

X 

N 

boo - 

eoo -- 

40 0 

200 

0- /s.scm -.0-IA4 CM 

I I I I I I I I I I I I I l l I l l l l l l l l l 

-300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 3, 

Y AXIS, C M 

a . APRILMEASUREMENT 

b. OCTOBERMEASUREMENT 

Figure 2 . Graphic display of point location data from a Douglas-fir tree near Seattle, Washington, on whic h 

branching was not surveyed in detail . Bole graphed to show diameter to scale. 

168

ing methods . Every point in the sample spac e 

is thus located with respect to any arbitraril y 

defined three-coordinate system . Details of 

the procedure are presented elsewhere (West 

and Allen 1971) . 1 

One of the theodolites is equipped with a 

specially designed circular reticle for measuring 

branch or stem diameter, employing th e 

principle of stadia measurement . All measure d 

angles and reticle readings are recorded in th e 

field on specially designed data forms, an d 

trigonometric conversions of field data t o 

point locations and stem or branch diameter s 

are made by computer . Figure 2 is an exampl e 

of a computer graphic of one of the trees o n 

the Thompson site . 

Description of the Sample 

The Douglas-fir stand, located on the A. E . 

Thompson Research Area in the Cedar Rive r 

watershed, lies some 64 km southeast of 

Seattle, Washington . The Research Area is 

described in detail by Cole and Gessel (1968) . 

Measurements were obtained from a group o f 

eight contiguous trees on each of two proximate 

(not contiguous) permanent research 

plots (designated 1 and 2 on the Researc h 

Area) within an even-aged 40-year-old plantation 

. Plot 1 received three applications o f 

nitrogen as ammonium sulfate (NH 4 SO 4 at 

the rate of 222 kg/ha in October 1963 , 

October 1964, and May 1970 . Plot 2 was left 

untreated as a control . 

The eight trees selected for measuremen t 

on each plot were selected first by choosin g 

an arbitrary point within each plot, and the n 

taking the eight trees nearest to each point a s 

sampling trees. The only controlling criterio n 

placed upon location of the starting point o n 

each plot was that it should be far enough 

within the plot to avoid inclusion of boundary 

trees in the sample . Measurements on the 

sample trees were made in April 1970, befor e 

bud burst, and then again in October 1970 , 

after the apparent end of the growing season . 

Data were taken to include the coordinate 

location and the diameter (outside hark) at 

the following points : 

a) At the base of the tree, defined as being 

at the duff line or ground line, as wel l 

as could be determined . Diameters were 

measured with a tape . 

b) Diameters at breast height (d .b.h . ) 

150 cm above the duff line were 

marked with a ribbon for subsequent 

remeasurement . 

c) The bole at every fourth whorl where 

limbs were still present (diameters wer e 

calculated from reticle readings) . 

d) The base of the live crown, defined as 

the lowermost whorl at which more 

than 50 percent of the branches hel d 

green leaves . The location of a whorl i s 

defined as the approximate centroid of 

branch emergence ; diameter measurements 

on the bole are made just below 

the lowermost branch as well as just 

above the uppermost branch of th e 

whorl (fig. 3). (Diameters calculated 

from reticle readings . ) 

e) At every fourth branch whorl within 

the live crown, or at least one branc h 

whorl within the middle one-third of 

the live crown . 

f) The topmost whorl in April and that 

same whorl plus the new topmos t 

whorl in October . 

g) The top of the leader, at the base of th e 

terminal bud whorl . 

1 E . E . Addor and H . W . West. A technique fo r 

measuring the three-dimensional geometry of standing 

trees . U .S. Army Waterways Experiment Station , 

Vicksburg, Mississippi . Unpublished . 

Figure 3. Definition of branch whorl location, an d 

location of diameter measurements at whorl . 

169

h) Point locations and diameters were 

measured for various defined points o n 

crown branches but will not be discussed 

here . 

A total of 30 turning points (instrument 

set-ups) were established for the April survey 

and all were referenced to a common coordinate 

system . The same points of reference 

were used again in October, but no deliberate 

attempt was made to duplicate the surveyin g 

sequence, e .g., to site each point on the tree 

and to measure each diameter from the same 

turning point . Nonetheless, the sequence tha t 

was adopted for the first survey was approximated 

during the second survey, as a result of 

constraints within the stand . Thus the 

sampled points were mostly viewed from th e 

sample angles during both surveys . Since both 

surveys were referenced to the same coordinate 

system, the reported coordinate locations 

of surveyed points are in theory exactl y 

comparable, so that any difference in th e 

reported location of a point represents a displacement 

of that point by wind action , 

growth, or survey error . 

Results of 

Theodolite Survey 

Crown Cover and Stand Densit y 

The crown cover was essentially closed an d 

the branching structure relatively dense . The 

ground area occupied by the eight sample d 

trees on each plot was determined in April b y 

traversing the ground points representing th e 

outer crown limits of the outermost trees o f 

the group . The crown coverage so determine d 

was 31 .2 m 2 on plot 2 (unfertilized) an d 

44 .0 m 2 on plot 1 (fertilized) representing a 

crown area per tree of 3 .9 m 2 and 5 .5 m 2 , 

respectively, or a density of approximatel y 

2,567 and 1,818 trees per hectare . 

In this same stand, in October, 1965, Dic e 

(1970) destructively analyzed 10 trees from a 

0.0045-hectare (45 m 2 ) plot, which is 4 .5 m 2 

per tree, or approximately 2,222 trees per 

hectare . These values lie reasonably between 

our values for the unfertilized and fertilized 

trees. Unfortunately we are not certain ho w 

the crown boundaries of his trees relate to th e 

boundaries of his 45 m 2 plot . 

The problem of the true relation of the 

crown cover per tree (tree mean area) to 

sample plot boundaries is controversial (Greig - 

Smith 1964) . Supposedly, tree randomness 

with respect to sample-plot boundaries should 

balance the excluded and included portions of 

included and excluded trees, but the true relation 

is apparently complicated by both plot 

size and plot shape . A crown-limit traverse is 

relatively simple with a theodolite survey and 

is easily converted to area by the computer . It 

should therefore be worthwhile to examine 

whether such a procedure would resolve th e 

problem of the relation between sample plo t 

boundary and tree crown boundary in th e 

determination of crown cover, stand density , 

or tree mean area . 

Patterns in Diameter Measurement s 

Examination of the diameter data from th e 

theodolite survey suggests that the unfertilized 

trees exhibited greater increment on th e 

upper portion of the bole than on the lower , 

whereas the fertilized trees showed approximately 

equal growth pattern throughout th e 

length of the bole . Such patterns seem reason - 

able because of the difference in / stan d 

density. Similar patterns have been reporte d 

in unthinned and thinned stands of Douglas - 

fir surveyed with an optical dendrometer over 

a 2-year period (Groman and Berg 1971) . 

Certain sources of inaccuracies in diameter 

measurements with the theodolite syste m 

should be mentioned . First, diameters calculated 

from reticle readings are dependen t 

upon accurate measurements of the distance . 

Second, interpolation errors from this source 

may be important when estimating smal l 

branch diameters with the reticle . Countering 

these disadvantages is the possibility of 

measuring diameter at any point on a stem or 

branch regardless of the direction or angle of 

inclination . Other instruments such as the 

optical dendrometer have no reticle inscripted 

and thus are restricted to measuring bole 

diameter. 

170

Patterns in Point Displacemen t 

Measuring changes in the physical structur e 

of vegetation consists primarily of simply 

measuring the displacement of defined points 

over a specified time interval . Obviously, an 

error in determining the location of a poin t 

either at the beginning or at the end of th e 

time interval will result in an error in th e 

measurement of displacement . 

Measurement errors may stem from a 

variety of sources, depending upon th e 

methods of measurement . Since theodolite 

surveying is dependent upon calculation o f 

point locations by trigonometric relations , 

both instruments must be precisely sighted o n 

the spot to be located . Disparities in th e 

assumed location of the target point wil l 

cause errors in the calculated location of th e 

point. Horizontal disparities will cause horizontal 

and vertical errors according to th e 

angle of convergence and the slope of lines-ofsite, 

while a vertical disparity will not locate a 

point at all . 

To reduce errors from this cause, a spottin g 

laser (fig. 1) can be used to project a bright 

orange spot a few millimeters in diameter 

onto the tree at the selected target point . Thi s 

provides a definitive target for sighting th e 

theodolites at one given time, but it does not 

resolve the problem of relocating the exact 

point of measurement for periodic remeasurement 

. The spotting laser was at the Thompson 

Site during both the April and October 

surveys, but it was inoperable much of the 

time. Periods of its use and nonuse may 

account for some of the patterns in the data . 

Other sources of error include the usua l 

reading and transcription errors by instrument 

men and note keepers. Errors from these 

various sources may or may not be critical , 

depending upon their magnitude and frequency, 

and the special purpose for which th e 

survey is being made . For the purpose of 

monitoring subtle changes in the vegetation 

structure over a brief time period, even ver y 

small errors may be important . Gross 

anomalies in the data may be identified and 

approximately corrected during data editing 

and preliminary analysis, but small error s 

regardless of source may not be distinguish - 

able from true displacement. 

For the purpose of discussion, any difference 

between the calculated location of a 

defined point from one observation to 

another (specifically, for the present case , 

from April to October), in any coordinat e 

direction, may be defined as an "apparent displacement" 

of that point in that direction . I t 

can then be defined that the apparent displacement 

always consists of two components 

: True displacement resulting fro m 

changes in the shapes of the trees, and errors 

resulting from inaccuracies and mistakes i n 

instrument reading, note keeping, and calcula - 

tions . Hereinafter, these latter will be referred 

to collectively as "survey error ." The question 

to be resolved, then, is what proportion o f 

apparent displacement can be attributed t o 

each of these two components . Data from the 

surveyed Douglas-fir stand provide a n 

opportunity to examine the theodolite 

surveying system with respect to thes e 

problems . 

Figure 4 is a set of graphs of the apparent 

displacement in the xy (horizontal) plant of 

defined points at various levels on the tree 

boles, as measured from April to October . 

They are : (A) at the base of the tree, (B) at 

the base of the live crown, (C) at an arbitrary 

intracrown whorl, and (D) at the whorl that 

was defined as the topmost whorl in Apri l 

(i .e., at the base of the April leader, three 

trees are omitted from the intracrown whorl 

data due to omissions in the survey) . With few 

exceptions, the apparent displacement in the 

horizontal plane at the base of the tree i s 

within plus or minus 3 cm, with a slight systematic 

bias in the positive x direction and in 

the negative y direction, but with the points 

for the unfertilized and fertilized trees reason - 

ably well interspersed . Since trees are 

anchored at the base, it may be assumed that 

any apparent displacement of the tree axis in 

the horizontal plane at that level must represent 

a surveying error . Therefore this sligh t 

systematic error at this level on these tree s 

may be interpreted as a horizontal error i n 

relocating the established origin of the coordinate 

system. The absence of a separation i n 

this plane at this level between the apparent 

displacement of the unfertilized and fertilize d 

171

+10 

q 

A 

0 -29 ~---o-- - 

D 

-A &A 

>. 

o"f 

I 

0 0 

o 

o FERTILIZE D 

q UNFERTILIZE D 

I I 1 

1 

q +37 

o 

-~ 

-1 0 

1 q + 1 9 

o--> 

of 

0 

I 1 

-10 0 +10 -10 

APPARENT DISPLACEMENT, C M 

0 

(x) 

+10 

Figure 4 . Apparent displacement of the tree boles in x and y (the horizontal plane) at various levels on th e 

sampled trees, from April to October : (A) at the base of the trees, (B) at the base of the live crowns, (C) a t 

an intracrown whorl, (D) at the base of the April leader (based on data from table 2 in West and Allen 1971) . 

172

trees indicates that the coordinate system wa s 

surveyed across the intervening distance between 

the two plots (about 20 m) with negligible 

error in this plane. If the data are adjusted 

for the horizontal error in relocation of 

the coordinate system origin, then the displacement 

error is quite small, and it must b e 

conceded that remeasurement of the horizontal 

angles to the trees has been achieve d 

with a fair degree of success, despite the 3 0 

turning points used in accomplishing th e 

survey . 

Two explanations may be suggested for th e 

increasing scatter of points in the horizontal 

plane with increasing height on the tree , 

shown on figure 4B, C, and D. First, it may be 

assumed that surveying errors have increase d 

with increasing elevation of lines-of-site, or 

second, it may be assumed that the positio n 

of the boles are less stable in the horizontal 

plane at higher levels on the tree . Since point 

locations in the horizontal plane are calculated 

from horizontal angles, irrespective of 

angles of elevation, there is no reason t o 

assume that surveying errors should increas e 

in relation to elevation. It follows therefore 

that errors in the location of points in th e 

horizontal plane at any elevation might b e 

equal to, but should not exceed, the errors in 

location of the base of the trees in this plane . 

It follows in turn that the apparent horizonta l 

displacement of points on the upper boles of 

these trees must be a true displacement . The 

pattern is consistent with what would b e 

expected as a result of movement of the trees 

by wind . 

Figure 5 is a set of graphs of the apparent 

displacement of z (the vertical plane, or elevation) 

for the same defined points on the bol e 

that are shown on figure 4, respectively . 

These show a considerable scatter in th e 

apparent vertical displacement at the base o f 

the live crown, moderate scatter in apparen t 

vertical displacement at the intracrown whorl , 

and again a relatively close clustering of 

apparent displacement at the base of the April 

leader . 

This pattern of variation may be attributed 

to relative differences in the difficulty o f 

identifying the defined points with respect 

to elevation. This difficulty is more or less 

inherent in the definition of the points ; that is 

to say that "at or near the duff or ground 

line" is less definitive than is "the centroid of 

the branch whorl," whereas the topmos t 

whorl (base of the leader) is the smallest an d 

most definitive of the defined points. The 

differences in vertical scatter of points for th e 

base of the live crowns and for the intracrow n 

whorl may be attributed to errors of approximating 

the exact elevational location of th e 

latter through obscuring branches and foliage . 

These inconsistencies of identification hav e 

important implications with regard to measurement 

of length relations, such as the tota l 

height of trees or the ratio of bole length t o 

length of live crown (although, of course, th e 

significance of the implication is determine d 

by the magnitude of the dimensions and th e 

special purpose for which the relations ar e 

being measured) . 

It was observed above that the apparent 

displacement of defined points on the unfertilized 

and the fertilized trees were reasonably 

well interspersed with respect to the horizontal 

plane . With respect to elevation, how - 

ever, the data exhibit systematic tendencie s 

that require explanation . Specifically, figure 

5B shows a seemingly strong tendency toward 

a positive apparent displacement of about 6 

or 7 cm for the base of live crowns on th e 

fertilized trees, while the apparent displacement 

of this point on the unfertilized trees 

appears to be randomly dispersed about zero . 

A possible explanation is that a vertical error 

was committed in extending the coordinat e 

system across the distance (approximatel y 

20 m) between the two groups of trees . However, 

if this were the case, then the sam e 

apparent vertical displacement must occur a t 

all other levels on the fertilized trees ; if it 

does not, then its absence from other level s 

must be accounted for . The data for the intracrown 

whorl (fig. 5C), though somewhat 

more scattered than the data for the base o f 

live crown, do indeed suggest a similar disparity 

between central tendencies for the tw o 

groups of trees, but a similar disparity is no t 

obvious at the base of the trees (fig . 5A), nor 

at the base of the April leaders (fig . 5D) . It 

may be that for the base of the tree, the disparity 

is simply obscured by the scatter of the 

173

30 

2 2 0 

V 

A B C D 

0 

q 

0 q 

0 

0 

0 

q 0 

A& 

II 

0 

A 

0 FERTILIZE D 

q UNFERTILIZE D 

-30 

Figure 5 . Apparent displacement of points on the tree boles in x and y as a function of z (elevation), from Apri l 

to October: (A) at the base of the trees, (B) at the base of the live crowns, (C) at an intracrown whorl, (D) a t 

the base of the April leader (based on data from table 2 in West and Allen 1971) . 

174

data; there is no obvious explanation for it s 

absence at the base of the April leader . 

Conclusions 

The preliminary survey of Douglas-fir tree s 

at the Thompson Site suggests that theodolite 

surveying is adaptable to the purpose o f 

detecting and monitoring subtle patterns o f 

change in vegetation structures . In terms of 

quantity and quality of data obtained for th e 

energy expended, this procedure appears t o 

be commensurate with any other known tre e 

measurement system, and it offers advantage s 

not offered by any other system . First, be - 

cause the location of every point in the 

sample is determined relative to an arbitrar y 

point defined as the origin of the coordinate 

system, the apparent displacement of any 

point on the tree is independent of th e 

apparent location of any other point on th e 

tree. Second, displacement of points in any 

direction can be measured with this procedure, 

such as, for example, the tips o f 

branches radially disposed about the stem an d 

growing at various angles from the vertical . 

Finally, and of considerable importance, thi s 

system is entirely nondestructive and is there - 

fore well suited to continued monitoring o f 

growth trends over extended time periods . 

Theoretically, the precision of the technique 

is within millimeters, since it is basicall y 

the same as is used for precision engineerin g 

surveying. There are, however, a few practica l 

limitations on the attainable precision . 

Following are a few observations about particular 

problems . 

A possible solution to the problem of 

target point identification would be to clim b 

the trees prior to the initial survey and afi x 

permanent sighting targets at points of interest. 

Also, a network of similarly smal l 

definitive targets could be establishe d 

throughout the sample area for use as permanent 

reference points, one of which could b e 

used to define the origin of the coordinate 

system. A series of carefully controlled 

experiments should be designed specifically t o 

determine the effects of operator error on the 

limits of attainable precision under different 

kinds of working conditions . 

Dense crown branches and foliage may 

place constraints on the usefulness or convenience 

of this method in some kinds o f 

vegetation. 

When precision is required, surveyin g 

should be avoided during periods of adverse 

weather conditions. Tarpaulins can be suspended 

over the instruments so that work 

may be carried on during rain or snow, but 

these conditions also affect visibility within 

the forest . Winds that are strong enough to 

cause movement of the trees will obviously 

increase the probability of meaningless 

apparent displacements, and should be 

avoided . 

As of this writing, the system has been use d 

exclusively for the dimensioning of trees . Th e 

principles upon which it is based, however , 

are universally applicable mathematical relations. 

Therefore the technique should b e 

eminently suited to the study of a variety o f 

ecological problems involving relations that 

can be described in a three-coordinate system . 

These might include, for example, the structure 

of bird rookeries, the spatial arrangemen t 

of epiphyte or parasite plants, or flower and 

fruit distributions . 


The work on which this paper is based wa s 

authorized and financed by the Recreation and 

Environmental Branch, Office of the Chief of 

Engineers, Washington, D .C., in cooperation 

with the Coniferous Forest Biome, U .S. Analysis 

of Ecosystems, International Biological Program. 

This is Contribution No . 33 to the 

Coniferous Forest Biome. I thank my referees , 

whose constructive criticisms were helpful in 

arriving at tenable explanations for some perplexing 

patterns in the data . 


Cole, D . W., and S. P. Gessel . 1968 . Ceda r 


pathways, rates, and processes of elementa l 

cycling in a forest ecosystem . Univ . Wash . 

Coll. For . Res . Monogr . Contr. No . 4, 53 p . 

175

Dice, Steven F . 1970. The biomass and nutrient 

flux in a second growth Douglas-fir ecosystem 

(a study in quantitative ecology) . 

53 p. Ph.D. thesis on file, Univ . Wash . , 

Seattle . 

Greig-Smith, P . 1964 . Quantitative plan t 

ecology. Ed. 2, 256 p., illus . Washington , 

D.C . : Butterworths . 

Groman, William A ., and Alan B. Berg . 1971 . 

Optical dendrometer measurement of increment 

characteristics on the boles of 

Douglas-fir in thinned and unthinned 

stands. Northwest Sci . 45(3) : 171-177 . 

and H. H . Allen . 1971. A technique 

for quantifying forest stands fo r 

management evaluations . U .S . Army Eng . 

Waterways Exp . Stn. Tech . Rep . M-71-9 , 

29 p ., illus . 

West, H . W ., R . R . Friesz, E . A. Dardeau, Jr . , 

and others . 1971 . Environmental characterization 

of munitions test sites : Vol. 1 , 

Techniques and analysis of data . U .S. Army 

Eng. Waterways Exp . Stn. Tech. Rep . 

M-71-3, 31 p ., illus . 

176



Estimates of biomass and fixed 

nitrogen of epiphytes from 

old growth Douglas fir 

Lawrence H . Pike and Diane M . Tracy 

Department of Botan y 

Oregon State Universit y 


Martha A . Sherwoo d 

Department of Biology 

University of Orego n 

Eugene, Orego n 

Diane Nielse n 

Division of Ecology and Systematic s 

Cornell University 

Ithaca, New Yor k 

bstract - 

Epiphytes are sampled concurrently with measurements of surface area of trunk and branch systems of 

old-growth Douglas-fir (Pseudotsuga menziesii) . Crude predictions of epiphyte biomass in branch systems are 

corrected by more detailed sampling of a subset of branch systems. Nitrogen analyses enable conversion of 

epiphyte biomass to the total amount of nitrogen present in the epiphytes. 

Introduction 

Epiphytic lichens and mosses are a conspicuous 

component of forest ecosystems in th e 

Pacific Northwest . Because of their ability t o 

concentrate materials from the environment 

and the ability of some of them to fi x 

atmospheric nitrogen, their importance i n 

nutrient cycling within the system may be 

greater than their contribution to total biomass 

would suggest . 

In old-growth Douglas-fir forests, epiphyt e 

biomass is expected to be in a steady state ; 

epiphyte growth is balanced by litterfall, in 

situ decomposition, and consumption b y 

herbivores. Annual turnover of epiphytic 

lichens varies from 5 to 25 percent (Edward s 

et al. 1960, Pike 1971) . 

Epiphyte-fixed nitrogen accounts, at least 

in part, for the nitrogen needed for th e 

growth of these epiphytes, and may represen t 

a significant input to the forest ecosystem . 

Nitrogen is added to water flowing over tre e 

surfaces through decomposition of epiphyte s 

and leakage from nitrogen-fixing epiphytes , 

and enters the soil from the epiphyte syste m 

via throughfall, stemflow, and litterfall . 

Measurements of biomass are necessary t o 

relate process studies of epiphytes to thei r 

contribution to forest mineral cycles on an 

ecosystem level . Estimating epiphyte biomass 

by felling and sampling selected trees is 

neither possible nor desirable in old-growt h 

Douglas-fir forests (where trees may approac h 

100 m in height) because such felling i s 

destructive, not only of the host tree, but also 

of the epiphytes one wishes to study . 

This paper outlines the methods and give s 

177

preliminary results of techniques develope d 

for sampling epiphytes on old-growt h 

Douglas-fir in the H . J. Andrews Experimenta l 

Forest . The relatively nondestructive samplin g 

is an extension of the procedure for estimating 

tree structure and biomass described b y 

Denison et al . (1972) . 

The Study Tree 

Methods 

Epiphyte sampling by the methods de - 

scribed here has been carried out on a single 

old-growth Douglas-fir tree . This tree i s 

located on a north-facing slope in watershe d 

10 of the H . J. Andrews Experimental Forest , 

75 km east of Eugene, Oregon . The base of 

this 65 m tall tree is at an elevation of 500 m . 

The surrounding stand of old-growth Douglas - 

fir has an understory which includes wester n 

hemlock (Tsuga heterophylla) and vine mapl e 

(Acer circinatum) . Epiphytic lichens an d 

bryophytes are present on the understory 

trees and shrubs as well as the overstory 

Douglas-fir . 

Sampling of Epiphytes 

For sampling, the tree was divided into 

trunk(s) and branch systems . The trunk is the 

main vertical axis of the tree ; branch systems 

are sets of branches leaving the trunk at the 

same point and are made up of axes (>4 c m 

in diameter) and branchlets (

where C is the percentage cover by epiphytes , 

r is the radius at the base of the axis, and h i s 

the length of the axis . The bulk of epiphyte 

biomass for these branch systems is assume d 

to occur on the axes; if this is not the case , 

the biomass of epiphytes on the many small 

branch systems, which are made up entirely 

of branchlets, may be significantly underestimated 

. 

Since EIV is based partly on subjective estimates 

of length of axes and percentage cove r 

by epiphytes, the relationship between EI V 

and epiphyte biomass may be expected t o 

vary from worker to worker and from sampling 

period to sampling period, and for thi s 

reason will be treated separately for each tree . 

However, it should be possible to correct 

importance values so that correlations may be 

obtained that will hold across a set of trees . 

Sampling Branch Systems 

The basic sampling unit for epiphytes o n 

axes in the branch systems is a "cylindrat, " 

which is analagous to a quadrat but runs completely 

around the axis so that two edges ar e 

fused . Our 1 dm cylindrats include the entir e 

surface for a distance of 1 dm along an axis . 

Thus the surface area of the cylindrat varies , 

depending on the diameter of the axis in th e 

region sampled . 

Cylindrats were spaced along an axis with a 

distance of 4 dm from the center of on e 

cylindrat to the center of the next . Th e 

distance from the trunk to the first cylindra t 

sampled along a main axis was varied from 0 

to 1, 2, and 3 dm so as to avoid errors whic h 

would be introduced by horizontal zonation 

on the axis near its origin from the trunk . 

Estimates of epiphyte cover were made, an d 

then the epiphytes were stripped from th e 

cylindrat and bagged . Axis diameter at th e 

center of each cylindrat was measured t o 

enable calculation of surface area of the axes . 

Branchlets (

anch system . Total epiphyte weight for a 

branch system is the sum of the weights o n 

axes and on branchlets . 

Nitrogen Analyse s 

Samples analyzed for nitrogen conten t 

were collected in watershed 10 in August an d 

October 1971 . Samples were air dried, an d 

nitrogen analyses were performed by Denni s 

Lavender of the Forestry Sciences Laboratory, 

Oregon State University, using th e 

Kjeldahl method . Air drying samples analyze d 

for nitrogen content avoids losses of nitroge n 

that may occur with ovendrying . In order to 

enable expression of the nitrogen contents o n 

an ovendry-weight basis, air dried samples of 

epiphytes were ovendried at 100 0 C to determine 

weight loss on drying . 

Results 

The results presented here are preliminary 

results from the first tree sampled in water - 

shed 10 of the H . J. Andrews Experimenta l 

Forest . The estimates are crude . They are presented 

to give an idea of how the methodology 

is being applied and the order of magnitude 

of results that are being obtained . Thes e 

preliminary results are being used in improving 

and refining the sampling strategy . 

Biomass of Epiphytes on the Trun k 

Total epiphyte biomass on the trunks of 

tree 1 is estimated to be 4.5 kg (table 1) . 

Nearly 90 percent of this is bryophytes, an d 

nearly 50 percent is found within 8 m of th e 

ground. Lichens contribute the bulk of th e 

epiphyte biomass on the trunk from about 5 0 

to 60 m from the ground . Epiphyte biomas s 

per unit area for the trunk as a whole i s 

0.31 g/dm 2 , and is much higher than thi s 

figure only within 8 m of the ground and a t 

the top of the second trunk, where large 

patches of Lobaria oregana (Tuck.) Mull. Arg. 

were encountered . 

Biomass of Epiphytes on the Branch System s 

The frequency distribution of branch sys - 

tems by EIV is presented in figure 1 . 

Epiphyte biomass on axes of the five 

branch systems sampled in detail ranged fro m 

0 to 198 g (table 2) ; that on the branchlets 

ranged from 1 to 48 g (table 3) . The five 

branch systems show a relationship betwee n 

EIV and epiphyte biomass (fig . 2) . Values 

from the least-squares regression line wer e 

used to convert the number of branch system s 

in an EIV class to an estimate of the tota l 

epiphyte biomass represented by that clas s 

(fig. 3) . 

The low importance value classes 

(EIV < 10 ), although representing man y 

branch systems, make only a small contribution 

to the epiphyte totals for the tree . Abou t 

one-half of the epiphyte biomass is contributed 

by branch systems with EIV abov e 

35 . In this connection, the preliminary nature 

of these results must again be emphasized a s 

no branch system with an EIV higher than 3 2 

was selected for detailed sampling . 

Epiphyte Biomass for the Whole Tre e 

Total epiphyte biomass for the tre e 

sampled is estimated to be 18 .3 kg ; 13 .8 kg , 

or about 75 percent of the total, is on the 

branch systems (table 4) . Assuming that the 

distribution of epiphyte biomass by specie s 

on the five branch systems studied is the sam e 

as the distribution overall, 50 percent of the 

total epiphyte biomass for the tree is bryophytes 

(table 4) . However, there is a decrease 

in the proportion of total epiphyte weight 

represented by bryophytes from larger 

diameter to smaller diameter stem sections . 

Bryophytes make up 86 percent of th e 

epiphyte biomass on the trunk, 47 percent of 

that on the axes of the branch systems 

(>4 cm diameter), and only 3 percent of that 

on the branchlets (

Table 1.-Epiphyte biomass on the trunk of tree 1 in watershed 10 ' 

Species 

y 

Heigh t 

(m) 

Trunk 

surface 

area 

L 

tl " o 

rz es 

CI .o0 

o 

tl tl 

tlb0 

o 

~l o 

All 

species 

din ' 

g 

g 

g/dm z 

TRUNK 1 : 

0-4 1,500 300 470 340 + 1,120 0 .75 

4-8 1,240 80 210 740 - 1,040 .83 

8-12 1,130 20 200 90 30 340 .30 

12-16 1,110 50 150 40 20 + 270 .24 

16-20 1,060 140 140 70 20 360 .34 

20-24 1,010 20 110 10 - 10 - - 150 .1 5 

24-28 960 + + 10 -- 10 .0 1 

28-32 900 20 60 20 90 .1 0 

32-36 860 + 10 10 + + 20 .0 2 

36-40 820 200 30 50 + + 290 .3 6 

40-44 750 + + + + 10 .0 1 

44-48 670 10 110 10 130 .1 9 

48-52 580 + 30 + + + 40 .0 7 

52-56 490 + + + 10 50 - 70 .1 3 

56-60 402 - 50 10 -- + 60 .1 4 

60-63 250 70 10 10 + 10 - 100 .3 9 

TRUNK 2 : 

41-4 5 440 10 + 10 - + - + - 10 .0 3 

45-49 310 60 + 20 - - 10 - 90 .28 

49-51 120 - + + - - 10 + 10 310 330 2 .84 

Total 14,610 850 1,570 1,480 120 20 10 110 20 10 10 320 4,530 

' Weights for each species of epiphyte are to the closest 10 g ; + indicates the presence of less than 5 g . 

Entries have been rounded and will not necessarily add to the total . 

18 1

Table 2.-Epiphyte biomass on axes of branch systems sampled ' 

Branc h 

system 

Distance from 

beginning of axis 

Section of axi s 

Diamete r 

at base 

Surface 

area 

Epiphyte biomas s 

dm cm dm 2 g/dm 2 g 

7 2 

33 0-3 .5 9 .0 9 .62 0 0 

3.5-7 .5 8 .5 10 .37 0 0 

7.5-11 .5 8 .0 9 .74 .01 . 1 

11.5-15 .5 7 .5 8 .80 .02 . 2 

15.5-19 .5 6 .5 8 .17 .02 . 1 

19.5-23 .5 6 .5 7 .38 .02 . 1 

23.5-27 .5 5 .2 6 .28 .02 . 1 

27.5-31 .5 4 .8 5 .66 .02 . 1 

31.5-33 .5 4 .2 2 .59 .04 . 1 

Total 68 .6 . 9 

61 0-2 .5 8 .0 6 .28 .24 1 . 5 

2.5-4 .5 8 .0 5 .34 .30 1 . 6 

4.5-8 .5 9 .0 11 .00 .31 3 . 4 

8.5-12 .5 8 .5 10 .21 2 .50 25 . 6 

12.5-16 .5 7 .8 9 .27 2 .97 27 . 6 

16.5-18 .5 7 .0 4 .30 1 .48 6 . 4 

0-3 .0* 4 .5 4 .01 .25 1 . 0 

Total 50 .4 67 . 1 

93 0-1 .5 12 .0 5 .54 .52 2 . 9 

1.5-5 .5 11 .5 14 .14 1 .13 16 . 0 

5.5-9 .5 11 .0 13 .35 1 .17 15 . 7 

9.5-13 .5 10 .2 11 .94 .97 11 . 5 

13.5-17 .5 8 .8 11 .15 1 .30 14 . 4 

17.5-21 .5 9 .0 11 .15 1 .68 18 . 7 

21.5-25 .5 8 .8 10 .52 2 .03 21 . 4 

25.5-29 .5 8 .0 9 .11 1 .28 11 . 1 

29.5-33 .5 6 .5 8 .48 1 .40 11 . 8 

33.5-37 .5 7 .0 8 .80 2 .26 19 . 9 

37.5-41 .5 7 .0 7 .39 1 .38 10 . 2 

41.5-45 .5 4 .8 5 .97 2 .25 13 . 4 

45.5-49 .5 4 .8 5 .81 2 .39 13 . 9 

49.5-53 .5 4 .5 5 .34 .54 2 . 9 

53.5-56 .0 4 .0 3 .14 .36 1 . 1 

0-2 .0* 6 .5 4 .01 .23 . 9 

2.0-6.0* 6 .2 7 .38 .35 2 . 6 

6.0-10.0* 5 .5 6 .28 .26 1 . 7 

10.0-14.0* 4 .5 5 .34 .10 . 5 

0-4.0** 4 .5 5 .34 1 .38 7 . 3 

Total 160 .2 198 . 0 

116 0-2 .5 10 .0 7 .66 .04 . 3 

2.5-6 .5 9 .5 11 .50 .23 2 . 6 

6.5-10 .5 8 .8 10 .56 .37 3 . 9 

10.5-14 .5 8 .0 9 .55 .67 6 . 4 

14.5-18 .5 7 .2 8 .61 .69 6 . 0 

18.5-22 .5 6 .5 7 .67 .25 2 . 0 

22.5-26 .5 5 .7 6 .72 .26 1 . 7 

26.5-30 .5 5 .0 5 .78 .24 1 . 4 

30.5-32 .0 4 .2 1 .93 .12 . 2 

Total 70 .0 24 .4 

1 Asterisks (*, * *) denote secondary axes . 

2 No axes greater than 4 cm diameter . 

182

NZ. 4 

50 100 150 

EPIPHYTE IMPORTANCE VALU E 

Figure 1 . Frequency distribution of branch systems by epiphyte importance value . EIV is an estimate of th e 

total area (dm 2 ) covered by epiphytes on the axes of the branch system . 

Table 3 .-Number of branchlets and epiphyte weights on branchlets 

for branch systems sampled from tree 1 in watershed 1 0 

Total weight of 

Branch Number of epiphytes on Mean epiphyte Total number of 

system branchlets branchlet number weight per branchlets in 

number sampled - branchlet branch system 

1 2 3 

Estimate d 

total weight of 

epiphytes on 

branchlet s 

g ----- - g g 

7 1 0.40 0.40 3 1 . 2 

33 3 .01 0 .08 0 .51 .20 13.5 ± 1 .5 1 2 . 7 

61 1 15 .99 15 .99 3 48 . 0 

93 1 4 .14 4 .14 4 16 . 6 

116 3 .14 5 .62 1 .72 2.48 12.5 ± 1 .5' 31 .0 

' Estimated . Total number of branchlets not recorded . 

18 3

1 .5 

50 


Figure 2 . Relationship between epiphyte importance value and total epiphyte biomass for the five branc h 

200 - 

• 

T 

0 10 2 0 


Figure 3 . Estimates of total epiphyte biomass on the branch systems in each EIV class . 

184

Table 4.-Biomass estimates (kg) for epiphytes on trun k 

and branch systems of tree 1 in watershed 1 0 

Branch system s 

Epiphyte Trunk Total 

Axes 

Branchlets 

BRYOPHYTES 

Hypnum circinale 1 .6 2 .6 4. 2 

Dicranum spp . 1 .5 1 .5 3. 0 

Other mosses .1 0 .1 . 2 

Scapania bolanderi .8 .1 1 . 0 

Other liverworts .5 . 6 

Bryophyte total 3 .9 4 .9 .1 8.9 

LICHEN S 

Sphaerophorus globosus .1 1 .8 1 . 9 

Other lichens with green 

algal phycobionts .2 .4 .3 . 9 

Lobaria oregana .3 3.2 3 .0 6 . 5 

Other lichens with Nostoc 

phycobionts .1 . 1 

Lichen total .6 5 .4 3 .4 9 . 4 

TOTAL 4 .5 10 .3 3 .5 18 .3 

a much higher nitrogen content than thos e 

which do not (table 5). In two of these associations, 

Lobaria oregana and Peltigera 

aphthosa (L.) Willd ., Nostoc is a secondary 

phycobiont located in cephalodia. Levels of 

nitrogen in the nitrogen-fixing lichens are 

similar to those previously reported (Pike 

1971). The nitrogen contents of the mosse s 

and nonnitrogen-fixing lichens are comparabl e 

to those reported by Rodin and Bazilevic h 

(1967) from tundra and conifer ecosystems , 

but are much lower than those reported fro m 

the agricultural Willamette Valley (Pik e 

1971) . 

Using these values of nitrogen content, th e 

biomass estimates were converted to estimate s 

of the total quantity of nitrogen in th e 

epiphytes on this one tree (table 6). Thes e 

results indicate that 65 percent of the tota l 

epiphyte nitrogen is located in lichens an d 

that 55 percent of the total is found in a 

single species, Lobaria oregana . Since one-half 

of the Lobaria oregana occurs on branchlets, 

more than 25 percent of the epiphyte nitrogen 

is found there . 

Discussion 

Our estimate of 18.3 kg of epiphyte biomass 

on the one Douglas-fir tree sampled i s 

considerably higher than the average 0 .3 kg 

per tree found in a stand of Picea engelmannii 

and Abies lasiocarpa in British Columbi a 

(Edwards et al. 1960) and the 0 .4 to 1 .3 k g 

per tree found in stands of Pinus banksiana 

and Picea mariana in Saskatchewan (Scotter 

1962) . This high epiphyte biomass is relate d 

to the tremendous size of old-growth Douglasfir. 

When wet, the epiphyte load on this tree 

is probably three to four times the dry weight 

and may be a significant factor affectin g 

branch fall. (See Barkman (1958) for wate r 

capacity of lichens and bryophytes . ) 

There are approximately 60 old-growt h 

Douglas-fir trees per hectare of forest in 

18 5

Table 5 .-Nitrogen contents of common epiphytes in water - 

shed 10. Analyses were performed on air-dry material; 

results are expressed on an ovendry-weight basis 

Epiphyte 

Percent nitroge n 

Lichens with green algal phycobionts : 

Alectoria sarmentosa 0 .49 

Hypogymnia enteromorpha .50 

Hypogymnia imshaugii .66 

Platismatia glauca .4 1 

Platismatia herrei .50 

Platismatia stenophylla .52 

Sphaerophorus globosus .4 2 

Lichens with Nostoc phycobionts: 

Lobaria oregana 1 .9 3 

Peltigera aphthose 2.84 

Pseudocyphellaria anomala 3.07 

Pseudocyphellaria anthraspis 2 .6 2 

Sticta weigelii 3 .78 

Bryophytes : 

Dicranum scoparium .87 

Hypnum circinale .95 

Isothecium spiculiferum 1 .1 0 

Table 6 .-Estimates of the total quantity of nitrogen containe d 

in the epiphytes on tree 1 in watershed 1 0 

Epiphyte 

Nitrogen 

g 

BRYOPHYTES 

Hypnum circinale 40 

Dicranum spp . 2 6 

Other bryophytes 1 7 

Bryophyte total 8 2 

LICHEN S 

Sphaerophorus globosus 8 

Other lichens with green algal phycobionts 4 

Lobaria oregana 12 7 

Other lichens with Nostoc phycobionts 4 

Lichen total 14 3 

TOTAL 225 

186

watershed 10 (C . T. Dyrness, personal communication). 

Our 18.3 kg of epiphytes, per 

tree would then correspond to 1 .1 metric tons 

per hectare, a figure well within the range of 

values reported from northern conifer forests 

(Edwards et al . 1960, Scotter 1962, Rodin 

and Bazilevich 1967). Our estimate is only for 

the overstory Douglas-fir trees, however, an d 

does not take into account the considerabl e 

biomass of epiphytes that is located on understory 

trees and shrubs . 

For comparison, our estimate of the tota l 

dry weight of needle biomass on the tree 

sampled is 84 kg, a figure 4 .6 times as high as 

the estimate of epiphyte biomass . 

Our finding, that a large proportion of 

epiphyte biomass is present on small branch - 

lets, indicates the importance of adequate 

sampling of this part of the tree and demonstrates 

the importance of nondestructiv e 

sampling because the branchlets are particularly 

likely to be destroyed when a large, old - 

growth Douglas-fir tree is felled . Since th e 

contribution from the branchlets is significant, 

EIV should be modified to include a 

component from the branchlets to avoi d 

underestimating the portion of epiphyte biomass 

located on small branch systems . 

Our methodology does not include estimates 

of biomass of crustose lichens or free - 

living algae. We have observed, even on twig s 

smaller than 1 cm in diameter, that cover of 

crustose lichens is regularly greater than 5 0 

percent of the total surface area of the twigs . 

Our estimates must be considered underestimates 

since they include only bryophytes an d 

foliose and fruticose lichens . 


We would like to acknowledge the assistance 

of Fred Rhoades, Jane McCauley, and 

Tom Denison in the field sampling and o f 

Mary Vreeland, Karen Barrett, and Kare n 

Berliner in sorting and weighing the epiphytes . 

This project began (June-August 1970) as a 

student project supported by the National 

Science Foundation Undergraduate Researc h 

Participation Program (Grant No. GY-7641 ) 

to the Department of Botany, Oregon Stat e 

University . Subsequent work was supporte d 

by National Science Foundation Grant No . 


U .S. Analysis of Ecosystems, International 




Barkman, J. J. 1958 . Phytosociology and 

ecology of cryptogamic epiphytes . 628 p . , 

illus. Assen, Netherlands : Van Gorcum. 

Denison, William C., Diane M. Tracy , 

Frederick M. Rhoades, and Martha Sherwood 

. 1972 . Direct, nondestructiv e 

measurement of biomass and structure i n 

living old-growth Douglas-fir . In Jerry F . 


Waring (eds.), Proceedings-research on 


p. 147-158, illus. Pac. Northwest 

Forest & Range Exp . Stn., Portland, Oreg. 

Edwards, R. Y., J. Soos, and R . W. Ritcey . 

1960 . Quantitative observations o n 

epidendric lichens used as food by caribou . 

Ecology 41 : 425-431 . 

Pike, L . H. 1971. The role of epiphytic 

lichens and mosses in production and nutri - 

ent cycling of an oak forest. 172 p ., illus . 

Ph.D. thesis, on file at Univ. Oreg ., Eugene . 

Rodin, L. E., and N . I. Bazilevich . 1967 . Production 

and mineral cycling in terrestrial 

vegetation . [Trans. from Russian .] 288 p . , 

illus. Edinburgh : Oliver and Boyd . 

Scotter, G . W. 1962. Productivity of arborea l 

lichens and their possible importance t o 

barren-ground caribou (Rangifer arcticus) . 

Arch . Soc . Bot. Fenn . " Vanamo" 16 : 

155-161 . 

187



Litter, foliage, branch, and ste m 

production in contrasting lodgepole 

pine habitats of the Colorado Front Rang e 

William H . Moir ' 

Colorado State University 

A bstrac t 

Harvest data are presented from a 70-year-old naturally thinned stand of lodgepole pine (Pins contorta 

Dougl.) in the subalpine P . contorta/Vaccinium myrtillus habitat and from five stands of similar-aged lodgepole 

pine in the drier P . contorta/Geranium fremontii habitat. These latter stands include both naturally thinned and 

artificially uniform-thinned treatments . Stand structure, natural mortality, biomass components, growt h 

increments of stem wood, branches, and foliage, and net primary production of these materials are given in 

tables. Litter production over a 4-year period is also reported for each stand ; sources of variation in litte r 

production are discussed. 

Introduction 

Forests of lodgepole pine (Pinus contorta 

Dougl.) are extensive in the central and northern 

Rocky Mountains and portions of the 

Cascade Range . Studies on net primary production 

are lacking, however, in virtually al l 

habitats in which this pine is the major re - 

source. Biomass studies in certain lodgepole 

pine habitats in Alberta have been publishe d 

(Kiil 1968, Johnstone 1971), and in Colorado, 

the biomass of forest floor humus and 

pine foliage has been reported (Moir and Grier 

1969, Moir and Francis 1972) . These studies 

were not extended to net primary productivity 

. Our rather extensive knowledge of net 

stem wood production in the central Rocky 

Mountains (Myers 1967) does not reveal th e 

relationship between productivity and th e 

lodgepole pine habitats or include branch an d 

foliage production. I report below, therefore , 

results of a 4-year study in Colorado concerning 

aspects of net primary productivity i n 

mostly 70-year-old natural stands of lodgepole 

pine . 

Methods 

Six stands on the east slopes of the Front 

Range in Boulder County, Colorado, wer e 

studied . Stand LH2 occurs in the subalpine P . 

contorta/Vaccinium myrtillus habitat; the 

others occur in the drier, montane P . 

contorta/Geranium fremontii habitat. Habitat 

features and pine population statistics ar e 

given by Moir (1969), Moir and Francis 

(1972), and table 1 . In each stand a rectangular 

study plot was established in an area 

where the forest structure appeared homogeneous. 

Plot size varied with stem density 

(table 1). Annual litter fall was periodicall y 

measured from randomized 0 .25 m 2 micro - 

plots permanently located within each plo t 

(Moir and Grier 1969) . All surface organi c 

material in the microplots was collected a t 

4-month intervals in 1969, and in late August, 

the remaining years . Mineral contamination a t 

collection was minimized by computing litter 

'Present address: Box 31, Rodeo, New Mexico 

88056 . 

189

Table 1 .-Some population characteristics of Pinus con torta in stand s 

on the east slopes of the Colorado Front Range 

Stand 

and 

year 

Age' 

Plot 

area 

Number of stems by d.b.h. classes 

0 .2 .5 2 .5-5 5-7 .5 7 .5-10 10.15 15-20 20-2 5 

Density 

Mortalit y 

Years m 2 cm Stems/ha Stems/ha x yr 

2 .2 71 8 4 

1966 130 136 27 2 35,700 

1970 82 135 25 4 29,300 1,60 0 

4 .1 72 12 6 

1966 3 35 56 38 5 10,900 

1970 1 24 49 39 8 9,600 32 5 

1 .1 72 12 6 

1966 1 26 46 30 10 9,000 

1970 24 35 36 12 1 8,600 100 

1 .3 103 15 0 

1966 2 7 5 25 8 3,10 0 

1970 5 5 23 12 3,000 2 5 

4 .3 71 37 5 

1966 29 1 1 12 45 2 1,600 2 

1970 25 2 7 45 8 1,650 0 

LH2 77 15 0 

1966 2 4 10 18 19 4 3,80 0 

1970 2 4 9 18 20 4 3,800 0 

' Reference year 1970 . Stand age is based upon the oldest trees . 

2 This stand was thinned in 1934-35 . Density refers to stems over 2 .5 cm d .b .h . 

fall as loss-on-ignition (Moir and Grier 1969) , 

and the tendency of debris of the forest floo r 

to "drift" into the microplots was offset b y 

slightly increasing the effective area of th e 

microplot during computation . 

In each of plots 1 .1, 4.3, and LH2 of table 

1, the stems of lodgepole pine were placed 

into five diameter (d .b.h.) classes, with an 

equal number of stems in each class. In September 

1969, the tree at the midpoint of eac h 

d .b .h. class was harvested (Ovington, Forrest , 

and Armstrong 1967), giving five trees fro m 

each plot. Current and second year foliage 

and twigs were separated from each other an d 

from older material. Branches with any livin g 

foliage whatever were classified as living an d 

separated from dead branches at harvest time . 

From each stem replicate disks were obtained 

at the stem base, at 3 to 4 dm ., and just belo w 

the lowest living branch (crown base) . From 

each disk I measured diameter inside bark , 

diameter outside bark (d .o .b.), wood density 

(volume determined by water immersion) , 

bark density at 102° C., and average radial 

growth during each of growing seasons 196 8 

and 1969. In addition, measurements of 

19 0

d.o .b. were made at periodic lengths along th e 

felled stems. Since this harvest procedure i s 

not effective- for estimating cone productio n 

(Ovington, Forrest, and Armstrong 1967 , 

Lotan and Jensen 1970), this is not include d 

in the study . Biomass values are reported at 

700 C. ovendry weights except for bole materials 

(stem wood and bark) which are reporte d 

at 102° C. ovendry weights . 

Stem biomass was calculated by the Smalian 

equation, multiplying each volume segment 

by the appropriate wood density observed 

from the disks . Stem biomass increments wer e 

computed as the difference between Smalian 

biomass for the years 1969 and 1968 . Similarly, 

bark biomass for each bole was deter - 

mined by the Smalian difference betwee n 

diameters inside and outside bark, using bark 

densities for each volume segment . The increment 

of living branches was calculated using 

estimative ratios (Whittaker and Woodwel l 

1968), assuming that relative branch increments 

are approximated by relative ste m 

wood increments . Branch mortality was computed 

by a canopy displacement model 

(Madgwick 1968). The computation assume s 

a canopy steady state in which the upward 

displacement of the envelope of green foliag e 

within a stand leaves a wake of dead branche s 

that were alive the previous year . The average 

displacement of foliage was further assume d 

to be the mean height growth of each stand . 

Stem mortality was computed from the mortalities 

of table 1, in which the biomass o f 

dead stems during the mortality period wa s 

determined from the basic allometric relation - 

ship between stem biomass and d .b.h . 

Calculations were based on the equation s 

below : 

Biomass and growth increments . 

H = total stem length (ft ) 

T = stem length to crown 

(lowest living whorl ) 

d i , d2 , d 3 

= diameter inside bark at ground 

level, 1 ft, and lowest living 

whorl respectively (inch) . 

P 1, P2, P3 = wood densities at ground level , 

1 ft, and lowest living whorl 

(g/cc) . 

k = conversion constant t o 

metric = 0 .0772 2 

Ar' 1 , Ore, Ara = mean radial growth (inch) at 

ground, 1 ft, and crown base. 

Ws = biomass (kg) at 102° C 

II. Living branches. 

AWb = Wb ~ s 

Ws 

+ new twig biomass 

Branch mortality (per tree) = Wb(AH ) 

T ' 

where AH . is average height growth for 

stand, T' is stem length within the crown 

over which branch mortality is observed , 

and Wb is the live branch biomass along 

stem portion T' , Wb is total live branc h 

biomass of the crown (along stem lengt h 

H-T) . 

III. Basic allometric equation . 

W = k 1 D k 2 , for biomass W and DBH of D , 

and regression constants , 

k 1 and k 2 . 

In the case of plot 4 .3 where a clear trend of 

stem wood increment with d .b.h . exists, the 

plot increment of stem wood was computed 

by linear regression over the independent 

variable, d .b.h . 

I. Stem . 

Ws = k[(di + di )P1 + (di + 43) (T - 1)p2 

+(d3 + .75 2 ) (H-T-3)ps ] 

AWs = W s - Ws- 1 , where Ws- 1 is obtaine d 

by substituting 

di - 2Ori for di (i = 1,2,3) . 

Litter Productio n 

Results 

Tables 2 and 3 give annual litter productio n 

in the six stands over the 4-year period . The 

mean annual litter production was 0 .4 6 

191

Table 2 .-Annual litter fall in Colorado stands of Pinus contorta . 

Values are given as loss on ignition weights . 

Year 

Stand 

2 .2 4 .1 1 .1 1.3 4 .3 LH2 Mean 

T otal 

Number 

of samples 

1967 0 .46 0 .47 17 1 

1968 0.37 0 .47 .36 0 .63 0 .30 0 .38 .40 5 2 

1969 .41 .56 .50 .88 .52 .69 .59 5 2 

1970 .25 .39 .37 .45 .30 .54 .37 52 

Mean .34 .47 .42 .65 .37 .52 .46 

1 A fire in our drying oven caused the loss of many samples . Only the samples from stand 1 .1 entirel y 

escaped the fire and can be used for comparison with other years. 

Table 3 .-Constituents of the 1967-68 annual litter fall, as los s 

on ignition in Pinus contorta stands in Colorado 

Stand 

Numbe r 

of samples 

Needles Branches 9 Cones d Cones Bark 

% kg/m 2 % kg/m 2 % kg/m 2 % kg/m 2 % kg/m 2 

2 .2 7 66 0 .24(18) 1 18 0.07(65) 12 0.05(59) 1 0 .00(58) 2 3 0 .01(20 ) 

4 .1 7 64 .30(10) 13 .06(105) 18 .08(98) 1 .00(90) 5 .02(28 ) 

1 .1 7 68 .24(11) 10 .04(29) 15 .05(52) 1 .00(45) 6 .02(36 ) 

1 .3 8 54 .34(28) 6 .04(60) 34 .21(103) 2 .01(44) 5 .03(59 ) 

4 .3 15 83 .25(17) 2 .01(105) 9 .03(72) 3 .01(35) 2 .01(70 ) 

LH2 8 72 .26(14) 10 .04(55) 10 .04(94) 4 .02(21) 4 .01(56 ) 

Mean 52 68 .27 10 .04 16 .07 2 .01 4 .0 2 

1 

Coefficient of variation in parentheses . 

2 Trace quantities are less than 0 .005 kg/m2 . Rounding errors may cause the percentages for each stand to 

total around 100 . Only trace quantities of nonpine debris are found in each stand . 

192

kg./m 2 . Both tables show, however, that 

numerous sources of variation in litter fal l 

occur in lodgepole pine forests. Variatio n 

within plots is evident (table 3) . Generally 

needle fall is less variable from microplot to 

microplot (CV from 10 to 18 percent except 

in low density stand 1 .3) than other components 

of the annual litter production . 

Branches and female cones exhibit very high 

spatial variability, and the latter is furthe r 

enhanced (beyond the values of table 3 ) 

through activities of the common pine squirrel , 

Tam ias c i u r u s h udsonicus . Generally th e 

percentages of needles, branches, etc . are 

rather uniform within and between stands , 

but marked departures from mean percent - 

ages are found in the low density stands 1 . 3 

and 4.3 . These departures probably resul t 

from the lessened sloughing of branches a t 

lower portions of the bole . Thus, branches 

comprised only 2 percent of the annual litte r 

fall in stand 4 .3, but 18 percent in high 

density stand 2 .2. The very high cone accumulation 

in stand 1 .3 is partially explained by 

fire thinning around 1900 which remove d 

many smaller, noncone-bearing trees, and 

partly by the chance distribution of collectio n 

sites near prominent cone-bearing trees . The 

mean annual litter fall at the bottom of tabl e 

2 indicates that stands with 'least canopy mas s 

(2 .2 and 4 .3) have lowest litter fall, and thos e 

with greatest canopy mass (1 .3 and LH2) have 

the highest litter fall . This relationship is no t 

clear, however, when only a single year ' s 

annual litter fall is considered . 

In all stands, 1969 was the maximum litter 

fall year. Nevertheless, year-to-year differences 

are not consistent from stand to stand . 

Thus, 1970 was the year of least litter production 

in stand 2 .2 but not LH2 . Seasonal variations 

are also significant. The inclusive perio d 

from early June to late August gave 38 per - 

cent of the total annual litter production i n 

1969, but only about 26 percent of the total 

from the same stands the following year . 

While some materials such as branches may b e 

shed somewhat uniformly through the year , 

others such as needles showed weak seasonal 

peaks from June through October . In only 

one year was there a conspicuous "peak sea - 

son" of needle fall in late August and September; 

in general, each stand was shedding 

needles continuously throughout the year . 

Biomass and Net Growth Increment s 

The live components of the abovegroun d 

pine crop are given in table 4. From the data 

the coefficients, k l and k 2 , of allometric 

equations of the form, W = k l D k 2 , were computed 

by regression (W is biomass in kg and D 

is d .b.h. in cm from columns of table 4) . 

These allometric equations and the entire 

population of d .b.h. values in each of plots 

4.3, 1 .1, and LH2 were used to compute th e 

stand biomass values of table 5 . This biomas s 

and the mean increments of table 4 were use d 

to compute the "new growth" values of tabl e 

6. The mean values of table 2 and the appropriate 

percentages of litter constituents i n 

table 3 were multiplied to give the litter fal l 

results of table 6. Stand LH2 in the P. 

contorta/V. myrtillus habitat is clearly th e 

most productive . In this stand the proportion 

of green needles to living shoot parts is very 

high in average to dominant tree sizes, as 

shown in the last column of table 4 . By contrast 

stand 1 .1 in the P. contorta/G. fremonti i 

habitat (Moir 1969) has in general a lowe r 

ratio of green needles to living shoot parts i n 

average to dominant trees . Current year 

fascicles comprise a greater end-of-season proportion 

of the green canopy (25 to 31 per - 

cent) in stand 1 .1 than in stand LH2 (17 to 

27 percent) . Most trees in the latter stand retain 

fascicles for several years longer than 

stand 1 .1 . Stand 4 .3 occurs in the same habitat 

as stand 1 .1 but was thinned in 1934-3 5 

to about 2 .5 x 2.5 m stem spacing . The consequence 

was to reduce light competition an d 

encourage lateral branch growth . Surviving 

stems retained a high proportion of needle s 

(11 to 17 percent of the shoot weight), an d 

abscission would not occur until about th e 

5th year. By 1969 this stand had age 

reached a "closed" canopy condition ; exces s 

needles were abscissing as suggested by th e 

high needle percentage of table 5 and by th e 

negative net needle production in table 6 . 

Further discussion of foliage characteristics of 

these stands is given by Moir and Franci s 

(1972) . 

193

Table 4. Biomass and growth increments from harvested Pinus contorta trees 

Stand and tree D .b .h . 

Total 

shoot 

Bole 2 

Stem wood Living branches Green needles 

Ws 

p /Ws Wb WWb Wn L\/Wn Wn/S 3 

cm • kg kg % kg % kg %+ % 

LH2 5 19 .3 145 111 .2 103 .1 2.2 12 .4 4 .9 12 .3 22 8 . 9 

4 16 .5 91 68 .0 63 .1 1 .4 7 .8 3 .3 5 .8 18 7 . 1 

3 12 .4 49 42 .0 38 .8 2.2 2 .7 4 .8 2 .7 17 5 . 7 

2 10 .0 27 23 .4 21.0 1.5 1 .1 3 .6 1 .0 27 3 .9 

1 7 .6 11 9 .2 8 .5 .7 .4 3 .9 

Mean 1 .8 4 .2 2 1 

1 .1 5 10 .0 23 17 .9 16.1 1 .9 1 .4 5 .5 1 .4 25 6 . 3 

4 7 .6 13 9 .6 8.6 1 .3 .8 4 .5 .8 30 6 .7 

3 7 .6 13 10 .8 9 .8 2.2 .7 5 .7 .8 28 6 .4 

2 5 .1 4 3 .5 3 .1 .8 .13 4 .6 .16 31 4 . 1 

1 3 .6 2 1 .8 1 .6 .05 .02 1 . 1 

Mean 1 .6 5 .1 2 8 

4 .3 5 14 .0 44 24 .2 22 .5 1 .7 6 .5 5 .1 4 .9 25 12 .6 

4 12 .7 30 19 .3 18 .1 1 .8 4 .2 4 .3 3 .0 24 10 .6 

3 11 .7 25 17 .4 15 .8 1 .9 3 .2 5 .3 2 .6 27 10 . 8 

2 10 .7 18 10 .7 9 .7 2 .2 2 .5 4 .8 2 .6 20 15 .5 

1 8 .9 15 9 .0 8.6 2 .3 2 .0 7 .5 2 .4 26 17 . 1 

Mean 5 .4 24 

Includes bole, living and dead branches, cones, and needles . 

2 Stem wood plus stem bark . 

3s refers to total shoot less dead branches . 

Table 6 suggests that organic production 

and turnover in the three stands is approximately 

steady (Kira and Shidei 1967) . The 

only meaningful net annual accumulation is 

the stem wood component (and the unmeasured 

root wood). Of course, my branch mortality 

model assumed a steady state condition 

for these stands, but the assumption is not 

disturbed in table 6 where new growth and 

branch mortality are balanced by the observed 

contribution of branches and needle s 

to the litter fall . 

The inherent photosynthetic capacity o f 

stand LH2 is high by virtue of the great lea f 

biomass and surface display of current-year 

foliage (Moir and Francis 1972) . The tw o 

stands of the drier habitat have considerably 

less foliage. Stand 4.3 has only half the increment 

of current-year foliage as stand 1 . 1 

(table 6). Since current-year foliage has 

greater photosynthetic capacity than older 

pine foliage (e .g., Larson and Gordon 1969) , 

this difference may account for the reduce d 

amount of net stem wood production in stand 

194

Table 5 .Pinus contorta stand biomass, 1970, from harvest s 

on the east slopes of the Colorado Front Rang e 

Compartment 

Stan d 

4 .3 1 .1 LH 2 

kg/m 2 96 kg/m 2 kg/m 2 96 

Live pine crop : 

Bole' 3 .24 46 9 .44 57 21 .68 5 8 

Live branches .75 11 .83 5 2 .11 6 

Roots 1 .782 25 2 4 .28 26 9 .68 2 6 

Green needles .50 7 .84 5 1 .74 5 

Cones .35 5 .57 3 .24 1 

Dead branches .39 6 .57 4 1 .69 4 

Total 7 .01 100 16 .53 100 37 .14 10 0 

Standing dead timber 3 0 1 .06 .1 5 

Forest floor humus4 2 .85 3 .16 3 .8 6 

Ground flora s .01 0 .005 

Total 9 .87 20 .75 41 .18 

' Stem wood plus bark . 

2 Biomass of roots with diameters over about 5 mm were computed on the basis of nine excavated roo t 

systems . Percentages and totals in stands 1 .1 and LH2 assume root biomass to be about 25 percent of the liv e 

pine crop (Rodin and Bazilevich 1967) . 

3 Above ground parts only (bole plus branches) . 

4 After Moir and Grier 1969 . 

5 Shoot parts only, from harvests in August 1967 . 

4.3 despite both stands having identical leaf 

area indices of 4 .5 m 2 /m 2 (Moir and Francis 

1972) . 

Conclusions 

Pine stands in the Lodgepole Pine Zone of 

Colorado (Moir 1969) are not very productive 

per unit of current-year leaf biomass . It is instructive 

to compare the net primary productivity 

of stand LH2 (the most productive in 

this study) with pine stands having simila r 

steady state quantities of annual foliage production. 

Unfortunately, there are very few 

published studies of net primary productivit y 

for older natural stands of Pinus (Art and 

Marks 1971, Kira and Shidei 1967) . However , 

several studies within young plantations o r 

natural stands do permit comparison, as give n 

in table 7 . It is clear that annual stem an d 

branch production in the P. contorta stand i s 

well below that of the three younger pine 

types. Numerous studies within developin g 

pine plantations have led workers to conclude 

that within a few years after canop y 

closure gross productivity levels off and i n 

following years net production declines a s 

stand respiration continues to increase. Forrest 

(1969) concluded that the weight of 

foliage in developing P. radiata plantation s 

(the oldest of his series of stands is given in 

table 7) is probably constant after about 1 2 

years. However, I am not convinced on th e 

19 5

Table 6.-Annual turnover from certain shoot component s 

of Pinus contorta stands in Colorado ' 

Component 

Stand 

4 .3 I 1 .1 LH 2 

------------------ kg/m2 

Bol e 

New growth 0 .06 0 .15 0.3 9 

Stem mortality 0 - .02 0 

Net .06 .13 .3 9 

Branches 

New growth .04 .04 .0 8 

Branch mortality -.03 - .01 - .0 3 

Litter fall - .01 - .04 - .0 5 

Net 0 - .01 0 

Needles 

New growth .12 .24 .3 7 

Litter fall - .32 - .29 - .3 8 

Net - .20 - .05 - .01 

' Litter fall is determined by the equation : 

Li = fiL, where L is the 3- or 4-year mean (table 2) and fi is the fractional 

component of needles or branches from table 3 . 

Table 7.-Some comparisons of annual increments in shoo t 

components of contrasting pine stand s 

Stand Age Density 

Annual production 

Foliage Stem Branches 

References 

yr sterns/ha metric tons/ha 

P. contorta 77 3,800 3 .7 3.9 0 .8 This study 

P. radiata 12 1,560 3 .1-3 .3 10 .8 2 .9 Forrest 196 9 

P. virginiana 17 5,750 4 .3 5 .8 3 .6 Madgwick 196 8 

P. densiflora 33 2,340 3 .4 8.7 2 .1 Hatiya et al . 196 5 

196

asis of the structure of the different pin e 

crops (tables 1 and 5 and comparable descriptions 

from references in table 7) that the 

three lodgepole pine stands of this study carry 

appreciably greater burdens of respiratory 

biomass (Yoda et al. 1965). The major limitation 

of pine productivity in the Central and 

Northern Rocky Mountains may be the generally 

infertile soils and inimical continental 

climates. Under these environmental conditions 

not even the thinning treatment (stand 

4.3) and subsequent surge of productivity i n 

surviving stems gave increased yield on an are a 

basis (table 6) . 

Management decisions in the Colorad o 

Lodgepole Pine Zone must recognize th e 

inherent low productivity of P. contorta . 

Where wood production must continue t o 

have management priority, efficiency o f 

utilization can be increased at least 10 percent 

if live branches can be economically include d 

in the harvest (table 5) (Young 1968) . Management 

can also be directed toward uneven - 

aged stands for purposes of esthetics, enhanced 

groundcover production, or better 

game utilization. Small, irregularly shaped 

clearcut areas are also an attractive management 

possibility for long-term planning in this 

low productivity region . This study suggests 

strongly that because of its very low productivity, 

especially when compared against th e 

productivity of intensively managed pin e 

plantations, lodgepole pine stands in environments 

of site index of 90 or less should b e 

regarded as only minor resources of woo d 

harvest. A greater percentage of such land i n 

the lodgepole pine region of North Americ a 

might be devoted to recreational usage . 

A much neglected management tool in 

lodgepole pine forests is prescribed burning . 

Periodic fires were an important feature of 

lodgepole forests (Moir 1969) . This study and 

others (Kiil 1968) reveal that high quantities 

of slowly decomposable materials build up 

within natural stands . Fairly high quantities 

of nutrients become locked up in the forest 

floor humus (Moir and Grier 1969) ; this 

humus together with the shaded condition of 

the forest floor in closed stands has adverse 

effect upon the ground flora (Basile and 

Jensen 1971, Moir 1966). Controlled fire has 

at least four possible beneficial effects in 

lodgepole pine forests : (1) The reduction of 

fuel and wildfire probability, (2) stimulatio n 

of ground vegetation, (3) a nutrient "pulse" 

effect stimulating tree production, (4) tree 

thinning and conversion to uneven-age d 

stands. The possible use of controlled surfac e 

fires in P. contorta should be given 

considerably more attention . 


Work reported in this paper was supporte d 

by the National Science Foundation, Grant No . 

B020357 in cooperation with the Coniferou s 




Biome . 


Art, H. W., and P. L. Marks. 1971. A summary 

table of biomass and net annual primary 

production in forest ecosystems o f 

the world . In Forest biomass studies, 15th 

IUFRO Congr., p. 3-32 . Orono : Univ. 

Maine Press . 

Basile, J. V., and C. E. Jensen . 1971 . Grazing 

potential on lodgepole pine clearcuts i n 

Montana. USDA Forest Serv . Res. Pap. 

INT-98, 11 p . Intermountain Forest & 

Range Exp . Stn ., Ogden, Utah . 

Forrest, W. G. 1969. Variations in the accumulation, 

distribution and movement of 

mineral materials in radiata pine plantations. 

276 p. Ph .D. thesis on file at 

Australian Nat . Univ., Canberra . 

Hatiya, K., K. Doi, and R. Kobayashi. 1965 . 

Analysis of the growth in Japanese red pin e 

(Pinus densiflora) stands. A report on th e 

matured plantation in Iwate Prefecture . 

Govt. For. Exp. Stn. Bull. (Tokyo) 176 : 

75-88 . 

Johnstone, W . D. 1971 . Total standing cro p 

and tree component distributions in thre e 

stands of 100-year-old lodgepole pine . In 

Forest biomass studies, 15th IUFR O 

Congr., p. 81-89 . Orono : Univ. Maine Press . 

197

Kiil, A. D. 1968. Weight of the fuel complex 

in 70-year-old lodgepole pine stands of different 

densities . Can. Dep. For. & Rural 

Dev., For. Br. Dep . Publ . 1228, 9 p . 

Kira, T., and T. Shidei . 1967 . Primary production 

and turnover of organic matter in different 

forest ecosystems of the wester n 

Pacific . Jap . J. Ecol . 17: 70-87 . 

Larson, P. R., and J . C. Gordon . 1969 . Photo - 

synthesis and wood yield . Agric. Sci. Rev . 

7 : 7-14 . 

Lotan, J . E., and C. E. Jensen . 1970. Estimating 

seed stored in serotinous cones of 

lodgepole pine. USDA Forest Serv . Res . 

Pap. INT-83, 10 p . Intermountain Forest & 

Range Exp . Stn ., Ogden, Utah . 

Madgwick, H. A. I. 1968. Seasonal changes in 

biomass and annual production of an old - 

field Pinus virginiana stand. Ecology 49 : 

149-152 . 

Moir, W . H . 1966. Influence of ponderosa 

pine on herbaceous vegetation . Ecology 47 : 

1045-1048 . 

. 1969. The lodgepole pine zon e 

in Colorado. Am. Midland Nat. 81 : 87-98 . 

and R . Francis . 1972. Foliage 

biomass and surface area in three Pinus 

contorta plots in Colorado . Forest Sci. 18 : 

41-45, illus . 

and H. Grier . 1969. Weight an d 

nitrogen, phosphorus, potassium, and cal - 

cium content of forest floor humus o f 

lodgepole pine stands in Colorado . Soil . 

Sci. Soc. Am. Proc . 33 : 137-140 . 

Myers, C. A. 1967. Yield tables for manage d 

stands of lodgepole pine in Colorado an d 

Wyoming. USDA Forest Serv . Res. Pap . 

RM-26, 20 p . Rocky Mountain Forest & 

Range Exp . Stn ., Fort Collins, Colo . 

Ovington, J. D., W . G . Forrest, and J . S. Armstrong. 

1967 . Tree biomass estimation . In 

Primary productivity and mineral cycling i n 

natural ecosystems-symp . Am. Assoc . 

Advan. Sci. 13th Annu . Meet . & Ecol . Soc . 

Am. (New York) Proc., p. 4-31 . Orono : 

Univ . Maine Press . 

Rodin, L . E., and N . I. Bazilevich . 1967 . [Production 

and mineral cycling in terrestrial 

vegetation.] 288 p. London : Oliver & 

Boyd . (English translation .) 

Whittaker, R . H ., and G. M. Woodwell . 1968 . 

Dimension and production relations of 

trees and shrubs in the Brookhaven Forest , 

New York. J . Ecol . 56: 1-25 . 

Y o d a, K., K. Shinozaki, H. Ogawa, K . 

Hozumi, and T . Kira . 1965. Estimation of 

the total amount of respiration in wood y 

organs of trees and forest communities . J . 

Biol . (Osaka) 16 : 15-26 . 

Young, H. E. 1968 . Challenge of complete 

tree utilization . Forest Prod. J. 18 : 83-86 . 

198



Small mammal and bird populations 

on Thompson site, Cedar River : 

parameters for modeling 

Sterling Miller, Curtis W . Erickson, 

Richard D . Taber, and Carl H . Nellis 

College of Forest Resource s 



Abstract 

Preliminary estimates of small mammal and bird populations on the Thompson site in the Cedar Rive r 

watershed were made from 1969 to 1971 . Mammal populations were estimated through a kill-trap grid, and bird 

populations through systematic direct observations . The small mammal fauna, notably rich in insectivorou s 

forms, has as its most abundant members Trowbridge shrew (Sorex trowbridgei), vagrant shrew (Sorex vagrans) , 

shrew-mole (Neurotrichus gibbsi), Oregon vole (Microtus oregoni), and deer mouse (Peromyscus maniculatus) . 

Bird populations differed between summer and winter. In the summer the most abundant species wer e 

Swainson's thrush (Hylocichla ustulata), winter wren (Troglodytes troglodytes), Oregon junco (Junco oreganus) , 

black-throated gray warbler (Dendroica nigrescens), chestnut-backed chickadee (Parus rufescens), brown-heade d 

cowbird (Molothrus ater), and MacGillivray's warbler (Oporornis tolmiei) . In winter only two birds were 

common: winter wren and golden-crowned kinglet (Regulus satrapa). For these species estimates were obtaine d 

for abundance, biomass and, for the mammals, energy flow . 

Introduction 

This paper represents a first step in providin g 

the information on terrestrial vertebrate s 

which will be needed in the development o f 

ecosystem models for the Western Coniferou s 

Forest Biome . It covers the initial investigations 

on small mammal and bird populations , 

which have centered on the Thompson site of 

the Cedar River watershed, King County, Wash - 

ington . Consideration is given to methodology , 

results in terms of species populations, ecological 

role, biomass and energy, and the desirable 

directions for further investigations . 

The Thompson site consists of a second - 

growth Douglas-fir (Pseudotsuga menziesii) 

forest about 60-70 years old, on Barnesto n 

soils formed from glacial outwashes . Elevations 

vary from 210-310 meters . There are 

small variations in ecological conditions with - 

in the forest, as indicated by differences i n 

the understory vegetation . 

Every biotic community is composed o f 

individual species, and these must be identified 

and studied individually in building u p 

knowledge concerning the whole . Schwarz 

(1967, p. 225-226) outlines our task and it s 

difficulties as follows : 

In order to determine the role of the species 

in the energy metabolism of the ecosystem, 

we must, first of all, determin e 

with the necessary reliability the following 

parameters of the population : th e 

absolute numbers of animals ; their biomass; 

the structure of the populatio n 

(because the intensity of the matter an d 

energy metabolism of animals of different 

weight, sex, age, physiologica l 

199

condition differ) ; the turnover rate o f 

the population, and the intensity o f 

metabolism of different groups of animals 

. . . Whilst it is not difficult to estimate 

the intensity of metabolism of certain 

species in a calorimetric chamber, i t 

is practically impossible to estimate the 

loss of energy by a bat, a mole, or a 

dolphin in the process of their natura l 

life activity . 

In the light of these considerations we wil l 

examine our preliminary data and, through a 

comparison of actual and needful information, 

delineate the work still to be done . 

Small Mammal Studies 

Since small mammals are generally inconspicuous, 

and do not make readily detectable 

signs, the determination of species presence 

and especially population density is widely 

recognized as a difficult problem . 

The most commonly used method fo r 

estimating population densities in smal l 

mammals is to capture, mark, and release a 

sample, and then capture a second sampl e 

containing both marked and unmarked individuals 

. Several underlying assumption s 

(Leslie 1952) are : all individuals have equa l 

probabilities of capture, released individual s 

disperse randomly into the populations, an d 

there is no mortality, natality, or significan t 

ingress or egress during the experimenta l 

period . 

Since these assumptions were not war - 

ranted in our case, and since we needed to 

collect specimens for data on weight, foo d 

habits, and reproduction, we considered intensive 

removal methods . Grodzinski et al . 

(1966) proposed such a method for IBP smal l 

mammal studies. They suggested intensive 

kill-trapping following a prebaiting period on 

a defined grid of locations . This provided data 

for an estimation procedure in which a regression 

line, plotted for the number of animals 

caught each day (on the ordinate axis) against 

the cumulative number previously caught , 

would provide an estimate of populatio n 

density (the point where the regression line 

intersected the abscissa) . Earlier work on this 

approach was that of DeLury (1947), Hayn e 

(1949), and Zippin (1958) . 

The major flaw in the method of Grodzinski 

et al. (1966) is that it gives no accurat e 

estimate of the area sampled by the grid, be - 

cause the small mammals move some unknown 

distance to the trapping point . Sinc e 

the area sampled is not defined, populatio n 

densities cannot be determined . 

Adamczyk and Ryszkowski (1968) suggested 

that the sample grid be surrounded o n 

each side by an external belt of trapping locations 

to catch animals moving toward th e 

inner grid before they could reach it, thereb y 

controlling the "periphery effect ." 

The data provided by the external trappin g 

belt cannot be used, because the distance 

travelled by each animal before being caught 

is highly variable (Miller 1970) . 

Adamczyk and Ryszkowski (1968) recommend 

a 5-day prebaiting period to accusto m 

the mammals to the trap locations, followe d 

by a 5-day period of removal trapping . Their 

basic assumptions are : (1) all residents of the 

inner grid are captured, (2) prebaiting does 

not increase the resident density, and (3) individuals 

not resident on the inner grid are no t 

captured on the inner grid ; immigrants and 

individuals resident in the outer grid will b e 

captured, if at all, in the outer grid . 

Of these three assumptions, we could tes t 

only the second . Adopting a 5-day prebaiting 

period, we ran pairs of trap-lines-one pre - 

baited and one control-in each case . Total 

catches over 5 days were not different, 

though prebaited lines took a higher proportion 

of the catch on the first day (Mille r 

1970) . This finding corroborates that o f 

Babinska and Bock (1969), who showed tha t 

prebaiting did not significantly increase th e 

density of resident small mammals . 

Our trapping procedure was as follows : a 

5 .94-ha (12 .4-acre) sample area was divide d 

into 256 stations, 16 rows by 16 lines, wit h 

15 .2-meter spacing . Anchored at each statio n 

was a paper plate which was baited with oatmeal, 

millet, sunflower seeds, oats, and wheat 

for 5 days . Then the remaining bait was removed 

and two mouse traps were placed on 

each plate, one trap baited with a peanut 

butter-bacon grease mixture, and the other 

200

with a birdseed-Crisco mixture . Traps were 

run for 5 full days . 

Two such grids were established to sampl e 

the range of variability within the Thompso n 

site, and each was trapped twice, once in summer 

and once in winter, as follows : Grid I : 

July 22 to 27, 1969, and March 22 to 27 , 

1970 ; Grid II : September 6 to 11, 1969, and 

February 18 to 23, 1970 . The results (table 1 ) 

show several patterns . There is a difference 

between summer and winter populations i n 

almost all cases as well as a number of differences 

between plots . From other work on 

small mammal populations we would predict 

that there would also be variations from year 

to year . 

Since this method assumes that all th e 

mammals on the inner grid are caught, it i s 

important that trapping be highly efficient . 

Study of the trapping data (Miller 1970) suggests 

that winter populations may not be as 

trappable as summer populations, so winte r 

estimates in table 1 may be low . We have 

planned further work to increase trapping 

efficiency and so improve the accuracy of 

these estimates . 

Meanwhile, it is possible with our presen t 

data to obtain at least preliminary values for 

small mammal biomass (table 2) . Note that 

the differences between the two plots in tota l 

biomass are rather small, and that the dro p 

from summer to winter, in total biomass, is 

Table 1 .-Small mammal density estimates from grid-trapping (number per hectare ) 

Species 

Grid I Gri I I 

July March September February 

Sorex trowbridgei 16 .8 3.0 8.2 3 .0 

S. vagrans 4 .3 2.6 1 .3 . 4 

Neurotrichus gibbsi 3 .5 .9 3.6 1 . 3 

Microtus oregoni 1 .3 .9 1 .7 1 . 7 

Peromyscus maniculatus 1 .7 1.3 2.2 1 . 3 

Total 27.6 8.7 17 .0 7 .7 

Table 2.-Small mammal biomass (grams per hectare) 

Species 

Grid I 

Gri d If 

July March September February 

Sorex trowbridgei 82 .3 18.5 41 .0 16 . 5 

S. vagrans 27 .4 17 .8 6.7 2 . 7 

Neurotrichus gibbsi 27 .9 7.7 51 .4 11 . 1 

Microtus oregoni 19 .8 14 .1 28.6 27 . 9 

Peromyscus maniculatus 27.7 22.0 36.6 22 . 2 

Total 185 .1 80 .1 164 .3 80.4 

20 1

proportionally about the same for each plot . 

Future studies will tell us what fractions of 

this biomass are attributable to variou s 

nutrient substances ; that is, what the chemical 

constitution of these small mammal bodies is . 

We can already make preliminary estimate s 

of energy flow through these small mamma l 

populations. The initial energy source for ecosystems 

is sunlight. It is fixed by green plants 

and may be passed to populations of plant - 

eating animals. Within animal population s 

energy is used in respiration and tissue production 

and may also be passed on to othe r 

(predatory) animal populations . Much less 

energy is expended in tissue production tha n 

respiration . 

We can combine published data on the 

caloric value of animal tissue and on smal l 

mammal metabolic rates with our estimates o f 

biomass and information on populatio n 

dynamics to estimate the energy consumed i n 

tissue production and respiration for ou r 

small mammal populations . 

The energy incorporated in tissue production, 

for small mammals, has been estimate d 

to average 1 .5 kcal/gm (Gorecki 1965) . Th e 

energy expended by the animal in its dail y 

activity-the metabolic rate-varies by species . 

The values used in our calculations are given 

in table 3 . 

We do not as yet have a detailed knowledg e 

of the dynamics of these small mammal populations. 

However, we now know enough t o 

establish upper and lower limits of energ y 

Table 3 .-Metabolic rates used in calculating the energy 

expended in respiration by small mammal s 

Species Cal/gm/day Authority 

Sorex trowbridgei 806 Pearson (1948 ) 

S. vagrans 920 Pearson (1948), Gebczyski (1965 ) 

Neurotrichus gibbsi 800 Authors' estimate 

Microtus oregoni 500 Estimate based on value fo r 

red-backed vole (Pearson 1947) 

Peromyscus maniculatus 440 McNab (1963) 

Table 4.-Estimated energy flow (respiration + tissue production ) 

through small mammal population s 

(kcal per hectare per year ) 

Species Minimum Maximu m 

Sorex trowbridgei 5,250 15,75 0 

S_ vagrans 3,470 7,42 0 

Neurotrichus gibbsi 2,760 6,81 0 

Microtus oregoni 4,300 6,43 0 

Peromyscus maniculatus 4,450 6,86 0 

Total 20,230 43,27 0 

202

flow for these species by using data on reproduction, 

seasonal population density and 

growth of young (Miller 1970) . These estimates 

are given in table 4 . If necessary, we 

can strive for more accuracy in future estimates 

through a more detailed knowledge -o f 

the dynamics of these populations . 

The role and foraging stratum of mammal s 

in the forest ecosystem can be roughly categorized 

as shown in table 5. Secondary consumers 

are relatively numerous in the small 

mammal populations pointing to the abundance 

of their invertebrate foods in the litter 

and soil of the forest . 

In categorizing such roles, we must recognize 

that the primary consumers do on occasion 

eat other animals, and that the secondar y 

consumers do eat plants . It remains for us t o 

delineate in more detail the nature of th e 

seasonal diet for each species, as well as the 

rates of consumption, and the amounts an d 

composition of the excretory products . 

Table 5 .-Names, foraging strata, and consumer roles of mammals o f 

the Thompson site, Cedar River watershed, Washington 

Scientific name 

Common name 

Foraging 

stratum 

Consumer 

role' 

Canis latrans Coyote G II ° 

Cervus canadensis Elk (wapiti) G,S I ° 

Chiroptera Bats C II° 

Eutamias townsendi Townsend's chipmunk G,S I° 

Lepus americanus Snowshoe hare G,S I° 

Lynx rufus Bobcat G II° 

Microtus oregoni Oregon vole G I° 

Mustela erminea Shorttail weasel G II ° 

Mustela frenata Longtail weasel G II ° 

Neotoma cinerea Bushytail woodrat G,S I ° 

Neurotrichus gibbsi Shrew-mole B,L II ° 

Odocoileus hemionus Black-tailed deer G,S I ° 

Peromyscus maniculatus Deer mouse L,G I ° 

Sorex trowbridgei Trowbridge shrew L II° 

Sorex vagrans Vagrant shrew L II° 

Tamiasciurus douglasi Chickaree G,S,C I° 

Ursus americanus Black bear L,G,S I° 

Zapus trinotatus Jumping mouse G I° 

1 B = soil layer ; L = litter layer ; G = ground layer, under 1 foot ; S = shrub layer, 1 to 6 feet ; C = crown layer , 

area occupied by living crowns of forest overstory . 

21° = primary consumer, eats mostly plant material ; II° = secondary consumer, eats mostly animal matter . 

20 3

Bird Studies 

The forest birds were sampled both summer 

and winter on a 6 .0-ha plot chosen t o 

represent the Thompson site . This plot was 

divided into fifteen 0 .4-ha sample units on a 

map of the whole plot . A series of transect s 

was made through the whole plot on each 

sample day by Erickson who noted each bird 

observation by location and species . Durin g 

the summer sample period (May and June) , 

many breeding males were singing, whic h 

facilitated their detection . During the winter 

observation period (December), there had t o 

be more dependence on sight than sound . 

Five days of observation were spent at eac h 

season, and all observations took place during 

mornings when there was no heavy rain. Th e 

results for the two seasons are given in table 

6 . 

Some less common, but in some case s 

rather large birds, are also found on th e 

Thompson site . During the coming year we 

will be able to make population estimates of 

these species, adding substantially to our total 

estimate of bird biomass . 

The contrast between summer and winte r 

populations is much sharper for birds than fo r 

small mammals . This has implications for ecosystem 

modeling, since some birds reproduc e 

on the Thompson site but winter elsewhere , 

presumably suffering some mortality, an d 

thus constituting a one-way movement o f 

energy and nutrients out of the Thompso n 

site . From a knowledge of body weights w e 

have calculated the biomass of common bird s 

on the Thompson site by season (table 6) . 

Energy-flow estimates for these bird populations 

cannot yet be made because neithe r 

population estimates of the larger birds no r 

annual cycles of abundance of each bird species 

are yet available for the Thompson site . 

This must be one of our next topics of study . 

The role and foraging stratum of birds is 

roughly categorized in table 7. As in the case 

of small mammals, some of these birds shif t 

seasonally in their consumer-roles . A diet of 

invertebrates is essential for the young of al l 

species. Also, shifts in the relative abundanc e 

of available food will presumably be reflecte d 

in food habits . These are also topics on which 

more work must be done . 

Table 6.-Abundance (number per 100 hectares) and biomass (grams per hectare) 

of the most common birds by species and seaso n 

Species 

Abundance 

Biom as s 

Summer Winter Summer Winter 

Molothrus ater 26 0 11 .4 0 

Junco oreganus 43 0 9.9 0 

Troglodytes troglodytes 36 33 3.5 3 . 1 

Hylocichla ustulata 16 0 5.8 0 

Dendroica nigrescens 43 0 4.5 0 

Parus rufescens 23 0 2.3 0 

Regulus satrapa 0 33 0 2 . 0 

Oporornis tolmiei 26 0 1 .6 0 

Total 213 66 39.0 5 .1 

204

Table 7.-Names, foraging strata, and consumer roles of birds of th e 

Thompson site, Cedar River watershed, Washingto n 

Scientific name 

Common name 

Foraging Consumer 

stratum' role 2 

Bombycilla cedrorum Cedar waxwing C I° 

Bonasa umbellus Ruffed grouse L,G,S,C I° 

Bubo virginianus Great homed owl G II° 

Colaptes cater Red-shafted flicker G,C II° 

Columba fasciata Band-tailed pigeon L,S,C I° 

Corvus brachyrhynchos Common crow G II° 

Corvus corax Common raven G II ° 

Dendragopus obscurus Blue grouse L,G,S,C I° 

Dendrocopos villosus Hairy woodpecker S,C II° 

Dendroica nigrescens Black-throated gray warbler S,C II° 

Empidonax spp . Empidonax flycatchers S,C II° 

Hylocichla ustulata Swainson's thrush L,G II° 

Ixereus naevius Varied thrush L,G II° 

Junco oreganus Oregon junco L,G I ° 

Loxia curvirostra Red crossbill C I° 

Molothrus ater Brown-headed cowbird G I° 

Oporornis tolmiei MacGillivray's warbler S,C II° 

Parus rufescens Chestnut-backed chickadee S,C II° 

Passerella iliaca Fox sparrow G I° 

Perisoreus canadensis Gray jay G,C II ° 

Pipilo erythrophthalmus Rufous-sided towhee G I ° 

Piranga ludoviciana Western tanager S,C I ° 

Regulus calendula Ruby-crowned kinglet C II ° 

Regulus satrapa Golden-crowned kinglet C II ° 

Selasphorus rufus Rufous hummingbird G,S I ° 

Sitta canadensis Red-breasted nuthatch S,C II ° 

Sphyrapicus varius Yellow-bellied sapsucker S,C I ° 

Spinus pinus Pine siskin C I ° 

Troglodytes troglodytes Winter wren G,S II ° 

Turdus migratorius Robin L,G II ° 

Vermivora celata Orange-crowned warbler S,C II ° 

Vireo gilvus Warbling vireo S,C II° 

B = soil layer ; L = litter layer ; G = ground layer, under 1 foot ; S = shrub layer, 1 to 6 feet ; C = crown layer , 

area occupied by living crowns of forest overstory . 

2 1° = primary consumer, eats mostly plant material ; II° = secondary consumer, eats mostly animal matter . 

20 5

Further Studies 

The data provided for small mammal populations 

will be more useful if we are able t o 

make more accurate population density estimates 

and if we determine the yearly, seasonal, 

and geographical fluctuations in 

density. Further, a more detailed understanding 

of population dynamics will permit a 

more accurate estimate of energy flow, and 

set the stage for studies of uptake, storage , 

and loss of chemical materials by smal l 

mammal populations . 

Of course, the same sort of informatio n 

must be obtained for the large mammals , 

which include several important primary consumers. 

For large mammals the role of movement 

must be studied, since seasonal movements 

entail a translocation of materials, an d 

patterns of mortality provide loci of nutrien t 

release . 

For birds the study of movements will also 

be important, since long-distance annua l 

migrations provide a potential mechanism for 

the annual loss of materials from the ecosystem 

. 

A prime consideration in all such studie s 

will be the establishment of confidence limits 

for the data, and estimates of additiona l 

sampling necessary for results of a stipulate d 

level of accuracy . Then we will be in a position 

to respond to the needs of modeling in a 

realistic way with regard to the actual field o r 

laboratory work which provision of any particular 

piece of information will require . 

In addition to these inventories of vertebrate 

animals in the forest ecosystem, we will 

seek an understanding of the effects that 

animal populations may have on the populations 

of plants which constitute their food , 

i.e ., the control function exerted by animal s 

over plant populations. A knowledge of suc h 

relations, properly quantified, will help pro - 

vide the basis for models of the dynamics o f 

plant populations . 


This study was supported by the Department 

of Game, State of Washington ; Institute 

of Forest Products, University of Washington ; 

Ecology Training Grant from the Nationa l 

Science Foundation to the Zoology Department, 

University of Washington ; and by 

National Science Foundation Grant No . 


U .S. Analysis of Ecosystems, Internationa l 


36 to the Coniferous Forest Biome, IBP . We 

also wish to thank the personnel of the Cedar 

River watershed and the Seattle Water Department 

for their assistance and cooperatio n 

throughout the study . 


Adamczyk, K ., and L . Ryszkowski. 1968 . 

Estimation of the density of a rodent popu - 

lation using stained bait. Acta Ther . 

13(17) : 295-311 . 

Babinska, J ., and E . Bock. 1969 . The effect of 

prebaiting on capture of rodents . Act a 

Ther. 14(19) : 267-270 . 

De Lury, D . B . 1947 . On the estimation o f 

biological populations. Biometrics 3 : 

145-167 . 

Gebczynski, M . 1965 . Seasonal and age 

changes in the metabolism and activity o f 

Sorex araneus L . Acta Ther. 10(22) : 

303-331 . 

Gorecki, A. 1965 . Energy values of body in 

small mammals . Acta Ther . 10(23) : 

333-352 . 

Grodzinski, W ., Z. Pucek, and L . Ryszkowski . 

1966. Estimation of rodent numbers b y 

means of prebaiting and intensive removal . 

Acta Ther . 11(10) : 297-314 . 

Hayne, D . W. 1949 . Two methods for estimating 

population from trapping records . J . 

Mammalogy 30(4) : 399-411 . 

Leslie, P. H. 1952. The estimation of population 

parameters from data obtained by 

means of the capture-recapture method. II . 

The estimation of total numbers . Biometrika 

39(3&4): 363-388 . 

McNab, B . K. 1963 . A model of the energy 

budget of a wild mouse . Ecology 44(3) : 

521-532 . 

206

Miller, S. 1970. Small mammal populations i n 

a Douglas-fir forest : Cedar River, Washington. 

102 p. M .S. thesis on file at Univer . 

Wash ., Seattle . 

Pearson, O . 1947 . The rate of metabolism o f 

some small mammals. Ecology 28(7) : 

127-145 . 

1948 . Metabolism of small mammals 

with remarks on the lower limit of 

mammalian size . Science 108(2794) : 44 . 

Schwarz, S. S. 1967 . Indirect methods of estimating 

field metabolism of mammals . In K . 

Petrusewicz (ed.), Secondary productivity 

of terrestrial ecosystems, p. 225-239 . 

Warsaw, Poland : Panstwowe Wydawnictwo 

Naukowe . 

Zippin, C. 1958. The removal method of 

population estimation . J. Wildlife Manage . 

22(1) : 82-90 . 

207

Terrestrial Process Studies 

209



Terrestrial process studies i n 

conifers: a review 

Abstract 

R . B . Walker, Department of Botan y 

D . R . M . Scott, College of Forest Resource s 

D . J. Salo, Department of Botan y 

aqd 

K . L . Reed, College of Forest Resources 



A few studies on the physiological processes of conifers go back as much as a century, but most informatio n 

has been accumulated during the past 40 years . These efforts have involved most of the widely distribute d 

species to some extent, but limited information is available even on those species most studied, and very little 

on the remainder. With Douglas-fir (Pseudotsuga menziesii), the most studied species of our region, considerable 

information is available, but insufficient for process modeling . Of other species of the Western United State s 

much less is known . Thus intensive studies ofphysiological processes will be necessary in the Coniferous Biom e 

effort on each of the species of principal importance . In this paper, the current status of information on th e 

following topics is reviewed: CO 2 assimilation and respiration; transpiration, water conduction and wate r 

deficits; translocation of photosynthates; and mineral nutrition . Stomatal behavior and leaf resistance are 

considered with respect to the various gas exchanges . Energy budgets, nutrient cycling, growth, and modeling o f 

processes are closely related to the foregoing topics, but are covered elsewhere in this symposium. 

Introduction 

Central in the goals of the IBP is the measurement 

of productivity and production, and 

particularly the understanding and predictio n 

of these for vegetation and ecosystems . It 

follows that a basic knowledge of the physiological 

and related processes involved is 

essential to reaching these goals . In the Coniferous 

Forest Biome our concern is naturally 

centered around the conifers, although associated 

deciduous and herbaceous species are 

also important. However, this review concentrates 

on the conifers as the species demanding 

our principal efforts . 

Experimentation with conifers, which began 

about a century ago and was taken up by 

various workers periodically, has been included 

in several general reviews in the Encyclopedia 

of Plant Physiology (Leyton 1958, 

Huber 1956, Stalfelt 1956, Pisek 1960), an d 

discussed in detail by Kramer and Kozlowski 

(1960) . Since then, a number of reviews concerning 

woody plants have covered various 

aspects: viz. food relations (Kozlowski an d 

Keller 1966), mineral nutrition (Baule an d 

Fricker 1967), and translocation (Zimmermann 

and Brown 1971) . Of course, the results 

of many experiments with broadleafed wood y 

plants are pertinent to conifers . Further, general 

principles established with herbaceou s 

plants may be applicable to conifers, although 

this carryover is more tenuous and in each 

case usually demands experimental verification . 

This review will concentrate on findings o f 

the past decade which, together with the 

work discussed in the reviews mentione d 

above, serves as the background on which th e 

studies of the Coniferous Biome are based . 

Included is considerable unpublished material 

211

from theses. The principal processes an d 

subjects which will be covered in sequence are 

(A) assimilation of CO 2 and respiration , 

(B) transpiration, water conduction, an d 

water deficits, (C) translocation of photosynthates, 

and (D) mineral nutrition . Stomatal 

status and leaf resistance are considere d 

in relation to various aspects of gas exchange . 

Energy budgets, nutrient cycling, growth, and 

modeling of processes are considered b y 

other authors in this symposium, so are no t 

emphasized in this paper . 

Assimilation of CO 2 

and Respiration 

The principal emphasis in the IBP on assessment 

and understanding of dry matter 

production naturally focuses attention o n 

photosynthesis and the conditions and factors 

which influence its rate . Likewise the magnitude 

of respiratory losses is of complementary 

interest. Both short-term and long-term influences 

on the rates of these processes are o f 

importance in estimating seasonal and annual 

totals . Further, both inter- and intra-specifi c 

variability must be taken into account . Regardless 

of the factor under study, methods of 

individual measurement assume marked 

importance, and carrying over these measurements 

commonly made on portions of the 

plant to the entire forest stand is even mor e 

important. The latter aspect is considered i n 

the papers of this symposium on terrestrial 

modeling. 

Measuremen t 

The appearance in late 1971 of the detailed 

and authoritative book, Plant Photosyntheti c 

Production, Manual of Methods (Sestak, 

6atsky, and Jarvis, eds .), makes a detailed 

consideration of principles and techniques o f 

measurement unnecessary here . Thus a general 

consideration of methods and their applicability 

in the Coniferous Biome wil l 

suffice . 

Many useful studies have been performe d 

under controlled-environmental conditions , 

using assimilation chambers of various types 

(Jarvis et al . 1971) . Such studies may be criticized 

on the basis that plants grown or hel d 

for study in controlled-environment rooms or 

boxes may behave differently from materia l 

growing under natural conditions . Thus a 

number of studies have been devoted to ga s 

exchange of conifers in the field (e .g., Botkin , 

Woodwell, and Tempel 1970 ; Gentle 1963 ; 

Helms 1965, 1970 ; Hodges 1967 ; Hodges and 

Scott 1968; Kunstle 1971; Ungerson and 

Scherdin 1965; and Woodman 1971) . Also 

the extensive unpublished studies on Picea 

abies (L.) Karst of W . Koch' and of O . L . 

Lange and E . -0. Schulze 2 should be mentioned. 

To eliminate the influence of greenhouse 

or controlled-environment on the 

plants, several workers have used seedlings o r 

saplings brought in from the field or garden , 

or made measurements on excised branches 

from field growing trees (Brix and Ebell 1969 ; 

Keller 1971 ; Parker 1963; Pisek, Larcher , 

Moser, and Pack 1969 ; Pisek and Kemnitzer 

1968 ; and Poskuta 1968) . 

All of the measurements in the studies 

cited in the previous paragraph were made using 

assimilation chambers of various kinds . 

Difficulties and potential errors associate d 

with use of such enclosures are well covere d 

by Larcher (1969a) and Jarvis et al . (1971) . 

The major concerns pertain to excessive 

boundary layer resistances if stirring is inadequate, 

to significant departures of irradiatio n 

and energy balances from those of unenclose d 

foliage, and to errors inherent in the measuring 

itself . 

Techniques other than those using assimilation 

chambers are available (Larcher 1969a , 

Denmead and Mcllroy 1971) . Harvest techniques 

give direct measurements of dry matte r 

accumulation, but must pertain with conifer s 

to time intervals of several weeks or months . 

Thus they cannot give information on responses 

to various controlling factors, nor d o 

they give information on respiratory losses . 

However, apart from their intrinsic value , 

they may serve as an independent check on 

I Forest Botanical Institute, University of Munich , 

Germany . 

2 Botanical Institute II, University of Wiirzburg , 

Germany . 

212

the validity of extrapolating from assimilatio n 

chamber measurements to forest stands . Thus 

Botkin, Woodwell, and Tempel (1970) noted 

that values of production calculated fro m 

assimilation chamber data were about 30 per - 

cent above harvest values for the Brookhave n 

oak-pine forest . Meteorological techniques 

have been applied to conifer stands in a few 

cases (Baumgartner 1969, Denmead 1969 , 

Kinerson 1971) . These can give values for en - 

tire stands without the influence of enclosures, 

but require large even stands, level 

topography, and uniform atmospheric conditions, 

thus measurement of either short or 

long term influences of factors affecting 

assimilation is very difficult. The use of 

meteorological data in combination wit h 

physiological information has potential fo r 

mathematical modeling and prediction of 

CO 2 assimilation (Reed and Webb 1972) . 

Difficulties arise in defining and measurin g 

the surfaces active in reception of light and i n 

CO 2 exchange. The general problems an d 

techniques for these assessments, particularl y 

in relation to broad-leafed plants and grasses , 

are covered in the review of Kvet and Marshal l 

(1971) . The needle leaves of conifers usuall y 

have longitudinal rows of stomata on th e 

lower surface or on the interior surfaces i n 

fascicled species . Thus it is common to consider 

the "one-side" area as accounting fo r 

most of the gas exchange . Areas can be attained 

by tedious linear measuring and counting 

methods (Kvet and Marshall 1971 , 

Madgwick 1964), but often dry weights o f 

foliage only have been reported to avoid thi s 

time-consuming effort. The introduction of 

the glass-bead technique (Thompson an d 

Leyton 1971) should encourage the determination 

of surface areas as well as dr y 

weights so that data can be compared between 

species . Further, because of the mathematical 

relations of gas diffusion and leaf 

resistances, most models of photosynthesi s 

require that values of photosynthesis b e 

expressed on a unit leaf area basis (Reed and 

Webb 1972) . The foliar distribution in crown s 

and stands must be obtained for extrapolation 

of gas exchange data, and several description s 

of such methods are available (Stephens 1969 , 

Kinerson 1971) . 

Factors Affecting CO 2 Assimilation 

and Respiration 

In general, the principle of limiting factor s 

will apply to the relative importance of th e 

environmental and endogenous variables 

which control photosynthetic and respirator y 

rates. Thus in consideration of an individua l 

variable or factor, its effect may be masked at 

particular times by the limiting influence o f 

another factor. In the discussions of any individual 

factor which follows, it is implicit that 

the influences described are those of the 

factor when it is believed to be limiting the 

metabolic process (Stalfelt 1960) . 

Light Intensity 

Numerous studies in controlled-environmental 

systems and using excised branche s 

have shown that conifers exhibit typical 

limiting-factor response to light, with saturation 

dependent on the species and growing 

conditions (Pisek 1960) . For those species of 

principal interest in the Coniferous Biome a 

number of values for saturating light intensity 

have been reported (table 1) . These values are 

in general agreement with the range of 

1,850-3,200 ft-c stated by Polster (1967b) t o 

give maximal net photosynthesis in bot h 

broadleafed and coniferous trees, except for 

the value quoted by Kiinstle (1971) . Further 

verification of light responses under field conditions 

are needed for adequate modeling 

even of Douglas-fir, Pseudotsuga menziesii 

(Mirb.) Franco. Western hemlock, Tsuga 

he te rophylla (Rafinesque) Sargent, Pin us 

ponderosa Laws., and the Abies species are 

much in need of further study with respect t o 

light . 

Recent studies on photorespiration and its 

enhancement with increasing light intensity i n 

spruce species (Poskuta 1968, Ludlow an d 

Jarvis 1971b) indicate the need to understand 

the variations of this process in the species o f 

special interest in the Coniferous Biome . This 

information will not be necessary for our 

earlier empirical models, but will become 

essential for more theoretically based model s 

in the future as well as being necessary for 

estimates of gross productivity . 

213

Table 1 .-Saturating light intensities for net photosynthesis in various species of conifers 

Description of plan t 

material 

Saturating ligh t 

intensity 

Author 

- - - foot-candle'-- - 

Douglas-fir (Pseudotsuga menziesii (Mirb .) Franco) : 

65-day-old seedlings grown with 1,100 ft-c 3,000 Krueger and Ferrell , 

long day ; temp 19°C day, 9°C night 196 5 

100-day-old seedlings grown at 18°C and 1,000 ft-c 2,500 Brix, 196 7 

1-year-old seedlings-outdoor grown ca . 2,300 Krueger and Ruth, 196 9 

4-year-old seedlings grown in cans in the garden 2,000-3,000 Fry, 196 5 

2-year-old seedlings-nursery grown 1,600 Hodges, 196 5 

Current year foliage from 10-year-old ca. 4,600 Kunstle, 197 1 

trees (upper crown July-Aug . ) 

Western hemlock (Tsuga heterophylla (Rafinesque) Sargent) : 

2-year-old seedlings-nursery grown 1,400 Hodges, 1965 

1-year-old seedlings-outdoor grown Krueger and Ruth, 196 9 

in light shade 2,30 0 

in heavy shade 1,25 0 

Noble fir (Abies procera Rehd .) : 


Sitka spruce (Picea sitchensis (Bong .) Carr .) : 

1-year-old seedlings-outdoor grown Krueger and Ruth, 196 9 

in light shade 2,30 0 

in heavy shade 1,25 0 


3- to 4-year-old seedlings and branches 1,450-1,930 Ludlow and Jarvis, 

from 10-year-old forest trees 

1971b 

1 1 foot-candle (ft-c) = 10 .8 lux ; 6,750 ft-c = 1 cal cm-2 min-e ; 1 Wm-2 = 1 .43 x 10-3 cal cm-2 min-l . 

Temperature 

The concomitant responses of respiratio n 

and gross photosynthesis to temperature , 

modification by light and moisture regime s 

and adaptation, as well as special influences o f 

low and high temperatures, all make it difficult 

to clearly indicate the influence of 

temperature alone on these processes (Krame r 

and Kozlowski 1960, Kozlowski and Kelle r 

1966, Larcher 1969b, Polster 1967b) . Thu s 

Pisek and his collaborators (1969) quote d 

rather considerable ranges of optimal temperatures 

at 10,000 Lux for a list of Europea n 

species. The total ranges for the conifers i n 

this list extended from 9° to 22°C, with the 

means showing a much narrower range of 12° 

to 15°C . Larcher (1969b) assembled data o n 

summer temperature optima for some 1 8 

conifers . These varied from the low range of 

12°-16°C for Pinus cembra L. to the hig h 

range of 20°-22°C for Abies balsamea (Mill.) . 

Data for species of our special interest i n 

the Coniferous Biome are not abundant . For 

Douglas-fir seedlings values of 20°-25°C a t 

saturating light (Krueger and Ferrell 1965) , 

and 18°-20°C at 1,000 ft-c (Brix 1967) hav e 

been reported . For 10-year-old saplings of thi s 

species growing in the garden we noted a 

somewhat lower range of 12°-17°C . 3 This 

3 R . B . Walker and D . J . Salo . Unpublished data . 

214

may be explained by the observation of Brix 

and Ebell (1969) that dry matter production 

of Douglas-fir shoots in bud dormancy had a 

broad optimal range of 12° to 24°C, with 

maximum increase in stem diameter and i n 

total plant dry weight at 18°C . 

Respiration is believed to follow the van't 

Hoff Q 1 0 relationship up to about 60°C , 

where enzyme inactivation ensues . This is supported 

by measurements of dark respiration , 

and by the fall off in net photosynthesi s 

above 20°-25°C (Pisek 1960, Larcher 1969b) . 

Studying current-year needles of 10-year-ol d 

Douglas-fir, Kiinstle (1971) found the Q1 0 of 

respiration of fully expanded needles to be 

about 2.2, but to be about 4 .0 in buds in late 

spring. Also Brix and Ebell (1969) noted a 

steady decline in dark respiration of Douglas - 

fir needles with years of age. However, more 

intensive study of respiration not only o f 

needles but of non-green parts will be neede d 

in our Biome effort for modeling production . 

This is even more pertinent for the specie s 

other than Douglas-fir . 

Since the evergreens may function through - 

out the year, special interest has long existe d 

in their responses to low temperatures (Pise k 

1960) . The low-temperature compensatio n 

points for conifers in the winter are in th e 

range of -5° to -8°C, which is also the expected 

range of freezing of the needle s 

(Larcher 1969b) . Again considering species of 

special interest in the Coniferous Biome, th e 

low-temperature compensation point fo r 

Sitka spruce proved to be -5°C (Ludlow an d 

Jarvis 1971b) . Respiration is very low a t 

temperatures below 5°C, and in our climat e 

such temperatures usually predominate onl y 

when light intensities are very low . Thus with 

light often limiting, temperature influences o n 

photosynthesis are small, and net assimilatio n 

in Douglas-fir was near the maximum fo r 

short times even at 0°C if illumination wa s 

less than 500 ft-c (see footnote 3) . However , 

the assimilation rate declined if a temperatur e 

under 2°C was maintained for an hour or 

more . 

Wintertime depressions in photosyntheti c 

capacity have been studied frequently in th e 

past (Pisek 1960) . These declines have bee n 

attributed to a combination of stomatal 

closure and a probable biochemical facto r 

(Parker 1963), to stomatal closure in Picea 

abies, Pinus sylvestris L., and Juniperus cornmunis 

L. in northern Sweden (Ungerson an d 

Scherdin 1965), to a breakdown of the photo - 

synthetic apparatus in Pinus taeda L. (Perry 

and Baldwin 1966), to a reversible effect 

associated with freezing of the needles an d 

presumed dehydration in Abies alba Mill . 

(Pisek and Kemnitzer 1968), to adverse needle 

water relations (Zelawski and Kucharska 

1967), and to a typical seasonal behavior in 

pines (McGregor and Kramer 1963) . Seasonal 

variations will be considered in more detail i n 

a subsequent section . 

There is some information on high temperature 

compensation points of photosynthesis , 

although nothing pertaining particularly to 

species of Western United States . The si x 

European conifers studied at Innsbruck (Pisek 

et al . 1968) varied from 46° to 50°C in heat 

resistance, and from 37° to 41°C in the maximum 

temperature for net photosynthesis . If 

our species fall into similar ranges, as seem s 

likely, there is little reason to be concerned 

with this aspect because only very occasionally 

would air temperatures reach the 

compensation value for photosynthesis in our 

region. Only if leaf temperatures should 

greatly exceed air temperatures on hot day s 

would this be a matter of interest. 

Carbon Dioxide 

Under favorable conditions of light intensity, 

temperature and moisture, carbo n 

dioxide concentration may sharply limi t 

photosynthesis . Thus Koch (1969) reported 

some threefold increase in yield of severa l 

conifers from raising CO 2 concentration to 

5X the normal atmospheric level in controlle d 

environment experiments. Also Fry (1965 ) 

recorded a measurable increase in net assimila - 

tion of Douglas-fir saplings when he increase d 

CO 2 from ambient (0 .035 percent) to 0 .0 5 

percent if light intensity was above 1,600 ft-c . 

Likewise Ludlow and Jarvis (1971b) foun d 

that net photosynthesis of Sitka spruce seedlings 

increased almost linearly with CO 2 concentration 

as it was varied from 0 to 0 .0 4 

percent, then continued to rise less steeply t o 

215

at least 0 .06 percent CO 2 . However, in th e 

forest it is not feasible to increase CO 2 concentration 

except for the natural contributions 

from the soil and litter . It is important 

to make periodic checks in the field using in - 

creased CO 2 in assimilation chambers to 

detect if the ambient CO 2 concentration is 

limiting, and to minimize the "draw down" of 

CO 2 in enclosures to avoid excessive deviations 

from assimilation rates of the unenclosed 

foliage. Also the atmospheric CO 2 concentration 

is known to be variable, and this 

should be monitored for taking into account 

any limiting effect on photosynthesis . 

Water Defici t 

A substantial literature documents the reductions 

in CO2 assimilation in various species 

caused by even moderate water deficits 

(Kozlowski and Keller 1966) . Some reductio n 

may occur in the afternoon with high transpirational 

rates even on well-watered soils , 

but lowered soil moisture is the principal 

cause of severe reductions . Using 4-yearold 

Douglas-fir saplings in containers, Fry 

(1965) found that depletion of bulk soil 

water to a potential of -5 bars reduced ne t 

photosynthesis to zero (compensation point ) 

and reduced respiration to about 10 percen t 

of normal. Since needle water potential 

dropped to about -30 bars, soil water potential 

at the root surfaces was probably considerably 

below -5 bars. Hodges (1965, 1967) , 

studying 2-year-old Douglas-fir seedlings in 

the field, found net assimilation closely correlated 

with leaf water potential . Using 7- to 

9-year-old saplings of about 0 .5-0.75-m height , 

Hinckley (1971) found net assimilation in 

Abies amabilis (Dougl.) Forbes and Abie s 

procera Rehd. to drop sharply with reduced 

water potential (pressure chamber method) o f 

the needles, reaching 10 percent of maximu m 

assimilation at about -10 bars needle water 

potential in the A . procera, and at about -1 3 

bars in the A. amabilis . 

Such declines in net assimilation caused b y 

reduced soil moisture and lowered leaf wate r 

potential are well recognized to be caused b y 

increases in stomatal and other leaf resistance s 

(Fry 1965, Slatyer 1967, Slavik 1971, Jarvis 

1971) . Considerable information has bee n 

accumulated on stomatal behavior in Douglas - 

fir (Fry 1965, Reed 1968, Phillips 1967) , 

ponderosa pine4 (Lopushinsky 1969), an d 

true firs (Lopushinsky 1969, Hinckley an d 

Ritchie 1970). However, meager informatio n 

is available on actual values for stomatal an d 

other leaf resistances in species of specia l 

interest in the Coniferous Biome . Fry (1965 ) 

calculated stomatal resistance in Douglas-fi r 

saplings from stomatal infiltration pressures . 

From these pressures ranging from 0 .25 to 

1.65 atm he calculated stomatal resistances o f 

2.3 to 32 sec cm-1 for CO 2 , and additional 

leaf resistances of 3 .7 to 83 sec cm-1 . Ree d 

(1972) found general agreement with thes e 

values. Ludlow and Jarvis (1971b) reported 

stomatal resistances in Sitka spruce seedling s 

in excess of 20 sec cm-1 at very low irradiance 

or at very low temperatures, with normal 

values of about 4 sec c m ' . Their data showed 

mesophyll resistances of 20 to 30 sec cm-1 or 

more at very low irradiances and at very lo w 

temperatures, with normal values of about 6 

to 8 sec cm- ' . Leaf resistance is such an 

important factor that understanding stomatal 

and other components of resistance in all species 

of special interest in the Biome is necessary 

for the development of reasonable 

models of photosynthesis and transpiration . 

Mineral Nutrition 

A number of experiments have established 

that mineral deficiency may limit photosynthesis 

(Kozlowski and Keller 1966, Baule and 

Fricker 1967) . In the Pacific Northwest, 

nitrogen deficiency is common in forests, thu s 

making determination of the influences o f 

this element very important in the Coniferou s 

Biome program. The stands at the Cedar Rive r 

Watershed show incipient N deficiency, s o 

care will be taken to assess the influence o f 

this element. Keller (1971) recently reporte d 

a 50 percent increase in CO2 assimilation i n 

needles of Picea abies seedlings with increas e 

in N content from 0 .41 to 1 .11 percent . Also 

Brix (1971) recently measured increases i n 

4 A . P . Drew, L. G . Drew, and H . C . Fritts . Unpublished 

data . 

216

net assimilation in excised branches fro m 

24-year-old Douglas-fir trees which had been 

fertilized with N . However, the increases wer e 

noted only at light intensities of 2,000 ft- c 

and above . 

Diurnal and Seasonal Pattern s 

(including adaptation ) 

Diurnal cycles or fluctuations in light, temperature, 

water vapor pressure gradients, an d 

other external factors are interactive with 

internal factors such as endogenous rhythm o f 

stomatal behavior, recovery from water deficit , 

and translocation. Therefore the response of a 

tree day by day is very complex . 

Gentle (1963) and Helms (1965) studie d 

diurnal responses of 38-year-old Douglas-fir i n 

detail over a 4-year period . Superimposed o n 

the daily responses to light and darkness were 

short-term and longer term fluctuations attributable 

to varying temperature, radiation and 

cloud cover, relative humidity, water deficit, 

and undetermined influences . Midday depressions 

were common in the summer but also 

occurred in cool autumn weather . Similar detailed 

studies of diurnal behavior of the other 

species important in the Biome are lacking . 

The diurnal endogenous changes in stomatal 

aperture have been studied in several 

species. Fry (1965) noted that saplings grow n 

from seed gathered at Pack Demonstratio n 

Forest, La Grande, Washington, did no t 

exhibit any closing of the stomata during th e 

day or night if water status was favorable . 

This was largely confirmed by Reed (1968) , 

although he observed a slight closing tendenc y 

at night. Phillips (1967) studied 38-year-old 

Douglas-fir trees at Pack Demonstration Forest 

and found that the stomata of leaves i n 

the lower crown closed at night. A geographic 

or provenance difference in stomatal response 

of Douglas-fir was demonstrated in the field 

in southern Oregon, where stomata were open 

at night in the spring and early summer, but 

closed at night from July through September 

(Reed 1972). Drew, Drew, and Fritts (see 

footnote 4) and Lopushinsky (1969) independently 

observed that the stomata of Pinus 

ponderosa are closed at night . Abies amabilis 

growing in the field was observed to hav e 

more open stomata in early morning than did 

Abies procera, indicating that the stomata of 

A. amabilis were probably more open durin g 

the night (Hinckley 1971). The stomatal 

behavior of western hemlock is unknown an d 

needs attention . 

The influence of season of the year ha s 

already been touched on in connection with 

winter depression of assimilation . In a more 

general context, seasonal patterns of CO 2 

assimilation were studied by Gentle (1963) 

and by Helms (1964, 1965) in the 38-year-ol d 

Douglas-fir stand mentioned above . As would 

be expected, winter assimilation was low and 

quite variable, depending on the light an d 

temperature prevailing. Spring and autum n 

performances were good, although somewha t 

below the summer . Higher temperatures presumably 

made respiration a good deal highe r 

in the summer than in spring and autumn and . 

particularly than in the winter . Also Wood - 

man (1968, 1.971) studied net assimilation in 

one of the trees of this lame stand during th e 

growing season of 1967. A somewhat different 

pattern can be expected in conifers 

growing in regions with very dry summers , 

with much Of the yearly total of photosynthesis 

occurring during the winter and spring . 

Such studies of seasonal patterns are imperative 

for adequate models of production in 

forest stands . This is particularly true becaus e 

a considerable number of variables in addition 

to those of the physical environment have 

marked influences (Kozlowski and, Kelle r 

1966). These include morphological variatio n 

(sun and shade foliage, stomatal distribution , 

cuticular thickness, and others), adaptatio n 

with respect to light intensity and temperature 

in par icular, interspecific and intraspecific 

variations, shading, dormancy, age of 

foliage, position in the crown, and the natur e 

of past seasons . Some information on these 

features is available for Douglas-fir. However 

very limited knowledge, gained mostly fro m 

studies with seedlings, is available in this connection 

for the other species of special concern 

in the Coniferous Biome programwestern 

hemlock, ponderosa pine, and one of 

the true firs l Hodges and Scott 1968, Kruege r 

and Ruth 1969, Ludlow and Jarvis 1971b t 

217

Pharis et al. 1967). Again Helms (1970) ha s 

studied CO 2 assimilation in ponderosa pine 

during the summer in the natural environment, 

and envisions extending this study t o 

all seasons of the year. Intensive work on CO 2 

assimilation of all of the other species named 

above from the standpoints of environmenta l 

factors and the other influences on their behavior 

will be needed for realistic modeling o f 

their photosynthetic production . 

Water Relationships 

Although water deficit was taken up in th e 

previous section as a potential limiting facto r 

in CO 2 assimilation and respiration, the general 

subject of water relationships is of concern 

in water balance, nutrient cycling, an d 

other soil-plant-atmosphere relations, as wel l 

as in its effects on plant metabolism . Also th e 

balance between water absorption and transpiration 

determines any water deficit in th e 

plant. Water relationships of woody plants 

were thoroughly reviewed by Polster (1967a) . 

Water Absorptio n 

The factors affecting water uptake in 

woody plants were discussed in detail i n 

Kramer and Kozlowski (1960) . In the usuall y 

well-drained and well-aerated soils of the 

western forests, temperature of the soil an d 

roots and the level of soil moisture are th e 

principal factors affecting water uptake in th e 

root zone. These factors are being regularl y 

monitored at the intensive study sites of th e 

Biome . Their influence will be of most concern 

in the magnitude of water deficits . Sinc e 

water uptake is wholly passive during times o f 

high water use, increased resistance to uptake 

in the root zone can be expected to enhanc e 

water deficits . Further, rate of water uptake 

and conduction may exert some effect o n 

mineral uptake (Slatyer 1967) . Investigation s 

of water uptake will probably not be a part o f 

the Biome studies, because of the difficultie s 

of measuring this quantity in large soil-roote d 

plants, and the near equivalence of absorptio n 

rates to transpiration rates . 

Water Conduction 

For woody plants, this aspect has bee n 

thoroughly reviewed, including special features 

of conifers, by several authors (Hube r 

1956, Kramer and Kozlowski 1960, Zimmermann 

and Brown 1971) . With reference to th e 

objectives of the Coniferous Biome, interes t 

in conduction pertains to the magnitude o f 

negative sap pressures (as a measure of water 

deficit), to its use as an indicator of transpirational 

fluctuations where actual transpirational 

measurements can not be made readily , 

to intra-tree water adjustments includin g 

potential storage in stems and branches, an d 

to its influence on rate of translocation of 

mineral nutrients . Nonetheless, rates of conduction 

will be inferred of necessity for ou r 

models from transpirational rates . However , 

some indication of the nature of potential 

storage of water in stems or branches may b e 

attained from measurements using sap velocity 

flow meters and by study of intra-crown 

variation in the sap pressure (Scholander technique) 

(Waring and Cleary 1967) of stems an d 

needles (Hinckley 1971, Ritchie 1971) . 

Transpiratio n 

The predominant interest in water relation - 

ships is with transpiration both from a hydro - 

logic viewpoint concerned with total plan t 

use, and from its influence on water deficit s 

and on resistance to water vapor and CO 2 

transfer. This major interest has resulted i n 

substantial work on the transpiration of conifers, 

and considerable attention to species o f 

interest in the Coniferous Biome . Thes e 

studies will be considered below under th e 

headings of measurement, influence of environmental 

factors, effects of leaf temperature 

and resistances, and the significance of 

water deficits . 

Measuremen t 

Although there are various methods fo r 

measurement of transpiration, in the fiel d 

with large plants the air-flow method using 

leaf enclosures, lysimeter systems, meteorological 

techniques, and tritiated water 

methods are most feasible. The air-flo w 

218

method, using psychrometers or infra-red ga s 

analyzers as the detecting instruments, is commonly 

applied to plant material enclosed i n 

assimilation chambers . The possible artifact s 

of the systems pointed out for CO 2 assimilation 

measurements (Jarvis et al . 1971) are 

even more critical in the case of transpirational 

assessments . This is true because of th e 

role of energy balance and resulting lea f 

temperature in determining the vapor pressur e 

gradient from leaf to atmosphere . Also 

adsorption and condensation problems ofte n 

make measurements difficult and less accurat e 

during cool moist periods either diurnal or seasonal, 

although rates are characteristically low 

under such conditions . The lysimeter technique 

is suited to diurnal and seasonal studie s 

without the artificiality of enclosing th e 

foliage (Fritschen 1972) . Root disturbance 

occurs in installation, but this should hav e 

little effect on transpiration and passive up - 

take of water . Meteorological methods, whic h 

measure evapotranspiration, also maintai n 

natural conditions around the foliage, bu t 

necessitate sizable stands of homogenous 

vegetation and uniform air conditions for bes t 

measurements. The tritiated water metho d 

(Kline et al . 1972) yields information on transpirational 

activity only over periods of hour s 

and days, but its advantage over the other 

methods lies in its applicability to very larg e 

trees, such as the 75-m-tall old-growt h 

Douglas-firs in the H . J. Andrews Experimental 

Forest. During certain periods o f 

1972, all four of these methods of transpirational 

measurement will be compared at th e 

Thompson Research Center (Cedar River) . 

The lysimeter study will be continuou s 

throughout the 1972 spring through autum n 

season, and the simultaneous measurement o f 

transpired water in the gas stream from the 

assimilation cuvettes will permit estimation of 

leaf resistance as well as giving data on transpirational 

rates . 

Influence of Environmental Factors 

General discussions of these factors an d 

their influences have been given by Stalfelt 

(1956), Kramer and Kozlowski (1960), an d 

Kozlowski and Keller (1966) . Emphasis here 

will be on more recent studies with conifers 

and especially with species of special concern 

in the Coniferous Biome . 

The dominant influence of the vapor pressure 

gradient from leaf to air on transpirational 

rate when stomata are open has long 

been recognized. This was verified by Ritchie 

(1971) in a study with Abies amabilis and A . 

procera of about 14-m height, using vapor 

pressure gradient and also saturation deficit of 

the atmosphere . On a statistical basis, th e 

latter value accounted for 56 percent of th e 

variation in transpirational rate on a seasona l 

basis . 

Depletion of soil moisture has also lon g 

been recognized as a major factor in the control 

of transpiration . This was verified by 

Mullerstael (1968) using three species of pine , 

as well as other evergreens, and he als o 

showed marked species differences . Hinckley 

(1971) showed great reduction in transpiration 

(sap velocity) in Abies amabilis and A . 

procera with soil water depletion. Reed 

(1972) developed a computer simulation o f 

transpiration in the field . He showed that lo w 

vapor pressure gradients limited transpiratio n 

in the spring, but increased stomatal resistance 

resulting from depleted soil moisture an d 

water deficits limited transpiration during th e 

summer. He also showed that the availabl e 

soil moisture can vary greatly from year t o 

year, which can cause considerable seasonal 

and annual differences in total transpiration . 

Although moderate air movement is commonly 

recognized as a factor which increase s 

transpiration because boundary layers are reduced, 

few studies have been conducted with 

high air velocities. Tranquillini (1969) varied 

wind velocity from 0 .5 to 20 m sec-1 , and 

observed that net assimilation of Pinu s 

cembra increased somewhat up to about 4 m 

sec-t , then declined . In Picea abies the decline 

started at 1 .5 m sec-t . Caldwell (1970), working 

with the same Pinus cembra plants , 

showed that stomatal aperture and transpiration 

rate were only slightly reduced by hig h 

wind speeds, but photosynthesis was reduce d 

considerably because of changes in needle display 

to the light. High wind would be expected 

to be of importance in the Coniferou s 

Biome only in extreme habitats . 

219

Effects of Leaf Temperature an d 

Leaf Resistances 

Leaf temperature establishes the vapo r 

pressure of water in the intercellular spaces o f 

the leaf, thus it is imperative to have goo d 

measurements of this value to accompan y 

transpirational (and net assimilation) estimations 

. Methods for determination of lea f 

temperature have been recently described i n 

detail (Perrier 1971) . 

Although transpirational rates may approach 

potential evaporation with favorabl e 

soil and leaf water status, commonly leaf 

resistances markedly impede water loss . This 

was well demonstrated by Waggoner an d 

Turner (1971) in a study of transpiration in 

Pinus resinosa Ait. with both natural and 

artificially induced stomatal closure markedly 

reducing transpiration . The separation of stomatal 

and other leaf resistance is very desirable 

when feasible (Jarvis 1971) . The influence 

of wax in the stomatal pores in reducin g 

transpiration in Sitka spruce was noted by 

Jeffree et al . (1971) . Similar wax accumulations 

increasing with age of needles i n 

Douglas-fir have recently been observed o n 

material collected from the 35-year-old tree s 

at the Thompson Research Center (Cedar 

River) . 5 In the Coniferous Biome studies , 

more information is needed on these resistances 

in all species for successful productio n 

models . 

Significance of Water Deficit s 

Conifer species differ in their ability to 

maintain photosynthetic production in th e 

presence of water deficits, and in their abilit y 

to control water loss under similar environmental 

conditions (Hodges and Scott 1968) . 

The status of woody plants can be assesse d 

over the season by periodic tests of the pre - 

dawn pressure chamber value (Waring an d 

Cleary 1967), and this used in establishin g 

moisture gradients and ecological tolerance s 

and distributions. In the Coniferous Biom e 

such assessments are needed throughout th e 

growing season over wide geographic areas 

5 P. Machno and K . L. Reed. Unpublished data 

with all species of primary interest, and this 

program is already well under way (Waring et 

al. 1972) . 

Translocation 

of Photosynthates 

and Relations to Growth 

The translocation of photosynthate in 

woody plants was thoroughly reviewed b y 

Kozlowski and Keller (1966), and very recently 

discussed in detail by Zimmerman n 

and Brown (1971) . 

Among the conifers the pines, especiall y 

Pinus resinosa, have received the greatest attention 

(Rangnekar et al . 1969, Dickman n 

and Kozlowski 1970) . The former group , 

working with 15-year-old trees in a plantation, 

found that each branch appeared to b e 

self-supporting, contributing more liberally t o 

its own growth than to the tree leader, whic h 

probably expands in part from reserve carbohydrates 

. Further, they suggested that th e 

products of cambial growth are largely de - 

rived from current photosynthate . The results 

of Dickmann and Kozlowski (1970), wh o 

labeled the 1-year-old needles of second-orde r 

branches of 20-year-old trees, are in general 

agreement with this, as they found that recovery 

from the various sinks was high and i n 

the order : 2d-year cones > terminal needles > 

lateral needles > terminal internode > latera l 

internodes > 1-year-old wood . 

Ross (1972) labeled different-aged branc h 

segments of 9-year-old Douglas-fir trees o f 

about 6-m height . He likewise was able t o 

discern source-sink relationships. The proportion 

of 14 CO 2 exported from needles was a 

function of their age, of the attractive "pull" 

of external sinks, and inversely of water 

stress . One-year-old needles exported a large r 

proportion of their photosynthate than di d 

current-year needles even late in the growin g 

season . The attractive "pull" of the sink s 

decreased in the order: elongating ne w 

1st-order internodes and needles, elongatin g 

new 2d-order internodes and needles, an d 

stem . The fact that 1-year-old needles mainly 

supplied adjacent new shoots and thus ex - 

220

ported little photosynthate basipetally in contrast 

with the preferential translocation of 

photosynthate toward the stem by 2-year-ol d 

needles may be explained by the ability of a 

sink such as the stem to mobilize photosynthate 

falling off with distance in a n 

approximately logarithmic manner . 

These studies give clues to the relationships 

between CO 2 assimilation in the leaves an d 

the formation and differentiation of new cells 

in apical and cambial growth . These relationships 

are not simple, and care must be take n 

to avoid oversimplified connections . For 

example, net assimilation rates can give an 

index of amounts of photosynthate availabl e 

for growth only if position and age of th e 

foliage studied are taken into account (Ros s 

1972). With caution, however, net assimilation 

rates taken in different parts of th e 

crown can be used to estimate daily mean 

assimilation and the total photosynthate production, 

on the basis of the whorls o f 

branches of different heights (Woodman 

1968, 1971) . 

In the Coniferous Biome studies, effort s 

must be made to bring together net assimilation, 

translocation, and terminal and lateral 

growth into a workable model . It may be 

necessary in such efforts to take into accoun t 

growth inhibitors and promoters in such a 

model, since Lavender and Hermann (1970 ) 

have pointed out the importance of thes e 

substances in their studies of light and photo - 

period effects in Douglas-fir . Although there 

is information to build on with respect to re d 

pine and Douglas-fir (see above), further wor k 

will be necessary with Douglas-fir to utiliz e 

the information effectively . Clearly, specifi c 

studies of translocation and growth in wester n 

hemlock, ponderosa pine, and a true fir ar e 

needed . 

Mineral Nutrition 

A number of books cover the general field 

of mineral nutrition and specifically that of 

forest tree species (Baule and Fricker 1967 , 

Bengtson 1968, Epstein 1972) . The aspects of 

particular concern in the Coniferous Biom e 

are mineral uptake and cycling, and the role 

of mineral elements in metabolic processes . 

Mineral cycling studies in second-growt h 

Douglas-fir forests have been carried out for 

over a decade (Grier and Cole 1972) . Mineral 

cycling is important in the nutrition of th e 

trees since limited nutrient capital is available . 

These studies need to be extended to th e 

other species of special importance in th e 

Coniferous Biome program . 

The importance of nitrogen in chlorophyll 

production and the close correlation between 

chlorophyll contents and CO 2 uptake were 

pointed out by Keller and Wehrmann (1967 ) 

with reference to Picea abies and Pinus sylvestris 

. The widespread occurrence of nitrogen 

deficiency in Douglas-fir in western Washington 

and Oregon (Gessel et at . 1965) give s 

reason for careful attention to this element i n 

studies of coastal Douglas-fir in the Coniferous 

Biome . Also it will be wise to make a 

sufficiently broad spectrum of mineral analyses 

to detect low levels of other element s 

which might be limiting photosynthesis or 

growth . 

The foregoing indicates that a large back - 

ground of information on terrestrial production 

processes in woody plants is already 

available. With respect to Douglas-fir, the 

species currently receiving most emphasis i n 

the Coniferous Biome, the background i s 

appreciable and useful, but often lacking i n 

specific data needed for the construction of 

mathematical models . These deficiencies must 

be made up for adequate modeling of processes 

in this species. With other species of 

special concern in the Biome-western hemlock, 

ponderosa pine, and true firs-muc h 

more background information as well as 

specific data is needed in order to develo p 

effective production models . 

A cknowledgments 


ported in part by the University of Washingto n 

and in part by National Science Foundatio n 


Biome, U .S. Analysis of Ecosystems, Inter - 

national Biological Program . This is Contribution 


221


Baule, H., and C. Fricker. 1967 . The fertilizer 

treatment of forest trees . (Transl . from 

German .) 259 p. Munich : BLV Verlagsges . 

Baumgartner, A. 1969. Meteorological approach 

to the exchange of CO 2 betwee n 

the atmosphere and vegetation, particularl y 

forest stands . Photosynthetica 3(2) : 

127-249 . 

Bengtson, G . W . (ed .) . 1968 . Forest fertilization 

: theory and practice . 306 p . Muscl e 

Shoals, Ala . : Tenn. Valley Authority . 

Botkin, D., G. M . Woodwell, and Neal 

Tempel. 1970. Forest productivity estimated 

from carbon dioxide uptake . 

Ecology 51 : 1057-1060 . 

Brix, H . 1967 . An analysis of dry matter production 

of Douglas-fir seedlings in relatio n 

to temperature and light intensity . Can . J . 

Bot. 45 : 2063-2072 . 

1971 . Effects of nitrogen fertilization 

on photosynthesis and respiration in 

Douglas-fir. Forest Sci . 17 : 407-414 . 

and L . F. Ebell . 1969 . Effects of 

nitrogen fertilization on growth, leaf area , 

and photosynthesis rate in Douglas-fir . Forest 

Sci. 15 : 189-195 . 

Caldwell, M. M . 1970. Plant gas exchange a t 

high wind speeds . Plant Physiol. 46(4) : 

535-537 . 

Denmead, O . T. 1969. Comparative micrometeorology 

of a wheat field and a forest 

of Pin us radiata . Agric. Meteorol. 6 : 

357-371 . 

and I. C. Mcllroy. 1971. Measurement 

of carbon dioxide exchange in th e 

field . In Z . Sestak, J . 6atsky, and P . G . 

Jarvis (eds.), Plant photosynthetic production, 

manual of methods, p . 467-516 . Th e 

Hague : Dr . W . Junk N .V . Publishers . 

Dickmann, D . I ., and T. T. Kozlowski. 1970 . 

Mobilization and incorporation of photo - 

assimilated 14 C by growing vegetation an d 

reproductive tissues of adult Pinus resinosa 

trees. Plant Physiol . 45 : 284-288 . 

Epstein, E . 1972. Mineral nutrition of plants : 

principles and perspectives, 412 p . New 

York: John Wiley and Sons . 

Fritschen, L . J. 1972. The lysimeter installation 

on the Cedar River watershed . In 


ecosystems-a symposium, p . 255-260, illus. 



Fry, K . E. 1965. A study of transpiration an d 

photosynthesis in relation to the stomatal 

resistance and internal water potential i n 

Douglas-fir. 192 p. Ph.D. thesis on file , 

Univ . Wash ., Seattle . 

Gentle, S . W. 1963 . The effect of local weather 

conditions on the assimilation of carbo n 

dioxide by Douglas-fir in western Washing - 

ton. 122 p . Ph .D. thesis on file, Univ . 

Wash., Seattle . 

Gessel, S . P., T. N . Stoate, and K . J. Turnbull , 

1965 . The growth behavior of Douglas-fi r 

with nitrogenous fertilizer in western Washington 

. Res . Bull. No. 1, 204 p . Coll. For . , 

Univ . Wash., Seattle . 

Grier, Charles C ., and Dale W . Cole. 1972 . 

Elemental transport changes occurring during 

development of a second-growth Douglas-fir 

ecosystem . In Proceedings-research 

on coniferous forest ecosystems-a symposium, 

p . 103-113, illus . Pac. Northwest 

Forest & Range Exp . Stn., Portland, Oreg . 

Helms, J. A. 1964. Apparent photosynthesi s 

of Douglas fir in relation to silvicultural 

treatment . Forest Sci . 10: 432-442 . 

1965. Diurnal and seasonal patterns 

of net assimilation in Douglas-fir, a s 

influenced by environment . Ecology 46 : 

698-708 . 

1970. Summer net photo - 

synthesis of ponderosa pine in its natura l 

habitat. Photosynthetica 4(3) : 243-253 . 

Hinckley, T . M. 1971 . Plant water stress an d 

its effect on ecological patterns of behavior 

in several Pacific Northwest conifer species , 

177 p . Ph .D . thesis on file, Univ . Wash . , 

Seattle . 

and G . A. Ritchie . 1970 . Within - 

crown patterns of transpiration, wate r 

stress, and stomatal activity in Abies 

amabilis . Forest Sci . 16: 490-492 . 

222

Hodges, J . D. 1965. Photosynthesis in forest 

tree seedlings of the Pacific Northwest 

under natural environmental conditions . 


Seattle . 

1967. Patterns of photosynthesis 

under natural conditions . Ecology 

48: 234-242 . 

and D. R. M. Scott. 1968 . 

Photosynthesis in seedlings of six conifer 

species under natural environmental conditions. 

Ecology 49 : 973-980 . 

Huber, B . 1956. Die Gefassleitung. In W . 

Ruhland (ed .), Handb . Pflanzenphysiol . III : 

541-582 . 

Jarvis, P. G. 1971 . The estimation of resistances 

to carbon dioxide transfer . In Z . 

Sestak, J. Jatsky, and P . G. Jarvis (eds .) , 

Plant photosynthetic production, manua l 

of methods, p. 566-631 . The Hague : Dr. W . 

Junk N .V. Publishers . 

J. Catsky, F. E. Eckardt, W . 

Koch, and D. Koller. 1971 . General principles 

of gasometric methods and the mai n 

aspects of installation design . In Z . Sestak , 

J. Catsky, and P . G. Jarvis (eds.), Plant 

photosynthetic production, manual of 

methods, p . 49-110 . The Hague : Dr. W . 

Junk N .V . Publishers . 

Jeffree, C. E ., R. P. C. Johnson, and P . G . 

Jarvis. 1971 . Epicuticular wax in the stomatal 

antechamber of Sitka spruce and its 

effects on the diffusion of water vapour 

and carbon dioxide. Planta 98 : 1-10 . 

Keller, Theo. 1971. Der Einfluss der Stickstoffernahrung 

auf den Gaswechsel der 

Fichte. Allg. Forst- u . Jagdzt. 142(4) : 

89-93 . 

and J. Wehrmann . 1963 . C02 

assimilation, Wurzelatmung and Ertrag vo n 

Fichten and Kiefersamlingen bei unterschiedlichen 

Mineralstoffernahrung. Mitt . 

Schweiz . Anst. Forstl. Versuchsw . 39(4) : 

217-242 . 

Kinerson, R .,Jr. 1971 . Selected aspects of gas 

and energy exchange in a young, matur e 

Douglas-fir stand . 91 p. Ph.D. thesis on file , 


Kline, J . R ., M . L. Stewart, and C . F. Jordan . 

1972 . Estimation of biomass and transpiration 

in coniferous forests using tritiated 

water . In Proceedings-research on coniferous 

forest ecosystems-a symposium, p . 

159-166, illus. Pac. Northwest Forest & 

Range Exp. Stn ., Portland, Oreg . 

Koch, W. 1969 . Untersuchungen caber di e 

Wirkung von CO 2 auf die Photosynthese 

einiger Holzgewachse unter Laboratoriumsbedingungen. 

Flora 158 : 402-428 . 

Kozlowski, T., and Theo. Keller . 1966. Food 

relations of woody plants . Bot. Rev . 32(4) : 

293-382 . 

Kramer, P . J., and T . Kozlowski . 1960 . Physiology 

of trees. 642 p . New York : McGraw- 

Hill Co . 

Krueger, K . W., and W . K. Ferrell . 1965 . Comparative 

photosynthetic and respiratory 

responses to temperature and light by Pseudotsuga 

menziesii var. menziesii and var . 

glauca seedlings . Ecology 46 : 794-801 . 

and R . H. Ruth. 1969 . Comparative 

photosynthesis of red alder, Douglas - 

fir, Sitka spruce, and western hemloc k 

seedlings . Can . J. Bot . 47: 519-528 . 

Kunstle, E . 1971. Der Jahresgang de s 

CO 2 -Gaswechsel von einj ihrigen Douglasien-trieben 

in einem 20 ja"hrige m 

Bestand. Allg. Forst- u. Jagdzt. 142(4) : 

105-108 . 

Kvet, J ., and J. K. Marshall. 1971 . Assessment 

of leaf area and other assimilating surfaces . 

In Z . Sestak, J. Catsky, and P. G . Jarvis 

(eds.), Plant photosynthetic production , 

manual of methods, p . 517-555 . The 

Hague : Dr . W. Junk N .V. Publishers . 

Larcher, W. 1969a. Physiological approache s 

to the measurement of photosynthesis i n 

relation to dry matter production by trees . 

Photosynthetica 3(2) : 150-166 . 

1969b . The effect of environmental 

and physiological variables on the 

carbon dioxide gas exchange of trees . 


Lavender, D. P., and R. K. Hermann. 1970 . 

Regulation of the growth potential of 

Douglas-fir seedlings during dormancy . 

New Phytol . 69 : 675-694 . 

Leyton, L. 1958 . The mineral requirements of 

forest plants . In W . Ruhland (ed .), Handb . 

Pflanzenphysiol . IV : 1026-1039 . 

223

Lopushinsky, W . 1969. Stomatal closure i n 

conifer seedlings in response to leaf moisture 

stress . Bot . Gaz . 130 : 258-263 . 

Ludlow, M . M ., and P . J. Jarvis. 1971a . Methods 

for measuring photorespiration i n 

leaves . In Z . Sestak, J. Catsky, and P . G . 

Jarvis (eds .), Plant photosynthetic production, 

manual of methods, p . 294-315 . Th e 

Hague : Dr. W . Junk N .V . Publishers . 

and P . J. Jarvis. 1971b . Photo - 

synthesis in Sitka spruce . 1 . General characteristics 

. J . Appl . Ecol . 8(3) : 925-953 . 

McGregor, W . H. D ., and P . J. Kramer . 1963 . 

Seasonal trends in rates of photosynthesi s 

and respiration of loblolly pine and whit e 

pine seedlings . Am . J. Bot. 50(8) : 760-765 . 

Madgwick, H . A. 1964 . Estimation of surface 

area of pine needles with special reference 

to Pinus resinosa . J. For . 62 : 636 . 

Miillerstael, H . 1968 . Untersuchungen uber 

den Gaswechsel zweijahriger Holzpflanze n 

bei fortschreitender Bodenaustrocknung . 

Beitr. Biol. Pflanz. 44(3) : 319-341 . 

Parker, J. 1963. Causes of the winter decline 

in transpiration and photosynthesis in some 

evergreens. Forest Sci . 9 : 148-166 . 

Perrier, A. 1971 Leaf temperature measurement 

. In Z . Sestak, J . Catsky, and P . G . 

Jarvis (eds.), Plant photosynthetic production, 


Hague : Dr . W. Junk N .V. Publishers . 

Perry, T. O., and G . W . Baldwin. 1966 . Winter 

breakdown of the photosynthetic apparatus 

of evergreen species . Forest Sci. 12 : 

298-300 . 

Pharis, R. P., H. Hellmers, and E . Schuurmans. 

1967. Kinetics of the daily rate of 

photosynthesis at low temperature for two 

conifers. Plant Physiol. 42(4) : 525-531 . 

Phillips, R. A. 1967. Stomatal characteristic s 

throughout a tree crown . 67 p . M .S . thesis 

on file, Univ . Wash., Seattle . 

P i s e k, A . 1960. Die photosynthetische n 

Leistungen von Pflanzen besonder e 

Standorte : immergrune Pflanzen . In W . 

Ruhland (ed.), Handb . Pflanzenphysiol . 

V 2 : 415-459 . 

and R. Kemnitzer . 1968 . Der 

Einfluss von Frost auf die Photosynthes e 

der Weisstanne (Abies alba) . Flora 157 B : 

314-326 . 

, W . Larcher, W . Moser, and I . 

Pack . 1969 . Temperaturabhangigkeit an d 

optimaler Temperaturbereich der Netto - 

Photosynthese . Flora 158 B: 608-630 . 

, W . Larcher, I . Pack, and R . 

Unterholzner . 1968 . Temperaturmaximu m 

der Netto-Photosynthese and Hitzeresistenz 

der Blatter. Flora 158 B: 110-128 . 

Polster, H . 1967a . Wasserhaushalt . In H. Lyr , 

H. Polster, and H . -J . Fiedler, Geholzphysiologie, 

p . 147-196 . Jena : G . Fischer Verlag . 

1967b. Photosynthese, Atmung , 

and Stoffproduktion . In H . Lyr, H . Polster , 

and H. -J . Fiedler, Geholzphysiologie, p . 

197-246 . Jena : G . Fischer Verlag . 

Poskuta, J . 1968 . Photosynthesis, photorespiration 

and respiration of detached 

spruce twigs as influenced by oxygen concentration 

and light intensity . Physiol . 

Plant . 21 : 1129-1136 . 

Rangnekar, P . V ., D . F . Forward, and N . J . 

Nolan. 1969 . Foliar nutrition and wood 

growth in red pine : the distribution of 

radiocarbon photoassimilated by individual 

branches of young trees . Can . J. Bot. 47 : 

1701-1711 . 

Reed, K. L. 1968 . The effects of sub-zero 

temperatures on the stomata of Douglas-fir . 

79 p . M .S. thesis on file, Univ . Wash . , 

Seattle . 

1972 . A computer simulation 

model of seasonal transpiration in Douglas - 

fir based on a model of stomatal resistance . 

133 p. Ph.D. thesis on file, Oreg . State 

Univ ., Corvallis . 

Reed, Kenneth L ., and Warren L. Webb . 

1972 . Criteria for selecting an optima l 

model : terrestrial photosynthesis . In 


ecosystems-a symposium, p . 227-236 , 

illus. Pac . Northwest Forest & Range Exp . 

Stn., Portland, Oreg . 

Ritchie, G. A. 1971 . Transpiration, water 

potential and stomatal activity in relatio n 

to microclimate in Abies amabilis an d 

Abies procera in a natural environment . 


Seattle . 

Ross, Stephen D . 1972. The seasonal an d 

diurnal source-sink relationships for photo - 

assimilated 14 CO 2 in the Douglas-fi r 

224

anch. 170 p . Ph .D . thesis on file, Univ . 


Sestak, Z ., J . Catsky, and P . G . Jarvis (eds .) . 

1971 . Plant photosynthetic production , 

manual of methods . 818 p . The Hague : Dr . 

W. Junk N .V . Publishers . 

Slatyer, R . O . 1967 . Plant-water relationships . 

366 p . London : Acad . Press . 

Slavik, B . 1971. Determination of stomatal 

aperture . In Z . Sestak, J . Catsky, and P . G . 

Jarvis (eds .), Plant photosynthetic production, 


Hague : Dr . W . Junk N .V . Publishers . 

Stalfelt, M . G . 1956 . Die stomatare Transpiration 

and die Physiologie der Spaltoffnungen 

. In W . Ruhland (ed .), Handb . Pflanzenphysiol. 

III : 351-426 . 

1960 . Die Frage der begrenzenden 

Faktoren . In W . Ruhland (ed .), Handb . 

Pflanzenphysiol . V 2 : 213-254 . 

Stephens, G . R . 1969 . Productivity of red 

pine. 1 . Foliage distribution in tree crowns 

and stand canopy . Agric . Meteorol . 6 : 

275-282 . 

Thompson, R . B., and L . Leyton . 1971 . Method 

for measuring the leaf surface area of 

complex shoots . Nature 229 : 572 . 

Tranquillini, W. 1969 . Photosynthese and 

Transpiration einiger Holzarten bei verschieden 

starkem Wind . Centralbl. fur da s 

Gesamte Forstw . (Wien) 86(1) : 35-48 . 

Ungerson, J ., and G . Scherdin . 1965 . Untersuchungen 

uber Photosynthese and 

Atmung unter nattirlichen Bedingungen 

wa`hrend des Winterhalbjahres bei Pin u s 

silvestris, Picea excelsa, and Juniperus cornmunis 

. Planta 67 : 136-167 . 

Waggoner, P . E ., and N . C . Turner . 1971 . 

Transpiration and its control by stomata i n 

a pine forest . Bull . Conn . Agric. Exp. Stn . 

726, 87 p . 

Waring, R . H ., and B . D . Cleary . 1967 . Plant 

moisture stress : evaluation by pressur e 

bomb . Science 155(3767) : 1248-1254 . 

, K. L . Reed, and W. H. Emmingham 

. 1972 . An environmental grid fo r 

classifying coniferous forest ecosystems . In 

Proceedings-research on coniferous fores t 

ecosystems-a symposium, p . 79-91, illus . 

Pac . Northwest Forest & Range Exp . Stn . , 


Woodman, James N . 1968 . The relationshi p 

of net photosynthesis to environment with - 

in the crown of a large Douglas-fir tree . 18 8 

p . Ph .D . thesis on file, Univ . Wash ., Seattle . 

1971 . Variation of net photo - 

synthesis within the crown of a large forest - 

grown conifer . Photosynthetica 5(1) : 

50-54 . 

Zelawski, W ., and J . Kucharska . 1967 . Winter 

depression of photosynthetic activity i n 

seedlings of Scots pine (Pinus silvestris, L .) . 

Photosynthetica 1(3-4) : 207-213 . 

Zimmermann, M . H ., and C . L. Brown . 1971 . 

Trees : structure and function . 336 p . New 

York : Springer-Verlag Inc . 

225



Criteria for selecting an optimal model: 

terrestrial photosynthesis 

Kenneth L . Ree d 



Seattle, Washington 98196 

an d 

Warren L . Web b 

School of Forestr y 


Corvallis, Oregon 9733 1 

A bs tract 

In theory, there exists an infinite number of models of a given system. These models differ in resolution, 

scope, descriptive and predictive power. Because the models described in the literature generally reflect th e 

personal perspectives and goals of the researchers who developed the models, it is helpful to be able to evaluat e 

the potential of a given model to provide the information needed in another research problem . Several criteria 

for selection of models or modeling approach are suggested in this paper. The use of the criteria is illustrated by 

discussion of several mathematical models of photosynthesis taken from the literature . The models discussed 

include a linear regression model, an energy budget model, and models based on theory of enzyme kinetics and 

gas exchange . 

Introduction 

A great many models are available that de - 

scribe photosynthesis, some of which will be 

discussed below . Because of the many models 

now available, we had hoped that we woul d 

not have to develop a new model-that w e 

need only choose or modify an existing model . 

Thus it was necessary to consider certain criteria 

by which to judge the applicability of a 

given model to the purposes of the IBP . 

Several of the criteria will be discussed be - 

low. This list is probably not complete, but i s 

suggested as a guide to model selection, modeling 

approach, or both . Because most scientists 

are modelers in a broad sense, this discussion 

is intended primarily for those scientists 

who are least familiar with formal aspects 

of modeling . 

Criteria for Model Selectio n 

Because plant growth and presumably community 

composition are greatly influenced b y 

photosynthetic capacity, that phenomeno n 

has been the object of a great deal of study . 

Photosynthesis is only one component of the 

biochemical system resulting in plant growth . 

Failure to consider other factors in growth, a s 

well as the fact that photosynthesis can be 

uncoupled from growth, results in the general 

failure of attempts to predict growth from 

photosynthesis. However, a model of photosynthesis 

is an important component of a 

growth model and because the process o f 

photosynthesis is a physical-chemical system , 

attempts to model photosynthesis should b e 

consistent with some basic tenets of systems 

theory as well as physical reality . 

227

Systems 

A system, as defined by Klir (1969), is imposed 

upon an object-a segment of nature 

(the earth, a community, a tree, an automobile, 

a computer)-by the observer from a distinct 

point of view. Everything that does not 

belong to the object is the environment . The 

boundary between the object and its environment 

cannot be clearly defined ; thus th e 

delimitation of the object is somewhat arbitrary 

and reflects the personal perspectives o f 

the modeler . Because we usually cannot stud y 

an object in its entirety (because of its complexity), 

we observe or measure values of certain 

quantities . The choice of quantities to 

measure depends on what we consider to b e 

of interest or important to the given purpos e 

(Klir 1969) . 

Most scientists are familiar with the concept 

of a system, especially as it is applied i n 

thermodynamics. For example, the laws of 

thermodynamics were derived by studyin g 

idealized closed systems in equilibrium . In 

dealing with such closed systems it is possible 

to study independent state variables such as 

pressure, temperature, and volume, which de - 

fine the state of the system. A closed thermodynamic 

system can exchange heat and work 

but not matter with its environment (Daniels 

and Alberty 1967) . 

As Ludvig von Bertalanffy (1969) pointe d 

out, biological systems are open system s 

whose structure is maintained by energy, in - 

formation, and matter flow through the systems. 

The behavior of the system is the resul t 

of the interactions of the elements of th e 

system and the flow through the system . Because 

the thermodynamic system is closed 

and idealized, and because the state variable s 

are independent (i .e., one can be varied with - 

out affecting another), it is convenient and 

correct to describe the system in terms o f 

total differentials; for example, the relation 

of internal free energy, E, of a system to 

temperature and volume, T and V, can be expressed 

(Daniels and Alberty 1967) : 

E = f(T, V) 

dE _ 

(3T 

) V dT + () T dv (1) 

Equation 1 implies that the variables E, T, 

and V are independent ; one can be varied 

while the others are held constant, but mor e 

importantly, equation 1 implies summativity 

in the system where the behavior of a summa - 

tive system is the physical sum of the behavior 

of the parts . Because this approach is useful 

in thermodynamics and other special 

areas, biologists often attempt to explai n 

observed phenomena by assuming that the 

elements of all systems are independent (vo n 

Bertalanffy 1969) . We in plant physiology 

have accepted as a standard procedure th e 

isolation of a plant in a growth chamber, an d 

the study of the response of the plant to on e 

factor while holding the others constant . This 

approach is epitomized by Cleary (1970) who 

derived the following model of photosynthesis : 

the n 

P = f(M,T,L,N,Pr) 

dP= (3P) dM y {('r,) dT 

am T,L,N P dT M .L.N.P 

+ . 

. + dpr~ dP r 

1, 7'. L,N 

(2 ) 

where P = photosynthesis, M,T,L,N, and Pr 

are moisture, temperature, light, nutrition an d 

preconditioning effect, respectively . This 

model assumes a summative system analogou s 

to a closed thermodynamic system . 

This assumption is not totally valid in a 

biological system like photosynthesis . Light , 

for example, has an effect not only on the 

photochemical reactions within the leaf, but 

also is converted to heat, which influences 

leaf temperature. Likewise, leaf temperature 

is affected by transpiration rates, which is in 

turn affected by temperature, moisture status 

of the leaf and air, and so on, but equation 2 

does not account for these interactions. While 

this model is inadequate in that respect , 

Cleary (1970) does recognize that photo - 

synthesis is not a simple light-and-temperatur e 

related phenomenon but represents a complex 

interaction of diverse factors, all of whic h 

should be taken into account in order to full y 

understand the process . Some of the factor s 

listed by Cleary (1970) probably are independent, 

namely, Pr and N. 

228

Von Bertalanffy (1969) calls nonsummative 

systems Gestalten ; for example, the nonadditivity 

of mixing equal volumes of concentrated 

sulfuric acid and water. Von Bertalanffy 

provides another simple example where he 

points out that the voltage of three isolated 

conductors would be different from that i f 

they were interconnected . Biological quantities 

often act very much like the charge i n 

von Bertalanffy's example in that their values 

in concert are different from their isolated 

values . Thus, it is essential for a biological re - 

searcher to understand the distinction between 

summative systems and Gestalten . It is 

advisable to consider a biological system to b e 

a Gestalt unless it can be proved otherwise . 

We do not believe that the concepts of systems 

can be overemphasized . It is not necessary 

for each researcher to be adept at formulating 

systems models, but he should be awar e 

that few things in a biological system operat e 

independently, and that it is difficult if no t 

impossible to understand biological phenomena 

if we simply take the one factor-on e 

response approach. As a result, in modelin g 

biological systems, the single most importan t 

concept is that the model must be a syste m 

model, consistent with the concepts of 

systems theory . 

Theoretical Validit y 

We view a model as a tool to aid in under - 

standing and prediction . Because there are, i n 

theory at least, an infinite number of model s 

that can apply to a given system, we must 

select that model which provides the greatest 

understanding and predictive power withi n 

certain limitations . If our system of interest 

can best be explained by physical principles , 

then the model should have physical validity , 

but there are systems which are not explain - 

able from the physical paradigm, for example , 

animal behavior . In other cases, the physica l 

theory is inadequate and the approach may b e 

fruitless . 

Given a physical-chemical system, like 

photosynthesis, a physically valid model is usu - 

ally a good choice . As an almost trivial example, 

photosynthesis, P, has been expressed b y 

Chartier (1966) as a function of light : 

P 1 bL (3 ) 

where a/b is P at light saturation, and a is the 

slope of the curve (fig . 1) at zero light since 

a = dP/dL(1 + bL) . Lommen and coworkers 

(1971) express the same phenomenon as : 

PmL 

P m 

(L) = 1 + K L /L 

where PmL is photosynthesis at light and C O 2 

saturation and KL is that light intensity at 

which Pm(L) = 1/2 PmL (fig. 2) . 

a/s 

P 

Figure 1 . Photosynthesis as a function of light . From 

Chartier (1966) A = slope of line O,X . Descriptio n 

in text . 

PM (L) 

PML 

------------------ - 

K 

(4) 

Figure 2 . Photosynthesis as a function of light . Fro m 

Lommen et al.(1971) . Description in text . 

L 

L 

229

It is obvious that in essence equation 4 i s 

identical to equation 3, where PmL = a/b an d 

K = 1/b . Both models describe the light dependence 

of photosynthesis equally well ; on e 

model is not more complex than the other . 

Yet, we would choose equation 4 because th e 

parameters have more physical meaning . Th e 

curve could also be described by : 

P=00 + a 1L+ 32 L 2 + 03 L3 . . . (5) 

which has even less appeal because it is nearl y 

impossible to determine the physical meanin g 

of the parameters ((3 i). 

It could be argued that equation 4 is littl e 

more valid than equation 3. Equation 4 is 

derived from the Michaelis-Menton equatio n 

that describes the rate of a single enzymati c 

reaction as a function of a substrate concentration 

. Photosynthesis is not a true Michaelis- 

Menton case because it is an integration o f 

photochemistry and a chain of enzymati c 

reactions . The reductionist might argue that 

we should model photosynthesis as a functio n 

of the photochemical and the enzymatic reaction 

rates . Further, it is obvious that temperature, 

substrate availability, and several othe r 

factors are important in photosynthesis. Also , 

since equation 4 considers only light an d 

(indirectly) CO 2 effects, it falls short of ou r 

first criterion, that the model be a system 

model insofar as possible. These points brin g 

up our third criterion : resolution level . 

Resolution 

According to Klir (1969) every quantity w e 

observe must be determined in space . That is , 

CO 2 concentration, incident radiant energy , 

and temperature may be specified at som e 

point in space . This specification may be irrelevent 

for some studies (e .g., the location of 

a laboratory may be trivial with respect to th e 

study) but when concerned with a weathe r 

forecast, it would be important to identif y 

the spot on earth where temperature is measured. 

We may also specify the interval and 

accuracy of our measurements, giving the 

space-time resolution level (Klir 1969) . 

It may be convenient to think of the spatial 

resolution level as the hierarchical level of 

resolution (von Bertalanffy 1969) . Biological 

systems can be placed in hierarchies corresponding 

with levels of organization (table 1) . 

0 . ?mo 

Table 1 .-Levels of organization 

1. Subatomi c 

2. Atomi c 

3. Molecular 

4. Subcellular 

5. Cellular 

6. Organ or tissue 

Thus a system can be thought of as a component 

in a larger system, and composed of sub - 

systems (von Bertalanffy 1969) . Each level of 

organization seems to have certain properties 

unique to that level, making prediction o f 

system behavior from the viewpoint of th e 

lower systems somewhat hazardous . This concept 

is implicit in the idea of nonsummativit y 

of systems. The reductionist might argu e 

against this view, but the limits of reductionism 

are obvious . It is impossible to predict 

all the behavior of a tree from the molecular 

level . If we hope to discover general ecosystem 

principles comparable to the fundamental 

principles of physics, probably we wil l 

have to emphasize study from the ecosyste m 

level. The ideal gas law, for example, was 

derived from observations of the behavior o f 

10 23 particles acting in concert. In ecosyste m 

study, we are among the particles; a holistic 

view is needed . 

There are three levels of resolution under 

consideration for the photosynthesis model : 

the leaf, tree, and stand levels . Each level will 

require a different model and approach ; we 

cannot extrapolate the leaf level model to a 

tree by multiplying the results by the numbe r 

of leaves in the crown . The temporal resolution 

of the models will be somewhat flexibl e 

at the leaf level, but will probably be on a 

daily basis for the stand model . 

Practical Considerations 

7. Organism 

8. Community 

9. Global 

10. Solar Syste m 

11. Galactic 

12. Universe 

13. ? o 

There are several practical consideration s 

230

which may be of importance in model selection 

or approach . Some nonlinear systems of 

differential equations are extremely difficult 

to solve and their solution may consume a 

great deal of computer time . Parameter estimation 

is often a significant problem in man y 

models. Some models require highly precis e 

measurement of input variables, reducin g 

their practical applications, but their theoretical 

contributions may be important . On the 

other hand, we sometimes settle for use of a 

linear regression model when the system is 

poorly understood and it is much easier t o 

measure the variables than to determine their 

exact interrelations . Such models are useful in 

science, but have some built-in dangers, e .g . , 

the hazard of extrapolation of linear regression 

models . 

One danger of reliance upon regressio n 

models is that the confidence in the model i s 

greatest at the mean and diminishes at th e 

extremes. In fact, it is possible to fit a curvilinear 

function to data of a given range, only 

to have the model become totally inadequate 

at the extremes . 

For example, net photosynthesis as a function 

of temperature can be described as a 

symmetrical quadratic (Pisek and Winkle r 

1958, Webb 1972) (fig. 3) . 

values. Thus equation 6 is valid only in th e 

range of temperature within which the 

parameters were estimated . 

The failure of a model to predict syste m 

behavior at extremes is not necessarily fatal ; 

such models are common in physics, e .g., the 

ideal gas law and Newtonian physics . It i s 

necessary to understand the limitations of 

one 's models to know when the system deviates 

from the model . 

In choosing a model, it is important t o 

understand the assumptions and limitations of 

the model . For example, it is questionable to 

describe a biological system with a model tha t 

assumes a closed reversible system . Th e 

assumptions implicit in the model should be 

compatible with present knowledge of th e 

system of interest . 

This is not to say that models of different 

systems cannot be used to model another 

system. The idea of isomorphism of system s 

models is central to general systems theory 

(von Bertalanffy 1969) . Thus thermodynami c 

models, for example, may be of great utility 

in certain biological systems models . It re - 

mains for the researcher to be sure that the 

systems are isomorphic so that such model s 

are applicable . 

Pn = X30 + R1 T - 32 T2 ( 6) 

But extrapolation of the curve in either direction 

will lead to progressively more negativ e 

G(T) 

Examination of 

Some Models 

of Photosynthesis 

The discussion above dealt with criteria fo r 

selection of models . It would be useful to discuss 

some models described in the literature . 

Leaf Environment Mode l 

a 

Figure 3 . Net photosynthesis, G(T), as a quadrati c 

function of temperature . 

T 

Botkin (1969) simulated photosynthesis in 

an open oak-pine forest near Brookhave n 

National Laboratory . He developed a linea r 

model of net photosynthesis : 

Pn=c 0 + 01 T+ (3 2 ln S 

+ Q3 (ln s) 2 + Q4 T2 + (3 5 T 1n S (7) 

231

where T = leaf temperature, ° K 

S = solar radiation, gcal cm-2 min 

Pn = net photosynthesis rate , 

mg CO 2 (g dry wt)-1 hr 1 

(3i = parameters to be estimate d 

Equation 7 was derived from equation 8 and 

equation 9 : 

P(L)=a+b 1nS (8) 

which had been used to fit photosynthesis dat a 

from several herbaceous species (Blackman an d 

Rutter 1946, Blackman and Wilson 1951) . 

Equation 8 was coupled with equation 9 suggested 

by data of Pisek and Winkler (1958) , 

Krueger and Ferrell (1965), and others : 

G(T) = a + bT + cT2 (9 ) 

Data from two species of oak were obtained 

from infrared gas analysis and wer e 

used to estimate the parameters of equation 7 

by stepwise multiple regression analysis . Som e 

of the terms of equation 7 were nonsignificant, 

and were thus discarded . The final 

model was : 

Pn =Qo +(3 1 T+(3 2 In S+(3 5 T In S (10 ) 

Having estimated the parameters of equation 

10, the model was used to predict photo - 

synthesis of oak in the field . Their model gav e 

fair to good agreement with subsequentl y 

measured net photosynthesis . 

In terms of the criteria discussed above , 

equation 10 is generally inadequate . It doe s 

allow for interaction between two variables but 

fails to take into consideration other factor s 

that affect photosynthetic rate (stomatal behavior, 

micrometeorological conditions, plan t 

nutrition) . The model does satisfy the requirement 

that the Pn model be solved as a functio n 

of systems variables, that is, Pn = f [S, T] . 

In terms of the other criteria, the model fall s 

short of having a great deal of theoretica l 

validity in that the function is one of convenience 

rather than having general physicalchemical 

meaning . It is unnecessary to discuss 

in detail Botkin's model with respect to the 

other criteria . Like most photosynthesis models, 

it is specified at the leaf level, where th e 

inputs to the leaf are measured, and the 

relations between the leaf subsystems are 

inferred . 

While his model has many failings, it doe s 

have one virtue. If one were interested in 

comparing Pn = f[T,L] in two species under 

identical conditions, this model may be adequate, 

given that some other factor, e .g . , 

stomatal resistance, is not limiting . The fact 

that the parameters of the model were estimated 

by least-squares allows the use o f 

statistical tests useful in comparing such data . 

Energy Budget Mode l 

Idso and Baker (1968) based their model of 

photosynthesis and their earlier model (Ids o 

and Baker 1967) on that of Gates (1965 ) 

which is an energy budget model where precis e 

measurement of incoming radiant energy is 

equated with outgoing energy . Thus, at equilibrium, 

the amount of energy leaving a leaf i s 

equal to that coming in . Gates ' (1965) leaf 

energy budget model is : 

where 

(1 +r) (S + s) (Rg' +R a) 

as 2 +a t 2 (11 ) 

-e taTZ ± C±LE= 0 

a s = mean total absorbance of plant to 

sunlight and skyligh t 

r = reflectance of underlying ground 

or plane surface to sunlight and 

skylight, S and s 

at = absorbance of plant to thermal 

radiation, Rg and Ra 

e t = emissivity of plant to thermal 

radiatio n 

Tl = leaf temperature ° K 

C = energy gained or lost by convection 

cal cm-2 min- 1 

L = latent heat of evaporation ca l 

cm-2 miri- l 

E = transpiration rate of leaf gm cal - 2 

min- 1 

S, s, Rg, Ra = various short and lon g 

wave radiation inputs ca l 

cm-2 min 1 

232

The first three terms of equation 11 represent 

the radiant energy available to the plant . 

By elaborating equation 11 and solving for Tl , 

Gates was able to predict leaf temperature as 

a function of the incoming energy less th e 

outgoing energy . From data of other worker s 

he developed a family of curves for photo - 

synthesis with respect to temperature and radiant 

energy . Given these curves and by calculating 

Tl, Gates predicted photosynthesis of 

several species . 

Idso and Baker (1967) constructed a similar 

family of curves for sorghum. In a subsequent 

paper (Idso and Baker 1968), they used 

these curves to predict Pn on four different 

types of days based on input from their 

energy budget calculations . They did not 

measure Pn directly, so there was no direct 

verification of the models . 

In terms of our suggested criteria fo r 

analysis of models, it is obvious that th e 

energy budget model is consistent with th e 

systems approach . However, the relation of Tl 

and absorbed energy to Pn is empirical an d 

incomplete, and constitutes the weak point o f 

this model . Thus, from a standpoint o f 

theoretical validity, the Pn model suggeste d 

by Idso and Baker (1967, 1968) suffers fro m 

drawbacks identical to those of Botki n 

(1969). Should a more satisfactory model o f 

Pn as a function of light and temperatur e 

become available, however, it can be incorporated 

into the leaf energy budget model . The 

energy budget model itself is sound and has 

great promise . It is probably sufficiently 

general to apply to any plant, and the variou s 

terms can be refined where more specifi c 

component models are needed . 

Unfortunately, from a practical standpoint , 

the input data necessary for such a model ar e 

very difficult to obtain, requiring a great deal 

of sensitive instrumentation. This limitatio n 

would preclude the applicability of the mode l 

to study of some ecosystems where suc h 

measurements would be most difficult o r 

impossible . Further, the resolution of th e 

model is great, and such variables as inciden t 

solar angle and leaf geometry are important . 

Extrapolation of this model to a tree or to a 

stand would require prodigious volumes o f 

input data and highly sophisticated subsystem 

models . 

In summary, the high-resolution energ y 

budget approach, while having considerabl e 

appeal from the systems and theoretica l 

standpoint, is probably limited in its utility t o 

theoretical studies and studies of simple r 

systems . The basic approach could be modified 

by altering the resolution level of the 

components and by perhaps adding som e 

stochastic models . In this way, the energy 

budget approach could be used in modeling a t 

a lower level of resolution . Such a model , 

being simpler, could then be used at the tre e 

and stand levels of organization . 

While the work of Gates (1965) and hi s 

colleagues has contributed greatly to th e 

understanding of the energetics of the leaf, it 

still remained to develop a model of the 

process of photosynthesis itself that could 

satisfy our second criterion, that of theoretical 

validity . This problem was attacked by 

Brown (1969) . 

Process Models 

Brown assumed that if Pn is a first-orde r 

reaction, then equation 12 would hold : 

where P = rate of CO 2 

leaf area 

P = KI[CO 2 ]g (12 ) 

exchange per uni t 

K = capacity of acceptor site in th e 

leaf to fix CO 2 at the site of 

photosynthesis 

I = incident radiatio n 

[C0 2]g = concentration of CO 2 at the fixation 

site 

Brown also assumed that I follows Beer' s 

law within the leaf , 

x = Ie" ax (13 ) 

where I is incident radiation at the leaf 

surface, a is the extinction coefficient, x is 

distance from the leaf surface, and Ix is 

radiation at x . Brown incorporated equation 

12 into Gaastra's (1959) derivation from 

Fick 's law of diffusion representing steadystate 

diffusion of CO 2 from the air to the 

leaf : 

233

where 

giving 

P = C a Cc 

Er 

(14 ) 

Ca = concentration of atmospheri c 

CO 2 g cm- 3 

Cc = concentration of CO 2 at the 

chloroplast g cm 3 

Er = resistance to gas diffusion 

sec cm- 1 

D[CO 2] a 

P= 1 + D/KI 

where D is the integral exchange coefficien t 

D= 1 = 1 

ra +rs +rm E r 

where ra is boundary layer resistance, rs is 

stomatal resistance, and rm is mesophyll 

resistance to CO 2 diffusion in sec cm-2 . 

Equation 15 is gross photosynthesis, not Pn , 

and thus equation 15 should be corrected for 

respiration : 

Pn = K[CO 2 /al - R 

(16) 

I + K1/D 

where R is respiration . 

These models assume that resistance t o 

transfer of CO 2 from respiration site to the 

photosynthesis site is negligible . The variable 

nature of rs must also be considered, requirin g 

an additional model of stomatal behavior . 

Brown (1969) notes that the complexity of 

equation 16 becomes great when other components 

are added, and states that it is 

necessary to solve equation 16 by arriving at 

the functional relationships of the components 

by independent means and substituting 

them into the general model . 

If the components are truly independent , 

such a tactic is valid . If not, error is introduced 

by failing to consider changes in th e 

values of the parameters due to interaction, a 

violation of our systems criterion . 

Lommen and coworkers (1971) took an 

approach similar to that of Brown (1969) . 

They began with Gaastra 's (1959) derivatio n 

from Fick's law of diffusion now familiar to 

us. The biochemical relations are described b y 

a Michaelis-Menton type equation that de - 

scribes the rate of a single enzymatic reaction . 

Their model for photosynthesis as a function 

(15) 

of chloroplast concentration of CO 2 is : 

Pm 

P 1 + K/Cc 

where P = photosynthesis g cm-2 sec- 1 

Pm = photosynthesis at saturating CO 2 

g cm-2 sec- s 

K = concentration of CO 2 at the 

chloroplast when P = 1/*Pm 

Cc = concentration of CO 2 at the 

chloroplast g cm- 3 

Equation 14 is solved for Cc and substitute d 

into equation 17 giving photosynthesis as a 

function of atmospheric CO 2 concentratio n 

and stomatal resistance : 

(Ca + K + E rPm) 

P= 

2Er 

(18) 

- [(Ca + K + ErPm)2 - 4CaErPm] lie 

21 r 

They also derived a similar though mor e 

complex model giving photosynthesis as a 

function of Ca, K, Pm, W, and two series of 

resistances, S 1 , S 2 ; W is respiration . 

They also derived submodels of photosynthesis 

as a function of light and temperature , 

Pm(L,T) = 

PmLT G(T) 

1 + KL/L 

(17 ) 

(19 ) 

similar to equation 4 but incorporating a term 

G(T) representing the temperature dependence 

of photosynthesis . The authors implie d 

that equation 19 could be incorporated int o 

equation 18 and their other model, but the y 

did not do so, and dimensional analysis o f 

equation 18 with PmLT and KL substituted 

for Pm and K, shows that such a substitutio n 

is physically unrealistic. Hence the model o f 

Lomrnen et al . (1971) is inadequate for 

predicting photosynthesis as a function o f 

light and temperature as well as CO 2 concentration 

and stomatal resistance . 

Chartier (1966, 1969, 1970) also developed 

a complex model of net assimilation derive d 

along the lines of that by Lommen et al . 

(1971) . Beginning with the fundamental for m 

(Chartier 1970) : 

234

P = F+ R 

where F = net assimilation rate 

per unit leaf area 

R = respiration rate 

and by analogy to Gaastra's (1959) Fick's la w 

function, he derived : 

C-F(ra +rs +rm)-nRrm 

F + R = 1 

aE C-F(rQ +rs +rm)-nRrm +rx 

(20) 

The only terms unfamiliar to us are rx, n, aE, 

and R . Here R is respiration in light, and n is 

the fraction of respiratory flux that is mixe d 

in the intercellular spaces (n < 1) . The term 

aE represents conversion of light to photosynthate 

where E is incident light energy, an d 

a is the efficiency of light energy conversion . 

The term rx represents resistance to carboxyla - 

tion, a parameter which includes biochemical 

restraints caused by mineral nutrition, age o f 

the leaf, etc . 

Chartier's (1970) model differs from tha t 

of Lommen et al. (1971) in that the effect of 

light is incorporated into the model . Th e 

effect of temperature must be included, perhaps 

by a multiplicative term (Lommen et al . 

1971, Webb 1972) . Chartier's model gives a 

quadratic solution for F, as in Lommen et al , 

but many of the terms in both models ar e 

difficult if not impossible to measure in a 

field study . Further, respiration is represente d 

by a single term ; an oversimplification requiring 

further work . 

Conclusions 

In order to be consistent with the systems 

viewpoint, photosynthesis must be treated as 

a Gestalt, or nonsummative system. The 

models described above violate this criterion 

to some extent, most often by failing to 

incorporate an important factor in the model . 

We had hoped to be able to use one of the 

models by Lommen et al. (1971) or by 

Chartier (1970) as a tool in our field research , 

but two considerations prohibit this : (1) both 

models have unmeasurable (in the field ) 

terms, (2) neither model is complete, i .e . , 

neither expresses photosynthesis as a function 

of all the known important factors . 

Consequently, it will be necessary to develop 

a model of photosynthesis as a function 

of light, temperature, leaf resistance and 

ambient CO 2 concentration with some simplifications 

from the above models which will 

increase the utility of the models. The 

parameters in this model will be estimate d 

from data after the manner of Webb (1972 ) 

by nonlinear least-squares . The final model 

will have much the same utility as a regressio n 

model, but will not be linear and th e 

parameters where possible will have physical 

meaning. Thus, the models will be develope d 

with certain specific goals in mind, necessitating 

development of different models for th e 

tree and stand levels of resolution . 

A cknowledgments 

The work reported in this paper was 

supported by National Science Foundatio n 


Biome, U .S . Analysis of Ecosystems, International 

Biological Program . This is Contributio n 

No. 38 to the Coniferous Forest Biome, IBP . 


Blackman, G . E ., and A. J. Rutter. 1946 . 

Physiological and ecological studies in th e 

analysis of plant environment . I. The light 

factor and the distribution of the bluebel l 

(Scilla non-scripta) in woodland communities. 

Ann. Bot., N.S. 10: 361-390 . 

and G . L. Wilson . 1951. Physiological 

and ecological studies in the analysi s 

of plant environment . VI. The constancy 

for different species of a logarithmic relationship 

between the net assimilation rat e 

and light intensity and its ecological significance. 

Ann . Bot., N.S. 15: 63-94 . 

Botkin, D . R. 1969 . Prediction of net photo - 

synthesis of trees from light intensity and 

temperature. Ecology 50: 854-858 . 

Brown, K . W. 1969 . A model of the photosyn - 

thesizing leaf . Physiol . Plant 22 : 620-637 . 

Chartier, P. 1966. Etude theorique de 1 'assimi - 

lation brute de la feuille . Ann . Physiol. veg . 

8: 167-196 . 

235

1969. Assimilation nette d'un e 

culture couvrante_ II. La reponse de l'unite 

de surface de feuille . Ann. Physiol. Veg. 

11 : 221-263 . 

1970. A model of CO 2 assimilation 

in the leaf . In Prediction and measurement 

of photosynthetic productivity, 63 2 

p . Wageningen, The Netherlands : Centre 

for Agricultural Publishing and Documentation 

. 

Cleary, B . D. 1970. The effect of plant 

moisture stress on physiology and establishment 

of planted Douglas-fir ponderosa pin e 

seedlings. 85 p. Ph .D. thesis on file, Oreg. 

State Univ., Corvallis . 

Daniels, F ., and R . A. Alberty . 1967 . Physical 

chemistry . Ed . 3, 767 p. New York : Wiley . 

Gaastra, P. 1959. Photosynthesis of cro p 

plants as influenced by light, C0 2 , temperature, 

and stomatal diffusion resistance . 

Med. Landb . Hogesch . Wageningen, Th e 

Netherlands 59 : 1-68 . 

Gates, D. M . 1965. Energy, plants, an d 

ecology . Ecology 46 : 1-13 . 

Idso, S . B ., and D. G. Baker . 1967. Metho d 

for calculating the photosynthetic respons e 

of a crop to light intensity and lea f 

temperature by an energy flow analysis o f 

the meteorological parameters . Agron. J . 

59 : 13-21 . 

and D . G . Baker. 1968. The 

naturally varying energy environment and 

its effects on photosynthesis . Ecology 49 : 

311-316 . 

Klir, G. J. 1969. An approach to general 

systems theory . 323 p . New York: Van 

Nostrand-Reinhold Co . 

Krueger, K. W., and W . K . Ferrell. 1965 . 

Comparative photosynthetic and respiratory 

responses to temperature and light by 

Pseudotsuga menziesii var . menziesii and 

var . glauca seedlings . Ecology 46 : 794-801 . 

Lommen, P. W ., C. R . Schwintzer, C. S . 

Yocum, and D . M. Gates . 1971 . A model 

describing photosynthesis in terms of gas 

diffusion and enzyme kinetics. Planta 98 : 

195-220 . 

Pisek, A., and E . Winkler . 1958 . Assimilationsvermogen 

and Respiration der Ficht e 

(Picea excelsa Link) in verschiedene r 

Hohenlage and der Zirbe (Pinus cembra L .) 

and der alpinen Waltlgranze. Planta 51 : 

518-543 . 

von Bertalanffy, L . 1969 . General syste m 

theory . 289 p. New York: Geo. Braziller . 

Webb, W. L. 1972. A model of light and temperature 

controlled net photosyntheti c 

rates for terrestrial plants . In Jerry F . 

Franklin, L. J. Dempster, and Richard H. 

Waring (eds.), Proceedings-research o n 


p. 237-242, illus. Pac. Northwest 

Forest & Range Exp . Stn ., Portland, Oreg. 

236


Bellingham, WashingtonMarch 23-24, 197 2 

A model of light and temperature 

controlled net photosynthetic rates 

for terrestrial plants 

Warren L . Web b 

Forest Research Laboratory 

School of Forestr y 



Abstract 

Steady-state relative net CO2 exchange was modeled in terms of a temperature-dependent respiration function 

and light- and temperature-dependent photosynthesis function. The parameters of the model were 

evaluated using the laboratory CO 2 exchange data of a group of 40 red alder seedlings (Alnus rubra Bong.). Th e 

model is continuous and well-behaved in the temperature region of 0-50°C for light energy between 0 .0 and 1 . 0 

ly/min total short-wave radiation. 

Introduction 

Consumer populations depend upon the 

chemical energy that is converted from sola r 

energy by plants in the ecosystem . The rate of 

conversion, or net photosynthesis, is in tur n 

dependent upon the genetic information avail - 

able to each plant and its immediate environment. 

Many factors influence net photosynthesis 

but, as Schulze (1970) found from hi s 

work in a beech stand, when water stress is 

not appreciable, the radiation and temperature 

regimes of the plant largely regulate net 

photosynthesis . This paper presents an empirical 

model of steady-state net CO 2 exchange i n 

terms of a light (L) and temperature (T) con - 

trolled gross photosynthesis function (Ps) an d 

a temperature-controlled dark respiration (Rs ) 

function . 

Net CO 2 flux entering the leaf, or ne t 

photosynthesis (Psn), is conceptualized as th e 

difference between carbon fixed in the photo - 

synthetic process and that lost during respiration, 

Psn = Ps - Rs (Larcher 1969) . The two 

terms, net photosynthesis and net CO2 exchange, 

are used synonymously in this paper . 

Although the former is less general in that it 

usually applies only to CO 2 exchanges occurring 

in the light, it is more mechanically 

viable . 

Fluxes of CO 2 are easily measured and 

many investigators have reported on the ne t 

photosynthetic response of plants to temperature 

and light (Heath 1969, Rabinowitc h 

1969, Milner and Hiesey 1969) . Figure 1 illustrates 

the net photosynthetic response of 

small plants or individual leaves to increasin g 

temperature or light . The linear portion of the 

light curve represents the rate of the photo - 

chemical reaction at the chloroplast and is 

largely invariant among species (Rabinowitc h 

1969) . This linear slope can be extrapolated 

through zero CO2 flux, called the light compensation 

point, and the negative absciss a 

intercept interpreted as dark respiration 

(Chartier 1969) . At high light levels, photo - 

synthesis becomes saturated and its rate i s 

dependent upon factors such as temperatur e 

237

Figure 2. Effect of temperature on dark respiration . 

Figure 1 . Leaf CO2 exchange as related to light an d 

temperature . 

and water potential that influence the carbon - 

fixing enzymatic reactions . 

Net photosynthetic response to temperature, 

also shown in figure 1, is generally a 

symmetrical form (Pisek et al . 1969, Moone y 

and Harrison 1969). Pisek's data have show n 

that the slope of the curve and the point of 

maximum shifts depending upon species an d 

season but the response remains symmetrical . 

The responses to light and temperatur e 

seem well-defined. A model using light an d 

temperature as independent variables shoul d 

conserve the known responses to each as wel l 

as include any interaction . To be useful, th e 

model should also predict CO 2 losses that 

occur below the light-compensation point a s 

well as respiration losses occurring durin g 

darkness . This requires that a respiratio n 

response be explicitly introduced into th e 

model. 

Figure 2 represents a generalized dark respiration 

response to temperature for steady - 

state conditions. The response at low an d 

intermediate temperatures is exponential i n 

keeping with the van 't Hoff Q 1 o rule for 

chemical reactions (Forward 1960). Abov e 

some high temperature, a decline resultin g 

from enzyme denaturation occurs which 

tends to become irreversible as the time o f 

exposure to high temperature increases (For - 

ward 1960, Longridge 1963) . 

Net CO 2 uptake, or net photosynthesis , 

can now be expressed in terms of temperature-controlled 

respiration and light- an d 

temperature-controlled photosynthesis . For 

purposes of this model, net CO 2 fluxes int o 

the leaf are assigned a positive value whil e 

CO 2 losses are considered negative . 

Psn(L,T) = Ps(L,T)-Rs(T) (1) 

The functional form of the model can b e 

expanded around the following exponential 

expression that is representative of the ligh t 

curve in figure 1 . 

Psn(L,T) = Bo(T) (1 - exp(B 1 L)) - Rs(T) (2 ) 

Rs(T) is negative and represents dark respiration 

. Bo(T) represents the light-saturate d 

asymptote as a function of temperature . Th e 

exponential coefficient (B 1 ) determine s 

response behavior at low light levels and is 

characteristically independent of temperature . 

Bo (T) can be represented by a quadratic function 

of the following form : 

Where : 

B o (T) = Bo + B i (T-B ) 2 (3 ) 

Bo : maximum photosynthesi s 

B ; : slope coefficient ; negative algebraic sign 

B2 : temperature of maximum photosynthesi s 

Respiration as a function of temperature 

(Rs(T)) can be modeled with the followin g 

expression : 

Rs(T) = Boexp(BI (T-B.) - exp(BI (T-B . )) ) 

Bo /exp(1) : maximum respiratio n 

B1 : slope coefficient ; positive sign 

B 2 : temperature of maximum respiratio n 

238

This function is asymptotic to zero respiration 

at both low and high temperatures an d 

exhibits an exponential increase in respiratio n 

followed by a rapid decline at high 

temperatures . 

The completed functional expression fo r 

Psn(L,T) is : 

Psn(L,T) = -Boexp[B 1 (T-B2 )-exp(B 1 (T-B2 ))] 

+ [ B 2 + B3 ( T -B4 ) 2 ] [1 - exp(B s L ) ] 

(4) 

The intent of this model is to predict ne t 

photosynthesis in light and CO 2 losses in 

darkness. The model is not useful for predicting 

gross photosynthesis without including 

some function for respiration in the light . 

Until the experimental techniques for obtaining 

such measurements are developed, th e 

light respiration term must remain implicit i n 

the gross photosynthesis function . This limit s 

the use of the model although most C 3 plant s 

will respond similarly . A different model may 

be required for C 4 plants such as grasses that 

require a much higher light level to saturat e 

photosynthesis . 

Methods and Materials 

Values for the six parameters in the model 

were determined with a least squares fit of th e 

data using an algorithm developed by Marquardt 

(1963) . Net photosynthetic data were 

taken with a controlled environment syste m 

developed by Webb (1971) . The test organis m 

for the model was red alder (Alnus rubra 

Bong.) . 

Forty red alder seedlings, 1- to 2-years-old , 

were removed from the field and propagated 

in a nutrient culture before transferring them 

to the gas-tight controlled environment chamber. 

The root system of the seedlings was en - 

closed in a nutrient flow system with temperature 

controlled at 11°C ± 1°C . The plants 

were maintained in this system for 1 month at 

conditions of light and temperature consisten t 

with those in the greenhouse . 

Carbon dioxide absorption rates were measured 

by monitoring the CO 2 depletion in the 

gas-tight environment chamber with a Beckman 

IR gas analyzer. Atmospheric CO 2 levels 

in the chamber were maintained between 32 0 

and 335 p .p.m. Light energy was varied from 

0.06 to 0 .68 ly/min (total shortwave radiation) 

at air temperatures from 5 to 30°C . For 

each of five light levels, CO 2 uptake was 

measured while temperature was changed a t 

the rate of 1° per 5 minutes beginning at 

15°C and proceeding upscale to approximately 

30°C . Measurements were then made 

while decreasing temperature at the same rat e 

until near 5°C at which time temperature wa s 

increased again up to 16°C . All CO 2 uptak e 

measurements were made during 3 successiv e 

days. Relative humidity varied between 6 5 

percent at low temperatures to 75 percent at 

high temperatures . 

Results 

A portion of the data is plotted in figures 3 

and 4 . The photosynthetic response to radiation 

in figure 3 is characterized by a linea r 

response at low light followed by the usual 

light-saturated response at high radiation . Thi s 

is consistent with the findings of many other 

investigators . The saturation value for ne t 

photosynthesis increases with temperature , 

but the increase is not linear . 

Figure 4 shows the net photosynthetic 

response to temperature for constant radiation 

of 0 .19 ly/min . A quadratic function wa s 

fit to the data and the parameters with their 

respective standard errors are listed in figure 

4 . At 0.19 ly/min, net photosynthesis has a 

maximum of 20 .5°C. Although these data 

represent the temperature range between 5 .5 ° 

and 27°, the data of Pisek et al . (1969) indicate 

a nearly symmetrical response betwee n 

5° and 40°C. Extrapolation of the quadratic 

beyond this temperature range should b e 

done cautiously although Phillips an d 

McWilliams (1971) have measured zero CO 2 

fluxes at high temperatures . It may be that 

high temperatures increase respiration until it 

exceeds the photosynthetic capacity . 

Figure 5 illustrates the response surface 

generated by expression (4) which is the net 

239

w 10 0 

0 

U) 

w 80 

1- 

z 

>- 

u) 

0 60 

I-- 

0 

22 ° 

i6 ° 

Psn t Rs (TI + Bo (T1 (I - e B2y 

° 

0 

t- 40 

w 

z 

w > 

2 0 

J 

w 

cc 

0 

1 

0 .2 0 .4 0.6 0 .8 

TOTAL SHORTWAVE RADIATIO N 

Figure 3 . Light effects on net photosynthesis of red alder at three temperatures . 

0 O 

w 4 0 

z 

w 

>_ 

1- 20 

J 

w 

cc 

I 

10 20 30 40 

TEMPERATURE, ° C 

Figure 4 . Net photosynthetic response of red alder to temperature at 0 .19 ly/min total s.w . radiation . 

240

The important interaction between light 

and temperature can best be shown by numerical 

example . Table 1 shows the predicted 

temperature of maximum net photosynthesis , 

or optimum temperature, and the peaks of 

net photosynthesis for each of five levels of 

light. Both the optimum temperature and th e 

peaks of net photosynthesis increase with 

light. The latter is exemplary of photosynthetic 

light saturation . 

Table 1. Predicted effect of light on 

maximum net photosynthesis 

Light Relative Temperature of 

ly/min maximum Psn maximum Psn 

Figure 5 . Response surface of net CO 2 exchange generated 

by Psn(L,T) = -513.3 exp[ .088(T-47.7) - 

exp .088(T-47.71] + [187 .3 - .105(T-41 .41 2 1 

[1 - exp(-9.59L)] . 

photosynthetic response to both light an d 

temperature . The functional form of the surface 

is included in figure 5 along with th e 

values of the 6 parameters obtained from a 

least-squares fit of the CO 2 exchange data. 

The average deviation of the function fro m 

the data was ±9 .1-pct . and the r 2 was 0 .97 . 

Note that the light saturation phenomeno n 

of photosynthesis is conserved in this model 

as well as the symmetrical response to temperature. 

Although there are no data below 5 ° 

and above 30°, the model seems well-behave d 

in the regions outside the data. As light decreases 

the gross photosynthesis functio n 

tends toward 0, and the dark respiration function 

begins to dominate . Below the light 

compensation point CO 2 fluxes become negative. 

At zero light only respiration occurs, an d 

because the function goes asymptotically t o 

zero at both low and high temperatures, the 

model never predicts an uptake of CO 2 in 

darkness . 

0 .07 27 .8 1 6 

.15 65 .5 2 0 

.19 76 .2 2 1 

.25 90 .8 2 2 

.68 100 .0 23 

The increased temperature of maximu m 

net photosynthesis associated with additiona l 

light can be explained as an interaction between 

the gross photosynthesis function an d 

dark respiration . At low light, photorespiration 

is minimal and photosynthesis is largely 

independent of the temperature increases that 

step-up dark respiration . Therefore, dark 

respiratory CO 2 losses from cellular maintenance 

are not compensated by photosynthesis 

as temperatures rise and maximum net photo - 

synthesis (photosnythesis minus respiration ) 

occurs at a relatively low temperature . As 

light increases, photosynthesis and its depend - 

ence on temperature increases . This allows 

photosynthesis to keep pace with temperature-controlled 

dark respiration and maxi - 

mum net photosynthesis then shifts to a 

higher temperature . 

Although the model is empirical, it has 

several noteworthy features . The model is 

continuous over a wide range of independent 

variables and is therefore more mathemati- 

24 1

cally tractable than discontinuous or piece - 

wise continuous models . This continuous feature 

should facilitate linking this model t o 

others . 

The general net photosynthetic response of 

plants to light and to temperature has been 

demonstrated with many different specie s 

under various preconditioning treatments . 

The general responses characterized in figur e 

1 predominate with various shifts in the pea k 

and magnitude of the response curves . These 

shifts can be described mathematically by 

determining CO 2 exchange at certain point s 

and using these data to define parameters in 

the basic model . 

In the Biome program, CO 2 exchange 

studies are being conducted on several species 

and these data will further test the model pro - 

posed here. A necessary addition to the model 

is an explicit evaluation of light respiratio n 

from which techniques for data acquisitio n 

are now being contemplated . 


Thanks are due Michael Newton and Duan e 

Starr for their contribution to the model an d 

to Jack Colby for his programming assistance . 


ported by the PHS grant 2P10-Es00210 i n 

cooperation with the Coniferous Forest 





Chartier, P . 1969. A model of CO 2 assimilation 

in the leaf . In Prediction and measurement 

of photosynthetic productivity, p . 

307-315 . US/IBP Tech. Meet. Proc . , 

Trebon . 

Forward, Dorothy F. 1960 . Effect of temperature 

on respiration . In W . Ruhlan d 

(ed .), Encyclopedia of plant physiology, p . 

234-258 . Berlin : Springer-Verlag. 

Heath, O .V .S. 1969 . The physiological aspect s 

of photosynthesis . 309 p. London : Heinemann 

Educ . Books Ltd . 

Larcher, W. 1969 . Physiological approaches t o 

the measurement of photosynthesis in relation 

to dry matter production by trees . 


Longridge, J. 1963. Biochemical aspects of 

temperature response. Annu. Rev. Plant 

Physiol. 14 : 441-462 . 

Marquardt, Donald W . 1963 . An algorithm for 

least-squares estimation of nonlinear 

parameters. J. Soc. Ind. Appl. Math . 11(2) : 

431 . 

Milner, Harold W., and William H . Hiesey . 

1963. Photosynthesis in climatic races o f 

Mimulus . I. Effect of light intensive and 

temperature on rate . Plant Physiol . 39(2) : 

208-213 . 

Mooney, H. A., and A. T. Harrison. 1969 . The 

influence of conditioning temperature on 

subsequent temperature-related photosynthetic 

capacity in higher plants . In Prediction 

and measurement of photosynthetic 

productivity, p. 411-417 . US/IBP Tech . 

Meet. Proc., Trebon . 

Phillips, V . R., and J. R. McWilliams . 1971 . 

Thermal responses of primary carb o 

enzymes from C 3 and C 4 plant adapted to 

contrasting temperature environments . In 

M . D. Hatch, C . B . Osmund, and R . O . 

Slatyer (eds .), Photosynthesis and photorespiration, 

p . 97-109. Conf., Canberra, 

Australia, Nov. 1970 (Proc .). New York : 

Wiley-Interscience . 

Pisek, A., W. Larcher, W. Moser, and Ida 

Pack. 1969. Kardinale Temperaturbereiche 

der Photosynthese and Grenz-temperaturen 

des L e b ens der Blatter verschiedener 

S permato-phyten. III. Temperaturabhangegkeit 

and optimaler Temperaturbereich 

der Netto-Photosynthese. Flora 158 : 

608-630 . 

Rabinowitch, E . 1969 . Photosynthesis . 273 p . 

New York : John Wiley & Sons, Inc . 

Schulze, Ernst-Detlef. 1970. Der C02 - 

Gaswechsel der Buche (Fagus siluatica L . ) 

in Abhangrgkeit von den Klimafaktoren i m 

Freiland. Flora 159 : 177-232 . 

Webb, Warren L. 1971 . Photosynthetic response 

models for a terrestrial plant community, 

79 p. Ph.D. thesis on file, Oreg. 

State Univ., Corvallis . 

242



Energy flux studies in a 

coniferous forest ecosystem 

Abstract 

Lloyd W. Ga y 

Associate Professor of Forest Climatolog y 

Department of Forest Engineerin g 



The fluxes of thermal energy between the atmosphere and a young Douglas-fir forest were measured durin g 

two contrasting summer days, one cloudless and one overcast. The energy budget components were evaluated 

by the Bowen ratio method, with ceramic-wick psychrometers at the 26 .16 m, 28.16 m, or 31 .16 m levels. The 

maximum height of the tallest trees was 28 m, and the general level at the top of the closed canopy was abou t 

22 m. Daily totals of the energy budget components (cal/cm 2 ) under cloudless skies on July 29, 1971, were : 

solar radiation, 584; net radiation, 410 ; change in storage, 5 ; convection, -135; and latent energy, -280. The 

albedo was 0.09 on both the clear and the overcast day . Analysis of the overcast conditions of July 31, 1971 , 

yielded the following values: solar radiation, 171 ; net radiation, 134 ; change in storage, 6 ; convection, -39; and 

latent energy, -102. 

Problems of measurement and analysis are discussed. These include the storage term in the biomass, and th e 

small gradients of potential temperature and vapor pressure above the canopy . Clear day gradients at noon, for 

example, were in the order of -0.03°C m-' and -0.03 mb m-' . Techniques are presented for minimizing measurement 

errors. 

Introduction 

The level of biological activity at the surface 

of the earth is closely associated wit h 

cycles of energy and mass . The cycles of mas s 

and of energy are virtually interchangeabl e 

concepts. Indeed, the transpiration an d 

photosynthetic components of the mass cycl e 

can be studied through examination of th e 

cycle of energy, with the energy required t o 

change the phase of H2 0 and CO 2 serving as 

the connecting link . 

The magnitudes and phase relationships of 

the mass and energy cycles are affected by th e 

characteristics of the surface, and by the stat e 

of the atmosphere . The properties of vegetation, 

particularly of low, cultivated ecosystems, 

have been investigated thoroughly i n 

a variety of studies that have clearly demonstrated 

many advantages for energy budge t 

evaluations of transpiration and photosynthesis 

(Baumgartner 1965) . The advantage s 

include sensitivity, mobility, and the benefit s 

to be gained by use of a nondestructive technique 

. The application of these techniques t o 

the forest ecosystem appears feasible and useful. 

A number of studies have already been 

reported, but in general the effects of forest s 

upon the cycles are not yet well known 

(Baumgartner 1971, Tajchman 1971) . Coniferous 

forests, as a class, are good absorbers of 

solar radiation. The roughness of coniferou s 

crowns also appears to effectively enhanc e 

mixing in the atmosphere near the top of th e 

canopy. These factors, combined with the 

large surface area of canopies, make forests 

into very efficient exchange surfaces for water 

vapor, carbon dioxide, and energy . 

Studies of the fundamental cycles of energ y 

and mass have begun at the Cedar River site i n 

the Coniferous Forest Biome. A variety of 

interrelated studies are planned in coopera - 

243

tion with physiologists, soil scientists, hydrologists, 

and meteorologists . The wor k 

reported here is of preliminary research int o 

the exchange of thermal energy between a 

young Douglas-fir forest and the atmosphere . 

The energy exchange processes have bee n 

evaluated during two contrasting conditions : 

the first during clear, warm weather characterized 

by a large energy input to the forest , 

and the second during cool, overcast conditions 

with a relatively small energy input to 

the forest. The objectives are to define th e 

energy transfer characteristics of the youn g 

forest under conditions of both high and lo w 

rates of transpiration . Study of these contrasting 

conditions will help us to understand the 

processes that control the exchange of energy 

and water vapor between the forest canop y 

and the atmosphere . 

Experimental Methods 

The evaluation of thermal energy exchange 

in the forest is simple in concept, requirin g 

that appropriate boundaries be defined abou t 

the forest site, that the quantities of energ y 

crossing the boundaries be identified, and that 

appropriate measurements be made at periodic 

intervals . The periodic samples can the n 

be combined to yield estimates of energy ex - 

change for the desired time interval . Th e 

selection of an appropriate model is an important 

step in evaluating the thermal energy 

exchange . 

Energy Transfer Models 

Several review articles have discussed th e 

application of various energy transfer model s 

to the problem of evaluating the exchange between 

the forest and the atmosphere 

(Baumgartner 1965, Federer 1970, Fritsche n 

1970) . 

The basic problem concerns the transformation 

of energy by the forest from radiant 

to nonradiant forms . The total energy thus 

transformed is called net radiation, Q* ; this 

equals the quantity of radiation of all wavelengths 

absorbed at the surface, minus th e 

radiation lost by reflection and emission . The 

net radiation is transformed into either a 

change in stored heat in the soil and biomass , 

G; a flow of convective (sensible) energy between 

the forest and the air, H; or a flow of 

latent heat, AE, that is associated with a flux 

of water vapor E . The amount of energy 

transformed in photosynthesis, P, is of great 

importance in productivity . However, since i t 

amounts to only a few percent in terms of the 

net radiation flux over the forest (Denmead 

1969), P will be neglected in this study . 

The rate and direction of energy transfer 

depends upon the relative energy states of th e 

canopy and the atmosphere and upon th e 

availability of radiant energy which is derived 

primarily from the sun . The state of energy is 

determined by temperature and by vapor concentrations. 

Motion of the atmosphere may 

also enhance energy transfer . Energy transfer 

models thus use measurements of temperature, 

vapor concentration, wind, and radian t 

energy in order to determine the flow of 

energy between the forest and the atmosphere . 

The "Bowen ratio" model was selected for 

use in this study because of relative simplicity 

in analysis and application, and because of its 

general acceptance based upon tests over 

other types of vegetation . The method was 

first derived by Bowen (1926), and has been 

adapted for energy transfer studies by a 

number of workers (Fritschen 1965, Tanner 

1960) . 

The Bowen ratio model has been thoroughly 

described elsewhere, but a short discussion 

here will help to place its applicatio n 

into perspective . First, we must sum th e 

thermal energy fluxes active in the forest i n 

accordance with the principle of conservatio n 

of energy, to obtain 

Q*+G+H+XE+P=0 . (1) 

The polarity convention considers fluxes t o 

the surface as positive . After neglectin g 

photosynthesis, equation 1 can be solved for 

latent energy to yield 

XE = - (Q* + G) / (1 + (3) ( 2) 

where i3 is the Bowen ratio of convective hea t 

to latent heat (H/XE). The Bowen ratio can be 

written 

244

(3 = H/AE=yo$/De (3 ) 

where y is the psychrometric constan t 

(y ^ 0 .66 mb/°C at sea level) and Da and 

De represent the measured differences i n 

potential temperature and in vapor pressure at 

two levels in the atmosphere above, but near , 

the surface . 

The Bowen ratio model thus provides a 

rationale for partitioning the measured suppl y 

of thermal energy into convection and laten t 

heat, based upon measurements of temperature 

and vapor concentration at only two 

levels above the canopy surface. Wind measurements 

are not required for the analysis , 

although they are often useful in interpretation 

of the results . The supply of thermal 

energy (Q* + G) is measured, so limits ar e 

placed on the estimates of convection and 

latent heat with this method . 

The disadvantages of the model are primarily 

associated with instrumentation ; 

accurate measurements are required in order 

to measure the small gradients in temperatur e 

and vapor found near the forest canopy . In 

addition, the basic relationship in equation 2 

becomes undefined whenever 0=-1 . This 

normally occurs infrequently, and ordinarily 

only for short periods near dawn or dusk 

when the amount of available energy i s 

limited. The magnitude of H and XE wil l 

normally be small at such times . 

Site Conditions 

The energy exchange studies were carrie d 

out in the broad, flat valley of the Ceda r 

River, near Seattle, Washington. The soil and 

stand conditions have been described by 

Fritschen (1972) . The stand is second-growth 

Douglas-fir approximately 35 years old, with 

an average of 570 trees per ha . The stand 

canopy is relatively level ; the height of the 

tips of the tallest trees in the vicinity of th e 

experimental site is about 28 m . The site is 

adjacent to the lysimeter tree described by 

Fritschen (1972) . A series of energy transfer 

studies are planned at this site in the future , 

using the lysimeter tree, meteorological 

towers, and a variety of models for evaluatin g 

energy exchange processes . 

Field Measurement s 

Net radiation, stored heat, and temperature 

and vapor pressure measurements were obtained 

with a mobile data acquisition that ha s 

been described by Gay . l The truck-mounted 

system includes sensors and supports, cabling , 

and a digital data logger with resolution of 

0.001 percent (0 .1 microvolt on a 10 millivol t 

scale). Ceramic wick, wet-bulb psychrometers 

of the basic design of Lourence and Pruit t 

(1969), as modified by Gay (n .d.), were used 

for the temperature and vapor pressur e 

measurements. 

The sensors were mounted on a 33 .5 m 

(110 ft) tall, triangular TV tower about 0 .3 m 

(1 foot) in width. Wind, temperature, and 

vapor pressure measurements were made a t 

six levels, respectively, 26 .16, 27 .16, 28 .16 , 

29 .16, 30 .16, and 31 .16 m above the fores t 

floor. The radiation budget components were 

measured from a height of 30 .66 m above th e 

floor. The tip of the tallest tree in the vicinity 

of the tower extended to 28 m, though th e 

bulk of the crowns were below 24 m, and th e 

general level of crown closure was about 20 t o 

22 m above the floor . Five soil heat flux disks 

were installed at the -2 cm level, just beneat h 

the surface of mineral soil on the forest floor . 

Observations began on July 27 and continued 

through August 1. The sensors were 

sampled at 5-minute intervals during the day , 

and at 10-minute intervals at night . The observation 

period spanned a range of weather conditions 

that included one clear and one completely 

overcast day. The data acquisitio n 

system performed well during this period . 

Problems in Forest Energy 

Budget Analyses 

A variety of measurement problems are en - 

countered in energy budget studies. In addition, 

forests have unique characteristics o f 

scale and mass that affect the application o f 

I L . W. Gay. An environmental data acquisitio n 

system . National Conference on the Forest, Weather , 

and Associated Environment, Atlanta, Georgia, May 

18-19, 1971 . Mimeo., 8 p . Abstracted : Bull. Am . 

Meteor . Soc . 52 : 202-203 . 

245

the basic energy budget model given in equation 

2 . The gradients of temperature an d 

vapor above the rough, porous forest canop y 

are very slight, and measurement difficultie s 

increase with small gradients . Another major 

problem is related to the difficulty of measuring 

changes in stored energy in the biomass of 

the forest . These problems will be placed int o 

perspective at the Cedar River site, for the y 

affect the analyses and the subsequent interpretation 

of the results . 

Evaluation of Gradients 

Gradients of temperature, vapor concentration, 

and wind are small above fores t 

canopies, even though the transfer of energy 

and mass may be proceeding at high rates . 

The small gradients result from mixing induced 

by the mechanical turbulence create d 

by the rough canopy surface . In addition, the 

exchange surfaces are distributed through a 

considerable canopy depth so that the source s 

of heat and vapor are diffuse, rather tha n 

being concentrated as in the dense canopies of 

crops or other low vegetation . The smal l 

gradients above the forest require tha t 

extreme care be taken in the development o f 

suitable instrumentation, and in the experimental 

design controlling the deployment o f 

sensors . 

The Bowen ratio model assumed tha t 

steady state conditions prevail, i .e., the value s 

of the variables do not change with respect t o 

time during the period of analysis . This i s 

partially satisfied by averaging the values ove r 

the period of an hour before applying the 

model . Integration into hourly means als o 

reduces the small random component of error 

associated with the measurements. It does 

not, however, reduce biases that are introduced 

by small differences among sensors . 

Such biases can be a source of serious error , 

particularly with small gradients that exis t 

above the forest . Bias errors must be handle d 

by techniques other than averaging . 

Two approaches have been used to cope 

with the problems involved in the measurement 

of small gradients above forests . In the 

first approach, the two levels of measuremen t 

required for the Bowen ratio are separated by 

a relatively large distance (10 m) in order to 

increase the differences that are being measured 

to a level commensurate with the sensitivity 

of the data system (Galoux et al . 1967 , 

Storr et al. 1970) . In the second, reversing 

sensors have been used to cancel the effect of 

any small biases that may exist between 

sensors (Black and McNaughton 1971) . 

A large vertical separation of the sensors 

introduces questions of the representativenes s 

of the measured gradients which should represent 

the effect of the underlying surface . It is 

quite possible that a sensor placed 10 m abov e 

the canopy may be measuring properties o f 

the atmosphere that are derived from a surface 

other than the one under investigation . 

In contrast, a pair of reversing sensors, place d 

near the canopy, offers an excellent means fo r 

evaluating small gradients . However, the 

reversing mechanism introduces new problem s 

of design for operation and support in tal l 

forests . 

A graphical approach was used in this study 

to minimize sensor errors in the gradien t 

measurements used for the Bowen ratio 

analysis . Initially, the mean hourly values o f 

potential temperature were plotted agains t 

the associated values of vapor pressure t o 

yield 24 plots (1 per hour) for each day, with 

each plot containing six points (one for eac h 

measurement level) . The plots will be straigh t 

lines if similarity exists between the gradient s 

of potential temperature and vapor concentration, 

providing there are no errors of measurement 

(Tanner 1963). Since similarity is an 

assumption of the Bowen ratio method, thes e 

plots were used to separate out the instrument 

levels that exhibited small offset s 

throughout the day and to identify thos e 

levels appropriate for use in the Bowen rati o 

analysis . 

The similarity plots for July 29 are shown 

for levels 1, 4, 6 in figure IA, and for levels 1 , 

2, 3 on July 31 in figure 1B . The potentia l 

temperature (e-) scale on the ordinate differ s 

between the 2 days. The vapor pressure (e ) 

scale is shown in the legend . Note that the plo t 

for each hour has been normalized by subtracting 

the & and e value at the bottom level fro m 

the observations at each of the other levels . 

Therefore each plot actually shows the incre - 

246

0.4 

0.2 

v 0 

0 

w 

IX -0.2 

I- 

cr -0.4 

w 

-0.6 

~ 0.04 

A ..29 

JULY 

HOUR (PDT) -->- y y g i 

0 I 2 3 4 5 6 7 1I I 1 1! I r 17 .19 202122 23 2 4 

9 10 II 12 13 14 15 16 o 18 

B. 31 JULY 

r 

a 

I 

2 3 4 5 6 7 8 9 

HOUR (PDT)----4 .- 

10 

I I 

2.03.122! J 

1213 14 15 16 17 18 19 ° 1 23 24 

0- I 

u 

-0.04 • MB 

VAPOR PRESSURE, M B 

Figure 1 . Similarity between gradients of potential temperature and vapor pressure . A. Clear weather on Jul y 

29, 1971 . Levels 1, 4, and 6 . B . Overcast weather on July 31, 1971 . Levels 1, 2, and 3 . 

ment of a and e with respect to the firs t 

measurement level near the canopy . 

Though several points can be deduced fro m 

such similarity plots, the most important conclusion 

concerns adequacy of data . The linearity 

at the selected levels confirms their 

acceptability for the Bowen ratio model , 

although an unexplained offset is evident at 

level 2 during the 1000-1200 hours on Jul y 

31 . 

Once the data is judged acceptable, on e 

notes that the slope of the lines is A&/Ae ; this 

is directly proportional to 0, the Bowen ratio , 

as shown in equation 3 . Thus the relative 

slope of the similarity plots is an index to th e 

way that the surface is partitioning the net 

energy supply into convection and evaporation. 

Further, the quadrant of each hourl y 

plot indicates the sign of 0, and the direction 

of the H and XE fluxes . 

For example, as plotted in figure 1A and 

1B, both fluxes will have the same sense in 

quadrants I and III where the slope is positiv e 

((3>0). In quadrant I, H and XE are both directed 

away from the surface and are negative 

by the sign convention adopted earlier . In 

quadrant III, H and XE are both positive, a s 

their sense is toward the surface . The slop e 

and the (3 coefficient are negative in quadrant s 

II and IV. In quadrant II, H is toward th e 

surface (advection) while XE is directed away , 

while the fluxes in quadrant IV have th e 

opposite sense . The similarity plots can thu s 

provide three things : a ready indication of the 

magnitude of the Bowen ratio ; an indicatio n 

of the direction of the H and E fluxes ; and an 

identification of levels suited for the Bowe n 

ratio analysis . 

247

Evaluation of Stored Hea t 

Application of the Bowen ratio model t o 

forest measurements reveals a problem i n 

measuring the stored heat term in equation 2 

that is a direct consequence of the nature of 

the forested surface . The scale of the forest 

elements makes the change in heat storage i n 

the biomass difficult to measure, and man y 

studies (Baumgartner 1956) have treate d 

these changes as negligible . This is certainly 

true on a daily basis, but an appreciable 

amount of energy appears to move into 

storage (-G) in the biomass in the early morning, 

and out again (+G) in the early evening . 

These amounts ordinarily will balance when 

totaled over the day, but comparisons amon g 

hourly totals may be in error unless the 

energy storage changes are estimated for th e 

biomass as well as for the soil . 

The estimates of biomass storage change 

were derived here by an indirect method that 

may prove useful in other situations as well . 

First, application of the Bowen ratio model to 

data collected during the early evening hours 

frequently resulted in positive estimates o f 

latent energy (condensation) when the vapo r 

gradient clearly indicated that evaporatio n 

was taking place . This apparent anomaly in 

the Bowen ratio estimate of XE could occur 

as a consequence of underestimating the 

storage term G . In order to obtain the proper 

sign on XE under these conditions, the estimate 

of G must be increased to a positiv e 

value that exceeds the absolute value of th e 

negative net radiation . 

Estimates of the minimum probable value 

of G were obtained in this manner for th e 

hours between sunset and the time near mid - 

night when the vapor gradient changed direction, 

indicating the beginning of condensation 

. As a second step, the crossover points 

between release (+G) and uptake (-G) o f 

stored energy were then estimated from m y 

experience and that of others (Grulois 1968 ) 

as being about 1 hour after sunrise and 3 

hours before sunset . The release of store d 

heat (+G) is then defined by the two cross - 

over points and by the magnitude during th e 

early evening hours . The third and final ste p 

was to estimate the magnitude of the gain in 

storage (-G) during the morning hours so tha t 

the daily integral approximately balanced . 

The final result is a much better estimate o f 

hourly changes in stored energy than could be 

obtained by direct measurement of the soil 

storage component alone. The magnitude of 

the stored energy change will be investigate d 

further during future studies . 

The major problems in application of th e 

Bowen ratio model to the forest appear to b e 

associated with adequate precision of th e 

measurements. The similarity tests demonstrated 

here appear useful in eliminatin g 

errors from the data . Further, the changes in 

stored energy can be estimated by an indirect 

method, based upon gradient measuremen t 

and knowledge of the time at which th e 

stored heat flux reverses sign . 

The Forest Energy Budget 

Energy budget analyses were developed fo r 

2 days that represent quite different amount s 

of available energy . The solar energy input to 

the forest was large on July 29, 1971, a da y 

characterized by clear skies and a warm mea n 

temperature (24.4°C). In contrast, July 31 , 

1971, was completely overcast with reduced 

levels of incoming solar radiation and a relatively 

cool mean temperature (17 .4°C) . 

Examination of the energy budgets unde r 

such contrasting conditions will improve our 

understanding of the basic processes tha t 

govern energy transfer between the forest and 

the atmosphere . 

The diurnal energy budget will first b e 

examined with respect to daily totals; a discussion 

of the relationships among hourly 

values of the energy budget components wil l 

then follow . 

The Diurnal Energy Budget 

The daily energy budget totals are presented 

for the separate periods of daylight (1 3 

hours) and night (11 hours) and for 24-hou r 

totals in table 1 . Dividing the daily totals into 

daylight and night portions enhances futur e 

comparisons that may be made with dat a 

collected under different daylengths at Ceda r 

River or elsewhere . 

248

Table L-Diurnal energy budget components ' 

Period 

July 29-clear 

July 31-overcast 

Q* G H XE @* G H X E 

cal/cm2 

Daylight 454 -43 -143 -263 141 -7 -38 -9 6 

Night -44 48 8 -17 -7 13 -1 - 6 

Daily total 410 5 -135 -280 134 6 -39 -102 

1 Totals are given for the daylight hours, 0630-1930 PDT ; night hours, 1930-0630 PDT; and the full day , 

0000-2400 PDT . 

There is a large difference in the radiatio n 

supply on the 2 days, as net radiation totale d 

410 cal/cm 2 under the clear skies of July 29 , 

and only 134 cal/cm 2 for the overcast conditions 

of July 31 . These totals include a stead y 

net loss of radiation at night, amounting t o 

-44 cal/cm 2 under clear skies, and -7 cal/cm 2 

under the overcast conditions . 

The net radiation term represents the 

energy converted from radiative to nonradiative 

forms by the forested surface . The shortwave 

radiation from the sun makes up th e 

largest component of the net radiation . During 

the 13 hours of daylight on the clear day , 

the forest received 584 cal/cm 2 of solar radiation 

and reflected 55 cal/cm 2 . Under overcast 

skies, the forest received 171, and reflected 

16, calories/cm 2 . The albedo was 0 .09 on 

both days . 

Since the gain and loss of longwave radiation 

also enters into the supply of radian t 

energy, it is not helpful to calculate a shortwave/net 

radiation ratio as an index of efficiency 

of conversion . However, the lo w 

albedo value (0.09) emphasizes the efficienc y 

with which the Douglas-fir canopy absorbs 

solar radiation. This low reflectivity is similar 

to values reported for other coniferous canopies 

(Stewart 1971), and is much lower tha n 

the 0 .2-0.25 albedo values that commonl y 

prevail over crops and other low vegetatio n 

(Monteith and Szeicz 1961) . As noted by 

Baumgartner (1971), forests are effectiv e 

absorbers of solar radiation . 

The changes in stored energy (G) tabulate d 

in table 1 for the 24-hour period are nea r 

zero, which is in accord with observation s 

reported elsewhere . This term represents the 

changes in heat storage of both the biomass 

and the soil. Most of the storage changes ar e 

attributed to the biomass ; the indirect 

methods used to estimate the changes in 

storage have been described in an earlier section 

. 

I estimate that -43 cal/cm 2 went into storage 

during the daylight hours on the clear da y 

and that 48 cal/cm 2 came out of storage during 

the night. The storage changes on the 

cloudy day proceeded in a similar direction , 

but the magnitudes were much smaller . 

The storage term at Cedar River appear s 

large because of the large quantity of the biomass 

there . The biomass is as yet unmeasured , 

however. Attempts will be made to measure 

the storage flux directly in future experiments . 

The convective flux for the clear day 

totaled -135 cal/cm 2 , directed away from the 

forest into the atmosphere . A slightly larger 

amount, -143 cal/cm2 , was lost during daylight, 

but 8 cal/cm2 was gained by the canopy 

at night when the canopy temperature s 

dropped below that of the air. Under overcas t 

sky conditions, -38 cal/cm 2 were lost over the 

full day . 

24 9

Latent energy was the largest dissipation 

term on each of the days, totaling -28 0 

cal/cm 2 for July 29 and -102 cal/cm 2 on July 

31 . There was a net loss of latent energy b y 

night, as well as by daylight, for both days . 

The evaporation equivalent of the laten t 

energy total was about 0 .5 cm on July 29 , 

and 0 .18 cm on July 31 . 

The Bowen ratio ((3 =H/AE) is a measure of 

how the surface partitions the energy suppl y 

between sensible and latent heat . The mean 

daily value of 13 was 0.48 for the clear day , 

and 0 .38 for the overcast day. The differenc e 

in (3 between days is not large, but it suggest s 

that the forest partitioned more of the energy 

supply into convection on the clear day tha n 

on the overcast day . From another viewpoint, 

the ratio of XE/Q* was 0 .67 on the sunn y 

day, and 0 .76 on the overcast day . This is i n 

the direction that one might expect for a 

stand of vegetation that receives a large inpu t 

of energy . 

Hourly Energy Budgets 

The phase relationships among the energy 

budget components can be examined with the 

aid of figure 2 which shows the hourly value s 

on July 29, and figure 3 which shows th e 

hourly values on July 31 . Each plotted poin t 

represents the midpoint of an hourly mean . 

The daytime, night and daily totals in table 1 

were integrated from the rates shown in thes e 

two figures . The 2 days exhibit different characteristics, 

so they will be discussed separately . 

The symmetry of the bell-shaped net radiation 

curve on July 29 confirms that the skie s 

were cloudless on that day . The maximu m 

intensity occurred during the hour centere d 

on 1300 hours PDT, which closely corresponded 

with solar noon . The net radiation 

values became positive about 1 hour after sun - 

rise and remained positive until shortly befor e 

sunset, indicating the hours in which there 

was a surplus of energy that might be dissipated 

through the other energy budget components. 

The net radiation was negative 

throughout the night, as the surface lost 

energy to the atmosphere. The greatest net 

radiation loss occurred at 2200, about 1 hou r 

after sunset at a time when the forest wes 

rapidly losing the absorbed solar radiatio n 

that had been stored during the day . 

The phase of the fluxes is also of interest , 

4 8 10 12 14 16 18 20 22 24 

TIME, HR (PDT ) 

Figure 2. Energy budget components under clear skies . Symbols: net radiation, Q* ; change in heat storage of 

soil and biomass, G ; convection, H ; latent energy, AE . 

250

4 8 10 12 14 16 18 20 22 24 

TIME, HR (PDT ) 

Figure 3 . Energy budget components under overcast skies . Symbols : net radiation, Q* ; change in heat storage 

of soil and biomass, G ; convection, H ; latent energy, XE . 

as G, H, and XE all lag behind Q* . Let us 

consider the stored heat flux first . It reache s 

its peak flow into the biomass and soil (-G) i n 

midmorning, and reverses to flow out of th e 

biomass (+G) in the late afternoon and earl y 

evening. The change in stored heat appears t o 

provide a significant source of energy to th e 

surface throughout the night . 

The sensible heat flux, H, reaches its maxi - 

mum about two hours after G, but still a n 

hour before solar noon . Sensible heat i s 

directed away from the surface during day - 

light (-H), but reverses in direction as convection 

begins to provide energy (+H) to th e 

surface during the night . During this period , 

the canopy cools below air temperature du e 

to longwave emission . 

The latent energy flux reached its maxi - 

mum about two hours after solar noon . 

Evaporation continued well into the night ; 

only during the early morning hours did a 

rather small amount of condensation tak e 

place . 

The marked phase shift between sensible 

and latent energy is of interest, as many 

studies have shown these two fluxes to be i n 

phase with net radiation (Baumgartner 1956 , 

Denmead 1969, Rauner 1960) . The Douglas - 

fir forest, in contrast, partitioned the energy 

available at the surface into sensible and 

latent energy on a preferential basis . This 

partition was on a 1 :1 basis during the morning, 

but latent energy was apparently favored 

at the expense of sensible energy during th e 

afternoon . A similar pattern is evident i n 

measurements over a young Douglas-fir forest 

near Vancouver, B .C. (Black and McNaughto n 

1971), and over a mixed hardwood forest 

(Grulois 1968) . 

The phase shift in latent energy into the 

afternoon is probably related to a vapor pressure 

deficit which has an afternoon maximu m 

on clear days . Stewart and Thom 2 have concluded 

that the latent energy flux from thei r 

pine forest site in England is controlled more 

by the vapor pressure deficit than by th e 

supply of available energy . This conclusion is 

based upon their evaluation of the interplay 

between the relatively large internal resistanc e 

to transfer and a small external resistance ; the 

ratio for the pine site was in the order o f 

20 :1 . 

Conclusions 

The observations reported here are a n 

initial contribution toward the problem of 

evaluating the flow of energy and mass between 

the atmosphere and the young Douglas - 

fir forest at the Cedar River site . 

2 J . B. Stewart and A . S. Thom . Energy budgets i n 

pine forest. Institute of Hydrology, Wallingford , 

Berkshire, U . K . Unpublished manuscript, 1972 . 

251

Forested surfaces are generally considere d 

to be effective energy exchange surfaces . Th e 

results confirm that this young stand has a 

high absorptivity for solar radiation, with a n 

albedo of 0 .09 for both clear and overcas t 

conditions . This high absorptivity contribute s 

to the large net radiation values that wer e 

measured under clear skies . 

The role of the forest in dissipating the net 

radiation is of particular interest. The porous , 

aerodynamically rough canopy is effective i n 

transferring sensible and latent energy into 

the atmosphere . The large quantity of forest 

biomass may also involve an amount of store d 

thermal energy that is of significance durin g 

short periods, even though the daily totals are 

small . Summed over a 24-hour period, evapotranspiration 

was about 280 cal/cm 2 min (0 . 5 

cm water equivalent) or about two-thirds o f 

the net radiation that was transformed unde r 

clear skies . Evapotranspiration was relativel y 

larger on an overcast day, about three - 

quarters of net radiation, although the tota l 

amount of latent energy (102 cal/cm2 min, or 

0.18 cm water equivalent) was considerably 

lower . 

These results provide initial estimates o f 

the amounts of energy transferred during extreme 

conditions under cloudless and unde r 

overcast skies . The exchange of energy an d 

mass depends not only upon the energy inpu t 

to the forest, however, but also upon the 

physiological response of the vegetation . No w 

that the instrumentation system and the 

analysis model have been tested at this site , 

subsequent research will include a range of 

environmental conditions . Instrumentation 

development and model testing will continue 

in cooperation with the lysimeter installatio n 

and eddy flux system of cooperating investigators 

. Ultimately, analysis and interpretatio n 

of the energy transfer studies must include 

physiological as well as physical characteristics 

of the stand . 


The work upon which this publication i s 

based was supported in part by funds provided 

by the U .S. Department of Interior, 

Office of Water Resources Research, as 

authorized under the Water Resources Re - 

search Act of 1964, and administered by th e 

Water Resources Research Institute, Orego n 

State University ; and the National Scienc e 

Foundation Grant No . GB-20963 to the 

Coniferous Forest Biome, U .S. Analysis o f 


This is Contribution No . 40 to the Coniferous 

Forest Biome and Paper 844, Forest Researc h 

Laboratory, School of Forestry, Oregon State 

University . 


Baumgartner, A. 1956 . Investigations on th e 

heat- and water economy of a young forest. 

Translated from Ber . Deut . Wetterdienst 5 : 

4-53 . Commonwealth Sci. & Ind. Res . 

Organ . Translation 3760, Melbourne , 

Australia, 1958 . 

1965. The heat, water and car - 

bon dioxide budget of plant cover : methods 

and measurements . In F. E . Eckardt 

(ed.), Methodology of plant eco-physiology: 

Proceedings of the Montpellie r 

Symposium, p. 495-512 . Paris: UNESCO . 

1971 . Wald als Austauschfakto r 

in der Grenzschicht Erde/Atmosphare . 

Forstwiss. Cbl. 3: 174-182 . 

Black, T. A., and K. G. McNaughton . 1971 . 

Psychrometric apparatus for Bowen-ratio 

determination over forests . Bound . Laye r 

Meteorol . 2 : 246-254 . 

Bowen, I. S. 1926 . The ratio of heat losses by 

conduction and by evaporation from an y 

water surface . Phys. Rev . 27 : 779-787 . 

Denmead, O. T. 1969 . Comparative micro - 

meteorology of a wheat field and a fores t 

of Pinus radiata . Agric. Meteorol. 6 : 

357-371 . 

Federer, C. A. 1970. Measuring forest evapotranspiration-theory 

and problems . USDA 

Forest Serv . Res . Pap . NE-165, 25 p . 

Northeast. Forest Exp . Stn ., Upper Darby , 

Pa . 

Fritschen, L . J. 1965 . Accuracy of evapotranspiration 

determinations by the Bowe n 

ratio method . Bull. Int. Assoc . Sci. Hydrol . 

10: 38-48 . 

252

1970. Evapotranspiration an d 

meteorological methods of estimation as 

applied to forests . In J . M. Powell and C . 

F. Nolasco (eds .), Proceedings of the Third 

Forest Microclimate Symposium, p. 8-27 . 

Can. For. Serv ., Calgary, Alberta. 

1972. The lysimeter installation 

on the Cedar River Watershed . In Jerry F . 


Waring (eds.), Proceedings-research on 


p. 255-260, illus . Pac. Northwes t 


Galoux, A., G . Schnock, and J . Grulois . 1967 . 

Les installations ecoclimatologiques . 

Travaux Stat . Rech . Eaux et Forets , 

Groenendaal-Hoeilaart, S4rie A, 12, 52 p . 

Gay, L. W. (n .d.) On the construction and use 

of ceramic-wick psychrometers . In R. W . 

Brown and B . P. Van Haveren (eds .) , 

Psychrometry in water relations research . 

Utah Agric . Exp. Stn. (In press . ) 

Grulois, J. 1968 . Flux thermiques et evapotranspiration 

au tours d 'une journ4e 

sereine . Bull. Soc . r. Botanique Belg . 102 : 

27-41 . 

Lourence, F. J ., and W. O. Pruitt . 1969 . A 

psychrometer system for micrometeoro 

l o gy profile determination. J. Appl . 

Meteorol . 8: 492-498 . 

Monteith, J. L., and G. Szeicz. 1961. Th e 

radiation balance of bare soil and vegetation. 

Quart. J. Roy. Meteorol. Soc. 87 : 

159-170 . 

Rauner, Yu. L. 1960 . The heat balance of forests. 

Izvestiya Akademii Nauk SSSR, Seriya 

Geograficheskaya. 1 : 49-59 . Translated by 

Israel Prog. Sci. Transl. No. 1342 . U .S . 

Dep. Commerce, Washington, D .C . 

Stewart, J . B. 1971. The albedo of a pine forest. 

Quart. J. Roy . Meteorol. Soc. 97 : 

561-564 . 

Storr, D., J . Tomlain, H . F . Cork, and R . E . 

Munn. 1970. An energy budget study 

above the forest canopy at Marmot Creek , 

Alberta, 1967 . Water Resour . Res. 6 : 

705-716 . 

Tanner, C . B. 1960. Energy balance approac h 

to evapotranspiration from crops . Soil Sci . 

Soc . Am . Proc . 24: 1-9 . 

1963 . Basic instrumentation an d 

measurements for plant environment an d 

micrometeorology . Soils Bull . 6, various 

pagination . Madison : Univ . Wis . 

Tajchman, S . J. 1971 . Evapotranspiration an d 

energy balances of forest and field. Water 

Resour. Res . 7: 511-523 . 

253



The lysimeter installation on 

the Cedar River Watershed 

Leo J . Fritsche n 

Associate Professo r 

College of Forest Resource s 



Abstract 

A lysimeter was built around the root ball of a 28 m Douglas-fir (Pseudotsuga menziesii) tree. The container , 

tree, and soil weigh 28,900 kg. The sensitivity of the weighing mechanism is 630g which is equivalent t o 

0.06 mm of water. The installation will be used to study evapotranspiration and volume changes in relation t o 

soil water potential and atmospheric demand; to test cuvette and meteorological methods ; determine canop y 

interception ; and to assess the effects of irrigation and fertilization . 

Introduction 

Understanding the process and quantifying 

the rate of water transfer within a forest ecosystem 

has become increasingly more important 

in recent years . This concern has originated 

from both applied and basic directions . 

For example, it has been estimated that by 

1980, six major regions will have no wate r 

reserve. These regions are Colorado River , 

South Pacific, Great Basin, Upper Ri o 

Grande, Pecos River, and Upper Missour i 

River (Colorado River Association 1966) . 

Furthermore, interbasin transfer is bein g 

practiced in at least two areas at the presen t 

time . 

Increasing need for water conservation, th e 

possibility of interbasin water transfer, and 

the treatment of watersheds to increase yiel d 

or change quality has demonstrated the nee d 

for additional research on water use by various 

types of vegetative cover. For example , 

clearcutting a north-facing Coweeta watershe d 

resulted in a 1st year increase streamflow o f 

40.2 cm and a stabilized increase of 23 .8 cm . 

Although the 1st year 's increase from a south - 

facing watershed was only 15 cm which de - 

creased to insignificance by the 3d year 

(Swift, unpublished data), it is believed that 

the different microclimatic influences upon 

the vegetation and resulting evapotranspiration 

will account for the discrepancy in wate r 

yield. To understand fully the processes involved 

and to predict results from futur e 

cuttings require detailed investigations relating 

water use to its availability and to atmospheric 

conditions . 

Short-period (1 hour or less) determinations 

of evapotranspiration are necessary t o 

study the complex soil-plant-atmospheric relations. 

Many methods or combination o f 

methods have been employed to determin e 

evapotranspiration. The major ones includ e 

watershed runoff ; measurement of soi l 

moisture depletion by either gravimetri c 

sampling, with resistance blocks, or with neutron 

soil moisture meters ; use of weighin g 

lysimeters ; and application of meteorologica l 

models . Other methods have employe d 

percolation tension plates, stemflow measurements, 

and enclosures . Of these methods only 

weighing lysimeters, enclosures, and meteorological 

models are capable of yielding hourl y 

results . Enclosures alter the microclimate an d 

255

may bias the results . Meteorological method s 

must be tested in the areas of anticipated us e 

(Slatyer and Mcllroy 1961) . Such testing has 

not been accomplished in forested areas . 

Thus, weighing lysimetry appears to be th e 

only method capable of yielding the neede d 

short period accurate rates . 

Weighing lysimeters have been used to determine 

evapotranspiration of agriculture 

crops (McIlroy and Angus 1963, Pruitt and 

Angus 1960, Harrold and Dreibelbis 1951, Va n 

Bavel and Meyers 1962), but they haven't 

been used for natural vegetation, such a s 

brush and trees, which have extensive root 

systems . 

The use of weighing lysimeters to deter - 

mine evapotranspiration of trees is feasibl e 

only if the trees are uniformly spaced such as 

in a plantation . Installation is easier when the 

root system is naturally restricted, thereby 

reducing the size of the lysimeter and it s 

deadweight. Under these conditions, it would 

be possible to obtain accurate short period 

evapotranspiration rates, thus enabling a 

mechanistic examination of evapotranspiration 

in relation to the determining meteorological 

and soil moisture conditions . 

Establishment of evapotranspiration from a 

single tree or group of trees in a specific location 

is important but does not yield answer s 

for other types of vegetation in different 

climatic zones . Installation of lysimeters in 

many locations is not possible and may not be 

feasible because of the cost . However, meteorological 

methods if tested can be employed to 

establish evapotranspiration rates wher e 

lysimeters are not feasible. Results fro m 

meteorological methods can be tested against 

the results from a properly installed weighing 

lysimeter. This has been accomplished for 

agriculture crops (Tanner 1967, Pruitt 1963 , 

Fritschen and Van Bavel 1963, Fritsche n 

1965) . However, differences in scale factor s 

such as canopy height, density, and roughnes s 

demand additional testing before these 

methods can be used over natural vegetation . 

The purpose of this paper is to discuss th e 

establishment of the weighing lysimeter in a 

Douglas-fir forest for the primary purpose of 

studying the complex relation between evapotranspiration 

rates and the determining 

meteorological and soil moisture conditions . 

In later stages the installation will be used t o 

evaluate meteorological methods for determining 

evapotranspiration to be used i n 

other areas . 

The Site 

The lysimeter installation is located on the 

Cedar River Watershed near Seattle, Washing - 

ton. The soil is a Barneston, gravelly, loamy 

sand originating from glacial outwash laid 

down at the end of the Vashon glacial period 

(Poulson and Miller 1952) and generally restricts 

the root system above the 3-foot depth 

(Gessel and Cole 1965) . The lateral extent of 

the root system is largely restricted to th e 

basic spacing of the trees . The area is relatively 

level and has a uniform canopy density 

making it desirable for micrometeorological 

investigations . 

The trees are 35-year-old Douglas-fir (Pseudotsuga 

menziesii) which regenerated naturally 

after logging. The average tree spacing i s 

5.8 m, resulting in 231 trees per 4,041 m 2 

consisting mostly of Douglas-fir with a fe w 

hemlock (Tsuga heterophylla) and maple . 

Trees in the 5 to 30 cm d.b.h. classes occur 

with the greatest frequency . Ground vegetation 

consists of bracken fern (Pteridiu m 

aquilinum), salal (Gaultheria shallon), red 

huckleberry (Vaccinium parvifolium), an d 

mosses, mainly (Eurhynchium oregonum) . 

The Lysimeter 

The lysimeter was constructed around th e 

root ball of a 35-year-old, 28 m tall an d 

38 cm diameter Douglas-fir tree . The lysimeter 

consists of two right cylinder containers, 

one located within the other . The inner - 

most container in which the tree is located i s 

366 cm in diameter and 122 cm deep, havin g 

a surface area of 10 .5 m 2 (fig. 1). The container 

at "field capacity " with tree and soi l 

weighs 28,900 kg (fig . 2) . 

The inner container is resting on a hydraulic 

transducer located on the bottom of th e 

outer container. The hydraulic transduce r 

consists of eleven 15 m lengths of 6 .35 c m 

256

Figure 1 . Walls of the inner container being fitted together 

prior to lowering down around the root ball . 

Figure 3. Butyl rubber tubing used for weighing coiled 

on the bottom of the outer container . 

Figure 2 . Completed inner container with soil and tre e 

shown raised 90 cm for installation of the hydrauli c 

transducer . 

butyl rubber tubing with a valve stem vulcanized 

30 cm from one end . The tubing was 

filled with air-free water and was coiled o n 

the bottom of the outer container starting at 

the center (fig . 3). Holes were cut in the bottom 

of the outer container to allow the valv e 

stems to pass through . Thick-walled tubing 

(9.5 mm I .D.) was attached to each valve stem 

and extended beneath the bottom of th e 

outer container to a shutoff manifold . The 

manifold is connected to a vertical standpipe ; 

thus, the weight of the tree, soil, and inne r 

container is reflected in the height of th e 

water column in the stand pipe . A water-fille d 

dummy standpipe is located adjacent to the 

active standpipe for the purpose of temperature 

compensation and counter balance . 

Weight changes are detected by a ±6 .9 mb differential 

pressure transducer located betwee n 

the active and the dummy standpipes . The 

signal is recorded on an automatic data logging 

system . 

To provide for drainage of excess water , 

eight filter candles were installed around th e 

periphery of the inner container about 10 c m 

from the bottom. The filter candles were 

5 cm in diameter by 20 cm ceramic tube s 

with a bubbling pressure in excess of 25 0 

millibars. The filter candles are connecte d 

through a manifold system to an evacuate d 

reservoir. The outflow from the drainage 

system is measured with a modified typ e 

tipping bucket rain gauge . 

To prevent the tree from blowing or falling 

over during and after construction, it wa s 

guyed from a yoke located at 10 m to the 

base of adjacent trees. After construction four 

climbable (30 cm triangular) towers were 

built around the lysimeter and the tree guye d 

to the towers with horizontal cables. The 

cables were loose enough to allow 15 c m 

motion at 10 m . 

The sensitivity of the lysimeter is 630 g 

which is equivalent to 0 .06 mm of water o r 

22 ppm. The lysimeter installation at Tempe , 

Arizona (Van Bavel and Meyers 1962), has a 

sensitivity of 20 g or 0 .02 mm of water o r 

257

7.4 ppm . The lysimeters at Coshocton, Ohi o 

(Harrold and Dreibelbis 1951), have a sensitivity 

of 2,268 g or 0 .25 mm of water or 

38 ppm and the lysimeter at Davis, Californi a 

(Pruitt and Angus 1960), has a sensitivity o f 

907 g or 0 .03 mm of water or 20 ppm . 

The Weather Station 

Located adjacent to the lysimeter tree an d 

on a tower 33.5 m in the air are the meteorological 

sensors of the weather station . Thes e 

consist of a solarimeter, a net radiometer, air 

temperature sensor, vapor pressure sensor , 

wind direction and speed sensors, and a 

tipping bucket rain gauge . In addition to thes e 

parameters, soil temperature is measured a t 

four depths inside and outside of the lysimeter. 

The signal from these sensors is re - 

corded automatically on a digital magneti c 

tape data logging system. At the present time , 

five of the signals are being integrated continuously 

: solar radiation, net radiation, rain - 

fall, windspeed, and the drainage from th e 

lysimeter. The integrals of these signals an d 

the other parameters are recorded on the 

magnetic tape at hourly intervals to conserv e 

the magnetic tape's supply . With hourly 

records, the tape supply will last for 30 days . 

The magnetic tape (6 .4 mm) is converted t o 

12.7 mm computer compatible tape and the n 

analyzed at the University of Washingto n 

Computer Center with the Burroughs 550 0 

computer . 

The Proposed Uses of the 

Lysimeter Facility 

Since the lysimeter was installed during th e 

summer of 1970 and the completed facility 

was not completely checked out unti l 

recently, very little data is available for discussion. 

However, it may be enlightening to 

discuss the proposed uses of the facility . 

An Evapotranspirimeter 

The lysimeter installation is expected to 

yield short period rates of evapotranspiratio n 

from the 28 m Douglas-fir tree . The accurac y 

of the measured evapotranspiration is withou t 

much question . How representative the dat a 

are is questionable unless the water potential 

of the root mass is kept equal to the wate r 

potential of the root mass of adjacent trees . 

Since the root mass is restricted in a 366 c m 

diameter container, the amount of soil avail - 

able for water and nutrient extraction by thi s 

tree is slightly less than adjacent trees . Durin g 

the summer months, water should be with - 

drawn from the lysimeter container at a faster 

rate than from the adjacent soil . Therefore , 

water will have to be added to the lysimeter 

container to maintain a water supply equal t o 

that in the adjacent area. During the winter 

months, the opposite will be true . The bottom 

of the lysimeter container inhibits dee p 

percolation . When rainfall exceeds evapotranspiration, 

water will have to be removed fro m 

the container in order to prevent a buildup of 

water in the bottom of the container . 

Evapotranspiration as a Function of Potentia l 

In order to study the mechanism of transpiration 

from a Douglas-fir tree, the evapotranspiration 

rates will be determined in relation 

to the soil-water potential and the atmosph 

eric evaporative demand . During thes e 

studies, soil moisture potential will be deter - 

mined with a series of thermocouple psychrometers 

installed in the lysimeter container . 

The atmospheric evaporative demand will b e 

calculated from meteorological parameters . 

Tissue Volume Change s 

Dendrometer bands will be installed at various 

locations on the tree to study the swellin g 

and shrinking of the tree in relation to water 

potential and evapotranspiration . At the sam e 

time, stomata aperture and plant stress will b e 

determined using leaf resistance meters an d 

thermocouple psychrometers or Scholande r 

bombs. 

Test Cuvette Technique 

While the detailed studies of evapotran - 

258

spiration are being conducted, cuvettes will b e 

installed at various locations within the crow n 

to monitor the spacial variation of transpiration. 

In addition, the rates of transpiration , 

determined with the cuvettes, will be compared 

with the lysimetric transpiration rate t o 

determine how representative the cuvett e 

technique is for determining transpiration o f 

the tree . 

Test Meteorological Methods 

Meteorological methods such as energy balance, 

aerodynamic, and eddy correlation techniques 

have been used to determine evapotranspiration 

from agriculture crops. The 

methods need further testing in forestry be - 

fore they can be utilized on a wide scale . Th e 

evapotranspiration rates from the lysimeter 

tree, if representative of other trees, can b e 

utilized for testing the meteorologica l 

methods . 

Profiles of meteorological parameters suc h 

as radiation, temperature, humidity, carbo n 

dioxide, and windspeed will be monitored . 

These parameters will be utilized to calculat e 

the transfer coefficients (eddy diffusivities) o f 

sensible heat, latent heat, momentum, an d 

CO 2 . In addition, transpiration and photo - 

synthesis by layers will be calculated fro m 

these parameters . 

Canopy Interception of Precipitatio n 

Interception of precipitation either as rain , 

dew, or snow will be represented as a weigh t 

increase by the lysimeter . The amount of 

precipitation intercepted by the crown of th e 

tree can be studied by providing a waterproof 

cover for the soil surface to prevent th e 

throughfall from being recorded as a weigh t 

increase . The throughfall will. be measure d 

separately . If stemflow is measured, then th e 

rest of the weight increase can be attribute d 

to canopy interception . It is obvious that thi s 

method can be utilized to determine th e 

amount of rainfall and snowfall interceptio n 

in relation to intensity, duration, and previou s 

wettings . 

However, this technique may be more useful 

in determining the amount of moisture 

extracted from the atmosphere as dewfall-at 

present, an unknown quantity. During the 

summer months, energy budget calculations 

demonstrate that Douglas-fir trees, under the 

conditions of the site, would be stresse d 

within 2 to 3 weeks after the last rainfall if 

they transpired at a potential rate, unless transpiration 

is supplemented by evaporation of 

dewfall. Dewfall may be a very important 

parameter in the survival of these trees . 

Fertilization and Irrigation 

At the present time considerable interest i s 

being expressed in using forested areas for disposal 

of sewage sludge . The benefit of irrigation, 

sludge, or other fertilizers to tree growth 

could be tested with the lysimeter installatio n 

by comparing the photosynthesis and transpiration 

of the tree being studied with th e 

adjacent trees . 

Summary 

During the summer of 1970, a lysimeter 

container (366 cm in diameter and 122 c m 

deep) was built around the root mass of a 

28 m Douglas-fir tree at the Cedar Rive r 

Watershed near Seattle, Washington . This container 

was placed on top of the hydraulic 

transducer which is capable of weighing th e 

tree, soil, and container, a mass of 28,900 kg , 

with a sensitivity of 630 g or 0 .06 mm of 

water . A weather station was located on a 

33 .5 m tower adjacent to the lysimeter tree . 

The proposed uses of this installation are : 

(1) as an evapotranspirimeter, (2) as a standard 

for testing cuvettes, (3) as a standard for 

testing meteorological methods, (4) to stud y 

evapotranspiration in relation to soil moisture 

potential and atmospheric evaporative demands, 

(5) to study the shrinking and swellin g 

of the plant tissue in relation to evapotranspiration, 

(6) to determine interception of 

rainfall, snow, and dewfall by the crown, an d 

(7) to study the effects of irrigation and fertilizer 

upon the growth of the tree . 

259



ported by National Science Foundation Gran t 

No . GB-20963 to the Coniferous Forest Biome , 

U .S. Analysis of Ecosystems, International 

Biological Program. This is Contribution No . 



Colorado River Association. 1966 . Newsletter . 

Colo. River Assoc . Los Angeles, Calif. Sept. , 

p. 4 . 

Fritschen, Leo J . 1965 . Accuracy of evapotranspiration 

determination by the Bowe n 

ratio method. Bull. Int. Assoc. Hydrol . 

10(2) : 38-48 . 

and C . H . M . Van Bavel. 1963 . 

Experimental evaluation of models of 

latent and sensible heat transport over irrigated 

surfaces. Int . Assoc. Sci. Hydrol . Pub . 

62 : 159-171 . 

Gessel, Stanley P ., and Dale W. Cole. 1965 . 

Influence of removal of forest cover on 

movement of water and associated elements 

through soil . J. Soc. Am. Water 

Works Assoc . 57 : 1301-1310 . 

Harrold, L . L ., and F . R. Dreibelbis . 1951 . 

Agricultural hydrology as evaluated by 

monolith lysimeters . U .S. Dep. Agr . Tech . 

Bull. 1050 : 135 . 

Mcllroy, I . C., and D. E. Angus . 1963. Th e 

A s p e n d ale multiple weighed lysimeter 

installation . Australian Commonwealth Sci . 

& Ind. Res. Organ., Div. Meteorol. Phys . 

Tech. Pap. 14, p. 30 . 

Poulson, E . N., and J. T. Miller. 1952. Soi l 

survey of King County . USDA Soil Ser. 

1939, No. 31, p. 106 . 

Pruitt, W. O. 1963. Application of several 

energy balance and aerodynamic evaporation 

equations under a wide range of stability. 

Final Rep., USAEPG No . DA-36 - 

0390SC-80334, p . 107-124 . 

and D. E. Angus . 1960. Large 

weighing lysimeter for measuring evapotranspiration. 

Trans. Am. Assoc. Agric . 

Eng. 3(2) : 13-18 . 

Slatyer, R. O., and I . C. Mcllroy . 1961 . Practical 

microclimatology . Australian Commonwealth 

Sci. & Ind. Res. Organ . Various 

pagination . 

Tanner, C . B. 1967. Measurement of evapotranspiration 

. R. M. Hagan, H. R. Haise , 

and T. W. Edminster (eds .), In Irrigation of 

agricultural lands. Am. Soc . Agron . 

Monogr. 11 : 534-574 . 

Van Bavel, C . H. M., and L . E . Meyers . 1962 . 

An automatic weighing lysimeter . Agr . Eng . 

43 : 580-583, 586-588 . 

260



Initial steps in decomposition 

of Douglasfir needles 

under forest conditions 

Abstract 

P. L . Minyar d 

Graduate Student i n 

Forest Pathology 

Douglas-fir (Pseudotsuga menziesii) needle decomposition was evaluated after 6 and 12 months exposure t o 

forest floor conditions by characterizing solubility changes of saccharides, waxes and oils, and cellulose. Th e 

results lead to the hypothesis that primary processes of decomposition are initiated on the outside of the needl e 

with the solubilization of waxes and, as the waxes are depleted, the cellular constituents are solubilized. 

and 

C . H . Drive r 

Professor of 

Forest Pathology 




Introduction 

Approximately 4 percent of the biomass of 

the Douglas-fir consists of needles (Dic e 

1970) containing a portion of the carbon to 

be cycled within the ecosystem . Some genera 

of the Moniliales have been reported that are 

capable of utilizing cellulose and breakin g 

down needle waxes to release carbo n 

(Macauley and Thrower 1966 ; Martin and 

Juniper 1970) . The decomposition of needl e 

litter and the release of carbon in a forest is 

important in the nutrient cycling processe s 

within a conifer ecosystem . In an effort to 

study carbon cycling pathways and the par t 

played by the processes of the decompositio n 

of Douglas-fir needles the following investigation 

was conducted . 

Primary decomposition is defined a s 

changes in solubility of the needle constituents. 

Therefore to study decompositio n 

techniques were used to detect changes i n 

solubility . Simplistically, needles consist of an 

outer layer of cutin and waxes and an inne r 

cellular layer made up of cellulose and protoplasmic 

contents. Needle constituents wer e 

divided into three classes : (1) compound s 

extractable with hexane (primarily cuticl e 

waxes) (Martin and Juniper 1970), (2) those 

extractable with hot water (saccharides) 

(Campbell 1952), and (3) cellular components 

soluble in 1 percent HaOH (hemicellulose ) 

(Campbell 1952) . 

Materials and Methods 

Needles used in this study were collected i n 

the following manner . Screen traps wer e 

attached to a tower at random intervals i n 

height from just below the crown of the tree s 

to ground level . Needles were collected fro m 

the traps aseptically, placed in sterile containers, 

and transported to the laboratory for 

analysis and further study. In addition needle s 

from the traps were placed in nylon mes h 

bags on the forest floor and allowed to de - 

compose. After exposure for 6 and 12 months 

the needle samples were analyzed in th e 

following manner . 

Samples were oven dried for 24 hr . at 

105°C and stored in airtight containers unti l 

analyzed . 

Three 8-g replications of each treatmen t 

sample of needles were extracted with boiling 

n-hexane in a Soxhlet extractor for 6 hr . and 

oven dried for 24 hr. at 105°C to remove the 

hexane and weighed . The amount of material 

261

extracted was determined by the weight differences 

before and after extraction . 

Water soluble materials were determined in 

a similar manner using boiling water . 

As employed by Cowling (1961), a 1 per - 

cent NaOH solubility test was used to determine 

rate of cellulose decomposition . 

In an effort to isolate fungi involved in the 

decomposition processes needles were cultured 

in two ways : A portion of each sample 

was plated on 2 percent malt extract agar and 

onto an antibiotic medium consisting of 2 

percent malt extract agar plus 1 :30,000 Rose 

Bengal and 50 p .p.m. streptomycin. A second 

portion of each sample was washed in sterile 

water for 2 minutes and the needles plate d 

onto these same agars . The plates were incubated 

at 26°C for 7 to 14 days. Isolation s 

were made from the resulting fungi . 

Results 

Table 1.-Frequency of occurrence of 

genera of fungi isolated from 

needle sample s 

Months 

Genera 

0 Phomopsis sp . 

Mucor sp . 

Penicillium spp . 

Trichoderma spp . 

Botrytis sp . 

Graphium sp . 

6 Phomopsis sp . 2 

Mucor sp . 3 

Penicillium spp . 4 

Trichoderma spp . 4 

Aspergillus spp . 3 

12 Penicillium 4 

Trichoderma spp . 4 

Mucor sp . 4 

1 1 = Very infrequent 

2 = Infrequen t 

3 = Frequent 

4 = Very frequent 

Frequency o f 

isolation' 

Discussion 

Several genera of fungi have been reporte d 

to exhibit the capability to utilize waxe s 

(Martin and Juniper 1970) . From the data i n 

figure 1 it appears that fungal action on th e 

needle waxes causes these materials to be - 

come more soluble in hexane after a period o f 

time. There is a large weight loss in hexan e 

extractables after a six month period and a 

smaller weight loss after 12 months . I t 

appears from this trend that during the first 

6-month period, waxes are acted on initiall y 

during primary processes of decomposition . 

This was followed during the second 6-mont h 

period by the processes of increasing solubilization 

of cellulose and saccharides . 

Dice (1970) reported that annual production 

of organic matter for Douglas-fi r 

litter is 359 kg/ha . The density data in figure 

1 shows that at 12 months there is a 60 per - 

cent reduction of organic matter or approximately 

213 kg/ha . 

The fungi listed in table 1 are being studie d 

to determine what role they play in th e 

physiochemical processes illustrated in figure 

1 . 

1 . 6 

1 . 5 

1 . 4 

1 . 3 

1 . 2 

1 . 1 

1 . 0 

.9- 

.8 / 

.7 / 

. 6 

. 5 

.4 

.3 

. 2 

. 1 

0 

6 1 2 

MONTHS 

- DENSITY 

- HEXANE ELTRACTAN.ES 

- I3T WATER EXTRACTABLES 

1% MOM EXTRACTAHLES 

1 . 6 

1 . 5 

1 . h 

1 . 3 

1 . 2 

1 . 1 

1 . o 

Figure 1 . Density and weight loss of extractables of 

Douglas-fir needles after decomposition . 

. 9 

. 8 

. 7 

.6 

.5 

.4 

.3 

. 2 

.1 

0 

3 

262

These preliminary results suggest the following 

hypothesis . Initially the primary 

processes of decomposition occur on the surface 

of the needle with the decomposition o f 

waxes, and as the waxes are depleted, the cell - 

ular constituents are acted upon . 




Grant No . GB-20963 to the Coniferou s 



No . 42 from the Coniferous Fores t 

Biome . 


Campbell, W. G. 1952 . Wood chemistry . L. E . 

Wise and E . C . Jahn (eds), 2 : 1343 p . Ne w 

York: Reinhold Publ. Corp . 

Cowling, E . B. 1961 . Comparative biochemistry 

of the decay of sweetgum sap - 

wood by white rot and brown rot fungi . 

U .S. Dep. Agric . Tech . Bull. 1258 : 2 . 

Dice, Steven F . 1970. The biomass and nutrient 

flux in a second growth Douglas-fir ecosystem. 

165 p ., illus. Ph .D . thesis, on file at 


Macauley, B . J., and L. B. Thrower . 1966 . 

Succession of fungi in leaf litter o f 

Eucalyptus regans . Brit. Mycol . Soc. Trans . 

49 : 509-520 . 

Martin, J. T., and B . E . Juniper. 1970. Th e 

cuticles of plants . 347 p ., illus. New York : 

St. Martin's Press Inc . 

263



Seasonal and diurnal patterns of 

water status in Acer circinatum 

James P. Lassoie 

Research Assistant 

and 

David R . M. Scott 

Professor of Forestry 




Abstract 

Diurnal patterns of water status were monitored in Acer circinatum on clear days during the summer of 

1971 . Sap velocity, branch water potential, and relative leaf resistance to water vapor diffusion were used t o 

characterize the water status of a typical tree clump . Within the clump studied, air temperature and relative 

humidity were measured and used to calculate atmospheric vapor pressure deficits . Daily potential evaporation 

and soil moisture were calculated from U.S. Weather Bureau precipitation and air temperature data using th e 

Thornthwaite method. Results suggest that the seasonal course of plant water status is dependent upon changes 

in soil moisture levels, possible changes in leaf morphology, and differences in diurnal patterns of water statu s 

resulting from differences in atmospheric moisture demand . 

Introduction 

Vine maple, Acer circinatum Pursh, is a 

major subordinate species within the Coniferous 

Biome ranging from the southern mainland 

coast of British Columbia to the coasta l 

section of northern California . It is seldo m 

found growing with a single stem ; rather, it 

grows in clumps, forming a bushy mass . It is 

very shade-tolerant and is usually found unde r 

a tree canopy (Lyons 1964, Anderson 1968) . 

Despite the prevalence of vine maple in this 

area, it seems that little is known of it s 

physiological processes . 

Numerous authors have illustrated the con - 

trolling influence of internal water regimes o n 

physiological processes such as photosynthesis, 

respiration, translocation, and growt h 

of forest trees and herbaceous plants (re - 

viewed by Crafts 1968, Gates 1968, and 

Zahner 1968) . Also, transpiration, resulting i n 

water loss from the soil, influences the total 

hydrologic cycle and water balance of forest 

stands (Rutter 1968) . Therefore, understanding 

the water relations of a major subordinat e 

species, such as vine maple, may be quit e 

important to the study of productivity an d 

water balance of forests within the Coniferou s 

Biome. This study investigated diurnal pat - 

terns of internal water status in vine maple 

during the summer of 1971 . Besides supplying 

general information on the water relations of 

vine maple, the study also supplied basic data 

which will be used in the development of a 

plant water status submodel for the Coniferous 

Biome . The data presented are also 

representative of some of the types of data 

presently being collected for conifers within 

the Biome (Walker et al . 1972) . 

265

Methods 

The study was conducted at the Allen E . 

Thompson Research Center located about 5 5 

km southeast of Seattle at an elevation o f 

about 120 m in the foothills of the Washing - 

ton Cascades . The Research Center is abou t 

24 ha in size and is located on the western 

portion of the Cedar River watershed, in a 

40-year-old second-growth Douglas-fir, Pseudotsuga 

menziesii (Mirb.) Franco, plantation . 

Dominant trees are about 24 m in height . 

Principal subordinate species, besides vin e 

maple, are salal, Gaultheria shallon Pursh , 

Oregon grape, Berberis nervosa (Pursh) Nutt , 

sword fern, Polystichum munitum (Kaulf.) 

Presl., bracken fern, Pteridium aquilinum (L .) 

Kuhn var . pubescens Underw., and variou s 

species of moss . The climate is typical of lower 

elevations in western Washington . The aver - 

age temperature in July is 16°C and in January, 

3°C ; the average annual precipitation i s 

about 145 cm, with the majority falling during 

the winter months . The soil is a gravelly , 

sandy loam derived from Pleistocene glacial 

outwash. The vegetation, climate, geology , 

and soils in this area have been more completely 

described by Cole and Gessel (1968) . 

In the vine maple clump investigated, ther e 

were eight live stems, averaging about 3 .7 cm 

in diameter at breast height. Since the stems 

were quite supple, a spreading crown form resulted 

with stem lean in all directions . This 

straggly, crooked form is common for vin e 

maple growing in the shade (Lyon 1964) . The 

tallest erect stem in the clump was about 

10 m high. The clump was a moderately 

heavy Douglas-fir canopy . 

The water status of the clump was characterized 

by measurements of sap velocity , 

branch water potential, and leaf diffusio n 

resistance . Sap velocity of one major stem in 

the clump was estimated using a heat puls e 

velocity (HPV) meter and techniques described 

by Hinckley (1971) . Numerous investigators 

have shown that HPV is a valid 

indicator of relative changes in the transpiration 

rate of forest trees (Skau and Swanso n 

1963, Wendt et al . 1967, Hinckley 1971) . 

Readings were taken automatically every half 

hour and sap velocities calculated as described 

by Marshall (1958) . Branch water potentia l 

(>Vb) was estimated using the pressure chamber 

apparatus and techniques described b y 

Scholander et al. (1965) . Many investigators 

have successfully used this technique to estimate 

VI) of trees in the field (Waring an d 

Cleary 1967, Klepper 1968, Hinckley an d 

Scott 1971) . In this study, pressure chambe r 

readings of at least three twig samples were 

taken and averaged for each observation period 

during the day. Samples were obtaine d 

from shaded portions of the crowns betwee n 

the heights of about 0 .5 and 2.5 meters . Eac h 

sample had at least two leaves attached to th e 

twig. Leaf resistances (RL) to water vapor diffusion 

(transpiration) were measured with a 

leaf resistance meter and techniques describe d 

by Van Bavel et al . (1965) . Resistances wer e 

calculated using the meter sensitivity to temperature 

relationship described by Van Bave l 

et al. Therefore, though the readings prove d 

to be adequate in indicating relative diurna l 

differences, their absolute magnitudes may b e 

in error by a constant amount . Resistances fo r 

two or three intact vine maple leaves (obtained 

from similar locations as OD samples) 

were measured and their values average d 

periodically during the day . 

Early morning and evening readings wer e 

usually impossible because of the very high 

resistances involved causing serious errors in 

measurement (Van Bavel et al . 1965). Probably, 

these high resistances are primarily due 

to nearly complete stomatal closure durin g 

these periods . Readings of HPV, 1'b, and R L 

were taken between about 0600 and 160 0 

Pacific Standard Time (PST) . 

Within the clump studied, relative humidity 

and air temperature (Ta) were measured i n 

the shade with a sling psychrometer approximately 

every hour between about 0600 and 

1600 PST . These data were used to calculat e 

the vapor pressure deficit (VPD) of the atmosphere. 

Daily potential evaporation (PE) and 

daily soil moisture (SM) in the top 90 cm of 

soil were calculated for July, August, an d 

September from precipitation (P) and air temperature 

data by the Thornthwaite metho d 

(Thornthwaite and Mather 1957) as modifie d 

by Machno (1966) . The data used in these 

calculations were taken from the clima - 

266

I 

- 5 

-l o 

-1 5 

HPV 

1 

10 . 0 

I . 6 

E 

U 

7 . 5 

5 . 0 

1 . 2 

0 . 8 

2 . 5 

. 

0 . 4 

3 0 

Ta 

VPO 

I 1 L 

3 0 

2 0 

10 

1 0 

- I 

0800 1200 1600 0800 1200 1600 0800 1200 600 

Time of Day (PST ) 

Figure 1 . Daylight patterns of branch water potential (lib), heat pulse velocity (HPV), leaf resistance (RL), ai r 

temperature (Ta), and atmospheric vapor pressure deficit (VPD) for July 21, August 22, and September 22 , 

1971, at the Thompson Research Center, Washington . 

tological records of the U .S . Weather Bureau 

Station at Landsburg, Washington, located 

about 3 km west of the Research Center a t 

approximately the same elevation . 


Water status and environment data examined 

in this paper represent typical data collected 

on clear days during the summer of 

1971 . The days presented as examples are 

July 21, August 24, and September 22 (fig . 

1) . An expected overall pattern for plant 

water status on these days is discernible in the 

results. During the early morning "predawn," 

the air was cool and fairly humid ; RL was 

high suggesting that stomata were relatively 

closed ; 'b was high while HPV was quite low . 

As the day progressed, Ta increased, thus in - 

creasing VPD and intensifying the evaporative 

demand . As stomata began to open, probabl y 

in response to increasing light levels, R L 

values began to decrease . Since VPD had in - 

creased and stomata had begun to open, th e 

tree began to transpire causing an increase i n 

HPV readings. Continued water loss resulted 

in a decrease in t/b until about midday . In th e 

afternoon, RL increased because of stomatal 

closure probably due to decreasing lib . With 

this came a corresponding decrease in HP V 

and an increase in branch water potential . 

The summer was marked by an abnormall y 

wet June, with more than 10 cm of precipitation 

having fallen. A rainless period followe d 

during most of July and August ; only on 1 

day, August 21, did P surpass PE (fig . 2). Th e 

drought ended with a rainy period during th e 

first 2 weeks of September . This was followed 

by 2 weeks without rain before another stor m 

occurred near the end of September . The dry 

period during July and August produced a 

26 7

1 5 

FIGURE 2 

1 0 

18 0 

FIGURE 3 

u 

160 

ILO 

-, 

- 0 

= 120 

I00 

SM 

- 3 . 0 

July 6 July 26 

Aug 1 5 

Time of year (1971 ) 

Sept L Sept 2 L 

Figure 2 . Precipitation minus potential evaporation (P-PE) calculated by the Thornthwaite method for July , 

August, and September 1971 at the Thompson Research Center, Washington . 

Figure 3 . Soil moisture (SM) in the top 90 cm of soil calculated by the Thornthwaite method and "predawn " 

branch water potential (*1y ) ) for July, August, and September 1971 at the Thompson Research Center , 

Washington . 

steady drying of the soil from field capacit y 

to a water potential well below -1 .0 bar (fig . 

3). The rain in early September recharged th e 

soil to a point near field capacity . For the top 

90 cm of this soil, field capacity is near 17 . 7 

cm of water while the -1 .0 bar point is abou t 

14.2 cm (Knutsen 1965) . 

Since plants obtain water primarily fro m 

the soil and lose water to the atmosphere , 

their internal water status will depend greatly 

upon the SM supply and the atmospheric demand. 

The relationship of plant water status 

to SM is illustrated by examining "predawn " 

branch water potential (*fib) periodically 

throughout the study period . These measurements 

were taken just prior to sunrise whe n 

'b is usually the highest of the day and 

should most accurately reflect soil water po - 

tential (Slayter 1967, Klepper 1968, Waring 

1969). Therefore, a continuous decrease in 

*lPb should have occurred during the rainles s 

period in July and August because of the 

steadily decreasing SM (i .e ., decreasing soil and 

water potential) during this period . Quite the 

opposite was observed (fig . 3) . During the 

period, "predawn" branch water potential decreased 

steadily to a minimum on August 3 

and then continually increased through 

August . This was surprising as *fib seemed to 

respond quickly to any change in SM, as illustrated 

by the immediate decrease in I1b during 

the temporary rainless period in September. 

Also, maximal and minimal points withi n 

this general trend proved to be significantl y 

different at the 1-percent level (table 1) . The 

apparent anomaly is probably explainable i f 

268

Table 1 .- Tukey's w Multiple Comparison Test for seasonal "predawn" branc h 

water potentials (*fib) showing significant differences at th e 

1-percent level. Values underlined are not significantly different. 

Date 7/15 9/15 8/24 8/31 8/17 8/22 7/21 7/27 8/ 3 

*>!ib 0.8 0.8 1.2 1.0 1 .4 1 .4 1 .4 1 .7 2 . 2 

certain plant and environmental changes 

which occurred through the summer an d 

which decreased daily transpirational wate r 

losses are taken into account . 

The differences in daily plant water status 

which occurred on clear days during the stud y 

period would help explain the *0b trend observed 

(fig. 1). These differences seemed primarily 

due to a general decrease in the atmospheric 

demand through the study period (fig . 

2) . Generally, higher average Ta's and VPD' s 

and higher PE's occurred earlier in the summer, 

despite higher SM levels . A higher evaporative 

demand produced higher average HPV' s 

and greater water loss, thus developing lowe r 

branch water potentials . The later part of 

August was marked by relatively lower Ta's , 

VPD's and PE's, even though SM had de - 

creased. Hence, lower transpirational losse s 

occurred and higher 4/b developed . Durin g 

this period, a general increase in diffusion resistance 

was also evident . 

It is also possible that changes in leaf morphology 

during the summer could have helpe d 

produce a trend of increasing *fib during a 

period of continually decreasing SM . Chlorosi s 

in leaves, leading to necrosis before abscission , 

has been shown to be induced by drought , 

vascular disease, and normal senescence in 

woody and herbaceous plants (Kramer and 

Kozlowski 1960, Talboys 1968) . Such manifestations, 

characteristic of "drought hardening" 

in plants, can produce decreased transpiration 

and photosynthesis rates due to persistent 

stomatal closure even while leaves are fully 

turgid (Talboys 1968, Turner 1969) . By the 

middle of August, the leaves of the vine mapl e 

clump investigated had developed numerou s 

chlorotic spots; necrosis became evident during 

September. Therefore, increasing *1Pb during 

August could have been influenced by th e 

systematic reduction in transpiration in portions 

of the leaves . 

Conclusion 

The diurnal patterns of water status in vin e 

maple were examined for clear days during 

the summer of 1971 . As expected, the seasonal 

pattern of plant water status was de - 

pendent upon a combination of certain soil , 

plant, and atmospheric factors which affecte d 

water uptake and loss by the plants . Soi l 

water potential, which steadily decreased during 

the rainless periods, should have progressively 

limited water availability and resulte d 

in the development of lower "predawn" plan t 

water potentials. However, during a continuous 

drying cycle, changes in certain plant an d 

atmospheric factors seemed to influenc e 

water loss greatly . Early in the summer , 

greater water losses occurred due to highe r 

evaporative demands, and the plants were 

unable to recharge fully during the night eve n 

though adequate soil moisture was probably 

available . The development of chlorotic spots 

suggested that during a prolonged drought, o r 

as the leaves approached abscission, the ability 

of the entire clump to transpire decreased . 

Also, as the atmospheric demand lessened 

later in the summer, less water loss occurre d 

269

during the day. Therefore, despite low soil 

water supply, greater nighttime recharging 

was possible because the deficits developed 

were less . 




No . GB-20963 to the Coniferous Forest 



No . 43 to the Coniferous Forest Biome , 

IBP . 

The authors are also indebted to Mr . Peter 

K. McCracken and Mr . Peter S . Machno for 

their assistance in various phases of this study . 


Anderson, H . C. 1968. Growth from and distribution 

of vine maple (Acer circinatum) 

on Marys Peak, western Oregon . Ecology 

50(1): 127-130 . 

Cole, D . W., and S . P. Gessel. 1968. A pro - 

gram for studying the pathways, rates an d 

processes of elemental cycling in a fores t 

ecosystem . Coll. of Forest Resour . Contrib . 

No. 4, 54 p ., Univ. Wash., Seattle . 

Crafts, A . S . 1968. Water deficits and physiological 

processes . In T. T . Kozlowski (ed .) , 

Water deficits and plant growth . Vol. 2 . 

Plant water consumption and response, p . 

85-133 . New York : Academic Press . 

Gates, C . T. 1968. Water deficits and growth 

of herbaceous plants . In T . T . Kozlowsk i 

(ed.), Water deficits and plant growth . Vol . 

2 . Plant water consumption and response , 

p. 135-190 . New York : Academic Press . 

Hinckley, T. M . 1971 . Estimate of water flo w 

in Douglas-fir seedlings . Ecology 52(3) : 

525-528 . 

and D . R. M. Scott . 1971. Estimates 

of water loss and its relation to environmental 

parameters in Douglas-fi r 

saplings . Ecology 52(3) : 520-524 . 

Klepper, B . 1968 . Diurnal patterns of water 

potential in woody plants . Plant Physiol . 

43: 1931-1934 . 

Knutsen, S . K. 1965. Hydrologic processes in 

30- to 35-year-old stands of Douglas-fir an d 

alder in western Washington . 167 p .,illus . 

M .S. thesis, on file at Univ . Wash., Seattle . 

Kramer, P. J., and T. T. Kozlowski . 1960 . 

Physiology of trees. 642 p., illus. New 

York: McGraw-Hill . 

Lyons, C . P . 1964. Trees, shrubs, and flower s 

to know in Washington . 211 p., illus . 

Toronto, Can . : J. M. Dent and Sons . 

Machno, P . S. 1966. Water balance and soi l 

moisture studies in western Massachusetts . 

136 p., illus. M .S. thesis, on file at Univ . 

Mass ., Amherst . 

Marshall, D . C . 1958. Measurement of sap 

flow in conifers by heat transport . Plan t 

Physiol . 33 : 385-396 . 

Rutter, A. J . 1968. Water consumption b y 

forests. In T . T . Kozlowski (ed.), Water 

deficits and plant growth . Vol. 2. Plant 

water consumption and response, p . 23-84 . 

New York : Academic Press . 

Scholander, P . F., E . D . Hammel, E . D. Bradstreet, 

and E . A. Hemmingsen. 1965. Sa p 

pressure in vascular plants . Science 143 : 

339-346 . 

Skau, C . M., and R . H . Swanson . 1963 . An 

improved heat pulse velocity meter as a n 

indicator of sap speed and transpiration . J . 

Geophys . Res. 68 : 4743-4749 . 

Slayter, R . O. 1967 . Plant-water relationships . 

366 p ., illus. New York : Academic Press . 

Talboys, P . W . 1968 . Water deficits and vascular 

disease . In T . T . Kozlowski (ed .), Water 

deficits and plant growth . Vol. 2 . Plant 

water consumption and response, p . 

255-311. New York : Academic Press . 

Thornthwaite, C . W ., and J . R. Mather . 1957 . 

Instructions and tables for computing potential 

evapotranspiration and the wate r 

balance . Climatol. Drexel Inst . Technol . 

Publ. 10(3) : 1-311 . 

Turner, N. C. 1969. Stomatal resistance to 

transpiration in three contrasting canopies . 

Crop Sci. 9(3) : 303-307 . 

Van Bavel, C . H. M ., F. S . Nakayama, and W . 

L. Ehrler. 1965 . Measuring transpiration resistance 

of leaves. Plant Physiol . 40 : 

535-540 . 

Walker, R. B., D. R. M . Scott, D . J. Salo and 

K . L. Reed. 1972 . Terrestrial process 

270

studies in conifers-a review . In Jerry F . 

Franklin, L . J . Dempster, and Richard H . 

Waring (eds.), Proceedings-Research o n 

coniferous forest ecosystems-A symposium, 

p. 211-225, illus . Pac. Northwes t 


Waring, R. H . 1969. Forest plants of the east - 

ern Siskiyous: their environmental an d 

vegetative distribution . Northwest Sci . 

43(1) : 1-17 . 

and B. D. Cleary. 1967. Plan t 

moisture stress : evaluation by pressure 

bomb . Science 155 : 1248-1254 . 

Wendt, C. W., J . R. Runkles, and R . H. Haas . 

1967. The measurement of water loss by 

mesquite (Prosopis glandulosa var . glandulosa 

Torr.) using the thermoelectric method 

. Soil Sci. Soc. Am. Proc. 31(2) : 

161-164 . 

Zahner, R . 1968 . Water deficits and growth o f 

trees . In T . T. Kozlowski (ed.), Water deficits 

and plant growth . Vol. 2. Plant wate r 

consumption and response, p . 191-254 . 

New York : Academic Press . 

271



Development and testing of a n 

inexpensive thermoelectrically 

cooled cuvette 

David J . Salo, Department of Botan y 

John A . Ringo, Department of Electrical Engineerin g 

James H . Nishitani, Department of Botany 

an d 

Richard B . Walker, Department of Botany 



Abstract 

An economical temperature-controlled cuvette system has been developed to monitor CO 2 assimilation an d 

dark respiration in the crowns of mature Douglas-fir (Pseudotsuga menziesii) trees at the Cedar River Thompso n 

site. Temperatures in the Plexiglas assimilation chambers are controlled with thermoelectric coolers, and CO 2 

concentration changes are measured with a differential infrared gas analyzer. Preliminary field data sugges t 

cuvette temperatures can be maintained within ±1 °C of ambient even under high insolation conditions . 

Introduction 

The "cuvette technique" has been used fo r 

field assessment of carbon dioxide assimilation 

since the 1930's (Bosian 1933), but the 

findings of these studies have often bee n 

criticized. These criticisms of cuvette methods 

have usually resulted from the difficulties of 

maintaining a plant-chamber environmen t 

closely approximating that outside the en - 

closure . 

Cuvette Problems 

Problems encountered have included th e 

development of "deep" boundary layers alon g 

leaf surfaces with the concomitant establishment 

of abnormal C0 2 , vapor pressure, an d 

temperature gradients . Additionally, CO 2 

concentrations deviating markedly fro m 

ambient levels have frequently occurred an d 

water condensation in chambers and in ai r 

conducting lines has been of particular con - 

cern at night, during cold weather periods , 

and during periods of intense thermoelectri c 

cooling of cuvette bases . But probably th e 

most difficult problem to overcome has bee n 

that of chamber heating during periods o f 

high insolation . In efforts to achieve temperature 

control under these conditions, a variet y 

of chamber designs, fabrication procedure s 

and materials, and cooling techniques have 

been utilized . 

A common practice employed by many investigators 

has been the use of various thin 

plastic film "skins" to cover cuvette frames . 

Bourdeau and Woodwell (1965) used 8 mi l 

polyvinyl chloride (PVC), Ritchie (1969) use d 

2 mil PVC, and Hodges (1965) used 5 mil 

polypropylene . These efforts have met with 

varying degrees of success because plastics 

differ markedly in CO 2 permeability, as well as 

infrared and visible light transmittance . As an 

example, though the IR transmissivity o f 

polyethylene is relatively good its visible light 

transmittance is less than that of either Plexiglas 

(methyl methacrylate) or PVC . It is also 

273

permeable to both CO 2 and water vapor. On 

the other hand, visible transmittance of Plexiglas 

and PVC is very good while PVC infrared 

transmissivity is better than Plexiglas and glass , 

though not as good as that of polyethylene . 

Other methods of reducing temperature 

buildups have included the rapid conductio n 

(700 - 1,500 1 hr ') of air through assimilation 

chambers (Lerch 1965) and the circulation 

of water, water-ethanol, or waterethylene 

glycol solutions through transparen t 

jackets surrounding them . 

Lange's (1962) and Ritchie's (1969 ) 

" Klapp-Kuvettes" were attempts to reduc e 

temperature increases by alternately openin g 

and closing the chambers during experimenta l 

periods. Ritchie observed air temperatur e 

buildups of 5° to 7° C and leaf temperatur e 

increases of up to 17° C when his cuvette s 

were closed. By alternately opening and closing 

them, these overtemperatures could b e 

reduced by more than 50 percent . 

Currently, the most convenient method o f 

achieving cuvette temperature control appear s 

to be through the incorporation of thermoelectric 

coolers into chamber designs . Usin g 

the Peltier principle to regulate temperature , 

Siemens Corporation (Erlangen, Germany ) 

has manufactured a highly reliable and sensitive 

temperature and humidity controlle d 

system whose utility and flexibility have bee n 

demonstrated by Schulze (1970) . However , 

cost per unit prohibits Biome acquisition of a 

sufficiently large number for simultaneou s 

sampling of CO 2 exchange in different crow n 

locations, on different age classes of foliage , 

or on different plants . Therefore, routin e 

monitoring must be accomplished with a les s 

expensive system, reserving the Siemen s 

equipment available for "factors studies ." 

Prototype Development 

To complete a thorough daily and seasonal 

sampling program which will provide sufficient 

data to describe assimilation, dar k 

respiration, and transpiration in representativ e 

individual Douglas-fir (Pseudotsuga menziesii) 

at the Cedar River Thompson site, a syste m 

composed of several reasonably priced cooled 

cuvettes is considered highly desirable . As the 

first step in producing a functional system, a 

single prototype "two cuvette-power supply - 

temperature controller-fan controller " 

package (fig . 1) was fabricated from materials 

costing approximately $675 . 

Figure 1 . Prototype cuvette, power supply, temperature 

controller, fan controller package . 

Two 5-liter chambers constructed of 

3/8-inch Plexiglas are fitted with 1/8-inc h 

aluminum bases, removable tops, and smal l 

0-6 volt d .c . fans to mix the air and preven t 

boundary layer buildups . The aluminum floor 

of each chamber is securely bolted to the col d 

plate of an 80 watt Cambion (Cambridge , 

Mass .) Model 7250-1 "Forced Convectio n 

Thermoelectric Assembly" (TEA) . The Mode l 

7250-1 coolers operate in a range between 0 

and 18 volts and 0 to 6 amps (larger TEA' s 

are also available) and are powered by d .c . 

supplies each of which provides power fo r 

two cuvettes, two fans, two temperature 

controllers, and two thermoelectric coolers . 

Each control circuit has been develope d 

around a temperature sensing silicon resisto r 

(sensistor) bridge (originally diodes fashione d 

from transistors were used as temperature 

sensors), the imbalance in which determine s 

the current output to the thermoelectri c 

device . 

274

Experimental System 

Summer field trials for the cuvette wer e 

conducted at the Allen Thompson Researc h 

Center during 1971 . The open system (fig . 2) 

utilized included an air intake approximately 

25 m above the ground on a tower located 

3 m from that used as an access to study 

trees. Air was drawn to the ground, through a 

20-liter "mixing reservoir," to assure uniformity 

and prevent "noisy" carbon dioxid e 

analyzer output, and then passed through a 

manifold for distribution to two Model 506 R 

Reciprotor (Copenhagen) pumps . One of 

these pumped a comparison airstream to th e 

URAS II (Hartmann and Braun, Frankfurt ) 

infrared gas analyzer housed 46 m away in th e 

permanent site building and the other move d 

air to the cuvette located at 17 .2 m, back to 

the ground, through a flow meter, and then t o 

the differential analyzer. The attenuating 

reservoir, manifold, pumps, and flow mete r 

were located in a shelter at the tower base a s 

was the power supply-temperature controllerfan 

controller. A Honeywell recorder and the 

URAS II, however, were located in the site 

building. Two 0 .004-inch copper-constanta n 

thermocouples were used for temperature 

sensing inside the cuvette and one 22-gage 

thermocouple was positioned in a shaded 

location outside . 

In addition to cuvette data, meteorologica l 

data were provided by instruments located o n 

the adjacent meteorological tower . Scholander 

bomb and pressure infiltrometer (Fry 

and Walker 1967) samples were routinel y 

taken to provide estimates of water potentia l 

and stomatal aperture. These data were correlated 

with cuvette and meteorological dat a 

collected during each trial . Sampling wa s 

carried out at cuvette level and illustrativ e 

values reported here represent averages-tw o 

to three twigs for Scholander pressures (Ps ) 

and five to seven needles for stomatal infiltration 

pressure (Pstom) . 

During preliminary trials, the cuvette temperature 

was maintained within at least ± 

1 .5° C of ambient under conditions of high 

AI R 

INTAKE 

CUVETTE 

-1 

PUMPS 

MIXING --Lf 

RESERVOIR 

-r 

l~I 

_ 1 

FLO W 

METE R 

J 1 

1 

URASII 

FLO W 

METER 

MANIFOL D 

Figure 2 . Experimental system used during 1971 field trials . 

275

insolation (1 cal cm 2 min-1 ) and generally 

control was within ± 1° C . With the recent 

installation of aluminum heat sinks on th e 

cuvette floor, incorporation of "sensistor " 

(Texas Instruments) temperature sensors, an d 

additional electronic balancing of the system , 

improvement of this performance is 

anticipated . 

Representative 

Experimental Results 

Summer data obtained with the prototyp e 

system agree with those of other investigators 

in suggesting a close relationship betwee n 

assimilation, stress, and stomatal behavior. To 

illustrate this, results of a single day's run 

have been included (fig. 3) . 

On August 3 and other clear days sampled , 

light did not appear to limit photosynthesis ; 

maximum assimilation rates occurred between 

1030 and 1130 hours (PST) ; CO 2 uptake 

3 

2 

EXPT 7I-CR 2 

RUN 5 (BRANCH 2 ) 

8/3/7 

37 1 

DF 

decreased between 1130 and 1230 but wa s 

accompanied by continued stress increase an d 

only small changes in stomatal aperture ; varying 

degrees of afternoon stomatal closure 

followed and coincided with further depression 

of assimilation and relief of stress . 



in part by National Science Foundatio n 

Grant No . GB-20963 to the Coniferous Forest 





Bosian, G . 1933 . Transpirations-und Assimilationsbestimmungen 

an Pflanzen des Zentralkaiserstuhls 

. Z. Bot . 26 : 209-294 . 

° 

° CQ2 ASSIMILATION 

(mg CQ2 dry g' hr -' ) ° 

1 5 

10 

5 

Pstom 

`~_~ 

Ps 

erg'°~ 

[atmospheres) 

1 .5 

-°~-~ - 1.0 

0. 5 

0.5 

0.0 

3 0 

\ Re °2 

(cal cm miri 1 ) 

10 _ °-0-0--0-0~ 

CUVETTE TEMPERATURE ('C ) 

1 , 1 1 1 

06 

08 10 12 14 1 6 

TIME OF DAY 

Figure 3 . Representative data obtained during cuvette field trials . 

276

Bourdeau, P. F., and G. M. Woodwell. 1965. 

Measurements of plant carbon dioxide 

exchange by infrared absorption under con - 

trolled conditions in the field. In F. E . 

Eckardt (ed .), Methodology of plant 

ecophysiology, p . 283-289 . UNESCO . 

Fry, K. E., and R . B. Walker . 1967 . A pressure-infiltration 

method for estimatin g 

stomatal opening in conifers . Ecology 48 : 

155-157 . 

Hodges, J . D. 1965 . Photosynthesis in forest 

tree seedlings of the Pacific Northwest 

under natural environmental conditions . 

177 p. Ph.D . thesis on file, Univ . Wash . , 

Seattle . 

Lange, O . L. 1962 . Eine "Klapp-Kiivette" zur 

C0 2 -Gaswechselregistrerung an Blattern 

von Freilandpflanzen mit dem URAS . Ber. 

Deut. Bot. Ges. 75 : 41-50 . 

Lerch, G. 1965 . On the problem of cuvetteclimate 

. In Bohdan Slavik (ed .), Water 

stress in plants, p. 291-293 . The Hague : Dr . 

W. Junk N .V. Publishers . 

Ritchie, G. A. 1969. Cuvette temperatures 

and transpiration rates . Ecology 50 : 

667-670 . 

Schulze, E . -13 . 1970 . Der CO2 -Gaswechsel der 

Buche (Fagus siluatica L.) in Abhangigkeit 

von den Klimafaktoren im Freiland . Flora 

159 : 177-232 . 

277

Aquatic Process Studies 

279



Studying streams as a 

biological unit 

James R. Sedel l 

Department of Fisheries and Wildlife 


Abstract 

The case for viewing and studying streams from the standpoint of processors of materials and energy instea d 

of exporters from the forest is elaborated. The unique opportunity for stream and terrestrial biogeochemica l 

investigators to cooperate and be coinvestigators on the same stream systems is explored and specific programs 

are identified. 

The shift in emphasis from standing crop sampling and instantaneous ingestion and growth information t o 

process studies utilizing radioisotope material balance experiments and carbon flux experiments is explained . 

Short term experiments using radioisotopes as food markers are described and discussed as to their usefulness in 

determining the effect of food quality on ingestion rates and assimilation efficiencies . 

11 

The Stream as a 

Biological Unit 

The primary aquatic interface with th e 

terrestrial component occurs in small water - 

shed streams . There is a general consensu s 

that the stream itself probably contribute s 

little to the conservation of nutrients (compared 

to the total terrestrial system of whic h 

they are a part) since relative biomass level s 

are low and export is a pervasive feature o f 

streams, particularly during freshets . Suc h 

ideas, although partially correct, have helpe d 

to perpetuate out-moded views advanced by 

terrestrial ecologists around the turn of th e 

century concerning streams . The opening 

paragraph of a paper by Shelford and Edd y 

(1929) 1 is sadly appropriate in 1972 : 

Modern ecologists . . . have considered 

that the development of communitie s 

gives the clue to their dynamics and relations 

to each other. Most plant ecologists 

have usually assumed that there are n o 

permanent fresh water communities . 

This assumption is based upon negative 

evidence . Streams change their locations , 

and it is essentially their abandoned posi- 

' V. E. Shelford and S . Eddy. Methods for the 

study of stream communities . Ecology 10 : 382-392 , 

1929 . 

tions that become ponds and develo p 

into land communities . Streams are 

permanent as long as the existing climat e 

endures, and this is the condition under 

which land communities reach a climax . 

The state of the art of stream ecology ha s 

clearly upheld Shelford and Eddy's hypotheses 

that permanent stream communitie s 

exist, undergo successional development , 

reach and maintain a relatively stable condition, 

and manifest seasonal and annual differences, 

i .e., streams are bona fide biological 

units. To continue further, streams are highl y 

evolved biological units. As Hynes (1970) has 

pointed out, almost every taxonomic group of 

invertebrates is represented in, on or near th e 

substratum of lotic environments . In contrast 

to lakes and ponds there are several group s 

which occur only in lotic systems and more 

which reach their maximum development an d 

diversity there . This is quite probably a consequence 

of the permanence of streams as compared 

with lakes and ponds . Rivers rarel y 

disappear so they are not evolutionary trap s 

(Hynes 1970) . 

The attitudes of terrestrial ecologists 

toward streams have not basically change d 

since the paper of Shelford and Eddy (se e 

footnote 1). A recent symposium at Orego n 

State University on forest land uses and 

281

stream environments (Krygier and Hall 1971 ) 

indicates that attitudes toward streams b y 

policymakers and foresters may be changing . 

However, with the notable exceptions of 

Chapman (1966) and Lantz and Hall (personal 

communication), the analysis of the 

effects of logging on streams seldom consider s 

the stream directly from a biological point o f 

view . 

In perhaps the most complete study on a 

watershed to date, the Hubbard Brook Experimental 

Forest in New Hampshire (Liken s 

et al. 1970, Bormann et al . 1969), the capacity 

for streams to alter and process the various 

kinds and forms of chemicals was not considered. 

The various forms in which nitroge n 

enters a stream (organic, both dissolved an d 

particulate, and inorganic) and their fate s 

have not been carefully explored for eithe r 

undisturbed or clearcut watersheds . On e 

might expect that more nitrogen is lost to th e 

terrestrial portion but retained in the stream 

portion of the forest ecosystem than Liken s 

and Bormann have shown . Indeed, Fishe r 

(1970), working on the same watershed, has 

shown that 80 percent or more of the particulate 

fraction that enters a small stream is 

processed in the stream . Kaushik and Hyne s 

(1971) and Triska (1969), in studies on th e 

fate and residence times for leaves in streams , 

also indicate that processing and not export is 

the dominant process . 

Objectives of the Research on 

Andrews Watershed Streams 

The inclusion of streams in the Coniferou s 

Biome Study represents a unique opportunity 

for stream biologists and terrestrial biogeochemical 

investigators to cooperate on an d 

investigate the same stream systems, thu s 

exploring together long neglected problems . 

We have taken advantage of this opportunit y 

on the Andrews forest streams by emphasizing 

major terrestrial-aquatic couplings in our 

stream investigations such as : (1) obtainin g 

estimates of allochthonous inputs and exports 

both particulate and dissolved ; (2) documentation 

of successional changes in microflora 

and the identification of specific metaboli c 

activities carried out during leaf decomposition 

; (3) determining the significance of roo t 

return of nutrients via riparian vegetatio n 

with the aid of radioactive isotopes and mas s 

balance experiments; and (4) estimating th e 

importance of fish and amphibians as the to p 

carnivores in some of the small watershe d 

streams . 

The long-range objective of the Andrew s 

stream program is to elucidate the role of a 

stream in the functioning of the watershe d 

ecosystem. The inclusion of fish, which ar e 

certainly quantitatively significant in some o f 

the streams, makes it apparent that the othe r 

components and their relationships will b e 

defined as well . The determination of how th e 

relationships between productivity and structure 

at the various trophic levels are altered 

by various degrees of land use-in this instance, 

logging-will be a prime objective . Because 

of the relatively short-term nature of 

the research, logged vs . unlogged comparison s 

will be based primarily upon simultaneou s 

observations over a range of conditions rathe r 

than on the conventional before-and-after approach. 

The Andrews forest provides a wid e 

range of conditions from which to sample ; including 

virgin watersheds, watersheds wher e 

streamsides have been recently logged, and 

watersheds where streamside vegetation has 

grown after past logging. There are obvious 

problems in this comparative approach . We 

will have to look at a number of "undisturbed" 

streams to establish a sort of natura l 

variance in order to distinguish betwee n 

natural vs. imposed variation . 

The Andrews Approach 

to Stream Studies 

At present the emphasis is focused on th e 

role of the biota in the small streams in th e 

Andrews forest . We are presently determinin g 

the structure of the stream, its energy sources , 

and its energy processing components . This 

descriptive work is both necessary and tedious 

and will continue into year 3 . The method s 

and approaches being used are standard for 

282

this type of stream work with some modifications 

to cover the peculiarities of our specifi c 

streams . 

A significant problem not being dealt with 

in year 2 is the dimension of the dissolved 

organic fraction (DOM) and the utilization of 

that fraction. The DOM from leaf leachate , 

algal excretion, and soil solution represents a 

significant organic pool which turns ove r 

quite rapidly, even though a large amount o f 

rather refractory matter may be expected . K . 

W. Cummins (personal communication), 2 

attempting to model woodland streams in a 

deciduous forest, estimated that about 50 per - 

cent of the total DOM input was reflocculated 

into fine particulate organic matter (FPOM) . 

This occurred by flocculation around air 

bubbles, colloidal settling, and chemical precipitation 

. The other half was metabolized by 

microbes of which he estimated 50 percen t 

passed off as respiratory CO 2 and 50 percent 

went into production of microbial FPOM . 

The DOM compartment is a vital coupling between 

the dynamics of the stream biota an d 

such biome investigations as stream hydrolog y 

and biogeochemistry . 

Little is known concerning the man y 

changes in quantity and quality of DOM that 

are produced by the biological processes in 

and near streams . The functional role of DO M 

in stream metabolism is also understoo d 

poorly and that information which is availabl e 

is essentially circumstantial. Inferences hav e 

been drawn from physiological studies o f 

bacterial and algal cultures under laboratory 

conditions. Technological difficulties hav e 

delayed the transition from laboratory t o 

field investigations . 

The approaches planned, to tackle th e 

dynamics of DOM in forest streams, involv e 

an investigation of the processes of microbia l 

decomposition of allochthonous material , 

measuring the quantity and quality of DO M 

involved, and examining the fluxing betwee n 

physical and biotic compartments in an d 

along the streams . 

The variety of stream conditions in th e 

2 K. W. Cummins . Narrative for a stream energy 

budget model . Unpublished manuscript on file a t 

Kellogg Biological Station, Michigan State University , 

Hickory Corners . 5 p ., 1970 . 

Andrews watersheds provide excellent situations 

in which to measure the types an d 

amounts of DOM and hopefully develop a 

budget. Integral to the measurements will be a 

study of the functional microbial groups involved 

in the decomposition of the large 

organic material such as leaves, woody pieces , 

and fish carcasses. The expertise of an investigator 

with knowledge of the biochemistry of 

microbial decomposition will be required . 

The necessary techniques for studyin g 

stream processes and conceptualizing streams 

in general have not as yet been adequately 

developed for coniferous systems . However , 

major components with the forest strea m 

systems can be identified (fig . 1). Contrary to 

the opinion of others in the biome, th e 

periphyton and decomposer units can be use - 

fully approached as a "black-box ." The complexity 

can be used advantageously to measure 

disappearance of substrate and release o f 

products, and some understanding of ho w 

ecosystems work and respond to perturbations 

can be gained. Radioactive isotope 

experiments will allow us to look at coupled 

events and processes in the stream . The 

parameters will be determined by the kinetics 

of the isotope . We can refine the experiments 

to look at what features of the system deter - 

mine these parameters and how . Carbon-1 4 

and phosphorus-32 will be used in mass 

balance techniques developed by Saunder s 

(1969) and Saunders and Storch (1971 ) 

(closed chambers, short-term experiments 

with C 14 ) and in situ P 32 techniques developed 

by Nelson et al . (1969) and Elwood and 

Nelson (personal communication) . 3 Isotopes 

of selected elements will be incorporated into 

biogenic material such as salmon carcasses an d 

the pathways and turnover rates of these elements 

following decomposition will be investigated 

. 

In addition to the decomposition studies 

already mentioned, an estimate of the large 

woody pieces (greater than 1mm) above, in , 

and near the stream bed will be made . This 

will be done in cooperation with terrestrial 

biomass investigations on the Andrews forest . 

3 J. W. Elwood and D. J . Nelson . Measurement o f 

periphyton production and grazing rates in a strea m 

using a 32 P material balance method . Oikos (in press) . 

283

S, - 4ydrolo9i c 

S2 = I¢rr¢strIa l 

S3 = -A9uattc 

- p hysical -Processe s 

Figure 1 . Stream system categories and relationships . 

284

The primary production and energy source s 

estimations will utilize the aforementione d 

P32 material balance method . The techniqu e 

consists of computing a material balance o f 

P3 2 following a 30-60 min pulse release. Total 

standing crop and effective stream botto m 

area can be calculated by equating the quantities 

of P 3 2 per unit area of substrate on th e 

stream bottom and per unit weight of periphyton 

on these substrates to the total quantities 

of P 3 2 retained in the study reach of the 

stream . 

The retained P3 2 in the stream will be subsequently 

lost in both dissolved and particulate 

forms through three processes : (1) as P 3 2 

is replaced by stable phosphorus throug h 

metabolic turnover ; (2) periphyton containing 

P 3 2 may be sloughed from substrates and 

transported out of the study reach ; and 

(3) particulate P 32 released to the stream 

from primary and secondary consumers and 

drift of consumers . 

Rates of change in the various compartments 

can be estimated by monitoring th e 

P 3 2 in the periphyton and stream water over 

time. The production rate of periphyton an d 

the grazing rate of periphyton can then be 

estimated . 

Nutrients could be lost from the stream t o 

the riparian portion of the stream study re - 

search. The P32 technique allows one to 

measure this movement of nutrients from the 

aquatic to the terrestrial compartment . 

The rationale for carbon flux experiments 

are discussed in another paper in this symposium 

by Lighthart and Tiegs (1972) . We 

will not discuss this approach now, except t o 

say that we think the technique can b e 

adapted for use in stream research . 

An additional study in the year 3 progra m 

will be the role of mosses in the stream bed in 

fixing energy and cycling nutrients . Their 

associated invertebrate fauna make them an 

important compartment in the streams 

trophic structure . 

Studies on the production of benthic in - 

vertebrates will be completed . In the analysis 

of the data particular attention will have to be 

paid to "strategies of survival" used by the 

various functional groups of invertebrates . 

Population regulation must be expected to 

differ quite widely in response to the degre e 

of stability of the environment . Freshets in 

the Andrews forest occur frequently but unpredictably. 

Food supplies may disappea r 

almost entirely, and animals must be able t o 

withstand and recover from very great fluctuations 

in numbers . They must be adapted to a 

great lack of constancy . 

Special attention will be paid to the chironomid 

larvae, pupa and adults . This group 

which represents the greatest numbers, specie s 

and probably biomass of all of the aquati c 

insect groups, has been somewhat neglected i n 

the first 2 years . Using a foam gathering method 

the associated exuvia, pupae, and emergin g 

adult midges will be identified and quantified . 

W. P. Coffman (University of Pittsburgh) wil l 

be consulted on the identity of the midges . 

Feeding experiments with the major invertebrate 

species will determine ingestion an d 

assimilation rates as affected by the qualit y 

and quantity of food . 

Meaningful data concerning feeding habit s 

of selected consumers and the relative nutritional 

values of different constituents of th e 

periphyton are extremely rare . Gut analysis of 

insects are often misleading as to major sourc e 

of nutrition due to differential digestion an d 

assimilation rates . Sede11 4 

has reported a technique 

which allows one to determine if th e 

type of microflora on the natural substrate s 

affect the rates of ingestion of stream invertebrates. 

Substrates from the stream are eithe r 

differentially sterilized or are surface sterilized 

and then reinoculated with an algal , 

bacterial, or fungal community or any combination 

of the three treatments . The treate d 

substrate is then soaked for 1 to 3 hours in a 

stream water solution of Co b 0 . The co" is 

adsorbed onto the aufwuchs and the specifi c 

activity of the food substance determine d 

(mg/cpm) . The Co b ° is a gamma emitter an d 

can be easily detected without killing th e 

animal. This being the case sensitive experimental 

designs can be tried which use th e 

individual as a block, thereby eliminatin g 

4 J. R. Sedell. Feeding rates and food utilization of 

stream caddisfly larvae of the genus Neophylax 

Trichoptera : Limnephilidae) using °Co and 4 C. In 

~ . J. Nelson (ed.), Symposium on radioecology : Proceedings 

of the Third National Symposium at Oak 

Ridge, Tenn. May 8-10, 1971 . (In press .) 

285

ingestion variation between individuals an d 

ingestion variation between periods . By coupling 

this method with C 14 labeling of different 

compartments of the aufwuchs one can 

determine assimilation, and the major source 

of nutrition or at least that food which i s 

most easily assimilated . 

Predator-prey experiments have been de - 

signed to examine the roles of macroinvertebrates, 

amphibians, and fish in the stream , 

and their influence on the community composition 

and numbers of aquatic insects . Th e 

higher consumer portion of the study will als o 

relate the contribution of allochthonous an d 

autochthonous production to the elaboration 

of fish flesh and to determine how the relationship 

may be altered by logging or other 

land management practices . 

The role of amphibians in the consume r 

dynamics of the stream areas in which the y 

are located is being investigated by a member 

of the terrestrial consumer group . This effort 

will strengthen the aquatic-terrestrial couplin g 

studies . 

The Andrews stream program for years 3 

and 4 will be focused on aquatic-terrestria l 

couplings . Apart from viewing the stream as 

receiving the bulk of its energy from forest 

litter, the extent to which the stream delay s 

the export of minerals and nutrients or may 

return them to the forest is not known . 

Stream research during years 3 and 4 will be 

geared to provide some of these answers . 


The author and stream program are indebted 

to J. R . Donaldson who has been a 

constant source of personal encouragemen t 

and an active proponent of a vigorous strea m 

program. Our present stream program is an 

expansion of one very ably organized over the 

past 2 years by J . D. Hall. Our approach to 

studying streams as ecosystems reflects i n 

many ways the ideas of K . W . Cummins, wh o 

has followed the program and given generously 

of his ideas since its inception . 



No. GB-20963 to the Coniferous Forest 





Bormann, F . H., G . E. Likens, and J. S . 

Eaton . 1969 . Biotic regulation of particulate 

and solution losses from a forest ecosystem 

. Bioscience 19 : 600-610 . 

Chapman, D . W. 1966. The relative contributions 

of aquatic and terrestrial primary producers 

to the trophic relations of strea m 

organisms . In K . W . Cummins, C . A. Tryon , 

Jr., and R . T. Hartmen (eds .), Organismsubstrate 

relationships in streams, p . 

116-130 . Spec . Publ. No. 4, Pymatuning 

Lab. Ecol., Univ. of Pittsburgh, Pittsburgh . 

Fisher, S . G. 1970 . Annual energy budget of a 

small stream ecosystem : Bear Brook, West 

Thorton, New Hampshire. 97 p . Ph.D . 

thesis on file, Dartmouth Coll., Hanover . 

Hynes, H. B. N . 1970. The ecology of running 

waters . 555 p. Toronto, Can . : Univ . 

Toronto Press . 

Kaushik, N . K., and H . B . N . Hynes . 1971 . 

The fate of the dead leaves that fall int o 

streams. Arch. Hydrobiol . 68 : 465-515 . 

Krygier, J. T., and J. D . Hall. 1971 . Forest 

land uses and stream environment, proceedings 

of a symposium at Oregon State University, 

Oct. 19-21, 1970 . 252 p . Continuing 

Education Publication, Corvallis, Oreg . 

Lighthart, Bruce, and Paul E . Tiegs. 1972 . Exploring 

the aquatic carbon web . In Proceedings-research 

on coniferous forest ecosystems-a 

symposium, p . 289-300, illus . 


Portland, Oreg. 

Likens, G . E., F. H. Bormann, N . M . Johnson , 

and others. 1970. Effects of forest cuttin g 

and herbicide treatment on nutrient budgets 

in the Hubbard Brook watershed ecosystem 

. Ecol . Monogr . 40 : 23-47 . 

Nelson, D. J., N. R. Kevern, J . L . Wilhm, an d 

N. A. Griffith. 1969. Estimates of periphyton 

mass and stream bottom area usin g 

phosphorus-32 . Water Res . 3 : 367-373 . 

286

Saunders, G . W. 1969 . Some aspect of feeding 

in zooplankton, p . 556-573 . In Eutrophication 

: causes, consequences, correctives . 

Natl . Acad . Sci., Washington, D .C . 

and T. A. Storch. 1971 . Coupled 

oscillatory control mechanism in a plank - 

ton system. Nature New Biol . 230 : 58-60 . 

Triska, F. J. 1970. Seasonal distribution of 

aquatic hyphomycetes in relation to the disappearance 

of leaf litter from a woodlan d 

stream . 189 p. Ph .D. thesis on file, Univ. 

Pittsburgh, Pittsburgh . 

287



Exploring the 

aquatic carbon web 

Bruce Lighthart and Paul E . Tiegs l 

Western Washington State Colleg e 

Bellingham, Washington 9822 5 

Abstract 

An aquatic carbon web containing the six compartments dissolved inorganic carbon (DIC), phytoplank ton, 

zooplank ton, dissolved organic carbon (DOC), detritus, and chemoorganotrophic bacteria is discussed . Tentativ e 

methods are presented for measuring the pool size and kinetics about each compartment at one depth in the epiand 

hypo-limnions during the four seasons in Lake Washington, Lake Sammamish, Lake Chester Morse, an d 

Lake Findley. 

Introduction 

It is a primary concern to man to be able t o 

prepare predictive mathematical models of 

organic forms in aquatic systems if he is going 

to understand and ultimately develop the 

tools to wisely manage his water resources . To 

prepare such models, it is desirable to evaluate 

the pool size (standing stock) and flux (kinetics) 

of energy (Lindeman 1942) in all the 

fractions or compartments of the system . 

Models may be based on measurements o f 

energy or of several materials, e .g., carbon , 

phosphorous, and nitrogen ; however, both 

approaches are fraught with technical difficulties. 

Of the materials named, carbon i s 

recommended because it is a major component 

of all organic matter and is cycle d 

through a relatively well known web in th e 

biosphere . In addition, it is possible to follo w 

its movement with c" and other simpl e 

analytical techniques . 

In the aquatic realm, the carbon web ma y 

be thought to form a cyclical system made u p 

of a series of interrelated compartments . I n 

this discussion the carbon web will be limite d 

to a consideration of carbon in the followin g 

compartments (fig . 1): (1) Dissolved inorgani c 

carbon (DIC), (2) Phytoplankton, (3) Zoo - 

plankton, (4) Detritus, (5) Dissolved organi c 

carbon (DOC), and (6) Chemoorganotrophi c 

bacteria (Stonier et al . 1963). The cycle is 

initiated when dissolved inorganic carbo n 

species such as bicarbonate (or other chemical 

species in the aquatic carbonate syste m 

(Stumm and Morgan 1970)) are fixed durin g 

the photosynthetic process into organic 

matter by the photosynthetic organism (primarily 

phytoplankton and secondarily photo - 

synthetic bacteria) . Phytoplankton form the 

first link in the carbon food chain and ar e 

either eaten by herbivorous zooplankton, die 

and become part of the detritus compartment, 

or excrete dissolved organic carbo n 

(DOC). As much as 80 percent of the algal 

photosynthate may be excreted as glycolic 

acid and other compounds into the wate r 

(Hutchinson 1957, Fogg 1963, 1965, Hoo d 

1970). Subsequently, microorganisms may 

take up the DOC transported by the turbulen t 

water and metabolize it to inorganic carbo n 

forms. Saunders (1957, 1971) and Saunder s 

and Storch (1971) have found that th e 

h eterotrophic bacteria and phytoplankto n 

form a coupled oscillating pair in which th e 

bacteria increase in activity during the day in 

response to the daylight production and re - 

lease of the algal photosynthate into the surrounding 

waters . The herbivorous zooplankton 

(and carnivorous zooplankton) also 

1 Bruce Lighthart is Director, Institute for Fresh- , 

water Studies, and Assistant Professor, Department of 

Biology . Paul E . Tiegs is with Institute for Freshwater 

Studies . 

289

ATMOSPHERI C 

CARBO N 

DIOXIDE 

■ 

DISSOLVED 

INORGANI C 

PH'TOPLAN KTQN ` ( 

_ BACTERI A 

290

contribute to detrital compartment as a resul t 

of their fecal deposits or their death and decomposition. 

Ultimately DOC and detrital 

carbon are decomposed by bacteria or zoo - 

plankton feeding on bacteria/detritus mixtures, 

to dissolved inorganic carbon as a result 

of respiratory oxidative decompositio n 

processes . 

In nature the standing stock in each compartment 

is the dynamic equilibrium level 

attained as a result of input/output rate s 

about the compartment . The quality of th e 

standing stock will also vary as the input / 

output mechanisms change . 

With the advent of the carbon-14 tracer 

techniques, the transports of carbon to and 

from each compartment in an aquatic ecosystem 

is possible . For example, Steemann- 

Nielsen (1952) initiated such studies with C' 4 

bicarbonate to estimate the rate of primar y 

production in the ocean . Goldman (1963 ) 

outlined the use of this method in fres h 

water . 

Sorokin (1966), Shuskina and Monako v 

(1969), and Johannes and Satomi (1967) use d 

labeled phytoplankton to determine the rat e 

at which zooplankton feed and process thei r 

waste . The production of dissolved organic 

carbon (DOC) by phytoplankton (Fogg 1965, 

Saunders and Storch 1971) and uptake by 

zooplankton and bacteria have been success - 

fully measured using carbon-14 tracers. 

Hobbie and Wright (1965) and Wright an d 

Hobbie (1966) refined the enzyme saturatio n 

technique of Parson and Strickland (1962) a s 

measured by a carbon-14 tracer, to evaluat e 

the pool size and potential flux of dissolve d 

organics through the bacterial (heterotrophic ) 

compartment in the aquatic web . With 

carbon-14 labeled detritus Sorokin (1966 ) 

measured assimilation by zooplankton . H e 

prepared the detritus by homogenizing radio - 

actively labeled algal cultures . Others hav e 

used autoclaved, labeled natural and culture d 

algal populations 2 (Bell and Ward 1970) . 

Labeled detritus may also be prepared, b y 

foam separating C-14 labeled dissolve d 

organics from the liquid phase (Baylor an d 

Sutcliff 1963) . However, there is evidenc e 

2 G . W. Saunders, personal communication . 

that this method will work only in sea water . 

For the first time, Saunders 3 recently combined 

the aforementioned methods to measure 

simultaneously all of the previously liste d 

carbon fractions and their fluxes in a "Compartment 

Analysis" scheme (Patten 1968 , 

Atkins 1969). He has made this type of 

analysis at one location and one instant in 

time at several lakes . 

Compartment Analysis, in essence, is performed 

by establishing a series of known fractions 

or compartments in an experimenta l 

confine (e .g., glass carboy or plastic bag) and 

following a tracer such as carbon-14 as a func - 

tion of time as it proceeds through each compartment 

in the confined system . Measurement 

of the radioactive carbon as it passe s 

through each compartment can be used t o 

evaluate the rate of incorporation of carbo n 

into each compartment in succession, and b y 

isotopic dilution or total carbon conten t 

analysis, the pool size estimated . These data 

may be used to describe mathematically th e 

carbon change in compartments as a functio n 

of time, e .g ., 

d( q 1) _ - k l q l and 

dt 

d( q 2 ) 

dt =k l q 1 -k 2 q 2 

where (4 1 and q 2 are the specific activities o f 

the first and second compartments and k l 

and k2 are the rate constants into each compartment. 

The compartments could very wel l 

be phytoplankton and zooplankton or DOC , 

bacteria and zooplankton, etc. The specifi c 

activities rather than tracer amounts will b e 

used in mathematical models . 

It must be emphasized that the confinement 

of an aliquot of the sampled water during 

the compartment analysis will yield value s 

prevailing for (1) standing stock at the time o f 

sample confinement, and (2) uptake value s 

for that particular sample during the incubation 

period . Thus, it is very important whe n 

and where test water samples are chosen . For 

example, if lake water were collected for a 

compartment analysis of the carbon fraction s 

during maximum expected respiratory and 

3 Unpublished data . 

291

photosynthetic activity with respect to sea - 

son, time of day, and depth, one might expec t 

measurements to approach the maximu m 

values that these variables would attain in 

situ . 

Reviews and key papers concerning variou s 

aspects of the carbon cycle in the aquati c 

realm include the following : zooplankto n 

physiology (Corner and Coney 1968), primary 

productivity (Ryther 1963, Steemann - 

Nielsen 1963, and Strickland 1965), dissolve d 

organic carbon (Provisoli 1963, Duursma 

1965 ), detritus (Parsons 1963), bacteria 

(Wood 1965, Brock 1966, and Lighthart 

1969), and food chains (Riley 1963) . Also 

some detailed methods are given in IBP Hand - 

books 12 and 17, Primary Production i n 

Aquatic Environments and Secondary Productivity 

in Fresh Waters (Edmondson and 

Winberg 1971) . 

To this point, the discussion has been wit h 

what might be called a primary system, the 

components in the carbon web . This system is 

controlled by another [or] secondary system , 

the physical and chemical environment . Som e 

of the environmental factors operating on th e 

primary system are listed in table 1 . 

Because factors in the secondary syste m 

regulate changes in the primary system, meas - 

urements of the primary system may allow u s 

to interpret the major controlling variables o f 

the water bodies . 

Table 1 .-Some environmental factors operating 

on the aquatic carbon we b 

Temperature 

Light intensity, quality, and duration 

pH 

Trace elements 

External inputs of organic matter , 

i.e ., nutrients, antibiotics, and 

growth factors 

Phosphorou s 

Silica 

Calciu m 

Environmental confine s 

Water circulation 

Purpose 

It is proposed that, when possible, one station 

will be occupied on each of four lakes i n 

the Cedar River watershed : Lake Washington , 

Lake Sammamish, Lake Chester Morse, an d 

Lake Findley, during the four seasons of th e 

year. While occupied, the flux and pool size s 

of six carbon compartments will be measured 

in the epi- and hypo-limnions . The purpose of 

this investigation is to determine whether the 

four lakes are distinguishable in terms of the 

fluxes and pool sizes of the six carbon compartments 

with the methods outlined below . 

Methods 

Carbon pools in each compartment will b e 

made by : (1) estimating the protoplasmic 

carbon pool from cell volume, and (2) from 

infrared spectrophotometric measurement o f 

oxidized cellular carbon . Carbon flux between 

compartment pools will be estimated by 

following the transfer of carbon-14 added to 

specific compartments . 

Carbon Pool Analysi s 

It is necessary to determine the "cold " 

carbon pool size, both to evaluate its standin g 

stock, and in calculating specific activities . 

Inorganic carbon in the lake water will b e 

measured either by determining the tota l 

alkalinity (American Public Health Association 

1971), pH, and temperature, and the n 

consulting the appropriate tables (Saunders e t 

al. 1962) for total available inorganic carbon , 

or by direct measurement using an infrared 

CO 2 analyzer (Menzel and Vaccaro 1964) . 

The dissolved organic carbon in the water 

samples will be determined on the filtrate of a 

O.22µ Millipore filtered sample of the test 

water using the method of Menzel and Vaccaro 

(1964) . Carbonates will be driven off th e 

filtrate prior to analysis by sparging th e 

phosphoric-acid-acidified filtrate with nitrogen 

gas . 

The carbon content of the planktonic algae 

will be made either from dry weight estimation 

by calculation from cell counts an d 

29 2

volume measurement, realizing that the protoplasm 

has a specific gravity of 1 .1, and is 

80-percent water, the carbon content is 5 0 

percent of the dry weight (Lund 1965) ; or by 

regression with chlorophyll-a measuremen t 

(Steemann-Nielsen and Jorgenson 1968) ; or 

by both . 

(1) 

Algal carbon 

(protoplasmic volume) X 

(protoplasmic specific gravity ) 

X (protoplasmic P .C. water 

content) X (P.C. dry weight 

carbon) 

Zooplankton carbon will be measured using 

the Menzel and Vaccaro (1964) method of 

individually picked or 15012 mesh net filtered 

samples . 

Bacterial carbon will be estimated as th e 

product of cell volume (assumed 0 .5 X 112 ) 

times cell density (1 .1) times water content 

(80 percent) times the dry weight carbon con - 

tent (50 percent dry weight) of fluorescence 

(Strugger 1948) phase contrast microscope 

enumerated cells . 

Detritus carbon will be evaluated as the 

sum of the bacterial and algal dry weight, an d 

seston ash weights, subtracted from the total 

seston dry weight . The weight of the total 

seston will be determined via combustion o f 

quadruplicate 0.812 tared glass fiber filtere d 

and dried water samples less zooplankton . It 

is assumed that 50 percent of the volatil e 

seston is carbon (equation 2) . 

(2) 

Detritus carbon (seston D.W. - zooplankton) - 

(algas D .W. + bacterial D.W. + 

seston ash weight) X (P .C . 

seston D .W. carbon ) 

Detritus carbon analysis will be attempted 

by infrared analysis of combusted total sesto n 

after the zooplankton have been remove d 

(equation 3) . 

(3 ) 

Detritus carbon = (total seston carbon - 

zooplankton) - (algal carbo n 

+ bacterial carbon ) 

Carbon Flux Analysis 

Carbon flux will be measured as the trans - 

Table 2.-List of compartment systems isolated for carbon flux measurements 

Bottle Added tracer Compartment system 

1 NaH 14 CO 3 Dissolved inorganic carbon 

Phytoplankto n 

Dissolved organic carbo n 

2 C-starch (U14 ) Dissolved organic carbo n 

with or without 

Bacteri a 

algal hydraulysate 

Dissolved inorganic carbon 

3 NaH 14 CO 3 Dissolved inorganic carbo n 

Phytoplankto n 

Zooplankto n 

4 Experimentally Dissolved organic carbo n 

generated 

Bacteri a 

DOC-Cl 4 

Zooplankto n 

(Dissolved inorganic carbon) ? 

5 Experimentally Detritu s 

generated 

Zooplankto n 

Detritus-C 14 

(Dissolved inorganic carbon)? 

29 3

18L SAMPLE LAKE WATER 

LIGH T 

ACIDIFIED TO pH 3 

WITH H 3 P04 AN D 

SPARGE WITH N 2 

1 

ADD ALIQUOT T O 

COCKTAIL AND COUNT 

Figure 2. Flow diagram of methods to be used to determine carbon flux from DIC to phytoplankton to DOC , 

and detritus to zooplankton . 

294

fer of the indicated carbon-14 tracer labele d 

substrate through the isolated systems tabulated 

in table 2 . Each of the five systems will 

be isolated in either glass carboys, if it is a 

light-requiring system, or in blacked plasti c 

collapsible bottles, if a dark system. Th e 

bottles will be incubated in situ . 

The flux of carbon from DIC to DO C 

through the phytoplankton will be determined 

as the uptake during a daylight perio d 

of a carbonate spike added to a glass carboy 

of lake water incubated in situ . The zooplankton 

will be filtered out of this water with a 

0.1 mm mesh Nitex net and used in anothe r 

experiment. Samples will be withdraw n 

hourly from the submerged bottle through a 

tube to the water surface . The Millipore filtered 

(0 .22/1 porosity) samples will be used t o 

separate the phytoplankton from the filtrate . 

The filtrate containing the DOC c", after 

acidification and nitrogen sparging, will b e 

dried in a scintillation vial and counted afte r 

moistening . Algal respiration will also b e 

evaluated either as the slope of the phytoplankton 

carbon-14 loss immediately afte r 

darkening the bottle, or in separate classica l 

light/dark bottle measurements (fig. 2) . 

Upon completion of the foregoing experiment, 

a portion of that carboy of lake wate r 

will be used to evaluate the carbon flux fro m 

DOC generated by the phytoplankton in it, t o 

DIC via bacteria . This water will be cleared of 

organism and labeled DIC by first Millipor e 

filtration and then nitrogen sparging . Subsequently, 

fresh lake water , less its zooplankto n 

from the original sample site will be added to 

the " clean " hot-DOC containing water 

incubated in situ in the dark. Particulate (i .e . , 

bacteria) and DIC containing activity in th e 

bottle will be monitored for a further 2 4 

hours (fig. 3) . 

The zooplankton initially filtered from th e 

water for the DIC to algae to DOC experiment 

(the first experimental description) wil l 

be added to a darkened plastic carboy containing 

sterile radioactive detritus so the 

carbon flux from the detritus to zooplankton 

and possibly to DIC compartments may b e 

measured . After 4 and 8 hours, one-half of 

the carboy will be filtered and the animal s 

counted numerically and radioactively. The 

three zooplankton data points so obtained 

will be graphed to evaluate the uptake rate 

(fig. 2) . 

The exact method of sterile detritus production 

awaits further experimentation to 

evaluate the chemical composition and 

optimum zooplankton feeding rate on 

detritus produced in several ways . 

The flux measurement of carbon fro m 

phytoplankton to zooplankton will be evaluated 

in a lighted glass carboy like the on e 

described for DIC to algae to DOC measurements 

except that the zooplankton will no t 

be removed from the water . Hourly phytoplankton 

samples, and 4- and 8-hour sample s 

of one-half the vessel will be filtered for 

subsequent numerical and radiological zoo - 

plankton counts . The mean specific activity 

of the phytoplankton food source for th e 

zooplankton may be evaluated as the mea n 

specific activity of algal carbon during th e 

experiment (fig . 4) . 

Finally, the flux of dissolved organic 

carbon through bacteria to zooplankton wil l 

be determined as the rate of carbon 1 4 

labeled DOC (i .e., starch and/or C 14 labeled 

algal protein hydraulysate) uptake by th e 

bacteria (particles passing through a 0 .1 Nitex net) incubated in situ in a 

darkened plastic carboy of lake water. 

Samples will be withdrawn hourly for th e 

bacterial-uptake and one-half the carboy filter 

after 4 and 8 hours for the zooplankto n 

measurements. Zooplankton will be filtere d 

out onto a 0 .1-mm mesh Nitex net (fig . 5) . 

All carbon-14 measurements will be made 

on a Packard Tri-Garb Model 3375 scintillation 

counter using the channel s ' ratio method . 

Schindler's (1966) method for phytoplankton, 

and Ward et al .'s (1970) method for zoo - 

plankton, detritus, and bacteria will be used . 

More details of proposed methods may be 

found in Edmondson and Winberg (1971) , 

and Parsons and Strickland (1965) . 

The numerical methods that will be used i n 

the analysis of the data generated by thi s 

study will be presented in future communications, 

but are anticipated to include : (a) statistical 

evaluation of the differences betwee n 

flow and pool sizes of the biome test lakes as 

295

1500 m l 

LAKE WATER 

2 L FRO M 

PHYTOPLANKTO N 

EXPERIMEN T 

Figure 3. Flow diagram of methods to be used to determine carbon flux from DOC to bacteria to DIC . 

296

18L LAKE WATER 

LIGH T 

50m1 HOURLY 

DURING DAYLIGHT 

1/2 CARBOY AT 

4 AND 8 HOUR S 

WASH GUTS 1 H R 

IN JUST TURBI D 

YEAST SUSPENSION 

NUMERICA L 

(AND/OR SEPARATE SPECIES ) 

AND RADIOACTIVE COUNT S 

ZOOPL AN KTO N 

Figure 4. Flow diagram of methods to be used to determine carbon flux from DIC to phytoplankton t o 

zooplankton . 

297

18L LAKE WATE R 

PRESERVE WIT H 

FORMALI N 

SPARGE WITH N 2 

10 MINUTES 

DRY ALIQUOT, AD D 

COCKTAIL AND COUN T 

D .O .C . 

NUMERICAL 

(AND OR SEPARATE SPECIES ) 

AND RADIOACTIVE COUNT S 

ZOOPLANKTO N 

Figure 5 . Flow diagram of methods to be used to determine carbon flux from DOC to bacteria to zooplankton . 

298

a function of time and space, (b) carbo n 

balances within the web and between the tw o 

test depths station, (c) carbon transfer efficiencies 

between compartments, (d) synthesis 

of our data together with other biome measurements 

that have small time and space 

resolutions, and (e) carbon flow and pool size 

input to mathematical models in preparation . 


The authors want to thank Professor Freid a 

B. Taub for her active encouragement of this 

study, and Dr. G. W. Saunders, Jr., whos e 

work generated this scheme of methods out - 

lined in this paper . 






No . 46 to the Coniferous Forest 

Biome and Contribution No . 16, Institute for 

Freshwater Studies, Western Washington State 

College . 


American Public Health Association. 1971 . 

Standard methods for the examination o f 

water and waste waters . 874 p . Washington , 

D .C . 

Atkins, G . L . 1969. Multicompartmen t 

models for biological sciences . 153 p . 

London: Methulen and Co . Ltd . 

Baylor, E . R., and W . H. Sutcliff, Jr . 1963 . 

Dissolved organic matter in seawater as a 

source of particulate food . Limnol . 

Oceanogr. 8(4) : 869-871 . 

Bell, R. K., and F . J. Ward, 1970 . Incorporation 

and organic carbon by Daphnia pulex . 

Limnol. Oceanogr. 15(3) : 713-726 . 

Brock, R. D. 1966. Principles of microbia l 

ecology. 306 p. Englewood Cliffs, N .J . : 

Prentice-Hall . 

Corner, E. D. S., and C . B. Coney . 1968 . Biochemical 

studies on the production of 

marine zooplankton . Biol. Rev. 4(3) : 

393-426 . 

Duursma, E . D. 1965. The dissolved organic 

constituents of sea water . In J. P. Rile y 

(ed.), Chemical oceanography, 433 p . Ne w 

York: Academic Press . 

Edmondson, W . T., and G . G . Winberg (eds .) . 

1971 . A manual of methods for the assessment 

of secondary productivity in fres h 

waters. IBP Handb . No . 17, 386 p . Philadelphia 

: F. A. Davis Company . 

Fogg, G. E. 1963 . The role of algae in organic 

production in aquatic environments. Brit . 

Phycol . Bull. 2(4) : 195-204 . 

1965. Algal cultures and phytoplankton 

ecology . 126 p. Madison : Univ . 

Wis . Press . 

Goldman, C. R. 1963. Measurement of primary 

productivity and limiting factors i n 

freshwater with carbon-14 . In M. S. Doty 

(ed .), Proc . Conf. on primary productivit y 

measurement, marine and freshwater . AE C 

Inf. T .I .O . F633, p . 103-113 . 

Hobbie, J . E., and R. T . Wright . 1965 . Bio - 

assay with bacterial uptake kinetic : glucose 

in freshwater. Limnol. Oceanogr. 10(3) : 

471-474 . 

Hood, D . W. 1970. Symposium on organi c 

matter in natural waters . Inst. Mar. Sci . , 

Univ . Alaska Occas . Publ . 1, 625 p . 

Hutchinson, G . E . 1957 . A treatise o n 

limnology . Vol. I, 1015 p. New York : Joh n 

Wiley & Sons, Inc . 

Johannes, R . E., and M. Satomi. 1967. Measuring 

organic matter retained by aquatic 

invertebrates. J. Fish. Res. Board Can . 

21(11) : 2467-2471 . 

Lighthart, B . 1969. Planktonic and benthic 

bacteriovorous protozoa at 11 stations i n 

Puget Sound and adjacent Pacific Ocean . J . 

Fish. Res . Board Can . 26 : 299-304 . 

Lindeman, R . L. 1942. The trophic-dynami c 

aspect of ecology . Ecology 23 : 399-418 . 

Lund, J. W. G. 1965. The ecology of th e 

freshwater phytoplankton. Biol. Rev . 40 : 

231-293 . 

Menzel, D. W., and R. F. Vaccaro. 1964. The 

measurement of dissolved organic and particulate 

carbon in water . Limnol . Oceanogr . 

9(9): 138-142 . 

Parsons, T . R. 1963. Suspended organic mat - 

ter in seawater . In M. Seres (ed .), Progres s 

299

in oceanography, Vol . 1, p . 205-239. Ne w 

York: Pergamon . 

and J. D. H. Strickland. 1962 . 

On the production of particulate organic 

carbon by heterotrophic processes in se a 

water. Deep-Sea Res . 8 : 211-222 . 

Patton, B . C. 1968. Ecological modeling with 

analog and digital computers . Am. Inst . 

Biol . Sci., 40 p . 

Provasoli, L. 1963. Organic regulation of 

phytoplankton fertility . In M . N. Hill (ed .) , 

The seas, p . 165. New York : Interscienc e 

Publ . 

Riley, G. A. 1963. Theory of food chain relationships 

in the ocean . In M. N . Hill (ed .) , 

The seas, p . 438. New York : Interscience 

Publ . 

Ryther, J . H. 1963. Geographic variations i n 

productibility . In M. N. Hill (ed .), The seas, 

Nauk. USSR 18(4) : 955-958 . 

Sorokin, Y . I. 1966 . Carbon-14 method in the 

study of the nutrition of aquatic animals . 

Hydrobiology 51(2) : 209-224 . 

Stanier, R. Y., M . Doudoroff, and E . A . 

Adelberg . 1963. The microbial work . 75 3 

p. Englewood Cliffs, N .J . : Prentice Hall . 

Steemann-Nielsen, E . 1952. The use of radio - 

active carbon for measuring organic production 

in the sea. J. Conserv. Intern. 

Explor . 18 : 117-140 . 

1963. Productivity, definitio n 

and measurement . In M. N. Hill (ed.), The 

sea, p. 129. New York : Interscience Publ . 

and E. G. Jorgensen . 1968. The 

adaptation of plankton algae. I. General 

part. Physiol. Plant. 21 : 401-413 . 

Strickland, J. D. H. 1965. Production o f 

organic matter in the primary stages of the 

p. 347. New York : Interscience Publ . 

marine food chain . In J. P. Riley and G . 

Saunders, G . W. 1957 . Interrelations of dissolved 

organic matter and phytoplankton . 

Skirrow (eds.), Chemical oceanography, p . 

712 . New York : Academic Press . 

Bot. Rev. 23(6) : 389-410 . 

and T. R. Parsons . 1965 . A 

1971 . Coupled oscillations in 

the metabolism of bacteria in lakes . 

(Abstr.) Am. Soc. Microbiol . (Bacteriol . 

Proc.), p . 53 . 

and Thomas A . Storch . 1971 . 

Coupled oscillatory control mechanism in a 

planktonic system . Nature New Biol . 

230(10) : 58-60 . 

, F. B. Trama, and R. W. Bachmann. 

1962. Evaluation of a modified 

carbon-14 technique for shipboard estimation 

of photosynthesis in large lakes . Univ . 

Mich., Great Lakes Res. Div . Publ . No . 8, p . 

61 . 

Schindler, D . W. 1966 . A liquid scintillation 

method for measuring carbon-14 uptake i n 

photosyntheses. Nature (London ) 

211(505) : 844-845 . 

manual of sea water analysis . Fish . Res . Bd . 

Can . Bull. 125, p. 203 . 

Strugger, S. 1948 . Fluorescence microscope 

examination of bacteria in soil . Can.J. Res . 

26 : 188-193 . 

Stumm, W ., and James J . Morgan. 1970 . 

Aquatic chemistry : an introduction emphasizing 

chemical equilibria in natural waters . 

583 p. New York : John Wiley & Sons, Inc . 

Ward, F . J ., B . Wong, and G . G . C. Robinson . 

1970 . A liquid scintillation procedure for 

determining the effect of size on self - 

adsorption of carbon-14 in Daphnia pulex . 

Limnol. Oceanogr. 15(4) : 648-652 . 

Wood, E. J. F . 1965 . Marine microbial 

ecology . 243 p . New York : Reinholt . 

Wright, R . T., and J. E. Hobbie . 1966 . Use of 

glucose and acetate by bacteria and algae in 

Shuskina, E. A., and A . V . Monakov . 1969 . aquatic ecosystems . Ecology 47(3) : 

The use of radiocarbon for the separation 

of planktonic animals into trophic levels . 

447-464 . 

300



Dynamics of nutrient supply and 

primary production in 

Lake Sammamish, Washington 

Eugene B . Welc h 

an d 

Demetrios E . Spyridakis 

Department of Civil Engineering 



Abstract 

Lake Sammamish, which lies about 12 miles east of Seattle, Washington, with moderate depth (mean 17 .7 m ) 

and area (19.8 km 2 ), ranks third in productivity of the four Cedar River drainage lakes and can be classified 

mesotrophic. While diversion of over one-half the P from nearby Lake Washington during 1963-67 was followe d 

by reduction in winter mean P content and a rapid shift from eutrophy to mesotrophy (Edmondson 1970) , 

mean winter P content and measured characteristics of plankton response have not changed in Lake Sammamis h 

following a diversion of similar magnitude . Annual nutrient budgets suggest a reduction in sedimented P sinc e 

diversion but little change in the quantity of P released from anaerobic sediment . P availability in the water 

column (winter mean content) appears to be controlled by Fe precipitation to a greater extent than in Lak e 

Washington. Experiments in situ show that N and P are equally limiting to summer phytoplankton productivity , 

but as found in Lake Washington, P may be of more long-term significance. 

Introduction 

The supply of limiting nutrient is probabl y 

the most significant factor that determine s 

the trophic status of a lake . This relationshi p 

can be affected by several factors, particularl y 

those that control nutrient supply such as 

morphometric and hydrologic conditions . 

Vollenweider (1968) has related supply of N 

and P to trophic status of 20 lakes and foun d 

that the relationship was dependent on lake 

morphometry expressed as mean depth . This 

relationship illustrates that reasonably reliable 

predictions about the primary production and 

trophic status of lakes are possible with in - 

formation on the supply of limiting nutrient 

and lake morphometry . Such predictability i s 

valuable in the management of man's encroachment 

on lakes. Of still more value is 

knowledge of the response rate of trophi c 

status indicators to changes in nutrient 

supply . 

This paper presents preliminary findings on 

a nutrient supply budget for Lake Sammamish 

near Seattle, Washington, and ho w 

alteration in that supply by sewage diversio n 

has affected nutrient limitation and primar y 

production . The response in Lake Sammamish 

to nutrient diversion is compared to that in 

Lake Washington and mechanisms are hypothesized 

to explain the delayed recovery in Lak e 

Sammamish . Their trophic status is compare d 

to that of other study lakes in the Cedar River 

drainage, Findley Lake and Chester Mors e 

Lake, with the ultimate intent of refining predictive 

relationships among trophic status , 

nutrient supply and morphometric factors i n 

lakes. Prediction of the rate of response o f 

trophic status indicators to change in nutrien t 

supply should also be enhanced by detaile d 

comparison of manipulated Lakes Sammamish 

and Washington . 

301

Methods 

The water column in Lake Sammamish wa s 

sampled at 2-week intervals in spring, summer, 

and early fall and monthly in winter during 

1970 and 1971 . Phytoplankton biomas s 

and productivity and nutrient content wer e 

determined in samples collected from severa l 

depths. Nutrient supply from surface water s 

was estimated by monthly sample collectio n 

and flow measurements from two major an d 

11 minor streams entering the lake . 

Primary productivity was determined i n 

situ according to procedures described b y 

Goldman (1961) . Water samples inoculate d 

with C 14 were incubated at four depths for 4 

hours and the results reported as integrate d 

productivity in the photic zone extrapolate d 

to daily rates assuming a 1 :1 relationship with 

incident light. Data are reported from on e 

centrally located station in each of the lakes . 

Methods of Strickland and Parsons (1968 ) 

were followed for N, P, and Chlorophyll a 

(Chl a) analyses in water . Total and orthophosphate 

phosphorus were determined spectrophotometrically 

as a phosphomolybdat e 

complex. Reactive silicate was also deter - 

mined from a silico-molybdate complex . 

Nitrate and nitrite were determined spectrophotometrically 

following reduction in a 

cadmium-copper filled column and are reported 

together as NO 3 -N. Chl a was determined 

with a Turner Model 110 fluorometer . 

Cations were determined by atomic absorption 

techniques and anions by routine procedures 

(American Public Health Associatio n 

1971). All analyses except for total P wer e 

performed on filtered (0.451.t poresize) water 

samples . 

Surface sediments (surface to 10-cm depth ) 

were collected with a Peterson dredge in the 

four lakes . Surface sediments in Lake Sammamish 

were sampled more extensively (2 6 

samples from 26 different depths) than those 

of the other three lakes where only four to si x 

samples were collected . Analyses for total C , 

N, and P contents of air-dried sediments wer e 

performed with procedures of Baker (1970 ) 

for C, Bremner (1960, 1965) for N, and 

Delfino et al. (1969) for P. The phosphomolybdate-ascorbic 

acid method of Strickland 

and Parsons (1968) was used for the determination 

of extracted P . Results are expresse d 

on an ovendried (104°C) basis . 

Bioassays to determine the limiting nutrient(s) 

were conducted in large (0 .21 m z X5m) 

plastic cylinders submerged in the lake for 7 

days. Nitrogen, P, C, and Si were added t o 

experimental bags separately and in combination 

. Phytoplankton response was determine d 

by daily measurements of productivity rat e 

and Chl a concentration . Significance of 

response was judged from results of analysi s 

of variance using a factorial design and 

Dunnett 's test at the 95-percent level of confidence 

(Steel and Torrie 1960) . Data are 

graphed as integrated values over time to indicate 

total production . 

Lake Sammamish 

Trophic Status 

Lake Sammamish is considered mesotrophic 

judging from measurements of phytoplankton 

productivity, biomass (Chi a), hypolimnetic 

oxygen deficit, and concentration s 

and loading of N and P . Guidelines for thes e 

characteristics are suggested in table 1 fo r 

judging the trophic status of a lake . With on e 

exception, values for these characteristics i n 

Lake Sammamish fall in between ranges typical 

of oligotrophy and those of eutrophy . Th e 

exception, mean winter ortho PO 4 -P, i s 

greater than the level often considered indicative 

of subsequent summer nuisance alga l 

blooms (Sawyer 1952). Of probably mor e 

significance than winter concentration, how - 

ever, is annual supply of P . In this regard , 

Lake Sammamish lies clearly between safe 

and danger limits of eutrophication, her e 

construed to suggest oligotrophy an d 

eutrophy, respectively (Vollenweider 1968) . 

These ranges of measured characteristic s 

are at best only guidelines for judging the 

trophic status in lakes . Eutrophication o f 

lakes is a complex process and is affected b y 

climate, basin morphology and soil type . Differences 

in these factors could lead to considerable 

inconsistencies in judging trophi c 

status using the limited number of characteristics 

in table 1 . On the other hand, the 

302

Table 1.-Suggested criteria for judging trophic status of temperate 

lakes and respective values for Lake Sammamish 

Item Oligotrophic Eutrophic Lake Sammamish 

Chlorophyll a ug/l ' (growth season mean) 0-4 10-100 7 . 1 

Primary productivity 2 mgC/m 2 • day 

(growth season mean) 30-300 1,000-3,000 77 0 

Hypolimnetic 02 deficit 3 in mg 02 /cm 2 • day 

(mean rate) < .025 >.055 .05 3 

Ortho P04 -P in ug/14 (winter mean) >10 12 (total P, 26 ) 

NO 3 -N in ug/l4 (winter mean) >300 14 2 

Total P annual supply s in g/m 2 • yr for 

mean depth 17 .7 m .26 .2 0 

Total N annual supplys in g/m 2 • yr for 

mean depth 17 .7 m 4 .00 3 .8 7 

' Based on data from several lakes in Canada and Unite d States . 

' Modified from Rhode (1969), including data from Schindler and Nighswander (1970) . 

3 After Mortimer (1941, 1942) . 

°Modified from Sawyer (1952) . 

5 After Vollenweider (1968) . 

Table 2.-Comparison of trophic status indicators (May-August means) in 

Lake Sammamish surface water with other lakes of the Ceda r 

River drainage, Western Coniferous Biome, Washington 

Lake Total P P04 -P N0 3 -N Si Chl a Productivity Secch i 

ug/l mg C/m 2 • day m 

Findley' 4 .9 1 .0 3 .0 76 0.3 370 1 5 

Chester Morse 5 .1 1 .0 16 .3 373 1 .6 520 7 . 3 

Sammamish 48 .0 7 .0 86 .0 1,100 7 .1 770 3 . 5 

Washington 2 18 .7 1 .1 56 .5 9 .5 1,070 2 .3 

July and August only because of earlier ice over . 

2W. T . Edmondson, personal communication . 

30 3

measurements of algal response and the tw o 

most commonly limiting nutrients probably 

represent a minimum list of variables most 

pertinent to the eutrophication process . To 

increase the general value of such guidelines , 

the data base should be more extensive in 

addition to including sediment constituents . 

Cedar River Drainage Lakes 

Comparison of nutrient and plankton characteristics 

in the four lakes of the study are a 

(table 2) shows that Lake Sammamish is inter - 

mediate in trophic status. Data from May t o 

August show Lake Sammamish to be slightl y 

more eutrophic than Lake Washington based 

on mean nutrient concentrations, but less 

eutrophic based on algal density and productivity 

indices. Judging trophic status fro m 

nutrient content during the growing seaso n 

can be misleading . Lake Washington is actually 

more than twice as enriched as Lake Sammamish 

based on annual P supply (0 .48 versu s 

0.20 g/m2 ), which conforms to the differences 

in productivity . 

Of more interest in table 2 than the comparison 

of Lake Sammamish and Lake Washington 

is the striking contrast in the tota l 

series. Findley Lake is clearly oligotrophic , 

while Lake Washington near sea level is apparently 

transitional between mesotrophy an d 

eutrophy . Lake Chester Morse and Lake Sammamish 

are intermediate in elevation an d 

trophic status, but nearly as widely separate d 

as are Findley and Washington (for other 

morphometric data, see Taub et al . 1972) . 

Data from these four lakes alone should pro - 

vide a substantial framework for prediction of 

trophic status from lake nutrient supply an d 

morphometry information, which is important 

to lake management . 

The two oligotrophic lakes are also widely 

different than the mesotrophic lakes in sediment 

characteristics and ionic composition . 

Table 3 summarizes data on major chemical 

ions for the four lakes . A graded sequence in 

water chemical composition from the lakes in 

the upper drainage to Lake Washington is 

readily apparent. A four- to tenfold increase 

in concentration is observed with most 

chemical parameters when comparing Findley 

and Chester Morse Lakes to Lake Sammamish 

and Lake Washington . These radical differences 

in chemical quality of the lake waters i n 

the Cedar River Drainage are primarily due to 

diversified human use and different geologic 

formation of the lake basins . 

Results of analyses for total C, N, and P i n 

samples of surface sediments are presented in 

table 4 . The larger C concentrations and the 

higher C/N ratios in the sediments of Findley 

and Chester Morse Lakes when compared 

with those values in the lakes of lower elevation 

appear to reflect trophic status and are 

most probably due to differences in allochthonous 

and autochthonous inputs . In th e 

two oligotrophic lakes, most of the organic C 

in sediments is derived from allochthonou s 

sources which are relatively resistant to mineralization, 

partially as a result of low N availa - 

Table 3.-Average summer (1971) chemical ion content in surfac e 

waters of Cedar River drainage lakes (Barnes 1972 ) 

Lake HCO 3 S0 4 Cl Ca Mg Na K Specific conductanc e 

mg/1 micromhos/cm at 25° C 

Findley 9 .8 0 .4 0 .6 1 .1 0 .3 0.8

Table 4.-Average C, N, and P contents of surface sediments from 

Cedar River drainage lakes (Bauer 1971, Horton 1972) ' 

Lake C 2 N P 

C/ N 

ratio 

N/P 

ratio 

percent mg/1 

Findley 8 .98 5 .55 1 .09 16 .18 5 .14 

Chester Morse 6 .12 3 .71 1 .56 16.41 2 .3 7 

Sammamish 5 .11 4 .82 1 .32 10.60 3 .6 5 

Washington 4 .22 3 .71 2 .13 11 .37 1 .74 

' Values are in terms of ovendried (104°C) sediment . 

2 The sediments from all four lakes contain less than 0 .1 percent CO3 - C on an ovendried basis . 

Table 5 .-Comparison of nutrient supply in Lakes Sammamish and Washington 

before and after diversion with special reference to Vollenweider' s 

(1968) nutrient supply limitations with respect to mean depth ' 

Item 

Lak e 

Sammamish 

Lak e 

Washingto n 

Area (km2 ) 19 .8 87 .61 5 

Volume (km3 ) .350 2.88 4 

Maximum depth (m) 31 64 

Mean depth (m) 17 .7 32 . 9 

Flushing rate (year) 2 .2 3 . 0 

Prediversion annual total P supply (kg) 11,160 92,600 

Prediversion annual total P supply per surface area (g/m 2 ) .56 1 .0 6 

Percent total P income diverted 65 2 5 5 

Postdiversion annual total P supply (kg) 3,906 41,70 0 

Postdiversion annual total P supply per surface area (g/m 2 ) .20 .4 8 

Vollenweider's danger limit of P supply 

for respective mean depth (g/m 2 ) 

.26 .4 2 

Prediversion annual inorg. N supply (kg) 49,100 246,10 0 

Prediversion annual total N supply per surface area (g/m 2 ) 3 4 .96 5 .6 3 

Percent inorg. N diverted 22 1 2 

Postdiversion annual inorg . N (NO3 -N) supply (kg) 38,298 216,56 8 

Postdiversion annual total N supply per surface area (g/m2 ) 3 3 .87 4 .3 3 

Vollenweider's danger limit of N supply 

for respective mean depth (g/m 2 ) 4 .00 6.0 0 

1 Emery, Moon, and Welch, unpublished data . 

2 Estimated on the basis of population equivalent nutrients diverted and prediversion annual income to the 

lake . 

3 Total N values are estimated by doubling inorg . N values . 

30 5

ility . Autochthonous sources are more significant 

in the mesotrophic lakes . The P conten t 

of all lake sediments was relatively high, with 

the highest occurring in Lake Washington . 

The smaller N/P ratios observed in the sediments 

of Lake Washington are undoubtedly 

due to the higher sediment P concentration . 

Nutrient Supply an d 

Primary Productio n 

The trophic status of a lake is probably 

most closely related to the supply of macro - 

nutrients N, P and C, which may not be easil y 

indicated by concentration . Although growth 

rate of phytoplankton is related to the concentration 

of a limiting nutrient, productio n 

is related to the supply rate of that nutrient 

(Dugdale 1967) . The trophic status in 20 well - 

known lakes in the world has been related to 

annual loading of N and P and mean depth 

(Vollenweider 1968) and with continued refinement 

such a relationship promises to be 

even more useful to management . As indicated 

earlier, Lake Washington is more productive 

than Lake Sammamish, and this tren d 

correlated with P supply or annual loading , 

but not growing-season concentrations . Concentrations 

during complete mixing in th e 

winter more nearly represent available supply 

than do concentrations during the growin g 

season. Alteration in the nutrient supply 

should change productivity, but the rate o f 

that change may depend upon physical characteristics 

of the basin . 

Effect of Nutrient Diversion 

Sewage was diverted from Lake Washington 

r 

16 (1965 8 1970 ) 

101- -0-- -0,s 

p \ 

0 V I I } I I I I I I 

SEPT OCT NOV DEC IAN FEB MAR APL MAY JUN JUL AU G 

Figure 1 . A comparison of surface water total phosphorus for 1964-65 and 1970-71 in Lake Sammamish . 

Shaded regions represent differences between means for the period covered by the horizontal lines . Arrows 

to the right indicate winter-spring averages of surface water total phosphorus for Lake Washington fo r 

prediversion (1963) and postdiversion (1971) conditions (unpublished data-Emery, Moon, and Welch) . 

306

during 1963 to 1967 and from Lake Sammamish 

in 1968, by the Municipality o f 

Metropolitan_ Seattle . The diversion remove d 

about 55 percent of the annual external P 

supply and 12 percent of the inorganic N 

from Lake Washington . Phosphorus and N 

external supply into Lake Sammamish was 

reduced by about 65 and 22 percent, respectively. 

Although the N supply to Lake Washington 

is not greatly different than that t o 

Lake Sammamish, Lake Washington receive d 

about twice the supply of P than Lake Sammamish 

(1 .06 vs. 0.56 g/m 2 ) before diversio n 

and this difference is still maintained afte r 

diversion (0 .48 vs. 0.20 g/m 2 ) (table 5). Th e 

diversion brought the P supply to Lake Washington 

to near Vollenweider's danger limit fo r 

eutrophication and below the danger limit i n 

Lake Sammamish (table 5). The prediversio n 

N supply did not exceed the danger limi t 

nearly as much as did P in either lake so N 

diversion may be considered less significan t 

than P. According to the alteration in P sup - 

ply and if P is most significant in these lake s 

as Vollenweider's relationship shows, the n 

p h y t o p l ankton productivity and biomas s 

should have been reduced in both lakes . 

The mean winter (December to April) total 

P concentration in the surface waters of Lak e 

Washington decreased over 60 percent following 

sewage diversion (Edmondson 1970) . 

Although diversion was not complete unti l 

1967, winter P concentrations began gradually 

decreasing soon after the 4-year diversio n 

process was initiated in 1963 . In contrast, 

little difference can be seen in the 197 1 

winter mean P concentrations in Lake Sammamish 

3 years after diversion in 1968 (fig . 

1) . 

Phytoplankton biomass quickly responde d 

to the reduction in mean winter P0 4 -P content 

in Lake Washington as shown by Edmondson 

(1970) (fig. 2) . Chl a decreased in 

direct proportion to P0 4 -P, while the other 

macronutrients, C and N, varied independent 

of Chl a. A significant change in phytoplankton 

biomass, production or water clarity ha s 

not been observed in Lake Sammamish.' In 

I R. M. Emery, C. E. Moon, and E . B. Welch , 

unpublished data . 

one respect this is gratifying because winter 

mean NO 3 -N and total P concentrations, 

which should indicate available supply, also 

have not changed . In another respect, the delayed 

responses of winter mean P content to 

diversion of over one-half the annual suppl y 

to the lake suggests that factors controllin g 

these winter levels in the two lakes are different 

in either kind or magnitude . 

Factors controlling winter P concentrations 

in Lake Sammamish are not yet understood , 

but comparison of seasonal changes in total P 

and morphological characteristics between the 

two lakes offers a hypothesis . Winter P content 

remained high until the spring diato m 

pulse in Lake Washington following which a 

moderate decrease was observed (see footnot e 

1). In Lake Sammamish, total P content 

normally increased to peaks as high as 70 t o 

100 pg/1 following turnover in November . Instead 

of remaining high until the sprin g 

diatom pulse in April, as it does in Lake Washington, 

total P decreased during the winter in 

Lake Sammamish before the diatom pulse 

(fig. 1). The surface water in Lake Sammamish 

has been observed to become cloudy 

with particulate matter during turnover an d 

remain that way for a month or two before 

clearing. Phosphorus may be sorbed by this 

particulate matter and removed in shallower 

Lake Sammamish (mean depth 17 .7 m), while 

in deeper Lake Washington (mean dept h 

37 m), particulate matter from the bottom is 

not so readily mixed to the surface . In support 

of this, iron content during and followin g 

turnover is higher in Lake Sammamish than i n 

Lake Washington particularly in the hypolimnion 

. The lower residual P content in late 

winter in Lake Sammamish is undoubtedl y 

due to P sedimentation through interactio n 

with relatively greater amounts of iro n 

(Horton 1972, Shapiro et al . 1971) . 

Recovery rate in Lake Sammamish may b e 

slower than in Lake Washington because th e 

former had never attained the enrichment o r 

productivity level of the latter . Rate of recovery 

might have also been rapid in Lak e 

Sammamish if prediversion annual supply had 

been as great as that in Lake Washington . 

There may exist a control threshold level of P 

in the lake, above which alteration by manip - 

307

125 

• 

1933 1963 1964 1965 1966 1967 1968 1969 

Figure 2 . (Printed with permission of W. T. Edmondson and Science magazine-Copyright 1970 by th e 

American Association for the Advancement of Science .) Mean winter (Jan. to Apr.) values in surface water; 

summer (July and Aug .) values of Chl a in surface phytoplankton . The 1963 values, plotted as 100 percent , 

were (in micrograms per liter) : P, 57 ; N, 428 ; and ChI a, 38 . Unconnected points show winter means (Jan . 

and Feb .) of bicarbonate alkalinity and free CO 2 in surface water (25 .3 and 3 .2 mg/liter in 1963) . 

ulation of external supply would be relativel y 

rapid while below that level the respons e 

would be very slow . 

Response to Short-Term Nutrient Change 

To relate external nutrient supply rate to 

production and trophic status in the lake, the 

nutrient of most significance must be known . 

Is phosphorus the most significant nutrient i n 

Lake Sammamish as has been the case i n 

Washington? This problem was approached i n 

two ways: (1) by observing the seasonal 

change in N/P ratio compared with productivity 

rate and biomass and (2) in situ experimental 

addition of macronutrients separately 

and in combination to plastic bags . Generally 

speaking, if the ratio of N :P by weight is 

308

greater than 7 .2:1 (16:1 by atoms) then P 

should be limiting further growth and its addition 

should increase production with th e 

reverse case true for N . If the ratio is aroun d 

7 .2 :1, then both nutrients would be require d 

to increase production. Of course, this can 

only be approximately true because th e 

atomic ratio of 16N :1P in algal cells is an 

average and interaction among nutrients ma y 

occur such that ratios different than thi s 

average may stimulate uptake of one or mor e 

nutrients . 

Lake Sammamish has one major phytoplankton 

outburst which occurs in the sprin g 

and is dominated by diatoms . In 1970, th e 

outburst occurred during May (fig . 3), whic h 

for unknown reasons was about 1 month later 

than normal. Productivity followed a similar 

trend as biomass (Chi a), but carbon assimilation 

(productivity) per unit biomass wa s 

noticeably greatest prior to the biomass peak . 

During this outburst, NO 3 -N decreased by 

more than 500 pg/l while ortho P04 -P decreased 

by 8 µg/l at most, a removal ratio of 

more than 60 :1 (fig. 4) . If this ratio of removal 

represents uptake by phytoplankton , 

which it most likely does, then it is weighted 

heavily in favor of N relative to the expected 

requirements of cells. Inspection of figure 4 

suggests that prior to the outburst, P was i n 

shortest supply relative to needs, but after - 

ward, N actually decreased at the surface to a 

level less than P, and N should then have been 

limiting. The great change in N relative to P, if 

all a result of plankton uptake, suggests tha t 

the supply of P0 4 -P was much greater than 

indicated by the concentration decrease. This 

supply could have come from rapid recyclin g 

from the particulate phase and, thus, the relatively 

constant P0 4 -P content only represented 

the difference between supply and demand 

of P . 

The change in mean ratio of N :P in the 

photic zone is compared to surface productivity 

and biomass in figure 3 . Although it is 

clear from the ratio that P was in shortes t 

supply relative to needs prior to the outburst 

and that the uptake was still apparentl y 

weighted in favor of N, the mean ratio in th e 

photic zone following the outburst remaine d 

near the 7 .2 :1 required value . Thus, if th e 

average values in the photic zone best reflect 

the consequence of nutrient uptake by plank - 

ton, then both N and P would appear almos t 

1600 - 2 8 

I 24 

Q 1200 

- 80 

N E 

U 

rnE 

12 

800 

8 

4 

ti 

1 

- 0 

F 

M 

I I 1 I 

A M J J A 5 0 N D 

1970 

Figure 3 . Primary productivity and chlorophyll a in the surface water and mean N/P ratio in the photic zone o f 

Lake Sammamish during 1970 . 

309

8 0 

600 

500 

cn 400 

z 

z 300 

0 

Z 200 

100 

0 

J I- M A M J J A S 

1970 

Figure 4 . N0 3 -N and P0 4 -P (ortho) in the surface water of Lake Sammamish during 1970. 

0 

equally limiting relative to needs during th e 

summer . 

Nutrient limitation was also determined by 

adding N, P, C, and Si singly and in combination 

to plastic bags suspended in the lake . 

Nitrogen and P additions to bags increased th e 

concentrations from a mean of 3 pg/l N and 8 

µg/l P to 237 pg/l N and 21 pg/l P, a fina l 

ratio of about 11N :1P by weight. Silicon was 

added as the third macronutrient in 1970 that 

raised concentrations in the bags from a mean 

of 0.2 mg/l to 1 .3 mg/l, while C was added i n 

1971 that raised the concentration from 7 . 5 

mg/l to 16 .9 . 

Results of the two experiments are show n 

in figure 5 . Response is measured in terms o f 

Chl a averaged over 7 days and C 14 productivity 

averaged over the first 2 days in 197 0 

and the first 4 days in 1971 . In no case did 

the single addition of a nutrient significantly 

stimulate C 14 uptake or biomass accumulation. 

The phytoplankton response was significantly 

greater than the control only when N 

and P were added together. This suggests that 

the ratio of available N :P prior to addition of 

N or P must have been near the required rati o 

of the plankton cells . This is corroborated by 

the mean N:P ratio in the photic zone durin g 

summer being very close to the 7 .2:1 mean 

considered typical for phytoplankton cells . 

Interpretations of limiting nutrients fro m 

seasonal data and results of experiments must 

be used with caution to predict effects of 

nutrient change in a lake . The conclusion that 

N and P are equally limiting to phytoplankton 

in Lake Sammamish as a result of experimental 

additions is based on measure d 

changes in biomass . This is largely because the 

response is determined over several day s 

allowing ample time for biomass accumulation. 

Growth rate may or may not change i n 

response to nutrient addition or change de - 

pending upon where the raised limiting nutrient 

level is on the growth rate-nutrient concentration 

response curve . However, biomass 

will respond usually to an increased suppl y 

rate of limiting nutrient if other factors are 

not limiting growth. The change in the N : P 

ratio is observed after the net result of nutrient 

uptake and loss has occurred . This can b e 

taken to indicate that an increase or decreas e 

in the supply of limiting nutrient will potentially 

alter biomass accordingly if othe r 

environmental factors are optimum . 

310

2 .0 

1 .0 

0 

Si P-N N-Si P-Si P-N-Si Contr . 

(A) 1970 BAG EXPERIMENT S 

r 

9 

6 

3 

0 

P N C P-N N-C P-C P-N-C Contr . 

(B) 1971 

BAG EXPERIMENTS 

Figure 5. Response of phytoplankton to added nutrients in in situ experiments in Lak e 

Sammamish as measured by C 14 assimilation and Chl a content averaged over time . 

Significant differences (95-percent level) from the control are indicated by dotted line s 

(unpublished data-Emery, Moon, and Welch) . 

311

As the results of Edmondson (1970) sho w 

in figure 2, a continual reduction in winte r 

mean P04 -P content lead to a proportional 

long-term decrease in mean summer algal biomass. 

The winter means indicate the supply o f 

nutrient available for subsequent productio n 

in summer . This may not mean that in short - 

term experiments PO 4 -P was always limiting 

biomass production. In fact, Edmondson 

(1970) showed from the N :P ratio in surface 

water that prior to sewage diversion N wa s 

more limiting than P and as P began to de - 

crease more than N, then P became limiting . 

This indicates that relationships betwee n 

annual changes in supply of limiting nutrien t 

and mean biomass are more predictable than 

can be obtained from short-term experiments . 

Seasonal alterations in the N :P ratio, as occur 

in Lake Sammamish, can result in different 

nutrients limiting at different times as a resul t 

of interaction between biological uptake an d 

chemical factors. The nutrient that limits 

most of the time and is, therefore, most influential 

on lake production can be best indicated 

from annual data. Although N and P 

appear equally limiting in Lake Sammamis h 

during August, P supply is considered th e 

most critical factor controlling long-term production 

as has been the case in Lake Washing - 

ton. Although plankton biomass in Lake Sam - 

mamish has not responded to change i n 

external P supply, the springtime concentration, 

which also represents supply for summe r 

growth, has also not changed in contrast t o 

the situation in Lake Washington . 

Internal-External Nutrient Supply 

The internal supply of nutrients, principally 

from sediments, may also partly explai n 

the slow recovery in Lake Sammamish fro m 

diversion of part of the external supply . Th e 

total P budget in a 30 m deep water column 

of Lake Sammamish was calculated for the 

period of November 15, 1970, to Novembe r 

19, 1971, as a function of inflow to and outflow 

from the lake, and of seasonal variatio n 

in P concentration in the lake water colum n 

(table 6). These calculations were made in a 

manner similar to those employed by Vollenweider 

(1968) for data provided by Bachofe n 

(1960) on Lake Baldeggersee . The differences 

in the values of P concentration in the lak e 

water column at the onset and the end of th e 

stratified period, corrected for external P sup - 

ply, represent the relative contribution o f 

sedimented P in the water column . The onse t 

of stratification in Lake Sammamish wa s 

designated to coincide not with 0 2 depletio n 

in the hypolimnion but rather with the firs t 

Table 6.-Exchange of total P between water and deep sediments in Lake Sammamis h 

during one aerobic and one anaerobic period ' 

Relative hypolimnion oxygen 

Aerobic period, 

Nov . 15, 1970- 

Anaerobic period , 

May 21, 1971- 

conditions May 21, 1971 Nov. 19, 197 1 

(188 days) (182 days) 

Change in total P content o f 

representative column of water, g/m 2 -0 .870 +1 .07 0 

Lake retention from surface loading, g/m 2 + .134 + .26 5 

Algebraic difference 

= amount released by sediments, g/m 2 + .80 5 

or 

= amount taken up by sediments, g/m 2 -1 .00 4 

Mean daily rates of release or uptake, g/m 2 day - .006 + .004 

' From Moon (1972) . 

312

detectable increase in concentration of hypolimnetic 

total P (25-30 m from the lake water 

surface) over that observed in late spring. 

Thus, the increase in the hypolimnetic total P 

concentration from about 15 pg/1 in Ma y 

1971 to 25 pg/1 in June 1971 was taken t o 

indicate the onset of lake stratification . 

Similarly, the end of lake stratification coincided 

with the period at which maximu m 

changes in hypolimnetic total P concentration 

were observed : from about 80 pg/1 on November 

19, 1971, to 15 µg/1 on November 26 . 

A net P deposition in the sediments o f 

Lake Sammamish is indicated from the data 

in table 6 . For the period of May 21, 1971, to 

November 19, 1971, P was released by the 

sediments at a rate of 0 .004 g/m 2 -day . This is 

nearly three times greater than the corresponding 

external supply of 0.0015 g/m 2 -day 

for the same period . However, only a portion 

of the P internal supply from the sediment 

will find its way to the trophogenic layer of 

the lake. Normally, most of the sediment - 

released P will remain locked in the hypolimnion, 

and enrichment of the epilimnion by 

sediment phosphorus will depend on edd y 

diffusion and other mixing mechanisms conditioned 

by exposure, morphometric, an d 

hydrologic factors; or by the release of gase s 

from the sediments . 

It is difficult to appreciate the relative importance 

of this P internal supply to the 

trophic dynamics of the water column . Very 

few exact calculations of phosphorus internal 

supply have been made that would permi t 

accurate comparisons with these results . Internal 

P supply from sediments in Lak e 

Baldeggersee, a eutrophic lake in Switzerland , 

of 0.009 to 0.01 g/m 2 -day (Vollenweider 

1968) is nearly double that observed in Lake 

Sammamish . This fact alone, however, canno t 

be employed to explain differences observe d 

in the trophic status of these two lakes . Sediment 

P release rates similar to the ones measured 

in the deep water column of Lake Sammamish 

would probably have a greater influence 

on the trophic status of small lakes or , 

for that matter, the shallow waters of Lak e 

Sammamish . However, preliminary calculations 

of sediment P released in the shallower 

(10 m deep) waters of Lake Sammamish, 

where oxidized conditions predominate, indicate 

a P internal supply at least one-half that 

measured in deep water . 

The P uptake by Lake Sammamish sediments 

during the mixed period (November 

15, 1970, to May 21, 1971) was 0 .00 6 

g/m2 -day. Comparison of Lake Sammamish P 

budgets before and after sewage diversion 

indicates a nearly threefold decrease in P up - 

take by sediments in the period of November 

to May following autumn lake destratification. 

More striking than this observation is the 

fact that the rates of sediment P release fo r 

the period of May to November for the year s 

1964-65 and 1970-71 were identical, 0 .004 

g/m 2 -day. This fact will indicate that th e 

mechanism that controls the sediment- P release, 

primarily the anoxic conditions in the 

hypolimnion, has not changed appreciably 

following sewage diversion. The prevailing 

anoxic conditions in Lake Sammamish durin g 

late summer and fall bring about large in - 

creases in hypolimnetic Fe concentration s 

(Horton 1972), which upon autumn overturn 

cause rapid disappearance of P from the water 

column. Low total Fe concentrations in th e 

hypolimnion of Lake Washington (one-fifth 

to one-tenth of that in Lake Sammamish 2 ) 

likely account for the low removal of P fro m 

Lake Washington following autumn overturn 

and maintenance of high winter P concentrations. 

In the winter months many more than the 

monitored 12 small creeks discharge int o 

Lake Sammamish ; however, their relative contribution 

to the P inflow is small when compared 

to the largest stream, Issaquah Creek , 

which contributes more than 75 percent an d 

90 percent of the measured surface P inflow 

to the lake during the winter and summe r 

months, respectively. Contributions from 

ground and rainwater, or from direct soi l 

surface runoff, are not included in the calculations 

presented in table 6 . The only surfac e 

outflow from Lake Sammamish is through th e 

Sammamish River. Measurements of P out - 

flow from the lake were based on P measurements 

in that river . 

Although the deep water column total P 

2W. T. Edmondson, personal communication . 

313

udget in Lake Sammamish indicates a significant 

release of sediment P to the overlying 

waters, the relative contribution of internal P 

supply to the trophogenic layer, when compared 

to external supply is far from clear . 

Additional annual P budgets in Lake Sammamish 

and in other lakes of varying nutrient 

income, but with similar morphometric an d 

hydrological characteristics, will provide a 

more accurate delineation of the role of sediments 

in supplying P to the water column . In 

particular, measurements of P internal suppl y 

for Lake Washington, which is a lake that i s 

morphometrically a macroscale of Lake Sammamish, 

contains relatively smaller amount s 

of Fe and P in hypolimnetic waters during th e 

summer stagnation periods, and is mor e 

eutrophic than Lake Sammamish, will provid e 

a unique test case . External P supply that re - 

mains in the epilimnic and metalimnic layer s 

of Lake Sammamish, at a rate of 0 .001 5 

g/m 2 •day during the summer stagnatio n 

period, apparently provides a more readil y 

available source of P to the trophogenic laye r 

of the lake than does hypolimnetic P . A more 

accurate evaluation of the relative importance 

of internal P supply to Lake Sammamish productivity 

must await a better understanding 

of the mixing mechanisms . 


Assistance was provided by graduate students 

in Civil Engineering working on individual 

research projects as part of the IBP program. 

In particular, these are R . S. Barnes, D . 

H. Bauer, R . M. Emery, G . R. Hendrey, M. A . 

Horton, and C . E. Moon . 





International Biological Program. Support was 

also furnished by an Environmental Protection 

Agency Research Fellowship, No . 

5-Fl-WP-26, a Public Health Service Training 

Grant and an Office of Water Resources Re - 

search Allotment Grant, No . A-045 Wash . 

This is Contribution No . 47 to the Coniferou s 



American Public Health Association. 1971 . 

Standard methods for the examination o f 

water and wastewater . Ed . 13. New York . 

Bachofen, R. 1960. Stoffhaushalt and Sedimentation 

in Baldogger-und Hallwilersee . P . 

1-118 . Zurich : Juris-Vorlay . 

Baker, R. G . 1970. Initial reports of the dee p 

sea drilling project . Vol. IV, p . 746-748 . 

Washington, D.C . : U .S. Govt . Print. Office . 

Barnes, R. S. 1972 . Trace metal survey o f 

Lake Washington drainage from alpine 

regions to lowland lakes . 150 p . M .S. thesis 


Bauer, D. H. 1971 . Carbon and nitrogen in 

the sediments of selected lakes in the Lak e 

Washington drainage . 91 p. M .S. thesis on 

file, Univ. Wash., Seattle . 

Bremner, J. M. 1960. Determination of N in 

soil by the Kjeldahl method . J. Agric . Sci . 

55: 11-33 . 

. 1965. Total N . In C . A. Blac k 

(ed.), Methods of soil analysis, Part 2, p . 

1149-1178 . Madison, Wis . : Am. Soc . 

Agron . 

Delfino, J . J., G. G. Bortleson, and G . F . Lee. 

1969. Distribution of Mn, Fe, P, Mg, K, Na , 

and Ca in the surface sediments of Lake 

Mendota . Environ . Sci. Technol. 3 : 

1189-1192 . 

Edmondson, W . T. 1970. Phosphorus, nitrogen 

and algae in Lake Washington afte r 

sewage diversion . Science 169 : 690-691 . 

Goldman, C . R . 1961 . The measurement o f 

primary productivity and limiting factors in 

freshwater with carbon-14 . In M. S. Doty 

(ed.), Proceedings of the conference on 

primary productivity measurements, marine 

and freshwater, p . 103-112 . U .S. At . 

Energy Comm ., Washington, D .C . 

Horton, M . A. 1972 . The role of the sediment s 

in the phosphorus cycle of Lake Sammamish. 

220 p. M .S. thesis on file, Univ . 


Moon, C. E. 1972. The effect of wastewater 

diversion on the nutrient budget of Lake 

Sammamish . 160 p. M .S. thesis on file , 

Univ . Wash., Seattle . 

Mortimer, C . H. 1941. The exchange of dissolved 

substances between mud and wate r 

314

in lakes . J. Ecol . 29 : 280-329 . 

. 1942 . The exchange of dissolved 

substances between mud and wate r 

in lakes . J. Ecol . 30 : 147-201 . 

Rhode, W. 1969. Crystallization of eutrophication 

concepts in northern Europe . In 

Eutrophication : causes, consequences, and 

correctives, p. 50-64 . Natl. Acad. Sci . , 

Washington, D .C. 

Sawyer, C. N. 1952 . Some new aspects of 

phosphates in relation to lake fertilization . 

Sewage & Ind . Wastes 24 : 768-776 . 

Schindler, D . W. 1971. Carbon, nitrogen and 

phosphorus and the eutrophication of 

freshwater lakes . J. Phycol . 7 : 321-329 . 

and J. E . Nighswander. 1970 . 

Nutrient supply and primary production in 

Clear Lake, eastern Ontario . J. Fish . Res . 

Board Can . 27 : 2009-2036. 

Shapiro, J ., W . T. Edmondson, and D . E . 

Allison. 1971. Changes in the chemical 

composition of sediments of Lake Washing - 

ton, 1958-1970 . Limnol. & Oceanogr. 16 : 

437-452 . 

Steel, Robert G., and J. H. Torrie . 1960 . 

Principles and procedures of statistics. 48 1 

p. New York : McGraw-Hill . 

Strickland, J . D. H., and T. R. Parsons . 1965 . 

A manual of seawater analysis . Ed. 2 . Bull . 

Fish. Res . Board Can . No. 125, 203 p . 

Taub, F. B., R. L. Burgner, E . B. Welch, an d 

D. E. Spyridakis. 1972. A comparative 

study of four lakes . In Jerry F. Franklin, L . 

J. Dempster, and Richard H . Waring (eds .) , 


ecosystems-a symposium, p. 21-32. Pac . 

Northwest Forest & Range Exp. Stn ., Portland, 

Oreg . 

Vollenweider, R . A. 1968. Scientific fundamentals 

of the eutrophication of lakes and 

flowing waters, with particular reference to 

N and P as factors in eutrophication . 182 p . 

Organ . Econ . Co-Op . & Dev . Dir. Sci. Aff . , 

Paris . 

315



Hydroacoustic assessment of 

limnetic-feeding fishes 

Richard E . Thorn e 

Fisheries Research Institut e 


Seattle, Washington 9819 5 

Abstract 

Hydroacoustic techniques have been applied at the University of Washington to determine the number and 

biomass of limnetic fishes in order to evaluate their role in the productivity of lake systems. The lakes are 

surveyed with high frequency, high resolution portable echo sounders . The echo signals are recorded o n 

magnetic tape and analyzed by a special computer program. Information on size and species composition is 

obtained primarily by net sampling, but acoustic determination of size appears feasible in some cases . 

Introduction 

Determination of the numbers or biomas s 

of fishes has been a continual problem in 

fishery management . Traditional techniques 

based on catch-per-unit-effort and tagging 

experiments have many inadequacies. Th e 

problem of assessing fish populations in lake 

systems is further compounded by the fac t 

that fisheries in lake systems are generally 

limited, highly selective, or nonexistent, s o 

that catch statistics are of little value . 

Determination of the numbers and biomas s 

of limnetic-feeding fishes in lake systems i s 

essential to assess recruitment, growth, mortality 

rates, distribution patterns and inter - 

actions with other trophic levels . As part of 

International Biological Program studies o f 

the Coniferous Forest Biome at the Universit y 

of Washington, this problem has been 

attacked through the application of hydroacoustic 

assessment techniques . 

History of 

Acoustic Assessment 

of Fish Populations 

Echo sounders have been used since th e 

mid-1930's to study the distribution and relative 

abundance of fish populations, especiall y 

in marine environments . However, prior t o 

about 1960, hydroacoustic studies of fis h 

populations were essentially dependent on 

subjective interpretation of echogram records . 

During the last decade, a variety of electroni c 

devices for automated signal processing has 

been developed. These advances, combined 

with improved data acquisition systems an d 

increased understanding of acoustic principles, 

have resulted in a number of successfu l 

applications of acoustic techniques to fish 

population estimations both in marine and 

fresh waters (Truskanov and Scherbino 1966 ; 

Cushing 1968; Thorne 1970 ; Thorne and 

317

Woodey 1970 ; Thorne, Reeves, and Millika n 

1971 ; Moose, Thorne, and Nelson 1971) . 

Acoustic Characteristics 

of Fish 

An echo sounder produces a pulse of sound 

from its transducer in the form of a spherically 

spreading cone whose dimensions are 

dependent on the type and size of the transducer. 

The distribution of sound energy transmitted 

in different directions is described b y 

the directivity pattern function and is maxi - 

mum in the direction perpendicular to th e 

transducer surface, termed the acoustic axis . 

where Ie is the intensity of the reflecte d 

sound measured lm from the target , 

and Ii is the intensity incident on th e 

target . 

Investigations into the relationship betwee n 

target strength and fish size indicate that, in 

general, the intensity of the echo from a fis h 

is proportional to its weight (Cushing et al . 

1963 ; Shishkova 1964) . 

Data-Acquisition System 

The equipment used on the lake studies a t 

the University of Washington includes an ech o 

sounder incorporated into a system by whic h 

target data is recorded on magnetic tape . A 

block diagram of the system is shown in 

figure 2 . The receiver-transmitter and the 

chart recorder is a Ross 200A Fineline ech o 

sounder with a frequency of 105 kHz and a 

transmitted pulse power of about 500 w . A 

transmitter pulse duration of 0.6 msec is 

generally used . The receiver amplifier include s 

a time-varied-gain circuit of 20 log R, where R 

represents depth . This circuit corrects for on e 

way spreading loss of signal intensity wit h 

depth . 

Figure 1 . Typical transducer directivity pattern . 

An example of a directivity pattern of a transducer 

is shown in figure 1 . A common way to 

describe the width of a sound beam is to us e 

the half value angle, that is, the angle at whic h 

the sound intensity has dropped to one-hal f 

(-3 in decibels) of the value it has on th e 

acoustic axis. Since the cross-sectional area of 

the cone increases with range or depth, the 

intensity of the sound within the cone correspondingly 

decreases in proportion to th e 

square of the depth . 

When sound is reflected by a fish target , 

the intensity of the reflected sound is proportional 

to the incident sound intensity and is 

dependent on characteristics of the fish target. 

A measure of the magnitude of the echo 

from a fish is the target strength, TS, which i s 

defined as 

TS = 10 log (Ie/Ii) 

Chart 

recorde r 

Powe r 

supply 

interface 

amplifie r 

Receiver 

transmitte r 

Transduce r 

Top e 

recorder 

I 

Oscilloscop 

e 

Figure 2 . Block diagram of data-acquisition system . 

Two modifications of the echo sounder receiver 

were made so that it could be usable in 

the data collection system . A transistorized 

isolation amplifier, which prevents loading of 

the echo sounder receiver circuitry by associated 

equipment, was built and installe d 

318

within the receiver-transmitter unit . The isolation 

amplifier couples echo signal data an d 

synchronization pulses to the other components 

of the system . The second necessary 

modification to the receiver was the installation 

of a vernier type control that allows precise 

adjustment of the receiver gain . 

An interface amplifier is used to connec t 

the signal output of the echo sounder to th e 

input of the tape recorder. A direct connection 

between the two units is not possible be - 

cause of bandpass limitations of the tap e 

recorder. Target data from the output of th e 

echo sounder are of a 105 kHz frequency , 

whereas the maximum frequency response of 

the tape recorder at unity gain is about 8 kH z 

at a tape speed of 3 3%4 IPS. The interface amplifier 

converts the 105 kHz output frequenc y 

of the echo sounder to a frequency of 5 kH z 

by the use of chopper and filter circuits . 

The transducer produces about an 8-degre e 

circular beam to the 3dB points . The transducer 

is generally mounted on a towin g 

vehicle suspended from the side of the boat . 

The towing vehicle can be used from smal l 

outboard vessels as well as large boats, thu s 

allowing complete portability of the system . 

Survey Design 

Surveys are primarily conducted at night , 

since most limnetic-feeding fishes exhibit pronounced 

diel behavior patterns. The fish are 

generally dispersed in midwater at night bu t 

are either on the bottom or schooled i n 

deeper water during the day . The number and 

spacing of transects depends on the variability 

of the population and the degree of accurac y 

required. Studies of optimal survey procedures 

have not yet been carried out in detail . 

Transect design, especially in initial surveys , 

has been governed primarily by the area to b e 

covered and the available time . During surveys 

of Iliamna Lake, a large sockeye-producing 

lake in Alaska, transects were spaced abou t 

six miles apart, while in Quinault Lake, a 

small sockeye-producing lake in Washington , 

transects were spaced about 3/4 mile apart. In 

Lake Washington, where considerable data 

have been collected, survey coverage is about 

one transect per mile . The primary consideration 

in this design was the number of transects 

which could be run in a single night . If 

greater precision is required, the same transects 

are repeated a second night . During 

studies of the hake (Merluccius productus) 

population in Port Susan, Washington, w e 

found that about 40 transects with an average 

length of about 3,500 m over a 45 million m 2 

area were required to produce a precision o f 

about ±15 %, and an additional 40 transects 

reduced it to ±10 % (Thorne et al . 1971) . 

Basic Techniques 

Data Analysis 

Techniques of processing acoustic data fo r 

abundance estimation can be broken dow n 

into two types : echo counting techniques , 

which count the numbers of individual tar - 

gets, and echo integration techniques, whic h 

relate fish density to integrated target voltage 

(Thorne and Lahore 1969) . An examinatio n 

of the theoretical variance associated with th e 

two techniques has been conducted by Moos e 

and Ehrenberg (1971) . Principles of echo integration 

and its application to fish populatio n 

estimates are described by Thorne (1970 , 

1971), Thorne and Woodey (1970), an d 

Thorne, Reeves, and Millikan (1971) . Echo 

counting techniques depend on the ability t o 

resolve almost exclusively individual fish tar - 

gets, thus are limited to relatively low densities. 

However, echo counting techniques hav e 

the advantage that they may also provide dat a 

on size distribution of targets . Applications of 

hydroacoustic techniques to lake systems at 

the University of Washington have generally 

included a combination of the two techniques. 

Basic analysis of data is done by integration, 

but the integrators are calibrated an d 

size information derived by echo counting . 

Echo Countin g 

Echo counts are made by observation of 

specific depth intervals on an oscilloscope 

while the tape recording of a transect i s 

played back. The transecting boat speed and 

sounder pulse rate are such that nearly all fis h 

319

targets are insonified several times . The tru e 

number of fish is determined by countin g 

only the peak echo amplitudes from eac h 

series of returns from a single fish (corresponding 

to the location of the fish nearest 

the acoustic axis) . Counting errors are directl y 

proportional to the concentration of fish . 

When several fish are observed within th e 

counting stratum simultaneously, it is difficult 

to keep track of the various targets . 

Under these conditions an alternative counting 

technique can be used . Instead of counting 

every fish target, the number of targets 

within the stratum can be noted for randomly 

selected pulses, and a mean number of target s 

per pulse determined for the transect . This 

number can be directly compared with th e 

sampling volume of the cone to determine th e 

mean density of fish along the transect . 

Determination of the Sampling Volum e 

The sampling volume of the sounder con e 

can be approximated from the directivity pat - 

tern. However, the volume is also dependen t 

on the size and depth of the fish targets, th e 

transmitter power and receiver gain of th e 

sounder and the minimum threshold fo r 

counting. The sampling volume can be directly 

determined from the number of times a 

target remains within the sounder cone as th e 

boat passes over the fish at a known speed . 

The width of the path of a fish through th e 

sound cone is determined from the formula 

w = 

boat speed (meters/sec) times duration in cone (pulses) 

pulse rate of sounder (pulses/sec ) 

where duration is the average number of time s 

an individual target is sounded upon . It can b e 

shown mathematically that the average length 

of parallel chords through a circle is 7r/4 time s 

the diameter. Thus the diameter of th e 

sounder cone at the depth of the fish target s 

is 4/7r times the average path width . The sampling 

volume of the cone on a single pulse i s 

then the cross-sectional area of the cone (irr 2 ) 

at the mean depth times the depth interval . 

The volume of water surveyed along a transect 

is the diameter of the sampling cone a t 

the mean depth, times the depth interva l 

times the transect length . 

Integration Procedure 

Since echo counting from the oscilloscop e 

is time consuming and its application limited 

to lower density situations, basic data processing 

is generally done primarily by echo integration. 

Determination of integrated voltages 

from the data collected on magnetic tape is 

done by use of a special integration system 

utilizing a small general-purpose computer . 

This system, called the Digital Data Acquisition 

and Processing System (DDAPS), integrates 

voltages from fish targets within several 

depth intervals simultaneously and calculate s 

the fish abundance using input calibration an d 

target strength data (Moose, Green, an d 

Ehrenberg 1971) . A block diagram of the 

system is shown in figure 3 . An example of 

the DDAPS output is shown in figure 4 . 

Tap e 

playe r 

Teletyp e 

printer 

Digita l 

squarin g 

circui t 

Figure 3 . Block diagram of digital data acquisitio n 

Upper limit of 

depth interval 

system (DDAPS) . 

Integrated 

(voltage) 2 

Rectifie r 

and 

filte r 

PDP- 8 

Computer 

Fish density 

(N/m 3 ) 

Numbe r of 

digital sample s 

6 3485 +0 .344950E-04 818 1 

12 6489 +13 .642376E-04 828 2 

18 38175 +0 .377970E-03 818 1 

24 24162 +0 .239227E-03 715 4 

30 3394 +0 .336040E-04 5 0 

36 0 +0 .000000E+00 0 

42 0 +0 .000000E+00 0 

48 0 +0 .000000E+00 

54 0 +0 .000000E+00 

Figure 4 . Example of DDAPS output . 

Analog t o 

digita l 

converte r 

Typically the average target strength is no t 

known precisely before processing . The density 

outputs of DDAPS are thus relative rathe r 

than absolute. Conversion to absolute density 

can be made either from measurement of tar - 

get strengths or by direct comparison of 

320

DDAPS output with determinations of absolute 

density by echo counting . An example 

regression of fish density from echo counts 

and integrated voltage is shown in figure 5 . 

3 6 9 12 1 5 

Fish targets per 1,000 m3 

Figure 5 . Relationship between integrated voltag e 

and fish density determined from ech o 

counts, Lake Washington, August 1971 . 

Sampling for Species and Size Compositio n 

The data processing methods described t o 

this point result in absolute estimates of the 

densities of fish targets within various dept h 

intervals along the various transects . These 

densities can be extrapolated over the respective 

volumes represented by the transects to 

obtain a total estimate of fish within the lake . 

The species and size composition is not derived 

directly from the acoustic data and must 

be obtained from net samples . Several types 

of nets have been applied at the University of 

Washington . In Lake Washington, which is 

accessible by large vessels, a program of sampling 

with a 10-ft Isaacs-Kidd midwater traw l 

has been conducted for several years . Development 

of midwater trawls for use from tw o 

outboard vessels is being conducted for lakes 

not accessible by large vessels . Vertical gil l 

nets of variable mesh size are also bein g 

applied . 

Size Determination from Acoustic Data 

As stated earlier, the target strength of a 

fish is a function of its size. Thus measurement 

of the size of the fish echoes gives an 

indication of the fish size . Unfortunately, the 

size of the fish echo is also dependent on th e 

orientation of the fish relative to the inciden t 

sound (its aspect) and upon its location with - 

in the sounder cone . The average effect of the 

directivity pattern can be sorted out, but , 

even so, the variability associated with the 

relationship between echo amplitude and fish 

size is quite high, and the precision of size 

determination by acoustic methods is uncertain 

at this stage . It does seem feasible at least 

to determine the number of larger fish . 

Dawson 1 was able to estimate the number of 

adult sockeye in Lake Washington by counting 

the large echo targets . The technique 

probably can be used to determine the number 

of large fish such as squawfish and pea - 

mouth chub in lakes . This ability would be 

extremely valuable since the large fish are 

likely to be undersampled by the midwater 

trawls . 

Future Modifications 

The basic hydroacoustic techniques for fis h 

population assessment have been derived an d 

are being applied in several lake systems i n 

Washington, British Columbia, and Alaska, including 

IBP-funded studies of the Coniferou s 

Forest Biome conducted primarily in Lake 

Washington and Lake Sammamish . Future 

modifications of the technique will be directed 

primarily toward refinement of 

methods of size discrimination from acousti c 

data, and further automation of data processing. 

A program for determination of the distribution 

of target strengths is presently unde r 

development with funding from Sea Gran t 

and the National Marine Fisheries Service . 

1 J . J . Dawson . Estimation of the 1971 Lake Washington 

sockeye salmon escapement by means of a n 

echosounder . Informal Progr . Rep . to Wash . Stat e 

Dep . Fish ., Univ . Wash . Fish . Res. Inst ., Seattle, 7 p . , 

1971 . 

321

This program will automatically calibrate th e 

integration program to absolute density, thu s 

eliminating the time-consuming echo countin g 

procedure . The data on target strength distribution, 

compared with the data from ne t 

sampling, will provide information on the precision 

and reliability which can be obtaine d 

from acoustic determination of size . 


Contributions of Dr . R. L. Burgner, Dr . O . 

A. Mathisen, Mr . E. P. Nunnallee, and Mr . J . 

Dawson to the preparation of the manuscript 

are gratefully acknowledged . 

The acoustic assessment techniques de - 

scribed in this paper have been developed primarily 

by an interdisciplinary group under th e 

Sea Grant Marine Acoustics Program and sup - 






Biome . 


Cushing, D . H. 1968 . Direct estimation of a 

fish population acoustically . J . Fish. Res . 

Board Can . 25(11) : 2349-2364 . 

F. R. H . Jones, R . B. Mitson , 

and others. 1963. Measurements of the tar - 

get strength of fish . Radio & Electron. Eng . 

25 : 299-303 . 

Moose, P . H., and J. E . Ehrenberg . 1971 . Variance 

of the abundance estimate obtained 

with 4 fish echo integrator. J. Fish. Res . 

Board Can . 28(9) : 1293-1301 . 

, J. Green, and J . E . Ehrenberg . 

1971 . Electronic system and data processing 

techniques for estimating fish abundance. 

Inst. Elec. & Electron. Eng. Conf. 

Eng. Ocean Environ . Proc ., p. 33-36 . 

, R. E. Thorne, and M . O . 

Nelson . 1971 . Hydroacoustic technique s 

for fishery resource assessment . Mar . Tech . 

Soc . J. 5(6) : 35-37 . 

Shishkova, E . V. 1964. Study of acoustical 

characteristics of fish . In Modern fishin g 

gear of the world, vol . 2, p. 404-409. London: 

Fishing News Books, Ltd . 

Thorne, R . E. 1970. Investigations into th e 

use of an echo integrator for measuring 

pelagic fish abundance. 117 p . Ph .D . thesis 


. 1971 . Investigations into the re - 

lation between integrated echo voltage and 

fish density. J. Fish. Res. Board. Can. 

28(9) : 1269-1273 . 

and H. W. Lahore . 1969 . Acoustic 

techniques of fish population estimatio n 

with special reference to echo integration . 

Univ . Wash. Fish . Res . Inst. Circ. 69-10, 1 2 

p. Seattle . 

, J. E. Reeves, and A . E . Millikan. 

1971. Estimation of the Pacific hak e 

(Merluccius productus) population in Port 

Susan, Washington, using an echo integrator 

. J. Fish . Res . Board Can. 28(9) : 

1275-1284 . 

and J . C. Woodey. 1970. Stoc k 

assessment by echo integration and its 

application to juvenile sockeye salmon i n 

Lake Washington . Univ. Wash. Fish. Res . 

Inst. Circ . 70-2, 31 p . Seattle . 

Truskanov, M . D ., and M . N . Scherbino. 1966 . 

Determination of density of fish concentrations 

by means of hydroacoustics. Ryb . 

Khoz. (Fish . Econ .) 8 : 38-42 . (U .S. Dep . 

C o m m er., Joint Publ. Res. Serv. TT : 

66-34723 .) 

322 * GPO-796-161

------------- ------ -

The FOREST SERVICE et ; U :' S D partment of Agriculture 

is dedicated to the pti lg oaf: mel# ;ease management of the 

Natio'h's forest resoi.Mi-it for srrst it ids of wood, water 

forage, wildlife, a ecretq: :7h? gh forestry research , 

cooperation with St s `and• 'pri v forest owners, an d 

management of the iortiil or sls'an¢! atonal Grasslands, i t 

strives - as directett N provide increasingl y 

greater service to a growtthfl

PE EIE[R-Rg RESEARCH ON - HJ Andrews Experimental Forest

Create successful ePaper yourself

Delete template?

Save as template?