(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Saval Ranch research design, integration and synthesis : modelling workshop report"

BLM LIBRARY 




88008528 




ADAPTIVE ENVIRONMENTAL ASSESSMENTS 



SAVAL RANCH RESEARCH DESIGN, INTEGRATION 

AND SYNTHESIS- 
MODELLING WORKSHOP REPORT 



USDI BUREAU OF LAND MANAGEMENT 
18th and C Streets, NW 
Washington, D.C. 20240 




SEP l7 



. 7 



- 



Memorandum 

To: Members of the Saval Pvanch Steering Committee, Participants in 
the Modelling and Integration VTorkshops, and Other Interested 
Parties 

Prom: Peter C. T^ent, Resource Sciences Staff 

Subject: Completion of Phase One of Saval Ranch Research Design, 
Integration, and Synthesis Contract 

I am pleased to provide for your use a copy of the Phase One Report for 
the Saval Ranch Research Design, Integration, and Synthesis contract, 

prepared by Adaptive Environmental Assessment, Inc. 

» 

This is a progress report and is thus Intended as a tool to help us further 
focus the research design and direction for the project and refine the 
concepts presented to improve the overall model. 

Direction for Phase Two of the contract will be discussed at the November 
meeting of the Steering Committee. One of the modellers will be present 
to interact with the committee. I hope all committee members will take 
the time to review this progress report prioi co the i^ctitius- 

Tentative plans have been made for a third technical meeting /workshop early 
in 1983. 




Enclosure: Phase One Report - Saval Ranch Research Design, Integration, 
and Synthesis 

cc: Official 

201 
201:PCLent:rls 9/17/82 653-9200 









SAVAL RANCH RESEARCH DESIGN, INTEGRATION 
AND SYNTHESIS — 
MODELLING WORKSHOP REPORT 



to 



Dr. Peter Lent 
USDI Bureau of Land Management 
18th and C Streets, NW 
Washington, D.C. 20240 



by 



Nicholas C. Sonntag 
David Marmorek 
Peter McNamee 
Timothy Webb 
Joe Truett 

Adaptive Environmental Assessment, Inc. 
P.O. Box 1745 
Grand Junction, Colorado 81502 



September 1982 



* 

A joint venture by LGL Ecological Research Associates, Inc 

and ESSA Environmental and Social Systems Analysts, Ltd. 



Bureau of Land Management 

Library 

Bldg. 50, Denver Federal Center 

Denver, CO 80225 



ACKNOWLEDGEMENTS 

On behalf of LGL Ecological Research Associates, Inc. (LGL) 
and ESSA Environmental and Social Systems Analysts Ltd. (ESSA) we 
express our appreciation for the cooperation and enthusiasm 
demonstrated by all Saval Program personnel involved in this 
project. They have made our participation in the project a 
p lea sure . 

We especially thank Peter Lent and Dick Eckert for their 
continuing assistance in maintaining the course of the project. 
Contracting Officer Jeff Petrino has made it easy for us to deal 
with project contract details. Saval Steering Committee Chairman 
Bob Papworth has been particularly helpful during workshop 
sessions, and J^L^yyi \^ <&{**><■' ?y _ *Oz T*c*jr& t~j Del Vail, was kind 
enough to spend a few days of his time attending the first 
workshop . 

Thanks also go to managers and scientists in the following 
organi zat ions : 

Saval Ranch Owners and Managers 

Without the cooperation and support of the Saval Ranch 
owners and management this project could not have been 
conduc ted . 

Saval Ranch Steering Committee 

The Steering Committee members involved in the project have 
been exceedingly helpful and cooperative. 

Saval Ranch Research Scientists 

Project Scientists from the University of Nevada at Reno, 
the U.S.D.A. Agricultural Research Service, the U.S. Forest 
Service, the U.S. Bureau of Land Management, the U.S. Soil 
Conservation Service and the Nevada Department of Wildlife 
contributed extensively to modelling workshops. They worked 
long hours in many cases to assemble the data to build the 
mode 1 . 

Non-Project Scientists 

Outside scientists contributed important information and 
review comments to portions of the program. Special thanks 
go to Clait Braun of the Colorado Division of Wildlife and 
to Fred Obermiller representing the National Cattleman's 
Association . 

Finally, we thank Jean Bench-Bra hmsteadt of LGL for the 
drafting, typing and editing efforts crucial to the production of 
the reports. 



TABLE OF CONTESTS 

Pag e 

STATEMENT OF PROGRESS, AUGUST 1982 iv 

EXECUTIVE SUMMARY v 

Introduction v 

Objectives of Work v 

Description of Work vi 

The Saval Ranch Model vii 

Vegetation vii 

Hydrology viii 

Livestock viii 

Wildlife viii 

R e c o mm ended Research Design . i x 

Future Directions x 

1 . INTRODUCTION 1 

1.1 Project Objective and Expected Products 1 

1 .2 Background and Study Area . . 2 

1.3 Project Strategy 5 

1.3.1 First Workshop Activities 6 

1.3.2 Beyond the First Workshop 7 

1.3.3 Research Design 8 

2. BOUNDING THE SAVAL RANCH SIMULATION MODEL 10 

2.1 Actions 10 

2.2 Indicators 12 

2.3 Space 12 

2.4 Time 14 

2.5 Submodel Definition 15 

2.6 Looking Outward 16 

3. VEGETATION SUBMODEL 18 

3.1 Classification of Vegetation 18 

3.2 Plant Biomass 23 

3.2.1 Weekly Growth 25 

3.2.2 Mortality 31 

3.2.3 Translocation of Biomass To and From Roots 31 

3.2.4 Percent Protein 34 

3.3 Range Site Available Water Capacity, Percent 

Cover and Erosion Potential 35 

3.4 Hay Production, Irrigation and Fertilization ... 42 

3.5 Plowing and Seeding, Burning, Spraying and 

Mining 44 

3.6 Review of Submodel Hypotheses 46 

4. HYDROLOGY SUBMODEL 49 

4.1 Soil Moisture 52 

4.1.1 Evapo transp ira t ion 52 

4.1.2 Snow Storage 56 

4.1.3 Runoff and Infiltration 56 

4.2 Erosion 57 

4.3 Freezing of Soil Water 60 



TABLE OF CONTENTS CONT. 

Page 

4.4 Streamflow 60 

4.5 Water Quality ........... 63 

4.6 Bank Stability 63 

4.7 Review of Submodel Structure and Hypotheses ... 65 

4.7.1 Time Resolution 65 

4.7.2 Important Hypotheses ...... 66 

4.7.2.1 Water 66 

4.7.2.2 Soil 68 

5. LIVESTOCK AND ECONOMICS SUBMODEL 7 

5.1 Cattle Movement 70 

5.2 Feeding 74 

5.2.1 Forage Selection 74 

5.2.2 Forage Consumption 79 

5.3 Cattle Growth 82 

5.4 Fall Sales 83 

5.5 Calving ........ 83 

5.6 Overwintering of Cattle 83 

5.7 Economics 86 

5.7.1 Ranch Revenues 86 

5.7.2 Ranch Costs 87 

5.7.3 Economic Benefits to Elko County 88 

5.8 Analysis of Range Utilization Effects on Cattle 
Growth 89 

5.9 Review of Submodel Hypotheses 91 

6. WILDLIFE SUBMODEL ...... 94 

6.1 Mule Deer 94 

6.1 .1 Population Structure 94 

6.1.2 Foraging ................. 96 

6.1.3 Health Index . 98 

6.1.4 Overwinter Survival 99 

6.1.5 Reproduction 99 

6.1.6 Hunter Mortality 100 

6.1.7 Deer Migration 102 

6.2 Sage Grouse 103 

6.2.1 Population Structure 103 

6.2.2 Reproduction 105 

6.2.3 Survival Rates 107 

6.2.3.1 Hens 107 

6.2.3.2 Males 109 

6.2.4 Hunting 110 

6.3 Review of Submodel Hypotheses 110 

6.3.1 Mule Deer 110 

6.3.2 Sage Grouse 112 

7. THE INTEGRATED MODEL 114 

7.1 The Question of Model Validation 114 

7.2 Model Output 116 



li 



TABLE OF CONTENTS CONT. 

Page 

7.2.1 Three Year Scenario 116 

7.2.2 Stocking Scenarios 119 

7.2.3 Feeding Preference Scenarios 126 

8. OUTSIDE THE MODEL 133 

9. FUTURE DIRECTIONS 136 

9.1 Model Improvements 136 

9.2 Data Management 139 

9.3 Research Design 143 

9.3.1 General Approaches 143 

9.3.2 Cross-Disciplinary Communication 145 

9.3.3 Disciplinary Research 147 

10. LITERATURE CITED 152 



in 



STATEMENT OF PROGRESS, AUGUST 1982 

Phase I of the Saval Ranch Research, Design, Integration, 
and Synthesis is now complete. Two workshops were conducted. 
The first workshop, held in November 1981, was the most intensive 
of labor and time. The workshop lasted five days and involved 
participants from most of the agencies working on the Saval 
Project. The major product was an initial Saval Ranch simulation 
model representing the dynamics of the biophysical system 
including livestock, soils, vegetation, hydrology, and wildlife. 
After the workshop a substantial model documentation and 
refinement period occurred based on participant responses at the 
end of the workshop. 

The second workshop was held in January 1982. Over a 3-day 
period the participants evaluated and modified the 1982 research 
plans using the refined model as a focus of discussion. 
Particular consideration was given to the timing, frequency, and 
spatial extent of the data collection. 

Following this meeting a draft report was prepared 
describing the work done, with recommendations. This report was 
circulated for review among Saval Ranch Steering Committee 
members and some other workshop participants. 

In June 1982 a week-long training session was held to better 
acquaint project scientists with the use and applications of the 
computer and the model. 

The original draft report has been revised according to 
reviewers' comments, and the final version follows this Progress 
Statement . 



IV 



EXECUTIVE SUMMARY 



Introduction 



The Saval Ranch Research and Evaluation Project (SRREP) is 
an interagency effort begun in 1978 through a cooperative agree- 
ment among the Bureau of Land Management (BLM), the Agricultural 
Research Service, the Forest Service, the Soil Conservation 
Service, and the owners of the Saval Ranch. The principal objec- 
tive of the project is to evaluate the effects of the Saval Ranch 
Coordinated Management Plan, involving a livestock grazing system 
and a series of rangeland improvement practices, on vegetation 
and livestock production, fish and wildlife habitats and 
resources, and water quality. A secondary objective is to extend 
the lessons learned from the Saval Ranch to other rangeland 
management situations. To meet this objective, research will be 
conducted on the Saval Ranch for approximately 15 years, enough 
to allow for a complete grazing rotation cycle. 

The Saval Ranch project, like other projects of its type, is 
difficult to focus and coordinate because many disciplines, 
organizations, and individuals are involved. Its scope touches 
on hydrology, animal science, rangeland management, economics, 
wildlife biology, and other concerns. Each agency and individual 
involved has unique perspectives, expectations, and biases. 
Although initially all research questions seem relevant and 
important to answer, careful evaluation is necessary to determine 
which questions are highest priority and can be explored under 
constraints of limited money and time. 

To avoid mult id isc iplinary research without synthesis and to 
provide periodic reevaluation of research direction, BLM desired 
an integrated, interdisciplinary research plan to help project 
managers and research scientists associated with SRREP plot the 
course of Saval Ranch research. Adaptive Environmental 
Assessment, Inc. (AEA) was contracted to assist in the develop- 
ment of the research plan. 

Objectives of the Work 

The objectives of the work conducted by AEA are to construct 
a systems model that will aid SRREP to (1) identify the signifi- 
cant hypotheses that should be tested to evaluate the impacts of 
the Saval Ranch Management Plan, and (2) provide an integrated 
research plan for SRREP that can be iteratively modified as 
project results are assimilated. The research plan should maxi- 
mize efficiency of research effort and interdisciplinary 
coor dinat ion . 



Description of Work 

The work is scheduled to span 2 1/2 years, from September 
1981 to February 1984. The contract calls for two modelling and 
integration workshops to be conducted within six months of 
contract award, and a series of smaller meetings between AEA and 
SRREP personnel over the following two years. 

The first two workshops have been conducted. The first 
workshop, held in November 1981, was the most lab* 



^otu\-u iuw\*^j. w a j w w u. & u i »-*. w u ^ v» «. w bu JoO wvs.k.rvoi.iv^* j. i ■ ^ ic ic vaut uo u a 

and understanding of the participants concerning livestock, 
economics, soils, vegetation, hydrology, and wildlife were 
synthesized and then integrated into a model, giving a conceptual 
picture of the biophysical and economic dynamics of the Saval 
Ranch system. Of particular importance at the first workshop was 
the identification of those variables and parameters which link 
the various disciplines outlined above. 

A substantial documentation and model refinement period 
occurred after the first workshop. Modifications to the model 
were made based on the responses of participants to the 
discussions at the first workshop and results of the first model. 
Refinements concentrated on better representing the relationships 
between disciplines, rather than on each particular discipline 
(e.g., the relationship between forage availability and quality 
and cattle growth, the relationship between soil water 
availability and plant growth, etc.). 



The second workshop, in January 1982, lasted three days and 
involved a smaller number of participants and staff. An impor- 
tant objective of the second meeting was to help evaluate and 
modify 1982 research plans using the refined model to help iden- 
tify critical hypotheses about the Saval Ranch system. Particu- 
lar consideration was given to the timing, frequency, and spatial 
extent of the data collections that would be needed to test the 
critical hypotheses. 



Subsequent to the second workshop, a draft report, 
describing the work done to date and giving recommendations for 
the direction of future research and modelling efforts, was 
written. The report was given to participants of both workshops 
for review, and has been revised in response to review comments. 



In addition to the two planned workshops, a session for 
helping to train Saval research scientists in the use and 
application of the computer and model was held in June 1982. 



VI 



The Saval Ranch Model 

The Saval Ranch model developed at the workshops is a dyna- 
mic state-dependent representation of the biological/physical/ 
economic ranch system. The model, structured using the FORTRAN 
language, is currently operational on an AMDAHL 470 computer 
using the MTS operating system. User interaction with the model 
is facilitated using an interactive graphics gaming package known 
as SIMCON. 

The spatial bounds of the model are the pastures making up 
the Saval project area as described in the Coordinated Management 
Plan (Alternative 2 of the Saval Ranch Project Environmental 
Assessment). Within each pasture, the model explicitly 
represents three broadly defined range sites (e.g., clay, loam, 
and riparian) that are an aggregation of those currently used by 
the vegetation research group. 




To improve the model user's ability to evaluate the effects 
of livestock grazing and range improvement practices on 
biophysical resources (i.e., vegetation, livestock, fish and 
wildlife), the range of feasible management alternatives, or 
actions, was defined and implemented within the submodels. Thus 
the user of the model can select and activate any collection of 
actions and run the model to help evaluate the effectiveness of 
those actions. Indicator variables (i.e., system attributes of 
interest to the user) can be selected by the user and then 
displayed graphically on a computer terminal to facilitate rapid, 
interactive gaming with the model. 

The chosen temporal horizon of the model was of the order of 
20 to 30 years with a within-year resolution of one week (i.e., 
the model time-step is one week). The four submodels are briefly 



described below. 



Vegetation 

The biomass of seven plant gr oup s--b it t er brush, big sage- 
brush, other shrubs, forbs, perennial decreasers and increasers, 
and cheatgrass — are predicted from week to week on each of the 
range sites within each pasture. Plant growth is assumed to 
depend on air temperature, soil moisture within three soil 
layers, and the amount of carbohydrate reserves stored in the 
roots. Changes in plant protein and shrub cover are also calcu- 
lated. The management actions included in this submodel are hay 
pasture irrigation and fertilization, and range plowing, seeding, 
burning, and spraying. 



Vll 



Hydrology 

The soil moisture in three soil horizons (0-10, 10-20, 
20-40 in) is predicted daily as a function of rainfall, snow pack, 
and the rates of infiltration and evapotransp irat ion. The stream 
flow in the Mahala and Gance creeks is determined as a function 
of runoff, and a water quality index (an indicator for cutthroat 
trout habitat) is expressed as a function of upland and stream 
bank er os ion . 



Livestock 

The livestock population is allocated to the Saval pastures 
according to the grazing scheme described in the Coordinated 
Management Plan. While in each pasture the cattle selectively 
graze each range site and plant group as a function of water 
availability and forage pa 1 a t ab i 1 i t y . Forage consumed is 

expressed as total digestible nutrients and used to determine the 
rate of cattle growth and reproduction. Overwintering cattle are 
fed exclusively on hay, which, although partly grow 
ranch, may require outside hay purchases. 



on the 



Ranch revenues are calculated as a function of calf and 
cattle sales and the costs of ranch operation, including any 
range management actions enacted by the model user. Indicators 
of the economic benefit of the ranch to Elko County are calcu- 
lated . 



Wildlife 

The numbers of mule deer and sage grouse are predicted 
annually as a function of the weekly changes in vegetation 
biomass and hunting pressure. 

Mule deer survival and reproduction are calculated in the 
model as a function of the total intake per animal over the 
spring, summer and fall periods. Mule deer overwinter off the 
Saval Ranch and their movement onto the ranch is a function of 
the timing of first greenup. Over the spring-summer period, the 
deer migrate towards the uplands, keying in on the timing of 
greenup in each pasture. 



Five 



groups of sage grouse leks are represented. Sage 
grouse nest success and reproduction are represented as a 
function of the average biomass of forbs and grasse 
aroundthelek 



s in the area 



around the lek. Survival from predators and other mortality 
factors are directly related to herbaceous plant biomass and the 
degree of shrub cover available during critical life history 
stages . 



vm 



Recommended Research Design 

Two traditional approaches to impact analysis research that 
could be applied on the Saval are evaluated. The first approach 
is monitoring, or repeatedly measuring, the status of components 
of concern (indicators) over time as the range management plan is 
carried out. The second is hyp o the s i s- t e s t ing , in which the 
response of indicators to specific management actions is measured 
in an effort to reveal the underlying functional relationship 
between components of the system. 

The latter approach is recommended as the basis for Saval 
Ranch research for a number of reasons: 

(1) By examining functional relationships between different 
components (e.g., the growth responses of cattle to 
changes in forage availability and quality), the essen- 
tial interdisciplinary connections between system 
components are clarified. It is usually these inter- 
disciplinary connections which become lost in a 
research project. 

(2) It is much easier to adapt and change the research 
design over time as the understanding about the Saval 
Ranch system changes and improves if functional 
hypotheses are being studied. Options in a monitoring- 
oriented research program invariably become foreclosed 
as a larger and larger data set is constructed through 
time . 

(3) Transferability of the research results to other areas 
will be much easier and more relevant if functional 
relationships are studied. For example, management 
plans for other ranches may involve different cattle 
stocking levels and rotation schemes, but the 
liliiiilikii between cattle growth and forage 
availability and quality (if defined and measured 
properly) will be the same. 




Given that the research should emphasize functional rela- 
tionships between disciplines, cross-disciplinary communication 
becomes vital. There must be a consistent view among SRREP 
managers and research scientists of what research is needed for 
the program to be effective and efficient. In order to insure 
this consistency, the Saval program should place strong emphasis 
on : 



IX 



(1) promoting frequent dialogue among field researchers, 
and between field researchers and project management; 

(2) coordinating data collection activities among the 
different disciplines; 

(3) insuring compatibility among researchers in the way 
that components are measured and in the units of 
measure; and 

(4) making efforts to overcome semantic difficulties in 
interdisciplinary communication. 

Ideally, the nature and design of the research will continue 
to evolve as some questions are answered and new ones become 
obvious. The model itself should help guide this evolution and, 
as new data surface and understanding improves, the model should 
become more realistic and useful as a management tool. 

Future Directions 

The three shorter workshops yet to come will serve to review 
research findings, refine the model, and define new areas of 
needed research. In the short term, the model will be 
transferred to the SRREP. In addition, some obvious model 
improvements will be made, perhaps by transferring portions of 
the large model to microcomputers more easily accessible to 
research scientists and project managers. Also, consideration 
should be given to a data-base management system. Particular 
consideration must be given to insuring that the data-base 
management system be structured to meet the interdisciplinary 
communication needs outlined above. 



1 . INTRODUCTIOH 

In September 1981 Adaptive Environmental Assessment, Inc. 
(AEA) received a contract from the USDI Bureau of Land Management 
(BLM) to assist in the research design, integration, and 
synthesis of a mul t idisciplinary research program conceived to 
evaluate the ecological and economic consequences of a livestock 
grazing and management system on a Nevada ranch. The research 
program — the Saval Ranch Research and Evaluation Project — is an 
ongoing interagency effort that commenced in 1978 through a 
master cooperative agreement among the BLM, the Agricultural 
Research Service (ARS), the Forest Service (FS), and the Soil 
Conservation Service (SCS). As of the contract award date a 
series of field studies, largely baseline in nature, had been 
init iated . 

The contract called for two modeling and integration 
workshops to be conducted within six months of contract award, 
and for a series of smaller-scale meetings between AEA personnel 
and Saval Project managers during the following two years. The 
two workshops have been conducted. This report describes 
contract performance to date, emphasizing workshop methods and 
results and giving recommendations for the direction of future 
research and modeling efforts. 



1.1 Project Objective and Expected Products 

The objective of this project is to facilitate the 
development of an integrated interdisciplinary research plan for 
the project managers and research scientists of the Saval Ranch 
Research and Evaluation Project. Products to be supplied to meet 
this objective are specified to be 

(1) a systems model that will assist Saval Project 
personnel to identify the significant hypotheses that 



should be tested to evaluate the impacts of the Saval 
Ranch livestock management plan, and 

(2) an integrated research plan for the Saval Project that 
can be iteratively adjusted through the life of the 
project (15 years) as the managers assimilate new 
project research results and refine the model. 

Hypotheses to be identified should be testable within the 
constraints of time, funding, physical environment, and ranch 
management. The research plan itself should promote the maximum 
in efficiency of research effort and interdisciplinary 
coordination. Moreover, it is BLM's desire that the planning 
strategy and the research results are applicable, to the extent 
possible, to economic and environmental impact analysis of ranch 
management in other parts of the Great Basin and elsewhere. 



1.2 Background and Study Area 

The principal objective of the Saval Ranch Research and 
Evaluation Project is to evaluate the effects of a livestock 
grazing management system and necessary range improvement 
practices on vegetation, livestock production, fish and wildlife 
resources and their habitats, watershed values, water quality, 
and other resources and values. 

A Steering Committee is responsible for the development of 
overall plans and actions needed to accomplish the project. This 
Committee consists of technical representatives from each Federal 
agency involved, University of Nevada Cooperative Extension 
Service, Nevada Department of Wildlife, Nevada Cattlemen's 
Association, and the Saval Ranch. 

The Committee determines inventory levels, management 
practices, improvements, monitoring systems, research needs, and 
economic evaluations. The Committee has assisted the land 
managers in preparation of a coordinated range management plan 



and the environmental analysis for the study area. All research 
studies, inventories, and management actions are cleared by the 
Committee to insure they do not disrupt or conflict with overall 
project objectives. 

The Saval Ranch (Fig. 1.1) is located about 40 mi north of 
Elko, Nevada. Elevation ranges from about 5800 ft on the eastern 
boundary along Highway 51 to about 8400 ft at the crest of the 
Independence Mountains on the west. The ranch and grazing 
allotment contains 14,000 ac of private land, 28,000 ac managed 
by the BLM and 17,000 ac managed by the Forest Service. 

Climate of the area is semi-arid with cold, moist winters 
and warm, dry summers. Mean daily air temperatures range from 
about 49°F at the low elevations to less than 42°F in the moun- 
tains. Annual precipitation averages about 9 inches in the 
valley and 18 inches in the mountains. 

Flood plain soils are very deep, dark colored, poorly 
drained, and calcareous. Terrace soils are dark colored, very 
slowly permeable to water, and have silica hardpans. Upland 
soils over flint-like bedrock are dark colored, moderately to 
slowly permeable to water, and occur on steeper slopes. 

Natural vegetation is typical of the northern part of the 
In t er m oun t a in region. Common shrubs are: sagebrush, rabbit- 
brush, bitterbrush, snowberry, ser vi c eb er r y , chokecherry, and 
mountain mahogany. Native grasses are represented by species of: 
needlegrass, bluegrass, squirre ltail , wheatgrass, fescue, broom- 
grass, and wildrye. There are also about 4,500 ac of crested 
wheatgrass, an introduced grass species. The most abundant 
broadleaf species are: balsamroot, phlox, aster, milkvetch, 
hawksbeard, groundsel, wyethia, and eriogonum. Stream-side vege- 
tation is characterized by aspen, willow, and stringer meadow 
habitats . 

Seven creeks originate on National Forest land and cross 
some or all of the study area. Three of these creeks have water 
most of the year. Channel cutting is active in spots in lowland 
areas and ranges from 1-4 ft in most areas to 6-12 ft in one 
reach of Mahala Creek. 




to 

> 

CD 

C 

OJ 

4J 

03 
(0 

0) 

u 



c 



C 


•H 
-P 

to 

u 





en 

-u 

•H 

C 
03 

SZ 


c 

03 

a; 



03 
> 

03 

cn 

CD 

-C 
Eh 



-H 
fa 



Mule deep and sage grouse are major game species, although 
16 other game species occur. The "threatened" Lahontan cutthroat 
trout occurs in Gance and California Creeks on National Forest 
land. Other fish species are sculpin, sucker, shiner, and dace. 
Non-game animals are represented by 155 vertebrate species, 33 
mammal, 112 bird, 7 reptile, and 3 amphibian. 



1.3 Project Strategy 

This project is built around a series of meetings between 
AEA and Saval management and research personnel. The first and 
most intensive of the meetings, upon which this report focuses, 
are called Adaptive Environmental Assessment and Management 
(AEAM) workshops. The simulation model and the research design 
recommendations come about as a consequence of these workshops. 

The AEAM workshop procedures have been developed over the 
past ten years by a group of environmental scientists and systems 
analysts at the University of British Columbia and the 
International Institute for Applied Systems Analysis (IIASA) in 
Austria to deal explicitly with interdisciplinary ecological 
problems (Holling 1978). They are intensive five-day workshops 
involving a team of four or five experienced simulation modellers 
and a group of 15 to 20 or more specialists. The focus of the 
workshop is the construction of a quantitative simulation model 
of the system under study. The development of the simulation 
model forces specialists to view their area of interest in the 
context of the whole system. This promotes an interdisciplinary 
understanding of the system, and allows ecological and 
environmental knowledge to be incorporated with economic and 
social concerns at the beginning of a strategic analysis rather 
than at the end of a design process. 

Simulation models require unambiguous information. In the 
workshop setting specialists are forced to be explicit about 



their assumptions. This objectivity exposes critical conceptual 
uncertainties about the behavior of the system, and identifies 
re search needs . 



1.3.1 First Workshop Activities 



The first step in the workshop is to clearly define and 
bound (limit) the problem. Lists are generated of de ve lop m en t 
actions (alternative controls available to management) and 
indicators (those measures [of economic benefit, environmental 
quality, etc.] used by management in evaluating system 
performance), and the conceptual limits of the model are definedo 
The model itself embodies the biophysical 'rules' required to 
transform the actions into indicator responses. 

The next step is to define the spatial extent and resolution 
required to adequately represent the system. Similarly the 
temporal extent or time horizon must be specified, and a usable 
time step or resolution must be agreed upon. This procedure 
makes the modelling problem more explicit, thereby facilitating 
the division of the system into manageable components or 
sub systems . 

The next activity of the workshop (called 'looking outward') 
focuses attention on the subsytems and those variables required 
by each subsystem from the other subsystems. In looking outward 
the standard questions of analysis are recast. Instead of 
asking, "What do you need to know to describe subsystem X?", the 
question is asked, "What do you need to know about all other 
subsystems in order to predict how subsystem X will behave?" 
This question demands a more dynamic view and forces one to "look 
outward" at the inputs into other subsystems. At completion of 
the 1 ooking- ou t w ar d exercise each subsystem has a list of 
"inputs" it needs from the other subsystems and a list of 
"outputs" it must provide to the other subsystems. 

At this point the workshop breaks into subgroups; one 
subgroup for each subsystem. Each subgroup is in charge of 



developing a submodel for the overall simulation model. One 
workshop facilitator wdrks within each subgroup and acts as the 
submodel programmer. Each subgroup has two basic charges: 
generation of output variables required by other submodels and 
generation of the indicator variables identified earlier. 

At the conclusion of the subgroup meetings the facilitators 
are charged with putting each of the submodels on the computer. 
The submodels are then linked together and run under a variety of 
scenarios to explore the consequences of various actions and 
hypotheses about system structure. The principal objective of 
this exercise in an initial workshop is to point out model 
deficiencies and identify areas requiring improved understanding 
and information. 

The workshop is generally concluded with a formal discussion 
of the research priorities identified during the development of 
the model. Ideally, these research priorities will help 
structure the data to be collected in any field seasons 
subsequent to the workshop. 



1.3.2 Beyond the First Workshop 

The first workshop is ideally followed by a period of 
independent work by the collaborating individuals (modelers and 
specialists) which will lead to a second workshop and possibly 
subsequent ones in a phased sequence. Early in the sequence 
workshops concentrate on technical issues, but later they focus 
more and more on communication to policy advisors and the 
constituencies. The emphasis on communication enables an 
effective and logical move to implementation of the research 
suggested by workshop exercises and the model. 

Throughout the workshop sequence, the simulation model 
serves as an expression and synthesis of not only new 
information, but also of the changing mental models of 
scientists, managers and policy makers. The involvement and 
interaction of these groups is essential; each group's learning 



the needs of each other group becomes as much of a product as 
does problem solving. 

Though the simulation model may to some extent replace 
individuals 1 mental models, it does not replace management. 
Management experiments, in which policies are designed both to 
explore opportunities and pitfalls as well as to fulfill 
immediate needs, are particularly valuable. They serve not only 
to test the simulation model's assumptions and predictions, but 
also tend to reveal new management strategies. To the extent 
that the simulation model withstands various tests (management 
experiments, historical data sets from other locations, etc.) its 
credibility as a predictive tool is enhanced. 



1.3*3 Research Design 

Research design recommendations that arise from workshop 
exercises are a consequence of 

(1) important gaps in knowledge that need to be filled to 
complete the simulation of how the system of interest 
functions , 

(2) what kinds and levels of research are feasible within 
existing time, funding, and management constraints, 

(3) what kinds of research approaches are likely to provide 
useful answers, and 

(4) (in the case of this and many other projects) what 
kinds of answers can be validly applied in other places 
and times . 



The simulation model cannot design the research; it can only 
suggest where the important data gaps are. Ideas for research 
design commence from dialogue between the modelers and the 



specialists as they examine the simulation. Experience in the 
kinds of research approaches that work, i.e, that tell w hat 
change s in ind ica t or s w ill occur in response to specified 
actions , is helpful background for this dialogue. We will 
recommend in this report general research approaches that have 
been found to work with other interdisciplinary studies based on 
AEA approaches. 



2. BOUNDING THE SAVAL RANCH SIMULATION MODEL 

The Saval Ranch Project is an interagency effort charged 
with evaluating the effects of a livestock grazing system and 
necessary range improvement practices on vegetation, livestock 
production, fish and wildlife resources and their habitats, 
watershed values, water quality, and other resources and values. 
To this end, description of the feasible management alternatives, 
or actions, and the measures used to evaluate the effects of 
actions on the biological/physical/economic system, or indica- 
tors, begin to describe the realistic limits of the system which 
will be considered during the workshop. The system to be 
simulated is further defined by placing the actions and indica- 
tors in a manageable spatial and temporal framework. 



2.1 Actions 

Actions, in the context of the Saval Ranch program, are the 
feasible human interventions which can alter the characteristics 
of range vegetation and soil, the cattle numbers and movement, 
and to some degree, the wildlife., It is important to define the 
actions as single interventions (e.g., plowing) rather than 
multiple, or a class of, interventions (e.g., vegetation manipu- 
lation). Particular components of the system may respond 
differently to specific actions which are often grouped into one 
generic category. 

At the workshop, the actions fell into five main categories, 
each being representative of a definable subsystem within the 
overall Saval Ranch system (Table 2d). Certainly some of the 
actions specified are more realistic than others and many of them 
are already being executed as part of the coordinated Management 
Plan. Actions listed but not currently being considered are 
included here for completeness. 

It is important to note that even though an action is listed 
under one category it does not preclude it being considered for 



10 



Table 2.1. List of actions developed at workshop. 



1. Vegetation Manipulation 

- plowing 

- seeding 

- burning 

- spraying 

- fertilization 

- chaining 

- biological control 

- moisture concentration (e.g., snow mgt., etc.) 



2. Livestock Manipulation 

- supplemental feeding 

- distribution of salt by riders 

- age ratio of herd (i.e., # of yearlings, 

calves , etc . ) 

- breeding season and methods 

- introduction of different breeds (i.e., sheep 

and cattle) 

- rotation scheme (i.e., deferred, rest, etc.) 



3. Wildlife Manipulation 

- hunting season 

- hunting permits 

- predator control 



4. Water Management 

- sediment ponds 

- dams 

- wells 

- water development (e.g., meadow improvement, 

irrigation) 



5. Economic 

- buying and selling hay 

- livestock movement to external pastures (i.e., 

off Saval) 

- employment and labor supply 

- amount of financing 



11 



another. For example, plowing and seeding, listed as vegetation 
manipulation, are actions likely applied to ultimately affect 
cattle production or wildlife habitat. 



2.2 Indicators 

Indicators are those measurements which individuals use to 
evaluate the state, or health, of a system. They are the links 
between the simulation model and the participants' "mental" model 
of the system. Different people use different measures of system 
performance and it is therefore important to compile a 
comprehensive set of indicators that represent the interests of 
all participating agencies and groups. This ensures that the 
simulation model remains relevant to the concerns of all 
part icipant s . 

The indicator list generated at the workshop reflected the 
major areas of focus (Table 2.2). Note that some of the indica- 
tors were not clearly defined at this stage. This was left as a 
subgroup charge and de f ini t ive definitions of the indicators 
included in the model are given in the submodel descriptions. 



2.3 Space 

The selection of a satisfactory spatial representation is 
invariably a difficult task in any modelling exercise. The scope 
of any model is a compromise between the specific (enabling 
examination of very specific hypotheses) and the general (ena- 
bling examination of the more general classes of hypotheses). 

At the first workshop the participants felt it was necessary 
to explicitly represent each pasture since the cattle graze each 
pasture as a group and on a fixed schedule each year, depending 
on the rotational scheme. The complicating feature however was 
that within each pasture there are very specific range sites 
which correspond to identifiable mixtures of vegetation and 



12 



Table 2.2. List of indicators developed at workshop. 



1 . Vegetation 



- hay production (t) 

- total forage production (Ib/ac/yr) 

- percent composition by weight of key species 

- vegetative cover 



Cattle 



- livestock production (lb/yr) by calves and 

cows 

- cattle weight ( lb/ individual) 



Wildlife 



4. Water 



- number of sage grouse and deer (i.e., females 

of breeding age) 

- hunter days on deer and sage grouse 

- non-consumptive user days 

- biomass of small mammals 

- diversity of non-game birds 



- water quantity (peak, low and total flows) 

- soil moisture (bars tension) 

- stability of stream channel 

- infiltration (cm/hr) 



5. Economics 



- net ranch revenue 

- non-market values 

- direct and indirect market values 

- hay sold 



13 



soils. These range sites are the major categorical framework for 
much of the vegetation work that has been carried out to date* 
The participants felt that the model must maintain a separate 
representation of each range site within a pasture (e.g., there 
are a maximum of 10 range types). However, it was agreed that 
the actual location of the range site in the pasture was not 
important* This therefore permitted application of an implicit 
spatial scheme within each pasture. For example, within the 
Lower Sheep Creek Pasture, 6% of the area is Loamy Bottom, 1% is 
Wet Meadow, 15% is Claypan 10-12 in, and 78% is Loamy 8-10 in. 
The fact that there may be two regions of Loamy 8-10 in making up 
the 78% is not considered in this scheme. 

At the end of the first workshop it was agreed to simplify 
the maximum range site specification within each pasture to 
three. This was carried out between the workshops and is the 
structure in the current model (see also Section 3). 



Development of a dynamic simulation model requires 
specification of a time horizon over which model projections are 
of interest, and a time step over which the change in value of 
the state variables will be calculated and displayed. Since the 
coordinated management plan to be applied to the Saval Ranch 
requires 12 years for one complete cycle of the livestock 
rotational scheme, it was agreed that ideally the model time 
horizon should be about two cycles to adequately evaluate the 
effectiveness of the management plan (i.e., 25 years). This does 
not however mean that the model can only simulate 25 year 
projections. The time horizon only acts as a guide for 
determining the temporal scale over which questions are being 
asked of the system. 

On the other hand, specification of a time step does limit 
(i.e., bound) the level of detail in the model. For example, a 
yearly time step would lose the fine scale growth characteristics 



14 



of the vegetation while a one hour time step is too detailed for 
evaluating livestock production. A suitable compromise was 
necessary . 

The participants chose a weekly time step for the period 
between March 15 and November 30 in each simulated year (i.e., 38 
weeks). This corresponds to the period of active vegetative 
activity and cattle on the range. The period from November 30 - 
March 15 was represented as one big time step during which the 
fine scale dynamics were considered at their average level. The 
choice of the weekly time step does not preclude the use of a 
different time step within each submodel if necessary. Only the 
integrated system is really constrained by the weekly 
specification from the point of view of model output and 
evaluation. (The hydrology submodel, for example, chose to use 
an internal daily time step [see Section 4]j the wildlife 
submodel used a yearly time step for sage grouse [see Section 6]. 



2.5 Submodel Definition 

Having defined the spatial and temporal bounds of the model, 
as well as the key inputs and outputs, the system was divided 
into four subsystems. The criteria for proper division are: (1) 
minimization of information transfers between subsystems (each 
submodel simulates a relatively isolated, self-contained part of 
the whole system), (2) efficient division of the expertise of the 
participants such that each subsystem represents the concerns of 
a particular subgroup of specialists, and (3) fairly equal pro- 
gramming workloads for each member of the workshop staff. 

The four subsystems are: 

- vegetation 

- hydrology and soils 

- wildlife 

- livestock and economics 



15 



Each participant joined one of the subsystems and helped 
conceptualize the dynamics describing how the components of the 
subsystem change with time. The linkages between the subsystems 
were defined through the looking outward procedure. 



2.6 Looking Outvard 

The purpose of looking outward is to define the pieces of 
information a particular subsystem requires from all other sub- 
systems in order to predict how that subsystem will behave 
dynamically. This is a qualitatively different question than the 
traditional, which requires lists of "factors which affect" a 
particular component of a system., The product of this exercise 
is an interaction matrix, with the columns specifying what infor- 
mation a subsystem requires from each of the other subsystems 
listed on the rows (Table 2.3). The diagonals are crossed out 
because those represent the internal dynamics of each subsystem, 
a task left to the subgroups to consider. 

In effect, each piece of information listed in the matrix 
represents a specific hypothesis. For example, the water 
subgroup stated it needed % cover by range site on each pasture. 
The hypothesis is that 1 cover has a significant effect on the 
movement and infiltration of water at the range site level. 



16 



a 
o 

Si 
rt 

a 

u 
o 

3 

JJ 
«S 

-a 

<u 

a 
o 

> 

0) 



u 
«3 

3 
jj 
3 


c 

■rt 

o 
o 



en 

cu 

ja 
id 

Eh 



Ed 

fa 
H 
J 

a 

H 

3 



co 

j 

CS M 

Ed O 

Eh CO 
< 

3 <* 



2 

c 

H 

Eh 

[J 
U 
fa 
> 



a 

X CO 

O U 

O H 

E- 2 

to o 
u z 
> o 

H U 




<0 \ 

a> to 

C M >» 

-h «j nj 

•0 

a) c 

H c s 

il-HH 
JJ 1-1 JJ — . 
(0 (Q -U 1) 

O CU 05 r-l 

•rt CJ -rt 

* u — s 
I 



(0 \ 

0) CO 

C H >i 

cu c 

H « J) 

JJ -H (H 

(0 A3 JJ CU 

U Qi«J>H 

=■*> M — S 
I 



rS !h 

JJ 

CD >, 
0>j3. 

Oj <- 
> 

«-< cu 

o e 

3 — . 

en en d) 

c c jj 

O -h 

•U U til 



I 



CJ 

o 

Eh 
CO 
Ed 
> 



c 
o 

■H 

JJ 

(0 

jj CD 

0) jj 

CT> -rt 
CU Ul 

> 

!+h a» cu u 

O 3 -U d) 

IQ O > 

en m h o - 

c cu o a. u 

O >i >J5 

jj ,q <*p <*> jj en 



cu » 
jj en 
•h 0) en 
en en cu 
en cu 

>1 rtJ u 
3 XJ h JJ 
C7< 



ja 

CU 3 



I 



I I 

cu 



CU — »-H 

JJ 1) 01 0! 



In- 
to S <d 

JJ U 

>i en d> 

ja is 

u en 

a> <^ j2 

> 3 

O CU U 

O Cuj3 

> en 

dp JJ — 
I 



CU 

C JJ 
(0 -W 

>-i >rt 
■H 

>»ja 
ja m 

u o 
S u 

< CU 

I I 





3 






j3 




1 


■h a o i 







JJ U. 3 (0 -H 




u 


(C O O CU H 




cu 


4.IW tl -H 
CU & C JJ 




>i 


C7>^ -H Cu >h 




* 


a) eu jj 3 cu 




,3 


> U 3 3 O M-* 




IM 


«-i to h cu a***-! 







O v. CuJj O 


B 


T3 


en o jj 


O 


en cu 


en s j3 U s jj 


■H 


3 


SO O Oi (0 en 


4-' 


3 


O JJ «3 -HO 


03 


JJ T3 


JJ —■ Q) <*> Cu O 


N 



z 
o 

H 

& 

Eh 
Ed 
U 
Ed 

> 





cu 




• 


u 




a, 3 




£ 


JJ 


>1 


0) 


en 


JJ 


jj 


A3 -H 




GUi-l 


t« 




IB 




tS 


3 


• 




cr 


a 


• 




■H J* 


u 


O 


3 


0) 


1) 




jj 


Sh 


>i 


rt3 


cu ja 


3 



>1 

.2 



I = 
O O 

A 

0) 
u - 

3 s 
JJ O 

en cm 
•« I 



X u 

(0 Cu 

S • < 

O 3 

U \ 3 

3 3 

•U OH 





dJ 

M 

CU 

CU OO^ I9HHJ — , 

S^jj£i-H.cnu<J-i in 

iH 3 J3 CU fH i-H 

ji jj ^( •. JJ CU -10 
cuen-Hs CuESJJCn 
11*00 9 HMO 3 
3 Cuen^HTajJ gJJ< 



i 



i 



I I 



OS CO 

Ed J 

£h bS H 

< O 

3 CO 



g 


0) 




c 


3 




+J 


n 




a 


u 




jj 






cu 


>i 




crua 




(D 


--^- 




> 


0) 







s 






3 


>~^ 


en 


en 


0) 


e 


3 


Jj 








■H 


4J 





<Sl 





>H 






0) 




• 


CO 




kl 


3 




>5 






V 


H 




in 


rt3 




> 


c 




(0 







T3 -H 






4J 




u 


to 




<0 


0) 




JJ 


IH 


en 


c 





>, 


3 


0) 


(0 


J3 


U T3 



I I 



Ed 
H 



17 



3. YEGETATIOH SUBMODEL 

The vegetation subgroup's responsibilities are shown in Figo 
3d and the sequence of submodel operations in Fig* 3.2. In 
addition to specifying the rules for change required to implement 
actions and produce required variables s the subgroup expended 
considerable effort synthesizing data to provide initial condi- 
tions and driving variables for the model. As a result of this 
latter responsibility, there was insufficient time during the 
first workshop to accurately conceptualize some of the important 
processes affecting plant growth. Representation of these pro- 
cesses was much improved during the second workshop, aided by a 
collapsed version of the model programmed on an APPLE micro- 
computer o 



3.1 Classification of Vegetation 

The model computes changes in the above ground live biomass 
(in Ibs/ac) of seven plant types (Table 3.1). All plant catego- 
ries except cheatgrass have "storage reservoirs" of potential 
above ground biomass in plant parts not accessible to grazing. 
For forbs and perennial grasses, the storage reservoir implicitly 
represents carbohydrate root storage; in shrubs the reservoir 
refers to storage in both the roots and above-ground woody 
material. This classification was considered at the start of the 
workshop to be the minimum necessary to reflect the effects of 
plant composition and abundance on beef production, wildlife food 
and cover requirements, and hydrological processes such as evapo- 
transpiration. The biomass of each plant type at the time of 
peak perennial grass biomass (late July) was estimated for each 
range site and pasture from data brought to the workshop (Fig. 
3.3). With the exception of wet meadows, the model assumes no 
difference in range site plant composition across most pastures. 
However, it is not known at this time whether this simplifying 
assumption is valid. Prior to the second workshop, the number of 



18 































co 










CJ 


CU 












cu 










co 


00 












cu 










~»^ 


3 












!-i * 










co 


CO 




a 








4-1 4-1 


3 








X 


u 




o 








,3 


O 








r— 1 






X 








00 


H 






CO 


s_^ 


>> 




co 








CO °H 


U 






rJ 




X 




AS 








XI 0) 


o 






fa 


* 


/— \ 




S-i 








3 3 


3 






a 


CU 


>> r-l 













l-i 3 + T3 






o 


a 


4-1 ■■r-l 




3 




-A 1 




X O >^ 4-1 









s 


>> 


•H O 








co| 




W =H XI 3 


u 






CO 


4-1 


cj co 




4-1 




(*\ 




4-» <u 


Cu 






35 




CO 




co 




ol 




o 3 6 








CO 


4J 


a - 




S-4 




HI 
<\ 

o| 




co -X 3 O CU 


B 








3 


CO 3 




•r-l 




T3 


cu -o -a -h > 


o 






Ofl 


CO 


CJ >H 




<4-i 




03 


co 3 4J o 


u 






w 


r— 1 


~*-^ 








CJ 


co 3 M -h g y-i 






X 


a 


S-i U 




4-1 


■ 


M| 


3 


co o a. co 








H 


r— \ 


0) cu 




CO 




Ql 


-o 


h h r-i -O 






O 


>> 3 


-J 4-1 








Z\ 


O 


oo oo cu a, -h 


cu 








x — I 


CO eg 




T> 




M| 


S-! 


00 S o 


> 






o 


CU 


3 3 




CU 




a 


>, T3 CO O CO 


o 






H 


3 4J 






a 








X 3 S-i O 


£ 








o o 


cu • 











>, 


CO O *W 


cu 






CO 


•r-l S«l 


1-4 3 




r-l 






CO 


U <4-l 4J O 


u 






w 


jj a. 


X -H 




CU 






X 


CU >•> 3 








J 


CO 


co *~s 




cu > 








> a. r-t co x 


co 






(S 


4J r>9 


r— 1 




4J CU 






en 


O O CO r-l CU 


CU 




_ 


<5 


CU 


•r-l CU 




•iH T3 






3 


O 3 4-i a. -3 


u 






r— 1 


oo -o 


CO 4.) 




co 






O 


CO O 3 


o 






cd 


cu 3 


> 'H 




r—i 






4-1 

1 


^ a 4_i s^s -h 
1 III 


CO 

1 






< 

> 

/ 


> CO 

1 


CO co 
1 




ture and range 
eluded in mode 


d 



S-4 

=5 

cn 










r-J 




















CjJ 


fa 

2 
1 










co 3 












Q 










CO -H 


c 










O 










Pu 













s 










4-1 


■H 










cq 










>. 


JJ 










3 


O 










X 3 


ItJ 










CO 












4J 












Cd 

O 

fa 










-X + 


OJ 










2 












cn 










O 
i— i 












0) 

> 










H 
< 


co 

w 

33 












<D 










H 












£ 










fa 

o 

fa 












4J 








*am 




> 




to 




r— 1 






4-1 

T3 












w 








^ 


w V 


•H 




co 


<D 






/* 










\ a. cu 


o ^ 




1—1 /— v 


C 






/ 










> >,H-I 


co r 




CO 4-1 


cn 






/ 










U -r-l 


O 




X M 


•H 






/ 








CO 


r— 1 


>» CN 




CO 1 


CO 






/ 








r-3 


4-J tJ 


X A 


e 


S cu 


cn 






/ 








w 


3 >-i 




3 


S-. 


m 






/ 








Q 


CO -H 


*— s •% 


= 


♦> CJ 








/ 








o 


-i 3 


r>9 : 


•H 


CU CO 


cn 






/ 








S 


a 


•^ o 


X 


CJ w 


(D 






/ 








CO 


-d 


CN 


CO 


3 


•H 






/ 








33 


>% 3 


4J 1 


E 


CO co 


4J 














CO 


X co 


3 O 

0) r-l 


>, 


O g 
CO 


•H 

r— 1 




00 


+ 








OS 


•3 CU 


4-J 


r— 1 


3 CU 


-r-l 


wj 


c 


r— 1 








fa 


CU r— i 


3 •> 


•H 


•H )-i 


-Q 


21 

01 


•H 


o 








SB 


S w 


O = 


CO 


4-i 


■H 


■X3 


u 








H 


3 4-1 


O O 


-o ^» 


3 co 


cn 


CU 


4J 








O 


co CO 


r— 1 


fa 


O 


c 


M 


CU 


3 










3 CJ 


CU V 


U-io 


r-l M 


a 


HI 


co 


3 O 








s 


O 


u ^ 


^ 


M-l S-I 


o 




o o 








o 


o >> 


3 




o 


cn 


<\ 


T3 


•H 








OS 


X 


4-1 S-I 


3 CU 


-3 fa 


CD 


C 


4-J i—l 








fa 


3 


CO CU 


CO S-i 


CU 


K 




CO 


CO CO 


+ 








o -;< 


•H >, 


CU 3 


4-» X 








00 N CJ 


00 






CO 


•rH /-S 


CO 


S w 


CO 4-1 






(20 


00 3 t-I .H 


3 






fa 


4-1 O 


e --i 


CO 


r-l J-l 


• 




3 


3 -H r-l 00 00 


•r-l 






r-4 


co ca 




>> S-i 


3 O 


rH 




•H 


•H >% -H 3 O 


3 






ea 


4-1 ' 


X 


r— 1 CU 


S 2 


• 




3 


3 CO 4-1 >H r— 4 


•r-l 






<d 


CU CO 


r— 1 4-1 


^ a 


3 


ro 




o 


M !-i W 3 


CO 






H 


oo x 


•r-l a 


cu S 


CJ T3 






r-^ 


3 3- 0) «H -H 


x 






od 


CU r— 1 


O cu 


CU 0) 


CJ 3 


• 




a. 

i 


Xi co u-i g x» 
1 1 1 1 1 


o 

i 






> 


> ^ 

1 


co -a 
i 


3 4J 

l 


CO co 

1 


-H 

fa 



19 




IMPLEMENT ACTIONS 

(PLOWING, SEEDING, ETCo) 
IF NECESSARY 



CALCULATE HAY PRODUCTION 
IN THREE HAY MEADOWS 
ON AUGUST I 



4* 



CALCULATE GROWTH AND NATURAL MORTALITY 
OF PLANTS BASED ON MOISTURE, 
TEMPERATURE AND TIME OF YEAR 



COMPUTE PLANT BIOMASS AT END OF WEEK 
INCLUDING GRAZING 



i 



PERFORM SHOOT— >ROOT OR 
ROOT— >SHOOT TRANSLOCATIONS 



i 



COMPUTE % COMPOSITION BY 
WEIGHT, % COVER, AND % PROTEIN 




Fig. 3.2. Sequence of vegetation submodel operations 



20 



Table 3.1. Plant types represented in vegetation submodel 



Plant type 



Main taxa considered in 
specifying rules for change 



Maximum 
biomass 
(Ib/ac) 



1 . Bitterbrush 

2. Big Sagebrush 

3. Other Shrubs 



4. Forbs 



5 . Perennial 

"decreasers" 



6 . Perennial 

"increaser s 

7 . Cheatgrass 



Pur shia tr identata 129 

Artemisia tr identata 2356 

Serviceberry ( Amelanchier 

alnif o lia ) , Snowberry 

( Symphor icarpos spp . ) , 

Chokecherry ( Prunus virginiana ) 4678 

Er iogonum spp., Cr ep is spp., 

Aster spp. 2148 

Idaho Fescue ( Festuca idahoens is ) , 
Bluebunch Wheatgrass ( Agropyron 
spicatum ) 800 



4474 
Bromus t ectorum 177 



This is the maximum biomass found in late July in all the 
range sites and pastures on the Saval Ranch. 



21 



E 

3 




X 

03 

E 

CD 

s: 

-U 



4J 

CD 
> 



03 

CD 

u 

CU 
H 
03 
U 
W 



CD 
-U 

•H 

en • 
C • 



jQ 
03 



0) 

en 

03 

a - 

0) 

>1-U 

-en 

01 
CD CD 

Cu en 
>i c 
4-1 <T3 
V-i 
4J 

c x: 

03 o 

rH 03 

a a) 

M-J C 

-H 

co en 

en en 

03 03 

E £ 


•H -H 

CQ JQ 






en 

Cm 



22 



range sites was reduced from ten to three based on the soil 
available water capacity. New plant biomasses and available 
water capacities were computed for the three new range sites as 
area weighted averages of the old quantities. This method of 
grouping range sites preserved much of the between pasture varia- 
bility (Tables 3.2 and 3.3). 

Since each simulated year begins March 15, the available 
mid-summer biomass data was reduced by a factor of 10 for 
estimated early spring biomass of each plant type. This factor 
of 10 was only a rough guess, but produced reasonable model 
behavior and is likely the right order of magnitude. 



Table 3.2. Grouping of range sites used for second workshop. 



New range site 
category 



Old range sites 



1 . clay 

2 . loam 

3 . riparian 



claypan 10-12 in, claypan 12-16 in, 
s . s lope 1 2-14 in 

loamy 8-10 in, loamy 10-12 in, 
loamy slope 10-14 in, loamy slope 
14-18 in 

loamy bottom, wet meadow, aspen 
woodland 



The model can easily be changed to group range sites differ- 
ently (e.g., by adding s. slope 12-14 in to the loam category). 



3.2 Plant Biomass 

The subgroup participants agreed that conceptualization of 
the growth dynamics of three plant groups would suffice: one for 
shrubs, one for forbs and perennial grasses, and one for 
cheatgrass . 



23 





o 

0) 
CO 

o 

AJ 

c 



hrj 

o 

Cn 
0) 

4J 

10 

o 

0) 

JJ 

•H 

CO 

0) 

c 

(5 

3 

0) 

c 
•a 



CJ 




M 




3 




4J 




(0 


• 


f0 


A 


a, >i 




•H 


>4J 


A 


g 




Q) 


ta 


1-t 


CO 


M 


rQ 


3 


2 


O 







■rt 


■O 


A 


c 




15 


4J 




c 


"to 


ft 


Or 


rH 


o 


ax: 




ta 


.* 


J! 


io 


S-l 


a> 





Eh 


3 


ci 




cn 




<U 




H 




A 




«j 




h 







. 




"0 jS 




c jj 




M 3 




• O 




Cd 05 




T3 .2 




C 4-) 




H h 




• 




U 2 




u a 




cu cu 




a. cu 




a,.s 




3 w 




u a 




co cy 




2 




JT, 




J w 




f3 




U rH 




cu m 




D.JS 




0) 03 




3 2 


O) 




S-l 




3 


0) (B 


4J 


H H 


CO 


"O 03 


(B 


■O JS 


a 


•H «J 




2 2 


>i 




,0. 




_^ 




OS 


ra 


x^ 


Sri iH 


0) 


<u 


£1 


3 £ 


rH 


O A3 


*"• 


i-3 2 


•f. 




ta 




en 




a 


17 


g 


C 


o 


•H 


•h 


• fH 


ca 


M U 




a 


.^ 


Q 


03 




Cj 




Q- 






CP 




c 




•H 




• fH 




2 H 




ia 




a 


m 


u w 


Eh a. 


H & 


C/3 £h 


H E-t 


« 2 


2 < 




« Cm 



00000 o 

in co <yi •*»■ <s\ h 

tN (SHH rH 

IO 



m o o o o 

H <N 00 •«■ 00 

r-» <n f- 1 t— 1 
n 



o o 



o o 



r» © 








00 O O O O 





in CftN H 


CN 


00 


»h nioi/i 


vo 


co h 






m ly H iH 




(M 






co 





CO o o o o 

<Ti flDNH 



.H O O © O 
VO CTi CO CN 



d O O O O 
iH 00 (N H 



OOOOO 

P-» 00 CM rH 

"5= H 



o 

(N 



o o 

CO 



o o 
00 



o o 
00 



o o 

CO 



CN O O O © 





CO <?\ VO CO 


en 


r» rH h 


p» 



o o 



m o o o o o 

ro CTi \o fO cj* 
o\ »h >h r*» 



o o 



— en 

05 3 
Q) M 
M X! 
18 S-l 
CD 

X 



en 
Si 

3 
£ S* 

3 (A 

S-! C 

J2 h 01 C 

IIVD V 

cue s-i m 



re 



en 
u 

CU rH 

CD (8 U) 18 



in 

S-l CO 

CU CO 



(8 

in 



18 

CU 

S-i 

y 

cu 

a) 13 



VM Q4 



18 

c cu 

3 U 

0) U 

S-i C 

CU -H 

a, 



m o o o o 

(JHBi-)OH 
"fT ft f) f\ 

co 



o 



«5» o o o o o 
00 in 00 r» r^ 

0> rH «* in 



o o 

m cn 



o o 
in 



o o 



noooo 

o 

r» 

ro 



m o o o o 

o <n rH r» r» 

<N in rH rH 



00000 
m m rH -"T 

CS m tH fH 



fl o o o o 

in fH f» fH 

VO <N rH 



© 
o 

co 



o 00 

o vo ** 

fH 



o 

VO 



o o 

VO fl 



o o 





V 










<u 


CU 




>i 


cs 


E 


B 


en 


03 


00 


3 


IB 


<8 


tH 


r- 


en 
en 
f8 


en 




O 



T3 
CU 

s 

3 
CO 

en 

03 



03 



03 JS 

— • en 

rB 3 
CU U 
U J2 
03 l-i 
CU 



O A 



en 

3 

en £. 
3 en 

-Q u us 
<u (U J2 
ty«J3 s-i 

03 4J O 



ai o * a 



en en 

s-i m en 

h cu h cu en 

03 to 03 en oj 

H 03 h 03 u 

C 0) c 

C S-l c 

CU <D 

u 
<D 

a 



'cr 



tt) "O 



P» o o o o o 
in <n vo n 00 vo 
rH pi m rH 

rH fH 



VO O 


OOO 


O 


O H 


&\ VO 


m 


rH 


CN "^ 


fH 



OOOO o 
«N V© in <M fH 00 

in *y tn r— 1 



(sj o o o o o 
o vs <y> in cm 00 
rn m cm m rH 



co o o 00 
o ao at vo 

cy\ rH m 



0000 o 

CN VO VO CM rH CPi 

■^ in rH 



vo o o 00 
co vo 
vo co 

<S3= 



o o 
d 1~- 



o o 

rH -rs- 



o o 

in r» 



o o 



o o 
in a\ 



o o 

■«"? CO 



m o o o o 
"0= cm cn c> -»r 
•^» ^> co 



o O 

o in 
m 



rs 

C8 

cu £ 
m en 
03 3 
.. V4 
J-Q 

H 

Q.i 

4-1 



OS 



en 
ia 
3 
JZ u 
en £ 
3 en 
M 



en en 

1-1 M 01 

fH cu rH cu en 

03 en 0; tn 03 
u 

c 
a 



03 
C CU 

Si u in c u . 
tu J3 cu u cu u 
ct£ hhih 

03 4-1 CJ "O 

en o *h a. 



cn 



18 

cu 
u 

03 
CU 

OJ-H £ 

a o 



0) 

4J 



a 



OJ 
I? 
C 
03 

U 

>, 
Xi 



2 


T3 

C 
CU 

o 

3-1 

Si 

JJ 

o 



o 

03 
<8 



B 
«J 

3 



E 
a> 

Ei 



cn 
3 

TJ 
(8 



3 

•H 

3 J2 
O 

c 



(8 

CU 

a. 

VM 

o 



c 
cu 

I 

cu 

c 
o 

cu 

J2 



m 

>N rH 
03 

j3 SZ 

O 

<U Sh 

jS 03 
61 2 



24 



For each plant group the weekly change in edible above- 
ground biomass (by range site and pasture) is calculated using: 

AB ■ G - M -. CE - CW - TR (Equation 3.1) 

where : 

AB ■ change in live above-ground biomass (lbs/ac) 
G ■ growth of live above-ground biomass (lbs/ac) 
M = death of live above-ground biomass (lbs/ac) 
(conversion to litter) 
CE = cattle consumption (lbs/ac) 
CW = wildlife consumption (lbs/ac) 
TR ■ translocation to or from roots (equivalent 
lbs/ac of above-ground live biomass) 

The amount eaten by cattle and wildlife is calculated by the 
Cattle & Economics and Wildlife Submodels, discussed in sections 
5 and 6 respectively. Note that the model does not account for 
cattle and wildlife eating different parts of the plant. Growth, 
mortality and root translocation are discussed in the following 
sect ions . 



3.2.1 Weekly Growth 

Above-ground plant growth (roughly equivalent to net 
photosynthesis) was assumed to depend only on temperature and 
soil moisture; light and nutrients were considered negligible as 
growth-limiting factors in northeastern Nevada. 

The weekly growth (in lbs/ac) of each plant type is computed 
by multiplying the biomass present the previous week by a % 
growth rate, calculated from: 



25 



reduction in reduction in 

PGR = PGRmax * growth rate * growth rate due 

due to temperature to soil moisture 

( Equat ion 3.2) 



where 



PGR = potential growth of above-ground biomass 

( %/week) 
PGRmax = maximum potential growth rate (%/week) 

under ideal temperature and soil moisture 

(e.g., in a greenhouse) 

The maximum potential weekly growth rates were varied to 
determine what values produced reasonable changes in plant 
biomass with and without grazing. At present the model uses the 
rates shown in Table 3.4. Note that: 

(1) temperature, soil moisture and e vapo t r an sp ir a t ion 
reduce the actual weekly growth rate to far below the 
levels in Table 3.4; and 

(2) changes in the representation of translocation (section 
3.2.3) will likely permit lower maximum growth rates to 
be used in the model. 



Table 3.4. Maximum potential plant biomass growth rates 
currently used in the model. 



Plant types 



Maximum potential 
weekly growth rates 



1. Bitterbrush, sagebrush 
and other shrubs 

2. Forbs and perennial 
grasses 

3 . Cheat grass 



150% 



300% 



300% 



26 



The reductions in growth rate due to temperature and soil 
moisture are computed as indices between zero and one (Figs. 3.4 
and 3.5). For this method to be successful, it is critical that 
the bars of tension of soil moisture reflect the variation in 
soil characteristics across range sites. Since soil moisture is 
passed from the hydrology submodel as a percentage of available 
water capacity it is necessary to convert moisture to bars of 
tension of soil moisture (Fig. 3.6). 

The model assumes that shrubs use water down to 40 in of 
soil depth, forbs and perennial grasses only the top 20 in, and 
cheatgrass only the top 10 in. To implement this in the model, 
the soil moisture levels used on the x-axis of Fig. 3.5 were 
selected according to the following rules: 

(1) Shrubs use the maximum soil moisture level within the 
three soil layers (0-10, 10-20 and 20-40 in). Shrubs 
experience mortality if all soil moisture levels are 
greater than 90% (greater than -.01 bars tension). 
Shrub survival through these moisture saturated weeks 
is assumed to equal the growth reduction factor in Fig. 
3.5. 

(2) Forbs and perennial grasses use the maximum soil 
moisture level within the top soil layers (0-10, 10-20 
in) . 

(3) Cheatgrass use only the soil moisture level in the top 
ten inches. 

This representation was used to reflect competition for 
water between plants. It is assumed that a high biomass of 
plants with shallow roots should prevent water from reaching the 
deeper layers, and thus remove the relative advantage held by 
deep-rooting plants (e.g., cheatgrass outcompeting perennials, or 
cheatgrass and perennials outcompeting shrubs). This mechanism 
requires careful linkage of plant growth and hydrological 



27 




Fig. 3.4. Percent of maximum plant growth rate achieved 
at different daily maximum temperatures 
(assuming optimum soil moisture) . 





ioo- 








^0.3 




E 

E 

k 80- 

o 

2 


/ 






© 


/ 






8? 60- 


/ 






UJ 

1- 

K 40- 


/ 






X 

o 20- 


J 






f>- 


/ - 0J ^ 




\J" 


-15 -10 -5 




SOIL MOISTURE (Bars Tension) 


Fig. 3. 


5. Percent of maximum plant growth rate achieved at 




different soil moisture levels (assuming optimum 




temperature) . 



28 











C^— 










• 










•H l—l 










ai h - 










C (D 










CD > 










4-> s- 










3 










^0 





















■m 










CO 










u "? 




< 









XI 5, 














•r™ 














rQ 1/1 














c £ 














03 













ai 


9 


< 


' ° 1-0 




>i-Q 


\ 








s». 


\ 








^ 


\ 


« Voo 


m 
-en 




03 > 

a-r- 


1 


H -o O rf - s 






«■£ 




MGE SI 

bottom 

eodow 
8-IO 
IO-I2" 
woodlan 

slope I 
slope I 

n IO-I2 
n I2-I6 


.0 





ter c 
iltern. 




RAI 

amy 

amy 
amy 

■ pen 

amy 
amy 

aypa 
sypa 






03 TO 




S\ °I°°S oo oo / 














0) 3 




YA — cvi rr <o Oi |C CO to' IT) > 

Va 






12 u 








^3 ^0 




v.* 






r-H 








■h -a 
> * 










«*■ 




V'A 


O 




4-1 I/) 




\ A 


"to 




c OJ 




\ "* 
\ '-^ 






§£ 




\ ■* 






un 




\ ".* 




»-» 


&-i 




\ A 
\ * 


m 


^■5 


a> 




-CM 


a ^ 




\ • \ 






c 




\ ' \ 




UJ 


r- * 




\'.\ 




cr 


£ ^ 




Y>^ 




=} 


CD - — 




O 


l- 


CD 




\ \ — 


-cvj 


C/J 

O 


5 • 

4J CD 




\ ••-.... "--0- 




S 


CD U 






i 
-£ 


O 


4-> 

a, w 








V) 


■H -rH 




^"- ,,, *—*^ B>> ^ ■ 1 


-in 




.6. Relationsh 
of soil mo 


I 


to 10 if) 






6 

i 


O ' T T 
(uoisuai sjog) 3ynisiow 1I0S 






m 

■H 



29 



processes. At the first workshop, the only linkage was by means 
of plant biomass - % cover relationship (discussed in section 
3.3) which caused infiltration to increase as plant biomass 
increased, but did not concurrently increase evapot ransp ir at ion. 

In the refined version of the model developed after the 
first workshop, plant growth rates are subject to an additional 
"water competition constraint". It is assumed that plant growth 
processes require a certain minimum level of evapotransp ir at ion 
(e.g., to keep leaf stomata open). The plant growth rate for 
each plant type decreases if less than the required level of 
evapo t ran sp ir a t ion occurred in the previous week, according to 
the following equation: 



GR = PGR * ET av /ET rq 



(Equat ion 3.3) 



where : 



GR 
PGR 

ET 



ET 



av 



rq 



growth rate of above-ground biomass (%/week) 
potential growth rate, from equation 3.3 
available evapo transp irat ion (in*wk *lb *ac) 
required evapotransp irat ion ( in*wk *lb *ac ) 



Evapo t ran sp irat ion available for plant type i (within a 
iven range site and pasture is given by: 



ET av £ = TET * 



where : 



B. 



EBi 
i-1 



(Equation 3.4) 



ET 



av , i 



TET 



evapotransp irat ion available for plant type i; 
total evapot ransp irat ion (in), computed in the 
hydrology submodel; 



30 



B^ = biomass of plant type i on the particular range 
site and pasture considered (i=l to 7 for seven 
plant types ) . 

Required evapotransp irat ion (ET ) is currently set at 0.003 
in per week per lb/ac of plant biomass, equivalent to 10 lb of 
water (transpired per wk) for each lb of plant biomass (dry 
weight). Model prediction of reasonable peak plant biomass was 
the only criterion used in selecting this value of ET ra ; this 
certainly does not constitute a theoretical justification. An 
improved procedure would be to compute an index of root activity 
at each soil depth layer, based on the percent biomass (and/or 
percent cover) of each plant type. Root activity is now used in 
the hydrology submodel in computing evapotransp irat ion (Equation 
4.2) . 



3.2.2 Mortality 

Mortality of some above-ground live biomass (conversion to 
litter) occurs whenever the weekly growth rate of a plant group 
(GR in Equation 3.3) is very low (Fig. 3.7). Winter survival of 
above-ground biomass is set at 20% for shrubs, forbs and 
perennial grasses, and 5% for cheatgrass. For cheatgrass, this 
5% represents the biomass carried through the winter in the seed 
pool. (Following preparation of the draft report participants 
pointed out that winter survival of above-ground biomass is 
generally much lower: 0% for forbs, 5% for perennial grasses, 
and 0% for cheatgrass unless there is fall germination. The 
model has not yet been run with these assumptions.) 



3.2.3 Translocation of Biomass To and From Roots 

Carbohydrate reserves can be an important buffer against 
grazing. In the first workshop, root storage was only considered 



31 



for forbs and perennial grasses; the model now also includes 
shrub carbohydrate storage in roots and inedible above ground 
woody material. 





10- 






"- 8 - 

us w 

(0 

o 
E 
o 






• 

CC. H 






2 2- 

o 
2 






n— 






u- 1 ii i i i i 

12 3 4 5 6 




Weekly Growth Rate (% Biomass) 


Fig. 3. 


7. Mortality rate of above-ground plant biomass at 




low plant growth rates. 



The model produced at the first workshop contained very 
simple rules to move "potential above ground biomass" to and from 
storage reserves. Seed production represented a critical 
"switching time": prior to seed production, the plant each week 
transferred a fixed proportion of its root stores to above ground 
biomass; after seed production the reverse occurred. The 
following three parameters therefore had a strong influence on 
the biomass of plants seen throughout the season (the values used 
in the workshop model are in brackets): 

IFDATE = time of seed production [the last week of 
July] 

PSTR ■ % of above ground biomass transferred to 

storage each week after seed production [10%] 



32 



PRTS = % of storage transferred above ground each 
week before seed production [10%] 

Root storage of "potential above ground biomass" was set 
initially at ten times the above ground biomass. 

Several refinements were made to the translocation rules 
following the second workshop: 

(1) BEFORE SEED SET: 

If cattle or wildlife are reducing above ground biomass 
through grazing, the amount of translocation from 
storage is computed to be just sufficient to maintain 
constant above ground live biomass (the "cropping 
effect"). If root storage is insufficient to replace 
all the above ground losses that occurred in a given 
week, 50% of the existing storage is shifted above 
ground per week. If above ground biomass has not been 
reduced by grazing, the percent of storage reserves 
shifted above ground (PRTS) is time dependent; between 
March 15 and the time of seed production PRTS decreases 
linearly from 20% to 0%. 

(2) AFTER SEED SET: 

The weekly increment in storage reserves is still 
computed as a fixed percentage of above ground biomass 
(PSTR), but the biomass above ground is not reduced. 
This change reflects the fact that after seed set, 
above ground dry weights remain relatively constant (in 
the absence of grazing). Translocation of storage 
reserves to above ground biomass does not occur after 
seed set, with or without grazing. 

Currently in the model storage reserves of forbs cannot 
exceed 400 lb/ac of potential above ground biomass, and perennial 



33 



grasses and shrubs can store up to 1000 lb/ac. Over the winter, 
respiration demands of the plant use up 70% of the reserves 
remaining on November 30. 



3.2.4 Percent Protein 

The change in percent protein over the growing season is 
structured as a function of time in any given year (Fig. 3.8). 
Future refinements should attempt to establish a representation 
of % protein that reflects more directly the actual plant growth 
taking place. For example % protein could be made a function of 
the weekly increment in live above-ground biomass. 



Percent Protein 
i i i i 


^ 

\ *s» 

\ 

\ 

\ _ _ \ ,. 


\ o o cneoicjiciss «o o /o 
\ ownnMao forbs and perennial grasses 
\ ©••••••••o shrubs 


WEEK 4 8 12 16 20 24 28 32 36 
MONTH MAR. APR , MAY • JUN « JUL . AUG . SEP • OCT * NOV « 


Fig. 3.8. Change in plant protein content over the growing 
season . 



34 



3.3 Range Site Available Water Capacity, 
Percent Cover and Erosion Potential 

In addition to plant biomass composition, each range site is 
characterized by available water capacities for three soil depth 
layers, a distribution of vegetation cover, and a soil erosion 
index (Table 3.5 in the first workshop and 3.6 currently). The 
available water capacities of each range site, which are kept 
constant over time, are important in determining the outcome of 
the competition between plant types for water at different soil 
depths. Range site erosion potential is also held constant 
across pastures and over time. 

It was agreed that the percent of grass, shrub and litter 
cover change over time, although it was not clear how to 
calculate percent cover from percentage weight composition of 
plant types. Note that percent cover here refers to the 
hydro logist 1 s use of the term (i.e., area of bare ground covered) 
where 100% is the maximum total cover. 

In the first workshop, it was assumed that the percentage of 
area under shrub cover was linearly proportional to the percent 
biomass of shrubs, and the areal percentage of grass cover was 
linearly proportional to the percent biomass of forbs, perennial 
increasers, perennial decreasers and cheatgrass. (Percent 
biomass refers to the percent of maximum biomass potentially 
present on the range site [see Table 3.1].) The relationships 
used for these conversions assumed the percentage cover values in 
Table 3.5 corresponded to the maximum percent biomass values for 
each range site across all pastures. These linear relationships 
and the single point used to form them are shown in Figs. 3.9 and 
3.10. 

This method of calculation of percent cover had several 
inadequac ies : 

(1) The range sites with the maximum percent biomass used 
in Figs. 3.9 and 3.10 may have levels of shrub and 



35 



T3 
CD 
co 

3 



05 
•H 

4-> 
G 
0) 

-P 

o 

Oj 

c 
o 

•H 
CO 

o 

5-1 
CD 



•H 
O 
CO 

TJ 
G 

03 

> 

o 
u 

CD 
-U 
•H 

CO 

a) 

tn 
C 
03 

M 



>^ 
4-' 
•H • 

u a. 
(d o 

a,G 

(13 01 

O M 

U 

u o 

CD 5 

-P 

rd 4-> 
5 CO 

d)'H 
HM-I 
rQ 

03 CD 
H£ 
•H -P 

03 

> G 



ft 

n 

CD 

rH 

43 
03 

Eh 



C Oj 
i— o i- 



o 
oo 



&_ 
CU 

> 
o 
u 

<u 

CD 
03 

+J 
c 
cu 
u 

S- 
0) 



O CD -S^ 
£_ +-» ^ 
CD O 

Q. 



'- 
O) 
+-> 
+-> 



u 
o 

OS 



-Q Q. 

S_ O 

<d c 

JC 03 
U 



s_ 

o3 

CQ 



aa .a 


> 


3 


Q. 


cu s- 


O 


0) JZ 


c 


S- CO 


03 


h^ 


(J 



> 

•1— 


•u 


1/1 


03 




(/I 




o 




<_> 


o 


03 


■s 


4-> 


< 




S- 



s- >> c 

03 *-> C 
2 O -r- 



a> < 
— - s_ 
a> 
>> >> 

M 03 



.a 
re 



03 CJ 
> <T3 >— 
< CLt- 
03 O 

u c/i 



O 



<u 



or 

a 
o> 
c 
re 

OS 



r> 


cr> 


p- 


CO 


r» 


r^ 


CM 


CN 


00 


CO 


OO 


"5T 


00 


«* 


ro 


ro 


ro 


ro 


CN 


CN 



m 



o 

(N 



ID 



CO 



O 



*3? 

rH 



ro 



y? 



LD 



CM 



U~> 



CM 

ro 



o 

CN 



ro 



ro 



CN 



O 



ro <jd ro 



^r 



O 



i— I 



CM 
rH 



CO 



O 

rH 



o 
ro 



CN 



LD 

ro 



o 


o 


CO 


ro 


o 


o 


O 


kO 


^ 


-?r 


"3" 


CN 


rH 


CN 


CN 


<* 



o 


rH 


O 


o 


O 


ro 


r-{ 


r» 


o 


KD 


in 




ro 


"^ 


•^ 


ro 


ro 


CN 


^r 


r-\ 


a) 




CD 


CD 


5 










5 


-p 




4-J 


4-1 


O 










O 


03 




03 


03 


rH 










rH 


M 




U 


u 














CD 


4G 


<D 


CD 


>1 


jG 






Xi 


>1 


T3 


tr> 


T5 


T3 


r4 


&i 


5 


5 


CT 


u 


o 


•H 


O 





CD 


•H 








•H 





g 


.G 


g 


g 


> 


£ 


rH 


rH 


X! 


> 


O 


<^> 


CO 


CO 




CN 


m 


<sQ 


•^r 


00 


rH 


rH 


o 


o 




r-\ 


o 


O 


rH 


O 


m 


• 


• 


• 


1 


• 


• 


• 


• 


• 


o 


O 


o 


o 




o 


o 


O 


o 


o 


ro 


KO 


ro 


rH 


CO 


in 


CO 


CO 


r» 


r^ 


rH 


rH 


rH 


r-i 


o 


rH 


o 


o 


rH 


o 



r» o 

r-\ r-\ 















I 


1 




5 






5 




2 




o 


"^ 


-0 


■5* 






CN 




<X> 




r-^ 


r-\ 


G 


r-A 


s 




rH 




rH 


s 






03 


1 







1 


E 


1 


CN 


CD 


CD 


rH 


CN 


4-J 


5 


O 


o 


CN 


H 


a. 


a. 


n 


rH 


4-> 





<-\ 


rH 


r-i 


1 





o 







Q 


T3 




1 




O 


<-{ 


rH 





CD 


Ja 


03 


G 


CO 


G 


rH 


CO 


CO 


3 


a 




CD 


03 




03 













>i 


e 


a 


>1 


04 


>1 


>1 


>1 


c 


r-\ 


S 




>i 


6 


>< 


s 


S 


g 


CD 


CO 


03 


+j 


03 


03 


03 


03 


fO 


03 


04 




O 


CD 


rH 


O 


rH 


O 


O 





0} 


• 


J 


S 


u 


J 


U 


rJ 


I-] 


J 


< 


Cfl 

• 



CN 



ro 



in 



CO 



r^ 



CO 



CTi 



36 



Table 3.6(a). Range site cover and soil erosion potential now 
used for the collapsed range sites (calculated 
as averages from original range sites). 



% COVER BY: 


Tr 


ee and 


shrub 




Herbaceous 






Soil Erosion 


RANGE SITE 




canopy 


Bare 


canopy 


Rock 


Litter 


Potential 


clay 




30 




45 


15 


5 


5 


0.34 


loam 




35 




20 


30 





15 


0.36 


riparian 




25 




5 


60 





10 


0.38 



Table 3.6(b). Available water capacities (in/in) now used for the collapsed 
range sites (by pasture). 









RANGE SITE AND SOIL DEPTH 








PASTURE 




CLAY 






LOAM 




RIPARIAN 






<10" 


10-20" 
0.13 


2 0-4 0" 
0.08 


<10" 


10-20" 
0.11 


20-40" 
0.08 


<10" 


10-20" 
0.15 


20-40" 


W. Darling 


0.13 


0:14 


0.15 


0.14 


E . . Darling 


0.13 


0.13 


0.08 


-0.14 


0.11 


0.08 


0.* 


0. 


0. 


Lower Mahala 


0.13 


0.13 


0.08 


- 0.14 


0.11 


0.08 


0.16 


0.16 


0.16 


Middle Mahala 


0.13 


"0.13 


0.08 


0.14 


0.12 


0.09 


0.13 


0.13 


0.10 


Upper Mahala 


0.13 


0.13 


0.08 


0.14 


0.12 


0.09 


0.13 


0.13 


0.11 


Lower Sheep 


0.13 


0.13 


0.08 


0.14 


0.11 


0.08 


0.13 


0.13 


0.10 


Upper Sheep 


0.13 


0.13 


0.08 


0.14 


0.12 


0.09 


0.13 


0.13 


0.10 


E. Ind. North 


0.08 


0.08 


0.01 


0.15 


0.14 


0.11 


0.15 


0.15 


0.15 


E. Ind. South 


0.08 


0.08 


0.02 


0.14 


0.11 


0.09 


0.15 


0.15 


0.15 



* no riparian in E. Darling. 



37 



<S=CD 



-<f 



C\* 



E 

.00 _ 

E *» 

«•- 

2* « 






•CM' 
o o o o o «== 



— cvi «i ■thci <o Kod*o>0" 



© 
p© 



r8 



© 




&3A00 AdONVO XN30d3d 



en en 

co c e 

03 Oi 



S -H 


u 




en 


<4-l 




•H 






12 U3 


U 




3 


CD 




E H 


> 




3 CU 







E 







•H 






X • 


>1 




03 CD 


a 




E -U 







■H 


c 




m-i en 


03 












CD 






4J CTi-U 




c c 


c 




CD 03 


CD 




u 


O 




u 


u 




CD £ 


CD 




a, u 


a 




03 






C CD 


TJ 




CD 


c 




CD U 


03 




2 






-U U-4 


CD 




CD 


JJ 




12 i-i 


•H 




CD 


en 




^ > 






a o 


CD 







en 




SZ 


c 




CO >i 


03 




^ a 


U 




s-i 






c 


c 




5 03 


•H 











+J 


en 




en jQ 


en 




U 3 


03 




•H S_i 


E 




m-j x; 







— ' en 


■H 


o 




12 


m 


en -u 




■ 


a c 


■U 


m 


•H CD 


c 




x: u 


CD 


» 


en u 


U 


a 


C CD 


i-i 


H 


o a 


CD 


fa 


•H 


a 




-P 13 




T3 


03 C 


E 


c 


rH 03 


3 


(B 


CD 


£ 




s-i en 


• H 


in 


12 


X 


• 


T3 3 


03 


rn 


CD J-i 


E 




E -C 




CD 


3 en 


4J 


t— 


en 





Q 


en m-i 


rH 


ra 


< 


O, Eh 


• 

CTi 






ro 






cn 






•H 






fa 







38 











en 










en 










03 <W 










s 





















■H u 










12 CD 










> 










e 










3 CJ 










S 




- = 






•H T3 




Vco 2 






X c 










03 3 




1 ' -a ' 

e "cm = 21 = ™ 






s 








14-I 01 




7. cos 2 








o-o~ 1 2J ■ 0000 











® E 00 c — <ffl «» * w 






-P c 










C CD 








CD O 




o„oooooo^ 






O u 




© S? — — O O w ° 


O 




U CD 

cd a. 


h. 


—cm* ro'*"iotb|Cco aid 






T3 



> 








c c 


O 








CD tO 


u 




.0 




CD 


c> 




0) 




5 - 

-u) en 


o 








cd en 


o> 






" CO 




.Q 03 
u 


CM —7 


CO 

\ IO «•' 







or kshop) 
d cheatg 




\ \ °* 1 1 




CO 


"2 C 




\. \ Hi ll 


.O 


CO 


03 




\ \\ \ 


(0 


< 

O 
CO 


ir st 

ses 
ite. 




N. \ \ 1 I 


O 




M-) en en 




\ O \\ \ . 


"»o 


H 


— ' 03 




\ ~ \ \ \' ti \ 




Z 


M CD 




\ \ \\ " \tl 




tu 



en CP en 









ae 


"H 1— I 03 






■*■ 


a. 


jG 03 s-i 




\ * \ \V\ 111 




en -h 










C C -C 










c 











■H CD 03 






■ 10 




jJ s-i CD 




h" \\ \ V\ HI 




-CM 
-O 




Assumed rela 
of forbs, pe 
grasses for 


O O 


• 1 ■ 1 1 

O O C 


1 







ad r» 


(O tO <)• (O jg _ 

U3A00 SSVH3 !N30iJ3d 






rH 

m 

• 

Cn 
■H 



39 



canopy cover very different from the "typical" 
percentages in Table 3.5. 

(2) The relationships between percent cover and percent 
biomass, although unknown, are almost certainly non- 
linear; percent cover may in fact be better related to 
the biomass per acre rather than percent biomass. 

(3) Similar range site types show roughly similar slopes of 
% shrub biomass vs. % shrub cover, but very different 
slopes for % grass biomass vs. % grass cover. 

There are good theoretical reasons and some field evidence 
to suggest that the percent cover should be related to above 
ground live plant biomass by a non-linear function. One simple 
method would be to use the equation: 




(Equat ion 3.5) 



where : 



PC ■ Z cover 

BC = base % cover (i.e. the % cover still present when 

there is zero live above ground plant biomass) 
BG = % bare ground 
B - live above ground plant biomass 

K = a constant - the biomass causing 50% cover if BC 
and BG are zero 

An example of this equation (setting BC - 10%, BG - 5% and K 
= 1000) is shown in Fig. 3.11. Data relating pho t o sy n t he t i c and 
woody biomass to crown width and area (Rittenhouse and Sneva 
1977) appears to support the general form of this relationship, 
and use of this equation in the APPLE model in the second 



40 









2* 








Cr- 








0) 3 

















i-i' 1 "". 








»t 








O-ro 








Q. 








S ^ 








55 








U <T3 








CM 








4-3 








u = 


E 

3 






0) <" 


£ 






s_ 






U Q- 


X 






cu 


o 






4J 5- 


E 






C^ 


ii 






CU O 






c 


55 






u 

QJ °° 

-a 


m 
en 


O 
O 




1 1 


rlf> 




en 


| \ 


CM 




ut in 
e and 


1 \ 


O 




0* > 


1 \ 


O 




R 5 





-O 


*"■? 


\ 


CM 








• \ 










! \ 




\ 


tf 


\ 




-Q 


^ E 


j \ 


O 

_o 

l£2 


CO 
CO 


ded 
sche 


i \ 




< 


M rrj 


| \ 




2 


CU 
£ i/J 


\ 







£ ,| ~ 


1 \ 


O 
O 


CO 


8 «? 


| \ 


-O 




ion rei 
ss . (Thi 


1 \ 


O 




4-1 re 


' \ 


-O 




O £ 


\ 


m 




e of fun 
imum bio 


1 • 1 1 1 • 


0000 






r-H X 00 

On nj CU 


O 00 CO <j» CM 






6 £ ,f ~ 

<S CU 


U3A00 !N30d3d 






Fig. 3.11. Ex 
of 
sp 



41 



workshop produced reasonable results. However, further 
discussion is needed in determining the parameter values to 
substitute in Equation 3.5 for each plant type; the large model 
currently still uses a linear function relating % biomass to % 
cover based on Table 3.6 (a). 

Tree cover is assumed constant at 30% in Aspen Woodland 
range sites, and 0% elsewhere. Percent canopy cover is 
calculated from percent shrub cover plus percent tree cover, and 
percent ground cover is equal to 1.4 times grass cover, to 
account for litter. A better method of accounting for litter 
would be to keep track of how much plant biomass dies each year, 
and then derive a relationship to show how percent cover by 
litter changes with the amount of litter present. 



3.4 Hay Production, Irrigation and Fertilization 

Since the hay meadows are relatively intensively managed 
systems, it did not seem reasonable to use the same rules for 
plant growth used in the pastures. Instead, annual hay 
production in the Tremewan, Gance and Haystack meadows was 
assumed to depend on: 

(1) The total flow from March 15 to August 1 in the Mahala, 
Gance and North Fork creeks respectively. (An 
alternative cutoff date is July 1.) 

(2) The degree of water management practiced on the meadows 
(quality of irrigation system). 

(3) The use of fertilizers. 

These factors are drawn together by means of the relation- 
ships shown in Fig. 3.12. Water management can be either "poor" 
(the present level) or "good", and fertilization can either take 
place to the "optimum level" under good water management, or not 



42 



I i 

o c o 

5 ° c 

O — o 

60 E 



c 
« 
E 
o 
o» 
o 
c 
o 
E 






*£ 



■0^-0 «- 
o £ o o 
o ° o o 



-T- 
rO 




><0 



-tO 



-«• C 



-IC 



5 

o 



3: 

Ui 
IT 
O 



-CM 



3 
< 



(00/SU04) Nouonaoad AVH 



c 



c 



•H 

4-> 

n3 

N 



U 
0) 
4-1 

C 
(C 

-P 
C 
CD 

e 

(D 

cn 
(0 

c 

03 

6 

CD 

4-) 

03 
5 



5 



r-l 
14-1 

-^ 

<D 

<D 
U 

CD 
4-> 
fO 

rH 
£ 

o c 
u 

03 -H 
4J 

14-1 O 

O 3 
T3 

cn 
4-) >-, 

U CI, 

CD 

y-i re 



CN 



cn 

■H 



43 



at all. These assumptions reflect the participants' belief that 
nutrient additions could only significantly benefit hay produc- 
tion under conditions of abundant water. The relationships shown 
in Fig. 3.12 are applied separately to each hay meadow and then 
added together to yield total hay production* 



3.5 Plowing and Seeding, Burning 8 Spraying and Mining 

Due to time constraints at the workshop, the simplest 
possible approach to implementing actions is used. The model 
does not disaggregate and quantify the processes determining the 
vegetational changes due to plowing and seeding, burning, 
spraying or mining. Instead, the user merely specifies (as 
driving variables) the number of acres and time that a certain 
location receives one or more actions, and the values of 
parameters that describe the expected outcome of the actions. 

For plowing and seeding, the user specifies at some time 
that a certain number of acres of land within a given pasture and 
range site(s) is "removed" from the available grazing area and 
placed in a "seeding reserve". The model reduces the acreage 
within each range site accordingly. The user must also specify 
when the seeding reserve will have reached a certain vegetation 
composition (typically after two years) and can be added back 
into the available grazing area. The model implements the 
changes in available forage by simply calculating (for each range 
site) the area-weighted average of the vegetation composition in 
the available grazing area and that of the seeding reserve. The 
model assumes that after two years, plowing and seeding yields a 
vegetation composition of: 

bitterbrush - Ib/ac 

sagebrush - 90 Ib/ac 

other brush - 90 Ib/ac 

forbs - 30 Ib/ac 



44 



perennial grasses: 
decreaser s 

mcrea ser s 
cheat gra ss 



- 800 lb/ac (crested wheatgrass) 

- 10 lb/ac 

- 2 lb/ac 



The expected biomass of crested wheatgrass after one year is 
250 lb/ac. 

This simple approach avoids the complexity of specifying the 
rules affecting seedling establishment, growth and competition, 
but, as a result, is insensitive to such factors as: 

(1) annual and seasonal variations in soil moisture; 

(2) the biomass of forbs, perennial grass and cheatgrass 
that might out-compete crested wheatgrass seedlings 
during the seeding period; and 

(3) the percent kill of sagebrush because of plowing. 

Future inclusion in the model of a dynamic representation of 
seeding must recognize that the processes and outcomes of 
competition for moisture between plant types as seedlings are 
quite different from the rules specified in Section 3.2.1 for 
established plants. 

Burning and spraying are represented simply by a transient 
percentage removal of shrubs at the specified time and location. 
The biomass of other plant types are not altered. 

Mining is represented in the same way as plowing and seeding 
(i.e., land is removed from grazing and later returned with 
different vegetation). The vegetation composition of land 
returned to grazing after mining and rehabilitation was not known 
at the time of the workshop. Presumably the period over which 
land is removed from grazing would be much longer than the 2-3 
years typical for plowing and seeding. 



45 



3.6 Review of Submodel Hypotheses 

A weekly time step is optimal e 

Though a weekly time step allows for detailed examination of 
vegetation processes, and flexibility in implementing different 
grazing schemes, many of the participants 8 "mental models" of 
vegetation dynamics are essentially seasonal or annual. It would 
be valuable to eventually construct a mini-computer vegetation 
model based on an annual time step. This would permit easier and 
faster examination of the overall (i.e., whole ranch level) 
impacts of different grazing schemes and consequences of 
alternative hyoptheses. Repeated runs of the existing weekly 
time step models could be used to help generate seasonal or 
annually-based functional relationships. 

Three range sites (clay, loam and riparian) are sufficient . 

Since the available water capacities and initial plant 
biomasses assigned to each of the three range sites are computed 
as a weighted average of the ten range site representations on 
each pasture, the spatial heterogeneity of vegetation and soils 
is reasonably well preserved. Sensitivity analysis of the 
importance of available water capacities, percentage cover 
assumptions and soil erosion indices could be used to determine 
the effects of grouping range sites on the overall behavior of 
the model. Also, the history of the range site may generally 
alter plant composition and potential growth. This should be 
considered in parameterization of the model. 

Seven plant types are necessary and sufficient . 

Each plant type in the model is functionally different from 
the others, either in its growth functions or its value to cattle 
and wildlife. It therefore does not seem appropriate to further 
reduce the number of plant types. Participants seemed generally 



46 



satisfied that this categorization was an adequate compromise 
between accuracy and parsimony. 

The maximum potential growth rates in Table 3.4 are 
reasonable . 

The growth rates appear too high, even allowing for the 
temperature and moisture growth constraints which are 
subsequently applied. The rates were set at these high levels 
after observing the devastating effect on plants of even moderate 
levels of grazing. Improvement in the translocation rules will 
likely allow more reasonable maximum potential growth rates to be 
used . 

If evapo transpiration is less than 0.003 in/wk per Ib/ac of 
plant biomass* plant growth is constrained. 

This arbitrary method of ensuring water limitation of plant 
growth should be replaced by an estimate of root activity at each 
soil layer, which is used in the hydrology submodel's calculation 
of evapotransp irat ion . 

A high biomass of plants with shallow roots can prevent 
water from reaching deeper layers. 

This assumption has been discussed by Walker et al. (1981) 
as the dominant mode of competition between shrubs and grasses in 
semi-arid areas. It is not certain whether this mechanism also 
applies in the Saval Project area. 

Plants and shrubs deplete carbohydrate reserv es prior to 
seed set and repl enish them afterwards. When grazing occurs 
prior to seed set, reserves are depleted in proportion to grazing 
intensity * but grazing after seed set does not deplete 
carbohydrates . 



47 



The stability of grazing systems generally increases with 
the size of plant reserves (Noy-Meir 1975). Hence, these 
translocation rules deserve close scrutiny, particularly the 
response of plants and shrubs to cropping^ Though the above 
rules are generally supported by the literature, fall regrowth 
can also deplete carbohydrate reserves (Garrison 1971). The 
"causes" and "effects" of storage reserve depletion need further 
clar if icat ion . 

Percent cover can be calculated from above ground biomass . 

This hypothesis is discussed in detail in Section 3.4. 

The actions of plowing and seeding always produce the plant 
composition shown in Section 3 C 6. 

This assumption is clearly invalid. Seedling competition 
has very different rules from the ones used in the model for 
mature plants. The plant composition after seeding is also 
highly dependent on variations in weather. 



48 



4. HYDROLOGY SUBMODEL 

The main responsibility of the hydrology subgroup was to 
produce a dynamic representation of soil moisture which would be 
sensitive to changes in the vegetative cover. In addition, it 
was necessary to represent water quality and flow rates in the 
upper Mahala and Gance creeks and produce estimates of bank 
stability and erosion. 

Figure 4.1 represents the detailed charge of the subgroup 
defined during the bounding and looking outward exercises. These 
inputs and the required outputs define the level of abstraction 
of the submodel. 

In view of the rapid fluctuations in rainfall and 
temperature it was decided that the weekly timestep used by the 
rest of the model would be insufficient to capture the important 
dynamics of infiltration and the snowpack. Therefore, a daily 
timestep is used within the submodel and values are averaged over 
weeks for use as indicators and for communicating with other 
submodels . 

The sequence of calculations is diagrammed in the form of a 
flowchart in Figure 4.2. Each of these steps is explained and 
discussed in further detail below. 

The bookkeeping for soil moisture within the model is 
carried out independently for each range type within each 
pasture. Runoffs within the Mahala and Gance basins are summed 
for calculation of the flow rates of the two streams at the USFS 
boundary . 

Daily temperature and precipitation estimates form two 
important driving variables. In the limited time available it 
was felt that these were best provided by generating 
representative time streams and using these every year. Wetter 
or dryer, hotter or colder years could be simulated by adjusting 
the standard time stream with a suitable multiplier. Daily 
temperature records were available for 1980, but daily rainfall 
figures were only available for the second half of 1979. An 
artificial series was therefore generated using monthly figures. 



49 



















co 

Cd 
Q 
O 

<£ 




CO 

c 
o 

N 
■H 
S=i 
O 
JS 

CU 
CU 
(4 


CU 

u 

3 

•U 

CO 
U) 
CU 

a. 

S 
cu 

id 


1—1 
•H 

o 




CO 








.* 








CO 


CO 


J£ 


g 


co 




osi 








C 










5 


.u 


3 






o 








CO 








eS 


o 




S 


U-) 




H 








XJ 








w 


f=-4 


C 


• i-i 







«J 








E 








32 


CW 


•H 


X 






a 








«0 








H 






CO 


4= 




H- 1 


>> 




CU 


CU 




G 




O 


g 


CU 


S 


4J 




a 


■u 


co 


M 


u 


G 









CO 


u 




a. 




2 


•H 


3 


3 


u 


o 


■H 




O 


CU 


3 


>. 


CU 




H 


<—! 


O 


iJ 


CO 


•H 


4J 




H 


M 


4J 


h 


-3 






t8 


I— 1 


CO 




co 


CO 






u 


co 


XL 








3 


U-4 


•H 


CO 


O 


4J 




CO 


05 


•H 


cu 


00 






cr 




o 


I—) 


M 


•H 




Ed 




o 


cu 


C 








s 


s 


^ 


CU 


a 




►J 


>. 


5 


3 


.H 


• 




u 


CO 




cs 




•H 




P3 


f— 1 






M 


a 




CU 


CD 


i— 1 


.u 


1—1 


O 




< 


^ 


i— t 


e 


CU 


3 




4J 


U 


»H 


CO 


•H 


Ci) 




H 


CU 


• h 


CO 


cu 







CtJ 


4J 


O 







M 




as 


cu 


O 


CU 


u 




5 

1 


33 
I 


co 
1 


3^8 

1 

\ 


co 

1 


0, 
i 

\ 


SUBMODEL 1 
CHANGE I 


<3 
> 

/ 


3 
l 


CO 

1 


e 
i 


CM 
1 


CT 

■Q 
>i 

cn 


rH 



S-I 

>i 

0) 


























4J 
















o ^ 














4-) 


















1ft 










T3 
CU 

c 
en 

•H 

CO 
















H 




^ 


v 


CU 

a. 




CU 

oo 


CQ 














/ 






x 


>. 




c 


(T3 














/ 






X 


4J 




CO 














J 


f 




CO 








u 


CO 












/ 






J 




CU 






<D 












/ 






w 


cu 


00 




>> 


<r ^ 












f 






a 


c 


c 




43 


4J 


















o 


o 


CO 






















s 


CO 


U 




>> 


•H 


















P3 






r-s 


4J 


^H 






c 












E3 


a 


-a 


CU 


•H 


■H 



















CO 


CO 


c 


u 


o 


jQ 






•H 














■ f) 


CO 


o 


CO 








4J 












pd 


u 




CO 


a 


-H 






CO 












tea 


CO 


CU 


~^ 


CO 


CO 


en 




M 












X, 


a, 


u 


CO 


CJ 


c 


Z 


iJ 


U 












H 


•r4 


3 


C 







o 


a 


C 












O 


U 


4J 


o 


u 


Qj 


M 


cu 


CU 
















co 


4J 


cu 


CO 


H 


S 


o 












S 


CU 


CO 


v.*' 


j_i 


CJ 


<u 


c 












o 


r— 1 


a. 




CO 


Q) 


< 


oo 
eg 

C 

CO 


o 
o 

0) 












a! 

CO 


•H 

a 

CU 


43 


U 
• H 

i—l 


CU 

r— t 


0$ 




s 


l_) 












w 


r— 1 


M 


»H 


43 


6 






3 












hJ 


JJ 


0) 


43 


CO 


H 




u 


4J 












pa 


•u 


> 


CO 


i— < 


« 




0) 


co 












< 


CO 


o 


-3 


•H CU 


^r 




4-) 


•H 












H 


CJ 


CJ 


O 


CQ Q, 






CO 


o 












OS 






S-i 


> ^ 






3 

1 


S 
1 












> 


1 


1 


CU 

1 


CO u 

1 


•H 
fa 



50 



LOOP OVER 
DAYS IN 
A WEEK 




»VSTART ' <r 



EVAPOTRANSPIRAT 



ion] 



SNOW PACK 

MANIPULATION 



INFILTRATION /RUNOFF 



^ 



CALCULATE 
FROZEN SOIL 




CALCULATE 
CREEK FLOWS 



WATER QUALITY 
INDEX 



BANK STABILITY 



CALCULATE 
WEEKLY 

TOTALS/MEANS 

J 




LOOP OVER 
PASTURES AND 
RANGE TYPES 



Fig. 4.2. Sequence of hydrology submodel operations 



51 



It was assumed that the USFS pastures received 2.3 times more 
rain than the BLM lands and that the temperatures were only 85% 
of those on the lower BLM lands. 



4ol Soil Moisture 

Figure 4*3 indicates the processes that were considered in 
the balance of soil water. The soil was considered in three 
horizons: 0-10 s 10-20 and 20-40 in; water was measured as inches 
of water per inch of soil (in/in) and was considered in terms of 
its availability to plants. Thus soil water in each horizon 
varies from zero (no available water) up to an Available Water 
Capacity (AWC) determined for each soil horizon in each range 
type. Water infiltrating the soil fills the horizons from the 
top down so that water is not added to the 10- to 20-in layer 
until the 0- to 10-in layer is filled to capacity. Water is 
removed from the soil through the mechanisms of evapotransp ira- 
tion. Evapotransp irat ion removes water from the top layers first 
while evaporation from the bare soil removes water only from the 
top layer. 



4.1.1 Evapotransp irat ion 

Evapot ran sp irat ion (ET) is calculated by first estimating 
the maximum potential ET (ETp) using the equation of Jensen and 
Haise (1963) which represents a full cover of alfalfa with non- 
limiting water: 



ETp ■ (0.014 * Daily mean temp 

in ° Fahrenheit 



37) * Radiation/580 in Langleys 
(Equation 4.1) 



52 



u_ 
u_ 
O 

z 



< 



Z 

o 

I- 
< 



o 

CL 



z 
< 



UJ < 



o 

z 

C/> 





1 




il :* • 




■i-Si*:;;; 




<c 


4 


;;;;&;;;: 




Pi 




o 




;:;;Z;;;; 




C/3 ! 







I 

o 

< 
> 

UJ 



< 

CO 

z 
< 

DC 



o 


UJ 






\- 


< 


^ 








< 


CD 


_i 




o 


2 


o 

CO 




u. 


O 






< 


<r 






> 


UL 






UJ 




EPTH (in) 







Q 






o 

CM 



o 



SZ 

o 

c 

(T3 
ft 



(0 
> 

CO 

QJ 
.G 
4J 

C 

o 

CD 
i-l 
3 

co 

•H 



e 



o 
en 

oi 

c 

•iH 
4-) 

u 
cu 

u-i 
iw 
03 

en 

cu 

CO 

en 

0) 

u 
o 

Oh 






en 

•H 
(J4 



53 



The solar radiation in Langleys was assumed to follow the sine 
curve shown in Fig. 4.4 and daily mean temperatures were input as 
a driving variable as described above. 

To estimate the actual ET by plants, it was necessary to 
make some assumption about how actively the plants were 
transpiring at any given time. Ideally this information would 
have come from the vegetation submodel. However, in the absence 
of this interaction the relationship shown in Fig. 4*5 was used. 

In the refined version of the model developed after the 
first workshop the actual evapotransp irat ion is sensitive to the 
amount of water in each soil horizon and to root activity using 
the relationships 



rop . act ive 1^ 3 

canopy * 3 * 2 

i s l 



root 

activity^ * % saturation^ 

(Equation 4.2) 



The root activity term allows for the interaction between the 
depth of roots and the level of water in the soil. For the 
purposes of this model, it is assumed that root activity in all 
soil levels is the same as the proportion of active canopy. 

Evaporation from bare soil was assumed to depend on the 
potential evapotransp irat ion and soil moisture in the top horizon 
according to the relationships 

in/in soil moisture ETp 
Evaporation ■ in top horizon * AWC 

(Equation 4.3) 

Thus when the top soil layer is full of water, evaporation occurs 
at the same rate as ETp and declines linearly with soil moisture. 
Total water loss is obtained by weighting ET and bare soil 
evaporation by the relative percentages of soil surface, covered 
and bare respectively. 



54 







800- 








_ L 










CO 










> 










UJ 










_i 


600- 








o 










z 










< 




/ 






-J 










** - ' 


400- 








z 










o 










H- 










< 










5 


200- 








< 










£T 


n . 










i 


AN 1 JUNE 


De'c 31 






DATE 




Fig. 


4.4. 


Pattern of solar radiation 
year on the Saval Ranch. 


levels through the 



100.-. 



80 - 




NOV 30 



Fig. 4.5. Pattern of activity of vegetative cover through 

the year on the Saval Ranch in terms of effective 
evapotranspiration levels (maximum = 100%) . 



55 



4.1.2 Snow Storage 

When the temperature is below 32°F, precipitation is assumed 
to occur as snow and is added to the snow pack for each area. 
When snow storage exists and the temperature is above 32°F, snow 
melt occurs according to the relationship of Stewart et al. 
(1975) : 

MELT (in) = (mean daily temp - 32)/10 (Equation 4.4) 

Melt water for each day is added to the precipitation and 
used for the calculation of infiltration and runoff. 



4.1.3 Runoff and Infiltration 

The calculation of runoff was based on the soil conservation 
service (SCS) curve number technique as described in Smith and 
Williams (1980). For daily precipitation (P), runoff is 
calculated using the relation: 

RUNOFF = (P-0.2*S) 2 /(P+0.8*S) if P > 0.2*S 

=0 if P < 0.2*S 

(Equation 4.5) 

where S is a retention parameter calculated from the curve number 
(CN) for the soil being considered and the relative water content 
of the soil: 

S - Smax (AWC - soil moi sture ) / AWC (Equation 4.6) 

Smax = (1000/CN) - 10 (Equation 4.7) 

In the version of the model produced during the first 
workshop only the water content of the top soil layer was 
considered in this calculation. Unfortunately this means that as 



56 



soon as the top layer is saturated, none of the lower layers can 
fill, because all precipitation runs off. In the period before 
the second workshop this was amended to be the mean water content 
(in/in) of the complete soil horizon. 

To ensure that runoff and thus infiltration would be 
sensitive to changes in vegetative cover the linear relationships 
of Branson et al. (1981) were used to relate curve number to 
per c ent cover : 



CN = A - 



* % cover 



(Equation 4.8) 



Values of A and B are given in Table 4.1 for each of the 
hydrologic soil groups. 

In the updated version of the model produced prior to the 
second workshop, the values shown in Table 4.2 were used. 

The amount of precipitation and snow melt left over after 
the removal of runoff represents infiltration and is added into 
the soil water. 



4.2 Erosion 

The quantity of soil eroded from each range type was 
calculated using the relationship: 



Soil loss (t/ac) =K*R*LS*C*P 



(Equation 4.9) 



where 



K = erodability factor (see Table 4.1) 
R = parameter related to storm depth and intensity 
LS = length slope parameter (see Table 4.1) 
C ■ parameter related to ground cover 
P ■ mechanical factor related to soil conservation 



57 








u 




T3 




>i 




jC 




'O 




c 




03 




co 




0) 




cu 




>i 




4J 




<U 




Cn 




c 




03 




S-4 




W 




3 









•H 




5-1 




03 




> 




u 







< 


M-) 


*-«. 




rH 


CO 


(U 


l-l 


TD 


<D 





-P 


s 


0) 




e 


IM 


03 





u 




03 


c 


CU 




■H 


5-1 


co 


CD 


s-i 


& 


0) 


6 


> 


3 




C 


JJ 




co 


CD 


S-l 


> 


•H 


u 


4-1 


3 


~— ' 


O 






w 


TS 


a 


C 


3 


03 







5-1 


CU 


en 


CU 







rH 


rH 


■H 


ra 







to 


i3 




-P 





CP 


>H 


C 


en 


CU 





J 


H 



<u 

rH 

j3 
03 

Eh 



CQ 





^ 




cn 




J 


£ 


»_* 


-P 




Cn 


<u 


c 


a 


cu 





j 


rH 




CO 



>i 



n 

03 



a 



o 


a 


•H 


3 


cn 





5-i 


rH 


01 







5-1 


H 


TS 


•H 


>i O 


Hfj 


CO 



<u 
a 
>i 
.p 

0) 

cn| 
C 

03 



u 


S-l 


<U 


cu 


> 


> 


O 





U 


o 


0\° 


oV» 


O 


o 


LD 


in 


V 


A 



IT) 






00 
CM 



o 



cn 



oo 

CN 



00 



LO 



CQ 



00 

rH 

I 



T3 

H C 
03 

<U rH 

o o 
o 

3 



CO 

>i 

E 

03 





C 
CU 

CO 

< 



00 

rH 

o 

o 



an 

00 



o cm cn GT* 



a 



«3« ^ 



I 

o 



CN 
H 




-P 

-P 




i 

CO 



CN 

rH 
I 

O 



CU CU 

o o 

H H 

CO 



>i !>f >i >i 

s e s e 



03 




03 




03 




03 




CO 

x: 

-P 
3 


CO 



cn 

H 

o 

o 



GTS 



o 


cn 


cn 


r* 


CN 


cn 


LO 


cn 


CTS 


r- 


CT\ 


cn 


rH 


cn 


CN 


m 


r* 


rH 


rH 


LD 



CN 



CN 


CO 


r^ 


cn 


r» 


CN 


00 


on 


r* 


r^ 


cn 


CN 


cn 


<a< 


cn 


CO 


CN 


t 


m 


cn 



o 



5 



03 
CU 

s 



cu 



CN 

rH 
I 

O 

rH 

c 

03 



<o 



CN 



c 

OS 



>i >i 

03 03 

H H 

u u 





■P 

cn 

C 

•H 


o 

u 

03 

s-i 
(U 

> 

o 


4J 

c 
cu 
u 
u 

<D 
Cu 

C 


CU 
CO 
03 
XI 

u 
CU 

.Q 

S 
a 
c 

> 

S-4 

3 
O 

CO 

o 

CO 
Cn 

c 

■H 

-P 
03 
i— I 
3 
O 

rH 

03 
O 

U 



<4-4 
CO 

u 

<u 

4J 

a> 

6 
03 
u 
03 



S-l 

cu 
> 



-p 

c 

CU 

o 

s-i 
CU 

a 



CQ 



Z 

u 



58 



Table 4.2. 



Length slope and curve number parameters for the 
second version of the model. 



Range type 



Clay 
Loam 
Riparian 



Length 
slope (LS) 


A 1 


B 1 


4.5 
2.8 
0.16 


83 
89 
91 


0.28 
0.18 
0.13 



Parameters for calculating SCS curve number based on percent 
cover according to: 

CN = A - B x percent cover 



Assuming that storms have a two hour duration, "R" was calculated 
us ing : 

R = 10 * Total Precipitation °« 037 ^ (Equation 4.10) 

The parameter "C" was calculated directly from percent ground 
cover using the relationship of Wischmeier and Smith (1978): 



= 10 -0.552 (1 + 0.029 % cover) 



( Equat ion 4.11) 



" P " was set at 1 for current purposes. Possible soil 
conservation actions would reduce the value of P and thus reduce 
so il loss . 

In general then, the amount of soil lost to erosion depends 
on two factors; a relatively invariant set of physical 
parameters, and the percent cover. The effect of percent cover 
in any given area is shown in Fig. 4.6. 

No attempt was made to determine the fate of eroded material 
and so there is no deposition of soil anywhere in the model. 



59 







CD 






SZ 






■U 






c 













(0 






cn 













H 






H 






■H 













cn 




o 


T3 




a .-<= = = =—, =j 


" 2 


c 




/ • 
/ 1 

/ ' 
/ ' 

/ ! 




03 
S-i 

> 







/ ' 

/ ! 


111 


c 



■H 








03 

4J 




Bo, «= cc r- .^ = .- . 1 «■ <n> w» » .=> <=» 


O H 

E in ^ 

UJ 


tu 
> 




/I 1 






o 






/ ' ' 


ac 


JJ 




/ ' ' 


UJ 


c 




t s 


0= 


CD 

u 

u 
CD 

a 

c 
cd 

CD 
5 




i i 


- O 


-P 




1 I — 1 


CD 

j2 


© 


m O o 




o 


Sf CVJ 


Q< 


CKS 




•H 




(uinunxDW jo %) SS01 1!0S 


Fig. 4.6. The relationsh 
Saval Ranch. 



60 



4.3 Freezing of Soil Water 

Due to constraints of time the dynamics of the freezing of 
soil were not considered in great detail. The maximum depth of 
freezing on the ranch was considered to be 30 in with a more 
normal figure being 6 in. It was assumed that the soil melted 
and froze at the bottom at a rate of 0.1 in per degree day. 

Change in 

frozen depth = 0.1 * (32 - Daily mean 

in inches temperature °F) 

(Equation 4.12) 

In the current version of the submodel there is no interaction 
between infiltration and soil freezing. It was felt tiiat when 
the soil surface was frozen, infiltration would be reduced by as 
much as 80%. However, the chosen representation of freezing 
dynamics did not allow determination of the state of the soil 
surface . 



4.4 Streaaflow 

The flow in the Mahala and Gance Creeks was considered down 
to the USFS boundary. These streams are important for hay 
production and their populations of cutthroat trout. The flow in 
the North Fork is not dependent on any processes internal to the 
Saval system and so was input as a driving variable. 

Virtually the entire upper Mahala basin is contained within 
the South East Independence pasture while the portion of the 
upper Gance basin not in this pasture is outside the Saval 
boundary. Given the runoff calculations described above and the 
area of each range type, total runoff can be calculated for each 



61 



of the basins within the Saval area (Table 4.3). The area of the 
Gance basin outside of the ranch is considered to have a uniform 
soil type and the same rainfall as the USFS pastures allowing 
calculation of runoff. 

No attempt was made to take account of the dynamics of water 
within the stream bed. 



Table 4.3. 



Areas of range types in Gance and Mahala Basins 
within South E. Independence Pasture. 



Range type 



Loamy slope 14-18 
Aspen woodland 
South slope 12-14 
Wet meadow 
Claypan 12-16 



Areas (mi ) 



Upper Maha 


ila 


Upp 


er Gance 


0.63 






1 .66 


0.85 






0.64 


1 .27 






2.19 


0.10 






0.16 


1 .34 






1 .97 


aval = 2.07 


mi 2 







Area of Gance Basin outside Saval 
Representative curve number = 80 



62 



4.5 Water Quality 

A water quality measure is required in the model solely as 
an indicator of cutthroat trout habitat for the wildlife 
subgroup. In view of this specialized function a decision was 
made to only consider sediment delivered to the streams. No 
attempt was made to dynamically model the deposition and 
resuspension of sediment within the stream although the water 
quality index was scaled by flow rates. 

Sediment delivered to streams has two components: 

(1) upland erosion, and 

(2) stream bank erosion. 

For upland erosion the model assumes that of the total quantity 
of soil eroded in a watershed only a proportion (0.2) reaches the 
streams, the rest being redeposited. 

Bank erosion is assumed to be a linear function of the 
number of cattle per mi in the riparian zone: 

Tons Soil 1 



Eroded/Day * 500 * # cattle/mi of riparian ( Equat ion 4.13) 

Thus, a water quality of zero indicates crystal clear water and 
higher values represent higher sediment loads and lower water 
quality . 



4.6 Bank Stability 

Bank stability is considered to depend on two factors in the 
riparian zone: number of cattle per mi and percent cover, (Figs. 
4.7 and 4.8) and are combined linearly: 

% Stable Bank = % Vegetative Cover - 0.2 * # cattle/mi 

( Equat ion 4.14) 



63 





I00-, 

as 

5 80- 
< 

S3 
UJ 

^ 6 °- 

< 

1- 

40- 
g 20- 

UJ 

0= 

0- 


s 


1 I J ■ i 




20 40 60 80 100 




CANOPY COVER (%) 


Fig. 4. 


7. Relationship between vegetation canopy cover and 




stream bank stability on the Saval Ranch when 




cattle are absent. 



100 




NO. 



200 300 400 500 
CATTLE PER MILE 



Fig. 4.8. Relationship between cattle per mile of riparian 
zone and stream bank stability on the Saval Ranch 
when vegetative canopy cover is 100 percent. 



64 



Percent vegetative cover is calculated as a weighted average of 
the percent covers of the various vegetation types: 

% Vegetative 4 * % Woody 3 * % Shrub 2 * % Grass 
Cover = Cover + Cover + Cover 



(Equation 4.15) 

This reflects the fact that woody vegetation holds the banks 
together best but shrubs and grass are still much better than 
bare soil. 

4.7 Review of Submodel Structure and Hypotheses 
4.7.1 Time Resolution 

The one overriding feature of the hydrology submodel is its 
use of a daily timestep. This level of resolution was considered 
by the subgroup to be necessary for a reasonable representation 
of the important dynamics of soil moisture. 

The main drawback of using a daily timestep is that it 
greatly increases the running time of the final model. This in 
turn makes gaming and manipulating the model more time consuming 
and ezpens ive . 

In view of the timescale of events (such as rain storms) a 
daily timestep is probably necessary to produce a precise 
prediction of soil moisture dynamics. However, because of the 
level of understanding of the relationships between vegetation 
and soil water, a cruder representation may be more useful in the 
framework of this model. 

The possibility of using a weekly timestep was investigated 
using the simple APPLE microcomputer model presented at the 
second workshop. This produces timestreams of soil moisture that 
resemble those produced by the larger model suggesting that a 
weekly timestep is feasible. 



65 



4.7.2 Important Hypotheses 

The hypotheses which relate vegetation, physical parameters 
such as temperature and solar radiation, and evapotranspiration 
and infiltration seem weak in this model. This is partly due to 
the poor linkage between the vegetation and hydrology submodels 
but there also appears to be a gap in understanding. Given the 
possible importance of water/plant interactions (Walker et al. 
1981) this would seem to be an important place to direct field 
studies . 



4.7 .2 .1 Water 

Maximum potential evapotranspiration is calculated using an 
empirical relationship based on daily mean temperature and solar 
radiation . 

This relationship may or may not be applicable to the Saval 
Ranch. Although it is largely a physical relationship, other 
factors such as wind and relative humidity are not considered. 

Evaporation from bare soil is the maximum potential 
evapotranspiration reduced in proportion to the p ere ent 
saturation of the top soil layer. 

This would seem to be qualitatively correct. If the top 
soil is dry then there will be minimal evaporation and if there 
are pools of water on the surface then evaporation will occur at 
some maximal rate. It seems likely that there is information in 
the literature on the maximal rate and the form of the 
intervening curve. 

Evapotranspiration by plants is the maximum potential evapo- 
transpiration reduced as a function of the activity of the canopy 



66 



and (in the second version of the model) the soil moisture con- 
tent and root activity in each horizon. 

This linkage between vegetation and evapot ransp ir a t ion is 
weak and needs to be reconsidered. At present, the percent 
active canopy and root activity figures are created in the 
hydrology submodel completely independently of the state of the 
vegetation. The simplest way to improve this relationship would 
be for the vegetation submodel to calculate root activities in 
each soil horizon based on the composition of the vegetation. 

Water is added to the soil by infiltration and fills the 
horizons from the top down. 

This seems reasonable although it seems probable that lower 
horizons start to fill before the upper ones are completely 
saturated . 

Curve numbers used to calculate infiltration rate are linear 
functions of percent cover. 

This is almost certainly incorrect but was the simplest way 
to include a perceived effect of percent cover on infiltration. 
Better relationships may be available in the literature but it 
seems likely that field studies will be necessary if this is to 
be improved on. 

Water is removed from the soil by ev apotranspiration from 
the top down. 

This hypothesis is all right as a first approximation but it 
reflects the lack of linkage with the vegetation submodel. 

Water does not diffuse b etw een soil horizons in the 
unsaturated state. 



67 



This is a commonly made assumption in hydrology models and 
does not seem to lead to large problems* 

Streamflow is the sum of all runoff in a given watershed* 

This hypothesis should give a qualitatively correct picture 
of variations in streamflow. But the absence of any subsurface 
flow and interchange with the stream bed results in a more 
variable flow than that observed,, 

The enow pack melts at the rate of 1/10 in for every degree 
day above freezing. 

This estimate is probably not accurate due to the effects of 
other factors, particularly rain* There is not however a high 
priority to improve this as the effect on the model would not be 
great . 

The presence of frozen soil has no effect on the 
infiltration rate. 

This is obviously false but to model this better requires a 
more accurate representation of the way in which soil moisture 
freezes, especially at the surface. 



4.7*2.2 Soil 

Soil loss in tlac is a function of the percent cover and the 
physical characteristics of a particular area. 

The applicability of this relationship to the Saval Ranch is 
unknown. It seems likely that the species composition of the 
vegetation may be important as well as the percent cover. (The 
percent cover vs. C factor values and subsequent relationships 
were derived from data for pasture and rangeland as a lumped land 



68 



type. The values were for a vegetation canopy of "tall weeds or 
short brush with average drop-fall height of 20 inches".) 

The fate of eroded material not delivered to streams (a 
constant proportion) is not considered . 

No attempt was made to model the transport of eroded 
material as the quantities involved were not considered to be 
s ignif icant . 

Bank stability is a simple function of p er cent cover 
{weighted so that woody plants are more effective) and number of 
cattle in the riparian zone. 

This relationship is almost certainly too simplistic and 
should be looked at in more detail. The vegetation information 
should be either available or easily collected, but the trampling 
effects due to animals are much more in question. Number of 
cattle per mi is not the only variable— season of use and dura- 
tion of use are likely to be very important. 



69 



5. LIVESTOCK AHD ECONOMICS SUBMODEL 

The livestock and economics subgroup was responsible for: 

(1) developing a conceptual framework relating range 
conditions and winter food supply to livestock growth 
and reproduction; and 

(2) using these dynamics of livestock production plus costs 
of management actions taken in all submodels to examine 
the economics of the Saval Ranch operation and some 
economic considerations pertaining to Elko County (Fig. 

The submodel had to be able to mimic the cattle movements 
specified in the Saval Management Flan but also had to be 
sufficiently flexible to mimic any other grazing system 
participants might wish to simulate. 

The livestock population is divided into three 
subpopulat ions : cows, calves, and yearlings. The subgroup did 
not have time to consider the consequences of stocking other 
domestic animals, other breeds of cattle, or salting as a means 
of livestock distribution control. The group considered the 
county economic effects of hunting effort on the Saval Ranch but 
did not incorporate the rules for change into the model because 
the wildlife subgroup only considered effects of constant hunting 
effort . 



5.1 Cattle Movement 

A major objective of the subgroup was to simulate movement 
of the cattle on the Saval project area according to alternative 
2 in the Management Plan and yet maintain enough flexibility to 
ensure the model could be used to mimic other realistic grazing 
schemes. The Management Plan grazing schedule is shown in Table 



70 

























CU 


































fl 





































































N 


































3 


































CO 






























CO 


i—i 


•r-l 






























-J 


3 


u 






























Ctj 


CU 


CO 






























Q 


4-1 


a 






























O 


CO 


•H 














<u 
















S 


CU 


I-l 














a 






CU 










33 


















>» 






3 










33 


3 


c 














4-1 




CU 


i—i 










CO 





•r-l 












CO 






3 


CO 












•r-4 














OS 


^ 




r-l 


> 










oj 


4-1 


en 










. 


O 


j2 




CO 












w 


CO 


>. 










en 


H 






> 


4J 




4J 






cr 


4-1 


CO 










<3 


-a 






CU 




-C 






H 


CU 


"O 










o 


1—1 




4J 


X 


-a 


00 






O 


00 












3 


H) 


o 




CU 


u 


i—i 


3 








CU 


CU 













Q 


en 


CU 


^ 


CO 


o 


O 






O 


> 


en 










<_i 


2 

H 


14-1 


3 
3 


u 

CO 


£ 


en 


-Q 






H 


CU 


3 








CU 
T3 


-Q 




cu 


CU 


e 


4-1 


^ 


>> 






CO 


I-l 


CU 













0) 


> 




o 


CO 


CO 






Pd 


CJ 


i—i 








CJ 


a 




Xi 


CU 

U 


4J 

O 


CU 

u 


JB 


-C 






33 


CO 


4-1 
4-1 








l-i 


CO 




. 




CU 


•H 


en 


en 






< 


• 


CO 








CU 


o 




en 


4J 


u 


T3 


C 


a 






H 


en 


CJ 








4-1 




-Q 


CU 


•r-l 


C 


o 









pi 


j3 










3 


•r-l 




l-H 

1 


a 
i 


T3 

1 


•H 
1 


4J 
1 


4J 
1 

\ 






< 

> 


i—i 
1 


1 








asture 

ubgroup 

entered into comp 


6 


c 



u 

CD 

C 
CD 

O 
















(J 

































Q W 
















Q. en 


4-J 
















2 a 














4J 


w 
















<: o 


2 














T3 >. O 


<u 
















s 

CJ) 3 


U 














3 ,£> 3 
CO 

"O 4-1 


> 

•H 
















O en 


as 
O 














CU CU 3 


r-l 
















H 














4J U -Q 


















CO O 

W r-l 














•H CU 

en -o T3 


CU 
















> S 
















•r-l CU 
















M O 


CO 

W 

rJ 

oi 














cu en l-i 


4-J 
















rJ 2 














00 3 cu 


















O 














3 O T3 


U-4 
















U 
Ed 














y ra 
ot c 
onsi 

















*m 






| 


fe 










CO 














•* 






1 


\ 










-O 3 CJ 


CD 

•r-l 












y 










\ 










i-l CN M 


4J 










CN 


s 








CO 


\ 












-r-l 
1—1 










en 

r- 1 










rJ 

CU 


\ 




















CO 










Q 














•r-l 










& 






r* 




O 














Ju 










■H 






CJ 




S 














•H 










a 






O 




33 














00 

c 




0) 




a 


CO 


en 




i-i 




s 










CN 






a 




o 




-a 




en 




CO 




rH 






en 






•r-l 


00 


en 


o 


CU 




CU 








C 






>, 









4J 


3 


co 


• 1-1 


CU 




>^ > 




ttJ 


3 


o 






CO 




a 






•h 


CU 


4J 


u 




co -r-l 




CU 


O 


•H 






-a 




CO 


CO 


■N 


-a 


en 


en 


-Q 




,C. i—l 




— I 


•r-l 


4-1 


4J 








cu 


2 


CO 


CU 




CU 










H 


4-1 


CO 


c 




CU 




O 


CU 


CU 


00 


S 


CU 




.. 




O 


CJ 


4-1 


CU 




en 




u 


i— i 


h 


<4-l 


s 


o 


u 




00 






3 


CU 


4J 




3 






H 


CO 




• 1-1 


-a 


CO 




a 




s 


-o 


00 


3 








cu 


o 




r— 1 


-o 




3 




•r-l 




o 


O 


CU 


O 


m 


r^ 




sz 


< 


5^ 


CO 


CU 


u 


l-i 




r-l 




pej 


S-i 


> 


CJ 


en 


CO 




Eh 




ja 


4J 


CU 


0) 


CU 




r^ 




Pti 


a 






>> 


3 








c 


u 


-C 


4-1 




CU 








CU 


C 


CO 


O 








oo 


CU 


-O 


4J 


1—1 


CN 


en 




co 


>^ 


I-l 


•r-l 


-o 


•r^ 








c 


e 




O 


CO 


00 


00 ">>. 




W 


CO 


CJ 


CU 




4J 




■ 




■r-l 


CU 


CU 






c 


s oo 




-I 


A 


CO 


4-1 


M 


CO 




r— | 




H 


1—1 


00 


J* 


X 


•r-4 


-i a 




33 




~^_ 


O 


CU 


CU 








o 


a. 


c 


o 


cj 


4J 


CJ -H 




< 


en 


• 


u 


4J 


l-i 




* 







a 


CO 


o 


o 


r-H 


3 >. 




H 


3 


en 


a 


3 


o 




CD 




4J 


3 


-C 


4-1 


4J 


CO 


V 3 




a! 


O 


X) 




3 


CU 








CO 

1 


en 
l 


CJ 

1 


en 
1 


en 

1 


en c. 
1 


-I _Q 
1 1 




< 
> 


4J 

1 


i—i 

1 


5-5 

1 


j3 
1 


u 

1 




CJi 
■H 
pt| 



71 



5.1. There are nine time slots to which the different pastures 
are allocated each year. The key to moving cattle according to 
the Management Plan is Table 5. 2. a which shows how the order of 
grazing changes from one year to the next. For example, a field 
grazed in the first slot in any year, (April 15 - May 8) is 
grazed in the eight slot the next year, (November 1 - November 
30). The field grazed in the eight slot in any year (November 1 

- November 30) is grazed in the second slot the next year (May 9 

- May 31), and so on. 

Table 5.1. Grazing schedule for the Saval Management Plan. 



Slot 



1 

2 

3 

4 

5 

6 
Wean 

8 
Rest 



Schedu le 



April 15 - May 8 
May 9 - May 31 
June 1 - June 15 
June 16 - June 30 
July 1 - August 15 
August 16 - September 30 
October 1 - October 31 
November 1 - November 30 
Not grazed for one year 



# weeks 



3 
3 
2 
2 
6 
6 
4 
4 



There is also associated with each pasture a maximum number 
of years that it can be grazed in the same slot within the order 
(Table 5.2.b). For example, the Independence pastures retain the 
same slot in the order for two years, while the Upper Mahala 
retains the same slot in the order for just one year. Table 
5.2.b shows how each pasture is grazed through 12 years of the 
accepted Management Plan grazing system. 

This framework therefore simulates the present Management 
Plan but can be changed to mimic any alternate grazing system by 
changing the slot to which a field changes (Table 5. 2. a), the 
number of years that a field can be grazed at the same position 
within the order (Table 5.2.b), and the number of pastures. This 
framework could be used to simulate simultaneous grazing by more 
than one herd, such as was proposed in alternative 3. To do so, 
the necessary information (Table 5. 2. a), would have to be derived 
for each herd. 



72 



LO 

03 

—I 

ig 

C-1 



X) 











01 0> H 








i_t>jS ui 


03 








S4J 3 










•H 


0) 








N C K 


E • 








(0 -H -H 


■H U 








u 


4J JJ 








trn J 











03 


3 








0» M *H 


•H «. 








C 3 0! 


» 








•rt « U 


TJ 








c w 


gj » 








•H fl ^ 


N ro 








s a,< 


ra 








u — • 


U «■ 








o> a) 


Cr> JJ 








*■'£ C 


03 








0> E-> "H 


03 03 








"3 


•H in 








0) 










h ■ — 1 


J* » 








C J- 


0) "9- 








u-t u <a 


0) 








0> -M 


kj 03 








£ *J 


U *J 








£ -M 0> 











JJ « £ 


Oar-I 








-1 O, Eh 


0) 0! 




H 




M 


0) 




O 




Cr< 


£ 0) 


en j 




tJl C o 


a: £ 


OS 


0} 




^H *-i 0] 


—I 


< 






(0 N 2 


U 4J 


w 


u 




A3 
- — i-l S-» 


0) 
S 3 


X 


§ 


< 
HHHHNOI \<-H 


f0 t_Ti 


O "H 


Et. 


w 


z 


0) 


J 


O 






i-l c 


T3 




z 




■• IB +i 


» 0) 


4k 


M 




= 03 


0) N 








0) >i B» M 18 








4J C 


O u 








bin 


M-J CTl 








>l<-4 H 


a» 








en ra 


Vj 03 








IB 


0) -H 








D> T3 JZ 








3 — 03 Eh <0 








■H .fl JJ 


H 








N 03 


(0 




E-< 




03 »«H 


• £ 




o 




h in h 


i-i IT] 




J 




0"> U. 


a s 




w 


■U 3 


ITS J-i 


0) 






03 ro 


H 0) 0) 


>i 0) 


E" a 


OOrHTP 03 VO in 0> fN 


to >»«a 


r-( 


X 2 


« 2 


> u 


>TJ 


w 


H 




(0 0) 


>-l TD 


Z 


Eh 




U3 > 


0) -H 








■h ai 


> 2 








0) j-> J3 


0) 








£ 3 4J 


% 








j-> o 


M • 








0) c 


0) rj 








U II) -w 


T3 J-i 








3 


U 0) 








-C 











3 O 0) 


^ 








N 


0) eo 








•H U (0 £ 








4J O U 


JJ » 




Eh 




IB UH tj> 


i-4 




O 




— 1 


0) 


E- 


hH 


c 


SOU 


On «■ 


Z 


W 


(0 


E Oh 


3 fN 


Ed 




<— in m^ uiio qjoo 


•HCU 


(0 


OS 


u 


S 


03 0) 


J3 » 


cc 


2 




3 JJ 


O 00 


13 


M 




o a 1 "-! 




U E* 




js a> ai 


O * 








E-i 03 i-i 


JJ iH 









Ed 
OS 

3 

eg 

< 

a- 















3 




JJ 










fH 


H 


18 




ta 


l-H 


CM 


ro 

JJ 
03 


•"J" 


m 

(N 


vo 

fN 


tti 

2 

3 

ra 


oo 


0) 
OS 


(N 


00 


0) 


ro 

JJ 

03 




in 

H 


CD 

2 

3 
IB 


rH 


"3* 


00 


.-H 


f 


0) 

05 


vo 

(N 


m 

(N 


0) 

3 
18 


fN 


ro 

JJ 
03 


H 


fN 


<"0 

JJ 

01 


■V 


in 


VO 


0) 

2 

3 


00 


0) 


rs 


00 


0) 

OS 


ro 

4J 
0] 


in 
CM 


VO 
fN 


0) 

2 

3 
ra 


-H 


T 


CO 


~i 


•f 


0) 


VO 


in 


0) 

3 
<0 


fN 


ro 

JJ 
01 


^H 


fN 


ro 

jj 

03 


•^ 


fN 


in 

fN 


0) 

2 

3 
ro 


CO 


0) 
OS 


(N 


00 


0) 

— 


ro 

jj 
03 


in 

H 


VO 

i-H 


0) 

2 

(0 


-H 


T 


CO 


pH 


T 


0) 
Cj. 


in 

fN 


VO 

fN 


0) 
2 

3 
<0 


fN 


ro 

JJ 
01 


i-H 


(N 


m 

JJ 
0) 


'^ , 




in 


0! 

2 

3 
re 


CO 


0) 

OS 


CN 


00 


0) 


ro 

JJ 

01 


VO 
fN 


in 

fN 


0) 

2 

3 
rO 


H 


^ 


CO 


r-i 


T 


0) 

a: 


in 


VO 


0) 

2 


fN 


ro 



0) 
0> 
u 

u 

a 
o> 

0) 

j3 
en 

u 

0) 

3 


J 



z en 



rfl C 3 



ra 

j3 
ra 

2 

u 

0) 

Qj 

a 

Z3 



(0 

ra 

2 



0) 
"O 
3 
0) 

CJ 

T3 
3 



0) 

3 
0) 

Qj ^ 

0) O 

T3 rO 

3 JO 

m ra 



JJ 

03 

CO 0> 

OS 



ai 

03 

u 
U 



2 td m < 



a 

03 

CD 



Vj 
03 

a 
a 

3 



73 



The actual Management Plan is flexible in the timing of 
cattle movement to and from pastures (i.e., +. one week). Parti- 
cipants stated this was designed to prevent over- or under- 
utilization of pastures. Cattle are usually moved after approxi- 
mately 50% utilization of the preferred plant types (perennial 
decreasers, forbs) has been reached. There was insufficient time 
during the workshop to include this state-dependent rule to 
determine cattle movement and the model presently considers the 
schedule in Table 5.1 as fixed. 



5.2 Feeding 



5.2.1 Forage Selection 

There was considerable discussion in the subgroup about how 
cattle feeding is influenced by availability of different types 
of forage, availability of water and the interrelation of these 
two factors. Two extreme hypotheses were proposed: 

(1) water availability is the dominant factor and cattle 
will consume low quality food near a water source 
before moving far from the water; and 

(2) availability of good forage is the dominant factor and 
cattle will roam far from a source of water in search 
of highly-preferred food types. 

The first hypothesis (hypothesis 1) was implemented in the 
model during the first workshop. Discussions later in the 
workshop revealed that an intermediate hypothesis (hypothesis 3) 
is probably more realistic; cattle will search far from water for 
preferred forage but will stay near water and feed on unpreferred 
forage if preferred forage is not available anywhere. This 
hypothesis was implemented between the first and second 



74 



workshops. Either hypothesis can be used when executing the 
mode 1 . 

Currently, the submodel considers two forage classes within 
each forage type: forage "near" water (within 1/2 mile of a 
water source) and forage "distant from" water. For each pasture 
participants estimated the percentage of forage "near" water 
(Table 5.3). This percentage is assumed to be the same for all 
range sites within each pasture, an invalid assumption for 
extremes such as riparian and south slope sites. Management 
actions such as construction of stock ponds, irrigation systems, 
and actions that change the land area cattle find near water, can 
therefore be crudely simulated by changing the percentages of 
land near water. 



Table 5.3. Percentages of pastures classified as "near" water 

for cattle. The percentages are assumed the same for 
all range sites within the pasture. 



Pasture 



Darling West 

Darling East 

Lower Mahala 

Middle Mahala 

Upper Mahala 

Lower Sheep Creek 

Upper Sheep Creek 

East Independence North 

East Independence South 



% Pasture "near" water 



100 

100 

100 

100 

100 

50 

65 

80 

65 



The distribution of grazing within a pasture is determined 
by three factors: range site, food preferences, and nearness to 
water. Under hypothesis 1, cattle attempt to meet all their 
nutritional requirements on particular food types, range sites, 
and nearness to water, following a preference order structured 
within each category (Table 5.4). Residual nutritional 
requirements are met on the next preferred combination of food 
types, range sites, and nearness to water. That is, any residual 
nutritional requirements will be met by changing distance to 



75 



Table 5.4. Cattle preferences for plant types, water availa- 
bility, and range sites. It is assumed that cattle 
will graze first on the most preferred combination of 
the three selection factors. 



2. 
3. 



Plants 

Perennial grass 
decreaser s 

Forbs 

Cheatgrass 



4. Perennial grass 
increaser s 

5 . Bit terbrush 

6. Other Shrubs 

7 . Sagebrush 



Water 



A. Near water 

B. Distant from 
water 



Range Sites 



I e Wet meadow 

II . Aspen woodland 

III. Loamy bottom 

IV. Loamy 8-10 

V. Loamy 10-12 

VI. Claypan 10-12 

VII. Claypan 12-16 

VIII. Loamy slope 10-14 

IX. Loamy slope 14=18 

X. South slope 12-14 



water first, plant type second, and range site last. This 
implies that cattle will stay on a particular range site and feed 
on all plant types near and distant from water before moving to 
another range site within that pasture. 

There are obvious inconsistencies in this conceptualization. 
For example, the proportion of range sites near to water is not 
constant across all range sites. Also, cattle are likely to 
select on the basis of nearness to water or food preferences, 
rather than range site. The fact that cattle appear to select 
riparian range sites over others is probably due primarily to 
proximity to water, rather than the kind of vegetation found in a 
riparian zone. Another problem is that cattle were assumed to 
first try to meet all their requirements on the most preferred 



76 



forage irrespective of the abundance of the forage, then all 
their residual requirements on the next most preferred, again 
irrespective of forage abundance and so on. It is more probable 
that cattle selection is made according to both relative 
abundance and simple preferences. 

In hypothesis 3 cattle selection is made according to both 
forage abundance and simple forage preferences. The model used 
is drawn from predation research (Holling 1959, Charnov 1973). 
The formulation given below calculates the biomass of each forage 
consumed by cattle given cattle intake rates, and forage 
preferences and relative forage abundances: 



a i,j P i,j F i C j 



Ej4 = L (Equation 5.1) 

1 + 2a njj P n>j F n h n>j 
n = l 



where : 



E^ • = the biomass of forage type i consumed by cattle type j 
(lb/ac) 
F^ = forage biomass (lb/ac) 
C: ■ # cattle of type j/ac 
P i,i = Preference by cattle type j for forage type i 

L = total number of forage types 
h Q • = the # weeks required to digest and process a pound of 
forage type n by cattle type j (the inverse of the 
weekly maximum intake rate) 
a^ j = the rate of effective search by cattle type j on forage 
type i (i.e., the area of ground animals cover in a 
week) 

There are 42 different forage types: 7 forage species over 
3 range sites over 2 water availability types. The preferences 
given in Table 5.5 are estimates made without consultation of 
grazing experts and likely bear no resemblance whatsoever to the 



77 



Table 5.5. 



Preferences by cattle (all types) for forage species, 
range sites and water availability for the alternate 
feeding model developed between workshops. The pre- 
ferences are directly proportional, i.e., forage near 
water is twice as preferred as forage distant from 
water . 



Prefer- 
Forage type ence Range site 



Prefer- Water Prefer' 
ence Availability ence 



Bitterbrush 





.1 


Clay 


1 


Near 


2 


Other shrubs 




.05 


Loamy 


2 


Distant 


1 


Sagebrush 




.05 


Riparian 


3 






Forbs 




.2 











Perennia 1 
grass 
decrea ser s 



Per enn ia 1 
gra ss 
increaser s 

Cheatgrass 



.1 
.2 



real world. These preferences are assumed fixed throughout a 
year. This assumption may be invalid; some participants stated 
at the second workshop that preferences for the plant types 
changed through a season. The rate of effective search was set 
to 126 ac/wk for all cattle types. (This assumes cattle travel 5 
mi per day and are able to begin recognizing different plant 
types at a distance of within 30 ft.) 

The form of equation 5.1 is given in Fig. 5.2. Cattle 
switch between forage types as their relative availabilities 
change . 



78 



CONSUMPTION 
» 


/^ TOTAL 

/S PREFERRED 

// ......... UNPREFERRED 








> 

PREFERRED FORAGE DENSITY 


Fig. 5.2. Form of the cattle feeding response curves devel- 
oped after the first workshop. Cattle are hypo- 
thesized to switch between forage types according 
to both relative preferences and availabilities. 



5.2.2 Forage Consumption 

Under both hypotheses, the consumption of forage is 
simulated by assuming cattle have a size dependent forage intake 
requirement (Fig. 5.3). This function does not take into account 
the observation of participants that food intake is also mediated 
by forage quality. Cattle can process a fixed amount of fiber 
per day, so that intake of high fiber food is lower than low 
fiber food. The simulated calf population feeds independently of 
the mother cows throughout the time on the range. (This is an 
invalid assumption for approximately the first four months of the 
calves' lives; during that time they gain their nourishment 
directly from the mother.) 

Under hypothesis 1, the total daily intake requirement is 
multiplied by the days per week and number of cattle in each type 
and then summed over the three cattle types to give a total 
weekly intake requirement for the herd. The total intake is then 



79 




250 500 750 

ANIMAL WEIGHT (lbs) 



000 



Fig. 5. 3 



Relationship between animal (cattle) weight and 
daily food intake. Food intake is the amount of 
food an animal of a given size can process in a 



day . 

tmtwmmmmm 



modified by the biomass of the most preferred forage type to 
ensure actual forage intake never exceeds the forage biomass 
available. The following is used to compute actual forage 
intake : 



EAT i = F i 



1 - e 



R 



( Equat ion 5.2) 



where 



R = total weekly herd requirements (lb/ac) 
F; ■ biomass of forage type i (lb/ac) 
EAT^ « biomass of forage type i eaten (lb/ac) by all cattle 

R/F; is essentially the instantaneous mortality rate of forage 
type i from cattle feeding and the terms inside the brackets 
represent the weekly mortality rate of forage from cattle feeding 



80 



(Fig. 5.4). The forage eaten (EAT-) is then allocated back to 
daily realized intakes per animal for each cattle type. The 
forage consumed by each cattle group is assumed directly 
proportional to the forage requirements for that cattle group 
relative to the total herd requirement. Therefore if adult cows 
require, say, 63% of the total forage requirements, they receive 
63% of the total forage consumed. 

Residual requirements are calculated for each cattle type by 
subtracting the weekly intake of each forage type from the weekly 
intake requirements , and are used in equation 5.2 on the next 
preferred type as determined from the preference scheme outlined 
in Sect ion 5.2.1. 

Under hypothesis 3, the forage intake requirement from Fig. 
5.1 is used to calculate h in equation 5.1 (e.g., if a cow 
requires 100 lb of forage per wk, it will take her, on average, 
1/10 0, or 0.01 wk to consume 1 lb of forage). E • • (equation 
5.1) is then calculated for each forage type and each cattle 
c yP e > given forage preferences and rates of effective search. 
EAT^ is then calculated using an analogue of equation 5.2: 



Hi 
UJ 

cr 

hi 

q2 



IOO-, 



80- 



60- 



tu o 
z° 40. 



UJ 

o 

< 
cr 
o 



20- 



W 



T 



T 



5,000 10,000 15,000 

FORAGE BIOMASS (lbs/acre) 



100 cows 




20,000 



Fig. 5.4. Relationship between forage biomass available and 

percent utilization of forage. The equation 

assumes no waste, such that all the forage util- 

^ ized is consumed by cattle. 



81 



EAT £ = Fi 




" j-1 



( Equat ion 5.3) 



EAT- is then allocated back to daily realized intakes per 
animal for each. cattle type in the same manner as in hypothesis 
1. Under hypothesis 3, equations 5.1 and 5.3 are used for each 
forage type to calculate forage lost by consumption and realized 
forage intake for each cattle type. 



5.3 Cattle Growth 

Forage biomass consumed by cattle is converted to total 
digestible nutrients (TDN) by assuming the relationship between 
protein content and TDN in Fig. 5.5. This function, which is 
assumed to be the same for all forage types, probably varies 
between different types of forage. 




CRUDE PROTEIN (%) 



Fig. 5.5. Relationship between percent crude protein content 
of food and percent total digestible nutrient con- 
tent. All seven food types were assumed to have 
the same relationship. 



82 



In the first workshop, the total weekly weight gain was 
calculated from the sum of the total weekly TDN consumption (Fig. 
5.6). Note that cattle can lose weight with inadequate forage 
consumption. Also, a smaller animal needs to consume less TDN 
than a large animal to gain the same weight. 



5.4 Fall Sales 

In the fall, all yearlings and 80% of the calves are sold. 
The remaining calves are retained over the winter and become 
adult cows the following spring. 

The sale of adult cows assumes the Saval Ranch has a fixed 
number of adult cattle it can graze on the range which is set by 
allowable animal unit months (AUMS). This number is maintained 
by buying or selling adult cows. Currently the ranch target is 
set to 1500. Therefore, if the number of adult cows plus calves 
retained exceeds the target then the excess cows will be sold. 



5.5 Calving 

Adult cow weight, when cows come off the range on November 
30, is used to determine the percent of cows that will calve the 
following spring (Fig. 5.7). Animals between 800 lb and 950 lb 
have a maximum calving rate of 85%. The declining calving 
success from 800 lb to 600 lb was hypothesized to reflect 
changing conception rates, while the drop in calving success with 
weights greater than 950 lb was included to reflect declines in 
birthing success for heavy cows. 



5.6 Overwintering of Cattle 

The cattle the ranch keeps over winter are assumed to be 
maintained strictly on hay, either grown on the Saval hay 



83 



ANIMAL WEIGHT (lbs) 



15 



z 
< 

X 



UJ 

5 



110 



770 



220 330 440 550 660 or mon 



-15 




T" 
42 



21 
WEEKLY INTAKE (lbs. TDN) 



Fig. 5.6. Relationship between weekly consumption of total 
digestible nutrients and weekly weight gain by 
different weights of cattle. 



84 




t r 

600 800 950 1050 

FALL COW WEIGHT (lbs) 



Fig. 5.7. Relationship between fall cow weight and calving 

success the following spring. Small cows have low 
conception rates; large cows have greater birthing 
problems . 



meadows, or bought. (This is not what actually happens; more 
nutritious food, such as cottonseed cake, is often fed to 
pregnant and poorly-conditioned animals over the winter period.) 
The winter hay required by a cow for maintenance is calculated 
using the intake function described earlier (Fig. 5.3). It was 
assumed that the ranch operators would attempt to bring cattle 
weights over winter to a target weight (currently set to 700 lb 
for adult cows and unsold calves) in the spring when they were 
turned out to the range. It was further assumed that it would 
take 15 lb of hay over and above the maintenance ration (Fig. 
5.3) to gain a lb of weight. The total winter hay required is 
therefore the maintenance ration plus any weight gain ration 
required. The user of the model has the option of buying all, 
part, or none of the deficit in hay supplies if ranch hay 
production is insufficient to meet total winter hay requirements. 
Any excess hay produced on the ranch is considered sold. 



85 



The growth of cattle over the winter period is calculated 
using the function in Fig. 5.8. 






< 

X 
C£ 

tu 



-1.0 




HAY CONSUMPTION (lbs) 



Fig. 5.8. Relationship between average daily hay consumption 
over the winter period and the average daily 
weight gain per cow over the winter period. 



5.7 Economics 

The subgroup used the dynamics of cattle production as a 
basis to simulate the economics of the ranch operation, and then 
used ranch revenues and costs to consider some aspects of the 
effect of the Saval Ranch operation on the economy of Elko 
County . 



5.7.1 Ranch Revenues 

Ranch revenues are assumed to come only from cattle and hay 
sales: 



86 



Revenues = £ ($/lb of cattle type i * lb cattle type sold) 
i-1 



($/t of hay) * (t hay sold) 



( Equat ion 5.4) 



The unit prices of these commodities are given in Table 5.6. 



Table 5.6. 



Unit prices associated with revenues and costs of 
ranch operation. 



It em 



A. Adult cows (buy or sell) 

B. Calves (buy or sell) 

C. Yearlings (buy or sell) 

D . Hay so Id 

E. Hay bought 

F. Growing hay 

G. Improving water availability to cattl 
H. Plowing and seeding 

I . Fert i lizing 

J. Fencing 

K. Chaining 

L. Burning 

M. Spraying and reseeding 



Price 



$0.40/lb 

$0.60/lb 

$0.63/lb 

$60/t 

$46/t 

$30/t 

$5/ac 

$40/ac 

$12/ac 

$2500/mi 

$15/mi 

$2/ac 

$24/ac 



5.7.2 Ranch Costs 



Ranch costs were calculated using the following 



Total fixed /cost of keeping # head 
Costs " costs + 1 head of cattle of cattle 



unit costs for each 
management action 



units of each manage- 
ment action taken 

( Equat ion 5.5) 



The unit costs of management actions are given in Table 5.6. The 
fixed costs were assumed to be $250,000 and the cost of keeping 
one head of cattle was set at $60. 



87 



The net income for the ranch was therefore: 

Net Income = Revenues - Total Costs (Equation 5.6) 

5.7.3 Economic Benefits to Elko County 

Three economic indicators were calculated from the economics 
of the ranch operation. The calculations were considerably 
simplified by the assumption that all commodities sold by the 
ranch are exported out of the country. The indicators are all a 
function of something called "ranch purchases", which is defined 
as those expenditures made by the ranch within Elko County. In 
the mode 1 : 

Ranch Total / Fixed 
Purchases = Variable + \* 05 * Costs/ (Equation 5.7) 

Costs 

The three economic indicators were generated as follows: 

Gross Values of 

County Output = 2 ' 5 * Ranch Purchases (Equation 5.8) 

Ranch Value Added = 0.15 * Ranch Purchases (Equation 5.9) 

County Value Added - 0.4 * Gross Value of 

County Output (Equation 5.10) 

The multipliers were estimated from economic input/output models 
of similar, but not identical, situations in eastern Oregon. 



88 



5.8 Analysis of Range Utilization Effects on Cattle Growth 

The cattle feeding and growth equations are the core to the 
submodel, for they link forage production and range land 
management actions to beef production and range revenues. The 
interaction and consequences of the feeding and growth processes 
can be investigated by looking at "isoweights" for cattle as 
functions of forage availability or protein content (Figs. 5.9, 
5.10). The model was used to calculate, for weights ranging from 
440 lb to 800 lb, the level of either forage or protein required 
to keep cattle at that weight. These "isoweights" were 
calculated over a period of one day; forage dynamics can 
therefore be assumed constant. 

Figure 5.9 shows these isoweights as a function of forage 
availability for different protein contents. As an example of 
how these graphs can be used, consider cattle weight gain on 
forage with a protein content of 9%. Cattle could never grow 
above about 6 80 lb with this diet. As the protein content of the 
food increases, less forage biomass is required to maintain or 
gain up to 770 lb. 

Figure 5.10 is a series of isoweights as a function of 
protein content for a series of forage biomasses. The vegetation 
submodel assumed all forage types have a protein content of 15% 
in the spring; all shrubs decrease to 10% and all other forage 
types decrease to 3% in the fall. A diet of anything less than 
about 5% protein, irrespective of forage availability, is 
disastrous. This implies that without shrubs in the fall, cattle 
would lose weight because the quality of all other forage types 
is too low to sustain any weight. Also, as long as forage 
biomass remains relatively high and protein content remains above 
5%, 70%-85% calving rates can be achieved. Over about 70 lb/ac, 
(and 92-15% protein content), change in forage biomass is 
relatively unimportant in affecting calving rate. It is only 
when forage biomass decreases to low levels that changes in 
quality of the forage strongly influences calving. 



89 



40% 28% 19% 

800™ 



PERCENT 

PROTEIN 



13% 



S 620 




120 160 

FORAGE BIOMASS (lbs/ocr«) 



200 



u 

o 

«-• 

z 



Fig. 5.9. Isoclines of constant cattle weight as a function 
of forage biomass for a series of protein contents. 
(Percent calving success at four cow weights are 

also shown . ) 



FORAGE 

BIOMASS (Ibs/oc) 
46 34 \ 23 




10 20 30 

PERCENT PROTEIN 



Fig. 5.10 



Isoclines of constant cattle weight as a function 
of percent protein for a series of forage biomass 
levels in lbs/acre. (Percent calving success at 
four cow weights are also shown, as in Fig. 5.9. 



90 



These isoweights also point out a potential danger in 
attempting to maximize livestock production by stocking heavily 
on very high quality forage. This would be in the upper left 
region of Fig. 5.9. Any decrease in forage biomass, due to 
drought or some other uncontrollable process would be 
catastrophic; it would take very little decrease in forage 
biomass at 15% protein to drastically reduce beef weight. The 
alternate strategy, production of high biomasses of lower quality 
forage, allows for lower animal weights but slower declines in 
weight with decreasing biomass. This strategy would on average 
produce fewer pounds of beef but would also buffer against 
unpredictable circumstances, such as drought years. The range 
under the first strategy would, in some years, produce very high 
amounts of beef and other years, very low amounts of beef. 



5.9 Review of Submodel Hypotheses 

The preferences given in Table 5.5 are correct. 

The consumption model developed between workshops is 
probably a more realistic representation of cattle feeding than 
the model developed at the first workshop. However, the 
parameters in Table 5.5 were derived independently of any input 
from range scientists. The feeding preferences in particular 
need to be better estimated. Existing literature could perhaps 
be used to estimate some of these preferences. 

Total digestible nutrients (TDN) is a good predictor of 
forage quality . 

Participants at the second workshop stated that acid 
detergent fiber (ADF) was a better index of forage quality. 
However, if crude protein, TDN and ADF can be quantitatively 
related to each other, any of the three can be used in the model. 



91 



Calves feed ind ep end ently of their mothers from the 
beginning of spring. 

This hypothesis is invalid but field research is likely not 
needed. The existing cattle feeding and growth functions could 
be used along with the assumption that calves consume no forage 
and gain a constant weight per week (from milk) until some time 
in the late spring or early summer. Adult cows' weight gain 
could then be adjusted according to food intake and the calf gain 
for which they have to consume food. 

Forage preferences do not change from spring to fall and are 
independent of forage quality . 

There is a confounding effect of preference and availability 
in the determination of what cattle (or other animals for that 
matter) eat. Apparent changes in preference, as measured by 
composition of diet, may in fact be due to changes in 
availability. Given the alternate hypothesis that forage 

preferences do change within the year, according to changing 
forage quality, the existing model could be modified to have the 
forage preferences changing weekly as a function of protein 
content . 

Fall cow weight is the best predictor of calving success . 

Cow condition through the winter period may also be 
important. It is conceivable that abortion rates would depend on 
how well the cows were maintained over the winter. 

Maximum food intake is solely a function of animal weight. 

Maximum fiber intake is probably solely a function of animal 
weight. Participants stated that animals could process fixed 
amounts of fiber per day; the actual quantity of forage consumed 
would depend on its fiber content. 



92 



All forage types exhibit the same relationship between % TDN 
and % crude protein . 

A review of the appropriate literature would likely reveal 
if this hypothesis was invalid. 



93 



6. WILDLIFE SUBMODEL 

The responsibilities assigned to the wildlife subgroup are 
summarized in Fig. 6.L Due to the limited time available at the 
first workshop, the subgroup succeeded in conceptualizing the 
dynamics of only the mule deer and sage grouse populations., At 
the second workshop, there was some discussion on the inclusion 
of jackrabbits, grasshoppers, and ground squirrels. Although 
these discussions were productive they did not leave us at a 
level of refinement suitable for actual inclusion in the model at 
this- t ime . 



6.1 Mule Deer 

Mule deer comprise one of the two important game species on 
the Saval project area. The entire project area lies within an 
area designated as mule deer key summer range in the Independence 
Mountains. Currently it is estimated that 350 deer summer on the 
project area. 

The deer also use the central pastures (Upper and Middle 
Mahala; Upper Sheep) and eastern pastures (Lower Mahala, Lower 
Sheep) as fall and spring range as they migrate to and from their 
wintering range, off the project area. 



6.1.1 Population Structure 

To simulate the mule deer population, two age groups are 
represented; fawns (0-1 yr) and adults (older than 1 yr). Each 
of these is further divided into male and female. Although this 
structure is overly simplistic the participants felt two age 
classes would serve the purposes of the first workshop, noting 
however, a future refinement should be the introduction of a 
yearling class. Survival and reproduction for each of these age 
classes is described in the following sections. 



94 













toO 




















C 














so 






•H 














3 






M 














•r-t 






OJ 














1-1 






4-1 














0) 






e 




o> 










4-1 






CD 




CO 




co 






c 










3 


CO 


■3 






0) 






cu 




O 


r— 1 


U 










0) 


CO 




u 


CO 


•H 






w 




co 


3 




too 


p 


X> 






0> 




3 


O 






£ 




CO 




0) 







u 




0) 


CO 


OJ 


as 




■a 




S-i 


00 




toO 


e 


£ 


O 








ao 






CO 




CO 


H 




0) 






01 




co 


.—i 


toO 


<J 




i— i 




0) 


00 






t— i 


1 


CJ 


u 


3 




bO 


CO 




c 


CO 


C 


1— 1 


<D 


£ 




CO 


co 







s 





Q 


0) 




oi 


CO 




<D 




CO 


3 


z; 


-o 


0> 


qo 




0) 


00 


CO 






i— i 




t— i 


CO 


U-l 


^ 


CO 


>^ 


<4-l 


4-1 




0! 


CO 




o 


CO 




CO 


o 


o 




i— I 


E 


60 




e 


toO 


-3 








3 


0) 


s 


>. 


V 


3 




co 


>% 




e 


u-i 


■ ^ 


4-1 


U-4 


•r-l 


S-i 


co 


J-l 








-a 


•r-l 




"3 


a; 


CO 


• l-» 




U-l 


U-l 


CD 


co 


U-l 


oi 


4J 


£ 


co 







O 


OJ 


3 


o 


0) 


C 





C 








S-i 


0) 




V-i 


3 


• i-i 


CU 




=*= 


^t 


-Q 


•3 


=«= 


X* 


XI 


J3 


TJ 











"3 










0) 










E 










3 










CO 










C 


C/3 











J 








o 


UJ 










a 








a 


o 








3 


S 











ca 








u 


p 








toO 


C/J 








4-1 


tsj 








c 


Ed 


i-i 






CO 


X 


CO 






r— 1 


H 


CU 






a. 


O 


>% 




0) 
CO 


J3 


O 


CO 


u 


3 


o CU 


H 


>-, 


<u 


O 


CO 4-4 




CO 


01 


S-i 


<u -H 


CO 


-3 


■3 


toO 


.— i 


w 








C4-I T3 


hJ 


U 


cu 


<v 


O .-< 


pq 


0) 


p- 1 


too 


•H 


< 


4-1 


3 


CO 


co 2 


^4 


c 


E 


co 


c 


VA 


3 






o >, 


< 


J3 


l 


1 


4-1 Xl 


> 











o 
u 
en 

-Q 



W 



4-4 



:/: 




2 




a 




i— i 


c 


H 


o 


U 


CO 


< 


CO 




CU 




co 




00 




C 




•r* 




4J 




c 




3 




.C 



J 


u 


w 


o 


Q 


2 


O 


< 


s 


X 


PQ 


u 


3 




CO 


ai 




O 


w 


fa 


fe 




1— 1 


CO 


J 


w 


Q 


J 


J 


3 


i— i 


as 


2 









CU 


r" 














a 


O 










w 




>•, 


CO 










J 


CO 


4-1 


CU 


CO 








IJJ 


CU 






CU 








a 


t-i 


3 


>, 


CU 








o 


CO 





X 


u 








s 




■r-l 




4-1 




T3 




23 


3 


4-1 


<U 






3 




3 


CO 


CO 


M 


-3 




CO 




CO 


•H 


4-1 


3 


c 










S-l 


0) 


4-1 


CO 




3 




tti 


CO 


00 


CO 




X 


O 




w 


0. 


CU 


CO 


CO 


CU 


•H 




X 


•rH 


> 


a x 


•3 


4-1 




H 


i-l 






3 


3 


CO 




C 




X 


X 


M 


•r-l 


4-1 






3 


u 


CJ 


,3 




•r-l 




s 


•H 


CO 


CO 


co 


>^ 


CL 




o 




CU 


CU 




4-1 


•H 




as 


0) 






« 


■r4 


o 


0) 


fa 


r-l 


U-l 


3 


co 


r— 1 


CU 


S-i 




4-1 





O 


CU 


CO 


u 


3 


CO 


4J 






co 


3 


a, 


■uJ 


w 


CO 


CO 


u 


CO 


a" 




CO 


J 


CJ 


co 


CU 


CO 




>, 


M 


S3 




CO 


> 


S-i 


M 


^-i 


0) 


< 


U-l 


E 


o 


00 


01 


^: 


G. 


>— i 








CJ 




4-1 


0) 


E 


OS 




•H 




U-l 


CO 


0) 


CU 


< 


=K= 


X 


r>? 


O 


3 


3 


u 


> 

















T3 



0) 
-C 

4-> 



4J 

(D 
C 
cn 

•H 

CO 
03 

CO 
CU 



CO 

c 



a 

CO 
CU 



en 

•rH 
fa 



95 



6.1.2 Foraging 

Conceptualizing the quality, quantity and timing of mule 
deer foraging is an extremely complicated task that has been 
attempted by a number of investigators ( C oop er r id er , pers. 
comm.). Mule deer have very definite preferences for habitat and 
food types that change over the year depending on the condition 
of the range. Although many ungulates appear to selectively 
forage for their preferred plant species, under conditions of low 
availability of the preferred species mule deer have starved to 
death with full rumens (Cooperr ider , pers. comm.). Further, the 
varied phenology of the plants makes the timing of range 
utilization an important determinant in the quality of the forage 
and the deer's nutritional intake. 

For the workshop submodel, a relatively simplistic view was 
taken towards the mule deer's foraging behavior. First, it was 
assumed that the actual biomass of food ingested per animal was 
constant; for fawns 5.6 kg/wk, and for adults 9.1 kg/wk. 

This implies that ingestion is independent of the quality of 
the vegetation available and, given there is adequate vegetation 
on the range, the biomass of that vegetation. This of course 
ignores the concept of searching and handling times which 
eventually must apply to a deer - vegetation interaction. 
However, the participants felt it was reasonable to assume the 
total biomass of available vegetation would always far exceed the 
total demands of the mule deer whatever the species mix on the 
range. This does not however guarantee the deer's nutritional 
requirements will be satisfied since the species making up the 
vegetation, although ingested by the deer, may be nutritionally 
worthless. This is accounted for in the model. 

The biomass of a particular plant group ingested by the mule 
deer each week is calculated as follows: 



total bio ma s s of 
group i ingested 



P • * B • 
lm i 



total 



■P. * B^ demand/ anima 1/wk 



# mu 1 e 
deer 



( Equat ion 6.1) 



96 



where : 



'• = relative preference for plant group i in month m 
B- ■ biomass of plant group i 



The plant preference parameters P • are expressed as 
positive numbers with P£ m ■ indicating a plant group is not 
eaten, P£ m = 1 indicating indifference and P£ m > 1 indicating a 
preference. The difference in plant phenology and other 
differentiating characteristics are reflected in monthly 
variation of the preference parameter for any particular plant 
group (Table 6.1). Note that this representation of ingestion 
will always result in the same total ingested biomass across 
plant groups as long as at least one B- > total demand. In the 
unlikely event that this does not hold, then all the B- is 
ingested . 



Table 6.1. 



Mule deer preference indices for the seven plant 
groups provided by the vegetation submodel (A. 
Cooperrider, pers. comm.). 



Spring 
MAM 



Summer 



Fall 



Bitterbrush 0.1 0.1 
Other Shrubs 0.2 0.2 



Sagebrush 
Forbs 
Gra ss-Dec 
Gra ss-Inc 
Gra s s-Ann 



0.1 
0.2 
0.2 



J JU A 

5.0 5.0 5.0 
10.0 10.0 10.0 



i 
5.0 
2.0 





5 .0 
2.0 



0.2 0.2 

20.0 20.0 20.0 

0.01 10.0 10.0 

0.01 10.0 10.0 

0.01 20.0 20.0 



I 
5.0 
2.0 

0.01 0.01 0.01 0.01 0.2 0.2 
20.0 20.0 20.0 20.0 20.0 20.0 
0.01 0.01 0.01 0.01 0.01 0.01 
0.01 .01 .01 .01 .01 .01 
0.01 0.01 0.01 0.01 0.01 0.01 



Conversion of the ingested biomass into actual utilizable 
energy is accomplished using Table 6.2. As with the preference 
parameter, these conversion figures reflect the monthly variation 
in nutritional value of the plant groups. 



97 



Table 6.2. Energetic Value of each plant group to mule deer by 

month (kcal/g ingested) (A. Cooperrider, pers. comm.) 











Spring 






S u mm e r 






Fall 










M 


A 


M 


J 


JU 


A 


S 





I 


Bit t erbrus 


h 


1 


c2 


1 .45 


1 .6 


1.9 


1 


.9 


1 .9 


1.9 


1 .9 


1.9 


Other Shru 


bs 


1 


.9 


2.3 


2.4 


2.6 


2 


.6 


2.6 


2.4 


2.3 


1 .9 


Sagebrush 




1 


.9 


1.9 


1 .6 


1 .6 


1 


.6 


1 .6 


1 .6 


1 .6 


1 .6 


Forb s 




3 


.3 


3.3 


3.3 


3.1 


3 


.1 


3.1 


3.1 


3.1 


3.1 


Grass-Dec 




1 


.2 


3.1 


2.6 


2.3 


1 


.7 


1 .2 


1 .2 


1 .2 


1 .2 


Grass-Inc 




1 


.2 


3.1 


3.1 


2.6 


2 


.3 


1.7 


1 .2 


1 .2 


1 .2 


Grass-Ann 




1 


.2 


3 .1 


2.6 


1 .2 


1 


.2 


1 .2 


1 .2 


1 .2 


1 .2 



6.1.3 Health Index 

Currently the model assumes that the health status of the 
deer is a cumulative function of the food ingested between March 
and November, the time period during which the deer are on the 
Saval range. The winter range was bounded out of the model and 
currently has no influence on the simulated dynamics. This was 
not done without some concern on the part of the participants and 
was immediately flagged as a questionable assumption that 
requires further attention. 

The mule deer health index is simply the average weekly 
amount of "utilizable energy" ingested divided by a pre-spec if ied 
requirement, (i.e., 18.34 x 10^ kcal/wk for males; 26.92 x 10 3 
kcal/wk for females) where the average is taken over the March to 
November period. An index value greater than 1, implies the deer 
is entering the winter range in excellent health; a value less 



98 



than 1 implies a less than optimum condition which will have 
implications on winter survival and the following years 
reproduction success. 



6.1.4 Overwinter Survival 

Overwinter survival, a discrete event in the model, is 
represented as a function of the previous year's health index for 
each population group (Fig. 6.2). These survival rates are 
applied to the populations in the first week of each simulated 
year (i.e., March 15). 




Adult Males 

Adult Females and Fawns 



I 

0.5 

HEALTH INDEX 



Fig. 6.2. Overwinter survival of the mule deer as a function 
of the previous year's health index. 



6.1.5 Reproduction 

Reproduction, a discrete event in the model that occurs on 
June 1 each simulated year, is represented as a function of the 



99 



previous year's health index for the adult females (Fig. 6.3). 
Given the simple population age structure of this model, this 
would include the current yearlings (i.e., last year's fawns). 
The maximum number of fawns per adult female (i.e., 1.2) is the 
effective number in the fall after mortality due to predation and 
disease. The effective fawn + female to male ratio in the fall 
is set at 1.5:1. 




NOEX 



Fig. 6.3. Mule deer fawns per doe as a function of the 
previous year's health index. 



Note this treatment of reproductive success includes 
predation (primarily from coyotes) as a fixed proportion of the 
actual births. The participants felt this was acceptable for the 
Saval Ranch area at this time but did not want to preclude it as 
an issue that may need further attention. 



6.1.6 Hunter Mortality 

Currently deer hunting in the Saval area is controlled using 
hunting permits issued on a regional basis (i.e., Game Management 



100 



Unit; GMU). Each permit allows the hunter one buck kill per 
year. The current number of permits for the GMU is 500/yr. 
Since the Saval area is a fraction of the GMU (approximately 
25%), only that fraction of the permits are allowed to hunt the 
simulated deer population. The actual kill per effort (where 
effort is measured in hunter days) is represented as a function 
of the density of bucks on the Saval range (Fig. 6.4). Effort is 
the product of the number of hunters (i.e., .25 * # permits) 
which hunt in the Saval area and the average number of days each 
hunter spends hunting (i.e., 4 days/yr). Currently this effort 
is distributed evenly over the 37 weeks that the mule deer are 
actually on the Saval range (i.e., permitted hunters can hunt any 
time in the year). (The current hunting period is about 4 wk 
long, from early October to early November; the model can be 
changed to reflect that.) 







1.0- 






>- 
< 
a 


0.8- 


^^-^^^ 




K 
Ui 

z 

X 

V. 
_J 
-1 


0.6- 
0.4- 

0.2- 

. 


f 




iii i « i • i ■ i ' l 

46 200 400 600 800 1000 






NUMBER OF BUCKS 


Fig. 


6.4. 


Number of bucks killed per hunter effort as a 
function of the number of bucks in the Saval 
population . 



At the second workshop, there was considerable discussion on 
how we might structure and parameterize a more realistic 
representation of hunting mortality. Unfortunately, the outcome 



101 



was inconclusive and remains a refinement to be addressed at 
future meetings. 



6.1 o7 Deer Migration 

Approximately 25% of the mule deer overwinter east of the 
Saval project area; the remainder to the southwest. Since the 
preferred summer range for deer are the higher pastures (i.e., 
East Independence North and South) the deer must pass through the 
lower pastures in the spring and fall. This passage can take as 
long as three months depending on the amount of snow and rate of 
spring melt. In the model this migration is described in three 
distinct phases, over time; 

(1) deer overwintering east of the project area move onto 
the Lower Sheep and Lower Mahala pastures (about March 
15 - May 15) ; 

(2) deer overwintering to the southwest and those on the 
lower pastures move onto the Upper Sheep, Middle Mahala 
and Upper Mahala (about May 15 - June 15); and 

(3) all deer move on to the East Independence North and 
South pastures (about June 15 - November 1). 

Observation has shown that initiation of each phase is 
usually dependent on the amount of snow cover remaining and the 
start of "greenup" . This is simulated by checking each week 
whether the forbs in selected pastures are in a positive growth 
stage. If so then the appropriate deer population is placed on 
the pastures indicated above. 

During their passage across the lower and middle pastures, 
the deer are generally only found in the lower range sites (i.e., 
gullies, streambeds). Therefore in the model the mule deer 
browse only on vegetation on the loamy 8-10 and wet meadows 



102 



during phase 1 and loamy 10-12 and wet meadows during phase 2. 
In phase 3 they browse all range types. 

Currently the simulated movement of the deer back to their 
winter range occurs in the week after November 1. Any grazing 
which occurs during this week is not determined and has no effect 
on the deer's nutritional index. This simplification may require 
refinement if the return migration is thought to be a critical 
period for the mule deer. 



6.2 Sage Grouse 

Sage grouse comprise the other game species of concern on 
the Saval Project. A number of strutting grounds (leks) exist 
within and in close proximity to the project (Fig. 6.5). The 
minimum estimated adult breeding population of 670 sage grouse 
resides within or near the project area. Due to the apparent 
sensitivity of the sage grouse to disturbance of areas within 2 
miles of a strutting ground, there is concern that some of the 
management options (i.e., plowing) could result in a decline in 
the grouse populations in the Saval Ranch area. 



6.2.1 Population Structure 

Generally a distinct breeding population can be associated 
with each lek or group of leks. For the purposes of the model, 
the observed lek populations were aggregated into five sage 
grouse populations (Fig. 6.5). The initial numbers associated 
with each are shown in Table 6.3. 

Within each strutting ground population, three age classes 
are represented: chick, yearling and adult. These are further 
divided into male and female (i.e., cock and hen). 

Unlike the deer model, the sage grouse populations are not 
simulated on a week to week basis. Instead survival and 
reproduction rates are estimated for critical periods over the 



103 




4-> -t-> 
C (O 
0) r— 
</> 3 
O) O. 
S- O 

a. o. 

4) 

a> 

oo O) 

Q.+-> 
•t— (O 

p— a 

CI 



o s- 

C O) 

a: E 

r— S 
> 

o 

CD -i- 
jC -)-> 
4-> fO 

> 

s_ 
cu 
ui 
.a 
o 



c 
o 

in 

T3 

C 
3 
O 

en 

en 

c 



3 
S_ 
4-> 



cu 

4_ r— 

en o 

3 E 
O 



< 


-a c 


o 

en 


0) <U T- 




1 00 •!- 




3 <+- "O 




o •<- <U 




i s- +j +-> 




C7> C C 




a; cu 




aj -a i/i 




cn-r- <D 




(O s_ 




wi t/i O- 




^ a) 




(+- CU S- 




O r— 




!/) 




c o en 




O «r- C 



+j -I- a. 

IT3 U 3 
O O) O 
O Q. i- 
I OO CD 



LO 



C7> 



104 



year and used to modify the populations at the end of the period. 
The populations currently displayed by the model are the number 
of sage grouse in the early spring (i.e., March 15) of each year. 



Table 6.3. 



Number of sage grouse observed at each of the popula- 
tion sites indicated in Fig. 6.5. 



Strutting area 
Lek Number 

1 
2 
3 

4 

5 



Sage Grouse Seen 



Males 

8 

14 

53 

19 

170 



Fema les 

2 
1 
8 
6 
38 



6.2.2 Reproduction 

Both yearling and adult hens are sexually mature. The 
nesting period is usually from April 1 to June 15 each year with 
chick emergence occuring between June 1 and August 1. It is 
assumed that each hen nests during the April 1 to June 15 period 
and that some fraction of these nests successfully produce chicks 
(i.e., nesting success). This fraction is a function of the 
total average (over range sites and time) biomass of forbs plus 
grasses expressed in lb/ac on three range sites (loamy bottom, 
loamy 8-10, and claypan 10-12) between April 1 and June 15 (Fig. 
6.6). Because the hens require some forbs during the nesting 
period the nesting success is reduced to zero if the ratio forbs: 
(forbs + grasses) is less than 0.1. 

The number of chicks produced by each successful nest is a 
function of the total average (over range sites and time) forbs 
plus grasses on four range sites (claypan 10-12, loamy 8-10, 
claypan 12-16, loamy 10-12) in the period between June 1 and 
September 1 (Fig. 6.7). Since forbs are required by the sage 
grouse hens and chicks during this period the number of 



105 



100' 



CO 


80 


to 




III 




o 


6b 


o 




3 
CO 




© 




z 


35 


H 




CO 




UJ 


20 




— i — r 

100 150 200 300 400 

F0RB AND GRASS BiOMASS (Ib/oc) 



Fig. 6.6. Sage grouse nesting success as a function of the 
average biomass of forbs plus grasses. (The 
average is taken over range types and time.) 







5- 








4_ 






CO 
UJ 

z 
v. 

CO 

o 

X 

o 


3- 
2- 

1 








0™ 








( 


Tii r i 

) 100 200 300 400 500 






F0RB AND GRASS BIOMASS (lb/ac) 


Fig . 


6.7. 


Number of chicks per successful sage grouse nest that 
survive to the fall as a function of forb and grass 
biomass. 



106 



chicks/hen is reduced to zero if the ratio f orbs: (forbs + 
grasses) is less than 0.1. Note the number of chicks estimated 
in this calculation is the number that survive to recruit into 
the fall population each year, not the actual number successfully 
emerging from the eggs. 



6.2.3 Survival Rates 

Survival rates for both male and female sage grouse are 
represented within two distinct periods — 'March 15 to September 
30, and September 30 to March 15. This division was felt to be 
adequate to represent the dynamics of the grouse as a function of 
range condition. No attempt was made to directly represent the 
dynamics of predation on the grouse. Predation is implicitly 
represented by expressing survival as a function of the 
vegetation which provides cover from birds of prey and other 
predator s . 



6.2.3.1 Hens 

Survival of the sage grouse hens between March 15 and 
September 30 is structured as a function of the quality and 
quantity of the vegetation during that period. Three plant 
groups are required--f orbs, grasses, and shrubs (for cover). 
From the perspective of the sage grouse these plant groups are 
complementary and are all necessary to ensure optimal grouse 
surviva 1 . 

For the purposes of the model, hen survival is calculated in 
two steps. First, survival under optimal cover conditions is 
calculated as a function of the total average density of forbs 
plus grasses between March 15 and September 30 (Fig. 6.8). Then, 
to reflect deviations from optimal cover, the survival is 
modified as a function of the average percent shrub cover on the 
pasture (Fig. 6.9). Therefore, the effective hen survival rate 



107 



100. 

so- 
so* 

60- 



40- 



< 

> 

> 
as 

40 20. 

z 

UJ 

X 



1 I I I I 

100 200 300 400 500 
FORBS AND GRASS BiOMASS (lb/ac) 



Fig. 6.8. Sage grouse hen survival rate over the March 15 

to September 30 period, as a function of forb and 
grass biomass, under optimal cover conditions. 




T 

20 25 30 

SHRUB 



40 50 
COVER (%) 



Fig. 6.9. Cover factor for sage grouse as a function of 

average percent shrub cover during the March 15 - 
September 30 period. (Optimal cover has a factor 

af 1 J 



108 



between March 15 - September 30 is the optimal survival rate 
times the determined cover factor. 

Over the winter period the hen survival rate is fixed at 0.9 

6.2.3.2 Males 

The most vulnerable period for the male sage grouse is 
during the breeding season, March 15 - May 31. During this 
period the male stays on or near the. lek and exercises little 
caution in avoiding predators and little care in maintaining his 
health. It was felt by the participants that most of the male 
mortality occurs during this period and that this mortality could 
be expressed as a function of percent cover alone. Therefore the 
March 15 to September 30 male survival rate is expressed as a 
function of the percent shrub cover on the pasture containing the 
strutting ground, averaged over the March 15 - May 31 period 
(Fig. 6.10). 

Over the winter period the male survival rate is fixed at 
0.8. 



Fig. 6.10. 



65- 
60- 

50- 



£ 


40 


_i 


3 5 


4 




> 


30 


> 




(T 




3 




</} 


20 


iu 




-i 




< 




2 


10 



5 10 



i — i — r 



— i 1— 

5 20 30 40 
SHRUB COVER (%) 



50 



Male sage grouse survival rate over the March 15 
to September 30 period as a function of average 
percent shrub cover during the critical breeding 
period (March 15 to May 31) . 



109 



6.2.4 Bunting 

Sage grouse are one of the main game species on the Saval 
Project. During the workshop there was insufficient time to 
adequately represent the relationships between the number of 
hunters, the density of grouse, and the actual grouse kill. For 
the time being, hunting mortality is held fixed at 10% of the 
fall population. Since sage grouse hunting is of recreational 
value, this aspect of the model should certainly receive further 
refinement in the future. 



6,3 Review of Submodel Hypotheses 



6.3.1 Mule Deer 

The total biomass of vegetation ingested weekly by each mule 
deer is constant * independent of the density of available forage. 

Under this formulation the actual amount of forage ingested 
is independent of the density of total forage on the pasture. 
However the use of plant preference parameters does ensure that 
the deer will receive more of the higher preference forage if 
available. As a consequence, there is an indirect dependence on 
forage density that will influence the mule deer population 
dynamics. The adequacy of this style of representation should be 
tested . 

Related to this are the actual values of both the preference 
factors and the utilizable energy conversion factors. Currently 
there is a very loose parallel between the relative values of 
these coefficients and digestable protein. Effort should be 
spent refining these numbers. 



110 



The concept of a mule deer health index is an effective 
means of representing the feedback between v eg etation and mule 
deer survival. 

The disadvantage of this approach is that it is sensitive 
only to the average forage conditions over each simulated year. 
Any extreme conditions that could result in short or medium term 
starvation of the deer would most likely be lost. This is 
especially critical from the perspective of fawn survival, 
primarily in spring and early summer. 

Related to this is the need for herbaceous cover during the 
fawning period to ensure protection from the weather and 
predator s . 

The dynamics of the winter habitat is not critical in the 
determination of deer survival and reproduction . 

Currently the model assumes that all the determinants of 
deer survival and reproduction are a function of deer foraging 
success while on the Saval range. Overwinter survival is the 
product of a fixed rate (dependent on sex) and the previous 
year's health index. The question is whether this is a 
reasonable representation, or are the winter population dynamics 
the controlling feature of the system? If in reality the 
overwinter factors are critical, then this submodel is currently 
not representative of the Saval deer population and could be 
mi s leading . 

Ultimately, this is a bounding problem and is common to any 
analysis concerned with a migratory population. From the 
perspective of the Saval Ranch management plan, the problem will 
eventually need to be addressed. However, it is reasonable for 
the modelling effort to claim only sensitivity to the spring- 
summer-fall dynamics (i.e., those dynamics directly affected by 
the management plan) under conditions of fixed winter survival 
(i.e., given winter conditions are held constant, what are the 
effects on the deer population?). 



Ill 



6.3.2 Sage Grouse 

Grouse nesting success and number of chicks per successful 
nest are primarily a function of the abundance (i.e*, relative 
and actual) of forbs and grasses near the nest and in brood use 
areas . 

Currently chick production is a function of only the 
combined density of forbs and grasses during two critical periods 
(April 1 - June 15; June 15 - September 1). Forbs and grasses 
serve as a source of both food and cover. What is ignored is the 
possible effects of shrub cover in both quantity and spatial 
layout especially in the spring. 

This representation assumes no depression in chick survival 
because of heavy snows in the spring. 

The key factor determining the summer survival rate of the 
male grouse is percent shrub {sagebrush) cover during the mating 
season . 

This formulation is in effect a surrogate way of 
representing the effects of predation and low nutritional input 
on the male grouse during the period they are on the strutting 
grounds. Although a more dynamic representation of predation 
(i.e., functional and numerical response) would be more precise, 
the information to build such a model was not available. The 
appropriateness of this structure should be tested. 

The key factors determining the summer survival rate of the 
female grouse are percent shrub {sagebrush) cover and the total 
biomass of forbs and grasses averaged over the summer period. 

As with the male grouse this is a "short" way to represent 
the effects of predation and feeding success on the grouse. 
Since it is looking only at s ea s ona 1 'a v er ag e s it may be lacking 
sensitivity to short-term events. 



112 



Hunting mortality is a fixed proportional removal (i.e,» 
1035) of the fall population . 

It may be an inadequate representation of hunting mortality 
to assume that it is fixed. Since sage grouse are an important 
game species in the Saval area (which is ultimately the major 
reason they are of concern to the study) effort should be put 
into changing this formulation of loss due to hunting, perhaps 
one that evaluates 

(1) the number of grouse killed per unit of effort as a 
function of grouse density; and 

(2) the change in effort as a function of the kill per 
effort in the previous season. 



113 



7 . THE INTEGRATED MODEL 

Sections 2 - 6 of this report have provided a detailed 
description of the development of the Saval Ranch simulation 
model. Once the four submodels had been programmed and debugged 
they were linked together at the first workshop to form a 
complete system model. Refinements of the model were carried out 
prior to and during the second workshop in January 1982. This 
chapter discusses the current model's output. 



7.1 The Question of Model Validation 

The Saval Ranch model, like any model of a complex 
bio logical/phys ical/ economic system, is necessarily incomplete. 
It is a highly simplified representation of a real system, 
undoubtedly containing some conceptual biases and incorrect data. 
However, it does represent a synthesis of the knowledge of a 
diverse group of scientists and managers who work with the system 
on a day to day basis. It is very much their model and the 
assumptions and simplifications applied in building the model 
represents the state of knowledge of the system given the 
constraints of time and the objective to build an integrated 
model capable of exploring the implications of a wide range of 
management options. 

Traditionally the accepted form of model evaluation involves 
extensive validation during which, if the model passes various 
rigorous statistical tests, it is classified as "valid". In AEAM 
this task is approached from the opposite direction; 
invalidation. In truth no bio-physical model is valid since it 
is by necessity simple in comparison to the real world. Further, 
a model is a hypothesis describing how the modellers think the 
world behaves and if one wishes to remain true to the scientific 
method the objective of evaluation should be to disprove the 
hypothesis (i.e., the model). If one fails, this doesn't 
necessarily mean the hypothesis is true, it could mean the right 



114 



test has not yet been developed. In any case the more tests the 
model survives, the greater one's confidence in its predictions. 
But one must be careful not to believe the model, but rather use 
its output to suggest ideas and issues that require action (i.e., 
monitoring, research analysis, etc.). One cannot, a priori, 
identify the limits of predictive power or robustness, no matter 
how much effort goes into testing the model. There is invariably 
a real world process not included in the model that will 
eventually cause a divergence between model and observation. 
However, it is within this world of uncertainty that management 
decisions must be made. Models have proven to be useful in 
helping deal with this uncertainty. 

Evaluating the model in this fashion is partially intended 
to de-emphasize the quantitative nature of the model output and 
concentrate more on the qualitative aspects. We are not so 
interested in the projected numbers over the next 15 years but 
more in qualitatively how the system responds given the various 
perturbations imposed (i.e., management actions, natural 
perturbations, etc.). In other words we want to 

(1) develop an understanding of how robust the system is to 
stress , 

(2) determine if there are management actions that can 
partially or completely mitigate an adverse impact, and 

(3) come to some agreement on the quantity, quality and 
kinds of data we need to both improve our understanding 
of the system and help us monitor its "health". 



115 



7.2 Model Output 



7*2*1 Three Year Scenario 

Figures 7.1 and 7.2 are results of a three-year simulation 
of the full model with a stocking of 1,000 cows. The indicators 
in these figures are graphed weekly over the three years. The 
sharp declines in plant biomasses at the end of each year reflect 
the single large time step over the winter period from November 
30 to March 15 (Section 2.4). 

Figure 7.1a shows the above-ground and below-ground biomass 
and cattle consumption of "other shrubs" on the Darling West 
pasture. There is an increase in above-ground biomass and a 
corresponding decrease in below-ground biomass in the beginning 
of all years. Some time after seed set, below-ground biomass 
begins to increase up to a peak in late fall. Cattle consumption 
is very low and has no effect on growth. 

The dynamics of perennial grass decreasers follow 
essentially the same pattern (Fig. 7.1b). Cattle consumption is 
greater on this plant type, however, reflecting the higher 
preference of cattle for this plant type over "other shrubs" 
(Table 5.4). The greater cattle consumption influences the grass 
dynamics slightly, especially in the fall of the second and 
spring of the third years. 

Cow weights (Fig. 7.2a) remain relatively constant at about 
750 lb. Forage availability is not restricted (Figs. 7.1a,b); 
the inability of cows to grow to more than about 7 50 lb is likely 
due to the weight gain vs. TDN intake (Fig. 5.6). Large animals, 
in the model, may simply not be able to consume sufficient 
amounts of TDN to grow past 750 lb; the calculations of percent 
crude protein (Section 3.2.4) may be improper as well. 

Calf weights (Fig. 7.2a) show a similar pattern among all 
years, with a constant increase through spring and summer and a 
leveling off in late summer and fall at about 3 50 lb. The high 



116 



1200-1 ■«■■ Below Ground 



Above Ground 






<o 

< 

O 
CD 



600- 



Cottle Consumption 




Fig. 7.1a. Weekly above- and below-ground biomass of "other 
shrubs", and cattle consumption, on Darling West 
pasture, clay range site, with ranch stocking 
level of 1000 cows. 



1200. 



.a 



S 600- 



CO 

< 

O 
00 



Below Ground 
Above Ground 



Cattle Consumption 




2 
YEAR 



Fig. 7.1b. 



Weekly above- and below-ground biomass of peren- 
nial grass decreasers, and cattle consumption, 
on Darling West pasture, clay range site, with 
ranch stocking level of 1000 cows. 



117 



800-1 



OOOCdGQG 



Cows 
Calves 



n 

£ 400. 



x 
& 



UJ 

5 



2 
YEAR 



Fig. 7.2a. Weekly cow and calf weights on the Saval Ranch, 
with ranch stocking level of 1000 cows. 



> 

O 

< 
Q. 
< 
O 

<r 

UJ 

< 

Ul 

_J 

< 

< 
> 

< 



00 _| 



50. 



I ■ ■ ■ ■ a 




Fig. 7.2b. Weekly soil water expressed as percent of avail- 
able water capacity for three soil depths on 
Darling West pasture, clay range site, with 
ranch stocking level of 1000 cows. 



118 



fall calf weights in the model are likely due to good forage 
availabil ity . 

Fig. 7.2b shows the soil water content in the three soil 
layers as percentages of the available water capacity. Water 
tends to stay longer through the year with increasing depth and 
the top layer shows the greatest fluctuation in water content 
because of higher evapo t r an s p ir a t ion from a greater number of 
plant types (Sections 3.2.1, 4.1). 



7.2.2 Stocking Scenarios 

Stocking levels used in the model ranged from 
unr ea li s t ica 1 ly low to unr ea 1 i s t ic a 1 ly high; this range was 
included simply to test model extremes. Consequences of 
intermediate levels can be interpolated. Four scenarios were 
run, each with different stocking levels: 



NC 
Scenario 1 
Scenario 2 
Scenario 3 



no cattle 
1,000 cows 
1 , 7 50 cows 
2,500 cows 



Also, hunting was not applied to wildlife and no hay could be 
bought in these scenarios. 

These, and all later scenarios, were simulated for 12 years. 
Indicators were output once a year on an arbitrarily selected 
reference date (on about May 15). Changes in the indicators and 
other variables are still calculated weekly, but simply not 
plotted on the graphs. 

Representative plant types, range sites and pastures were 
chosen to examine the effects of the different stocking levels on 
range condition. Figure 7.3 shows "other shrubs" and perennial 
grass decreasers on clay sites in the Darling West pasture. 
Figure 7.4 shows "other shrubs" and forb biomass on riparian 
sites in the Middle Mahala pasture and Fig. 7.5 shows sagebrush 



119 



1200-1 



o 



< 
O 

S 



60 0- 



,-— No Cattle 
, 2500 Cows 




1750 Cows 
1000 Cows 



»**£ 



-r 

4 



T 
12 



YEAR 



Fig. 7.3a 



Annual "other shrub" biomass dynamics on Darling 
West pasture, clay range site for four cattle 
stocking levels on Saval Ranch. 



1200 



in 

.a 



C 600- 



CO 

< 
2 
O 



No Cattle 
2500 Cows 



... ff . 1750 Cows 
...... 1000 Cows 




YEAR 



Fig. 7.3b 



Annual perennial grass decreaser biomass dynamics 
on Darling West pasture, clay range site for four 
cattle stocking levels on Saval Ranch. 



120 



1200 



u 

a 

u» 600" 

a 



w 

CO 

< 
z 
o 

CO 



No Cattle 
2500 Cows 



1750 Cows 
1000 Cows 




YEAR 



Fig. 7.4a. 



Annual "other shrub" biomass dynamics on Middle 
Mahala pasture, riparian range site, for four 
cattle stocking levels on Saval Ranch. 



200 



o 

o 
\ 
to 

§ 600i 

CO 
CO 

< 
2 
o 

CD 



No Cattle 
2500 Cows 



1750 Cows 
1000 Cows 




Fig. 7.4b. 



Annual forb biomass dynamics on Middle Mahala 
pasture, riparian range site, for four cattle 
stocking levels on Saval Ranch. 



121 



200-1 



SO 

J3 



< 

s 

g 

CD 



600 



— — No Cattle 
— = 2500 Cows 



l ■> • * . . 1750 Cows 
... 1000 Cows 




-r 

4 



12 



YEAR 



Fig. 7.5a. 



Annual sagebrush biomass dynamics on E. Indepen- 
dence South pasture, riparian range site, for 
four cattle stocking levels on Saval Ranch. 



..-— No Cattle 
- 2500 Cows 






en 

< 
2 
O 
03 



600- 



!..*•• 1750 Cows 
,'..... 1000 Cows 




YEAR 



Fig. 7.5b. 



Annual perennial grass decreaser biomass dyna- 
mics on E. Independence South pasture, riparian 
range site, for four cattle stocking levels on 
Saval Ranch. 



122 



and perennial grass decreasers on riparian sites in the E. 
Independence South pasture. In all cases, increased stocking 
causes greater range degradation on pastures when they are occu- 
pied by cattle. However, grass decreasers in Darling West show 
no recovery between grazing periods such that it is highly 
degraded by year 12 under heavy stocking. Forbs in the Middle 
Mahala (Fig. 7.4b), although still heavily overgrazed, fare 
better under high stocking levels than under "moderate" stocking 
levels. This is because forbs are less preferred than perennial 
grass decreasers (Table 5.4) and enjoy a better competitive 
advantage over grass decreasers and cheatgrass for water when the 
last two plant types are heavily grazed. (It may be that this 
model prediction is unrealistic; some workshop participants 
suspect that there would be little or no forb growth under heavy 
stocking levels. The model may need more work in this area.) 

Cattle fare well in spite of the heavy range degradation 
(Fig. 7.6). Cow weights decrease only under very heavy stocking, 
while calving success remains near 75% and calf weight near 300 
pounds under all stocking levels. The ability of calves to 
maintain a good weight even under heavy stocking is because they 
are assumed to search as large an area as cows but to require 
less food intake (Section 5.2); they, in the model, enjoy a 
competitive advantage over cows who, in effect, have to find 
about twice as much food in the same area of range. 

The ability of cattle to fare well even under heavy stocking 
and high range degradation may be due to the assumption in the 
model that all above-ground biomass is utilizable by cattle, 
which is an assumption that should be modified in later versions 
of the model. Cattle weights and calving success would 
undoubtedly show greater response to different stocking levels 
under different forage utility assumptions. 

Deer show exponential growth in the absence of hunting (Fig. 
7.7a). The absence of any density-dependent processes on deer 
populations means the model is unstable, with deer numbers either 
going to infinity or to 0. Final deer numbers at the end of 12 
years decrease with increasing cattle stocking. This occurs 



123 







UODBDSliai fcOOU CDQOOO a «^ £ f\ 






800.. 


cows ■•"" l00 ° 

. • ,,««*"•»! Ktviwtm tiw^^^j i i mstiw^iij : liw 




m 

«*s*" 








X 
(9 


400 = 


CALVES 




UJ 

5 


0- 


■wiw*m+»«ri .!wm*»M*W"¥i«A*w*+m*Tmw^ 






1 4 8 12 






YEAR 


Fig. 


7.6a. 


Cow and calf weights under three different stock- 
ing levels on Saval Ranch. 







,.,*.-.. 1750 Cows 








lea- 








2 












w 

CO 

UJ 

o 
o 

(0 


se- 








z 

E 

< 


es 








4 8 12 






YEAR 




Fig . 


7.6b. 


Calving success under three different stock 
levels on Saval Ranch. 


ing 



124 



I500H 



IT 
UJ 

o 



750- 



No Cattle 
2500 Cows 
1750 Cows 
1000 Cows 




YEAR 



Fig. 7.7a. Total mule deer numbers under four cattle stock- 
ing levels on Saval Ranch. 



No Cattle 
2500 Cows 
1750 Cows 
1000 Cows 




Fig. 7.7b 



Total sage grouse numbers under four cattle 
stocking levels on Saval Ranch. 



125 



because decreasing forage availability to deer that arises from 
increased cattle consumption means a lower health index for deer 
(Section 6.1.3) and therefore decreased winter survival (Section 
6.1.4) and lowered reproductive success (Section 6.1.5). 

Sage grouse, under conditions of no cattle, exhibit two 
peaks and declines over 12 years (Fig. 7.7b). The declines 
probably occur because of a combination of lowered reproductive 
success through changing forb:forb + grass ratios (Section 6.2.2) 
and lowered hen survival rates from changing cover (Section 
6.2.3.1). With cattle present, the second peak is delayed by 2 - 
3 years. This occurs because cattle essentially maintain a 
higher forb: forb + grass ratio through their higher feeding 
preference for grass decreasers over forbs. 



7.2.3 Feeding Preference Scenarios 

Two scenarios were run with cattle feeding preferences 
different from those in Table 5.4: 



Scenario 4 - 



feed only on "near water" sites and on 
riparian range sites; plant type preferences 
are as in Table 5.4; 



Scenario 5 - 



cattle do not distinguish between different 
plant types, range sites, or water nearness 
(i.e., all preferences in Table 5.4 set to 
1). 



A stocking level of 2,500 cows was simulated, and again, 
wildlife hunting was not simulated and no hay was bought. It is 
realized that 2,500 cows is an unr e a 1 i s t i c a 1 ly high stocking 
level, but experience has shown that modeling extremes of the 
expected is a useful process for understanding functional 
re lat ionships . 



126 



In Run 4, the Darling West indicators are the same as in the 
no cattle run discussed previously (Fig. 7.4); cattle do not feed 
there because it is not a riparian range site. All other forage 
indictors do show response to cattle grazing but in no case does 
forage biomass show any marked difference to the previous run 
with 2,500 cows (Figs. 7.8 - 7.10). This absence of range 
degradation in either scenario is because: 

(1) cow and calf weights under the highly restricted cattle 
feeding scenario (Run 4) show declines to unrealistic- 
ally low levels (Fig. 7.11), meaning lowered consump- 
tion rates and therefore lower stress on the range; and 

(2) under the unrestricted feeding scenario, feeding 
pressure is distributed over the entire pasture instead 
of being concentrated on particular range sites with 
the result that all areas of the pastures are fed upon 
but all are fed upon very little. 

Deer have a slightly higher final density in the highly 
restricted cattle feeding scenario (Fig. 7.12a); this is likely 
because most areas of the ranch appear to fare slightly better in 
this scenario . 

Sage grouse (Fig. 7.12b), in the highly restricted feeding 
scenario, show a trend very similar to that of the previous no- 
cattle scenario (Figure 7.7b). This is because clay and loam 
range sites in this scenario are not utilized by cattle. Sage 
grouse hens use only these two range sites for reproduction 
(Section 6.2.2); there is, in effect, no interaction between 
cattle and sage grouse in this scenario. Sage grouse numbers in 
the unrestricted cattle preference scenario (Run 5) have a trend 
similar to previous cattle stocking scenarios, reflecting the 
interaction between changing forb:forb + grass ratios by cattle 
consumption and sage grouse reproductive success. 



127 



1200 



600 



< 
O 

£ 



Scenario 4 



Scenario 5 



./•' 



T 
4 



12 



YEAR 



Fig. 7.8a 



Annual "other shrub" biomass dynamics on Darling 
West pasture, clay range site under two cattle 
food preference scenarios. (See text this sec- 
tion for description of preference scenarios.) 



200-1 






- 600. 

CO 

< 
2 
o 



»— ->- Scenario 4 



Scenario 5 



YEAR 



Fig. 7.8b 



Annual perennial grass increaser biomass dyna- 
mics on Darling West pasture, clay range site 
under two cattle food preference scenarios. 
(See text this section.) 



128 



200 H — — Scenario 4 



2 600J 



co 
co 

< 

o 

CD 



Scenario 5 



\ 



12 



YEAR 



Fig. 7.9a. Annual "other shrub" biomass dynamics on Middle 
Mahala pasture, riparian range site, under two 
cattle food preference scenarios. (See text for 

scenar ios . ) 



BIOMASS (Ibs/ac) 

0) ro 

o o 

o o o 


— — Scenario 4 Scenario 5 

\ 
i 
t 

V 

r. 

V- A. A. z. 

V / \, 3 V. / v. 

><j X/ vyf ^ 




i i - r 

4 8 12 
YEAR 


Fig. 7.9b. Annual forb biomass dynamics on Middle Mahala 
pasture, riparian range site, under two cattle 
food preference scenarios. (See text for 
scenar ios . ) 



129 



I200«i— — Scenario 4 






CO 
CO 

< 
o 



600« 



Scenario 5 



•% 
\ 



YEAR 



1 



Fig. 7.10a. 



Annual sagebrush biomass dynamics on E. Indepen- 
dence South pasture, riparian range site, under 
two cattle food preference scenarios. (See text 
for scenar ios . ) 



1200. 






CO 
CO 

< 

2 
o 

CD 



600- 



— — Scenario 4 



\ 

x 

\ 

A 

A 
1 

\ 
1 

4 

\ 



Scenario 5 



\ 



■i" 

/ 

t 









I 

\ 

\ 



r 






\ 
% 



i 

I. 
#• 

// 

': 
#. 
#• 
v 

*- 

/ 



T 
12 



YEAR 



Fig. 7.10b. 



Annual perennial grass increaser biomass dyna- 
mics on E. Independence South pasture, riparian 
range site, under two cattle food preference 
scenarios. (See text for scenarios.) 



130 



....... Scenario 4 



Scenario 5 



800- 




M 

.a 
^400- 


i 
\ 
\ 
\ 
\ 
\ 
\ 
\ 


- 


• ' ' \ 

Calves^ '***...-'•"' v *»».^ "*"* 







Cows 



Calves 



Cows 



YEAR 



12 



Fig. 7.11a. 



Annual cow and calf weights under two cattle 
food preference scenarios. (See text this 
section for a description of preference 
scenar ios . ) 







■"■'■■■'■ Scenario 4 Scenario 5 






10 0. 






3 




50. 










Z 

> 

< 



0. 








n « 1 1 

1 4 8 12 






YEAR 


Fig. 


7 .lib. 


Annual calving success under two cattle food 
preference scenarios. (See text this section 






for a description of preference scenarios.': 



131 



.*.».—. Scenario 4 



Scenario 5 



1500" 



750- 



Ml 
UJ 

a 



<f 



S 



«*' 



***** 



><**' 



°4- 



4 S 

YEAR 



Fig. 7.12a. Total mule deer numbers under two cattle food 
preference scenarios. (See text this section 
for a description of preference scenarios.) 



J 



500 



-I, 



Ui 
IT, 

3 
C 

oc 

(5 

UJ 
O 

< 



*fc 



250- 



F i g . 1.12b. 



.— Scenario 4 
•••■ Scenario 5 



A 






i - 

i: 
I- 

i: 



>*. 



«..***■ 



# 
*• 



v 



ft 



YEAR 



Total saga grouse numbers under two cattle food 
preference scenarios. (Gee text this section 

for a description of preference scenarios.) 



132 



8. OUTSIDE THE MODEL 

Changes in opinions of Saval program participants about 
research needs and changes in program direction have already come 
about as a result of interactions of participants during work- 
shops. Most of the important changes in thinking and the 
immediate operational changes that have been evident were not 
based on outputs of the model (which are still tenuous at best), 
but were prompted by discussions among and within groups of 
individuals at the workshops. The focus and conclusions of many 
of these discussions, however, came about because of the require- 
ment to bring selected people together and to build a model. 
Examples of changes in attitudes and direction that seem to have 
evolved independently of model output are noted below. 

( 1 ) M ajor c aange s in the p lans for hydro logical research 
appear t o have taken place . These changes seem to have 
resulted from the apparent differences between the 
information needs of the vegetation researchers for 
hycrolcgical data, and the data being generated by the 
hydro legists . 

( 2 ) Changes in the m ethods , units o f m easure , a nd perhaps 
spat i & 1 resol u tion used for veg etat ion p roduc t ion. 
lives tock con su m p 'c ion and/or hy d ro logy research see m 
.Imminent - These expected changes are a consequence of 
the demonstrated need for the d a c a of tbe research 
projects co be coiapatible in space, in time of collec- 
tion, and./ or in the way field measurements are made, 

( 3 ) There has been in some cases a_n ad juBtneat of exist inK 
q-q i n ion ab out whether cattle and wildlife . ire co mpa t i- 
b le . At project outset, consensus seemed to held that 
management for cattle and management for wildlife were 
generally incompatible, and that determining compromise 
management options was the prime purpose of the Saval 



.133 



project. But conclusions reached during the workshops 
strongly suggested that selected range management 
options could benefit cattle, mule deer, sage grouse, 
and perhaps other wildlife species simultaneously. As 
an example, increases in habitat int e r sp er s ion and 
herbaceous plant production caused by judicious sage- 
brush control might increase cattle production, song- 
bird diversity, and grouse and deer production. 

(4) The importance of understand ing populat ion regulat ing 
factors for wildlife species became obvious as a. conse- 
quence o f bu i Id ing the m ode 1 . As a result of this 
realization it is anticipated that future research will 
change to focus more sharply on population regulating 
mechanisms, and on how the Saval grazing program 
affects them. 

( 5 ) Bu i Id ing t he m od e 1 ha s m ad e i t clear that econo m ics 
ana lysis, a s it current l y exists, canno t a c c o m o date 
va lues m easured in un its other than dollars . To 
develop an analysis that can comment on values in terms 
other than dollars (as is needed in the case of many of 
the wildlife species, for example), the economists need 
dollar equivalents of values that are currently per- 
ceived in other ways. 

( 6 ) During the course o f building f. he m odel, it beca me 
clear that environ m enta 1 changes resulting fro m the 
im p le m entat ion o f the S ava 1 M anage m ent Plan m ight b e 
readily over shadowed by the expect ed annua 1 or seasona 1 
variability caused b y w eather and other factors; t h i s 
will make it extremely difficult to separa t a the conse- 
quences o f the S a v a 1 M anage m ent P 1 an fro m "nor m al" 
chang e . This points out a major weakness of the pro- 
gram: the lack of control data. Because the program is 
new, good controls in time are lacking. Moreover, 



134 



nearby off-site areas potentially available as controls 
in space probably differ sufficiently from Saval tbat 
the comparability of data from the two would be 
limited. Innovative methods will be needed to provide 
useful experimental control, 

( 7 ) The need for a. m ore sophisticated data- m anage m ent 
sche me than currently exists ha s beco me obvious . 
Building the model has emphasized that the relatively 
informal methods of data storage and management tbat 
currently exist will hinder effective interdisciplinary 
transfer and use of the data. 

( 8 ) The need £ o_ more clearly define "quality" jja some o f 
the indicators has become evident . For example, it was 
agreed that songbirds and rodents were important 
wildlife groups, and that impacts on them caused by the 
Saval grazing plan should be measured. But a 
definition of quality in bird and rodent populations 
chat would enable decision-makers to decide if the 
observed impacts were "good" or "bad" needs to be 
finalized so research can be appropriately focused. 



135 



9. FUTURE DIRECTIONS 



9.1 Model Improvements 

A simulation model should be simple enough that the output 
can be logically deduced as the consequence of basic hypotheses. 
The existing model cannot easily perform this task, because of 
the expense and complexity of performing multiple simulation runs 
(the usual method of tracking down the causes of events). 

Some suggestions for submodel improvements and simplifica- 
tion are presented within the submodel reports and will not be 
repeated here. However the major recommendation that does 
warrant more discussion is a longer time step. Both the vegeta- 
tion and hydrology submodels could be adapted to function ade- 
quately on a longer time step. Although some detail would be 
lost, we feel the overall trends of the indicators would be very 
similar and certainly adequate to evaluate the effectiveness of 
the Coordinated Management Plan. One suggestion is that the 
model be structured to operate as an event driven model rather 
than a fixed time step. The events would be the movement of the 
cattle from pasture to pasture and the timing between these 
events could be determined or set internally in the model as a 
function of one or more state variables (for example when the 
cattle have grazed 50% of the available forage they move to the 
next pasture). Such a change with the "full" model is probably 
not warranted at this time. However it does present a reasonable 
alternative that may facilitate easier development of a reduced 
version of the model for operation on a micro-computer system. 

There is great potential value in simplifying the model down 
to a mic r o - com pu t er scale, particularly for educational and 
evaluative objectives. For a relatively small investment, the 
Elko offices or other group could make use of the model, as well 
as the statistical, plotting and other packages available for 
mini-computers. This is especially relevant to the ultimate 
objective of the Saval project, namely to provide a transferable 



136 



product (i.e., the model) that can be used in other locations. 
The low cost and ease of operation of micro-computers makes it 
feasible to transfer not only the modeling insights, but also the 
model itself. 

Once a micro-computer model is behaving credibly it should 
be possible for ranchers themselves to invest in a micro-computer 
and actually start using the model first hand. This could 
ultimately meet many of the educational and communications 
objectives of the Saval Project. 

Some of the submodels are presently incomplete in terms of 
scope of Saval Ranch research, because there was not enough time 
in the workshops to include all items of interest. This is 
particularly true of the wildlife submodel, which now includes 
detail on only mule deer and sage grouse. In addition to these 
there is significant concern about (and on-going Saval research 
to investigate) songbirds, small mammals (rodents, rabbits), 
fish, and others . 

Water quality (in terms of sediment delivered to streams) in 
the model is currently a crude index to fish habitat quality, but 
substantial refinements need to be made. It was suggested in the 
second workshop that macrobenthos species composition, abundance, 
and/or diversity would be better measures of habitat quality for 
fish, and easier to document in the field than sediment delivery. 
Tbis needs to be evaluated further. 

Environmental variables that regulate populations of 
g ra s s bopp er s , passerine birds, jackrabbits, Belding's ground 
squirrels, and rodent species diversity were discussed in the 
second workshop, but there was not time to enter the 
animal/habitat relationships into the model. Crude 

reprejentatxoc^ of important functional relationships between 
some of these (grasshoppers, jackrabbits, ground squirrels) and 
their habitats were developed (Fig. 9.1). These need to be 
refined and entered into the model, and relationships for song- 
birds end other rodents need to be developed. 



137 





Food Quality, Quantity 



Herbaceous Cover 



JACKRABBITS 



Vegetation Complexity (Diversity) 

GRASSHOPPERS 





Food Quality, Quantity 



Herbaceous Cover 

belding's ground squirrels 



Fig. 9.1. Suspected functional relationships between habitat variables 
and jackrabbits, grasshoppers, and Belding's ground squirrels. 
(These nave not been quantified or included in the Saval model.) 



138 



9.2 Data Management 

Much information is being collected in the Saval Project. 
The total data "bank" can be thought of as existing in three 
dimensions: subject area, space, and time. That is, information 
is being generated about different subjects (e.g., range science, 
wildlife biology, economics, etc.), at different spatial levels 
of resolution (from the whole ranch down to plant quadrats), and 
at different time intervals (daily, monthly, annually, etc.) 
(Fig. 9.2). Furthermore, the data needs to be comparable within 
and among the subject disciplines (Fig. 9.3). 

There are several advantages to a computerized, cross- 
referenced system of data storage (a "data base") for the Saval 
Project : 

(1) all data can be accessed from one location; 

(2) analysis of data across subject areas, between spatial 
areas, or through time is greatly facilitated; 

(3) tabulation or graphical display of results can be 
performed quickly. 

Data base systems consist of at least two components (Martin 
19 7 7): 

(i) a collection of interrelated data stored together with 
"controlled redundancy" (e.g., all data might be 
labelled with a date oc collection, pasture number 
and /or range site number) to serve multiple 
applications; and 

(2) a language and set of programs for adding new data and 
modifying and retrieving existing data. 



139 











• 










£ 










O 










c 










CO 










OS 










H 










03 














> 

03 

CO 




WSN^Sv 










\ 

\ 










J840M 




sz 
u 

c 


CD 

4-1 

o 

0) 

rH 
rH 


\ 


\\\\\ 








\ 


WW 




\ 


\ \\\ 














\ 


\ 


\ 


















< 




u 


\ 


\ 




















u 




\ 


> 




















oc 


en 




















a^HPliM 


< 


c 


in \ 




V 


\ 














» 


■rH 

CD 


O 1 


\ 














S0IU1OUC0S 

/aiuoo 


OD 

3 


03 

H 

03 

^3 

U-4 


w c \ 












2 - > 
















l/l 





o \ 












s;uo|d 




w 


7 s 
















c 













II 

day 
ranch 

pasture 
range site 
mpling site 


*- 

a 
i_ 

■o 
a 

3 

cr 






•H 

CO 

c 
e 

CD 


a 








i_' 


(0 










l/l L< U Hi 








CM 
■H 










fa 



140 









































rH 






C 








>N 1 
J-l O 


co 


u 






















QJ fO 






O w 








•H H 


09 


U 3 






















4J -H 






■H 3 








r-1 CJ 


j-i 


3 O 






















03 4J 




02 


u 








•r4 CO 


CO 


J-l -r-l 






















U 03 




Ed 


05 1— 1 








-O E 


U 


CO J-l 






















■H di 




H 


4J CO MH 








CO 


-Q 


•r-l CO 






















TD CO 




< 


•h c 








J-l CJ 


QJ 


O J-l 1-1 
























2 


ecip 
atio 
ream 








nk s 
nthi 


j-i 
M 

OJ 

> 


il m 
nten 
filt 






















C 
■H JC 

4-J 
CO 
5 jQ 






U JJ JJ 








CO QJ 


3 


O O 3 


























ix co n 








-Q XI 


•r-l 


CO CJ -H 
























A 


• • 








• • 




• * 





























































1-1 CJ" 








































u C 

< 




i 






1 


























w 


M 


U 








3 






















£ 




a 1 


1) 










O 








00 










1 




fO 






co 


01 










•H 








3 




1 






00 




« 






= 


T. 










J-i 
CO 








•H 
jj 




CO 


f—l 




3 
O 




4J -— I 

U (T3 






1) 


u3 










N 








CO 




•H 


rM 




CO 






Ed 


> 












•H 








QJ 




r^ 


CO 










Eh 


»H 


CD 






1) 




•—1 OJ 


QJ 






3 




•rH 


E 




co 




■n -u 




!— 1 


U 


CO 






CO 




•^ CO 


CO 










J-l 


co 




— 




c c 




J 


co O. 


3 






3 




JJ 3 


3 






QJ 




3 






=$fc 




u OJ 




a 


>, E 


C 






O 




3 O 


O 






co 






U-l 




"^ 




a. co 




j 


CO 2 


U 






5-. 




U 


U 






3 




r-l 


O 




>» 






>-i 


T3 CO 


00 






00 




1) oO 


00 











1) 






J-l 






2 


1 C 












U-l 








l-i 




QJ 


CO 


CO 


•r-l 




r— 1 CD 






U O 


0) 






OJ l-i 




•r-4 U 1-1 


QJ S-i 






00 




■O 


CO 


1— 1 


CO 




ro 






0) O 


00 






00 d 




r-H 00 QJ 


00 QJ 








co 




CO 


ca 


U CO 




> CO 






JJ 1 CO 


CO 






CO OJ 




•a CO 1) 


CO QJ 






OJ 


0) 


0) 


3 E 


5 


CJ -0 




03 -H 

CO 






C C >-, 


CO 






co -a 




r-l CO T3 


CO -0 






00 


jj 


r-i 


O O 


5 


> u 








3 co 












>H 








CO 


• ij 


3 


•r-l -H 


5 


• t-l •-< 








j3 C T3 


=*£. 




"; 




3 =*! =8= 


= rt: = rt= 






co 


co 


c 


JJ J3 


E 


T3 X3 




>1 




4 


• • 


• 




1* * 




• • • 


• • 






• 




• ' 


• 




• 




a.' 4J 










| 


























-C —I 




_ 








1 


























4-J rH 

•H 

M -Q 


T 


1 

!j 




c 



, 
























w 


en 


CO 

a> co £ 




•rJ 

J-l 


3 3 

•H J-l 




3 
O 






















fO 

M-4 Qj 


z; 


O 


3 a; 




CJ 


1 CO co 




•H 






















03 
T3 




t— 1 


C 3 JJ 




3 


1 00 -H 




j-i 






















H 


S 


a) i-t a 




"3 


1 U-l 




CO 
























O 


> C8 01 







4—1 ( 4—* ■ — ^ 




N 






















CD 


rJ 


§ 


qj > u 

U ■* 




U 

a. 


00! 00 J-l 


^, 


r-J 






















U CT 1 
CD C 


a* 



Ed 


jj -a 


co 

11 


^ 


• h|t-i a; 

QJ! OJ -H 


co 
3 


JJ 






















_C -h 

4J U 


H 


^--, 


O Jji -r-1 


3 "3 


CJ 


3l 3 T3 


O 


3 
























Ed 


i 3 rJ ^ 


r-J r— 1 







CJ 
























03 c 





c— 


! CO co U 
1 M E O 


CO O 

> co 


j-i 

CO 


QJ QJ 0J 

1 1 I I 1 T— 1 


-a 


5 






















CT- 0) 

'_4 


en 


f-l 


1 CL' 




qj 


J-l J-l JJ 


D 


JJ 






















03 QJ 




<: 


\U G u 


J-l ^ 


> 


j-i ! j-i j-i 


J-i 


jj 






















M 


O 


QJ O -H 


OJ CO 


•H 


co! co CO 


CO 


CO 






















4J U-l 






C C T? 


zl n 


— 1 


0| cj 


i-l 


y 




















03 QJ 


a 


^ 


































T3 i-J 
1 
cw CO CO 






















i 
















f 


s 









3 







3 




i 
| 




3 










CO QJ 

c 






•r-t OJ 


u 




i-H 11 


l-i 


QJ l-i 


O O 


1) 


i 




O 


QJ 




CO 


QJ 1-1 ■ -1 






J-l 00 


OJ 




1 J-i to 


OJ 


00 QJ 3 


3 >H 









•H 


O 




QJ 




s^ U <— l 




O 


O CO 3 


> 




! CJ CO 


c > 


C03 > O 


jj jj 


3 


i QJ 




jj 


3 




•rH 




r i 






3 >-i 







1 3 ii 





S-iO O -r-l *s 


' -r-l 


QJ 


1 > 




■H 


QJ --» 




CJ 








T3 'H 


a 




!T3 


•H O 


OH CJ JJ CO 


co 


l-i 


i-r-l 




CO 


U QJ 




QJ 




4- 73 -H 




< 


O "J-l J-l 






' O <*■• 


J-l 


"J-UJ -r-l CO 


rJ 


U 


JJ 







rJ O 


0/ 


a 




03 C U 




H 


-- a 


• 




i ^ 


O • 


CJ • CO CO 


qj ^^ a. 


3 


1 co 




Cu 


3 3 


CJ 


CO 




C 03 lO 




Cd 


C_ r-l 3 


00 




' C — 


3 00 


r-13 00 O E 


> co £ 


O 


: jj 


M 


E 


CJ U 


g 






*i—! 




U 


CO "3 


QJ 




! cc 


"3 GJ 


do <u a 


O QJ O 


O 


; 


CJ 


O 


cj co 


V 


U-l 




i— 1 QJ T3 




w 


> J-l 


> 




I >• J-i 


O > 


JJO > E •« 


cj a. 





■ 60 


j> 





O U 


03 









>l 


CO O J-i 






co 


l-i 


Or-I C rO 


>-. 




' CJ 


O 




l-i 


rQ 






0! CT 1 




1 


_C JJ Cj s^S 




3 J- 


a. 5-5 


JJQ. o-5 CJ ^-s 


B^S JJ S>? 


J-? 


^> 


CJ 


^? 


3^ a 


CO 


:**- 




03 T3 




1 


» • 


e 




• * 


• 


• • • 


• • 


• 


1 




• 


• 




A 




— ^ C 

- C (0 




















1 
















































U 1— 1 CO 




1 


CO 
































(13 rH 






QJ 
















"0 


















lj 4J QJ 




J 


(-1 
















3 










CO- 








QJ 03 > 




Ed 


3 X! 

J-l CJ 








0) 






QJ QJ 


3 



— 


JJ 


1 

c 




CO 
S-i 








■H jC OJ 




Ed 


co 3 








3 O 




QJ 


4J U 


u 


•1-1 


CO 


ra 




JJ 








~ 4-J r-J 




~3 


co "0 








J-l -0 




00 


■H 3 


00 


1-1 


1-1 


r4 
















1 Q~ ^1 








CO CO 




3 


co jj 




00 


T3 


JJ 




r— 1 












Jl 








CO OJ 




CO 


CO 


00 


CO 






CO 












< j (U 0) 








a. E 




l-i 


00 Tl 


3 


00 


3 


CJ 




3 








. 




M 


1 <-~ < — ' 














c 


•rA 


a 


a* 


'4J 




-O 








ro 




H 


a 








u >■, 




QJ 


•H £ 


u 


•H 




•rJ 




•r-l 












< 


-H j3 








1— 1 c-J 




r-" 


r- Jxi 


JJ 


a 


jj 


f—l 




5. 








CT 




0-, 


*J 3 








00 JT 




00 QJ 


Q. I—' CJ 


3 


Q. 


3 


-J 


jj 


■H 












en 


■ — 1 








G 




3 JJ 


£ -rH O 


1-1 


CO CO 


r-t 





T3 






, 






3 G 








•H i-l 




•r- -H 


CO 1— < 


<J 


UH 


■ r- 


OJ 


Fj 




1 ^"^ 






E JJ 








•0 O 




co co 


co en X. 


•.1 


u a. 


■s 


CO 


• J 










1 






If 




• 


* • 


* 


• '• 


»• 









Uh 



141 



Data should be stored so that they are independent of 
application programs which use the data (i.e., statistical 
packages, programs for graphing or spatial display, etc.). The 
storage scheme must represent the associations inherent in the 
data (i.e., not force awkward restructuring), and must be 
flexible to allow new data types, cross-references and 
applications to be added in the future. In addition, the basic 
structure of the data base should be easily understood by users 
with no training in programming. 

There are many data base systems currently available, which 
fulfill the above objectives to varying degrees and have widely 
varying computer hardware requirements. If the Saval Project 
decides to seriously examine the potential benefits and costs of 
alternative data base systems, detailed consultations with 
application and systems programmers are essential. Much time 
will be saved in the future if it is decided now what cataloging 
information needs to be recorded by all investigators. For 
example, data sheets used in the field or laboratory could be 
formatted for direct keypunching, saving both time and copying 
errors . 

It is probably not worthwhile for the Saval Project to 
pursue the development of cartographic data bases which, although 
allowing automated production of maps, require major investments 
of time and money in digitizing, storage, error correction and 
programming (Harvard Library of Computer Graphics 1979). 

Finally, it should be stressed that data bases are 
constructed incrementally. A detailed inventory of data being 
collected either new or in the future (specifying the frequency 
of collection, spatial resolution, variables being measured, and 
units of measurement) should be completed as soon as possible. 
(Terry Dai ley has already accomplished much of this.) The next 
step should be the compilation of a list cf all intended future 
data analyses (across subject areas, spatial areas and time 
periods), and the assignment of priorities to this list. 
Construction of a small "prototype" data base on a subset of the 
whole project's data would permit examination of the potential 



142 



benefits and costs of a larger system, without major investments 
of time or money. The prototype should be designed in such a way 
that pieces could be added to it incrementally. 



9.3 Research Design 

Emphasis in the workshops and in this report has been on 
building a simulation model, but the prime objective of research 
done on the Saval Ranch is not to build a model. It is to help 
evaluate the consequences of ranch management activities 
(actions), What the model-building does is help Saval Project 
scientists perceive what kinds of information are needed to best 
understand the consequences of activities. It is then up to the 
scientists to design and conduct research to best supply this 
inf ormat ion . 



9.3.1 General Approaches 

There are two kinds of approach for trying to measure the 
consequences of an operation such as the Saval Ranch management 
scheme. One, which we will call the 'monitoring 1 approach, is to 
select the components of concern (indicators) and try to measure 
how each changes over time as the range management plan is 
carried out. In theory this approach documents the cumulative 
effects of range management actions. The other approach, which 
ve will call 'hypothesis-testing', is to focus on how each 
indicator responds to a specific action at a specific time and 
place, and to eventually try to sum the results of each action to 
evaluate their cumulative effect. It is imperative that the 
weaknesses and benefits of each approach be evaluated, and the 
future research for the Saval planned accordingly. 

The 'monitoring' approach has two major problems. First, it 
is usually impossible to validate whether observed changes in the 
Indicators were caused by the actions or by unrelated phenomena, 



143 



because the variables that might affect the indicators are too 
many to measure. (A related problem is finding a control area 
that matches the treatment area in all respects except the 
treatment.) Second, the results of the monitoring approach are 
nearly impossible to extrapolate elsewhere, for essentially the 
same reasons: even should one be able to demonstrate a response 
to the treatment (action), what combination of factors caused the 
response will not be clear, and seldom or ever will the same 
precise combination of factors exist elsewhere. Experience 
suggests that these problems override most advantages the 
monitoring approach offers. 

The 'hypothesis-testing' approach likewise has two apparent 
disadvantages. First, the effects of each action must be added 
in some way to determine the overall effects of the ranch manage- 
ment operation. Second, the results may (as with the monitoring 
approach) have limited applicability elsewhere. But these prob- 
lems are more apparent than real. In the first place, one can 
usually examine (as has been done in the Saval workshops) the 
mechanisms that normally regulate each indicator, and find that 
very few of the planned actions are likely to significantly 
affect the indicator. Research can then focus on these few 
actions, and adding their effects is simple. And in terms of the 
transferability of site-specific data, as long as the inf ormat ion 
collected de scribe s ma i or f unct iona 1 re lat ionship a . it is usually 
broadly applicable (see Reichle 1975, Kerr and Neal 1976, Odum 
and Cooley 1976, Truett 1980). Moreover, because most actions 
are localized in time and space, suitable control areas (always 
required for rigorous design in impact analysis research) are 
frequently available nearby. 

The research plans that currently exist for the Saval Ranch 
seem to incorporate both research approaches. For example, the 
stated needs to conduct ranch-wide surveys (of vegetation trend, 
deer numbers, etc.), to study the effects of the ranching 
operation as a whole, and to find a control area outside the 
ranch suggest that researchers are proposing a monitoring 
approach. Other proposed studies (relationships between birds 



144 



and vegetation structure, and between cattle diet and forage 
composition, etc.) are hypothesis-testing types of inquiries and 
attempt to document functional relationships at selected sites. 

We suggest that, to respond to the stated project needs, 
research on the Saval Ranch emphasize hypothesis-testing studies 
that investigate functional relationships at selected sites. 
Only in this way can the research (1) go beyond measuring how the 
indicators changed, to suggest what caused the change, and (2) 
provide results that are readily applicable in other places and 
times. Moreover, evaluating impacts on an action-specific basis 
makes gaming with the computer model easier (most single-action 
scenarios can be handled by an APPLE mi c r o - co mpu t er ) and makes 
the computer output more understandable. 



9.3.2 Cross-Disciplinary Communication 

For research to be effective and efficient, there must be a 
consistent view among managers and research disciplines of what 
constitutes the problem. Briefly, the problem in the Saval ranch 
project appears to be: 

Ranch managers desire to make more money by maximizing 
the annual net production of grazing animals— cattle. They 
think the new management plan will promote this. Public 
resource managers (Bureau of Land Management, Forest 
Service, Nevada Department of Wildlife) want to be sure *:hat 
other renewable resources (e.g., soil productivity, selected 
fish and wildlife populations) do not suffer as a 
consequence. It appears desirable to conduct research to 
evaluate the plan's financial rewards, and its conflicts 
with other resources in such a way that both the research 
strategy and the results can be applied to the same kind of 
problem elsewhere. 



145 



Not only must the view of the problem be consistent, but 
strong emphasis should be placed on: 

( 1 ) Promoting frequent dialogue among field researchers and 
between field researchers and pro ject m anage m ent . 
Suggestions include (1) frequent meetings of all 
researchers to discuss their current efforts, new 
findings, and new proposed research plans, and (2) 
meetings between the field personnel and steering 
committee (during the regular steering committee 
seasons? ) . 

( 2 ) _C o. o. r. jL .in. iLiLi. JLiL J. _L £.1.<L H .§. A .§. !.£ J-.JL (data collecting) 
activities am ong the d i s c ip 1 ine s . For example, the 
proposed one-month observation to be made of cattle 
movement in North Independence Pasture could coincide 
with before-after measures of vegetation biomass at 
selected sites in the same pasture. 

(3) Insuring c o m pa t ib i 1 i t y am ong researchers in the w ay 
t hat c o m ponen t s are m ea sured and in the un its o f 
m easure, if cross-disciplinary use o f t he data i_s_ 
anticipated . Examples are: (1) If hydrologists need 
to know total canopy cover of vegetation regardless of 
the vegetative class, it is difficult for them to use 
data gathered by vegetation studies that measure only 
canopy cover of shrubs and basal cover of grasses, and 
thac segregate the data by plant class. (2) Unless 
wildlife people can convert grouse and deer to dollars, 
and it is unlikely they will be able to, the value of 
the wildlife will have to be judged outside the 
economics analysis. 

(4) Ma.k__in.g_ e.J._lo.r. t_JL t__o min._im_i.ze. inUr Ai-S^L IP. J.i.B.a.J-i 
communication prob lems caused by semant ic difficulties . 
For example, much of the terminology describing range 



146 



quality has evolved where cattle grazing has been the 
dominant use of rangeland. It describes how useful the 
range is to grazers, but not necessarily to anything 
else. Thus, practices currently called "range improve- 
ments" may or may not "improve" the range for animals 
other than livestock, and wildlife biologists and 
ranchers may have different perceptions of what such 
terms mean. 



9*3.3 Disciplinary Research 

The recommendations for disciplinary research that follow 
are based on several premises: 

(1) The purpose of the research is to evaluate effects of 
ranch management actions on selected indicators 
identified during the workshops. 

(2) The research design and output should have utility 
beyond the Saval to the extent possible. 

(3) Interdisciplinary coordination of research should be 
maximi zed . 

As suggested earlier, we believe that hypothesis-testing 
kinds of research, specific to actions and sites, should be 
instituted in all disciplines as soon as possible. In this and 
all impact analysis research, the smaller tht .-spatial scale, and 
the fewer the variables of interest, che easier it will be to 
isolate causes for observed changes. Ranch-wide inventories or 
monitoring programs are perhaps necessary for developing a 
baseline characterization, for tracking range condition changes, 
or for formulating important hypotheses, but monitoring -type 
programs will almost certainly fail to provide statistically 
valid and defensible answers about causes o f c hang; e s (i.e., to 



147 



determine whether observed changes can be attributed to operation 
of the Saval management plan). 

If these kinds of hypothesis-testing programs are adopted, 
there will probably be little need for "control" areas outside 
the ranch—they can be found (or created as exclosures, etc.) 
within the ranch. It is extremely doubtful if an adequate 
"control" ranch could be located anyway; there would be too many 
differences in unmeasured variables between it and the Saval that 
might affect the responses observed. 

Additionally we recommend that, as soon as possible, each 
disciplinary scientist review the literature in his field re lated 
t_o. t_b_e. ki.nd.8 o.__ iug.ctj.onal 2_iiJ_£__i.iL____Jl______ ________ e_me.r_g.e_d_ _____ 

im portant dur ing the w orkshop s . Certainly each person has 
already reviewed the general literature about his or her subject, 
but we are suggesting that each look in depth at a different kind 
_____ informatio n, J_.h_a_.t_ wh_Lc.h ____£__u_.s__e.__ h__,w __,__,__. J_Il<L__i_iLJL_____L 

identified _____ the workshops are regulated . As noted earlier, 

these kinds of functional relationships are generally much more 
conservative from place to place than are data describing levels 
of populations or other components, implying that literature from 
many other places and times are relevant. The importance of 
evaluating these kinds of literature to help researchers 
formulate important hypotheses and evaluate research results 
cannot be overstated. 

In a practical sense, we recommend that research focus in 
space on representative actions planned or already implemented on 
the Saval (e.g., improving irrigation systems in hay meadows, 
plowing sagebrush and seeding with crested wheatgrass, grazing 
pastures at given times and with given stocking rates, etc.). At 
the site of each action, the disciplines to be involved in 
research would be those that, judged by current knowledge, would 
hypothesize there to be a measurable effect of the action. For 
example, based on workshop discussions, it seems that a plowing 
and seeding operation might affect (1) vegetation production; (2) 
weight gain by calves; (3) numbers and/or diversities of grouse, 
deer, songbirds, and small mammals; and (4) water infiltration 



148 



and evapo t r an s p ir a t ion regimes. In this example, a research 
effort that involves these disciplines might then proceed to 
test, at appropriate sites, hypotheses about the impacts of 
plowing and seeding. 

A first step in the planning procedure would perhaps be to 
develop hypotheses about how expected actions would importantly 
affect each indicator. This would let each scientist determine 
which actions he might be interested in investigating. Examples 
of the kinds of hypotheses the workshop exercises suggested were 
important follow (some of these may have already been tested, and 

certainly there are others that need testing): 

♦ 

Vegetat ion 

Hay production can be doubled (or tripled, etc.) by 
improving the irrigation system without changing the 
annual amount of irrigation water used. 

Plowing sagebrush areas and seeding them to crested 
wheatgrass increases average annual biomass of fcrbs 
and cheacgrass produced. 

Formulas can be developed to predict (by season?) 
changing level of cattle use of each plant type 
(increaser, decreaser, etc.) with distance from water 
and elcpe of terrain. 



Hydro logy 

Flowing sagebrush areas and seeding them to eras ted 
vheatgrass increases average soil water availability in 
the top ten inches. 



149 



Water infiltration into the soil is significantly 
diminished when cattle graze pastures at current 
stocking rates in spring (or summer, or fall). 

Water infiltration into the soil is increased and total 
evapotransp irat ion is decreased when sagebrush stands 
are replaced with crested vheatgrass plantings. 



Cattle and Economics 

Weaning weights of calves (or, alternatively, calving 
success) are affected more by cow weights the preceding 
fall than by cow weights in early spring. 

Calf daily weight gain on a pasture is significantly 
increased when the total area of pasture more distant 
from water than 0.5 mi is reduced from 50% (or any 
given percent) to 0. 

Irrigation improvement (on given hay meadows) is a 
cheaper way (amortized over 15 yr) of acquiring winter 
hay than is buying hay from outside the ranch. 



Wildlife 

Plowing sagebrush areas and seeding them to crested 
wheatgrass decreases the amount of time deer/sage 
grouse use the areas during critical periods. 

Deer and cattle diet overlap, in pastures used 
simultaneously by both, is such that competition for 
food between deer and cattle is minimal regardless of 
the time of year the pasture is grazed. 



150 



Grazing selected pastures in spring (early summer) 
decreases the amount of time deer/sage grouse use them 
in spring and summer. 

Change in bird/small mammal species diversity as a 
consequence of a given management action is predictable 
on the basis of change in vegetation structure caused 
by the action. 

In summary, for results of research on the Saval Ranch to 
reliably evaluate the consequences of the Management Plan and be 
readily applicable elsewhere, the research scientists should 

(1) Test hypotheses related to impacts of specific actions 
at specific sites on selected indicators, 

(2) Shift away from approaches that attempt to monitor how 
indicators respond ranch-wide to the sum of management 
actions , 

(3) Clarify functional relationships that strongly 
influence the behavior of (i.e., 'regulate') the 
indicators, and that are sensitive to expected actions, 
and 

(4) Require a level of communication among disciplines that 
promotes interdisciplinary comp a t ab i 1 i t y in research 
goals, field methods, and data collected. 

Ideally, research design will continue to evolve as some 
hypotheses are tested and new hypotheses are developed. The 
model should guide this evolution and, as new data surface, 
become more realistic and hence more useful as a management tool. 



151 



10. LITERATURE CITED 

Branson, F.A., G.E. Gifford, K.G. Renard, and R.F. Hadley. 1981. 
Rangeland hydrology. Publication for Society for Range Manage- 
ment, Range Science Ser. No. 1, 2nd Ed. Kendall Hunt Publish- 
ing Company o 340 p. 

Charnov, ?. 1973. Optimal foraging: some theoretical considera- 
tions. Ph.D. Thesis, Seattle Univ., Seattle, WA. 

Garrison, G,A. 1971. Carbohydrate reserves and response to use. 
Pages 217-278 In: CM. McKell, J. P. Blaisdell, and J.R. Goodin 
(Eds.), Wildland Shrubs - Their Biology and Utilization. USDA 
For. Ser. Gen. Tech. Rep. INT-1. 

Harvard Library of Computer Graphics. 1979. Mapping software 
and cartographic data bases. Harvard University Laboratory for 
Computer Graphics and Spatial Analysis. 240 p. 

Holling, C.S. 1959. The components of predation as revealed by 
a study of small mammal predation of the European pine sawfly. 
Canad. Entomol. 91:293-320. 

Holling, C.S. (Ed.). 1978. Adaptive environmental assessment 
and management. John Wiley and Sons, New York. 377 p. 

Jensen, M.E., and H.R. Haise. 1963. Estimating e vapo t r an s p ir a- 
tion from solar radiation. J. Irrig., Drain. Div. 89:15-41. 

Eerr, S.R., and M.W. Neal. 1976. Analysis of large-scale ecolo- 
gical systems. J. Fish. Res. Board Can. 3 3 ( 9 ) : 20 83 -20 89 . 

Martin, J. 1977. Computer data-base organization. 2nd Edition. 
Prentice-Hall, Inc., New Jersey. 713 p. 

Martin, S. 1970. Relating vegetation measurements to forage 
consumption by animals. Pages 93-100 Ln: Range and Wildlife 
Habitat Evaluation — -A Research Symposium. USDA For. Ser. Misc. 
Publ. Nc. 1147. 

Noy-Meir, I. 1975. Stability of grazing systems: an application 
of predator -prey graphs. J. Ecol. 63:459-481. 

Oduu, E.P., and J.C. Cooley. 1976. Ecosystem profile analysis 
and performance curves as tools for assessing environmental 
impact. Paper presented at AIBS meeting, 2-3 June 1976, New 
Orleans, LA, 

Reich 1 e , D.E. 1975. Advances in ecosystem analysis. Bio 
Science 25( 4) : 257 -26 4. 

Xittenoouse, L.R.. and F.H. Sneva. 1977, A technique for esti- 
mating big sagebrush production. J. Range Manage. 3 C 1 ) : 6 8 - 
70. 



152 



Smith, R.E., and J.R. Williams. 1980. Simulation of surface 
water hydrology. In: W.G. Knisel (Ed.), Creams: A Field Scale 
Model For Chemicals, Runoff, and Erosion From Agricultural 
Management Systems. U.S. Dept. Agr. Cons. Rep. No. 26. 643 p. 

Truett, J.C. 1980. Beaufort Sea barrier is land- lagoon ecolo- 
gical process studies: Final Report, Simpson Lagoon. Part 7. 
Synthesis, Impact Analysis, and a Monitoring Strategy. 92 p. 
Res. Unit 467. I_n: Envir. Assess. Alaskan Cont. Shelf, Annu. 
Rep. Prin. Invest. BLM/NOAA, OCSEAP. Boulder, CO. 



Walker, B.H., D. Ludwig, C.H. Holling, and R.M. Peterman. 1981. 
Stability of semi-arid savanna grazing systems. J. Ecology 
69:473-498. 

Wischmeier, W.H., and D.D. Smith. 1978. Predicting rainfall 
erosion losses—a guide to conservation planning. DSDA Agr. 
Handbook No. 537. 58 p. 



153 









k [/i • oo en 










vO isj 25 '-n ^ | 


Cfl 




G 








03 (^ W ' 


c ° 




</> 








N) -P- W 




O 






r 


Un 


n J 




1 






O D 

> > 




00 *^ 


CO 1 






Z H 




2* 








m pi 





{a 










cn en 


























rt P9 
















d* h* 
















05 
















CO ?3 
















H- ft> 












CO 3 












1 o 










CD 


X ^ 




1 






O 


3 


CD 


1 




50 


o H 


o 








a) 


Q> (D 


50 








o 

PI 


a to 

I- 1 ft) 


TO 
O 








50 


l_i pj 
1— t-t 


m 


Bureau c 
Library 


)f Land Management 




Bldg. 50, 


Denver Federal Cente 


r 


Denver, ( 




^O 80225