4IEPA
United States
EnvironmQntal Protection
Agency
integrating Ecological Risk
Assessment and Economic
Analysis in Watersheds
A Conceptual Approach and
Three Case Studies
-------
EPA/600/R-03/140R
September 2003
Integrating Ecological Risk Assessment and
Economic Analysis in Watersheds:
A Conceptual Approach and
Three Case Studies
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH
/*Y~y Recycled/Recyclable
Printed witti vegetabte-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free.
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
ABSTRACT
This document reports on a program of research to investigate the integration of
ecological risk assessment (ERA) and economics, with an emphasis on the watershed as the scale
for analysis. In 1993, the U.S. Environmental Protection Agency initiated watershed ERA (W-
ERA) in five watersheds to evaluate the feasibility and utility of this approach. In 1999,
economic case studies were funded in conjunction with three of those W-ERAs: the Big Darby
Creek watershed in central Ohio; the Clinch Valley (Clinch and Powell River watersheds) in
southwestern Virginia and northeastern Tennessee; and the central Platte River floodplain in
Nebraska. The ecological settings, and the analytical approaches used, differed among the three
locations, but each study introduced economists to the ERA process and required the
interpretation of ecological risks in economic terms. A workshop was held in Cincinnati, OH in
2001 to review progress on those studies, to discuss environmental problems involving other
watershed settings, and to discuss the ideal characteristics of a generalized approach for
conducting studies of this type. Based on the workshop results, a conceptual approach for the
integration of ERA and economic analysis in watersheds was developed.
The objectives of this document (by chapter) are as follows:
describe the rationale, limitations, and contributions of the document (Chapter 1)
create a context for understanding by a diverse, technical audience (Chapter 2)
present a conceptual approach for integrating ERA and economics in the context of
watershed management (Chapter 3)
present and critically evaluate the methods and findings of the three watershed case
studies (Chapters 4-6)
identify research needed to improve the integration of ERA and economic analysis in
watersheds (Chapter 7).
This report is unique in its focus on the problem of ERA-economic integration and the
watershed management context and in its presentation of case studies. The conceptual approach
is used as a basis of discussion of each case study to illustrate how its particular methodological
advances and insights could be used to fullest advantage, both in the watershed studied and in
future integration efforts.
Preferred citation:
U.S. Environmental Protection Agency (USEPA). (2003) Integrating ecological risk assessment and economic
analysis in watersheds: A conceptual approach and three case studies. Prepared by the National Center for
Environmental Assessment, Cincinnati, OH. EPA/600/R-03/140R. Available from: National Technical
Information Service, Springfield, VA, PB2004-101634; and .
ii
-------
TABLE OF CONTENTS
Page
LIST OF TABLES ix
LIST OF FIGURES xii
LIST OF ABBREVIATIONS ,xv
PREFACE , .....xviii
AUTHORS, CONTRIBUTERS AND REVIEWERS xix
EXECUTIVE SUMMARY xxvii
1. INTRODUCTION 1-1
1.1 THE IMPORTANCE OF INTEGRATED, WATERSHED-LEVEL
ANALYSIS 1-1
1.2 GENESIS OF THIS DOCUMENT 1-5
1.3 OBJECTIVES AND ORGANIZATION . 1-7
1.3.1 Create a context for understanding by a diverse, technical
audience (Chapter 2) 1-7
1.3.2 Present a conceptual approach for integrating ERA and
economics in the context of watershed management (Chapters) , 1-8
1.3.3 Present and critically evaluate the methods and findings of
three case studies (Chapter 4-6) 1-8
1.3.4 Identify research needed to improve the integration of ERA and
economic analysis in watershed (Chapter 7) 1-8
1.4 RELATIONSHIP TO USEPA GUIDANCE DOCUMENTS. 1-9
1.4.1 USEPA Guidelines for Ecological Risk Assessment 1-9
1.4,2 USEPA Guidelines for Preparing Economic Analyses 1-9
1.4.3 USEPA Framework for Economic Assessment of Ecological
Benefits 1-10
1.5 LIMITATIONS 1-10
1.5.1 Lack of complete integration 1-10
1.5.2 Specificity to a watershed context 1-11
1.6 UNIQUE CONTRIBUTIONS 1-12
1.7 REFERENCES 1-13
111
-------
2. BACKGROUND: ECOLOGICAL RISK ASSESSMENT AND ECONOMIC
ANALYSIS IN WATERSHEDS AND THE NEED FOR INTEGRATION., 2-1
2.1 ECOLOGICAL RISK ASSESSMENT 2-1
2.1.1 Framework and methods for ecological risk assessment 2-2
2.1.2 Critiques of ecological risk assessment 2-11
2.1.3 Watershed applications of ecological risk assessment 2-14
2.2 ECONOMIC ANALYSIS 2-17
2.2.1 Welfare economics......... 2-17
2.2.2 Economic value 2-20
2.2.3 Cost-benefit analysis.... 2-25
2.2.4 Complementary analyses 2-26
2.2.5 Game theory ....2-28
2.2.6 Ecological economics 2-30
2.2.7 Applications of ecological economics.. ....2-31
2.3 ECOLOGICAL AND ECONOMIC ANALYSIS FOR WATER
QUALITY STANDARDS ....2-33
2.3.1 Water quality standards and ecological risk assessment 2-34
2.3.2 Water quality standards and economic analysis 2-38
2.4 THE NEED FOR INTEGRATION 2-41
2.5 REFERENCES 2-44
APPENDIX 2-A: DISCUSSION OF STATED PREFERENCE METHODS USED
IN TWO CASE STUDIES 2-59
APPENDK2-B: USING MULTIMETRIC INDICES TO DEFINE THE
INTEGRITY OF STREAM BIOLOGICAL ASSEMBLAGES
AND INSTREAM HABITAT 2-64
3. A CONCEPTUAL APPROACH FOR INTEGRATED WATERSHED
MANAGEMENT 3-1
3.1 EXISTING FRAMEWORK FOR WATERSHED MANAGEMENT 3-1
3.2 GUIDING CONSIDERATIONS FOR AN INTEGRATED
MANAGEMENT PROCESS 3-2
3.3 DIAGRAMING AN INTEGRATED MANAGEMENT PROCESS 3-7
3.3.1 Assessment planning 3-10
3.3.2 Problem formulation.... .3-11
iv
-------
3.3.3 Analysis and characterization of baseline risk 3-14
3.3.4 Formulation of alternatives 3-16
3.3.5 Consultation with extended peer community 3-18
3.3.6 Analysis and characterization of alternatives 3-18
3.3.7 Comparison of alternatives 3-20
3.3.8 Decision 3-21
3.3.9 Adaptive implementation ...3-21
3.3.10 Linkage to regular management cycles 3-22
3.4 EXAMPLES OF ANALYSIS AND CHARACTERIZATION
FOLLOWED BY COMPARISON OF ALTERNATIVES 3-23
3.4.1 Example 1: Cost-benefit analysis of all changes that can be
monetized, with qualitative consideration of other changes 3-23
3.4.2 Example 2: Use of stated preference techniques to effect
integration of ecological, economic and other factors 3-25
3.4.3 Example 3: Use of linked ecological and economic models to
dynamically simulate system feedbacks and iteratively revise
management alternatives 3-27
3.5 CONCLUSION 3-29
3.6 REFERENCES 3-31
APPENDIX 3-A: DISCUSSION OF EXISTING FRAMEWORKS THAT HAVE
BEEN APPLIED TO WATERSHED MANAGEMENT 3-38
4. EVALUATING DEVELOPMENT ALTERNATIVES FOR A
HIGH-QUALITY STREAM THREATENED BY URBANIZATION:
BIG DARBY CREEK WATERSHED 4-1
4.1 WATERSHED DESCRIPTION 4-2
4.2 ECOLOGICAL RISK ASSESSMENT 4-4
4.2.1 Planning 4-4
4.2.2 Problem formulation 4-6
4.2.3 Current status of analysis and risk characterization 4-8
4.3 ECONOMIC ANALYSIS 4-11
4.3.1 Research approach 4-12
4.3.2 Communicating the effects of urban development on
ecological endpoints 4-14
4.3.3 Communicating the effects of urban development on
economic and social services 4-17
4.3.4 Land use scenarios for framing expression of preference and
value in the stream 4-19
4.3.5 Eliciting monetary valuation 4-30
-------
4.3.6 Linking stream integrity to the development scenarios , 4-33
4.3.7 Linking stream integrity and willingness to pay ., 4-34
4.4 DISCUSSION ,. .. 4-37
4.5 REFERENCES 4-42
5. VALUING BIODIVERSITY IN A RURAL VALLEY: CLINCH AND
POWELL RIVER WATERSHED 5-1
5.1 WATERSHED DESCRIPTION 5-1
5.2 ECOLOGICAL RISK ASSESSMENT 5-4
5.2.1 Planning 5-4
5.2.2 Problem formulation..... 5-8
5.2.3 Risk analysis 5-12
5.2.4 Risk characterization 5-22
5.3 ECONOMIC ANALYSIS 5-25
5,3.1 Methods for valuing biodiversity and environmental quality 5-26
5.3.2 Integrating the choice model with the ecological risk
assessment. ....5-29
5.3.3 Results of economic analysis ....5-36
5.4 DISCUSSION ........5-43
5.4.1 Consultation with extended peer community 5-43
5.4.2 Baseline risk assessment 5-45
5.4.3 Formulation, characterization and comparison of
alternatives .........5-45
5.4.4 Adaptive implementation 5-49
5.5 REFERENCES 5-50
APPENDIX 5-A: EXCERPT FROM SURVEY ADMINISTERED BY THE
UNIVERSITY OF TENNESSEE: EXPLANATION OF
HYPOTHETICAL AGRICULTURAL POLICIES AND THEIR
POTENTIAL IMPACTS 5-53
APPENDIX 5-B: RANDOM UTILITY MODEL 5-56
VI
-------
6. SEEKING SOLUTIONS FOR AN INTERSTATE CONFLICT OVER WATER
AND ENDANGERED SPECIES: PLATTE RIVER WATERSHED 6-1
6.1 WATERSHED DESCRIPTION 6-1
6.1.1 Watershed resources and impacts of development 6-1
6.1.2 Watershed management efforts 6-7
6.2 ECOLOGICAL RISK ASSESSMENT 6-13
6.2.1 Planning 6-13
6.2.2 Problem formulation 6-15
6.2.3 Analysis 6-21
6.2.4 Risk characterization 6-25
6.3 ECONOMIC ANALYSIS 6-26
6.3.1 Model I: Determining who should provide and pay for
environmental water 6-29
6.3.2 Model II: Determining how much water to allocate to
environmental use 6-35
6.4 DISCUSSION 6-62
6.4.1 Assessment planning and problem formulation 6-62
6.4.2 Formulating alternatives, and baseline ecological risk
assessment 6-63
6.4.3 Analysis and characterization of alternatives, and comparison
of alternatives 6-64
6.4.4 Consultation with extended peer community 6-68
6.4.5 Decisions and adaptive implementation 6-69
6.5 REFERENCES 6-70
APPENDIX 6-A: SUMMARY OF SURVEY RESPONSE INFORMATION USED TO
CALCULATE UTILITY OF ENVIRONMENTAL MANAGEMENT
POLICY OPTIONS FOR THE CENTRAL PLATTE RIVER
FLOODPLAIN 6-80
7. CONCLUSIONS 7-1
7.1 ACHIEVING ECOLOGICAL-ECONOMIC INTEGRATION REQUIRES
A COHERENT STRATEGY 7-1
7.2 INTEGRATION REQUIRES ASSESSMENT PLANNING AND
PROBLEM FORMULATION TO BE INTERDISCIPLINARY 7-3
7.3 RESEARCH IS NEEDED ON THE DEVELOPMENT AND USE OF
INTEGRATED CONCEPTUAL MODELS 7-5
vii
-------
7.4 CLEARLY FORMULATED MANAGEMENT ALTERNATIVES
FACILITATE INTEGRATED ANALYSIS .7-5
7.5 CAREFUL EFFORT IS REQUIRED TO RELATE ECOLOGICAL
ENDPOINTS TO ECONOMIC VALUE , 7-7
7.6 THE APPROPRIATE TOOLS FOR ANALYSIS AND
COMPARISON OF ALTERNATIVES DEPEND ON THE DECISION
CONTEXT , ...7-11
7.7 RESEARCH IS NEEDED ON TRANSFERRING THE
VALUE OF ECOLOGICAL ENDPOMT CHANGES 7-14
7.8 THE ROLE OF ECOLOGICAL RISK INFORMATION IN THE
MEASUREMENT OF PREFERENCES REQUIRES FURTHER
RESEARCH 7-15
7.9 ' FINAL WORD 7-16
7.10 REFERENCES ..7-16
vui
-------
LIST OF TABLES
No. Title Page
1-1 Case studies of the integration of watershed ecological risk
assessment and economic analysis, funded by the USEPA in 1999 1-6
2-1 Daily's classification of ecosystem services with illustrative
examples 2-19
2-2 Methods for estimating values of environmental goods and
services ....2-23
2-3 Structure of a cost-benefit analysis 2-26
2-B-l Individual metrics constituting two indices of biological integrity
used by the Ohio Environmental Protection Agency 2-67
2-B-2 Primary and secondary metrics constituting the Qualitative Habitat
Evaluation Index (QHEI) used by the Ohio Environmental
Protection Agency 2-70
3-1 Typology of frameworks that have been applied to the processes of
watershed assessment and management 3-3
3-2 Important considerations in framework design, and resulting
design elements... 3-5
3-3 Categories (and some examples) of watershed management
measures ,3-17
3-4 Rough correspondence between the components of the conceptual
approach for ERA-economic integration and other selected
watershed management frameworks 3-30
4-1 a Relative effect of four housing development scenarios on the
four main causes of change in Big Darby Creek 4-20
4-lb Relative effect of four housing development scenarios on
socioeconomic outcomes in Big Darby Creek 4-21
4-2 Mean willingness to pay and confidence intervals for two
model specifications 4-32
4-3 Runoff-inducing condition and IBI per scenario 4-34
IX
-------
4-4 Estimated WTP per unit of BBI improvement over a 150-mi2 study
area for two model specifications 4-36
5-1 Outstanding ecological resources, environmental management
goal and management objectives for the Clinch Valley ecological
risk assessment , 5-7
5-2 Stressors and sources identified in the Clinch and Powell watershed 5-9
5-3 Attributes and attribute levels used in survey questionnaire 5-31
5-4 Sample question and choice set from survey questionnaire 5-32
5-5 Choice model variables and expected sign 5-34
5-6 Summary statistics 5-37
5-7 Results for conditional logit with CHOICE as dependent variable 5-39
5-8 Implicit prices, or implied willingness to pay for a given attribute
level as compared with the status quo....... 5-41
6-1 Participants in planning for the central Platte River floodplain
W-ERA ....6-14
6-2 Eleven environmental management objectives that are implicit in
and required to achieve the management goal.. 6-16
6-3 Principal stressors (and their primary sources) in the central Platte
River floodplain 6-17
6-4 Ecological assessment endpoints for the central Platte River
floodplain W-ERA 6-18
6-5 Selected assessment endpoints and stressors and the associated
risk hypotheses developed during problem formulation for the
central Platte River floodplain W-ERA 6-20
6-6 Welfare effects from supplying 140,000 acre-feet of
environmental water 6-34
6-7 Statements used in the household preferences survey to assess
respondent level of knowledge; answers regarded by researchers
as correct; and basis. Respondents were asked to rate
agreement/disagreement on a five point scale 6-37
-------
6-8 Descriptions of the three policy attributes and their respective
levels, a-e, that were evaluated in part 3 of the household
preferences survey, 6-39
6-9 Respondent classification into bargaining groups, by state. Based
on type of employment, interest-group affiliation, and attitude
regarding endangered species, a respondent could be classified
as either agriculture, environmental, both, or neither 6-49
6-10 Definition of Pareto efficient policy options: attribute levels
corresponding to each policy 6-52
6-11 Pareto efficient policy preferences, by state 6-53
6-12 Pareto efficient policy preferences, by bargaining group and
state 6-54
6-13 Comparison of preferred policy options between competing
interest groups 6-56
6-14 Results of bargaining models, all bargaining groups 6-58
6-A-l Degree of support for policy attributes, by state 6-81
6-A-2 Degree of support for policy attribute levels in Colorado, by
interest group ; 6-82
6-A-3 Degree of support for policy attribute levels in Nebraska, by
interest group 6-83
6-A-4 Degree of support for policy attribute levels in Wyoming, by
interest group 6-84
6-A-5 Policy attribute weights by bargaining group 6-85
XI
-------
LIST OF FIGURES
No. Title
1-1 Locations in the USA of five watershed ecological risk
assessment studies undertaken by USEPA and other partners.
Comparison economic analyses were undertaken at three of the
five locations 1-4
2-1 Framework for ecological risk assessment 2-4
2-2 Estimation of risk by comparing a cumulative frequency distribution
of exposure to a stressor and a stressor-response relationship;
ECX denotes stressor concentration affecting X% of test population 2-10
3-1 A conceptual approach for the integration of ecological risk
assessment and economic analysis in watershed management 3-8
3-2 Analysis and characterization of alternatives, folio wed by their
comparison, example 1: CBA of all changes that can be monetized,
with qualitative consideration of other changes 3-24
3-3 Analysis and characterization of alternatives, followed by their
comparison, example 2: use of stated preference techniques to
effect integration of ecological, economic and other factors.... 3-26
3-4 Analysis and characterization of alternatives, followed by their
comparison, example 3: use of linked ecological and economic
models to dynamically simulate system feedbacks and iteratively
revise management alternatives ., 3-28
3-A-l Framework for environmental health risk management..... 3-41
3-A-2 Framework for integrated environmental decision making 3-42
3-A-3 A framework for planning and project development of large dams,
including five key decision points at which specific criteria
should be evaluated 3-44
3-A-4 A watershed management model for the planning and implementation
of watershed projects 3-45
3-A-5 The USFS planning framework incorporates regular adaptive
management and situational planning processes 3-47
3-A-6 The watershed-based management cycle used by many states may
include TMDL development and implementation 3-49
xii
-------
4-1 The Big Darby Creek watershed in central Ohio, USA 4-3
4-2a Illustration of high density scenario (dots represent houses) 4-23
4-2b Illustration of low density ranehette scenario (dots represent houses) , 4-25
4-2c Illustration of low density cluster scenario (dots represent houses) 4-27
4-2d Illustration of present agriculture scenario (dots represent houses) 4-29
4-3 Techniques used for analysis, characterization and comparison of
management alternatives in the Big Darby Creek watershed, as
compared to the example shown in Figure 3-3 4-39
5-1 The Clinch and Powell River watershed in the eastern USA.
The study area is the portion of the watershed that is above Morris Lake.
Initial ecological study focused on Copper Creek. Towns where
discussions were held shown, as are urbanized areas 5-2
5-2 Comparison between historic (pre-1910) and present locations of
native mussel concentrations in the Clinch/Powell watershed; red areas
represent mussel beds 5-5
5-3 Simplified conceptual model showing major pathways between
sources (land use), stressors, and effects on the assessment endpoint
for native mussel species abundance and distribution and data
sources available 5-10
5-4 Fish community integrity as a function of agricultural land in a
riparian corridor of 200 m width and 1500 m length in Copper
Creek 5-15
5-5 Relationship between two instream physical habitat parameters,
clean sediment (substrate embeddedness) and instream cover, and
IBI score, where IBI is categorized as either poor (impaired) or
good (unimpaired) based TVA's criteria; fish community
impairment is associated with poorer habitat quality as measured by
these two parameters 5-18
5-6 Fish IBI (A) and maximum number of mussel species (B) in the
Clinch/Powell basin as a function of the number of stressors 5-19
5-7 Number of mussel species recorded over time at two sites in
Clinch/Powell watershed affected by large toxic point-source
discharge events..... ...5-21
xm
-------
5-8 Techniques used for analysis, characterization, and comparison of
management alternatives in the Clinch Valley Watershed, as
compared to the example shown in Figure 3-3 .....5-47
6-1 The watershed of the North Platte, South Platte and Big Bend
Reach of the Platte River in the great plains of the USA 6-2
6-2 Price of 10,000-acre foot increments of environmental water, and
cumulative cost, assuming different levels of political
compensation , 6-33
6-3 Techniques used for analysis, characterization, and comparison of
management alternatives in the central Platte River floodplain, as
compared to the example shown in Figure 3-3 6-65
xiv
-------
LIST OF ABBREVIATIONS
ASCA
ASCB
AWQC
BMPs
BOD
CA
CAFOs
CBA
CEA -
CENR
CERCLA
COD
CSO
cv
CVM
CWA
DEM
DO
DOT
DPSIR
DWR
EIA
EMAP
Alternative-Specific Constants - Option A
Alternative-Specific Constants - Option B
Ambient Water Quality Criteria
Best Management Practices
Biological Oxygen Demand
Conjoint Analysis
Confined Animal Feeding Operations
Cost-Benefit Analysis
Cost-Effectiveness Analysis
Committee on Environment and Natural Resources
Comprehensive Environmental Response, Compensation and Liability Act
Chemical Oxygen Demand
Combined Sewer Overflow
Compensating Variation
Contingent Valuation Method
Clean Water Act
Digital Elevation Models
Dissolved Oxygen
Department of Interior
Driving forces, Pressures, State, Impacts, Response
Department of Water Resources
Economic Impact Analysis
Environmental Monitoring and Assessment Program
xv
-------
EPT
ERA
ESA
FERC
GIS
IBI
ICI
KAF
KL
Mlwb
MRS
NEPA
NOAA
NFS
NRC
NRCS
NRDA
OECD
OEPA
POTWs
PRWCMT
QHEI
RUM
Ephemeroptera, Plecoptera, and Tricloptera index
Ecological Risk Assessment
Endangered Species Act
Federal Energy Regulatory Commission
Geographic Information Systems
Index of Biotic Integrity
Invertebrate Community Index
Knowledge Adjustment Factor
Knowledge Index
Knowledge Level
Modified Index of Well-Being
Marginal Rate of Substitution
National Environmental Policy Act
National Oceanic and Atmospheric Administration
Nonpoint Source
National Research Council
Natural Resource Conservation Service
Natural Resource Damage Assessment
Organization for Economic Cooperation and Development
Ohio Environmental Protection Agency
Publicly-Owned Treatment Works
Platte River Whooping Crane Maintenance Trust
Qualitative Habitat Evaluation Index
Random Utility Model
xvi
-------
SAB
TMDL
TN
TNC
TP
TVA
UAA
UN-L
USAGE
USEPA
USFS
USFWS
USGS
UT-K
W-ERA
WQS
WTA
WTP
Science Advisory Board
Total Maximum Daily Load
Total Nitrogen
The Nature Conservancy
Total Phosphorus
Tennessee Valley Authority
Use Attainability Analysis
University of Nebraska-Lincoln
U.S. Army Corps of Engineers
U.S. Environmental Protection Agency
U.S. Forest Service
U.S. Fish and Wildlife Service
U.S. Geological Survey
University of Tennessee-Knoxville
Watershed Ecological Risk Assessment
Water Quality Standards
Willingness to Accept
Willingness to Pay
xvn
-------
PREFACE
A national goal of the Clean Water Act is to achieve water quality that provides for the
protection and propagation offish, shellfish, and wildlife, wherever attainable. To ensure a
sound scientific basis for the protection of aquatic and other ecosystems and the diversity of
species they support, the USEPA published a Framework for Ecological Risk Assessment in
1992 and Guidelines for Ecological Risk Assessment in 1998. Since the early 1990s, the USEPA
has also urged the use of a "watershed approach" to aquatic ecosystem protection, which views
the geographic area encompassed by a watershed as the basis for monitoring, assessment and the
formation of management partnerships and action plans. The watershed is also the usual basis
for establishing total maximum daily loads (TMDLs) for impaired waters.
Under Executive Order 12866 and the Unfunded Mandates Reform Act, the USEPA is
required to document the costs and benefits of its major regulatory actions. To guide those
efforts it published, in 2000, the Guidelines for Preparing Economic Analyses, Additional
guidance for determining the economic benefits of ecosystem protection was provided in the
2002 Framework for the Economic Assessment of Ecological Benefits. More information is
needed, however, about the application of economic methods to local ecological protection
efforts, such as at the level of the watershed. Watersheds are varied settings in which the
ecological resources, stakeholder concerns, management partnerships and decision-making
arrangements tend to be unique, and flexible approaches to analysis and problem-solving are
required. Furthermore, while advances continue to occur in the methods of ecological risk
assessment and economics, the integration of these sciences remains problematic.
This technical report presents the results of USEPA-sponsored ecological and economic
research conducted in three locations: the Big Darby Creek watershed of Ohio, the upper Clinch
and Powell River watersheds of Virginia and Tennessee and the central reach of the Platte River
in Nebraska. The watershed management problems that were addressed and the study techniques
used differed from case to case, and they achieved varying degrees of success. The information
gained from these experiences has enabled the development of a generalized conceptual
approach for the integration of ecological risk assessment and economic analysis in watershed
management, which this report also presents.
This report will be useful to technical audiences interested in the science and practice of
watershed management and in the scientific and practical problems that underlie the integration
of ecology and economics. The conceptual approach that it presents provides useful insights for
the future design of integrated watershed assessments.
XVlll
-------
AUTHORS, CONTRIBUTERS, AND REVIEWERS
The National Center for Environmental Assessment (NCEA), within U.S. EPA's Office
of Research and Development, was responsible for preparing this document. Randall J. F.
Bruins and Matthew T. Heberling (NCEA) were the document editors.
CHAPTER AUTHORS
Chapter 1. Randall J. F. Bruins and Matthew T. Heberling
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 2. Randall J. F. Bruins and Matthew T. Heberling
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 3. Randall J. F. Bruins and Matthew T. Heberling
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 4. O. Homer Erekson
Bloch School of Business & Public Administration
University of Missouri, Kansas City, MO
Orie L. Loucks, Steven R. Elliott, and Donna S. MeCollum
Departments of Economics and Zoology
Miami University, Oxford, OH
Marc Smith
Ohio Environmental Protection Agency, Columbus, OH
Randall J. F. Bruins
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 5. Steven Stewart
Department of Hydrology & Water Resources
The University of Arizona, Tucson, AZ
James A. Kahn
Environmental Studies Program, Williams School of Commerce
Washington and Lee University, Lexington, VA
Amy Wolfe and Robert V. O'Neill
Environmental Sciences Division
Oak Ridge National Laboratory, Oak Ridge, TN
xix
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont.)
Victor B. Serveiss
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Washington, DC
Randall J. F. Bruins and Matthew T. Heberling
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 6. Raymond Supalla, Bettina Klaus, and John Allen
Department of Agricultural Economics
University of Nebraska, Lincoln, NE
Dennis E. Jelinski
Departments of Biology and Geography
Queens University, Kingston, Ontario
Osei Yeboah
Department of Agricultural Economics
Auburn University, Auburn, AL
Victor B. Serveiss
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Washington, DC
Randall.I. F. Bruins
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
Chapter 7. Randall J. F, Bruins
National Center for Environmental Assessment
U.S. Environmental Protection Agency, Cincinnati, OH
xx
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont)
EXTERNAL REVIEWERS
Darrell Bosch, Ph.D.
Department of Agricultural and Applied Economics
Virginia Tech
Blacksburg, VA
Robert Costanza, Ph.D.
Gund Institute for Ecological Economics
University of Vermont
Burlington, VT
Peter deFur, Ph.D.
Environmental Stewardship Concepts
Richmond, VA
INTERNAL REVIEWERS
Chapter 2. Anne Grambsch
Office of Research and Development
National Center for Environmental Assessment
Brian Heninger
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Sabrina Ise-Lovell
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Christopher Miller
Office of Water
Office of Science and Technology
Mark L. Morris
Office of Water
Office of Science and Technology
Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
xxi
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cent.)
William O'Neil
Office of Policy, Economics and Innovation
National Center for Environmental Economics
John Powers
Office of Water
Office of the Assistant Administrator
Keith Sargent
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Anne Sergeant
Office of Research and Development
National Center for Environmental Assessment
Victor Serveiss
Office of Research and Development
National Center for Environmental Assessment
Glenn Suter II
Office of Research and Development
National Center for Environmental Assessment
William Wheeler
Office of Research and Development
National Center for Environmental Research
Chapter 3, Wayne Munns
Office of Research and Development
National Health and Environmental Effects Research Laboratory
Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
William O'Neil
Office of Policy, Economics and Innovation
National Center for Environmental Economics
John Powers
Office of Water
Office of the Assistant Administrator
xxn
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont.)
Keith Sargent
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Anne Sergeant
Office of Research and Development
National Center for Environmental Assessment
Chapter 4.
Victor Serveiss
Office of Research and Development
National Center for Environmental Assessment
Brian Heninger
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Matt Massey
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Lester Yuan
Office of Research and Development
National Center for Environmental Assessment
Chapters. Susan Herrod-Julius
Office of Research and Development
National Center for Environmental Assessment
Matt Massey
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Chapter 6. Sabrina Ise-Lovell
Office of Policy, Economics and Innovation
National Center for Environmental Economics
xxm
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont)
Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Catriona Rogers
Office of Research and Development
National Center for Environmental Assessment
Chapter 7. Stephen Newbold
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Keith Sargent
Office of Policy, Economics and Innovation
National Center for Environmental Economics
Victor Serveiss
Office of Research and Development
National Center for Environmental Assessment
CLEARANCE REVIEWERS
Glenn Suter n
Office of Research and Development
National Center for Environmental Assessment
Michael Slimak
Office of Research and Development
National Center for Environmental Assessment
ACKNOWLEDGMENTS
The editors wish to acknowledge Bette Zwayer, Pat Daunt, Patricia L, Wilder, Dan
Being, Teresa Shannon and Lana Wood for their assistance in the preparation of this document,
and Ruth Durham, Donna Tucker and David Bottimore for management of document reviews.
We acknowledge Glenn Suter II, Chris Cubbison, Mike Troyer and Haynes Goddard for
participation in the review of grant proposals that formed the core of this research effort, and
xxiv
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont.)
Barbara Cook for invaluable assistance in grant management. Mike Troyer prepared maps
appearing in several chapters. We also acknowledge the important work of Suzanne Marcy, and
members of the USEPA Risk Assessment Forum, who initiated the watershed ecological risk
assessments that provided a basis for this research, and Victor Serveiss, who later assumed
leadership of the watershed ecological risk assessment effort. Finally, we acknowledge Jackie
Little and Nancy Keene of TN and Associates for their assistance in the organization of a
workshop held in 2001 in Cincinnati, OH, and the attendees of that workshop, many of whom
are further acknowledged below.
Chapter 3
The authors wish to acknowledge Glenn Suter II for many helpful discussions in the
development of this chapter.
Chapter 4
The authors wish to thank the members of the Darby Partners and members of the Big
Darby Creek Watershed Ecological Risk Assessment Workgroup for their participation in the
development of the risk assessment on which parts of this chapter are based. We also thank
attendees of a workshop held in July 2001, in Cincinnati, OH for their comments on an early
draft of this work, and in particular we acknowledge John M. Gowdy, Robert V. O'Neill, Ralph
_ Ramey, and David Szlag for their written reviews. The views expressed in this chapter are those
of the authors and do not necessarily reflect the views or policies of the U.S. Environmental
Protection Agency.
xxv
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (cont.)
Chapter 5
The authors wish to thank the members of the Clinch and Powell Watershed Ecological
Risk Assessment Workgroup for their participation in developing the USEPA assessment report,
upon which this manuscript is based, Dennis Yankee and Jeff White provided GIS support and
database management. We also thank attendees of a workshop held in July 2001, in Cincinnati,
OH for their comments on an early draft of this work, and in particular we acknowledge Leonard
Shabman, Charles Menzie, Glenn Skinner and James E. Smith for their written reviews. The
Ui
views expressed in this chapter are those of the authors and do not necessarily reflect the views
or policies of the U.S. Environmental Protection Agency.
Chapter 6
The authors wish to thank the members of the Middle Platte Watershed Ecological Risk
Assessment Workgroup for their effort in performing activities upon which this report is based,
and Nancy Pritchett for technical support. We also thank attendees of a workshop held in July
2001, in Cincinnati, OH for their comments on an early draft of this work, and in particular we
acknowledge Glenn Suter II, Haynes Goddard and Ann Bleed for their written reviews. The
views expressed in this chapter are those of the authors and do not necessarily reflect the views
or policies of the U.S. Environmental Protection Agency.
xxvi
-------
EXECUTIVE SUMMARY
1. INTRODUCTION
Aquatic ecosystems provide many services to human society, including the supply of
water, food and energy, the treatment of wastes, opportunities for recreation, and the provision of
habitat for many valued species. However, by altering stream corridors, changing patterns of
flow, introducing nonindigenous species, and releasing pollutants into these ecosystems, society
has diminished their ability to continue providing these services. Because aquatic ecosystems
have complex interactions with their surrounding landscapes, efforts to better manage and to
restore these systems often focus on watersheds as basic units for analysis.
This document is concerned with two types of analysis that are both important for aquatic
ecosystem management: ecological risk assessment (ERA) and economic analysis. Both have
been recognized as necessary, but they have been kept largely separate in practice, and this
separation can hamper management efforts.
Recommended procedures for carrying out ERA have been published by the U.S.
Environmental Protection Agency (USEPA) and are widely used for regulation and management.
ERA carried out at the spatial scale of the watershed is termed watershed ERA (W-ERA).
Watershed management choices involve complex and uncertain trade-offs of current and future
financial and ecological resources. Economics offers analytic frameworks for evaluating the
trade-offs involved in choices made by individuals, firms or society. However, the integration of
W-ERA and economic analysis entails theoretical, technical and procedural challenges (Section
1.1).
xxvn
-------
This document reports on a program of research to investigate the integration of ERA and
economics, with an emphasis on the watershed as the scale for analysis. In 1993, USEPA
initiated W-ERA in five watersheds to evaluate the feasibility and utility of this approach. In
1999, economic case studies were funded in conjunction with three of those W-ERAs: the Big
Darby Creek watershed in central Ohio; the Clinch Valley (Clinch and Powell River watersheds)
in southwestern Virginia and northeastern Tennessee; and the central Platte River floodplain in
Nebraska. The ecological settings, and the analytical approaches used, differed among the three
locations, but each study introduced economists to the ERA process and required the
interpretation of ecological risks in economic terms (Section 1,2),
The goal of the research reported in this document was to enhance the management of
aquatic ecosystems by piloting the integration of ERA and economic analysis in watersheds.
This document is intended for technically educated readers with an interest iii improving
environmental management, including academic, government, or private researchers, and local,
state, or federal environmental decision-makers. The objectives of this document (by chapter)
are as follows (Section 1.3):
create a context for understanding by a diverse, technical audience (Chapter 2)
present a conceptual approach for integrating ERA and economics in the context of
watershed management (Chapter 3)
« present and critically evaluate the methods and findings of the three watershed case
studies (Chapters 4-6)
* identify research needed to improve the integration of ERA and economic analysis in
watersheds (Chapter 7).
The topics discussed in this document overlap with the topics of three USEPA guidance
documents, the Guidelines for Ecological Risk Assessment, the Guidelines for Preparing
xxviii
-------
Economic Analyses and the Framework for Economic Assessment of Ecological Benefits. This
report is unique in its focus on the problem of ERA-economic integration and the watershed
management context and in its presentation of case studies. This research report should not be
construed as guidance, and it does not replace any of those guidance documents (Section 1.4).
Some limitations of this document should be recognized. First, while the case studies
provide insights into the problem of ERA-economic integration, these studies themselves were
not integrated in any ideal sense, since the ERA and economic components were carried out
separately. Second, the problem of integrating ERA and economic analysis for environmental
management in general has many facets, not all of which can be addressed in the watershed
context. Therefore, care should be taken in extending the findings of this document beyond that
context (Section 1.5).
Notwithstanding these limitations, this document makes several unique contributions for
environmental management. First, it helps risk assessors better understand how ERA procedures
can be integrated with economic analysis. Second, the risk assessment perspective employed in
this document also poses interesting challenges for the economist, since translating ecological
risks into terms amenable to economic analysis is difficult. Third, it enables a comparison of
three different approaches for ERA-economic integration. Finally, this document introduces, in
Chapter 3, a new conceptual approach for integrating ERA and economic analysis in the context
of watershed management (Section 1.6).
2. BACKGROUND: ECOLOGICAL RISK ASSESSMENT AND ECONOMIC
ANALYSIS IN WATERSHEDS AND THE NEED FOR INTEGRATION
This section provides an introduction to basic terms and concepts in ERA and economic
analysis and to some of their applications to watershed management. ERA is a scientifically-
based process for framing and analyzing the nature, probability and uncertainty of adverse
xxix
-------
effects from human-caused threats to ecological resources. Procedures described in USEPA's
Guidelines for Ecological Risk Assessment include four primary phases: planning, problem
formulation, analysis and risk characterization.
The planning process is a dialogue between risk assessors and risk managers and, where
appropriate, interested and affected parties (stakeholders). The dialogue clarifies the context of
the environmental decision facing officials or the public, the ecosystem management goals and
objectives (including the identification of what characteristics are valued), and the information
needs that the assessment should address. Problem formulation is a process of generating
preliminary hypotheses about how human activities may cause ecological effects. It requires the
identification of assessment endpoints (ecological entities that reflect the valued characteristics),
the development of one or more conceptual models (such as box-and-arrow diagrams of how
human activities may generate stressors, leading to effects on the endpoints), and the
development of an analysis plan. Analysis characterizes exposure and effects. Exposure
analysis describes sources of stressors, stressor transport and distribution, and the extent of
contact or co-occurrence between stressors and affected organisms. Effects analysis determines
what effects are thought to be elicited by a stressor, then examines the quantitative relationship
between the stressor and the response, the plausibility that the stressor may cause the response
(causality), and the links between particular measures of effect and the assessment endpoints.
Risk characterization unites information about exposure and effects, in order to first estimate
and then describe the risks of adverse effects of stressors. Risk characterization also describes
the adequacy of data, the strength of all available lines of evidence, and the uncertainties (Section
2.1.1).
Critics of the uses of ERA in decision-making have argued that assessments tend to rely,
too heavily on limited data, to oversimplify ecological complexities and to underestimate the
xxx
-------
likelihood of unexpected outcomes. They also have argued that assessors may be biased. Most
of these criticisms are addressed, however, if assessments establish an effective planning
dialogue, formulate problems appropriately, and carefully evaluate different lines of evidence, as
called for in the USEPA Guidelines (Section 2.1.2).
Watersheds have been used for over a century as a basis for the study and management of
water resources, and since the early 1990s USEPA has urged the use of a "watershed approach"
for the study and management of water quality problems. Conducting ERA on a watershed scale
makes sense whenever problems exist that are not addressed simply by the establishment and
monitoring of water quality standards (WQS). Examples include the presence of unusual or rare
habitats or species with atypical requirements; effects caused by multiple sources or stressors;
and effects due to stressors such as modification of flow or habitat for which WQS have not been
established or effects of unknown cause (Section 2.13),
Welfare economics is the study of agents (individuals, firms) making choices; it assumes
that they are trying to maximize their well-being (i.e., their welfare, also termed utility). In an
ideal market, agents' decisions would lead to an efficient outcome, or one in which all mutually
beneficial trades have been made. In real situations, however, characteristics of the market or of
the goods and services often make trade in the marketplace inefficient. Markets often fail to
allocate environmental goods and services efficiently, complicating efforts to estimate the values
5
of different levels of environmental protection (Section 2.2.1).
Therefore, economists have developed nonmarket methods for estimating economic
value, or people's willingness to pay (WTP) for these goods and services. A variety of methods
exist for estimating nonmarket values. They can be categorized according to how the data are
generated (i.e., whether preferences are revealed or stated). Revealed preference methods infer
values from data on actual market choices related to the good, such as travel to a recreational
xxxi
-------
site. Stated preference methods use data generated by placing individuals in hypothetical choice
settings, often by use of a questionnaire. The choice settings use descriptions of hypothetical
changes to environmental amenities in order to elicit values (Section 2.2,2). Two of the case
studies presented in this document used stated preference techniques. Chapter 4 describes a
contingent valuation method (CVM) model used to value alternative development approaches in
the Big Darby Creek watershed. CVM surveys ask individuals how much they would be willing
to pay for a specifically described nonmarket good. Chapter 5 uses conjoint analysis (CA) to
study social tradeoffs among riparian protection policies in the Clinch Valley. CA surveys ask
individuals to rank or choose their most preferred option from a set of nonmarket goods. Each of
the goods is described in terms of a common set of attributes, and one of the attributes is the cost
of providing the good (in order to estimate economic values).
Economic values can be incorporated into analyses to help support decisions about
environmental protection. Traditionally, a complete economic analysis consists of three
techniques: cost-benefit analysis (CBA), economic impact analysis and equity assessment. CBA
is the process of summing all the individual values, present and future, associated with a project
or policy. It provides a method to calculate whether the project or policy improves efficiency
based on whether there are positive or negative net benefits (Section 2.2.3). Economic impact
analysis is a process to quantify a variety of economic consequences of various actions. Equity
assessment allows economists to understand changes in the distribution of wealth due to a policy
or project (Section 2.2.4). A technique similar to CBA is cost-effective analysis, which ranks
alternatives that are expected to deliver comparable levels of environmental protection from
lowest to highest cost.
Game theory is a type of economic analysis that is concerned with human behavior and
can examine individuals interacting within a market or in situations of market failure. It entails a
xxxii
-------
theory of strategic behavior where an outcome depends on many individuals' strategies and the
current conditions of the situation (Section 2.2.5). Chapter 6 discusses the use of game models to
inform an interstate water negotiation in the Platte River watershed of Colorado, Wyoming and
Nebraska.
Ecological economics is a relatively new paradigm that has sought various ways to
incorporate into economic analysis the physical and biological limitations of the ecological
systems that underpin economic systems. In addition to efficiency and equity, it is also
concerned with determining the scale of economic activity that ecological systems can
sustainably support (Section 2,2.6). Analytic approaches have included various methods that
propose some biophysical commodity (e.g., land or energy) as a replacement for economic
welfare, as well as approaches that link ecosystem models to more conventional, welfare-based
economic models (Section 2.2.7).
Attempts at integrating ERA and economics under terms of the Clean Water Act (CWA)
have been limited. ERA procedures have been used to determine what CWA measures are
protective and whether they are physically attainable, whereas economic analyses have been
used primarily to determine what is cost-effective and financially attainable. Under the CWA,
states, tribes, and territories with approved WQS programs must establish designated uses or
goals for their water bodies. Scientifically-derived criteria are then adopted to protect the
designated uses. In general, WQS are based on a level of water quality that provides for the
protection of aquatic life (i.e., propagation offish, shellfish, and wildlife) wherever it can be
reasonably attained, not wherever it can be shown to provide positive net benefits. If WQS are
not being met, the costs to society of attainment can be substantial, but the benefits of attainment,
XXXlll
-------
often harder to measure, can be large as well. Therefore, methods for better understanding the
tradeoffs between the ecological and economic effects of WQS are of interest (Section 2.3).
The same holds for the attainment of nonregulatory goals. ERA is useful for determining
the likely ecological responses to various kinds of proposed management actions, and economic
analysis is useful for interpreting those ecological changes, and other changes, in terms of human
well-being - so that decisions are effective and beneficial. But the best results will be achieved
only if ERA and economic analysis are integrated, rather than compartmentalized. A coherent
integration approach is needed (Section 2.4).
3. A CONCEPTUAL APPROACH FOR INTEGRATED WATERSHED
- MANAGEMENT
Several frameworks have been applied to watershed management processes, but none has
addressed specifically the ERA-economic integration problem. An approach for this purpose
should be tailored accordingly, since existing frameworks vary widely in scope and purpose.
Some address only monitoring or assessment, and exclude decision-making, whereas others
describe planning and management processes more broadly. Some frameworks are for
situational use, in response to problems or opportunities, whereas others describe regular,
ongoing management processes. Frameworks also differ as to the extent to which they integrate
the natural and social sciences and in the roles stakeholders are expected to play (Section 3.1).
Other characteristics should also be considered in design of anew approach. According
to USEPA's Science Advisory Board, processes used for integrated environmental management
should be transparent (clearly understandable) to all parties; flexibly applied; dynamic
(interconnected and iterative); open and cooperative; informed by many different sources and
disciplines; and should reflect holistic, systems thinking (Section 3.2).
xxxiv
-------
This document presents a new conceptual approach for the integration of ERA and
economic analysis in watersheds (Figure ES-1). The approach is designed so as to recognize the
unique value of ERA, to be responsive to critiques of ERA, to incorporate key attributes of
economic thought, to be pluralistic in methodology, to incorporate adaptive management (i.e., a
learning-by-doing approach) and to link situational and regular management processes. It
borrows from USEPA's Framework for Ecological Risk Assessment, but modifies that approach
at every stage to integrate economic analysis.
Assessment planning is analogous to "planning" in ERA, except that identification of
the decision context is expanded to include determining who has the authority to make the
decisions and what criteria they expect to use. Problem formulation is also similar to that done
in ERA, except that economic as well as ecological assessment endpoints must be identified, and
the relationships that are diagramed in conceptual models must include hypotheses about how
the various management alternatives would affect the ecological and economic assessment
endpoints. Analysis and characterization of baseline risk corresponds to the ERA stages of
analysis and risk characterization but is limited to risks that exist now, or will occur in the future,
if no new management action is taken. Formulation of alternatives entails the development of
alternative action plans for achieving the watershed management objectives. It is required for
integrated analysis, since economic analysis generally requires the evaluation of alternatives.
Consultation with the extended peer community refers to deliberation with scientific peers as
well as with stakeholders who have practical knowledge that is relevant to the situation.
Analysis and characterization of alternatives is the stage in which the management
alternatives are assessed from the perspectives of ERA and economics (and possibly other
disciplines such as human health risk or sociocultural assessment). Ecological risk
characterization describes probabilities, magnitudes and severities of effects on ecological
xxxv
-------
ASSESSMENT PLANNING
(Stakeholders, Managers, Technical
Specialists Dialogue)
Y : rv
PROBLEM FORMULATION
Integrated
Conceptual,M'
m
CONSULTATION
WITH
EXTENDED PEER
COMMUNITY
: :' :'!: ' " ';,
mm : "-''';'^-.-; i
Negotiation
;Revi;:. '*< .
Shading indicates primary role
played by technical specialists
White indicates interaction of
stakeholders, martagers and
technical specialists
ADAPTIVE
IMPLEMENTATION
FIGURE ES-1.
A conceptual approach for the integration of ecological risk assessment and economic analysis in
. watershed management
xxxvi
-------
assessment endpoints. The economic component analyzes costs and benefits associated with the
management alternatives, including changes in ecosystem services. Comparison of alternatives
is the step in which the ecological, economic and other factors, both qualitative and quantitative,
are arrayed for comparison. Depending on the decision context, comparison methods could
include stated preference methods, methods for assigning weights to different factors according
to their importance, or methods for modeling a negotiation process. Because watershed
management issues are often complex, the decision stage is likely to involve multiple parties and
may take the form of negotiation. Adaptive implementation, in which management actions are
monitored for effectiveness and periodically reevaluated, can help ensure that objectives are met.
It can also provide a means whereby parties who are at odds can agree on an interim step that
will be reevaluated after an agreed period. New information acquired during adaptive
implementation may require earlier stages of assessment to be revisited. The activities of this
conceptual approach are carried out only when situational needs arise, but they may be most
effective when linked to regular activities such as those of the watershed management cycle used
by many states (Section 3.3).
The more technical steps of integration, occurring in the analysis and characterization of
alternatives and in the comparison phase that follows, can employ a variety of ecological and
economic analytic tools. For example, analysis and characterization could involve estimating
monetary values for as many ecological and other changes as possible and using CBA to
estimate the overall net benefit of each alternative. Comparison would involve examining the
net benefits of the alternatives, in light of their impacts, equity effects, and any other effects that
could not be quantified (Section 3.4,1). In another example of an approach that could be used,
ecological effects, (market based) economic effects and other effects could be quantified to the
greatest extent feasible in the analysis and characterization phase, and the most important of
xxxvii
-------
these changes could be used in the design of a broadly-based stated preference study. Variants
of this approach are used in Chapters 4-6 (Section 3.4.2), Another possible approach could
involve the use of linked ecological and economic models to allow ecological-economic
feedbacks and optimize the design of alternatives (Section 3.4.3).
4. EVALUATING DEVELOPMENT ALTERNATIVES FOR A HIGH-QUALITY
STREAM THREATENED BY URBANIZATION: BIG DARBY CREEK
WATERSHED
Located in central Ohio, Big Darby Creek is widely recognized for its unusual biological
diversity, including many rare and endangered fish and freshwater mussel species; local efforts
to protect the watershed are longstanding. However, agricultural land uses, and rapidly
increasing urban development in the eastern portions of the watershed near Columbus, threaten
the stream's ecological quality. The watershed was selected for W-ERA because of
broad interest in protecting the Big Darby Creek and because of Ohio's large water quality
database (Section 4.1).
W-ERA was initiated in 1993 by USEPA, the Ohio Environmental Protection Agency
and other partners. The management goal for the watershed, arrived at through planning
discussions with residents, resource managers, public agencies and private organizations, was
protecting and maintaining native stream communities of the Big Darby ecosystem. In the
problem formulation phase, "species composition, diversity, and functional organization of the
fish and macroinvertebrate communities" was chosen as the assessment endpoint. Preliminary
analyses (which were expanded to encompass other areas of the Eastern Corn Belt Plains
ecoregion in Ohio) showed a negative association between urban development and the functional
organization offish communities (as measured by the index of biotic integrity or TBI). Risk
characterization in the watershed has not yet been completed (Section 4.2).
XXXVlll
-------
The specific objectives of the economic case study initiated in 1999 by Miami University
were as follows: (a) to estimate the quantitative or qualitative ecological and socioeconomic
impacts of four land use scenarios (Preserve agriculture; Zone for low density, ranchette style;
Zone for low density, cluster style and Take no action, allow high density urbanization); (b) to
communicate these impacts to the public effectively, and to measure the overall economic value
corresponding to each scenario based on individual willingness to pay (WTP) and (c) to better
understand the particular contribution stream ecological condition makes to the value of a given
scenario.
Presentations were made to three samples of respondents (residents, near-residents, and
non-residents), explaining the scenarios and their likely impacts on stream ecological condition,
local economic well-being and local quality of life. Using the CVM, respondents were asked
their WTP to avoid the high density urban scenario given one of the other three remaining
scenarios. The results suggest that the cluster-style alternative was preferred to the agriculture or
ranchette alternatives. In addition, residents were willing to pay more than near-residents, and
near-residents were willing to pay more than non-residents. Researchers also made a preliminary
attempt to associate the WTP with a unit change in the BI (Section 4.3).
This case study demonstrated an effective use of the planning and problem formulation
processes to initiate a baseline W-ERA, as well as an effective use of ecological risk information
to frame a valuation question. Its value for decision-making still is limited, resulting in part from
}
the separate conduct of the ERA and economic components. For example, the planning and
problem formulation stages of W-ERA did not characterize a specific decision context. The
economic study did not provide enough information to estimate the net social benefits or equity
effects of the scenarios, because costs to current landholders were not estimated. To better
determine the applicability of WTP measured in this study to watershed management, a renewed
xxxix
-------
assessment planning process focusing on development decisions would be needed. Further work
also is needed to determine the component of WTP that is specifically attributable to ecological
effects (Section 4.4).
5. VALUING BIODIVERSITY IN A RURAL VALLEY: CLINCH AND POWELL
RIVER WATERSHED
Originating in southwestern Virginia and extending into northeastern Tennessee, the
watershed of the Clinch and Powell Rivers historically contained one of the most diverse fish
and mussel assemblages in North America. Most evidence suggests land uses such as mining,
agriculture, urbanization and other human activities are responsible for the decline and extinction
of many of these populations. This area was chosen as a subject of W-ERA because of its
remaining valued aquatic resources, the wealth of information already collected, interest from
many groups, and the multiple stressors present (Section 5.1).
W-ERA was initiated in 1993 by USEPA, the U.S. Fish & Wildlife Service (USFWS),
The Nature Conservancy and other partners. An interagency workgroup determined the
management goal to be "establish[ing] and maintain[ing] the biological integrity of the
Clinch/Powell watershed surface and subsurface aquatic ecosystem." The two assessment
endpoints selected were: (1) reproduction and recruitment of threatened, endangered or rare
native freshwater mussels and (2) reproduction and recruitment of native, threatened, endangered
or rare fish species. Analyses examined various correlations between land uses, instream habitat
quality, IBI and mussel diversity. The assessment found that stream reaches with high portions
of riparian areas in agriculture had poor in-stream habitat and low IBI values, stream reaches
close to mining activity had low IBI values, and stream reaches with many stressors present had
low numbers of mussel species and low IBI values (Section 5.2).
xl
-------
The economic analysis, initiated in 1999 by a team headed by researchers from the
University of Tennessee-Knoxville, addressed the difficult task of valuing potential changes in
biological diversity and other ecological services at risk in the watershed. As a focus of analysis,
researchers examined hypothetical, voluntary policies to restrict agriculture in the riparian zone
with compensation to farmers. Using a conjoint analysis (CA) survey, watershed respondents
were asked to choose between alternative descriptions of the watershed as a function of the
agricultural policy and certain other characteristics. Those characteristics dealt with recovery of
aquatic life, quality of sport fishing, prevalence of song birds, effects on agricultural income, and
cost per household.
The responses provide information on the quality-of-life trade-offs respondents were
willing to make among various ecological and economic characteristics of this watershed. The
resulting choice model provides the values respondents would place on a range of policy changes
similar to those identified in the survey. This ability to estimate welfare effects over a complex
set of ecosystem changes is an advantage of CA over other valuation techniques (Section 5.3).
The economic study made effective use of qualitative information from the W-ERA study
to design the CA survey, and the study demonstrates the flexibility of the CA method. However,
because the ecological and economic effects of the policies themselves were not quantitatively
characterized, these results are of limited use for policy evaluation without additional analyses.
Furthermore, as in the Big Darby Creek case study, the decision context relevant to the
establishment of riparian management policies (i.e., who makes these decisions and how they are
made) would need to be further explored before the usefulness of this approach for management
could be determined (Section 5.4).
xli
-------
6. SEEKING SOLUTIONS FOR AN INTERSTATE CONFLICT OVER WATER AND
ENDANGERED SPECIES: PLATTE RIVER WATERSHED
Nearly one-half million sandhill cranes and several million ducks and geese use the
central Platte River floodplain in Nebraska during their annual migration. Several species that
depend on its broad, braided channel and associated wet meadow habitats including the
interior least tern, the piping plover and the whooping crane are federally listed as threatened
or endangered. However, flow diversions and storage reservoirs that supply irrigation,
hydropower and recreation to the region's economy are jeopardizing these habitats and species,
sparking conflict among federal agencies and water users in Nebraska, Colorado and Wyoming.
USFWS has determined an amount of annual flow and an acreage of restored wet meadows
required for meeting species' needs; the states have negotiated lesser amounts, to be
implemented on a trial basis and monitored for ten years, but since they still disagree as to who
should provide those reduced amounts, action has been delayed for several years (Section 6.1).
Interest in protecting these ecological resources, and willingness of several agencies and
stakeholders to participate, led to the establishment in 1993 of a W-ERA workgroup. The
management goal was to "protect, maintain and, where feasible, restore biodiversity and
ecological processes in the central Platte River floodplain, to sustain and balance ecological
resources with human uses." Nine assessment endpoints were derived from this broad goal, but
analyses were completed only for grassland breeding bird diversity and abundance and sandhill
crane abundance and distribution. Habitat use by wet-meadow nesting species was maximized in
larger patches, suggesting that habitat fragmentation has adverse effects on these species. Use of
river segments by sandhill cranes was found to be a function of channel width and the proximity
of wet meadows. However, a characterization of the risks to these species, especially in relation
to stream flow, was not completed (Section 6.2).
xlii
-------
An economic analysis initiated in 1999 by the University of Nebraska-Lincoln studied
game theory as a means to identify policies that might help resolve the Platte River resource
management conflict. Two models were constructed. Model I demonstrated a simple auctioning
approach for supplying the needed water whereby the players (the three states) would have
incentives to reveal their true supply costs. Model II, a multilateral bargaining model, sought to
identify promising policy solutions by examining (a) different ways to provide additional water
and habitat, (b) how far the parties are willing to go toward meeting the USFWS requirement and
(c) how costs should be shared. Constructing the model required surveying a sample of
households in the three states to correlate attitudes on these policy questions to membership in
interest groups (state residency, agricultural, environmental). The survey also evaluated
respondents' level of knowledge about the factors affecting these species. The survey found the
greatest level of disagreement was between agricultural and environmental interests within a
state, rather than among states. Policies finding widest acceptance involved adaptive (trial)
implementation, minimization of impacts on agriculture, and a partial sharing of costs by
environmental interests (Section 6.3),
The limited interaction between risk assessors and economists in this case study resulted
in a divergence of analytic objectives and perspectives. The W-ERA did not address a particular
decision context, whereas the economic study developed a tool designed to inform a specific
negotiation process. The W-ERA studied habitat requirements of dozens of riparian-dependent
avian species while the economic analysis addressed only the needs of endangered species.
Game theory models may be well-suited to the support of ongoing negotiation because they can
respond quickly to changes in negotiating position and can suggest new solutions. However, the
solutions will not necessarily result in protection of species unless (a) interest groups are
informed about species' needs and choose to support them or (b) informational feedback from
xliii
-------
the adaptive implementation process addresses key questions and is used to update the policies
(Section 6,4),
7. CONCLUSIONS
The following conclusions are derived from evaluation of the case studies:
Achieving ecological-economic integration requires a coherent strategy, such as the
conceptual approach presented in Figure ES-1
Integration requires assessment planning and problem formulation to be interdisciplinary,
involving ecologists and economists (and other disciplines as needed)
Research is needed on the development and use of integrated conceptual models, i.e.,
models that include economic as well as ecological endpoints and show how management
alternatives are expected to affect those endpoints
Clearly formulated management alternatives facilitate integrated analysis by giving risk
assessors and economists a common basis for analyzing endpoint changes
» Careful effort is required to relate ecological endpoints to economic value, including
linking these endpoints to ecosystem services and devising methods for explaining
ecological measurements or indices to the public
» The appropriate tools for analysis and comparison of alternatives depend on the decision
context, and since decision situations in watershed management are varied, a variety of
tools are needed
Research is needed on appropriate means of transferring the value of ecological endpoint
changes from one watershed setting to others
* The role of ecological risk information in the measurement of preferences requires further
research, since individuals who are surveyed may be unfamiliar with an issue and may
form their preferences based on information provided in a questionnaire.
xliv
-------
1. INTRODUCTION
1.1 THE IMPORTANCE OF INTEGRATED, WATERSHED-LEVEL ANALYSIS
Aquatic ecosystems provide many services to human society. They mediate the supply of
water for drinking and other human uses; they assimilate wastes and provide food, energy, and
habitat for many valued species; they offer opportunities for transportation and recreation; and
they provide aesthetic values and inspiration, hi taking advantage of these services, humans
have stressed these ecosystems. Alteration of stream corridors, changes in patterns of flow,
introduction of nonindigenous species, and pollution by toxicants, nutrients, sediments, heat, and
oxygen-demanding substances have diminished aquatic ecosystems* ability to continue
providing the services that society values.
As social awareness has increased, efforts have been made to better manage and reduce
human impacts upon these ecosystems, hi the U.S., these efforts have included increased
regulation and mitigation of pollution; increased attention to the ecological impacts of water
resource projects; modification of agricultural practices and subsidies; and efforts by urban,
suburban and rural communities to better steward their aquatic ecological resources through
monitoring, planning and collective action. Most of these efforts have been accompanied by a
recognition that aquatic ecosystems have complex interactions with their surrounding
landscapes. As a result, the watershed increasingly is seen as a basic unit for aquatic ecosystem
analysis and management.
This document is concerned with two types of analysis that are both important for aquatic
ecosystem management: ecological risk assessment (ERA) and economic analysis. Both have
been recognized as necessary, and their use is provided for in law and regulation, yet because
1-1
-------
they arise from very different philosophical traditions they have tended to remain separate in
i *y
both theory and practice. ' This separation hampers environmental management. Analysts from
the respective traditions often fail to coordinate their efforts, lack the ability to understand one
another's terminology and approach, or disagree as to what is important, and they may provide
decision-makers with incomplete or confusing information. Decision-makers may also assume
that these analyses ought to be separate and fail to recognize the wealth of insight that their
effective integration could produce.
ERA has been defined as "a process for collecting, organizing and analyzing information
to estimate the likelihood of undesired effects on nonhuman organisms, populations or
ecosystems."3 Recommended procedures for carrying out ERA have been published by the U.S.
Environmental Protection Agency (USEPA),4 and the practice has been employed for a wide
variety of ecological problems and settings. For example, a 1999 report by the Committee on
Environment and Natural Resources (CENR) documented the use of ERA by five U.S. federal
agencies to regulate the uses of toxic substances and pesticides, for the control of
nonindiginous species, and to remediate and determine compensation for damage caused by
chemical releases.5 The general principles of ERA also underlie many important regulatory
protections for aquatic ecosystems in the U.S., such as state-issued water quality standards
(WQS), but watersheds themselves are not usually the subject of ERA. However, routine
management approaches, including the monitoring and enforcement of WQS, cannot address
certain kinds of aquatic ecosystem impairment. Some undesired effects are caused by human-
caused insults (hereafter termed "stressors") for which there are no standards; these include, for
example, introduced organisms and altered habitat. Some are a complex result of multiple kinds
of stressors; and in some cases the causes remain unclear without further study. Moreover, some
1-2
-------
aquatic ecosystems host unique resources (such as rare species or habitats) having special
requirements that are not adequately understood. In addition, it is often unclear, without focused
analysis, whether a given set of proposed actions to correct these problems will be effective. Li
these cases, an ERA that is carried out at the spatial scale of the watershed, here termed
watershed ERA (W-ERA), may be useful.
As described in Section 2.1, W-ERA focuses on the key ecological resources and
management goals for the watershed, rather than regulatory standards alone. The approach
directly engages stakeholders in the determination of assessment goals and scope, identifies all
relevant threats, and applies scientific methods to the identification of causes, risks and
uncertainties of adverse effects. The resulting information is intended to be useful for the design
>
of approaches for ecosystem protection or restoration, whether these measures are physical or
institutional, regulatory or driven by incentives, governmental or community-based or some
combination of these.
hi 1993, USEPA initiated W-ERA in five watersheds to evaluate the feasibility and
usefulness of this approach (Figure l-l).5' The outcomes from some of these assessments, and
their usefulness for management, have been described in the literature,7"12 and W-ERA guidance
has been made available as a web-based training unit.13 Prior to this document, however, no
information has been available on approaches for integrating economic analysis with ERA in a
watershed management context.
Economists study choices made by individuals or other entities relating to the allocation
of scarce resources across competing uses (see Section 2.2), and economic analysis sometimes
has been used jointly with ERA in support of decisions (see CENR5 and Chapter 3). Watershed
management choices involve complex and uncertain trade-offs of current and future financial
1-3
-------
aquoit
Bay, MA
Big Darby gcj
Creek, OH
Middle Platte
River, NE
Clinch Valley,
VA / TN
Location of watershed ecological risk assessment
(W-ERA)
Location of W-ERA and related economic analysis
FIGURE 1-1
Locations in the USA of five watershed ecological risk assessment studies undertaken by
USEPA and other partners. Comparison economic analyses were undertaken at three of the five
locations as indicated.
1-4
-------
and ecological resources. Economics offers an analytic framework for determining whether a
given choice appears to provide an overall benefit to society. Depending on the approach used,
economic analysis can also address impacts on affected parties, can illuminate negotiation
processes, and can help evaluate the long term sustainability of outcomes. However, the
integration of W-ERA and economic analysis, which is needed to realize these insights, entails
theoretical, technical and procedural challenges.
1.2. GENESIS OF THIS DOCUMENT
This document reports on a program of USEPA-funded research to investigate the
integration of ERA and economics, with an emphasis on the watershed as the scale for analysis.
In 1998, the National Center for Environmental Assessment of USEPA's Office of Research and
Development solicited applications for assistance to conduct case studies of the integration of
ERA and economic analysis. Research to be funded was required to include original economic
analysis conducted in collaboration with an ongoing ERA, to reflect the state of the science of
ERA and economics, and to be relevant to decision-making with respect to the problem being
assessed, hi 1999, following peer review of proposals, economic case studies were funded in
conjunction with three of the five aforementioned W-ERAs (Figure 1-1, Table 1-1).
The resulting case studies were quite different from one another. The ecological settings
and resources of concern differed among the three locations. The degree of progress made by
each W-ERA team prior to the economic study varied as well, and the methodological lenses
brought to these problems by the respective economic teams also varied considerably. But the
commonalities between these three studies were also considerable in that each involved the
watershed scale, each introduced economists to the ERA process, and each included the
1-5
-------
TABLE 1-1
Case studies of the integration of watershed ecological risk assessment and economic analysis,
funded by USEPA in 1999
Study Area
Big Darby
Creek
watershed, Ohio
Upper Clinch
Valley, Virginia
and Tennessee
Central Platte
River
floodplain,
Nebraska
Project Title
"Determining biodiversity values in
a place-based ecological risk
assessment"
"A trade-off weighted index
approach to integrating economics
and ecological risk assessment"
"A strategic decision modeling
approach to management of the
middle Platte ecosystem"
Principal Investigators
O. Homer Erekson and One L.
Loucks
Miami University, Oxford, Ohio
James Kahn and Steven Stewart
University of Tennessee-Knoxville
Raymond Supalla
University of Nebraska-Lincoln
1-6
-------
challenging task of interpreting ecological risks in economic terms, in a manner that would be
meaningful to decision-makers.
Building on those commonalities, a workshop was held in Cincinnati, OH in 2001 to
review progress on those studies, to discuss environmental problems involving other watershed
settings, and to discuss the ideal characteristics of a generalized approach for conducting studies
of this type. Based on the workshop results, a conceptual approach for the integration of ERA
and economic analysis in watersheds was developed.
1.3 OBJECTIVES AND ORGANIZATION
The goal of the research reported in this document was to enhance the management of
aquatic ecosystems by piloting the integration of ERA and economic analysis in watersheds.
This document is intended for technically educated readers with an interest in improving
environmental management, including academic, government or private researchers, and local,
state or federal environmental decision-makers. This section describes the specific objectives of
this document (by document chapter).
1.3.1 Create a context for understanding by a diverse, technical audience
(Chapter 2)
Because of the differences in approach between ERA and economic analysis, most
readers will not be familiar with the methods and terminology of both. Therefore, Chapter 2
provides background information on ERA (Section 2.1) and economic analysis (Section 2.2) and
their applications to watersheds, with special reference to their relationship to WQS programs
(Section 2.3). Readers already familiar with any of these topics may skip the corresponding
sections of Chapter 2.
1-7
-------
1.3.2 Present a conceptual approach for integrating ERA and economics in the
context of watershed management (Chapter 3)
Chapter 3 presents a conceptual approach for the integration of ERA and economic
analysis in watershed management. This approach serves as a point of reference for critical
discussion of the three case studies, and it is intended to be useful for the design of future studies
that inform watershed decision-making. The chapter first reviews a variety of procedural
approaches that have been applied to the study and management of watershed problems. It then
identifies the main considerations that should guide the design of a conceptual approach and
describes such an approach.
1.3.3 Present and critically evaluate the methods and findings of three case studies
(Chapters 4-6)
Chapters 4-6 present detailed discussion of work done in each of the three watersheds
(Table 1-1). The organization of these chapters reflects the development of these studies. In
each case, a W-ERA was initiated first, by USEPA and other governmental and
nongovernmental partners. The complementary economic study was initiated later, through a
research grant to an educational institution. Therefore, after an initial section describing the
watershed setting, the second section of each case study chapter describes the methods and
findings of the W-ERA. The third section is devoted to the economic study, and a fourth section
critically analyzes the success of the integration and the usefulness of results for improving
management decisions.
1.3.4 Identify research needed to improve the integration of ERA and economic
analysis in watershed (Chapter 7)
This final chapter reexamines the commonalities of these studies to draw general
conclusions with respect to the integration problem, and it identifies areas for further research.
1-8
-------
1.4 RELATIONSHIP TO EXISTING USEPA GUIDANCE DOCUMENTS
1.4.1 USEPA Guidelines for Ecological Risk Assessment
USEPA published a Framework for Ecological Risk Assessment in 1992, and
Guidelines in 1998.4 These documents provide the basis for ERA as currently practiced in
USEPA and many other organizations, A further guidance document that provides detail on the
development of management objectives in the ERA planning process is currently in draft form.15
These methods, summarized in Section 2.1, formed the basis for the W-ERA studies described in
this document. The conceptual approach presented in Chapter 3 is based on those methods, but
shows how they may be modified and extended to enable the integration of ERA and economic
analysis in a watershed management context.
1.4.2 USEPA Guidelines for Preparing Economic Analyses
These Guidelines16 describe how USEPA conducts economic analyses of its
environmental policies and programs, as may be required for their justification under Federal
statute or Executive Order. They present methods for deriving monetary estimates of the costs
and benefits of those policies or programs. By contrast, the present document addresses
watershed management processes, which are location- and context-specific and can encompass a
wide variety of decision-making approaches, from statutory to ad hoc, taking place within or
outside of Federal agencies and involving single- or multi-party decisions. These decisions can
be informed by various economic methods, not all of which develop monetary estimates.
Therefore, the present document serves a different purpose and audience. While it includes some
methods that monetize ecological costs and benefits, it is not limited to them.
1-9
-------
1.4.3 USEP A Framework for Economic Assessment of Ecological Benefits
This recently-developed Framework1 deals specifically with the problems of integrating
ERA and economic analysis, and in this context it is a valuable companion reference to the
present work. Like the present document, it provides information about ERA and economic
analysis to a multidisciplinary audience, and it discusses integration approaches. Like the
Guidelines for Preparing Economic Analyses, however, it is limited to the development of
monetary estimates as needed to support policy or regulation. Unlike the present work, it does
not address place-based management processes, and it does not evaluate case examples.
1.5 LIMITATIONS
1.5.1 Lack of complete integration
Although the subject of this document is the integration of ERA and economic analysis,
the case studies that it presents are not integrated in a complete or ideal sense. On one hand, the
efforts invested by USEPA and its partners to conduct W-ERA in a set of U.S. watersheds
offered a unique opportunity to sponsor complementary research in economic analysis. The
assistance award criteria ensured that the funded economic research would focus on key
elements of a W-ERA. Yet, as explained above, the economic studies were initiated several years
later than the W-ERA studies. There was collaboration between members of the W-ERA and
economic teams in each watershed, but because of the later starting point and separate funding
mechanism of the economic research, the teams were not unified. Further, the conceptual
approach for integration described in Chapter 3 was designed as an outcome of this research and
was not available at the outset. As a result, the initial planning and problem formulation work
conducted in each watershed did not include economists or consider their needs. While the
economic research teams had the benefit of groundwork laid by the W-ERA effort, they
1-10
-------
sometimes perceived watershed needs and goals differently, and some of these differences are
evident in this report. Finally, because of the difficulties involved in funding, coordinating and
completing large, multi-participant studies, the W-ERA studies themselves were not all
completed during the time frame of economic study, and in this manner as well the economists
did not obtain the full benefit of interdisciplinary collaboration. Therefore, these case studies
should be seen as providing a unique set of insights into the ERA-economic integration problem
but not as exemplars of such integration,
1.5.2 Specificity to a watershed context
The impetus for this research is the protection of aquatic ecological resources, which
often requires analysis at the level of the watershed. The problem of integrating ERA and
economic analysis for environmental management in general has many facets, not all of which
can be addressed in the watershed context. W-ERA tends to be resource-based; that is, it
identifies the ecological resources of concern in a given place and identifies the risks to those
resources. Economic analysis that is done in conjunction with W-ERA must address those risks,
and the particulars of the local decision context. By contrast, policies or regulations promulgated
at the national or state level (e.g., WQS, effluent guidelines) tend to address stressors or
categories of polluting activities occurring over a broad area, and therefore their risk and
economic assessments may have a different character. Furthermore, some integrated
assessments are for the purpose of setting priorities among different kinds of environmental
problems across different resources, stressors or media. Therefore, while the findings of this
document shed light on the overall integration problem, they should not be considered generally
applicable.
1-11
-------
1.6 UNIQUE CONTRIBUTIONS
Notwithstanding these limitations, this document makes several unique contributions for
environmental management. First, it places economic analysis into a context that is familiar to
risk assessors. Because it uses the specific procedures and terminology of ERA, it will help ERA
practitioners better understand how those procedures can be integrated with economic analysis.
The conceptual approach presented in Chapter 3 borrows heavily from USEPA's ERA
Framework. The case studies demonstrate how risk assessment outcomes - i.e., probabilities of
adverse changes in ecological assessment endpoints - figure into economic analysis, and they
sensitize the reader to the difficulties that economists face in using those results. They also
illustrate for risk assessors the importance of the "with-without" context that is familiar to
economists. Whereas risk assessors sometimes focus only on identifying risks associated with
current situations and trends, or on identifying exposure targets for reducing those risks,
economists most often focus on choices between alternative actions. Therefore, economists
demand a comparison of current and future risks "with and without" a given action. The
economist's perspective, evident both in the conceptual approach and the case studies, prods the
risk assessor to use ERA in a way that maximizes its value to decision-makers.
Second, the risk assessment perspective employed in this document also poses interesting
challenges for the economist. Economists sometimes use relatively vague statements about the
ecological improvements expected under a given policy to elicit the monetary amounts
individuals would pay to obtain the policy. ERA, on the other hand, uses best-available data and
methods to quantify the linkages between human activities, the stressors they produce, and the
ensuing effects on particular ecological endpoints. The resulting statements about risk are as
specific as possible about the nature and magnitude of effects expected, but they may also
1-12
-------
include description of uncertainties. Translating these statements into terms amenable to
economic analysis is difficult, as these case studies illustrate, but the challenge must be accepted
if these sciences are to be integrated.18
The document makes a further, useful contribution by allowing comparison of different
integration approaches. Two case studies used surveys to estimate economic values associated
with policies to protect watershed ecological resources, based on the assessment endpoints
identified in the W-ERA. One of these (see Chapter 4) valued those policies explicitly, using the
contingent valuation method, whereas another (see Chapter 5) did so implicitly, using conjoint
analysis (Appendix 2-A compares these methods). The third case study (see Chapter 6) used
economic game theory to identify policies most likely to resolve a longstanding conflict over the
protection of watershed resources. These differences in approach make the overall findings of
this document more robust.
Finally, this document introduces a conceptual approach for integrating ERA and
economic analysis, in the context of watershed management as practiced under the Clean Water
Act (see Chapter 3, and especially Figure 3-1). The approach draws its elements from existing
USEPA guidance, as well as from other environmental management frameworks developed by
various agencies and advisory bodies. By synthesizing these elements in a way that emulates yet
expands the ERA Framework, which is a familiar tool in the field of environmental management,
it communicates the essential principles of integration to an important audience.
1.7 REFERENCES
1. Norgaard, R., The case for methodological pluralism, Ecological Economics, 1, 37,1989.
1-13
-------
2. Shogren, J.F. and Nowell, C., Economics and ecology: a comparison of experimental
methodologies and philosophies, Ecological Economics, 5,101,1992.
3, Suter, G.W. et al., Ecological Risk Assessment for Contaminated Sites, Lewis Publishers,
Boca Raton, FL, 2000.
4. USEPA, Guidelines for ecological risk assessment, EPA/630/R-95/002F, Risk
Assessment Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
5. CENR, Ecological Risk Assessment in the Federal Government, CENR/5-99/001,
Committee on Environment and Natural Resources of the National Science and
Technology Council, Washington, DC, 1999.
6. Butcher, J.B. et al., Watershed level aquatic ecosystem protection: Value added of
ecological risk assessment approach, Project No. 93-IRM-4(a), Water Environment
Research Foundation, Alexandria, VA., 1997, 342 pp.
7. Diamond, J.M. and Serveiss, V.B., Identifying sources of stress to native aquatic fauna
using a watershed ecological risk assessment framework, Environmental Science and
Technology, 35,4711,2001.
1-14
-------
8. USEPA, Waquoit Bay watershed ecological risk assessment: The effect of land derived
nitrogen loads on estuarine eutrophication, EPA/600/R-02/079, U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC, 2002.
9. USEPA, Clinch and Powell Valley watershed ecological risk assessment, EPA/600/R-
01/050, U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment, Washington, DC, 2002.
10. Serveiss, V.B., Applying ecological risk principles to watershed assessment and
management, Environmental Management, 29, 145, 2002.
11. USEPA, Ecological Risk Assessment for the Middle Snake River, Idaho, EPA/600/R-
01/017, U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment, Washington, DC, 2002.
12. Valiela, I. et al., Producing sustainability: management and risk assessment of land-
derived nitrogen loads to shallow estuaries, Ecological Applications, 10,1006, 2000.
13, Serveiss, V., Norton, S., and Norton, D., Watershed ecological risk assessment, The
Watershed Academy, US EPA, 2000, on-line training module at
http ://www. epa. gov/owow/watershed/wacadem y/acad2000/ecorisk.
1-15
-------
14. USEPA, Framework for ecological risk assessment, EPA/630/R-92/001, Risk
Assessment Forum, U. S. Environmental Protection Agency, Washington, DC, 1992,
15. USEPA, Planning for Ecological Risk Assessment: Developing Management Objectives.
External Review Draft, EPA/630/R-01/001 A, Risk Assessment Forum, Office of
Research and Development, U.S. Environmental Protection Agency, Washington, DC,
2001.
16. USEPA, Guidelines for Preparing Economic Analyses, EPA-240-R-00-003, Prepared by
the National Center for Environmental Economics, 2000.
17. USEPA, A framework for the economic assessment of ecological benefits, Science
Policy Council, U.S. Environmental Protection Agency, Washington, DC, Feb. 1,2002.
18. Suter, G.W., Adapting ecological risk assessment for ecosystem valuation, Ecological
Economics, 14, 137,1995.
1-16
-------
2. BACKGROUND: ECOLOGICAL RISK ASSESSMENT AND ECONOMIC
ANALYSIS IN WATERSHEDS AND THE NEED FOR INTEGRATION
This document presents a conceptual approach and three case studies for the improved
integration of ecological risk assessment (ERA) and economic analysis in the management of
watersheds. This chapter lays necessary groundwork for the technically trained reader who may
not have a background in ERA or in economic analysis. It explains the basic elements of each
and their uses in watershed management, and helps the reader understand their uses in the case
studies.
Readers already familiar with the U.S. Environmental Protection Agency's (USEPA's)
Guidelines for Ecological Risk Assessment can safely skip Section 2.1.1, which summarizes the
steps of ERA, but should read Sections 2.1.2 and 2.1.3, on its critiques and watershed
applications, respectively. Similarly, readers acquainted with environmental economics need not
read Sections 2.2.1 through 2.2.4, which cover familiar theory and applications, but they may
want to read Sections 2.2.5, on game theory, and 2.2.6, on ecological economics. Section 2.3
discusses applications of ERA and economics to water quality standards (WQS) programs in the
U.S., and Section 2,4 offers concluding thoughts on the need for ERA-economic integration.
2.1 ECOLOGICAL RISK ASSESSMENT
This section discusses ERA and its relationship to watershed management. The goal is to
provide sufficient background to make the succeeding chapters understandable to non-
practitioners of ERA; it is not a comprehensive introduction to the topic. First, the origins of risk
assessment and ERA in particular are briefly discussed, and the steps of ERA are presented as
described in the USEPA's Guidelines for Ecological Risk Assessment.1 Some criticisms of ERA
2-1
-------
are then discussed, and finally some applications of ERA to the analysis and management of
environmental problems at the watershed scale are covered.
2.1.1 Framework and methods for ecological risk assessment
The U.S. Council on Environmental Quality has defined risk as "the possibility of
suffering harm from a hazard" where a hazard is "a substance or action that can cause harm"
and risk assessment as "the technical assessment of the nature and magnitude of risk." The
Presidential/Congressional Commission on Risk Assessment and Risk Management defined risk
as "the probability of a specific outcome, generally adverse, given a particular set of conditions"
and risk assessment as "an organized process ... to describe and estimate the likelihood of
adverse health outcomes.. .."3 Risk assessment thus includes both qualitative description (i.e.,
the "nature" of a possible "harm") and quantitation (i.e., of its "magnitude"). "Magnitude" can
apply both to the harmful effect itself (e.g., how many individuals or populations will be harmed,
and to what degree) and to the possibility that the harm will occur. "Possibility" encompasses
the concepts of probability (or likelihood) and uncertainty. In common usage the term "risk"
often equates to likelihood, but in risk assessment a naked probability has little meaning apart
from a qualitative and quantitative description of the probable harm and of the uncertainty
associated with both the harm and its probability. This document uses the term "adverse effects"
rather than "harm," and it uses "risk" to encompass the nature, probability and uncertainty of
adverse effects.
The terms "probability" and "uncertainty" are closely related. "Uncertainty with respect
to natural phenomena means that an outcome is unknown or not established and is therefore in
question."4 Uncertainty that is attributable to natural variability ("inherent uncertainty") is
considered irreducible and often is described using probability distributions. Uncertainty that is
2-2
-------
due to incomplete knowledge ("knowledge uncertainty") is considered reducible given additional
information,4'5*8
ERA is a scientifically-based process for framing and analyzing human-caused risks to
ecological resources,1'6"8 In some of its elements it follows a framework defined earlier for
human health risk assessment,9 but it differs because of special problems presented in the
assessment of ecological risks. The definition of "human health" is not especially problematic
for health risk assessors, and the general public places a high value on "human health" protection
measures (even if there is sometimes debate about what those measures should be).b Assessment
of risks vis-a-vis "human health" is therefore both scientifically meaningful and socially
relevant. Some ecologists have defined a parallel concept of "ecosystem health,"10'11 but the
appropriateness of this concept and the means to define and measure it are controversial among
ecologists,12"15 and there is no consensus among the general public about what constitutes
ecological health or in which instances, or in what forms, it must be preserved.16
2.1.1.1 Planning
Lacking such a clearly-defined reference point, ERA calls for an initial planning step that
includes the explicit establishment of ecosystem management goals (Figure 2-1).1 The planning
process is a dialogue between risk assessors and risk managers and, where appropriate, interested
and affected parties ("stakeholders"), to determine the goals and scope of the assessment.
However, according to USEPA,1 planning should be separated from the scientific conduct of the
8 By some definitions, inherent uncertainty is termed variability, and the term uncertainty is reserved for knowledge
uncertainty.126
While the World Health Organization has defined human health broadly as "a state of complete physical, mental
and social well-being and not merely the absence of disease or infirmity," health risk assessment as practiced by
environmental agencies is concerned only with hazards causing "damage," "injury" or "harm."2'3'12' "Human
health" for risk assessors is thus the absence of these adverse conditions.
2-3
-------
Planning
(Risk Assessor/
Risk Manager/-*
Interested Parties
Dialogue)
Ecological Risk Assessment
Characterization
of
Exposure
Characterization
of
Ecological
Effects
'-l
1
V)
Jo
(t>
oT
-o
3
o
V)
Communicating Results
to the Risk Manager
FIGURE 2-1
Framework for Ecological Risk Assessment (from USEPA1)
2-4
-------
risk assessment proper, to "ensure that political and social issues, though helping define the
objectives for the risk assessment, do not bias the scientific evaluation of risk." This separation
is consistent with a principle espoused by the National Research Council (NRC);9 however, its
appropriateness is explored further in Section 2.1.1.5 and Chapter 3.
ERA planners seek agreement on (1) the decision context, (2) management goals and
objectives, and (3) information needs. Characterizing the decision context entails understanding
the decisions faced by officials, groups or citizens regarding an environmental problem, as well
as the public values, the legal, regulatory, and institutional factors, the geographic relationships,
and the available risk management options that make up the context of those decisions. It also
includes identifying risk assessors, risk managers, other specialists, and interested individuals
and groups who should be involved in the planning process. Management goals are "general
statements about the desired condition of ecological values of concern"1 whereas management
objectives are sufficiently specific to allow the development of measures.17 Objectives must
identify "what matters" given the decision context (in other words, what valued ecological
characteristic should be protected), what protection requires, and what level of improvement, or
direction of change, is to be achieved. Examination of informational needs entails determining
whether an ERA is warranted and, if so, its scope, complexity and focus.17 Suppose, for
example, there were concerns over the decline of a sport fishery in a reservoir influenced by
municipal effluents and agriculture. Understanding the decision context may require listing the
potential regulatory or restorative actions that could be taken by officials, farmers, reservoir
users and other citizens throughout the watershed; involving individuals representing each of
those groups; and appreciating the values and the legal and economic interests held within each
2-5
-------
group. The management goal might be to maintain a viable sport fishery in the reservoir, and
objectives might entail a listing of desirable species to be maintained.
2.1.1.2 Problem formulation
USEPA1 defines problem formulation as "a process of generating and evaluating
preliminary hypotheses about why ecological effects have occurred, or may occur, from human
activities." It requires (1) the identification of assessment endpoints, (2) the development of one
or more conceptual models, and (3) the development of an analysis plan. Assessment endpoints
operationalize the valued ecological characteristics identified in the management objectives by,
first, identifying those that are both ecologically relevant and susceptible to human caused
stressors and, next, selecting specific ecological entities, and measurable attributes of those
entities, to embody those valued characteristics in the analysis. For example, if a management
objective was to maintain a viable fishery for a list of popular recreational species, then
assessment endpoints might include population size, mean individual size and recruitment for
those species.
A conceptual model is "a written description and visual representation of predicted
relationships between ecological entities and the stressors to which they may be exposed."1 The
visual representation usually takes the form of a box-and-arrow diagram illustrating
hypothesized relationships between sources (human activities that produce stressors), stressors
(chemical, biological or physical entities that can induce an adverse response), exposure
pathways, and receptors (ecological entities that may be adversely affected). An example is
presented in Chapter 5 (see Figure 5-3). Initial versions of the conceptual model for a complex
problem may be overly detailed; later versions can be simplified to emphasize only those
pathways that figure importantly in the analysis plan.
2-6
-------
The analysis plan identifies those hypotheses8 that are believed to be important
contributors to risk, or that can be feasibly reduced through management efforts. The plan
specifies data needs, data collection methods and methods for analysis of existing or newly
collected data in order to confirm, or quantify, the underlying relationships and estimate risks.
Referring again to the reservoir fishery example, if fishery declines are hypothesized to
result either from low dissolved oxygen concentrations caused by excessive nutrient inputs from
municipal and agricultural sources or from agricultural pesticide use, diagrams (and
accompanying text) would be produced illustrating these hypothesized sources and pathways of
pollutant transport to the lake. The ecological processes specific to each pollutant, nutrient
effects on dissolved oxygen levels, pesticide effects on aquatic food webs, and ultimate effects
on the assessment endpoints would also be diagrammed. Following an evaluation of existing
data, an analysis plan might call for the analysis of data on pesticide use in the watershed,
municipal effluent characteristics, water quality in the lake and its tributaries, and fish
populations.
2.1.1.3 Analysis
Analysis is "a process that examines the two primary components of risk, exposure and
effects, and their relationships between each other and [with] ecosystem characteristics."1
Exposure analysis describes sources of stressors, stressor transport and distribution, and the
extent of contact or co-occurrence between stressors and receptors. Exposure analysis may be
carried out using environmental measurements, computational models, or a combination of these.
The product of exposure analysis is an exposure profile describing the intensity, spatial extent
a Except as otherwise specified, "hypothesis" in this document refers to a "maintained hypothesis," or statement
thought to be true (i.e., an assumption).
2-7
-------
and timing of exposure. In effects analysis, the effects that are thought to be elicited by a
stressor are first identified. Effects of concern are then subjected to an ecological response
analysis, which examines the quantitative relationship between the stressor and the response, the
plausibility that the stressor may cause the response (causality), and links between particular
measures of effect and the assessment endpoints. In the sport fishery example, exposure analysis
would examine the magnitude, timing and spatial dynamics of nutrient inputs; it would also
characterize reductions in dissolved oxygen (DO) concentrations, since low DO constitutes a
secondary stressor potentially affecting the assessment endpoints. Exposure analysis would also
characterize the input, fate and transport and resulting water concentrations of pesticides used in
the watershed. Effects analysis would include a literature analysis to identify the kinds of effects
potentially caused by these stressors and to determine whether exposure-response relationships
had been estimated for the same or phylogenetically similar species. It would also evaluate the
possibility that the primary effects of one of these stressors on the food base are causing
secondary effects in the assessment endpoints. Effects analysis would also examine the relative
timing of exposures and observed effects of concern to determine whether there is a causal
relationship.
2.1.1.4 Risk characterization
Risk characterization is the process of uniting information about exposure and effects, in
order to first estimate and then describe the likelihoods of adverse effects of stressors. Risk
estimates range in sophistication from simple, qualitative risk ratings (e.g., high, medium or
low), used when information is limited, to comparisons of point estimates of exposure and
effective level, to comparisons of probability or frequency distributions of exposure and
response.
2-8
-------
Figure 2-2 illustrates the latter ease. The intensity of exposure to a stressor varies across
an assessed population of individuals, and this variability is expressed as a cumulative frequency
(curve on left). The fraction of individuals in a tested population that responded to a given
intensity of exposure also varied (curve on right). By aligning these curves on the same
exposure axis, it is shown that median exposure is below the median level of sensitivity by a
relatively large margin, and that 90% of individual exposures are below a level that caused a
response in 10% of individuals, albeit by a smaller margin. These data would suggest a very low
level of response is expected in the assessed population, as long as the test population adequately
represents the assessed population.
Risk descriptions that accompany risk estimates should discuss the adequacy and quality
of data on which the assessment is based, the degree and type of uncertainty associated with the
evidence, and the relationship of the evidence to the hypotheses of the risk assessment. For
example, the exposure and response distributions represented in Figure 2-2 may represent
inherent uncertainty, which cannot be further reduced, that is due to variability in the
environmental distribution of the stressor and in the sensitivity of organisms tested. But there
may be knowledge uncertainty associated with the data as well, if the number of exposure
measurements or organisms tested was too low to adequately characterize the variability or if
there were problems or biases associated with those measurements. There may be knowledge
uncertainty concerning whether the response of the wild assessment population is similar to that
of the test population, or whether the duration of the test and the endpoints examined were
sufficient to characterize the possible effects. Risk descriptions should evaluate all lines of
evidence, both supporting and refuting the risk estimates. They should also discuss the extent to
2-9
-------
0.90
S1
-------
which changes predicted in the risk assessment should be termed adverse, including the nature
and intensity of expected effects, their spatial and temporal scale, and the potential of affected
species or ecosystems to recover.
2.1.2 Critiques of ecological risk assessment
Using the steps of planning, problem formulation, analysis and risk characterization,
ERA seeks to provide a concise roadmap for science-based decision support - beginning with an
inclusive, policy-informed discourse, proceeding through a rigorous process of hypothesis
generation, data gathering and evaluation, and leading to a set of carefully delimited statements
about the probabilities of specific, adverse outcomes, to be provided to decision-makers. The
process is intended to be flexible; it can employ tiers of increasing specificity (e.g., from
screening-level to definitive), and sequences can be iterated as needed before proceeding to
subsequent steps (see Figure 2-1).
Nonetheless, ERA has been subject to various criticisms. Some of these pertain to
problems of application, others to methodology, and others to the premises underlying the role of
science in decision-making. Many are centered on the treatment of scientific uncertainty, and
several involve questions of whether science and policy can, or ought to, be separated. It is
important to consider these issues openly when the use of ERA is contemplated for decision
support partly to be aware of the potential for misuse of the ERA process, and partly to
acknowledge concerns that may be held by many stakeholders.
Some critics have charged that ecological risk assessors are prone to a rather sanguine
view of the process, in which long-term laboratory tests of properly chosen sentinel species are
assumed to yield results that are stable and adequately predictive of ecosystem responses (see
Power and McCarty18 and ensuing discussion).19"21 They argue that the variability in stress-
2-11
-------
response among species and among field sites is sometimes ignored, and that biological
regulatory mechanisms operating at the level of the field population or the ecosystem can
confound the conventional interpretation of laboratory test results. These criticisms highlight the
importance of using multiple lines of evidence (e.g., both field and laboratory observations) and
making a full presentation of assumptions and uncertainties when characterizing risk, as called
for in the ERA Guidelines.1
A common mistake in the analysis stage of ERA is ignoring statistical power - i.e., the
probability that a given experiment or monitoring study will detect an effect if it actually
exists.22"24 If hypothesis testing fails to reject the null hypothesis (no significant effect is
detected), statistical power analysis determines the level of confidence that can be placed in the
negative result; when power is low, a greater need for precaution is indicated.
Where the above criticisms pertain largely to ERA methods and applications, more
fundamental issues have also been raised (see especially papers from a symposium held in 1994
entitled "Ecological Risk Assessment: Use, Abuse and Alternatives"25 and calls for the use of
precautionary rather than risk-based approaches [e.g., the .Wingspread Statement on the
Precautionary Principle26]). Critics claim that (a) unintended ecological consequences of past
actions demonstrate that ecosystems are too complex to be predictable under novel conditions,
and (b) in view of these inherent uncertainties, it is immoral to rely upon the results of even a
well-conducted risk assessment if alternative (albeit more costly) courses of action exist that
appear to pose less hazard.27"28 A related argument (see the Wingspread Statement) adds that the
burden of removing uncertainty must lie with the proponent of any potentially risky action rather
than with society at large. These arguments sometimes portray even the unbiased risk assessor
as an enabling participant, who by virtue of his/her expertise lends a cloak of legitimacy to an
2-12
-------
intrinsically unjust process.29 More often, the assessor is portrayed as biased (e.g., holding a
narrowly reductionist worldview or having an organizational conflict of interest) or intentionally
deceptive. In the end, according to this critique, ERA is at best unreliable for decision-making
and at worst a tool to facilitate ecosystem exploitation.
Some of these criticisms pertain to governance structures themselves rather than to ERA
per se. If indeed the validity of the governance structure underlying an environmental
management effort is itself in dispute, then the trust that is necessary for an effective planning
dialogue may be impossible to obtain, and ERA may be ineffective. In most cases, however, if
an effective dialogue as described in the ERA Guidelines can be established, then many of the
practical and fundamental issues that critics raise can be accommodated, even where deep-seated
disagreements exist. As stated above, an effective planning dialogue clarifies the decision
context, including participant values, burden of proof, institutional factors and management
alternatives, and ensures that the assessment is not too narrowly conceived. Organizational
interests and biases can be made clear at this stage as well. The Guidelines also state that the
appropriateness of including stakeholders depends on the circumstances; in some cases, existing
law and policy might narrowly prescribe the terms for conducting an assessment. However, it is
unlikely that such a restriction ever is appropriate for assessment of problems in watersheds,
where there are multiple sources and stressors, a variety of resources to protect, various
regulatory authorities and incentive programs, and a need for broad community support.
In summary, through an inclusive planning dialogue and careful treatment of uncertainty,
an ERA conducted according to the Guidelines can address many of the practical and
philosophical criticisms that have been leveled against risk assessment. Further steps may need
to be considered as well. Whereas the Guidelines argue for a strict delineation of policy and
2-13
-------
science the planning process, where stakeholders may participate, remains "distinct from the
scientific conduct of [the] risk assessment" other scientists have argued that the limits of
science should be acknowledged not only at the planning stage but throughout the assessment.
When risk assessors are forced to make judgments that go beyond the limits of the data, as they
routinely do, they move from the realm of science into what Alvin Weinberg30 has termed "trans-
science." These judgments reflect the knowledge, experience and even cultural values of the
assessor,31 and they cannot, according to Weinberg, be viewed as free of bias. Funtowicz and
Ravetz32' likewise have suggested that as uncertainties, decision stakes and urgency increase,
problem-solving strategies correspondingly must progress from "applied science," to
"professional consultancy," to "post-normal science." Post-normal science does not pretend to
be value-free or ethically neutral, and it makes use of deliberation. The NRC 4 acknowledged
that deliberation, including interested and affected parties, in the problem formulation stage of
risk assessment can elicit insights that would not occur to assessors and managers alone, and they
called for deliberation involving decision-makers and interested and affected parties throughout
the risk assessment process. The participation theme will be discussed further in Chapter 3.
2.1.3 Watershed applications of ecological risk assessment
The use of the watershed as a geographic unit for planning and management of water
supply and flood control in the U.S. dates to the late 19th and early 20th centuries, but its use for
ecosystem protection is more recent.35 After the formation of the USEPA in 1970, the need for
such an approach grew steadily - as environmental regulatory programs proliferated yet were
spatially uncoordinated and lacked efficient mechanisms for sharing information. Also, during
this period point-source pollution problems were beginning to be solved through the issuance of
discharge permits, bringing to light the less tractable problems of nonpoint sources and habitat
2-14
-------
modification. Finally, in the 1990s environmental groups began to sue the USEPA over its
failure to go beyond the source-by-source issuance of discharge permits, in the thousands of
cases where these had proved insufficient to rectify water quality impairment. Dozens of court
actions, brought under the water quality standards provision of the Clean Water Act of 1972
(CWA), required the States or the USEPA to determine, on a whole-water-body basis, the total
maximum daily load (TMDL) allowable from all sources.
For these reasons, in the 1990s the USEPA began to encourage the use of a "watershed
protection approach" (later termed simply the "watershed approach") for evaluating and
managing threats to freshwater and estuarine ecosystems,3 "40 and they defined a framework for
that process (a discussion of this and other frameworks is presented in Chapter 3). This approach
provided an effective way of spatially delimiting ecological resources and the threats to those
resources, engaging stakeholders in protection efforts, and promoting management actions that
were concerted rather than piecemeal. Thus, the watershed protection approach focused on goal-
setting, partnerships and management. Early USEPA guidance on the approach did not describe
a role for ERA; there was an emphasis on procedures for calculating TMDLs,41 but these were
aimed at determining how to meet numeric water quality standards (WQS) rather than at
determining risks per se (see further discussion of WQS in Section 2.3). However, WQS do not
address several aquatic ecological problems, including those due to hydrologic modification
(e.g., water withdrawal, flow control, or development-related changes in runoff and recharge
patterns), stream channel modification, removal of riparian vegetation, and introduction of
normative species. Nor can they address chemicals for which no standards have been defined,
indicate which of several pollutants may be causing an observed impairment, nor indicate
whether a given protective or restorative measure, if implemented, will reduce the pollutant
2-15
-------
successfully. Nor can WQS adequately address problems whose severity is a function of spatial
scale or the interactions of multiple stressors. Even motivated and involved teams of citizen and
governmental partners can fail to achieve ecological improvements when risks in a watershed are
not adequately understood. These are questions ERA is geared to address.
Therefore, ERA has a significant role to play as a tool for watershed management.42 Five
watershed ecological risk assessment (W-ERA) case studies were initiated by USEPA in
199343'44 and results for several of these recently have been published.42"45"49 The case studies
were initiated to evaluate the feasibility of applying the ERA process to the complex context of
watershed management. Watersheds were selected for study on the basis of data availability,
identification of local participants, diversity of stressors, and significant and unique ecological
resources. The watersheds selected were the Big Darby Creek in central Ohio, the Clinch River
Valley in southwest Virginia and northeast Tennessee, the Platte River watershed in Colorado,
Wyoming and Nebraska, with special emphasis on the Big Bend Reach in south central
Nebraska, the Middle Snake River in south central Idaho, and Waquoit Bay on the southern
shore of Cape Cod in Massachusetts (Figure 1-1). These watersheds comprised different surface
water types, stressors, scales, management problems, socioeconomic circumstances, and regions.
An initial review of progress of these assessments through the problem formulation
stage44'50 found that ERA provided formal and scientifically defensible methods that were a
useful contribution to a watershed management approach. They also found that the analyses in
these five cases had not been as strongly linked to watershed management efforts as would be
desired. However, subsequent experiences from these assessments have suggested that
following W-ERA principles increases the likelihood that environmental monitoring and
assessment data are considered in decision-making. 2>51'52 The three major principles that proved
2-16
-------
most beneficial were (1) holding regular meetings between scientists and managers to establish
assessment goals and to share interim findings that could be of immediate value to managers, (2)
using assessment endpoints and conceptual models to understand and communicate cascading
effects and identify the most significant ecological concerns, and (3) combining data from many
sources into an overall analytic framework, within which multiple stressor analysis is made
feasible.42 Later chapters of this report will present the findings of economic studies that were
funded in 1999 in three of those watersheds in order to further utilize the ERA results and extend
their value for decision-making.
2.2 ECONOMIC ANALYSIS
This section discusses economic analysis in relationship to watershed management. As
with the preceding discussion of ERA, the goal is to provide sufficient background to make the
succeeding chapters understandable to the non-economist, rather than to provide a
comprehensive introduction to the topic. First, it describes welfare economics as the foundation
of environmental and natural resource economics, and the related concept of economic value.
Next, this section introduces some tools that are used for the valuation of environmental goods
and services, and some watershed-related applications of those tools. Then it introduces game
theory, a set of approaches for modeling decisions that are based on economic theory. Finally, it
discusses ecological economics, an emerging field that has criticized the mainstream economic
paradigm, and its potential contribution to the practice of watershed analysis.
2.2.1 Welfare economics
Economists study the allocation of scarce resources across competing uses. Like time
and money, the allocation of environmental goods and services entails important choices,
because all wants cannot be satisfied.
2-17
-------
Welfare economics is the study of agents who are making choices, under the given
assumption that they are trying to maximize their well-being (i.e., their welfare or satisfaction,
also termed utility). Economists focus on choices made by agents such as individuals or firms.
They assume individuals are rational that is, they make choices that maximize their well-being
subject to constraints on time and money and that firms maximize profits subject to technology
or resources. These decisions are examined through marginal analysis that is, by determining
how beneficial or costly one additional unit of a good or service would be to the agent.
In an ideal market, agents' decisions will lead to an efficient outcome, or one in which all
mutually beneficial trades have been made. In other words, under conditions of economic
efficiency, also termed Pareto efficiency, the distribution of resources among agents is such that
no one can be made better off without making someone else worse off.a Rarely, however, do
markets achieve efficient outcomes for environmental goods and services.53 More often,
characteristics of the market or of the goods and services make trade in the marketplace difficult.
Economists describe these as situations of market failure, and they may attempt to identify social
arrangements, including policies and institutions, for adjusting the distribution of resources in
order to improve efficiency.
Aquatic ecosystems provide many goods and services to humans (Table 2-1). Some of
these, like hydropower or bottled water, are traded in markets, yet imperfections in these markets
may lead to inefficiency and degradation. Others, including public goods such as recreational
fishing sites and ecological services such as aesthetics or groundwater recharge, may lack
markets entirely; economists refer to these as nonmarket goods and services.
' It should be noted that efficient outcomes are not always fair. The concept of equity is discussed in Section 2.2.4.
2-18
-------
TABLE 2-1
Daily's classification of ecosystem services with illustrative examples
Production of Goods
Regeneration Processes
Stabilizing Processes
Life-Fulfilling Functions
Preservation of Options
food (terrestrial animal and plant products, forage, seafood, spice)
pharmaceuticals (medicinal products, precursors to synthetics)
durable materials (natural fiber, timber)
« energy (biomass fuels, low-sediment water for hydropower)
industrial products (waxes, oils, fragrances, dyes, latex, rubber, etc., precursors to many synthetic products)
genetic resources (intermediate goods that enhance production of other goods)
cycling and filtration processes (waste detoxification and decomposition; soil fertility generation and renewal; air and water
purification
translocation processes (dispersal of seeds necessary for revegetatiort; pollination of crops and natural vegetation
coastal and river channel stability
compensation of one species for another under varying conditions
* control of the majority of potential pest species
moderation of weather extremes (such as of temperature and wind)
* partial climate stabilization
hydrological cycle regulation (mitigation of floods and droughts)
cultural, intellectual, and spiritual inspiration
aesthetic beauty .
existence value
- * scientific discovery
serenity
* maintenance of the ecological components and systems needed for future supply of these goods and services and others
awaiting discovery
K>
(Adapted from Daily, GC, Environ. Sci. and Policy, 3,333,2000 and as cited in USEPA, Planning for Ecological Risk Assessment: Developing Management Objectives, External Review Draft,
EPA/630/R-01/001A, June 2001.)
-------
Further inefficiencies in the market exist because aquatic ecosystems have been used as
waste receptacles; third parties are "external" to these market transactions, although they are
affected by them. Consider, for example, pollutant discharges by a firm into a river that is used
by downstream households for recreation; regular markets provide no mechanism to compensate
these third parties for the effects of these "externalities" and are therefore inefficient.
A final type of market failure occurs when the economic agents have incomplete
information, or differing information, about a good or service.54 Information may be incomplete
because not all the relationships within an aquatic ecosystem are fully known; for example,
decisions to pollute or to develop may be made without full understanding of the consequences.54
Asymmetric information may lead to strategic interaction among those involved, rather than
straightforward responses based on supply and demand,55"56
Recognition of these kinds of market failure has led to the development of natural
resource and environmental economics as specialized sub fields of welfare economics. Natural
resource economics examines the optimal allocation of scarce resources over time, including
both nonrenewable resources (e.g., minerals) and renewable resources (e.g., fisheries and water
resources).53 Environmental economics tends to focus on two main issues: regulating pollution
or damages as an externality, and valuing nonmarket goods.57
2.2.2 Economic value
co
At this point it is necessary to provide a clear definition of economic value. Freeman
defines economic value within the welfare economic framework. Because each individual is
considered to know how well off he or she is in a given situation, and each individual's well-
being depends on both private and public goods, then economic value of any particular good
should be based on the associated changes to individuals' well-being. In some cases, markets
2-20
-------
help define economic value, but in the absence of markets, or in cases of market failure, other
techniques are needed.
hi either situation, economic value is defined as the maximum of something someone is
willing to give up to get something else.58 It does not need to be measured in dollars (e.g., an
individual may be willing to give up the usefulness of a dam to obtain an increase in water
quality and better fishing), but the dollar metric allows economists to compare trade-offs to all
other goods. Willingness to pay (WTP) is a monetary measure of a welfare change or economic
value; it is the maximum amount a consumer would pay in order to obtain or avoid a particular
change. An alternative measure to WTP is willingness to accept (WTA), the minimum amount
of money an individual is willing to take to give up some change. Both WTP and WTA measure
value, but they are likely to differ for a number of reasons.58"* For example, they use different
starting points for the initial levels of well-being (for an improvement, WTP is measured by
starting at the individual's level of well-being before the improvement and WTA is calculated by
starting at the individual's level of well-being after the improvement). Also, WTP is constrained
by income while WTA has no upper constraint. Economists typically use WTP to value benefits
because it is easier to estimate.59
Economic value for environmental goods and services has been separated into use and
nonuse value. Use value applies when people get some satisfaction from personal utilization of
environmental goods and services; use can be direct or indirect use. An example of direct use is
enjoying the woods while hiking. To one who enjoys fishing for smallmouth bass, indirect use
may mean valuing crayfish because smallmouth bass eat them. The idea of nonuse value, first
introduced by Krutilla,61 comes from the notion that individuals can value environmental goods
and services regardless of whether they use the resource. For example, individuals in the U.S.
2-21
-------
are willing to devote resources to protecting Brazilian habitat for the endangered, golden
tamarind monkey, even though they do not ever expect to visit the area or to see the species. The
total economic value for a nonmarket good or service is the aggregate of these categories of
values.
Economists have developed a variety of methods for estimating nonmarket values.58 The
methods can be categorized according to how the data are generated (based on observed or
hypothetical behavior).62 Observed-behavior approaches, referred to as revealed preference
methods, infer values from data on actual market choices related to the public good. Table 2-2
briefly describes four revealed preference approaches. Revealed preference approaches require
market data, which limits the kinds of environmental goods that can be valued. The assumptions
on which these approaches rely also affect the results. The hedonic price method, which
examines the effect of differences in environmental quality on, for example, housing or job
markets (Table 2-2), assumes that all buyers in the market perceive these environmental
characteristics.
Hypothetical approaches, called stated preference methods, use data generated by placing
individuals in hypothetical choice settings. These methods are needed when no behavior can be
observed (or no other market data exist to infer value), such as to estimate nonuse values or to
value changes that have not yet occurred. These approaches typically use surveys that determine
WTP or WTA; Table 2-2 describes two such approaches. Stated preference methods typically
require more time and cost to develop and implement than revealed preference approaches, and
can be subject to bias. These biases can create uncertainty about whether respondents would
actually pay the amounts they indicate.
2-22
-------
TABLE 2-2
Methods for estimating values of environmental goods and services
Method
Description
Examples
Revealed preference methods (can estimate use values only)
Market
When environmental goods are
traded in markets, their value
can be estimated from
transactions.
The benefits of an oil spill cleanup that would
result in restoration of a commercial fishery can
be projected from changes in markets for fish,
before and after the spill, and their effects on
fishermen and consumers.
Production
function
The value of an environmental
good or service can be
estimated when it is needed to
produce a market good.
If an improvement in air quality would lead to
healthier crops, the value of the improvement
includes, e.g., the reduction in fertilizer costs to
produce the same amount of agricultural crops.
Hedonic
price
method
The value of environmental
characteristics can be
indirectly estimated from the
market, when market goods are
affected by the characteristics.
If an improvement in air quality improves a
regional housing market, its value includes
increases in housing value, which can be
measured by statistically estimating the
relationship between house prices and air
quality.
Travel cost
method
The value of recreational sites
can be estimated by examining
travel costs and time.
The value of a recreational fishing site to those
who use it can be estimated by surveying
visitors, to determine the relationship between
the number of visits and the costs of time and
travel.
Stated preference methods (can estimate both use and nonuse values)
Contingent
valuation
method
Individuals are surveyed
regarding their willingness to
pay for a specifically described
nonmarket good.
In a telephone survey, respondents are directly
asked their willingness to pay, via a hypothetical
tax increase, for a project that would reduce
runoff, improving the health of a particular
stream.
Conjoint
analysis
Survey respondents evaluate
alternative descriptions of
goods as a function of their
characteristics, so the
characteristics can be valued.
In a mail survey, hypothetical alternative
recreational fishing sites are described by type
offish, expected catch rate, expected crowding
and round-trip distance; respondents'
preferences are used to calculate value for
changes in each of the characteristics.
2-23
-------
Benefit transfer is an alternative to either stated or revealed preference methods. This
method estimates the value of environmental goods and services by transferring the results of
previous studies at different locations.64 For example, the value of clean water in Ohio could be
approximated using a number of different studies that estimate the value of reducing nutrients in
Pennsylvania waterways. Like stated preference methods, it can be used in the absence of
market data, but it is less expensive to implement. However, many factors need consideration to
determine whether benefit transfer will provide adequate information.59
To summarize, the choice of valuation technique depends on the values individuals have
for the good or service (i.e., use and nonuse), the availability of appropriate data, the researcher's
constraints (e.g., time and money), and the ability to minimiEe biases. For more detail on
revealed preference methods, stated preference methods and benefit transfer approaches (such as
CQ £'1
the theory, analysis and steps), the reader is referred to Freeman, Hanley and Spash and
Desvousges et al.65 For additional information on estimating ecological benefits, the reader
should see USEPA.59'66
Two of the case studies presented in later chapters of this document used stated
preference techniques (Table 2-2). Chapter 4 explores the use of a contingent valuation method
(CVM) model to value alternative development approaches in the Big Darby Creek watershed of
central Ohio, and Chapter 5 presents a study of the use of conjoint analysis (CA) to study social
trade-offs among development policies in the Clinch Valley of southwestern Virginia and
northeastern Tennessee. To prepare the reader unfamiliar with those methods, Appendix 2-A
discusses their differences more at length.
2-24
-------
2.2,3 Cost-benefit analysis
Cost-benefit analysis (CB A) is the process of summing the value of the individual
welfare changes, present and future, associated with a project or policy. The purpose is to assess
all changes that can be feasibly measured to determine whether society gains more than it loses.
If the benefits exceed the costs so that the gainers could potentially compensate the losers this
is termed the potential Pareto criterion the project or policy is said to improve efficiency. '
Under this criterion, it is considered irrelevant whether compensation actually occurs. The
procedure may be used prospectively, in planning, or retrospectively, to determine if planned
goals were met. CBA was originally developed to assess the net economic value of public works
projects, the outputs of which usually were market goods, and the goal of which was to produce
CO y;a
net social benefit. ' Some of the earliest examples of its use were for water resource projects
in the U.S.,63'67 so the relationship between CBA and watershed management is longstanding.
Hanley and Spash63 describe eight stages of CBA (Table 2-3). The first stage defines
what is to be analyzed, to reveal how the project or policy will cause change. The next stages
identify the relevant impacts and their physical characteristics, including applicable time
horizons, as necessary for economic comparison. For example, if stream restoration is
undertaken to improve stream ecological communities, then the time necessary to plant the
riparian zone; the duration of required maintenance; the lag period for fish population response;
and the type and magnitude of the response need to be determined. The process of economic
valuation is next. Its purpose is to express all changes in the common metric of dollars. Where
market prices of goods and service do not exist, or do not capture the full value, corrected or
"shadow" prices are calculated, as further discussed below. Negative effects of the project are
2-25
-------
TABLE 2-3
Structure of a cost-benefit analysis
1. Definition of project/policy alternatives
2. Identification of project/policy impacts
3. Which impacts are economically relevant?
4, Physical quantification of relevant impacts
5. Monetary valuation of relevant impacts
6. Discounting of costs and benefit flows
7. Applying the net present value test
8. Sensitivity analysis
Source: Hanley and Spash1
63
estimated as opportunity costs, or the lost value of a resource that cannot be used because of the
project.5 ' For example, if a firm chooses to pollute a river, an opportunity cost might be the
lost value of recreational fishing. The sixth step, discounting of cost and benefit flows, is
necessary when benefits and costs occur at different times, to translate all values into present
value. Present values can be compared; if the net present value is greater than zero the project or
policy is said to improve efficiency. If more than one project or policy is being compared, the
one with the largest net present value is said to be the most efficient or provide the largest
improvement in social welfare. The final stage is sensitivity analysis, which examines the
uncertainty of the relevant impacts and discount rate.
2.2.4 Complementary analyses
Traditionally, a complete economic analysis is comprised of three techniques: CBA,
economic impact analysis (EIA), and equity assessment.59'69 Where CBA provides information
2-26
-------
about economic efficiency, the other two techniques examine resource distribution. These two
latter types of analysis are briefly discussed in this section, as well as cost-effectiveness analysis
(CEA) and natural resource damage assessment (NRDA) as they relate to CBA.
Tietenberg68 defines impact analysis, whether environmental or economic, as a process to
quantify the consequences of various actions. By this definition, it is similar to CBA and CEA;
however, rather than transforming all changes into a single (dollar) metric, it simply organizes a
large amount of information for decision support. USEPA 9 defines E1A as a process to examine
the distribution of impacts (both positive and negative), usually by examining economic changes
across a variety of economic sectors.
Fair distribution is an important goal in both welfare and ecological economics, and
equity and efficiency are sometimes traded off. Because it relies on the potential Pareto
criterion, CBA is not concerned with whether the potential compensation actually takes place;
therefore a project by which society as a whole benefits may cause transfers of wealth, creating
winners and losers. Equity assessment allows economists to understand changes in distribution
of wealth due to a policy or project. According to USEPA's economic guidance,59 the first step
is to identify potentially-affected subpopulations; next steps may involve determining each
subpopulation's net benefits or the distribution of the net benefits among the subpopulations.
Most often, however, equity has not been a decision criterion in water resource projects, since, as
long as net benefits over society as a whole exceed zero, those subpopulations experiencing
positive net benefits theoretically could compensate the others. Research to investigate how
winners could compensate losers may be needed to better ensure equitable outcomes.70
CEA resembles CBA but considers only costs. It may be used in situations where the
estimation of benefits is infeasible (e.g., because of time or budget constraints) or too
2-27
-------
determination of "substantial" impact be based on the financial burden to affected households
(for a facility that is publicly owned) or to private-sector entities of installing additional pollution
controls; "widespread" impacts are those involving relatively large changes in socioeconomic
conditions throughout a community or surrounding area.1' Conversely, where water quality is
higher than required to meet designated uses, CWA antidegradation provisions prevent the
issuing of any permit that would result in a significant lowering of water quality unless necessary
to allow an "important" economic or social development; an "important" development is one that
would have "significant" and "widespread" impacts if foregone.115
USEPA also performs economic analyses of WQS. Cost analyses of federally
implemented regulations are required under Executive Order 12866116 and the Unfunded
Mandates Reform Act,117 and depending on the magnitude of the Federal action, a CBA may
also be presented. For example, the USEPA performed an economic analysis of the California
Toxics Rule which established numeric water quality criteria for toxic pollutants necessary to
I 1 o
meet the requirements of the CWA. Even in so doing, however, USEPA does not make the
promulgation of its WQS-related rules subject to an economic efficiency test (i.e., a
determination of whether benefits exceed costs), nor have states, tribes, or territories relied on
such a test for WQS. A 1983 proposed revision of WQS regulations that would have allowed
CBA to serve as a basis for changes in designated uses was discarded following public comment.
In spite of previous regulatory language that required states to '"...take into consideration
environmental, technological, social, economic, and institutional factors' in determining the
attainability of standards for any particular water segment," the agency recognized "inherent
difficulties" in balancing costs or benefits with achievement of CWA goals.119 USEPA Interim
Economic Guidance for WQS allows CBA to be presented as part of an economic impact
2-39
-------
analysis for UAA but suggests that the determination for assessing benefits be coordinated with
USEPA regional offices.115
Comprehensive efforts to integrate ecological and economic analyses have been rare due,
in parts to existing policy. In most cases, ecological analyses determine what measures are
protective and physically attainable, and separate economic analyses determine only what is
financially attainable. For example, where designated uses are not being attained, stakeholders
may be engaged in seeking least-cost mechanisms for meeting a TMDL target, but stakeholder
preferences with respect to the ecological or other benefits of attainment normally do not play a
role in identifying the target, or in downgrading the use. However, the NRC23 has criticized this
approach to WQS as "narrowly conceived" and has suggested that a "broadened socioeconomic
benefit-cost framework" be employed for use designation. Novotny et al.1 recommended the
use of CBA in UAA in cases where "the nonmarket impacts (especially on water quality
benefits) are likely to be large or the costs of incremental benefit very large," in spite of a lack of
guarantees that USEPA reviewers would accept such an analysis as persuasive.
Furthermore, stakeholder preferences come into greater play wherever the protection of
water quality is dependent on the integrity of riparian systems and adjacent uplands especially
in headwater systems. The CWA affords little Federal authority for controlling the physical
modification (other than dredging or filling) of streams or riparian systems or for the control of
nonpoint source (NFS) pollution resulting from upland land uses. Headwater systems, including
intermittent or ephemeral streams, while of critical ecological importance,121 are also very
numerous and highly subject to disturbance and may need to be protected through approaches
involving public cooperation and evaluation of benefit. For example, the Kansas Legislature122
has mandated that certain types of low-flow or intermittent streams be entirely exempted from
2-40
-------
CWA requirements except on those stream segments where the economic efficiency of
regulation can first be demonstrated. In Ohio, although the applicability of the CWA to
headwater streams has not been questioned, a need for stakeholder input as to the appropriate
level of protection is acknowledged.123
2.4 THE NEED FOR INTEGRATION
Risk and economics are unavoidably linked. In the post-Silent Spring era, U.S. society
entered into a number of social contracts that arguably combined elements of bold foresight and
naivete - foresight with regard to the importance of reducing ecological risks, but naivete with
regard to scientific nuance and cost. The 1973 Endangered Species Act required Federal
agencies to "insure that any action... is not likely to jeopardize the continued existence of any
endangered species or threatened species..." before the sheer numbers of endangered and
threatened species and their potentially overwhelming habitat protection or restoration costs were
well understood. Consider, for example, the substantial costs and far-reaching social disruption
that would be required to restore some endangered salmon runs in the Pacific northwest.124
Similarly, the 1972 CWA established as a goal "restoring and maintaining the chemical,
physical, and biological integrity of [the] Nation's waters" and called for achieving a level of
water quality that provides for the protection and propagation offish, shellfish, and wildlife, and
recreation in and on the water, "wherever attainable" (33 USC 1251) well before TMDL lawsuits
would require that longstanding water quality impairments be addressed and the lion's share of
blame would shift from big industry and sewage treatment plants to agriculture and urban
sprawl. There is now a wider recognition that reducing ecological risks is quite costly, and that
its costs are paid not only by big, discrete polluters but by society at large.
2-41
-------
Moreover, risks, as humans define them, have an economic dimension. This does not
imply that ERA should be limited by economics, or serve only as input to economic analyses, but
rather that any risk humans can recognize has economic implications. By definition, a risk
entails a probability of an "adverse" effect, or an effect that is contrary to what is desired.
Therefore, risk is defined with respect to human preference. Even in those cases where norms or
standards have been established by statute or regulation, subjective interpretation is often needed.
As stated earlier, it is not possible to precisely define terms such as "integrity" with reference to
ecosystems, and the "attainability" of a level of water quality is usually a function of cost. In
many instances, USEPA regulatory programs have been required to codify a particular
interpretation of these normative terms. Whenever it is allowable and practicable, however,
determining the preferences of interested and affected individuals can be a means to identify the
best alternatives and to ensure broadly based support for management efforts.
Since people's information about risks is usually incomplete, technical information about
risk plays an important role in informing those preferences. Furthermore, the form of the
technical information is critical. Compendia of monitoring data, problem reports or expert
opinions can all prove misleading because they do not provide for the rigorous and systematic
determination of, e.g., objectives, causative agency, and the probabilities and uncertainties
associated with projected outcomes. Risk characterization,, the last step in ERA, links each of
these elements in careful, informative statements. ERA is needed if economic analysis of
complex ecological problems is to be done.
Just as risks have an economic or preferential dimension, so decisions about actions to
reduce risks always entail trade-offs. This interrelationship of information, preferences and
effective management argues for the thorough integration of ERA and economic analysis. It
2-42
-------
should be obvious, furthermore, that an approach in which the disciplines are compartmentalized
rather than integrated will invariably lead to an analysis of poorer quality. Such an approach
would assume that the natural and social sciences do not bring differing lenses to the
understanding of goals and problems, and that the analytical requirements of each are mutually
grasped without difficulty. In fact, the fundamental relationship between the social and natural
sciences has long been a matter of philosophical dispute, and while dialogue between
economists and ecologjsts has dramatically increased in recent years it still must be assumed
that, in any new circumstance, conscious effort will be required to establish mutual
understanding between the disciplines and a concerted approach to environmental problem-
solving.
When is integrated analysis needed? ERA often is needed to determine the likely
ecological responses to proposed management actions, and economic analysis often is needed to
interpret those ecological changes, and other changes, in terms of human well-being - so that
decisions are effective and beneficial. Whenever both ERA and economic analysis are needed to
address a watershed management problem, the analytic processes should be undertaken in an
integrated fashion. Here, the term 'integrated' does not necessarily imply that any distinction
between the respective sciences is erased, or that either loses its essential character. It does
imply that these analytic processes will be mutually informed and fully coordinated. The
alternative, a piecemeal or haphazard process, is unlikely to serve decision-makers or
stakeholders as well.
To accomplish integration in practice, extensive interaction is needed between the
disciplines as well as with others who have relevant knowledge or a stake in solving the problem.
The form these interactions should take will vary according to the circumstances, but experiences
2-43
-------
from the field of environmental management can be drawn upon to identify certain principles,
and sequences of events, that help determine success. Chapter 3 will examine those experiences
and present a conceptual approach for integrating ERA and economic analysis in the
A
management of watersheds.
2.5 REFERENCES
1. USEPA, Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F, Risk
Assessment Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
2. Cohrssen, J. J. and Covello, V. T., Risk Analysis: A Guide to Principles and Methods for
Analyzing Health and Environmental Risks, U.S. Council on Environmental Quality,
Washington, DC, 1989.
3. PCCRARM, Framework for Environmental Health Risk Management,
Presidential/Congressional Commission on Risk Assessment and Risk Management.,
Washington, DC, 1997.
4. NRC, Risk Analysis and Uncertainty in Flood Damage Reduction Studies, National
Research Council, Commission on Geosciences, Environment and Resources, Washington,
DC, 2000.
5. Morgan, M.G. and Henrion, M., Uncertainty; A Guide to Dealing With Uncertainty in
Quantitative Risk and Policy Analysis, Cambridge University Press, Cambridge, UK, 1990.
6. Gentile, J.H. et al., Ecological risk assessment: a scientific perspective, Journal of
Hazardous Materials, 35, 241,1993.
7. Suter, G.W.IL, Ecological Risk Assessment, Lewis, Boca Raton, FL, 1993.
2-44
-------
8. USEPA, Framework for Ecological Risk Assessment, EPA/630/R-92/001, Risk Assessment
Forum, U. S. Environmental Protection Agency, Washington, DC, 1992.
9. NRC, Risk Assessment in the Federal Government: Managing the Process, National
Research Council, National Academy Press, Washington, DC, 1983.
10. Rapport, D.J., What constitutes ecosystem health?, Perspectives in Biology and Medicine,
33,120,1989.
11. Schaeffer, D J., Herricks, E.E., and Kerster, H.W., Ecosystem health I: measuring
ecosystem health, Environmental Management, 12, 445,1988.
12. Simberloff, D., Flagships, umbrellas, and keystones: is single-species management passe in
the landscape era?, Biological Conservation, 83,247, 1998.
13. Suter, G.W.IL, A critique of ecosystem health concepts and indices, Environmental
Toxicology and Chemistry, 12,1533, 1993.
14. Wicklum, D. and Davies, R.W., Ecosystem health and integrity?, Canadian Journal of
Botany, 73, 997,1995.
15. Lackey, R.T., Values, policy and ecosystem health, BioScience, 51, 437, 2001.
16. Hood, R.L., Extreme cases: a strategy for ecological risk assessment in ecosystem health,
Ecosystem Health, 4,152, 1998.
17. USEPA, Planning for Ecological Risk Assessment: Developing Management Objectives.
External Review Draft, EP A/630/R-01/001 A, Risk Assessment Forum, Office of Research
and Development, U.S. Environmental Protection Agency, Washington, DC, 2001.
2-45
-------
18. Power, M. and McCarty, L.S., Fallacies in ecological risk assessment practices,
Environmental Science and Technology, 31, 370A, 1997,
19. Mayer, F. et al, Letter to the editor, Environmental Science and Technology News and
Research Notes, 3, 116A, 1998.
20. Suter, G.W., Letter to the editor, Environmental Science and Technology News and
Research Notes, 3,116A, 1998.
21. Power, M. and McCarty, L.S., Authors' response, Environmental Science and Technology
News and Research Notes, 3,117A, 1998.
22. Peterman, R.M. and M'Gonigle, M., Statistical power analysis and the precautionary
principle, Marine Pollution Bulletin, 24,231, 1992.
23. NRC, Assessing the TMDL Approach to Water Quality Management, National Research
Council, National Academy Press, Washington, DC, 2001.
24, Suter, G.W., Abuse of hypothesis testing statistics in ecological risk assessment, Human
and Ecological Risk Assessment, 2, 331,1996.
'9.5. Mazaika, R., Lackey, R.T., and Friant, S.L., Special issueecological risk assessment: use,
abuse and alternatives, Human and Ecological Risk Assessment, 1, 1995.
26. Montague, P., Headlines: the precautionary principle, Rachel's Environment and Health
Weekly, 586,1998.
27. O'Brien, M.H., Ecological alternatives assessment rather than ecological risk assessment:
considering options, benefits and hazards, Human and Ecological Risk Assessment, 1, 357,
1995.
2-46
-------
28. O'Brien, M.H., Making Better Environmental Decisions, MIT Press, Cambridge, MA,
2000.
29. Pagel, J.E. and O'Brien, M.H., The use of ecological risk assessment to undermine
implementation of good public policy, Human and Ecological Risk Assessment, 2,238,
1996.
30. Weinberg, A.M., Science and its limits: the regulator's dilemma, Issues in Science and
Technology, 59, 1985.
31. Shabman, L. A., Environmental hazards of farming: thinking about the management
challenge, Southern Journal of Agricultural Economics, 22, 11,1990.
32. Funtowicz, S.O. and Ravetz, J.R., Risk management as a posmormal science, Risk
Analysis, 12, 95,1992.
33. Funtowicz, S.O. and Ravetz, J.R., A new scientific methodology for global environmental
issues, in Ecological Economics: The Science and Management ofSustainability, Costanza,
R. Ed., 1991,10,137.
34. NRC, Understanding Risk: Informing Decisions in a Democratic Society, Washington, DC,
1996.
35. Cole, R.A., Feather, T.D., and Letting, P.K., Improving Watershed Planning and
Management Through Integration: A Critical Review of Federal Opportunities, IWR
Report 02-R-6, U.S. Army Corps of Engineers, Institute for Water Resources, Alexandria,
VA,2002.
36. USEPA, The Watershed Protection Approach: an Overview, EPA/503/9-92/002, Office of
Water, U.S. Environmental Protection Agency, Washington, DC, 1991,
2-47
-------
37. USEPA, The Watershed Protection Approach: Annual Report 1992, EPA840-S-93-001,
Office of Water, U.S. Environmental Protection Agency, Washington, DC, 1993.
38. USEPA, Watershed Protection: a Statewide Approach, EPA 841-R-95-004, Office of
Water, U.S. Environmental Protection Agency, Washington DC, 1995.
39. USEPA, Watershed Protection: a Project Focus, EPA 841-R-95-003, Office of Water,
Environmental Protection Agency, Washington DC, 1995.
40. USEPA, Watershed Approach Framework, EPA840-S-96-001, Office of Water,
Environmental Protection Agency, Washington DC, 1996.
41. USEPA, Draft Guidance for Water Quality-Based Decisions: The TMDL Process (Second
Edition), EPA 841-D-99-001, U.S. Environmental Protection Agency, Office of Water,
Washington, DC, 1999.
42. Serveiss, V.B., Applying ecological risk principles to watershed assessment and
management, Environmental Management, 29,145,2002.
43. CENR, Ecological Risk Assessment in the Federal Government, CENR/5-99/001,
Committee on Environment and Natural Resources of the National Science and
Technology Council, Washington, DC, 1999.
44. Butcher, J.B. et al., Watershed Level Aquatic Ecosystem Protection: Value Added of
Ecological Risk Assessment Approach, Project No. 93-IRM-4(a), Water Environment
Research Foundation, Alexandria, VA., 1997, 342 pp.
45. USEPA, Ecological Risk Assessment for the Middle Snake River, Idaho, EPA/600/R-
01/017, U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment, Washington, DC, 2002.
2-48
-------
46. USEPA, Clinch and Powell Valley Watershed Ecological Risk Assessment, EPA/600/R-
01/050, U.S. Environmental Protection Agency, Office of Research and Development,
National Center for Environmental Assessment, Washington, DC, 2002.
47. USEPA, Waquoit Bay Watershed Ecological Risk Assessment: the Effect of Land Derived
Nitrogen Loads on Estuarine Eutrophieation, EPA/600/R-02/079, U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC, 2002.
48. Diamond, J.M. and Serveiss, V.B., Identifying sources of stress to native aquatic fauna
using a watershed ecological risk assessment framework, Environmental Science and
Technology, 35,4711,2001.
49. Valiela, I. et al., Producing sustainability: management and risk assessment of land-derived
nitrogen loads to shallow estuaries, Ecological Applications, 10,1006, 2000.
50. SAB, Advisory on the Problem Formulation Phase of EPA's Watershed Ecological Risk
Assessment Case Studies, SAB-EPEC-ADV-97-001, Science Advisory Board, U.S.
Environmental Protection Agency, Washington, DC, 1997.
51. Diamond, J.M., Bressler, D.W., and Serveiss, V.B., Diagnosing causes of native fish and
mussel species decline in the Clinch and Powell River watershed, Virginia, USA,
Environmental Toxicology and Chemistry, 21,1147, 2002.
52. USEPA, Report on the Watershed Ecological Risk Characterization Workshop,
EPA/600/R-99/111, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC, 2000.
53. Fullerton, D. and Stavins, R., How economists see the environment, in Economics of the
Environment, Robert Stavins Ed., W. W. Norton and Company, NY, 2000.
2-49
-------
54. Hurley, T. and Shogren, J,, Environmental conflicts with asymmetric information: theory
and behavior, in Game Theory and the Environment, Hanley, N. and Folmer, H. Eds.,
Edward Elgar, Northampton, MA, 1998.
55. Folmer, H. and de Zeeuw, A., Game theory in environmental policy analysis, in Handbook
of Environmental and Resource Economics, Jeroen CJ.M.van den Bergh Ed., Edward
Elgar, Northampton, MA, 1999.
56. Varian, H., Microeconomic Analysis, W.W. Norton and Company, NY, 1992.
57. Cropper, M. and Gates, W., Environmental economics: a survey, Journal of Economic
Literature, 30, 675, 1992.
58. Freeman III, A.M., The Measurement of Environmental and Resource Values: Theories
and Methods, Resources for the Future, Washington, DC, 1993.
59. USEPA, Guidelines for Preparing Economic Analyses, EPA-240-R-00-003, Prepared by
the National Center for Environmental Economics, 2000.
60. USEPA, Economic Analysis Guidelines , Federal Register, 65, 53008,2000.
61. Krutilla, J., Conservation reconsidered, American Economic Review, 57, 777,1967.
62. Mitchell, R.C. and Carson, R.T., Using Surveys to Value Public Goods; The Contingent
Valuation Method, Resources for the Future, Washington, D.C., 1989.
63. Hanley, N. and Spash, C.L., Cost-Benefit Analysis and the Environment, Edward Elgar
Publishing Company, Vermont, 1993.
2-50
-------
64. Boyle, K. and Bergstrom, J., Benefit transfer studies: myths, pragmatism, and idealism,
Water Resources Research, 28, 657, 1992.
65. Desvousges, W., Johnson, F.R., and Banzhaf, H.S., Environmental Policy Analysis With
Limited Information, Edward Elgar Publishing, Northampton, MA, 1998.
66, USEPA, A Framework for the Economic Assessment of Ecological Benefits, Science Policy
Council, U.S. Environmental Protection Agency, Washington, DC, Feb. 1, 2002.
67. Dasgupta, A.K. and Pearce, D.W., Cost-Benefit Analysis: The Theory and Practice,
MacMillan Press Ltd., London, 1978.
68. Tietenberg, T., Environmental and Natural Resource Economics, Addison-Wesley
Longman, Inc., New York, 2000, 19.
69. Hanley, N., Cost-benefit analysis of environmental policy and management, in Handbook
of Environmental and Resource Economics, Jeroen CJ.M.van den Bergh Ed., Edward
Elgar Publishing, Inc., MA, 1999.
70. Zilberman, D. and Lipper, L., The economics of water use, in Handbook of Environmental
and Resource Economics, Jeroen CJ.M.van den Bergh Ed., Edward Elgar Publishing, Inc.,
MA, 1999.
71. Mazotta, M., Opaluch, J., and Grigalunas, T., Natural resource damage assessment: the role
of resource restoration, Natural Resources Journal, 34 (Winter), 153, 1994.
72. Kopp, R. and Smith, V., Valuing Natural Assets: The Economics of Natural Resource
Damage Assessment, Resources for the Future, Washington, DC, 1993.
2-51
-------
73. Unsworth, R. and Bishop, R., Assessing natural resource damages using environmental
annuities, Ecological Economics , 11, 35,1994,
74. Perm, T. and Tomasi, T., Environmental assessment: calculating resource restoration for an
oil discharge in Lake Barre, Louisiana, USA, Environmental Management, 29(5), 691,
2002.
75. Gibbons, R., Game Theory for Applied Economists, Princeton University Press, Princeton,
NJ, 1992.
76. Nicholson, W., Microeconomic Theory, The Dryden Press, NY, 1992.
77. Norton, B., Costanza, R., and Bishop, R., The evolution of preferences: why "sovereign"
preferences may not lead to sustainable policies and what to do about it., Ecological
Economics, 24,193, 1998.
78. Frank, R.H., Passions Within Reason: The Strategic Role of the Emotions, Norton, New
York, 1988.
79. Kolstad, C., Environmental Economics, Oxford University Press, New York, NY, 2000.
80. Gowdy, J., Terms and concepts in ecological economics, Wildlife Society Bulletin, 28(1),
26, 2000.
81. Costanza, R. et al., An Introduction to Ecological Economics, St. Lucie Press, Boca Raton,
FL, 1997.
82. Costanza, R., What is ecological economics?, Ecological Economics, 1,1, 1989.
2-52
-------
83. Sahu, N. and Nayak, B., Niche diversification in environmental/ecological economics,
Ecological Economics, 11,9,1994.
84. Daly, H.E., Allocation, distribution, and scale: towards an economics that is efficient, just,
and sustainable, Ecological Economics, 6,185,1992.
85. Turner, R.K., Environmental and ecological economics perspectives, in Handbook of
Environmental and Resource Economics, van den Bergh, J. C. J. M. Ed., Edward Elgar
Publishing, Inc., MA, 1999.
86. Tacconi, L., Biodiversity and Ecological Economics: Participation, Values and Resource
Management, Earthscan Publications Ltd., London, 2000.
87. Toman, M., Pezzey, J., and Krautkraemer, J., Neoclassical economic growth theory and
'sustainability1, in The Handbook of Environmental Economics, Bromley, D. W. Ed.,
Blackwell Publishers, Maiden, MA, 1995.
88. Costanza, R., Embodied energy and economic valuation, Science, 210,1219,1980.
89. Odum, H.T., Environmental Accounting: Emergy and Environmental Decision Making,
John Wiley & Sons, New York, 1996.
90. J0rgensen, S.E., Integration of Ecosystem Theories: A Pattern, 1997.
91. Faber, M., Manstetten, R., and Proops, J., Ecological Economics: Concepts and Methods,
Edward Elgar, Northampton, MA, 1996.
92. Heuttner, D.A., Net energy analysis: an economic assessment, Science, 192, 101,1976.
2-53
-------
93. Shabman, L.A. and Batie, S.S., Economic value of natural coastal wetlands: a critique,
Coastal Zone Management Journal, 4, 231,1978.
94. Rees, W.E. and Wackernagel, M., Ecological footprints and appropriated carrying capacity:
measuring the capital requirements of the human economy, in Investing in Natural Capital:
The Ecological Economics Approach to Sustainability, Jarmson, A. M., Hammer, M.,
Folke, C., and Costanza, R. Eds., Island Press, Washington, B.C., 1994, 362.
95. Folke, C. et al., The ecological footprint concept for sustainable seafood production: a
review, Ecological Applications, 8, S63,1998.
96. Costanza, R., The dynamics of the ecological footprint concept, Ecological Economics, 32,
341,2000.
97. Costanza, R., Wainger, L., and Bockstael, N., Integrated ecological economic systems
modeling: theoretical issues and practical applications, in Integrating Economic and
Ecologic Indicators: Practical Methods for Environmental Policy Analysis, Milon, J. W.
and Shogren, J. F. Eds., Praeger Publishing, Westport, CT, 1995,45.
98. Reyes, E. et al., Integrated ecological economic regional modelling for sustainable
development., in Models of Sustainable Development., Faucheux, S. e. al. Ed., Edward
Elgar, Aldershot, UK., 1996,253.
99. Costanza, R. and Ruth, M., Using dynamic modeling to scope environmental problems and
build consensus., Environmental Management, 22, 183,1998.
100. Geoghegan, J., Wainger, L.A., and Bockstael, N.E., Spatial landscape indices in a hedonic
framework: an ecological economics analysis using GIS, Ecological Economics, 23,251,
1997.
2-54
-------
101. USEPA, Water Quality Standards Handbook, Second Edition, EPA 823-B-94-005a, U.S.
Environmental Protection Agency, Office of Water, Washington, DC, 1994.
102. USEPA, Integrated Planning and Priority Setting in the Clean Water State Revolving Fund
Program, EPA-832-R-01-002, U.S. Environmental Protection Agency, Office of Water,
Washington, DC, 2001.
103. USEPA, Nonpoint Source Program and Grants Guidance for Fiscal Year 1997 and Future
Years, U.S. Environmental Protection Agency, Office of Water, Washington, DC, 1996.
104. USEPA, National Recommended Water Quality Criteria-Correction, EPA 822-Z-99-001,
U.S. Environmental Protection Agency, Office of Water, Washington, DC, 1999.
105. Stephan, C.E. et al., Guidelines for Deriving Numerical National Water Quality Criteria for
the Protection of Aquatic Organisms and Their Uses., NTIS PB85-227049, U.S.
Environmental Protection Agency, Office of Research and Development, Duluth, MN,
1985.
106. USEPA, Final water quality guidance for the Great Lakes System; final rule, Federal
Register, 60,15365,1995.
107. Karr, J.R., Assessment of biotic integrity using fish communities, Fisheries, 6,21,1981.
108, DeShon, I.E., Development and application of the Invertebrate Community Index (ICI), in
Biological assessment and criteria: Tools for water resource planning and decision
making., Davis, W. S. and Simon, T. Eds., Lewis Publishers, Boca Raton, FL, 1995,15,
217.
2-55
-------
109. USEPA, Biological Criteria: Technical Guidance for Streams and Small Rivers. Revised
Edition, EPA 822-B-096-001, U.S. Environmental Protection Agency, Office of Water,
Washington, DC, 1996.
110. USEPA, Ambient Water Quality Criteria Recommendations: Information Supporting the
Development of State and Tribal Nutrient Criteria for Rivers and Streams in Nutrient
Ecoregion VI: Corn Belt and Northern Great Plains, EPA 822-B-00-017, U.S.
Environmental Protection Agency, Office of Water, Washington, DC, 2000.
111. USEPA, Nutrient Criteria Technical Guidance Manual: Rivers and Streams, EPA-822-B-
00-002, U.S. Environmental Protection Agency, Office of Water, Washington, DC, 2000.
112. OEPA, Association Between Nutrients, Habitat, and the Aquatic Biota in Ohio Rivers and
Streams., Technical Bulletin MAS/1999-1 -1, Ohio Environmental Protection Agency,
Division of Surface Waters, Columbus, Ohio, 1999.
113. OEPA, Total Maximum Daily Loads for the Upper Little Miami River. Draft Report.,
Ohio Environmental Protection Agency, Division of Surface Water, Columbus, Ohio,
2001.
114. USEPA, U.S. EPA NPDES Permit Writers' Manual, EPA-833-B-96-003, U.S.
Environmental Protection Agency, Office of Water, Washington, DC, 1996.
115. USEPA, Economic Guidance for Water Quality Standards, U.S.Environmental Protection
Agency, Office of Water, 2001, Available from http://www.epa.gov/ost/econ/.
116. The President, Executive Order 12866 of September 23,1993; Regulatory Planning and
Review, Federal Register, 58,1993.
2-56
-------
117. USEPA, OAQPS Economic Analysis Resource Document, U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Innovative Strategies and
Economics Group, Research Triangle Park, NC, 1999.
118. USEPA, Economic Analysis of the California Toxics Rule, EPA Contract No. 68-C4-0046,
1999.
119. USEPA, Water Quality Standards Regulation, Federal Register, 48, 51400,1983.
120. Novotny, V. et al., A Comprehensive UAA Technical Reference, Project 91-NPS-l, Water
Environment Research Foundation, Alexandria, VA, 1997.
121. Zale, A.V. et al.,. The Physicochemistry, Flora, and Fauna of Intermittent Prairie Streams:
A Review of the Literature, Biological Report 89(5), Fish and Wildlife Service, U.S.
Department of the Interior, Washington, DC, 1989.
122. An Act concerning the waters of the state; relating to classified stream segments and
designated uses of classified stream segments, Kansas Legislature, Substitute Senate Bill
204,2001.
123. OEPA, Fact Sheet: Clean Rivers Spring From Their Source: The Importance &
Management of Headwater Streams, Division of Surface Water, Ohio Environmental
Protection Agency, Columbus, Ohio, 2000.
124. Lackey, R.T., Salmon policy: science, society, restoration, and reality, Environmental
Science & Policy, 2, 369, 1999.
125. Little, D.E., Beyond Positivism: Toward a Methodological Pluralism in the Social
Sciences, Delittle@Umd.Umich.Edu, 2002, Available from http://www-
personaLumd.umich.edu/-delittle/BEYPOSIT.PDF.
2-57
-------
126. USEPA, Guiding Principles for Monte Carlo Analysis, Risk Assessment Forum, U.S.
Environmental Protection Agency, 1997.
127. WHO, Preamble to the Constitution of the World Health Organization, Official Records of
the World Health Organization, no. 2, World Health Organization, Geneva, 1946.
128, USEPA, Development Document for Final Effluent Limitations Guidelines and Standards
for the Pharmaceutical Manufacturing Point Source Category, 2002.
2-58
-------
APPENDIX 2-A
DISCUSSION OF STATED PREFERENCE METHODS USED IN TWO CASE STUDIES
This appendix discusses the differences between two stated preference methods used for
valuing environmental goods, the contingent valuation method (CVM) and conjoint analysis
(CA). CVM was used to value alternative development scenarios in the Big Darby Creek
watershed of central Ohio (Chapter 4), and CA measured the trade-offs among development
policies in the Clinch Valley of southwestern Virginia and northeastern Tennessee (Chapter 5).
CVM measures value directly by asking respondents' their willingness to pay, using a
specified payment vehicle (e.g., a change in the electric hill or in taxes), to avoid or obtain a
particular change. The question format could be open-ended (i.e., how much are you willing to
pay...?) or dichotomous-choice (i.e., would you be willing to pay $X amount: yes or no?).
Mitchell and Carson1 describe CVM as a "versatile tool for directly measuring a range of
benefits for a range of goods consistent with economic theory." Unlike revealed preference
techniques, which are limited to valuing existing goods at existing quantity and quality levels,
CVM can be used to measure both use and nonuse values of goods that may not presently exist.
As a result of compensation claims associated with the Exxon Valdez oil spill in Prince William
Sound, Alaska, the National Oceanic and Atmospheric Administration (NOAA) convened a
panel to conduct hearings on the validity of CVM.2 The panel established rigorous guidelines for
legally admissible studies. Nonetheless, the method remains controversial among some
economists because of its hypothetical nature. Several potential biases have been identified,3'4
and CVM models have had a mixed performance when subjected to internal and external validity
tests.5'6
2-59
-------
Whereas CVM typically measures the value of a good as a whole, CA induces
respondents to evaluate alternatives as a function of their attributes, so that the attributes can be
individually valued.7"9 For example, a respondent may be asked to state a preference (and
perhaps to rate the strength of the preference) between alternative streams for fishing. The
streams are said to vary as to the type of fishing, expected catch rate, expected crowding,
expected weather, and round-trip distance.10 One attribute, in this case driving distance, usually
is either a cost or a proxy for cost to allow estimation of WTP. By choosing one alternative, the
respondent reveals a (strength of) preference for that particular bundle of attribute values vis-a- ,
vis the others presented. By presenting a series of choice sets in which these attribute values are
varied, respondent preferences can be disaggregated and the contribution of each attribute to the
combined preference determined.
Environmental management alternatives and their fiscal, social and ecological results also
occur as bundles in the real world and can be analyzed using CA. The technique has been used
in environmental applications where attributes are cardinal (e.g., travel distance) or class (e.g.,
terrain type) variables associated with the economic and environmental elements of a choice,
such as a choice of recreational opportunity11"13 or electricity generation scenario. 4 If the key
features, both ecological and nonecological, that define each alternative can be expressed by the
selected attributes, then CA can be used to quantify the key sources of stakeholder preference
and to inform the design of an optimal alternative.
The multiattribute choice process employed in CA could avoid or reduce certain biases
associated with the bid process in CVM, especially if the choices presented were meaningful and
plausible to survey respondents.15 Potential difficulties with such an application include: (1) the
difficulty of constructing choice sets that encompass the needed range of potential management
options and outcomes; (2) the potential for confusing or fatiguing respondents if too many
2-60
-------
attributes or choice sets are presented; and (3) a lack of experience in applying the method to
evaluate indirect or nonuse values.
CA is similar to CVM, and therefore some of the same benefits apply, including the
ability to value goods that have not been observed yet (e.g., impacts of global climate change), as
long as they can be described adequately to the respondent. But whereas CVM results apply
only to the scenarios or goods described, CA results can be extrapolated to any good within the
range of attribute values used; even a good that was not specifically tested. It also avoids some
of the problems of dichotomous-choice CVM such as yea-saying (i.e., bias toward agreement).
However, CA has not been subjected to the same scrutiny as CVM, questionnaire design is
difficult, and the optimal design is unsettled.
REFERENCES
1. Mitchell, R.C. and Carson, R.T., Using Surveys to Value Public Goods: The Contingent
Valuation Method, Resources for the Future, Washington, D.C., 1989.
2. Arrow, K.J. et al., Report of the NOAA panel on contingent valuation, Federal Register,
58, 4602, 1993.
3. Diamond, P.A. and Hausman, J.A., Contingent valuation: Is some number better than no
number?, Journal of Economic Perspectives, 8, 45,1994.
4. Desvousges, W.H., Hudson, S.P., and Ruby, M.C., Evaluating CV performance: Separating
the light from the heat, in The Contingent Valuation of Environmental Resources,,
Bjornstad, D. J. and Kahn, J. R. Eds., Edward Elgar, Cheltenham, UK., 1996,117.
2-61
-------
5. Haiuneman, W.M., Valuing the environment through contingent valuation., Journal of
Economic Perspectives, 8,19, 1994.
6. Bjornstad, D. J. and Kahn, J.R., Characteristics of environmental resources and their
relevance for measuring value, in The Contingent Valuation of Environmental Resources:
Methodological Issues and Research Needs, Bjornstad, D. J. and Kahn, J. R. Eds., Edward
Elgar, Cheltenham, UK., 1996, 3.
7. Louviere, J. J,, Conjoint analysis modeling of stated preferences: A review of theory,
methods, recent developments and external validity, Journal of Transport Economics and
Policy, 22, 93, 1988.
8. Louviere, J.J., Relating stated preference methods and models to choices in real markets:
calibration of CV responses, in The Contingent Valuation of Environmental Resources,
Bjornstad, D. J. and Kahn, J. R. Eds., Edward Elgar, Cheltenham, UK., 1996,167.
9. Kahn, J.R., The Economic Approach to Environment and Natural Resources, Harcourt
Brace/Dryden Press, Fort Worth, TX, 1998.
10. Heberling, M., Valuing public goods using the stated choice method, PhD Dissertation
thesis, The Pennsylvania State University, State College, 2000.
11. Adamowicz, W., Louviere, J., and Williams, M., Combining revealed and stated preference
methods for valuing environmental amenities, Journal of Environmental Economics and
Management, 26, 271,1994.
2-62
-------
12. Adamowicz, W. et al,, Perceptions versus objective measures of environmental quality in
combined revealed and stated preference models of environmental valuation, Journal of
Environmental Economics and Management, 32, 65,1997.
13. Roe, B., Boyle, K.J., and Teisl, M.F., Using conjoint analysis to derive estimates of
compensating variation, Journal of Environmental Economics and Management, 31,145,
1996.
14. Johnson, F.R. and Desvousges, W.H., Estimating stated preferences with rated-pair data:
environmental, health and employment effects of energy programs, Journal of
Environmental Economics and Management, 34, 79,1997.
15. Hanley, N., Wright, R.E., and Adamowicz, V., Using choice experiments to value the
environment, Environmental and Resource Economics, 11(3-4), 413, 1998.
2-63
-------
APPENDIX 2-B
USING MULTIMETMC INDICES TO DEFINE THE INTEGRITY OF
STREAM BIOLOGICAL ASSEMBLAGES AND INSTREAM HABITAT
To determine if a stream provides a suitable environment for a robust biological
community, measurement of a set of chemical and physical water quality parameters (e.g.,
toxics, dissolved oxygen, temperature) is not sufficient. A chemical that is present, but not on
the monitoring list, may be affecting the stream community, and episodic exposures, which are
difficult to detect without continuous sampling, can also cause long term effects. Even high
quality water can fail to support robust communities if other factors affect the stream
environment. The physical habitat of the stream may have been altered (such as by
channelization) in a way that removes instream cover or substrate needed by organisms, and
barriers such as low-head dams may prevent migration or recolonization. Changes in stream
hydrology that result from watershed development or flow diversion can create flow conditions
that degrade the instream environment as well.
A goal of the Clean Water Act is "to restore and maintain the chemical, physical, and
biological integrity of the Nation's waters."3 The term "biological integrity" implies a concept of
wholeness that encompasses more than water quality alone. Indeed, to determine whether a
stream biological community is flourishing as expected, it makes sense to measure the
community itself. However, because biological communities are both complex and variable over
space and time, the list of aspects that could be measured is long, and the measurements are not
meaningful without interpretation. To establish an operational definition, various aggregate
indices have been designed that measure selected ecological parameters and express some aspect
'33 USC 1251 (a)
2- 64
-------
of "integrity" (e.g., of the fish or invertebrates assemblages present, or of the instream habitat) on
a simple numerical scale. While the concept itself remains controversial, and some argue that in
the aggregation of measures, information useful for assessment is lost rather than gained, the
approach has gained sufficient acceptance to become widely used in environmental monitoring
and regulation.3 Since watershed ecological risk assessments often must rely on the data that are
available, whether or not they are ideal, their application in stream assessments is common. Four
such indices that are referred to in later chapters of this document are briefly described here.
Methods used for computing an index vary regionally; they are modified to fit regional
ecological conditions. This description relies on methods used by the Ohio Environmental
Protection Agency (OEPA);4 methods applied in other locations, while not identical, are similar.
It should be noted that indices of biotic integrity are not necessarily useful for the study
or management of rare species. Although Karr and Chu state that the explicit inclusion of
threatened or endangered species in an index can improve their management,1 bioassessments
that are conducted for routine monitoring of stream condition may not have the spatial or
temporal intensity needed to detect them. Therefore, the indices, like those used by OEPA, may
not be designed to respond to the presence or absence of rare species. Furthermore, a low score
on one metric that is due to the absence of a rare species could be masked by high scores on
other metrics.
Another potential weakness of integrity indices is that the choice of sampling techniques
may be taxonomically limiting. For example, the Invertebrate Community Index, described
below, relies heavily on artificial substrates and its metrics mainly reflect organisms that
colonize those substrates. As a result, the presence and diversity of noninsect taxa such as
2-65
-------
crustaceans and mollusks, many of which are sensitive to human disturbance (e.g., see Chapters
4 and 5), are poorly reflected in the index.
Index of Biotic Integrity (IBI)
The IBI, originally developed by Karr, expresses the status of the stream fish assemblage
in a given location at the time of sampling. A stream reach of a given length is sampled by
electrofishing techniques, and captured fish are identified to the species level.6 To compute a set
of 12 metrics, species are categorized into various groupings including taxonomic family,
tolerance to pollution, feeding type, breeding type, and whether indigenous or exotic
(Table 2-B-l). Visible skin or subcutaneous disorders are also recorded; these include
deformities, eroded fins, lesions/ulcers and tumors. For each metric, a score of 5,3 or 1 is
assigned according to whether the sample approximates (5), deviates somewhat from (3) or
strongly deviates from (1) the reference value, or that value expected under minimally impacted
conditions. For most metrics, the reference value is scaled according to drainage area (i.e., the
area of the watershed above the point sampled), since fish assemblages in larger streams tend
naturally to be more diverse. The index is a sum of scores of the individual metrics, with a
maximum score of 60. The interquartile range (25th percentile - 75th percentile) of IBI for
wadable, warmwater reference sites in Ohio is 38-50.3'4
Modified Index of Weil-Being (Mlwb)
The Index of Well-Being, developed by Gammon7 and modified by OEPA,4 also
expresses the status offish assemblages. It uses the same sampling data as required for the IBI
but also requires determination of the total weight of each species in the sample. The index is
computed as follows:
" Wadable streams are those that can be sampled by personnel walking in the streams, but do not include headwaters
streams (drainage area < 20 mi2). Warmwater streams, which include most streams in Ohio, are those not capable of
supporting coldwater fauna such as trout.
2-66
-------
TABLE 2-B-l
Individual metrics constituting two indices of biological integrity used by the Ohio
Environmental Protection Agency
Metric #
1
2
3
4
5
6
7
8
9
10
11
12
Index of Biotic Integrity (IBI)a
Total number of indigenous fish
species
Number of darter species (Percidae)
Number of sunfish species
(Centrarchidae)
Number of suclcer species
(Catostomidae)
Number of pollution intolerant species
Percent abundance of tolerant species
Percent abundance of omnivores
Percent abundance of insectivores
Percent abundance of top carnivores
Total number of individuals
Percent lithophils (species requiring
clean gravel/cobble for spawning)
Percent with deformities, eroded fins,
lesions and tumors
Invertebrate Community Index (ICI)
Total number of taxa
Number of mayfly taxa
(Ephemeroptera)
Number of caddisfly taxa (Trichoptera)
Number of true fly taxa (Diptera)
Percent mayflies (Ephemeroptera)
Percent caddisflies (Trichoptera)
Percent Tanytarsini midges
Percent other true flies and non-insects
Percent pollution tolerant organisms
Number of EPTtaxa"
a Metrics listed are for wadable, nonheadwaters sites. For other sites, some metrics differ.
b EPT = Ephemeroptera (mayflies), Plecoptera (caddisflies) and Tricoptera (stoneflies). Index is determined only
from sampling of natural, not artificial, substrates.
2-67
-------
Mlwb = 0.5 In N + 0.5 In B + H (no.) + H (wt.)
where:
N = relative numbers of all species excluding species designated
"highly tolerant"
B = relative weights of all species excluding species designated "highly
tolerant"
H (no.) = Shannon diversity index based on numbers
H (wt.) = Shannon diversity index based on weight
and the Shannon diversity index is computed as:
- N N
where:
n, = number or weight of the ith species
N = total number or weight of the sample
The interquartile range of Mlwb for wadable, warmwater reference sites in Ohio is 8.3-9.4.
Invertebrate Community Index (ICI)
The ICI was developed by DeShon and others to determine the condition of the benthic,
or bottom-dwelling, invertebrate assemblage.4'6'8 Where there is sufficient stream flow, a device
consisting of a series of hardboard plates, spaced along an eyebolt, is submerged in the stream
and allowed to be colonized for a period of six weeks during the summer months. It is then
collected for laboratory enumeration and identification of the attached organisms. To augment
observations from the artificial substrates, a net is used to sample organisms occurring on natural
substrates. Where the artificial substrates cannot be used, the natural substrates are sampled
more extensively. When possible, individuals collected are identified to species, but sometimes
identification is only to the genus or a higher level. As with IBI, species are categorized into
2-68
-------
groups for calculation of the index. The ICI is composed of 10 metrics (Table 2-B-l) that are
scored as either 6,4,2 or 0 according to a relationship that varies with drainage area. These
relationships are more complex than those for fish. For example, diversity of certain groups first
increases and then decreases as drainage area increases. Like the IB I, the highest possible score
is 60. The interquartile range of ICI for reference sites in Ohio where artificial substrates could
be used is 36-4S.4
Qualitative Habitat Evaluation Index (QHEI)
The QHEI evaluates physical characteristics of stream habitats that are important to fish
and invertebrate communities.6'9'10 Six principal metrics compose the index, each having two to
five constituent measures (Table 2-B-2). The metrics describe the material covering the stream
bottom (substrate), areas where fauna can hide (cover), complexity and stability of the stream
channel (channel quality), naturalness and stability of the streamside environment
(riparian/erosion), variety of instream habitat types such as riffles, runs and pools (pool/riffle),
and steepness of the stream in the direction of flow (gradient). The maximum score is 100. The
interquartile range of QHEI for wadable, warmwater reference sites in Ohio is 68-78.
2-69
-------
TABLE 2-B-2
Primary and secondary metrics constituting the Qualitative
Habitat Evaluation Index (QHEI) used by the Ohio Environmental
Protection Agency
Metric
Substrate
Type
Quality
Instream Cover
Type
Amount
Channel Quality
Sinuosity
Development
Channelization
Stability
Riparian/Erosion
Width
Floodplain quality
Bank erosion
Pool/Riffle
Max depth
Current available
Pool morphology
Riffle/run depth
Riffle substrate stability
Riffle embeddedness
Gradient
Total Score
Score
20
0-
-5-
20
-3
20
0 9
1-
11
20
1
1-
1-
1
4
7
6
3
10
0-
0-
1
4
3
3
20
0-
6
_2 4
0-
0-
0-
2
4
2
-1-2
10
100
Source: Rankin
2-70
-------
References
1. Karr, J.R, and Chu, E.W., Restoring Life in Running Waters: Better Biological Monitoring,
Island Press, Washington, B.C., 1999.
2. Suter, G.W.U., Ecological Risk Assessment, Lewis, Boca Raton, FL, 1993.
3. USEPA, Biological Criteria: Technical Guidance for Streams and Small Rivers. Revised
Edition, EPA 822-B-096-001, U.S. Environmental Protection Agency, Office of Water,
Washington, DC, 1996.4. OEPA, Biological Criteria for the Protection of Aquatic Life.
Volume II: Users Manual for Biological Field Assessment of Ohio Surface Waters, WQMA-
SWS-6, Ohio Environmental Protection Agency, Columbus, Ohio, 1987.
5. Karr, J.R., Assessment of biotic integrity using fish communities, Fisheries, 6,21,1981.
6. OEPA, Biological Criteria for the Protection of Aquatic Life. Volume III: Standardized
Biological Field Sampling and Laboratory Methods for Assessing Fish and
Macroinvertebrate Communities, WQMA-SWS-3, Ohio Environmental Protection Agency,
Columbus, Ohio, 1987.
7. Gammon, J.R., The Fish Populations of the Middle 340 Km of the Wabash River, Tech. Rep.
86, Purdue University Water Resources Center, Lafayette, IN, 1976.
8. DeShon, J.E., Development and application of the Invertebrate Community Index (ICI), in
Biological assessment and criteria: Tools for water resource planning and decision making.,
Davis, W. S. and Simon, T. Eds., Lewis Publishers, Boca Raton, FL, 1995, 15, 217.
9. RanMn, E.T., The Qualitative Habitat Evaluation Index (QHEI): Rationale, Methods, and
Application, Ohio Environmental Protection Agency, Columbus, Ohio, 1989.
2-71
-------
10. Rankm, E.T., Habitat indices in water resource quality assessments, in Biological
Assessment and Criteria: Tools for Risk-based Planning and Decision Making, Davis, W: S.
and Simon, T. Eds., Lewis Publishers, Boca Raton, FL, 1995,13,181.
2-72
-------
3. A CONCEPTUAL APPROACH FOR INTEGRATED WATERSHED
MANAGEMENT
In Section 2.4 a rationale was presented for ecological risk assessment (ERA)-economic
integration in watershed management: (a) that both risks and actions to reduce risk have an
economic dimension, because they invoke preferences and trade-offs; (b) that technical
information about risks, as is provided by ERA, is necessary for the formation of informed
preferences; and (c) that the compartmentalization of disciplinary efforts leads to a poorer quality
of analysis. It was recommended that whenever both ERA and economic analysis are needed to
address a watershed management problem, they should be undertaken in an integrated fashion,
which means that they should be mutually informed and fully coordinated. The goal of this
chapter, then, is to develop a generalized, conceptual approach for achieving ERA-economic
integration in a watershed management context. The conceptual approach has a similar form
and purpose as existing frameworks developed by the U.S. Environmental Protection Agency
(USEPA) such as the Framework for Ecological Risk Assessment or the Framework for
Assessment of Ecological Benefits? This work draws from those and other frameworks, but the
term framework is aot used so as to emphasize that it is not intended to replace them.
This chapter first examines existing frameworks that have been used for watershed
management, then considers some guiding principles, and finally presents a new conceptual
approach that incorporates ERA into a well-integrated management process.
3.1 EXISTING FRAMEWORKS FOR WATERSHED MANAGEMENT
Various frameworks, emanating from the fields of risk assessment, environmental
monitoring, project planning, environmental regulation and natural resource management, have
been applied to watershed management processes, but none has addressed specifically the ERA-
3-1
-------
economic integration problem. Review of these frameworks reveals several characteristics by
which they differ, which will be seen later to have bearing on the integration problem. The first
of these is comprehensiveness with respect to the management process. Some frameworks
address only monitoring or assessment, stopping short of decisions, whereas others are for
planning and management as a whole, including decisions (and often, implementation,
evaluation and adaptation). The second has to do with the intended use. Some can be termed
situational, or responding to the advent of a problem or opportunity; others are for ongoing
management and may be termed regular. The third characteristic is disciplinary breadth. Some
frameworks are focused within the natural sciences whereas others emphasize both the natural
and social sciences. The final characteristic is the degree to which the process is open to
stakeholders, ranging from no explicit role to a role that entails negotiation rights. These four
characteristics have been used to create an illustrative typology of some existing frameworks
(Table 3-1). A discussion of each of these frameworks, in relation to the typology, is presented
in Appendix 3-A.
3.2 GUIDING CONSIDERATIONS FOR AN INTERGRATED MANAGEMENT
PROCESS
Given the existing frameworks, what considerations should guide the design (via
borrowing and adaptation) of an approach for ERA-economic integration? According to
USEPA's Science Advisory Board,3 the processes used should have the following
characteristics: they should be transparent (clearly understandable) to all parties; flexibly
applied; dynamic (interconnected and iterative); open and cooperative; informed by many
different sources and disciplines; and they should reflect holistic, systems thinking. Bellamy et
al.4 comment on the tendency for natural resource management efforts to fail to develop clear
goals, achieve an integrated perspective, match actions to objectives, and evaluate outcomes
3-2 '
-------
TABLE 3-1
Typology of frameworks that have been applied to the processes of watershed assessment
and management.2 Bold, bracketed numbers indicate degree of stakeholder integration in
the process;1* italics indicate an emphasis on the integration of natural and social sciences,0
Situational:
For project
design or problem
response
Regular:
For ongoing
management of
watershed
resources
Monitoring and Assessment Planning and Management
EMAP (Environmental
Monitoring and Assessment
Program) indicator design [0]
DPSIR (Driving forces,
Pressures, State, Impacts,
Response) indicator framework
design6'[0]
Guidelines for ecological risk
assessment [I]
Framework for the economic
assessment of ecological
benefits2[l}
Monitoring program with
12
cyclical redesign fO]
Society for Environmental
Toxicology and Chemistry's
ecological risk management
framework [1]
Framework for environmental
health risk management8 [2]
U.S. Army Corps of Engineers
project planning [2]
World Commission on Dams
planning and project development
framework10 [3]
USEPA's watershed project
guidance11 [3] .
Clean Water Act watershed
management cycle13 [2]
U.S. Forest Service land and
resource management planning
framework14'15 [2]
See Appendix 3-A for description of cited frameworks.
Bold, bracketed numbers are further explained as follows:
[0] - No explicit stakeholder role: process may be amenable to stakeholder involvement, but such involvement is not
described
[1] - Stakeholder-informed process: stakeholder involvement occurs primarily at the outset, as part of goal-setting
[2] - Stakeholder-engaging process: stakeholder involvement is sought throughout the process
[3J - Stakeholder-empowering process: process occurs at the initiative of stakeholders themselves; or framework
deals explicitly with issues of "power" and assigns specific rights to stakeholders
c Integration of social sciences denotes the use of scientific methodologies, not stakeholder inclusion alone. It
includes economics and the decision sciences.
3-3
-------
They develop a broad set of criteria for evaluating efforts that have been implemented. These
criteria are useful prospectively as well and are presented here as relevant to the development of
an integrated process. They state that an effective process
(a) addresses evaluation from a systems perspective, (b) links objective to consequence, (c)
considers the fundamental assumptions and hypotheses that underpin core policy or program
objectives, (d) is grounded in the natural resource, policy/institutional, economic, socio-
cultural and technological contexts of implementation in practice, (e) establishes practical and
valid evaluation criteria by which change can be monitored and assessed, (f) involves
methodological pluralism including both quantitative and qualitative methods to ensure rigor
and comprehensiveness in assessment, and (g) integrates different disciplinary perspectives
(i.e. social, economic, environmental, policy and technological).
Based on these ideas, issues raised in Chapter 2 and the examination of other
frameworks, a set of considerations that address watershed management generally, and
are also specific to the ERA-economic integration problem, are listed in Table 3-2. These
considerations, and the design elements resulting from each, are summarized below.
As was emphasized in Chapter 2, ERA has unique value as an ecologically informed
process that conceptually defines the ecological system at hand and the anthropogenic forces
acting upon it and that progresses, in structured and logical fashion, from ecosystem
management goals to the characterization of risks affecting those goals. An integrated
framework should retain the processes composing the analytic core of ERA, and the essentially
scientific character of the analysis should not be compromised. At the same time, in order to
secure broad participation leading to robust solutions, there must be sensitivity to the critiques of
ERA discussed in Section 2.1.2, particularly that ERA can be too narrowly focused: bearing the
mantle of "science" yet serving particular interests16 or lacking a clear link to management
efforts.17 These criticisms may be answered by an approach that emphasizes the comparative
3-4
-------
TABLE 3-2
Important considerations in framework design, and resulting design elements
Consideration
Specific points
Framework design element
Unique value of ERA
ecologically informed, biophysical in
nature
structured, deductive process
proceeding from goal > objectives
> hypotheses > analyses > risk
characterization
scientific character of analysis is
not compromised
retain core of ERA process
Sensitivity to critiques of
ERA
some stakeholders may perceive
overall process as unfair
assess broad range of alternatives (not
constrained set)
acknowledge limits of science
throughout assessment
acknowledge potential "winners"
and "losers;" extend negotiating
rights
comparative assessment of
alternatives
"deliberation" by "extended peer
community" throughout process
Key aspects of economic
thought
* individual preferences and trade-offs
are essence of value
« citizen sovereignty (as constrained by
mandate of representative government)
comparative assessment of
alternatives
stakeholders in process; analysis
of preferences
risk communication is inside
process
Methodological pluralism
neither ERA nor economic analysis
ascendent
both deliberative (constructive) and
logico-deductive processes can inform
decisions
extended peer community
decision is based on input from
multiple disciplines
Importance of adaptive
management
costs and uncertainties often high;
politics may not bear full
implementation
assume incremental, negotiated
decisions; include analysts here and in
subsequent steps
negotiation part of decision
process
adaptive management integral,
not accessory
Linkage of situational and
regular management
processes
both types of processes are needed
should be mutually supporting
linked cycles
3-5
-------
assessment of a range of management alternatives; that identifies stakeholder groups that are
likely to bear the respective risks and benefits of the alternatives; and that sees negotiation
among these groups as legitimate. Where uncertainties and decision stakes are high, the
1 Q
approach should acknowledge the limits of science by accommodating "deliberation" by
"extended peer communities" throughout the process. Scheraga and Furlow2 coined the term
"policy-focused assessment" to describe a scientific process that is constantly engaged with
stakeholders and decision-makers so that the results will be relevant to policy.
The incorporation of economics into the process implies there will be an increased
emphasis on the measurement of individual preferences, expressed as the willingness to make
trade-offs. This dynamic reaffirms the importance of comparative assessment of alternatives. It
also implies that risk communication, necessary for informing preferences, is an essential
component of an integrated process (whereas it is accessory to ERA); and that stakeholder
preferences will be analyzed in some form.8
Methodological pluralism21 is a relevant goal because the salient attributes of
environmental management problems are not adequately modeled by any single disciplinary
paradigm. The extended peer community should include multiple disciplines; ' both qualitative
and quantitative data collection methods may be needed; and deliberative as well as deductive
processes may be relevant. In the decision-making phase it may not be possible to reduce all
relevant factors to a single dimension: multiple objectives may need to be treated.
Adaptive management has been described as a "learn-by-doing" approach to
?*)
decision-making, in which both goals and approaches are subject to revision over time. When
For guides on sharing environmental information with the public, refer to USEPA;44'45 for useful information on
see S<
3-6
terminology for communicating ecological concepts, see Schiller et al.2 and Norton.4
-------
the process is applied to the implementation of a plan or policy, rather than the ongoing
management of a resource, the term "adaptive implementation" may be used. Analytical
frameworks often treat adaptive implementation as an accessory process - a post-analytic
feedback loop which acknowledges that uncertainty and complexity may prevent us from
precisely hitting the target qn the first try. Experience, however, suggests something more.
Where costs of remediation or restoration are high, the political will to take fully responsive
actions may be lacking, even where scientific knowledge is relatively adequate. Interested
parties might first negotiate a less costly, interim decision. Adaptive implementation could then
constitute an indispensable learning process through which a community gradually acquires
willingness to take more vigorous steps. As Holling et al.24 put it, "managers as well as scientists
learn from change," and the same can be said for other stakeholders. If so, it would be a mistake
to view negotiation as a purely nonscientific process taking place after the specialists have "had
their say." Rather, technical specialists should participate in the design of an incremental process
that yields information and employs evaluation criteria at each step. They should be expected to
play a supporting role during negotiations and to be actively engaged through the adaptive
implementation process.
Finally, because environmental management entails both regular and situational
processes, it may be important to examine how the problem-oriented process of ecological-
economic analysis, decision-making and adaptive implementation that is being developed herein
can best interact with ongoing resource management requirements.
3.3 DIAGRAMING AN INTEGRATED MANAGEMENT PROCESS
Figure 3-1 diagrams a conceptual approach that addresses each of the guiding
considerations listed in Table 3-2 and, in so doing, responds to each of the SAB and Bellamy
3-7
-------
fMHI AT
ASSESSMENT PLANNING
(Stakeholders, Managers, Technical
Specialists Dialogue)
. I rrRNATP/ES"
PROBLEM FORMULATION
UV
CHAi '
- OFSASELINERISK
Integrates
Conceptual
\NALYS!S & CHARACTER'.' -.
'-. OF ALTERNAtiVE
CONSULTATION
WITH
EXTENDED PEER
COMMUNITY
Negotiation ~ - Revised Design]
Shading indicates primary role
played by technical specialists
White indicates interaction of
stakeholders, managers and
technical specialists
ADAPTIVE
IMPLEMENTATION
FIGURE 3-1
A conceptual approach for the integration of ecological risk assessment and economic analysis in
watershed management
3-8
-------
criteria. The major components are discussed in the succeeding sections. In many respects the
approach is similar to the ERA Framework. However, ERA only estimates the likelihood of
adverse ecological effects, and it assumes that economic analysis, if needed, will be able to use
the assessment results. This approach modifies the ERA Framework at every stage of risk
assessment, beginning with the planning process, to ensure compatibility. In so doing, however,
the core scientific character of ERA is not compromised. The scope of planning and problem
formulation are broadened but the key steps of articulating ecological values, goals, objectives,
and endpoints are still carried out. Analysis and characterization of ecological risks is carried out
in a scientific manner as part of the analysis of management alternatives and sometimes also as
part of an assessment of baseline risks.
This conceptual approach would be placed hi the upper right cell of the typology
presented earlier (Table 3-1); that is, it is a situational process, triggered by need rather than
ongoing. However, it includes an adaptive implementation phase, which may continue, and it
can be linked to or used within an ongoing watershed management cycle. It is a planning and
management approach that includes decision-making and implementation; it is not limited to
providing information for decision-support. It generally assumes that stakeholders and decision-
makers will be involved in the initial stages and will remain engaged at some level throughout
the process, such as through consultations with an extended peer community, but that analysis
and characterization will be conducted by technical specialists. Depending on the decision
context,, stakeholders may be empowered to participate in or to make decisions (i.e., it would be
scored as [2] or [3] in Table 3-1). Each of these aspects is further discussed in the following
sections.
3-9
-------
The sequence of discussion is not necessarily that in which the process will occur. The
process may begin with assessment planning, initiated because a problem or opportunity has
been recognized. On the other hand, a proposal for one or more actions may have been
formulated that now requires full evaluation, or a study of baseline risks (i.e., present and future
risks, if no new action is taken) may have been conducted that demonstrates a need for actions to
be formulated and comparatively assessed. A separate step for the study of baseline risks is not
needed at all if the analysis of alternatives includes the no-action alternative. However,
assessment planning, problem formulation and formulation of alternatives all should be
completed prior to the assessment of alternatives and subsequent steps (although the reiteration
of these steps may be necessitated by later findings, or by intervening events).
3.3.1 Assessment planning
Assessment planning is analogous to "planning" in ERA and to "identifying problems
and opportunities" in the U.S. Army Corps of Engineers (USAGE) project planning process; it is
here termed assessment planning to distinguish it from the more encompassing terms "project
planning," used by USAGE, and "resource management planning" used by the U.S. Forest
Service (see Appendix 3-A). It is a stage that emphasizes discussions among analysts of multiple
disciplines (i.e., ecological, economic and others as needed), risk managers and, where
appropriate, stakeholders about values and goals. It is conducted as described in Section 2.1.1.1,
except on three major points. First, the identification of the decision context is somewhat
expanded. Besides identifying the decisions to be made and determining their context,
assessment planners must also determine who has the authority to make the decisions and what
criteria they expect to use. These are critical factors for the characterization and comparison of
alternatives; analysts need to know how the decision-makers view the decision situation so their
3-10
-------
comparisons comprise all the needed elements. For example, decision-makers may be
specifically constrained to consider, or not to consider, particular factors such as cost, equity or
threatened and endangered species, or to prioritize some factors vis-a-vis others.
Second, the scientific disciplines needed to address all important dimensions of the
problem should be represented in assessment planning. Besides ecology and economics, which
are the focus of this document, the watershed management problem may have implications for
human health, requiring the involvement of health risk assessors. In addition, sociocultural
issues such as environmental justice concerns or threats to cultural artifacts could require the
parallel involvement of additional disciplines (geography, cultural anthropology, archeology,
etc.), here and throughout the assessment process. These various analysts should help decision-
makers elucidate their time horizon of concern. Decisions have both short- and long-term
consequences, and ecological and economic time frames of analysis will need to acknowledge
the time horizons of the relevant processes involved, the decision-makers and the other
disciplines.
Third, not only must interested and affected parties be identified, but the ways in which
they may be benefited or harmed by the alternatives under consideration should be indicated
because, depending on the legal context, it may be necessary or advisable to accord them
negotiating rights, or to address compensation issues, in the decision process. This information
will also be useful if the negotiation process is to be modeled (e.g., using game theoretic
techniques, see Section 2.2.5).
3.3.2 Problem formulation
In the ERA Guidelines? problem formulation is a scientific process that is kept separate
from planning (see Figure 2-1, and refer to the discussion of problem formulation in Section
3-11
-------
2,1.1.2). As shown in Figure 3-1, however, it is separated from assessment planning by a dashed
line to indicate the tendency for these two steps to be closely associated in practice. For
example, conceptual models produced in problem formulation diagrammatically illustrate for
stakeholders and decision-makers the complex causes, nature and ramifications of ecological
f\ C
problems in watersheds, as is necessary for assessment planning.
The distinction between these two steps is further reduced here because of the need to
broaden conceptual models and assessment endpoints to include socioeconomic as well as
ecological impacts - an exercise that is likely to rely on repeated discussions with interested and
affected parties. In ERA, risk hypotheses, which are proposed explanations of relationships
between sources, stressors, exposure pathways, receptors and ecological effects, are the basis of
conceptual models (see Section 2.1.1.2). To include socioeconomic impacts, risk hypotheses
must be extended to include the changes in ecosystem services (see Table 2-1) that will be
associated with the changes in those endpoints. Finally, since the evaluation of alternatives is
also required for an integrated assessment, risk management hypotheses are needed as well; these
are proposed explanations of how management alternatives will affect sources, exposures, effects
and services.
Section 2.1 used the example of the decline of a hypothetical reservoir fishery to illustrate
the components of ERA. Section 2.1.1.2 listed population size, mean individual size and
recruitment of popular angling species as appropriate ecological assessment endpoints, and it
stated that conceptual models should diagram the ecological processes whereby the stressors
suspected of causing the decline, in this case agricultural pesticides and municipal and
agricultural nutrients, were thought to exert effects. Continuing that example, the integration of
economics at the problem formulation stage would require adding management alternatives to
3-12
-------
the conceptual model. In this example, suppose that a baseline risk assessment (see Section
3.3.3) had identified nutrient loadings to the reservoir as the actual cause of the decline, and that
risk management alternatives to be studied (see Section 3.3.4) included restricting further
sewerage connections to the municipal treatment plant, upgrading the treatment plant, instituting
an incentives program for riparian zone restoration, and conducting an outreach program to
encourage conservation tillage. Extending the conceptual model would require adding each of
these alternatives to the diagram and illustrating their expected effects on the ecological
processes relevant to the endpoints. Additional effects that might have ecological relevance
would be diagrammed as well, such as important beneficial or detrimental effects on species that
were not the original subject of the assessment. These might require defining additional
ecological assessment endpoints.
Economic effects of the alternatives must also be added to the model. Since the
ecological assessment endpoints in this example (fish species, population, size, etc.) are not
directly valued, the link to ecosystem services such as fishing success must be included in the
diagram, and assessment endpoints corresponding to the service changes (for example, value to
recreational users) must be added. Other economic effect pathways, such as the effects of plant
upgrade costs or land use changes on the local economy, also need to be included. Finally, other
kinds of changes expected to result from the alternatives, such as changes in human health or
quality of life, should also be indicated. Complete risk management hypotheses will consist of a
causal chain that extends from a given management alternative to each of the applicable
ecological and economic assessment endpoints.
The analysis plan, which is the final product of problem formulation, must include
procedures for evaluating the risk management hypotheses, including the efficacy of proposed
3-13
-------
management actions and the relationship between ecological responses and ecosystem services.
The plan must include quantification of the spatial and temporal extent of endpoint changes. (In
the reservoir example, ecosystem service improvements resulting from a management action
would depend on the size of the area over which the fishery was improved and the time required
to effect the improvement.) The plan must also include proposed methods for the comparison of
alternatives that closely reflect the needs of decision-makers, as determined during assessment
planning (see Section 3.3.7 for further discussion of comparison). Finally, the analysis plan and
other products of problem formulation (assessment endpoints and conceptual models) must be
verified with managers and stakeholders as being not only technically accurate but well-targeted
to the most important concerns. If members of these groups have been engaged throughout
assessment planning and problem formulation, they may have acquired in the process sufficient
technical knowledge to understand these products. If not, or if the economic methods to be used
later will require surveys of a broader audience or the general public, then careful work will have
to be done at this stage to build a risk communication capability. Steps may include developing
common-language terminology to express key ecological concepts,27 and using focus groups to
refine this lexicon and verify assumptions about the values held by the public or stakeholder
groups.
3.3.3 Analysis and characterization of baseline risk
If preexisting information is not sufficient, a separate study of baseline risks may be
conducted prior to the formulation of alternatives. Although definitions can vary slightly,
baseline risks are defined as the present and future risks to ecosystems or human health that
would occur if no new action is taken.28 Baseline risk assessment is a formal part of
environmental impact assessments conducted under the National Environmental Policy Act
3-14
-------
(NEPA) and site characterizations conducted under the Comprehensive Environmental
Response, Compensation and Liability Act ("Superfund"). Since NEPA requirements are
invoked only when an action is proposed, the action alternative and no-action alternative are
assessed in the same stage of environmental impact assessment, and baseline assessment as a
separate step is not needed. Under Superfund, on the other hand, baseline assessment is needed
to characterize the risks prior to remedial action design. In watershed management, a separate
baseline assessment as shown in Figure 3-1 may be required if the kind of management action
needed, or the need for any action at all, is unclear.
Characterizing baseline risks may also require characterization of harms that have already
occurred. Risks to socioeconomic well-being may also form part of this analysis, but these risks
are more easily addressed in comparative than absolute terms and are therefore likely to receive
limited attention at this stage. Methods for analysis and characterization of ecological risk were
discussed in Section 2.1.1; methods for the assessment of health risks are presented elsewhere.
Determining the magnitude and severity of ecological or health effects helps determine the need
for management actions. Determining causality and pathways of exposure provides information
useful in the design of management alternatives. Developing models of exposure and response,
and risk characterization approaches, establishes the methods that will be used in the
comparative analysis of management alternatives.
The generation of exposure scenarios may be an important part of baseline risk
assessment. Scenarios are often used to describe alternative circumstances for which risk will be
estimated. In some instances they help describe the range of the expected exposure conditions;
for example, an assessment of pesticide impacts on watershed resources may require setting up a
range of use scenarios to cover the different types of practices actually occurring in the
3-15
-------
watershed. Exposures resulting from all scenarios would then be used in the full characterization
of baseline risk. In other cases, scenarios result from alternative assumptions about an unknown
future; for example, alternative CC>2 emission assumptions and global climate models are being
used to establish alternative future climate scenarios for watershed risk assessment.30 These
scenarios are part of baseline assessment if they do not correspond to designed policies or
alternative management actions but rather form a positive basis for design of management
actions. On the other hand, some future scenarios are explicitly policy-based. For example,
Coiner et al. developed future scenarios for the Walnut Creek watershed of Iowa based on
alternative policies that respectively prioritized agricultural production, water quality and
biodiversity; and Hulse et al.32 developed scenarios for the Muddy Creek watershed of Oregon
reflecting different policies with respect to development density and conservation. Policy-based
future scenarios, which enable a normative comparison of policy outcomes, would be developed
as part of the next stage, "Formulation of Alternatives."
3.3.4 Formulation of alternatives
This phase entails the development of alternative action plans for achieving the watershed
management objectives. Depending on the nature of watershed problems and the management
goals, there is a wide array of management actions that may be considered at this step (Table
3-3). The planning process may include engineering design or policy development; the
discussion of specific techniques is beyond the scope of this report. Details of processes that can
be used for developing alternative plans are presented elsewhere.9'13'33"35 '3>33,34,34>34,35,35,35while
actions to reduce ecosystem risks are emphasized in this report, actions designed to reduce
human health risks or improve socioeconomic well-being may cause ecological changes and
therefore may also need to be evaluated according to the procedures in this chapter.
3-16
-------
TABLE 3-3
Categories (and some examples) of watershed management measures .
Control of point sources (source reduction, waste recycling, waste pretreatment, or improvement
of waste treatment infrastructure)
Control of urban or agricultural nonpoint sources (land use changes, runoff detention structures,
improved waste management, educational outreach programs)
Contaminant remediation (chemical spill cleanup, acid mine drainage treatment)
Stream channel and riparian restoration (tree planting, instream structures)
Species management (habitat creation, control of nonnatives, reintroductions)
Water resource development (irrigation, hydropower, recreation)
Improvement of other use values (access)
Strategies for adaptation to global change (land use changes to accommodate sea-level rise)
3-17
-------
To avoid bias toward preselected solutions, planning objectives and constraints should be
clearly established in advance, and a broad range of alternatives should be examined (see
Section 2.1.2). A given alternative should comprise not just the design of management actions
such as those listed in Table 3-3; long term success depends on establishing a planned system
that also includes implementation tools (such as permits, incentives and information) and
institutional and organizational arrangements (such as extension services). 5
3.3.5 Consultation with extended peer community
Funtowicz and Ravetz describe "extended peer communities" as including scientists
outside the specific discipline or practice at hand, and others lacking formal knowledge but
possessing practical, including local, knowledge (see Section 2.1.2). The term used here
includes interested and affected parties and decision-makers, in addition to scientific peers.
"Consultation" does not apply to the assessment planning phase, where interested and affected
parties are already an integral part of the process. It applies rather to components such as
analysis and characterization that are explicitly scientific. "Consultation" recognizes on the one
hand that these steps must be carried out by analysts with specialized knowledge, and on the
other that risk assessment often requires judgments that go beyond strict inference and are
therefore susceptible to bias. Consultation is a process in which technical information from the
assessment is discussed with the extended peer community for purposes of (a) identifying issues
or deficiencies in the assessment and (b) keeping interested and affected parties engaged during
1 r*
what can be a lengthy process. It is equivalent to the term "deliberation" as used by NRC.
3.3.6 Analysis and characterization of alternatives
In this stage the alternatives are assessed from the perspective of various disciplines
including ERA, economics and possibly others such as human health risk or sociocultural
3-18
-------
assessment, depending on the situation. In the diagram (Figure 3-1), the disciplines are shown as
jointly conducted, indicating at least an exchange of information and at best an integrated
analytic approach. However, it is by no means a requirement that the disciplines depart from
their characteristic approaches, as long as they are mutually informed. Since ecological and
economic time frames of analysis may differ, the time frame for each should be made explicit.
Analysis of alternatives is guided by the risk management hypotheses, indicating which
exposures and responses are likely to be affected by risk management. Those not expected to be
affected remain part of the baseline risk but are not included in the alternatives analysis. The
ecological risk component estimates the changes in exposure profiles likely to result from each
management alternative. Where management alternatives create new exposures, i.e., to stressors
that were not originally present (such as to sediments from project construction or to pesticides
used to control invasive species), additional exposure profiles and exposure-response
relationships beyond those of the baseline assessment must be developed.
Ecological risk characterization describes probabilities, magnitudes and severities of
effects on ecological assessment endpoints. These should be described both in absolute terms
and as changes with respect to baseline. Uncertainties in the effect estimates must be
characterized as well, and the uncertainties, as well as the other parameters, must be carried
forward into the economic analysis.26
The economic component analyzes costs (including financial and opportunity costs) and
benefits associated with the management alternatives. This includes, to the extent practicable,
the changes (with respect to the no-action baseline) in ecosystem services that are associated
with changes in the ecological assessment endpoints. This especially includes services that can
be quantified objectively, such as biophysical services (e.g., the production of food, fiber or other
3-19
-------
goods, and regeneration and stabilization processes) and services that are quantifiable by
revealed preference methods (e.g., many forms of recreation). It may also include life-fulfilling
functions (including functions corresponding to non-use values), if these can be quantified by
benefit transfer methods. The use of stated-preference or other subjective methods to quantify
these services is not ruled out at this stage, but for pragmatic reasons such efforts may best be
carried out as part of the subsequent, comparison phase. For example, if a stated preference
questionnaire were to be designed and administered, it may be possible, and therefore cost-
effective, to do so in such a way as to affect a multifactoral comparison, as described below.
3.3.7 Comparison of alternatives
This step is included in the conceptual approach based on the assumption that not all
factors important for decision-making can be objectively reduced to a single vector and that the
comparison step itself therefore is both subjective and nontrivial. Even if net economic benefit
to society, as determined by CB A, is an important criterion, there will usually be other
ecological, moral, political or legal factors that it cannot adequately encompass. Comparison is
the step in which these various factors are arrayed in terms as amenable as possible to those of
the legitimate decision-maker, be it an agency official, the collective of residents of a
jurisdiction, or individual landowners. Any process used to assign subjective weights to the
factors, or to enable individuals or groups to systematically compare the alternatives (based on
information about these factors and their subjective judgment) is considered to be part of the
comparison phase. Methods may include stated preference analyses (appropriate for large
groups of individuals) or decision-analytic approaches in which factors are weighted by technical
experts, or by representatives of interested and affected parties, acting either as individuals or
within consensus-seeking groups (see Morgan for a useful summary of non-monetary, multi-
3-20
-------
criteria evaluation methods). On the other hand, if the ultimate decision will be reached by
negotiation among parties with divergent interests, the comparison methods used might seek to
identify the alternative that the parties believe is the best they can hope to obtain, rather than the
one with optimal overall utility. The comparison process is carried out according to agreements
made during the assessment planning phase (in which the decision context, including
decision-makers and decision factors, was described) and the problem formulation phase (in
which the comparison methods to be used were verified).
3.3.8 Decision
Because environmental management problems in watersheds usually are
multidimensional, it is unlikely that a problem can be solved based on the actions or authority of
any single entity. Therefore, the decision process is likely to involve multiple parties. In spite of
the findings of the analysis and characterization of alternatives, or because of the associated
uncertainties, the parties may hold divergent beliefs about expected outcomes of a given
alternative, or even if they agree on technical issues they may have divergent incentives or
expectations regarding compensation. They may also have divergent interpretations of legal
constraints on the decision process. Therefore, a decision may entail less a consensus selection
among the alternatives than a negotiated redesign. Where implementation cost is a predominant
factor, negotiation may entail scaling back on a design or agreeing to a provisional schedule of
incremental implementation, conditioned on verification that performance criteria are being met.
Technical specialists therefore may be called on to assist the negotiation process.
3.3.9 Adaptive implementation
Because achieving agreement can be difficult, a provision for adaptive implementation
may therefore be indispensable to reaching a decision: it can provide a middle-ground approach
3-21
-------
that satisfies no one but provides a respite until confirmatory data are available. However, a
flexible or incremental approach does not constitute adaptive management unless several criteria
are met. Holling et al.24 recommended that experimental perturbations be designed to evaluate
specific questions. Walters 7 emphasized that perturbations need to be great enough to probe
system responses across domains of interest; cautious incrementation may not produce any
fj>j
usable information. The National Research Council (NRC) stated that adaptive management
must not only generate useful information but must specify the mechanisms by which the
information will be translated into policy and program redesign. Depending on the findings and
the nature of the agreement, evaluation of the data could lead to further action or could trigger
renewed negotiation; it could also invalidate certain assumptions of planning, problem
formulation or analysis, indicating that earlier stages of the process need to be reiterated. The
possibility of revisiting earlier steps in the assessment as more information is learned is indicated
by broken lines in Figure 3-1.
3.3.10 Linkage to regular management cycles
The process described here for integrated assessment is situational; i.e., it should not be
thought of as a cyclical process that can never be completed. By contrast, resource management
is ongoing, and the two processes can be mutually supportive. For example, the rotating basin
approach to CWA management (see Appendix 3-A) identifies priorities and needed actions,
which may call for a detailed, integrated assessment in situations where needed actions are
unclear or where regulatory approaches are insufficient. Stakeholder processes that may have
been established as part of that cycle can be drawn upon for the integrated assessment. The
rotating basin approach also establishes a long-term water-quality and biological monitoring data
base that can establish temporal trends and correlations in stressors and biological response that
3-22
-------
can be useful in establishing causation, exposure profiles and stress-response relationships. The
management alternatives to be considered in the integrated analysis can include (among other
measures) regulatory and incentive mechanisms provided for under the CWA, to be implemented
and monitored as part of the regular management cycle. Similarly, some watershed resources
(e.g., forest resources) are adaptively managed in an ongoing fashion (see Appendix 3-A, Figure
3-A-5). Integrated assessments can link effectively to an adaptive management cycle.
3.4 EXAMPLES OF ANALYSIS AND CHARACTERIZATION FOLLOWED BY
COMPARISON OF ALTERNATIVES
Planning and problem formulation (together with baseline ERA and formulation of
alternatives) lay the groundwork for a successful integrated analysis, but the technical aspects of
integration are encountered in the analysis and characterization of alternatives and in the
comparison step that follows (Figure 3-1). Because there are a variety of ecological and
economic analytic tools that could be applied in these stages, the specific elements of these steps
will also vary. This section provides examples to illustrate how ecological and economic
techniques might interact.
3.4.1. Example 1: Cost-benefit analysis of all changes that can be monetized, with
qualitative consideration of other changes
Cost-benefit analysis (CBA, see Section 2.2.3) is commonly used where decision-makers
are concerned about the net economic benefit to society of a given action (that is, to determine
whether economic efficiency is increased).38 As discussed in Section 2.3.2, CBA is required for
certain federal actions. In an integrated assessment where changes in economic efficiency will
be a key factor in the decision, the process may occur as diagramed in Figure 3-2. For each
management alternative, ecologists would quantify the changes expected in each ecological
assessment endpoint Changes that could not be quantified would be characterized qualitatively.
3-23
-------
: ANALYSE ,'irai
ify endpoint
ehartgelrwhere
feasible
QuanttTy Tirianc-Ial costs and
other changes
35olaci-,,V
other changes
effeess
'-l: -y
Analyze economic
aificfency,
Quanbry e.r
Feasible
Ascribe
Compare rvonquaulifieci
aocl other
nonmonettzed ch
efficient
/, Jmoact
Cotrtpa re fi pnqu a nti FI. n)
,^nd ofher
-d changes.'
OM'PARlSOri OF '.ALTERWATIV E.
FIGURE 3-2
Analysis and characterization of alternatives, followed by their comparison, example 1: CBA of
all changes that can be monetized, with qualitative consideration of other changes.
3-24
-------
Other analysts might examine quantitative or qualitative effects on health or quality of life, as
needed. Economists would look first at the financial costs of the alternatives and any effects that
could be determined from markets (for example, opportunity costs of land taken out of
agricultural production). Economists would then seek to monetize the effects estimated from the
ecological or other analyses, using revealed preference or benefit transfer methods wherever
possible (see Section 2.2.2 and especially Table 2-2). Due to their required time and cost, stated-
preference techniques would be used only if other methods were unsatisfactory (that is, if nonuse
values are important and/or reliable studies from similar settings are lacking). Based on this
information, economists would analyze economic efficiency, equity and impacts (Section 2.2.4).
This information, and information about effects that either could not be quantified or could not
be monetized, would be carried forward into the comparison step.
3.4.2 Example 2: Use of stated preference techniques to effect integration of
ecological, economic and other factors
i
In the example above, stated preference methods, if used at all, would monetize the
ecological changes associated with one or more management alternatives. Figure 3-3 diagrams
the use of stated preference methods to achieve a more broadly-based comparison, such as one
that includes the ecological, health, quality-of-life, equity, and impact dimensions of a choice.
For example, this could be accomplished using a contingent valuation method (CVM, see
Appendix 2-A) survey that explains the effects of the management alternatives (i.e., that
"frames" the alternatives) in each of those dimensions before asking individuals about their
willingness to pay (WTP) or to accept (WTA) (see Section 2.2.2). To design such a survey, each
of those dimensions would first need to be analyzed and characterized, with all effects quantified
to the extent possible. The technical findings would then need to be refined (such as through the
use of focus groups) into a format that highlighted only the most important factors and used
3-25
-------
-J-
bcological
Health DJ
1 Quantify endpoinf
changes * Her ft
feasible
Quantify financial costs
and ma.rket-bas.ed
economic effects
Quantify endpoint
; changes where
- fe-asibfc
Qijafitatn
descri&e
olhsrchanges
analyze
equity, economic impact
describe
olhei chantjes
Express equity effects, Impacts
in cornmor* language
it* common >
Estimah
social bcnents
iMRARISO&J 0^ ALTERMA1IV/ES-
FIGURE 3-3
Analysis and characterization of alternatives, followed by their comparison, example 2: use of
stated preference techniques to effect integration of ecological, economic and other factors.
3-26
-------
commonly understood language.27 A broadly-framed CVM approach that was similar to this in
certain respects was employed in the Big Darby Creek watershed case study presented in
Chapter 4.
A broad comparison could also be accomplished using a choice modeling method such as
conjoint analysis (CA, see Appendix 2-A). In this approach, focus groups would again be used
to identify the most important factors across those dimensions, and to establish common
terminology. Survey design would entail transforming those dimensions into choice attributes,
so that respondents' choices would reveal how the various dimensions contributed to WTP or
WTA. A method of this type was used in the Clinch Valley case study presented in Chapter 5.
3.4.3 Example 3: Use of linked ecological and economic models to dynamically
simulate system feedbacks and iteratively revise management alternatives
A disadvantage of sequentially integrated assessments, in which ecological changes are
estimated and then economically evaluated, is that there is no opportunity to simulate dynamic
interaction between economic and ecological processes. >4 In cases where the economic effects
of changes in ecosystem quality (such as effects on housing, recreational or agricultural
values41'4 ) will have an important influence on land use decisions and ecosystem quality, an
integrated system that models these feedbacks may enable a better understanding of the behavior
of the real systems. In Figure 3-4, models of the ecological processes affecting the assessment
endpoints are linked to a regional economic model in a manner that allows parameter feedbacks
over time. Once such a modeling system is established, management alternatives can be
simulated and iteratively revised to optimize their design according to a variety of criteria, such
as cost-effectiveness, equity and ecological risk. The example pictured in Figure 3-4 arbitrarily
assumes a case where ecological and economic models are linked and that other effects (e.g., on
health or quality of life) are estimated using other methods. The example further assumes that it
3-27
-------
XLY8IS S GHARACTEPIZAl ION.GF
Economics:
de1 ecoi.
.ss~£5 affecting
en f | points :; :
'V
es-ijffec
sndpniryts
I
Create (inketi model
ecological I
chfi
.Quantify
eco (<<->
changtss
, Adjust'mgaaDemerit !___
"3 alternatives, re he reie l_
endpi(int.c.hanges- |
Hcallli
1 'fjTere,
" isasible
Qualitative^
jte.scribfc
niter Sttshges
r.CMPARlSON OF£LTtRMAT!V±3
FIGURE 3-4
Analysis and characterization of alternatives, followed by their comparison, example 3: use of
linked ecological and economic models to dynamically simulate system feedbacks and iteratively
revise management alternatives.
3-28
-------
may be difficult to estimate net social benefit from such a modeling approach, since WTP or
WTA for nonuse values is not estimated, although in theory an appropriate benefit transfer
module could be added to the model, hi the comparison step, the modeling results for the
various management alternatives and/or for different optimization criteria could be described,
along with qualitative discussion of any effects that could not be quantified by the modeling
effort
3.5 CONCLUSION _
This conceptual approach does not represent a fundamental departure from existing
practice. Its steps correspond in large part to those of other frameworks (Table 3-4); they differ
as needed to emphasize the ERA-economic integration problem. However, the incorporation of
multiple disciplines into an integrated assessment process may create significant challenges of
communication, coordination and funding. Therefore the use of this approach is not appropriate
in all instances where ERA alone is called for. However, if decisions need to be informed on the
basis of both ecological risks and economics, an integrated approach, while more demanding, is
more likely to provide coherent information.
This conceptual approach is used in the following chapters as a vantage point from which
to analyze a set of case studies. As was mentioned in the previous chapter, the case studies that
will be presented in Chapters 4-6 were undertaken with a number of constraints, hi each case,
the involvement of economists came well after ERA had been initiated, and in one case the ERA
was never completed. Furthermore, the scope of these studies did not encompass the full span of
management activities, from assessment planning to adaptive implementation. Nonetheless, the
conceptual approach helps to illustrate how the methodological advances and insights from each
3-29
-------
TABLE 3-4
Rough correspondence between the components of the conceptual approach for ERA-eeonomic integration and other selected
watershed management frameworks2
Component of Conceptual Approach
for ERA-EA Integration (Figure 3-1)
Assessment Planning
Problem Formulation
Analysis and Characterization of
Baseline Risk
Formulation of Alternatives
Analysis and Characterization of
Alternatives
Comparison of Alternatives
Decision
Adaptive Implementation
Corresponding Component
Framework for Ecological Risk
Assessment '
Planning
Problem Formulation
Analysis
Risk Characterization
NAb
Analysis (reiteration)
Risk Characterization
(reiteration)
NA
NA
Framework for Integrated
Environmental Decision
Making 3
Phase I: Problem Formulation
Phase II: Analysis and
Decision Making
Phase III: Implementation and
Performance Evaluation
Watershed Management Mode! "3
Phase I: Assessment/Problem
Identification
Phase II: Planning
Phase III: Implementation
Phase IV: Evaluation
U.S. ACE Six-Step Planning
Process
Identifying Problems and
Opportunities
Inventorying and Forecasting
Conditions
Formulating Alternative Plans
Evaluating Alternative Plans
Comparing Alternative Plans
Selecting a Plan
NA
See Appendix 3-A for discussion of other watershed management frameworks
b Not applicable
-------
case study could be used to fullest advantage, both in the watersheds that were studied and in
other settings where similar methods could be applied.
3.6 REFERENCES
1. USEPA, Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F, Risk Assessment
Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
2. USEPA, A Framework for the Economic Assessment of Ecological Benefits, Science Policy
Council, U.S. Environmental Protection Agency, Washington, DC, Feb. 1, 2002.
3. SAB, Toward Integrated Environmental Decision-Making, EPA-SAB-EC-00-011, U.S.
Environmental Protection Agency, Science Advisory Board, Integrated Risk Project,
Washington, DC, 2000.
4, Bellamy, J.A. et al., A systems approach to the evaluation of natural resource management
initiatives, Journal of Environmental Management, 63,407, 2001.
5. USEPA, Environmental Monitoring and Assessment Program (EMAP) Research Strategy,
EPA/620/R-98/001, U.S. Environmental Protection Agency, Washington, DC, 1997.
6. Walmsley, J.J., Framework for measuring sustainable development in catchment systems,
Environmental Management, 29, 195,2002.
3-31
-------
7. Stahl, R.G. et al, Risk Management: Ecological Risk-Based Decision-Making, Society for
Environmental Toxicology and Chemistry, Pensacola, FL, 2001,
8, PCCRARM, Framework for Environmental Health Risk Management,
Presidential/Congressional Commission on Risk Assessment and Risk Management,
Washington, DC, 1997.
9. USAGE, Planning Guidance Notebook, ER 1105-2-100, U.S. Army Corps of Engineers,
Washington, DC, 2000.
10. World Commission on Dams, Dams and Development; A New Framework for Decision-
Making, Earthscan Publications Ltd., London, 2002.
11. USEPA, Watershed Protection: a Project Focus, EPA 841-R-95-003, Office of Water,
Environmental Protection Agency, Washington DC, 1995.
12. Timmerman, J.G., Ottens, J.J., and Ward, R.C., The information cycle as a framework for
defining information goals for water-quality monitoring, Environmental Management, 25,
229,2000.
13. USEPA, Watershed Protection: a Statewide Approach, EPA 841-R-95-004, Office of Water,
U.S. Environmental Protection Agency, Washington DC, 1995.
3-32
-------
14. USFS, National forest system land and resource management planning, Federal Register, 65,
67513,2000.
15. USFS, National Forest System Land and Resource Management Planning; Proposed Rules,
Federal Register, 67, 72769, 2002.
16. Pagel, J.E. and O'Brien, M.H., The use of ecological risk assessment to undermine
implementation of good public policy, flyman and Ecological Risk Assessment, 2, 238,1996.
17. Butcher, J.B. et al., Watershed Level Aquatic Ecosystem Protection: Value Added of
Ecological Risk Assessment Approach, Project No. 93-IRM-4(a), Water Environment
Research Foundation, Alexandria, VA., 1997,342 pp.
1.8. NRC, Understanding Risk: Informing Decisions in a Democratic Society, Washington, DC,
1996.
19. Funtowicz, S.O, and Ravetz, J.R., A new scientific methodology for global environmental
issues, in Ecological Economics: The Science and Management ofSustainability, Costanza,
R. Ed., 1991, 10, 137.
20, Scheraga, J.D. and Furlow, J., From assessment to policy: lessons learned from the U.S.
National Assessment, Human and Ecological Risk Assessment, 7,1227, 2002.
21. Norgaard, R., The case for methodological pluralism, Ecological Economics, 1, 37,1989.
3-33
-------
22. NRC, Restoration of Aquatic Ecosystems: Science, Technology and Public Policy, National
Research Council, Commission on Geosciences, Environment and Resources, Washington,
DC, 1992.
23. NRC, Assessing the TMDL Approach to Water Quality Management, National Research
Council, National Academy Press, Washington, DC, 2001.
24. Rolling, C.S. et al., Adaptive Environmental Assessment and Management, Wiley-
hiterscience, New York, 1978.
25. Serveiss, V.B., Applying ecological risk principles to watershed assessment and
management, Environmental Management, 29,145,2002.
26. Suter, G.W., Adapting ecological risk assessment for ecosystem valuation, Ecological
Economics, 14, 137,1995.
27, Schiller, A. et al., Communicating ecological indicators to decision-makers and the puhlic,
Conservation Ecology, 5, 19 [online], 2001.
28. USDOE, Use of Institutional Controls in a CERCLA Baseline Risk Assessment, CERCLA
Information Brief EH-231-014/1292, U.S. Department of Energy Office of Environmental
Guuidance, Washington, DC, 1992.
3-34
-------
29. van Leeuwen, CJ. and Hermens, J.L.M., Risk Assessment of Chemicals; An Introduction,
Kluwer Academic Publishers, Dordrecht, 1995.
30, Rogers, C.E., Julius, S.H., and Furlow, J., Assessment as a method for informing decisions
about water quality, aquatic ecosystems and global change, in Water Resource Issues,
Challenges and Opportunities: Part II: Using Science to Address Water Issues, 2002,10.
31. Coiner, C., Wu, J., and Polasky, S., Economic and environmental implications of alternative
landscape designs in the Walnut Creek Watershed of Iowa, Ecological Economics, 38,119,
2001.
32. Hulse, D. et al., Planning alternative future landscapes in Oregon: evaluating effects on water
quality and biodiversity, Landscape Journal, 19,1, 2000.
33. USEPA, Ecological Restoration: A Tool to Manage Stream Quality, EPA 841-F-95-007, U.S.
Environmental Protection Agency, Office of Water, Washington, DC, 1995.
34. U.S.Water Resources Council, Economic and Environmental Principles and Guidelines for
Water and Related Land Resources Implementation Studies, 1983.
35. Hufschmidt, M.M., A conceptual framework for watershed management, in Watershed
resources management: An integrated framework with studies from Asia and the Pacific,
Easter, K. W., Dixon, J. A., and Hufschmidt, M. M. Eds., Westview Press, Boulder, 1986,2,
17.
3-35
-------
36. Morgan, R.K., Environmental Impact Assessment: A Methodological Perspective, Kluwer
Academic Publishers, Boston, 1998.
37. Walters, C.J., Adaptive Management of Renewable Resources, Macmillan, New York, 1986.
38. USEPA, Guidelines for Preparing Economic Analyses, EPA-240-R-00-003, Prepared by the
National Center for Environmental Economics, 2000.
39. Lindner, M. et al., Integrated forestry assessments for climate change impacts, Forest
Ecology and Management, 162, 117, 2002.
40. Duraiappah, A.K., Sectoral dynamics and natural resource management, Journal of
Economic Dynamics and Control, 26,1481,2002.
41. Geoghegan, J., Wainger, L.A., and Bockstael, N.E., Spatial landscape indices in a hedonic
framework: an ecological economics analysis using GIS, Ecological Economics, 23,251,
1997.
42. Odom, D.I.S. et al., Policies for the management of weeds in natural ecosystems: the case of
scotch broom (Cytisus scoparius, L.) in an Australian national park, Ecological Economics,
44,119,2003.
43. Davenport, T.E., The Watershed Project Management Guide, Lewis Publishers, Boca Raton,
FL, 2002.
3-36
-------
44. USEPA, Considerations in Risk Communication: A Digest of Risk Communication As a
Risk Management Tool, EPA/625/R-02/004, National Risk Management Research
Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH, 2003.
45. USEPA, Risk Communication in Action: Environmental Case Studies, National Risk
Management Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH,
2003.
46. Norton, E.G., Improving ecological communication: the role of ecologists in environmental
policy formation, Ecological Applications, 8, 350, 1998.
3-37
-------
APPENDIX 3-A
DISCUSSION OF EXISTING FRAMEWORKS THAT HAVE BEEN APPLIED TO
WATERSHED MANAGEMENT
Table 3-1 presents a typology of frameworks that have been applied to the processes of
watershed assessment and management. This appendix discusses the frameworks listed in each
of the four cells of the typology, and it presents several applicable flow diagrams that serve as
background for the design of the conceptual approach presented in Figure 3-1.
Situational monitoring or assessment frameworks
Several frameworks pertain to monitoring or assessment that provide information for
decision-makers but do not include the decision-making process. ERA, per U.S. EPA's
Guidelines, is described in Section 2.1 and diagrammed in Figure 2-1. ERA is a situational
process for decision support; it is initiated in response to past, ongoing or potential future adverse
effects to ecological resources. ERA emphasizes the natural sciences and the separation of
science and policy. Stakeholder involvement may be important for development of management
goals during planning and, debatably, for problem formulation, but is considered inappropriate
for analysis and risk characterization. The results of risk characterization are communicated to
risk managers, but decision-making occurs outside the ERA process.1 A Framework for the
Economic Assessment of Ecological Benefits has been described by U.S. EPA which explores
the potential integration of ERA and economic valuation techniques; it has not been applied to
watershed management but is included in the typology as a point of reference.
Environmental monitoring is an essential component of watershed management, and
decisions about what to monitor implicitly are decisions about management. Most monitoring
programs are limited to the collection of natural science data, but some include economic and
3-38
-------
institutional indicators as well. An example of the former is USEPA's Environmental
Monitoring and Assessment Program (EMAP), which estimates status and trends of selected
ecological resources by monitoring indicators of ecosystem structure and function and by
measuring relationships between environmental stressors and impacts. An example of a broader
indicators framework is one developed by the Organization for Economic Cooperation and
Development (OECD).4 The DPSIR framework (see Table 3-1) calls for indicators of the social
and economic conditions that drive environmental changes, and the policy and management
responses to those changes, in addition to indicators of the environmental changes themselves.
Monitoring system design usually stresses input from managers but not other stakeholders. For
example, EMAP's indicator development process borrows several concepts (such as ecological
values, assessment questions, and conceptual models) from the ERA Framework but does not
assume stakeholder involvement.5'6
Regular monitoring or assessment frameworks
ERA generally is not a regular process; while its steps may be reiterated as more is
learned, it is not intended to be continuous. Frameworks for the set-up of monitoring systems,
including indicator design, usually depict a one-time (i.e., situational) process as well. However,
a cyclical (i.e., regular) redesign process can allow monitoring systems to adapt as knowledge
and management needs change.7
Situational planning and management frameworks
The Society for Environmental Toxicology and Chemistry has described an ecological
risk management framework composed of the following steps:8
» issue identification
goal setting
management options development
data compilation and analysis
3-39
-------
* option selection
decision implementation
tracking and evaluation.
The process is informed by stakeholders during goal setting, and effective communication with
stakeholders throughout the process is considered important. It assumes that economic analysis
will be involved in the decision, but processes for integrating ecological and economic aspects
are not discussed.
The Framework for Environmental Health Risk Management depicts a process that is
similar, albeit with a slightly different ordering of steps (Figure 3-A-l).9 Active engagement of
stakeholders is encouraged throughout the process, and it is suggested that stakeholders be
empowered to make decisions where allowable. While the framework is pictured as cyclical, it
should be viewed as situational (responding to problems) yet amenable to adaptive management
as necessary to implement effective solutions. A panel convened by USEPA's Science Advisory
Board (SAB), tasked with making recommendations on the integration of environmental
decision-making, presented similar ideas10 but depicted the process more appropriately as
unidirectional, albeit with feedback loops, rather than cyclical (Figure 3-A-2).
The U.S. Army Corps of Engineers (USAGE) uses a six-step planning process for civil
works projects, including those related to water resources and watersheds:
Step 1 - Identifying problems and opportunities
» Step 2 - Inventorying and forecasting conditions
* Step 3 Formulating alternative plans
Step 4 - Evaluating alternative plans
» Step 5 Comparing alternative plans
* Step 6 Selecting a plan.
The process includes decision-making but in most instances does not include implementation,
retrospective evaluation or adaptive management. Stakeholder involvement is intended to play
3-40
-------
Problem/
Context
Evaluation
Engage
Stakeholder
Options
Actions
Decisions
FIGURE 3-A-l
Framework for environmental health risk management (from PCCRARM9)
3-41
-------
Information
Expert Judgment
Values
- Information
-Expert Judgment
- Values
Legal and
Institutional Milieu
PHASE 1
PROBLEM FORMULATION
(What are the most important environmental risks?
What are our environmental goals?)
Risk Comparisons Goal Setting
Preliminary Options Analysis
I
PHASE II
ANALYSIS AND DECISION MAKING
(What are the best risk reduction opportunities?
How can we achieve our goals and objectives?)
Risk Assessment Screenlng/Selectkm
Options Analysis Performance Measures
PHASE III
IMPLEMENTATION and
PERFORMANCE EVALUATION
(How are we doing?)
Implementation Monitoring and Reporting
Information Evaluation
REPORT
CARD
(Is the nature
of the problem
changing?)
REPORT
CARD
(Are we meeting
our objectives?)
FIGURE 3-A-2
Framework for integrated environmental decision making (from SAB10)
3-42
-------
an important role in step 1, including in the selection of decision criteria; communication
channels are to be maintained throughout the process; and stakeholder consultation is to occur
after evaluation is completed and before plan selection. Evaluation of alternative plans includes
quantifiable national and regional costs and benefits as well as nonquantifiable environmental
and social impacts or benefits.
By comparison, a planning and project development framework developed by the World
Commission on Dams12 provides for a more extensive stakeholder role and for adaptive
management (Figure 3-A-3). Criteria ensuring, among other things, public participation,
assessment of ecological risks, and consideration of a comprehensive set of alternatives, are
checked at the conclusion of each development phase. Analyses of alternatives include the
identification of people who are affected when lands or other resources are put at risk by the
project, and negotiating rights with respect to the final decision are conferred according to risk
burden. The framework emphasizes compliance with negotiated agreements during and post-
construction. Finally, project operation is to be reviewed periodically and should adapt to
changes in the project context.
A four-step process for the planning and implementation of watershed projects (Figure 3-
A-4) was described by USEPA13 and more fully elaborated by Davenport.14 The process is
designed to be carried out through a partnership of government agencies and local stakeholders,
and it emphasizes involvement and action. The assessment and problem identification phase
consists of four parts - inventory, analysis, problem identification and goal-setting - and is
analogous to ERA. However, ERA assumes that analysis itself will require advance planning
and substantial time and resources to conduct and will result in a quantitative characterization of
risks, whereas the watershed project management approach emphasizes qualitative description of
3-43
-------
Needs Assessment
Validate the needs
for water and
energy services
Selecting Alternatives
Identify the preferred
development plai among
the lull range of options
Investigate studies
Criteria 2A
Policy, program, projects
Assign,
responsibility for
implementation
Project Preparation
Verify agreements are
in place before tender of
the construction contract
Criteria 3
Project Implementation
Confirm compliance
before commissioning
Criteria 4
I
Project Operation
Adapt to
changing context
Criteria 5
FIGURE 3-A-3
A framework for planning and project development of large dams, including five key decision
points at which specific criteria should be evaluated
(redrawn from World Commission on Dams12)
3-44
-------
PHASE 1
Assessment/Problem
Identification
PHASE 3
Implementation
PHASE 4
Evaluation
FIGURE 3-A-4
A watershed management model for the planning and implementation of watershed projects
(redrawn from Davenport14)
3-45
-------
the most critical problems and their causes. Natural science is used to identify problems, and
know-how, partnerships and consensus-building processes are used for making and
implementing decisions. Project analysis, including economic analysis, is not emphasized. Like
the Framework for Environmental Health Risk Management the process is pictured as circular;
we have grouped it with situational methods on the assumption that efforts will conclude once
conditions change. If a partnership is effective, however, an effort could be longstanding.
Regular planning and management frameworks
Several frameworks have been proposed for the regular and ongoing management of
watershed resources. These regular processes can spawn situational analyses which may be
portrayed as linked cycles.15 For example, the U.S. Forest Service (USFS) uses a planning
process (Figure 3-A-5) to guide the ongoing management of national forests and grasslands.16'17
The spatial scale of planning ranges from national to regional to local, and it can be done at the
watershed level if appropriate to the scope and scale of issues addressed. Existing plans
authorize site-specific management actions, and outcomes are monitored and evaluated
according to plan criteria in an adaptive cycle. New rounds of planning are undertaken after 15
years or as necessitated by issues or conditions. Stakeholders play an important role in the initial
development of goals and are encouraged to participate in subsequent steps; participation
opportunities are to be early, frequent, open and meaningful, and stakeholders may lodge
objections before decisions are taken. Information development includes baseline analyses of
both ecological and economic sustainability of current forest or grassland management practice.
Ecological analyses include the effects of current or anticipated human disturbance (as compared
to natural and historical human disturbance) upon ecosystem processes and system and species
3-46
-------
Initial consultation
Assessment of
sustainability and
plan revision
Planning
Identification of
issues, opportunities
or new information
Issues, \
Context *
Changes
Plan
Site-specific
Decisions
Adaptive
Management
Monitoring
and Evaluation
x*
T
Public
comment
** Amendment
* or Revision
FIGURE 3-A-5
The USFS planning framework incorporates regular adaptive management and situational
planning processes.
3-47
-------
diversity. Social and economic analyses examine the benefits provided by forest lands, social
and economic trends, and the society-forest relationship.
Many U.S. states have adopted a watershed management cycle, sometimes referred to as a
"rotating basin approach," for implementation of the regulatory requirements and other programs
of the Clean Water Act (CWA).18'19 Whereas the approach usually is adopted to improve State
agency efficiency, in most cases it has led to enhanced involvement of stakeholders as well, and
the trend is toward more localized, partnership-based approaches driven by multi-stakeholder
teams.19 Typically, the state is divided into major watershed units, and CWA activities are
implemented on a roughly five-year activity cycle that is staggered to begin in different years by
watershed (Figure 3-A-6). The cycle begins with monitoring and assessment and continues
through planning and implementation. "Assessment" as referred to here entails comparison of
monitoring data and Water Quality Standards (WQS), a process which should detect likely
adverse effects from stressors for which WQS have been determined but which falls short of risk
assessment per se (see Section 2.3). While economic or other social-science studies are not
precluded as part of this process, natural science is emphasized, hi theory, activities such as
review of designated uses, listing of impaired waters, issuance or review of point-source
discharge permits, and award of loans and grants for water quality improvement projects are
carried out in the implementation phase of this cycle, although in practice limited resources and
competing priorities make this difficult to accomplish.19 Total Maximum Daily Loads (TMDLs)
may be developed and implemented for high-priority impaired waters; here the TMDL process is
depicted as a situational cycle linked to the regular management cycle (Figure 3-A-6).
3-48
-------
Management
Cycle
Implementation
plan
Monitoring plan
& schedule
Allocate loads
Link targets
and sources
TMDL
Development
and
Implementation
Assigning
Priorities and
Targeting
Resources
Developing
Management
Strategies
FIGURE 3-A-6
The watershed-based management cycle used by many states may include TMDL development
and implementation (Adapted from USEPA18)
3-49
-------
REFERENCES
1. USEPA, Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F, Risk Assessment
Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
2. USEPA, A Framework for the Economic Assessment of Ecological Benefits, Science Policy
Council, U.S. Environmental Protection Agency, Washington, DC, Feb. 1, 2002.
3. USEPA, Environmental Monitoring and Assessment Program (EMAP) Research Strategy,
EPA/620/R-98/001, U.S. Environmental Protection Agency, Washington, DC, 1997.
4. Walmsley, J.J., Framework for measuring sustainable development in catchment systems,
Environmental Management, 29, 195,2002.
5. Barber, M.C., Environmental Monitoring and Assessment Program Indicator Development
Strategy, EPA/620/R-94/022, U.S. Environmental Protection Agency, Office of Research and
Development, Athens, GA, 1994.
6. Jackson, L.E., Kurtz, J.C., and Fisher, W.S., Evaluation Guidelines for Ecological Indicators,
EPA/620/R-99/005, U.S. Environmental Protection Agency, Office of Research and
Development, Washington DC, 2000.
7. Timmerman, J.G., Ottens, J.J., and Ward, R.C., The information cycle as a framework for
defining information goals for water-quality monitoring, Environmental Management, 25,
229, 2000.
3-50
-------
8. Stahl, R.G. et al., Risk Management: Ecological Risk-Based Decision-Making, Society for
Environmental Toxicology and Chemistry, Pensacola, FL, 2001.
9. PCCRARM, Framework for Environmental Health Risk Management,
Presidential/Congressional Commission on Risk Assessment and Risk Management,
Washington, DC, 1997.
10. SAB, Toward Integrated Environmental Decision-Making, EPA-SAB-EC-00-011, U.S.
Environmental Protection Agency, Science Advisory Board, Integrated Risk Project,
Washington, DC, 2000.
11. USAGE, Planning Guidance Notebook, ER 1105-2-100, U.S. Army Corps of Engineers,
Washington, DC, 2000.
12. World Commission on Dams, Dams and Development: A New Framework for Decision-
Making, Earthscan Publications Ltd., London, 2002.
13. USEPA, Watershed Protection: a Project Focus, EPA 841-R-95-003, Office of Water,
Environmental Protection Agency, Washington DC, 1995.
14. Davenport, T.E., The Watershed Project Management Guide, Lewis Publishers, Boca Raton,
FL, 2002.
3-51
-------
15. Cole, R.A., Feather, T.D., and Letting, P.K., Improving Watershed Planning and
Management Through Integration: A Critical Review of Federal Opportunities, IWR Report
02-R-6, U.S. Army Corps of Engineers, Institute for Water Resources, Alexandria, VA,
2002.
16. USFS, National forest system land and resource management planning, Federal Register, 65,
67513,2000.
17. USFS, National Forest System Land and Resource Management Planning; Proposed Rules,
Federal Register, 67, 72769,2002.
18. USEPA, Watershed Protection: a Statewide Approach, EPA 841-R-95-004, Office of Water,
U.S. Environmental Protection Agency, Washington DC, 1995.
19. USEPA, A Review of Statewide Watershed Management Approaches, Office of Water, U.S.
Environmental Protection Agency, Washington, DC, 2002.
3-52
-------
4. EVALUATING DEVELOPMENT ALTERNATIVES FOR A HIGH-QUALITY
STREAM THREATENED BY URBANIZATION: BIG DARBY CREEK WATERSHED
A vision for integrating ecological risk assessment (ERA), economics and watershed
decision processes has been presented in the previous chapter. The objective in this chapter is to
consider a case study in which certain elements of that conceptual approach (see Figure 3-1) are
implemented and field-tested with specific data. A large watershed in central Ohio, the Big
Darby Creek, provides the locale and basis for the study design.
In 1993, the Big Darby Creek watershed was selected by the U.S. Environmental
Protection Agency (USEPA) for one of five watershed ecological risk assessment (W-ERA) case
studies for several reasons: the substantial interest by organizations at the local, state and federal
level in protecting the watershed; the outstanding character of the aquatic biological resource; the
range of sources and stressors (agricultural nonpoint sources, urban nonpoint sources, permitted
discharges, etc.); the existence of a large, multiple year, watershed-wide database; and a
commitment by Ohio EPA (OEPA) to co-lead the risk assessment team.
In 1999, while the W-ERA was in the later stages of completion, a USEPA-funded study
was initiated by Miami University with the goal of integrating ERA and economic analysis to
further inform environmental management efforts in the Big Darby Creek watershed. The
methodological framework for this integrated research was rooted in a broadly based approach to
sustainability that encompasses, but extends beyond, ERA. This approach views economic
development as complementary with, rather than antagonistic to the maintenance of non-
renewable resources. As such, it argues that sustainable systems require coordination between
4-1
-------
ecological, economic, and social considerations in order to maintain overall system resilience.
Because the Miami University study was initiated well after the Big Darby Creek
W-ERA, using information from the latter but carried out by a separate team, the two efforts
were not integrated in an ideal sense (see Section 1.5.1), However, the research approach used
illustrates some of the advantages, as well as the difficulties, of integrated study. Section 4.1
describes the watershed setting, and Section 4.2 discusses the W-ERA effort and its findings.
The Miami University study is presented in Section 4.3, and Section 4.4 discusses these findings
in light of the larger integration problem.
4.1 WATERSHED DESCRIPTION
Big Darby Creek is a high-quality, warm water stream located in the Eastern Corn Belt
Plains ecoregion of the Midwest (Figure 4-1). The watershed encompasses 1443 km2 (557 mi2)
and is home to a diverse community of aquatic organisms including many rare and endangered
fish and freshwater mussel species. The Big Darby Creek watershed was given a conservation
priority by The Nature Conservancy (TNC) through its recognition as one of the "Last Great
Places" in the western hemisphere.2'3 The risks to ecological resources in the Big Darby
watershed derive from ongoing changes in agriculture and suburban land use.
The watershed drains portions of six counties in rural Ohio just west of Columbus.
Agriculture currently comprises 92.4% of the land use of the watershed. Cropland, most of
which is actively row-cropped, is the highest use (72%), followed by livestock pasture (8.6%).
However, suburban Columbus is expanding westward in the Big Darby watershed. Currently,
the western tributaries drain agricultural lands almost exclusively, whereas the eastern
4-2
-------
10 20 30 40 50 Kilometers
A
FIGURE 4-1
The Big Darby Creek watershed in central Ohio, USA. The Columbus metropolitan area is
expanding into the easternmost area of the watershed, where Hellbranch Run is especially
affected. Respondents surveyed in this study were drawn either from the watershed area,
Columbus, or Oxford.
4-3
-------
tributaries drain areas with increasing suburban and commercial/industrial land use. Urban
development recently has quadrupled in some areas, with significant negative consequences for
stream habitat. Although there have been recent improvements in fish and invertebrate indices in
the Big Darby Creek mainstem, the easterly Hellbranch Run shows degradation. ' A number of
stream reaches in the watershed have been listed as impaired and are subject to potential
regulation through development of total maximum daily loads (TMDLs, see Sections 2.1.2 and
2.3.1), mostly focused on phosphorous, nitrogen and sediment.
To the west from Hellbraneh Run, the urban and industrial impacts are generally not
greater than agricultural impacts, but given the present population of the region and the rapid rate
of development, urban water pollution problems are a risk for a large part of the Big Darby
watershed in the future. Without management, the increased frequency of damaging storm
runoff and associated pollutant loads pose risks to the uncommon species, game fish and general
aquatic system functioning. These are risks that could be reduced through best management
practices for both urban and agricultural runoff.5
4.2 ECOLOGICAL RISK ASSESSMENT
The phases of ERA as described in USEPA's Guidelines,6 i.e., planning, problem
formulation, analysis, and risk characterization, are summarized in Section 2.1.1. This section
describes the work that was conducted in each phase of the W-ERA for Big Darby Creek.
4.2.1 Planning
The OEPA database available for this assessment included standard water quality
parameters such as suspended and dissolved solids, pH, oxygen-demanding substances, nutrients,
ammonia and metals. It also included biological assemblage data describing the presence and
4-4
-------
abundance offish species and of macroscopic sediment-dwelling invertebrates (termed benthic
macroinyertebrates or benthos) collected by standard sampling procedures. Also available were
a set of descriptors of stream corridor condition, including condition of substrates, instream
habitat types (pools, riffles), channel stability and riparian zone vegetation. Multimetric indices
that provided a composite assessment of habitat or biological quality, based on these data,
included the Qualitative Habitat Evaluation Index (QHEI) for stream corridor condition; the
Index of Biotic Integrity (IBI) and the Modified Index of Well-being (Mlwb), which are
measures of the functional and structural organization of the fish community, respectively; and
the Invertebrate Community Index (ICI), which evaluates the structural organization of the
macroinvertebrate community. These indices have been used extensively by the OEPA to
establish biological criteria and to evaluate stream use attainment (see Section 2.3.1 and
Appendix 2-B for further description of these indices).7
Cooperators in the Big Darby Creek Watershed Ecological Risk Assessment included the
W-ERA team co-chairs from USEPA and OEPA and at various times representatives from The
Ohio State University, The Nature Conservancy, the United States Geological Survey (USGS)
and Operation Future, a conservation oriented farm group. Management goals for the risk
assessment were developed through review of pertinent regulations, discussions with residents
and resource managers, and meetings with the Darby Partners, a loose-knit group of over 40
public agencies and private organizations united by the shared goal of watershed protection. The
overarching risk reduction goal from these discussions was to "protect and maintain native
stream communities of the Big Darby ecosystem." Three specific objectives were seen as
necessary to meet this risk reduction target:
4-5
-------
1. Attaining criteria for designated uses throughout the watershed (see Section 2.3.1)
2. Maintaining OEPA's exceptional warm water criteria for all stream segments
having that designation between 1990 and 1995
3. Ensuring the continued existence of all native species in the watershed.
The risk management problem was to ensure that these specific objectives could be met.
The risk characterization would require understanding how various environmental factors might
prevent meeting these objectives.
4.2.2 Problem formulation
Ecological assessment endpoints are measurable attributes of valued ecological
characteristics. Two assessment endpoints were chosen for the Big Darby risk assessment:
1. Species composition, diversity and functional organization of the fish and
maeroinvertehrate communities
2. Sustainability of native fish and mussel species.
From a practical standpoint, the first of these endpoints could be evaluated utilizing three
composite indices (D3I, Mlwb and ICI) and the individual measures they comprise. It was
determined ultimately, however, that while some of the available data were relevant to the
second endpoint, the necessary information on life history and genetic diversity of native species
in the watershed was not sufficient for evaluating their sustamability,3 Therefore, only the first
endpoint was carried further.
A critical step in problem formulation is the development of a conceptual model. It
articulates the risk assessors' hypotheses on the relationships among the sources of stress,
stressors, effects, and endpoints. Six significant stressors were identified for this watershed as
4-6
-------
affecting the assessment endpoint: altered stream morphology, increased flow extremes,
sediment, nutrients, temperature and toxicants. A conceptual model illustrating the hypothesized
relationship between land use, sources of stress, the aforementioned stressors, subsequent
ecological interactions and the stressor signatures (i.e., characteristic changes in aquatic
community metrics) is presented elsewhere.3
Seven risk hypotheses were developed based on the relationships inherent in the
conceptual model:
1. No differences exist in community structure and function among the
subwatersheds
2. No differences exist in community structure and function among time periods
3. Community structure and function will decline downstream from identified point
sources
4. An increase in certain land uses or land use activities will result in a change in the
IBI and/or the ICI
5. An increase in certain land uses or land use activities will result in an increase in
the intensity or spatial or temporal extent of in-stream stressors
6. An increase in the intensity, or spatial or temporal extent of in-stream stressors
will result in a change in the biological community as quantified by ICI and IBI
metrics and species abundances
7. The pattern of response of the stream community can discriminate among the
different type of stressors.
4-7
-------
The first two were null hypotheses; analysis would determine whether they could be
statistically rejected. The other five were maintained hypotheses, thought to be true; analysis
would seek confirmatory or contradictory evidence.
4.2.3 Current status of analysis and risk characterization
The analysis to test these hypotheses was carried out in two phases. Hypotheses 1 and 2
were tested by analyzing historical biological assemblage data within the Big Darby Creek
watershed.4 Both hypotheses were rejected because certain spatiotemporal differences were
shown in the analysis. Time series analysis, which was feasible for fish community metrics and
IBI within the Big Darby Creek mainstem, indicated a general improvement over the time period
1979 - 1993, At the same time, spatial comparisons among the Big Darby, Little Darby and
Hellbranch Run subwatersheds revealed significant spatial differences for IBI, ICI and several
component metrics. In general, the Big Darby Creek mainstem showed superior biotic
condition; however, some of this difference could be attributed simply to its comparatively larger
drainage area. After correction for drainage area, many differences disappeared, but the biotic
condition of the urbanized Hellbranch Run remained lower than the mainstem according to
several measures.4 These findings, while encouraging for the watershed as a whole, were
consistent with concerns that suburban encroachment threatens watershed ecological resources in
the eastern portion of the watershed. However, without an ability to correlate biological
condition with stressors of concern or their sources, these results were of limited value for
assessing risks associated with likely future changes in the watershed.
By contrast, hypotheses 3 to 7 required the analysis of point sources, land uses and
stressors in spatial relation to biological data. Relatively few point sources of pollution are
4-8
-------
present in the watershed but most have shown negative effects on the mussel community for
some distance downstream. Migration of species within the fish community making up IBI
tended to remove the downstream effect. Thus, hypothesis 3 was confirmed for metrics focused
on invertebrate species, but not for free swimming migratory species such as fish.
Initial attempts to analyze stressor effects derived from land use patterns were
complicated by the watershed's relatively good water quality and higher than average IBI. The
narrow range of variability in the biotic metrics and the chemical and physical parameters seen in
the Big Darby needed to be assessed in the context of the greater variability of the region as a
whole. Therefore, Norton et al.8 analyzed biological, chemical and habitat data for the Big
Darby Creek and other comparably-sized watersheds within the Eastern Com Belt Plains
ecoregion in Ohio, among which a wider gradient of the stressors and subsequent responses
could be observed. Discriminant functions constructed using biological variables from this
larger dataset were used to separate site groups into high-, medium-, and low-stress categories
along stressor gradients. Analysis of the biological variables here did distinguish between
higher- and lower-quality sites classified on the basis of six different types of stressors: degraded
stream corridor structure; degree of siltation; total suspended solids, iron, and biochemical
oxygen demand (BOD); chemical oxygen demand (COD) and BOD; lead and zinc; and nitrogen
and phosphorus. Functions based on biological variables could also discriminate between sites
having different dominant stressors.8
Using somewhat different methods for their data aggregation and analysis, Gordon and
Majumder9 analyzed similar data, but they also included land use (dense urban, forested or
agricultural, as a percentage of each watershed) in an effort to develop regression models that
4.9
-------
could predict the ecological effects of future land use changes. A number of models showed
some ability to explain average watershed IB I. For a set of 137 watersheds, the regression model
explained 39,5 % of variance in the IB I when only stream corridor characteristics, land use and
stream order were included (N = 467), 47.4 % when an index of chemical pollution stress was
added to the model (N = 196), and 65.5 % when upstream IB I was added (to correct for spatial
autocorrelation, N = 177). Percent dense urban land use was a strongly negative predictor. For
the three models described, standardized regression coefficients for percent dense urban land use
(which relate the variance in that factor to the variance in JOB!) were -0,305, -0.258 and -0.179,
respectively. '
Therefore, hypotheses 4-7 were shown to hold true for the Eastern Corn Belt Plains
ecoregion, and the relationships found can reasonably be applied in the Big Darby Creek
watershed." These preliminary results suggested that fish and macroinvertebrate community
responses to land use, stream corridor habitat and various chemical stressors are predictable to a
degree, USEPA's efforts to apply these findings to the assessment of ecological risks in the Big
Darby Creek watershed are still ongoing. Additionally, because of the identified impairments to
some of its subwatersheds, Big Darby Creek is subject to the development of a TMDL by OEPA.
Similarly, in an effort to assist planners, environmental organizations, government agencies, and
concerned citizens, scientists and planners in The Ohio State University's City and Regional
Planning Program, working on a USEPA-funded grant, have created an interactive, geographic
information systems (GlS)-based screening tool to evaluate the biological effects of various
changes within the Big Darby Creek watershed and other watersheds within the Eastern Cornbelt
Plains ecoregion.10
4-10
-------
4.3 ECONOMIC ANALYSIS
The overarching goal for Miami University's integrated ecological and economic analysis
was to utilize the findings of ERA in an economic analysis that would be relevant to
environmental management decisions in the watershed. At the time of initiating our integrated
study in the Big Darby Creek watershed, the problem-formulation phase and early portion of the
analysis phase of ERA had provided a clear picture of current conditions and apparent threats.
Because the spatial scope of the analysis had to be expanded to all Eastern Corn Belt Plains
watersheds in Ohio, a full complement of stressor-response or source-response relationships was
not yet available, and the risk characterization had not been carried out. However, the following
sections show that sufficient information was available for meaningful analysis.
The objective for this integrated case study was to undertake an analysis capable of
informing decisions about reducing risks from suburban development. An independent modeling
study sponsored by Miami University's Center for Sustainable Systems Studies11 had quantified
the range of effects on hydrology, sediment transport, and nitrogen concentrations from changes
in land use. This study had considered three types of residential development in the Big Darby
basin and had found that two different types of low density development protected the stream
amenities very well. The analyses by Norton et al.8 also informed the selection of stressors
considered to be key influences on stream conditions following urbanization. Thus, the goal for
this case study was an integrated evaluation of ecological and socioeconomic impacts associated
with several land use approaches at the peri-urban fringe.
The specific objectives of the case study, therefore, were as follows: (a) to estimate the
quantitative or qualitative impacts of a set of land use scenarios on stream ecological condition,
4-11
-------
local economic well-being and local quality of life, (b) to communicate these impacts to the
public effectively, and to measure the overall economic value (see Section 2.2.2) corresponding
to each scenario based on individual willingness to pay (WTP), and finally (c) to better
understand the particular contribution stream ecological condition makes to the value of a given
scenario.
4.3.1 Research approach
Based on prior work in the Big Darby watershed,11'12 four development scenarios were
used to compare outcomes for stream amenities: (1) a most-likely case of high density,
conventional subdivisions using 1A- to 1-acre lots with water and sewer services discharging to
the Big Darby; (2) a low-density ranchette development on 3- to 5-acre lots with local water and
septic system disposal; (3) a low density cluster development, with intervals between clusters to
achieve the same housing density as ranchettes (e.g., as maintained through purchase or set-aside
of transferable development rights); and (4) a reference case of continued agriculture, which was
the predominant land use pattern actually observed in the 1990s.
A dichotomous-choice contingent valuation method (CVM) survey instrument (see
Section 2.2.2 and Appendix 2-A for a discussion of this method) was developed that allowed
presentation of technical information on how changes in stream amenities are induced or avoided
during land development, followed by expression of WTP for a certain outcome. Analysis was
also carried out to develop a quantitative relationship between the four land use scenarios and
stream biological integrity based on empirical relationships.
The survey approach involved in-person, multimedia presentations to noninteracting
groups of 30-50 respondents who completed a questionnaire. The instrument was designed
4-12
-------
according to Arrow et al.13 and implemented according to Dillman.14 In addition, multiple
stakeholders were brought into a pretest phase to gain insight about their viewpoints, as well as
their suggestions about refining the survey instrument.
The survey/presentation was divided into three parts. The first section asked respondents
their knowledge about and use of the Darby Creek before we provided information about this
watershed. An example of such a question is: "Do you believe some types of residential
development lead to increased soil erosion and runoff" of fertilizers and pesticides? " hi this case,
90% of respondents answered that they were aware of this issue. A follow-up question for those
who answered yes to the above question asked: "If so, do you think these runoff products do
significant damage to streams and water quality? " Again, a significant portion, 85%, of the
respondents answered yes. Finally, those respondents who answered yes to both questions were
asked: "If so, do you think these runoff products will do damage to fish and other species in the
stream?" About 85% responded that they were aware of the damage of the runoff.
The second section was designed to engage the respondents as they were presented the
effects that development might have on the environmental, social, and economic characteristics
of the area. These effects are discussed below, but the reader must note that as this material was
presented, respondents were asked questions about it. Many questions related back to the first
section. For example they were asked: "Did you know that the food base for many fish was tiny
insect larvae that live on the stream bottom? "' About 30% of the respondents were not aware of
this before the presentation, hi another example, respondents were asked: "Didyou know that
lawn and garden chemicals could affect the fish in the stream?" Consistent with the results in the
first section, 94% indicated that yes they were aware of this. The other 6% were made aware by
4-13
-------
the presentation. Thus it was possible to create a uniform minimum knowledge base across
respondents. The final section was the valuation and demographics questions. These results are
discussed in greater detail below.
The respondents were drawn from three different populations. "Residents" were defined
as people who live within the study area of the Big Darby Creek watershed, both farmer and non-
farmer. "Near-Residents" were people living outside the watershed but within the greater
Columbus, Ohio metropolitan area. "Non-Residents" were drawn from people in the area
surrounding Oxford, Ohio, a two-hour drive from Big Darby Creek. Residents and near-
residents capture the value attached by people who use the area for residence or recreation (use
value). These two groups also capture nonuse value if they value the watershed solely for the
benefit of acknowledging its existence or are willing to contribute to preservation of future use.
Non-residents may use the area for recreation, but at a much lesser rate than those in the
Columbus area. The primary value for this group was expected to be nonuse value (see Section
2.2.2).
Samples were drawn at random from zip codes contained within each of the targeted
areas. Respondents received a payment of $30 to cover their out-of-pocket travel costs and to
show appreciation for the time spent in the one-hour presentation and survey response. Each
respondent later received a mailed summary of survey results. The following sections develop
the sequence of topics covered in the presentation to survey respondents and provide details on
the valuation question.
4.3.2 Communicating the effects of urban development on ecological endpoints
Scientific understanding of the mechanisms by which residential development brings
4-14
-------
about change in streams generally can be reduced to four causal factor groups:
1. increased nutrients (which increase algal growth and affect the kinds offish)
2. increased sediment (which decreases light penetration and affects the food chain)
3. increased toxic substances (which cause mortality in the food chain and fish)
4. increased runoff and flooding (which allow bank erosion and sedimentation).
The most difficult challenge for this project concerned the need to have the public (represented
as survey respondents) understand the mechanisms inducing change in the stream well enough
for them to attach value to the outcomes they prefer. This question of informing respondents
about linkage mechanisms was addressed by presenting the following synthesis on watershed
processes and ERA.
4.3.2.1 Increased nutrients, leading to change in fish species
Nutrients were described as chemicals that enhance growth of plants, on land as well as
in the water. In high-density subdivisions, nutrients come from lawn fertilizers, from storm
water runoff and, at some "downstream" locations, from household sewage. Runoff also carries
soil and fertilizers from farmland, further enriching streams. The amount of nutrients entering
the stream has been shown to depend on the number of people living in an area and how they
manage their fields, lawns or gardens. Nutrient loading also has been shown to depend on the
amount of hard surfaces (roads and roofs) developed in a neighborhood.
The main effects of increased nutrients on the amenities to be valued are nuisance-level
growths of algae in streams. This increased growth can cause a change in water quality and the
kinds offish that live there. In enriched streams, fish species that feed on decaying stream-
bottoms (many minnows, carp and catfish) are favored over those predator fish (e.g., bass and
4-15
-------
sunfishes) that feed on small fish. If nutrient input is very high, "fish kills" can occur. It was
then possible to ask the respondents what types offish they would rather have in a stream, and
they were given the choice of (1) minnows, sunfishes and carp, or (2) bass, sunfishes and darters.
4.3.2.2 Increased sediment, leading to a decreased insect food base in
streams
Survey respondents also were told and shown that the amount of sediment entering a
stream from residential areas or farmland can vary widely, but poses a serious problem. During
initial construction, erosion from bare soil can be very high during heavy rainfall. Large
amounts of soil can enter the stream and remain there for years, despite being mobilized after
every major rain event. After construction, less sediment enters from residential areas than from
agricultural land that exhibits standard row crop cultivation, seasonally bare soils, and livestock
wading in and alongside streams.
The main effect of sediment is to decrease the quality of fish and invertebrate habitat by
filling small spaces between pebbles in the stream bottom that are normally home for insect
larvae. Such insects are the main food for many types offish, part of the rich ecological
diversity in the Big Darby Creek. Without these insects the number and kinds offish decrease.
Furthermore, very fine particles are shown to stay suspended in the flowing water, making it
cloudy and also decreasing the ability of fish to find their prey.
4.3.2.3 Toiic substances, changing the insect food base, fish species and
causing disease
In the section on effects from toxic substances in runoff, the survey respondents were
given information on how storm runoff washes pesticides from cropland, lawns or gardens into
streams. Such compounds also may come from spilled oil, gasoline, and other automotive
4-16
-------
chemicals present on roads and driveways. These chemicals often cause change in the numbers
and kinds offish in streams, favoring fish that tolerate these substances. Respondents were asked
whether they knew that lawn and garden chemicals could affect the fish in streams.
4.3.2.4 Changes in runoff and flooding patterns, decreasing habitat quality
and causing a shift to fewer, more tolerant species
Natural streams were described as having bends, pools and riffles, with logs and limb
"dams" all the way to their headwaters, thus slowing passage of water. Slow natural drainage
from the land also allows water to seep into the ground slowly after heavy rains, replenishing
ground water. However, with some residential development, streams are straightened, logjams
are removed and storm water drains quickly off the land, increasing the risk of downstream
flooding.
The effects of these physical changes were described in the presentation as increasing the
speed of water flow, causing further erosion from stream banks and increasing flood heights.
After the runoff, water flow can become quite low in the absence of a strong groundwater
recharge. These alternating high and low flows drastically change the quality offish habitat,
reducing biological diversity. Instead of many different depths and bottom types, the channel
becomes wide and shallow. During low-flow periods, water moving over or through a gravel
base becomes too shallow to be inhabitable. The resulting crowded conditions lead to increased
death rates for fish as they use up nearly all of the available oxygen.
4.3.3 Communicating the effects of urban development on economic and social
services
While the information above sought to frame certain values attached to the Big Darby,
respondents also derive other kinds of value from economic and social functions within the area.
4-17
-------
To isolate the value placed on ecological services, one must control for value related to the
economic and social services. Accordingly, the economic and social dimensions included in a
sustainable development framework1 also were briefly described.
In considering the value of economic services, the dominant endpoint is increased
economic well-being. Although many measures that contribute to economic well-being were
considered, the presentation focused on four economic outcomes: (1) dependence upon
agricultural employment; (2) distance to employment for non-agricultural workers; (3) provision
of retail services; and (4) impact on the local income base. Employment opportunities for
agricultural and non-agricultural workers can be expected to change significantly across the four
development scenarios. As residential development increases, agricultural employment
opportunities will decline, but there would be sufficient population growth to justify expansion
of retail services. Dependence upon commuting for non-agricultural work not only involves
travel costs and value of time dimensions, but also has feedback effects as commuters either
make purchases outside of the Big Darby or, conversely, bring higher incomes back to the local
area. This is one of the ways in which development would be expected to affect the local income
base, hi addition, the income profile of residents who would be expected to populate the study
area would vary under the different scenarios. Questions were included to capture respondent
preferences about these economic outcomes.
With respect also to social services, the ultimate endpoint is increased quality of life.
Among the many factors that contribute to quality of life, the presentation focused on four social
outcomes; (1) open space, (2) privacy, (3) public services, and (4) quality of education. These
factors vary among the different development scenarios as well. The change in open space and
4-18
-------
privacy during the transformation from rural to suburban could be a confounding variable of
importance to respondents. As residential development progresses, the availability of open space
for use in recreational activities and the degree of privacy begin to decline. In addition,
residential development not only brings a need for increased public services, such as police and
fire services, but also difference in access as the proximity to these services changes. Moreover,
the quality of publicly provided elementary and secondary education is likely to change with
increases in local income and property wealth, and as voter tastes for education change.
Questions were included to capture respondent preferences about these social outcomes.
4.3.4 Land use scenarios for framing expression of preference and value in the
stream
All the variables considered in the previous two sections vary among the land
management or development options, allowing an approach that estimates stakeholder value
through CVM surveys. The CVM questionnaire tries to focus on the unique amenities that could
be at risk while acknowledging that other factors come into play. The information provided to
survey respondents about physical stressors and ecological, economic, and social mechanisms
can affect the estimate of WTP in terms of the direction and magnitude of the potential bias.16'17
Thus, the survey instrument must have questions concerning preferences as well as values. To
facilitate an understanding of the contrasts in options and outcomes, maps, data and photographs
were used to frame WTP to conserve amenities described in each of the development scenarios
below.
For easy reference, survey respondents were provided Table 4-la and Table 4-lb which
show the levels of effect from each of the objective factors considered in the section on linking
mechanisms. In each case, a range of possible effects was described, categorized as low,
4-19
-------
TABLE 4-la
Relative effect of four housing development scenarios on the four main causes of change
in Big Darby Creek
Nutrient input
Sediment input
Toxin input
Change in flow
patterns
High Density
Development
Medium to high
Low to high
Medium to high
High
Low Density
Ranchettes
Low to high
Low to medium
Low to high
Low to medium
Low Density
Clusters
Low to medium
Low to medium
Low to medium
Low to medium
Agriculture
Medium to high
Medium to high
Medium to high
Medium to high
4-20
-------
TABLE 4-lb
Relative effect of four housing development scenarios on socioeconomic outcomes in Big
Darby Creek
High Density
Development
Low Density
Ranchettes
Low Density
Clusters
Agriculture
Economic
Outcomes
Agricultural
employment
Retail services
Distance to
employment for
non-agricultural
workers
Local income
base
Low to Medium
High
Low
High
Low
High
Medium
Medium
Medium to High
Medium
Medium
Medium
High
Low
Medium to High
Low to Medium
Social
Outcomes
Open space
Privacy
Proximity to
police and fire
services
Quality of
education
Low
Low
High
Medium to High
Medium to High
High
Medium
Medium
Medium to High
Medium
Medium
Medium
High
High
Low
Low
4-21
-------
medium, or high in both the script and color slides. These categories are intended to reflect
increasing levels of risk. For instance, low nutrient input would be that input leading to nutrient
concentrations in the stream that are in the range of the lowest 1/3 of the observed data on
nutrient concentrations. The factors are normalized such that when the effect reaches high
levels, there is risk to stream integrity.
4.3.4.1 High density development
The base case against which the respondents are asked to indicate preferences or WTP (to
avoid) is illustrated in Figure 4-2a. It shows a 4-mi2 area that includes both sides of the Big
Darby, not far off 1-70. It represents the conventional residential development that many people
expect based on the patterns already being seen in the Columbus area. The characteristics
defined into this high-density scenario are: 15% open or agriculture, 70% residential, 10% forest,
and 5% nature preserve. The lot size is about 1A to 1 acre and the residential density is 200
dwelling units per 100 acres of land.
Nutrient input is affected by storm water runoff that carries lawn fertilizers at certain
times of the year. In this scenario, the aggregate effect is expected to be medium to high.
Sediment input from this scenario will be high during construction periods, but then may be
fairly low. Toxin input will be medium to high depending on lawn and garden care practices,
and whether a storm water treatment system is in place to treat the chemicals scavenged from
roads and driveways. The pattern of stream flow, flood frequency and scouring is changed
considerably, mainly due to the very large increase in hard surfaces. The respondents were
asked several questions as to their preferences for avoiding associated enrichment, toxins and
extreme flow outcomes.
4-22
-------
Roads
Streams
Forests
Hills
FIGURE 4-2a
Illustration of high density scenario
(dots represent houses)
4-23
-------
4.3.4.2 Low density ranchette development
A second scenario, shown in Figure 4-2b, illustrates the same 4-mi area but with
development in the form of large lots, based on the patterns already observed in many suburban
areas. The characteristics defined for this "ranchette" residential development were: 10%
agriculture, 70% residential, 15% forest, and 5% natural preserves. The dwelling unit density is
20 units per 100 acres, with 3 to 5 acre lots.
The inputs of nutrients and toxins can vary from low to high in this scenario depending
on how much of each lot is left in natural vegetation and how the lawns are maintained. Some
nutrient input to the stream from septic tank seepage also is possible. When few pesticides are
used on lawns and much of the land is left in a "natural" state, then both nutrient and toxin input
will be much lower than in the high-density scenario. However, when large areas are maintained
as lawns using standard chemical lawn treatments, then both nutrients and toxins could be almost
as high as the high-density scenario.
Sediment input also will range widely, from low to medium, with some entering the
stream mainly during the construction phase, and tending to be much less over time. Changes in
stream flow peaks will be low to medium, and much less than the high-density scenario. In this
scenario, stream habitat will depend largely on the amount of forest and wetlands left near the
stream channel. In comparison to conventional agriculture, however, the overall change in the
Big Darby system from large lot development is likely to be positive. The survey respondents
were asked whether it is likely that residents of this ranchette type of development will leave
enough land in its natural state to protect Big Darby water quality, and whether they would be
4-24
-------
Roads
Streams
Forests
Hills
FIGURE 4-2b
Illustration of low density ranchette scenario
(dots represent houses)
4-25
-------
willing to pay slightly higher land and construction costs to guarantee that sediment input to the
creek is minimized by erosion barriers and sediment traps. They also were asked whether taking
over nearly all the farmland is a significant negative consideration for them.
4.3.4.3 Low density cluster development
A third scenario, shown in Figure 4-2c, illustrates the same 4-mi2 area, but with a
clustered development that keeps most of the land in agriculture. The characteristics defined into
this type of development are: 60% agriculture, 20% residential, 15% forest, and 5% nature
preserves. The dwelling unit density is 20 units per 100 acres, the same as for the ranchette
development.
Nutrient input from this scenario is shown to vary from low to medium depending
primarily on associated farming practices. The cluster housing developments would each
include their own sewage treatment system, possibly in the form of package treatment and
wetland wastewater application, with little input to the creek. Maintenance of lawn area also
would contribute little because of the small lot sizes for housing. Nutrient input from farms may
be insignificant, depending on fertilizer applications and the density of livestock.
Sediment input will vary here much as it does in the ranchette scenarios, with higher
input during construction, decreasing with time. Because the amount of bare land in hard-
surface roadways is less than in either of the other two residential scenarios, overall sediment
input even during construction will be low to medium, with the input determined by the amount
of land left in agriculture. Soil-conserving agricultural practices such as low-tillage could
decrease the sediment load even further. Toxin input will be lower in this scenario than for
either of the other residential developments because of the smaller area of lawns and hard
4-26
-------
Roads
Streams
Forests
Hills
FIGURE 4-2c
Illustration of low density cluster scenario
(dots represent houses)
4-27
-------
surfaces, but the range of agricultural practices largely will determine the level of toxins reaching
the stream. The altering of stream flow and flooding pattern is lower here than for the
agriculture or high density scenarios.
4.3.4.4 Agriculture land use
The final scenario is shown in Figure 4-2d. This scenario shows the land use and
residence density actually observed in the area in the early 1990's. The characteristics of this
"present landscape base case" are: 75% agriculture, 10% "residential" (including farm lawns),
and 15% forest. The dwelling unit density is 2 units per 100 acres.
The input of nutrients, sediment, and toxins in this scenario can be medium to high,
depending on local agricultural practices and the amount of livestock (see Table 4-la). The time
of cultivation and the amount of fertilizer and pesticide application also influence the amount of
sediment, nutrients and toxins in runoff reaching the stream. Certain farming practices can be
adopted to reduce fertilizer applications and minimize runoff after rain events. However, many
farmers in the Big Darby drainage area already use conservation tillage practices to reduce
nutrient, pesticide, and sediment inputs.
The altering of stream flow characteristics under this scenario is medium to high (relative
to a pristine, unfarmed condition), also depending on farming practices. Because the Big Darby
area is fairly flat, water does not flow to the stream quickly, and farmers are often anxious to
drain the water off their fields. Tile drainage systems and straight clean waterways have been
introduced locally, increasing water flow and transport of nutrients off the land. The survey
respondents were asked how important it is to them that a large portion of the Big Darby
watershed be retained in agricultural land use.
4-28
-------
Roads
Streams
Forests
Hills
FIGURE 4-2d
Illustration of present agriculture scenario
(dots represent houses)
4-29
-------
4.3.5 Eliciting monetary valuation
The four scenarios, and the ecological, economic, and social variables affected by
residential development in the hypothetical 4-mi2 area, were presented visually to groups of
about 30 respondents, who completed the survey questionnaire during several pauses in the
presentation. In the first part of each session, respondents were introduced to the potential
impact of development under each of the scenarios. Photographs taken within the Darby
watershed were used to illustrate these effects.
In the latter part of each session, respondents were asked to identify which of the four
scenarios they felt were most likely to occur and which they most preferred. This was followed
by a WTP question used in the CVM analysis. A map showing a portion of the Big Darby Creek
watershed was displayed, with a 150-mi area just west of Columbus highlighted as "facing
likely development over the next 20 years." The Darby watershed sample was drawn from this
area. Each respondent was then confronted with a choice between the high density base case and
one of the other development scenarios. This question was framed around the idea that a group
of citizens, along with government officials at both the local and state levels, had developed a
fund to ensure that development in the highlighted area of the Darby follows a path that would
lead to a specified state. It is proposed that monies for the fund would come from a hypothetical
check-off on Ohio State Income Tax forms similar to current donation opportunities for wildlife
and for natural areas. The respondents were asked if such a check-off were available, asking
them to contribute $_ to the fund, would they check YES or NO? The dollar amounts were
filled in by a random allocation within the questionnaire of amounts ranging from $1 to $100,
based upon results from focus group pretests.
4-30
-------
A method suggested by Loomis and colleagues was used to calculate mean WTP based
on survey results. For a particular landscape scenario, a core logit equation was formulated as
follows:
VOTE = f (FUND, INC, USEFREQ, AGE, Z), 4-1
where VOTE is a dummy variable indicating whether the respondent voted YES or NO on the
WTP question (preferring an alternative to the high density outcome), FUND is the respondent's
posed dollar value contribution, INC is household income, USEFREQ is the number of times per
year the respondent or family uses the Big Darby for outdoor activities, AGE is the age of the
respondent, and Z is a variable indicating special circumstances that might influence WTP. For
example, one question asked whether the respondent or a family member considered themselves
to be a farmer; another asked if the respondent was a member of an environmental group.8
Alternate specifications of the model were estimated using different respondent variables as the
basis for the core equation (i.e., Z was a dummy, YES/NO, variable either for "farmer" or for
"environmental group member"), then separately considering status of the respondent (Resident,
Near-Resident, Non-Resident), and finally by scenario type (Ranchette, Cluster, Agriculture).
The results can then be interpreted as the contribution of each of the variables towards an
individual's probability of contributing to the rand.b
Mean values for all the variables are used in conjunction with the estimated regression
coefficients from the logit regression to estimate a mean WTP. The resulting general estimates
from two alternate model specifications are shown in Table 4-2. The upper and lower bounds for
a In the sample of 766 respondents, 83 stated they were members of an environmental group, 66 said they were
farmers, 8 were both, and 625 were neither.
Details of these results are available upon request (Loucks, Erekson, Elliott, McCollum, and Bruins, submitted to
Landscape Ecology, 2002),
4-31
-------
TABLE 4-2
Mean willingness to pay and confidence intervals for two model specifications8
Sample
Entire
Resident
Near-Resident ,
Non-Resident
Ranchette
Cluster
Agriculture
Specification lb
Mean
WTP
$37.65
$49.82
$33.91
$25.99
$25.62
$67.05
$29.58
90%
C.I.
Min
. $28.64
$29.29
$23.37
$14.99
$17.15
$30.89
$20.86
90%
C.I.
Max
$58.18
u$ 156.09
$68.28
$80.57
$58.91
$261.17
$57.72
Specification 2°
Mean
WTP
$37.96
$51.44
$33.38
$25.45
$25.19
$69.73
$29.24
90%
C.I.
Min
$28.72
$29.15
$23.40
$15.06
$17.02
$27.33
$20.50
90% C.I.
Max
$58.94
$162.39
$67.19
$75.11
$54.47
$291.69
$57.09
aResidents, n = 322; Near-Residents, n = 319; Non-Residents, n 106
''Model specification includes dummy variable for "farmer"
°Model specification includes dummy variable for "environmental group member"
4-32
-------
the 90% confidence intervals are estimated using a simulation model with 10,000 random draws
of the estimated regression coefficients. As would be expected, the mean values are higher for
residents than for near-residents, and those are higher than for non-residents. In addition, the
WTP for a Cluster landscape alternative was significantly higher than for the Agriculture or
Ranchette alternatives.
4.3.6 Linking stream integrity to the development scenarios
The approach to linking stream ecological condition with the development scenarios
relied to some degree on aii empirical relationship between impervious surface area (a runoff
inducing condition) and IBI. Recent work by Yoder et al.19 showed that for the lowest quartile
of urbanization around Ohio stream sampling sites (with impervious surface of less than 4.3% of
watershed area), modal IBI is 42. This is just above the Warm Water Habitat criterion of IBI =
40 (and well below the Exceptional Warm Water Habitat criterion of 50)." For the second
quartile of urbanization (4.3 to 14,6% impervious), the modal IBI is 39.5. For the third quartile
(14.7 to 29.3% impervious), the IBI is 35.0, while for the fourth quartile (over 29,3%
impervious), the estimated mid-range IBI is 24, or highly degraded. This work suggests a likely
median of 3 percent impervious surfaces for rural agricultural land, and 20 percent or more for
urban areas, both reflecting a literal understanding of the term impervious surface: the total
surface area of roads, driveways, and roofs. These results also suggest a possible threshold for
serious degradation of IBI when impervious surfaces are at or above 20 percent. In addition, the
a Under OEPA's designation, exceptional warm water habitat differs from warm water habitat in having an
exceptional or unusual community of species when compared to reference sites (i.e., comparable to the 75*
percentile of reference sites on a statewide basis). More stringent biological criteria are established for exceptional
waters (see Section 2.3.1),
4-33
-------
majority of watersheds having more than 15% impervious surface do not meet the OEPA's
Warm Water Habitat Biocriteria.19
However, runoff hydrologists20"21 have over many years developed an empirical
relationship between modified surface conditions (such as cultivation, or residential lawn
surfaces) and the intensity of runoff induced. These papers show that intensive cultivation
creates runoff-inducing conditions in agricultural areas roughly equivalent to a moderate level of
impervious surfaces. Using a transformation based on the "curve numbers" adopted by the
hydrologists, a measure, "runoff-inducing condition," has been developed as shown in Table 4-3
that captures the conditions (and IB I) associated with each of the development scenarios.
TABLE 4-3
Runoff-inducing condition and IBI per scenario
Scenarios
Agriculture
Ranchette
Cluster
High
Density
Indicated Impervious Surface
Assumptions (after Yoder et al.)
3%
-
-
20%
Interpolated Runoff-Inducing
Conditions
16.9
16.3
17.0
21.3
Modal
IBP
42
43.0
41.8
35
"Interpolated from a graph linking the results of Yoder et al.; interpolated runoff-inducing condition, and
IBI. Details available from the authors.
4.3.7 Linking stream integrity and willingness to pay
There is a great deal of interest among environmental managers in determining the dollar
values that may be associated with changes in ecological condition. When respondents
expressed WTP to obtain one of the development scenarios, their valuation took into account the
4-34
-------
economic, quality-of-life and ecological ramifications of adopting that scenario in place of the
expected, high-density scenario, In this case study, those ecological changes were quantified as
units of IBI change. A multimetric index such as IBI has the potential to respond in complex
fashion to changes in water or habitat quality. The large number of metrics it includes, however,
and the functional complementarity among those metrics, apparently lend it a degree of
numerical stability. In practice, IBI often has been treated as having cardinal properties for
purposes of environmental analysis and regulation. In this section, the investigators probe the
implications of their data for associating a dollar value with a unit of change in IBI.
Table 4-4 provides preliminary estimates of the relationship between WTP and IBI
change in the 150-mi2 area considered in the survey. For example, in the case of respondents
considering the agriculture alternative, the change in runoff-inducing condition from high density
to agriculture (from 21.3 to 16.9) corresponds to an IBI improvement of from 35 to 42.
Respondents for the agriculture cohort had a mean WTP of $29.58, corresponding to the 7-point
improvement in IBI. Thus, an estimate of the WTP per unit of IBI for this cohort would be $4.23
per unit of IBI. The corresponding estimates ($9.86 per unit of IBI) for the cluster cohort were
more than double that of the agriculture cohort, and almost triple that for the ranchette cohort.
For many reasons, however, caution is necessary in interpreting these IBI-normalized
WTP values, since these results do not separate changes in ecological and related risks from
other environmental, economic and social changes associated with the development scenarios.
In fact, since the IBI changes associated with these three scenarios were similar in magnitude, it
is likely that the expressed differences in value between the scenarios were influenced by both
non-ecological factors and certain perceptions about ecological factors not captured by IBI.
4-35
-------
TABLE 4-4
Estimated WTP per unit of D3I improvement over a 150-mi2 study area for two model
specifications
Ranchette
Cluster
Agriculture
IBI
Improvement
8
6.8
7
Specification 1
Mean
WTP
$25.62
$67.05
$29.58
Mean
WTP/IBI
$3.20
$9.86
$4.23
Specification 2
Mean
WTP
$25.19
$69.73
$29.24
Mean
WTP/IBI
$3.15
$10.25
$4.18
4-36
-------
Analyses now underway are looking more closely at the respective, marginal contributions of the
ecological, economic, and social factors to WTP.
4.4 DISCUSSION
When the Big Darby Creek watershed ERA (Section 4.2) and economic analysis (Section
4.3) are considered collectively, the overall work has some of the ideal characteristics of an
integrated analysis as described in Section 3.3 and diagrammed in Figure 3-1. It also
demonstrates some of the problems that result when integration is not a goal from the outset.
Assessment planning involved a wide variety of partners and stakeholder groups,
resulting in clearly defined goals and objectives (see Section 4.2.1). The problem formulation
conducted as part of the ERA identified two ecological assessment endpoints, of which one
could be feasibly measured, and conceptual models developed relating human activities in the
watershed to stressors, to effects on endpoints, and to specific measures of effect. An analysis
plan for the evaluation of specific risk hypotheses was developed and substantial progress was
made toward the analysis and characterization of baseline risk. The ERA made use of data
collected as part of the statewide watershed management cycle (see Figure 3-A-6) and had begun
to provide empirical, stressor-response and source-response relationships that will be useful in
TMDL development.
The team conducting the economic analysis formulated a set of management alternatives,
in this case suburban development scenarios, focused on one of the more severe concerns
identified in assessment planning and problem formulation: stream degradation linked to urban
encroachment in the watershed's eastern portion. The subsequent steps, analysis and
4-37
-------
characterization of alternatives and comparison of alternatives, were similar in form to the
example shown in Figure 3-3 but with a number of important differences. As shown in Figure
4-3, they provided a qualitative analysis of the effects of each scenario on a set of important
stressors affecting instream biota and on economic and social services to watershed residents.
They did not examine the financial costs or other market-based effects of the management
alternatives. In that those costs would accrue to land holders who would have to forego valuable
development options, the analysis also did not address equity.
To compare the alternatives, the ecological, economic and social impacts of each scenario
were incorporated into an integrated CVM instrument. The comparison was effected using
monetary WTP associated with each scenario. That is, the economic analysis examined current
WTP to avoid development changes that were expected to take place at some time during the
next 20 years. Respondents were presented with a set of development alternatives and the
expected ecological, economic and social changes that would result from each.
The expected time frame for these effects was not made explicit, making interpretation of
the analysis difficult. The time horizon is important both for understanding the respondents'
preferences and for comparing the value of current effects to that of future effects (i.e.,
discounting the stream of future costs and benefits, see Section 2.2.3).22 Supposing, for example,
that respondents assumed most of the expected, high-density development would not occur for
10-15 years in any case - and thus that any benefits of funding an alternative would be similarly
delayed - they would have discounted their current WTP accordingly. If development actually is
likely to occur sooner than they assumed, WTP values measured in this study would be too
small. Similarly, if they assumed that the ecological effects of high density development would
4-38
-------
-I-
ANALYSIS & CHARACTERIZATION OF ALTERNATIVES
EeolQgical Risk Economics Other Analyses
Quantify endpoint
changes where
feasible
FyTJi
' : >l is
i%i".(j- ' "
Qualitatively
describe
other changes
Qualitatively analyze
equity, economic impact
Express primary changes
in common language
Qualitatively
describe
other changes
Express equity effects, impacts
in common language
.1
Attempt to compare
WTP and IBI change
Stated preference
study
Express primary changes
in common language
EsfllT' I
efc
COMPARISON OF ALTERNATIVE
FIGURE 4-3
Techniques used for analysis, characterization and comparison of management alternatives
in the Big Darby Creek watershed, as compared to the example shown in
Figure 3-3. White boxes and bold type show features included in this analysis.
4-39
-------
occur only much later than the other (economic and social) effects, and if this assumption was
incorrect, then the ecological benefits of the other scenarios would not matter as much as the
other changes and WTP for the more ecologically beneficial scenarios would be negatively
biased.
In a subsequent step, WTP was compared to estimated IBI change. This latter step was
of limited success, for reasons just discussed in the previous section, but with further analysis it
could provide information that is useful in other settings. In general, this integrated assessment
process provided decision support only (see Table 3-1); it did not include decisions or
subsequent implementation.
In future studies of this type, if estimates of WTP for a given IBI change are sought, a
more effective approach might be to elicit preferences for different fish community
characteristics and preferences for different housing densities using separate CVM questions
(within one survey) or representing these as separate attributes in a conjoint analysis study (see
Appendix 2-A). The next step would be to use these data, along with information on the effects
of the development scenarios on fish communities and the financial and market-based economic
effects of the scenarios, to assess the net social benefits of the scenarios. Such an approach
would be less reliant on establishing accurate respondent understanding of the ecological impacts
of housing scenario, and it would also allow adjustment for new knowledge about that
relationship without repeating the survey. It would also yield a more inclusive indicator (i.e., net
social benefit) than WTP alone.
4-40
-------
Nonetheless, neither WTP nor net social benefit estimates are necessarily the best
endpoint for housing-related decisions in the Darby watershed. In spite of the thoroughness of
the biophysical and socioeconomic framing of this CVM study, reviewers of this study at a
USEPA workshop held in July 2001 were pessimistic about its likely influence on development
decisions in the study area. They cited the substantial private gains to be made by developing
individual tracts to the maximum allowable number of housing units, the spatial fragmentation of
zoning authorities, and the tendency of zoning boards to respond to the wishes of property
owners and developers. In other words, in specific zoning or development decisions there is not
an effective mechanism for internalizing the negative externalities of high density development
manifested in statewide WTP. There was skepticism that the simple provision of WTP
information would make an impact. Although there is some Clean Water Act authority for
reducing the water-quality impacts of home construction, road construction and imperviousness,
it does not otherwise interfere with local land development.
Although the assessment planning effort that was carried out originally as part of the Big
Darby Creek ERA examined a broad suite of watershed problems, the reviewers' observations
suggested that this analysis did not adequately characterize the decision context (see Section
2.1.1.1) specific to suburban development. To better determine the applicability of WTP
measured in this study to development decisions in the Big Darby Creek watershed, the
assessment planning process would need to be revisited. Participants in a renewed process
should include members of zoning boards, farm owners, developers, and individuals representing
the residents', near-residents' and statewide interests in retaining the ecological, economic and
4-41
-------
social amenities of the area. They should also include OEPA officials responsible for addressing
local stream reach impairments. Interactions could involve the provision of information about
these amenities and the impacts of development, discussion of shared values and an attempt to
develop consensus goals for this portion of the watershed. Techniques used might include the
joint development of future scenarios for the area.23'24 Further analyses should include
development of TMDLs and implementation plans that consider alternative residential (or
industrial) development scenarios. Significantly, these plans should include efforts to develop
compensation mechanisms whereby those who partially or completely forego development
options are compensated, as is done under "transferable development rights" initiatives.
4.5 -REFERENCES
1. Erekson, O.H., Loucks, O.L., and Strafford, N.C., The context of sustainability, in
Sustainabitity Perspectives for Resources and Business, Loucks, O. L., Erekson, O. H.,
Bol, J. W., Gorman, R. F., Johnson, P. C., and Krehbiel, T. C. Eds., Lewis Publishers,
Boca Raton, 1999.
2. Zwinger, A., Darby Creek, Ohio: back home again, in Heart of the Land: Essays on Last
Great Places, Barbato, J. and Weineman, L. Eds., Pantheon Books, New York, 1994,
151.
3. Cormier, S.M. et al., Assessing ecological risk in watersheds: a case study of problem
formulation in the Big Darby Creek watershed, Ohio, USA., Environmental Toxicology
and Chemistry, 19,1082, 2000.
4-42
-------
4. Schubauer-Berigan, M.K. et al., Using historical biological data to evaluate status and
trends in the Big Darby Creek watershed (Ohio, USA), Environmental Toxicology and
Chemistry, 19, 1097,2000.
5. USFWS, Little Darby Creek Conservation Through Local Initiatives: A Final Report
Concluding the Proposal to Establish a National Wildlife Refuge on the Little Darby
Creek in Madison and Union Counties, Ohio, U.S. Fish & Wildlife Service, Ft. Snelling,
Minnesota, 2002.
6. USEP A, Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F, Risk
Assessment Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
7. USEP A, Biological Criteria: Technical Guidance for Streams and Small Rivers. Revised
Edition, EPA 822-B-096-001, U.S. Environmental Protection Agency, Office of Water,
Washington, DC, 1996.
8. Norton, S.B. et al., Can biological assessments discriminate among types of stress? a
case study from the Eastern Corn Belt Plains ecoregion, Environmental Toxicology and
Chemistry, 19,1113,2000.
9. Gordon, S.I. and Majumder, S., Empirical stressor-response relationships for prospective
risk analysis, Environmental Toxicology and Chemistry, 19, 1106, 2000.
10. Gordon, S.I., Arya, S., and Dufour, K., Creating a Screening Tool for Identification of the
Ecological Risks of Human Activity on Watershed Quality, Report to the U.S. EPA on
Cooperative Agreement # CR826816-01-0, City and Regional Planning Program, School
of Architecture, Ohio State University, Columbus, Ohio, 2001.
4-43
-------
11. Hume, H.G., Sustaining Biological Diversity and Agriculture in the Big Darby Creek
Watershed, Institute of Environmental Sciences, Miami University, 1995.
12. Zucker, L.A. and White, D.A., Spatial Modeling of Aquatic Biocriteria Relative to
Riparian and Upland Characteristics., Alexandria, VA, June 8-12, 571.
13. Arrow, K.J. et al., Report of the National Oceanic and Atmospheric Administration Panel
on Contingent Valuation, 58, Jan. 15, 1993, 4602.
14. Dillman, D.A., Mail and Internet Surveys: The Tailored Design Method, John Wiley and
Sons, New York, 2000.
15. Karr, J.R. and Chu, E.W., Restoring Life in Running Waters: Better Biological
Monitoring, Island Press, Washington, D.C., 1999.
16. Elliott, S.R. et al., Reliability of the Contingent Valuation Method, U.S. EPA Cooperative
Agreement CR-812054, University of Colorado, Boulder, 1989.
17. Knetsch, J.L., Environmental policy implications of disparities between willingness to
pay and compensation demanded measures of value, Journal of Environmental
Economics and Management, 18, 227, 1990.
18. Loomis, J. et al., Measuring the total economic value of restoring ecosystem services in
an impaired river basin: results from a contingent valuation survey, Ecological
Economics, 33, 103, 2000.
4-44
-------
19. Yoder, C.O., Miltner, R.J., and White, D., Using biological criteria to assess and classify
urban streams and develop improved landscape indicators, in National Conference on
Tools for Urban Water Resource Management and Protection, EPA/62 5/R-00/001,
Minameyer, S,, Dye, J., and Wilson, S. Eds., U.S. Environmental Protection Agency,
Office of Research and Development, Cincinnati, OH, 2000, 32.
20. Soil Conservation Service, Urban Hydrology for Small Watersheds, Technical Release
No 55, United States Department of Agriculture, Engineering Division, Washington,
D.C., 1975.
21. Soil Conservation Service, Ohio Supplement to Urban Hydrology for Small Watersheds:
Technical Release No 55., United States Department of Agriculture, Columbus, Ohio,
1981.
22. USEPA, A Framework for the Economic Assessment of Ecological Benefits, Science
Policy Council, U.S. Environmental Protection Agency, Washington, DC, Feb. 1, 2002.
23. Hulse, D. et al, Planning alternative future landscapes in Oregon: evaluating effects on
water quality and biodiversity, Landscape Journal, 19, 1, 2000.
24. Coiner, C., Wu, J., and Polasky, S., Economic and environmental implications of
alternative landscape designs in the Walnut Creek Watershed of Iowa, Ecological
Economics, 38,119, 2001.
4-45
-------
-------
5. VALUING BIODIVERSITY IN A RURAL VALLEY:
CLINCH AND POWELL RIVER WATERSHED
5.1 WATERSHED DESCRIPTION
The Clinch and Powell Rivers originate in mountainous terrain of southwestern Virginia
and extend into northeastern Tennessee, flowing into the upper reaches of the Tennessee River
(Figure 5-1). The Powell River originally was a tributary of the Clinch River, but both now flow
into the upper reach of Norris Lake. The Clinch and Powell River watershed above Morris Lake,
also referred to here as the upper Clinch Valley, covers 9,971 km2 and ranges between 300 and
750 meters in elevation. Historically, it contained one of the most diverse fish and mussel
assemblages in North America,1 yet most of these populations have declined dramatically or
been eliminated.2 The mainstem Tennessee River and many of its tributaries have been dammed,
resulting in the loss of habitat for many fish and mussel species, and therefore the upper Clinch
and Powell Rivers represent some of the last free-flowing sections of the expansive Tennessee
River system. Currently, the Clinch Valley supports more threatened and endangered aquatic
species than almost any other basin in North America.3 Despite implementing recovery plans for
most federally protected species in this basin, there is evidence that these species are either
declining or becoming extinct at an alarming rate due to impacts from mining, agriculture,
urbanization and other stressors.4
The Clinch Valley is a traditional rural Appalachian region. The areas are among the
poorest in their respective states, with coal mining, agriculture and scattered manufacturing the
primary industries. Although the area is very scenic, with a few exceptions tourism is poorly
5-1
-------
NC
120 Kilometers
A
FIGURE 5-1
The Clinch and Powell River watershed in the eastern USA. The study area is the portion of the
watershed that is above Norris Lake. Initial ecological study focused on Copper Creek. Towns
where discussions were held are shown, as are urbanized areas.
5-2
-------
developed. The regional coal and tobacco industries are in decline, and the "high tech" economy
has not found its way south of Blacksburg (Virginia Polytechnic Institute and State University)
or east of Knoxville (University of Tennessee/Oak Ridge National Laboratory). Many former
miners suffer from Black Lung Disease and other problems. School districts often have trouble
offering curricula that are comparable to the suburban school districts and finding qualified
teachers. Children often leave the region upon completion of their university education.
Transportation problems contribute to the area's economic isolation. Interstates 1-81 and
1-40 run parallel to the Clinch River, only one or two ridges east, and a quick glance at a map
might indicate that transportation is not a problem; however, getting from the Clinch Valley
communities to the interstate highways can be quite time consuming, often requiring more than
an hour's travel on rural roads. An additional one to two hours is required to reach the
Blacksburg/Roanoke area or the Knoxville area. Given the topography of the region, improving
the transportation system can conflict with protecting the Clinch River and its tributaries, as the
only place for roads is in the flood plains of the streams.
The people of the region do appreciate its environmental resources and are very active in
activities such as hunting, fishing and hiking. Evidence of this perspective was found in an
unpublished survey. Preliminary to ecological study of the watershed, local environmental
organizations surveyed several communities in the region in 1994 to determine their attitudes
and values. The results indicated strong interest in protecting local natural resources, but not at
the expense of building roads, attracting industry or creating new jobs.
A large amount of ecological information has been collected in this watershed over many
years, but much of it had not been analyzed prior to this work. Entities collecting environmental
data included The Nature Conservancy (TNC), Tennessee Valley Authority (TVA), U.S. Fish &
5-3
-------
Wildlife Service (USFWS), U.S. Geological Survey (USGS), Virginia Department of Game and
Inland Fisheries, and Virginia Department of Conservation and Recreation. Resource managers
suspected that mining, urbanization and agricultural activities were adversely impacting the
exceptional fish and mussel diversity. While several hypotheses have been advanced to explain
these species' decline in other watersheds,5 definitive answers as to their decline in this
watershed (Figure 5-2) have been lacking. Resource managers recognized that a comprehensive
examination of the available data was needed to evaluate the relative effects of different human
activities. Given the socioeconomic context of the Clinch Valley, it is also important to
investigate the ways the people of the region compare environmental protection with economic
development.
The following sections of this chapter describe studies carried out in the Clinch Valley by
the U.S. Environmental Protection Agency (USEPA) and its partners to improve management of
the areas unique ecological resources. Section 5.2 describes a watershed ecological risk
assessment (W-ERA), initiated in 1993 and carried out by an interagency workgroup. In 1999,
USEPA awarded a grant to the University of Tennessee for an economic study that would use the
results of the W-ERA and address decision-making needs; this study is described in Section 5.3.
Section 5.4 then examines the overall work in the light of a conceptual approach for ERA-
economic integration in watersheds (described in Chapter 3).
5.2 ECOLOGICAL RISK ASSESSMENT
5.2.1 Planning
The Clinch Valley ecological risk assessment7'6'8 was one of five prototype,
watershed ecological risk assessments (W-ERA) sponsored by the USEPA to further
5-4
-------
Historic mussel
distribution
KENTUCKY
Present mussel
distribution
Towns
Rivers
Counties
FIGURE 5-2
Comparison between historic (pre-1910) and present locations of native mussel concentrations in
the Clinch/Powell watershed; red areas represent mussel beds, (from Diamond et al.6)
5-5
-------
develop, demonstrate and test the use of the ecological risk assessment paradigm9 at the
watershed scale, (The reader is referred to Section 2.1 for more explanation of the procedures
and terminology of ERA). Like the other watersheds selected, the Clinch Valley was a candidate
for W-ERA because it contains valued and threatened ecological resources, has been the subject
of data collection efforts, is subject to multiple physical, chemical and biological stressors and
receives attention from several organizations working to protect its resources. Federal, state and
local managers had been working with scientists from Virginia and Tennessee to study the
distribution of aquatic resources in the Clinch Valley. The global significance of the faunal
(especially molluscan) diversity had drawn a great number of scientists to the area.
For this risk assessment, an interdisciplinary, interagency workgroup was established
in 1993 with representatives from USFWS, TV A, TNC, Virginia Department of Game and
Inland Fisheries, Virginia Cave Board, USEPA and USGS. Unlike in the other W-ERAs, a
broader stakeholder group was not convened. Information on attitudes and values from the
community survey mentioned in Section 5.1 was taken in lieu of direct stakeholder involvement.
Among six environmental concerns presented in that survey, "preserving our rare plant and
animal species" was rated lowest in importance, whereas "our water quality" was rated highest.
This information stood in some contrast to the urgency for biodiversity protection felt by
members of the interagency workgroup.
To focus the scientific information that would be analyzed in the Clinch/Powell
watershed, the workgroup identified outstanding ecological resources, developed a management
goal and identified a set of management objectives considered important to achieving the
management goal (Table 5-1). The workgroup agreed to focus the assessment on the
5-6
-------
TABLE 5-1
Outstanding ecological resources, environmental management goal and management objectives
for the Clinch Valley ecological risk" assessment
Outstanding ecological resources:
« The diversity and biological integrity of aquatic macroinvertebrates, especially the
unique native freshwater mussels
The diversity and abundance of the native fishcommunity
Environmental management goal and subgoals:
Establish and maintain the biological integrity of the Clinch/Powell watershed surface and
subsurface aquatic ecosystem.
Establish self-sustaining native populations of macroinvertebrates and fish
» Improve water quality in the rivers
* Establish and maintain functional riparian corridors of native vegetation
Safeguard water quality in a sustainable sub-surface ecosystem
Management objectives:
Create and maintain vegetated riparian zones in agricultural areas to intercept sediment,,
nutrient, and pesticide runoff; enhance fish habitat; reduce thermal stress in smaller
headwater streams; and exclude cattle from stream beds
* Create and maintain vegetated riparian zones in urban, industrial, and developed areas to
diminish sedimentation from storm water runoff and reduce instream habitat alteration
* Implement agricultural best management practices (BMPs) such as rotational grazing to
reduce sedimentation, pathogens, and nutrient enrichment instream
* Contain and treat runoff from mining activities to reduce pollutant load and
sedimentation instream
Install or improve sewage treatment facilities in streamside rural and urban communities
to reduce inputs of toxic pollutants, pathogens, and nutrients instream
Adequately treat industrial discharges to reduce input of toxic pollutants instream
Create and maintain storm water retardation and holding facilities for highways and
developed areas to reduce sedimentation runoff instream
From Diamond et al. and USEPA
10
5-7
-------
unimpounded stream segment above Norris Lake, since only that portion of the watershed
provided suitable habitat for the fish and mussel species of concern. The assessment would use
its limited funds to analyze data collected previously. Terrestrial and aquatic communities in
caves associated with karst, though unique and diverse in the watershed, were not examined in
this risk assessment because of insufficient information. The workgroup also recognized that
there were other possible sources of stress in the watershed, including competition from exotic
species (e.g., the asiatic clam Corbicula ftumined) and atmospheric deposition of contaminants.
They opted not to consider these sources in this assessment because their impacts are relatively
minor and they cannot be addressed by local managers.
5.2.2 Problem formulation
During problem formulation, the broad management goal of establishing and maintaining
biological integrity was more explicitly defined. Human-caused sources and stressors in the
watershed were listed (Table 5-2) and considered in detail.6 Assessment endpoints
corresponding to the outstanding biological resources were selected, and conceptual models were
drawn illustrating the pathways by which the endpoints may experience adverse effects. The two
endpoints selected in this assessment were; (1) reproduction and recruitment of threatened,
endangered or rare native freshwater mussels; and (2) reproduction and recruitment of native,
threatened, endangered or rare fish species.
Conceptual models developed by the workgroup traced the most important, hypothesized
pathways between sources, stressors, and direct and indirect ecological effects. For example, the
model for effects on mussels (Figure 5-3) shows agriculture, mining, silviculture and urban areas
to be sources of excess sediment. The resulting turbidity affects mussel survival and recruitment
-------
TABLE 5-2
Stressors and sources identified in the Clinch and Powell watershed
Stressor
Sources
Degraded Water Quality
Toxic chemicals
Pathogens
Nutrients
Catastrophic spills
Urbanization
Point-source discharges
Atmospheric deposition
Urbanization
Urbanization
Atmospheric deposition
Agriculture
Coal mining
Transportation
Agriculture
Agriculture
Physical Habitat Alteration
Sedimentation
Riparian modification
Instream destruction
Coal mining
Hydrologic changes
Transportation
Agriculture
Hydrologic changes
Agriculture
Hydrologic changes
Agriculture
Urbanization
Urbanization
Urbanization
Biotic Interactions
Exotic species introductions
Overexploitation
Accidental (Asiatic clam, zebra mussel)
Recreational (brown trout, rainbow trout)
Other biota
Over harvesting
Poaching
From Diamond et al.
5-9
-------
Conceptual Risk Model for Mussels
W
0)
0)
c
ro
£
Q
"ra
I
*>
c
HI
Instream Habitat ^\ ^ Sedimentation ")
DeKtnintion x -^^ ~^
DectBased
Allcxtithonous
Inputs to Stream
and Increased Heat
and Light
Reproduction, Recruitment of Threatened,
Endangered, or Rare Mussel Species
FIGURE 5-3
Simplified conceptual model showing major pathways between sources (land use), stressors, and
effects on the assessment endpoint for native mussel species abundance and distribution and data
sources available (adapted from Diamond et al,8).
5-10
-------
by interfering with filter feeding, and siltation smothers the substrates to which they attach.
Siltation also smothers benthic (bottom-dwelling) macroinvertebrates, the food source of
insectivorous fish, thereby reducing the availability of host species for the mussels' parasitic
larval stage, or glochidia, which must attach onto the fins, epidermis or gills of a suitable host
fish. A similar model (not shown) traced the pathways for risks to fish species.
Risk hypotheses to be evaluated in the analysis phase were developed for each endpoint,
and eventually consolidated to three, corresponding to two categories of stressors:
Physical Habitat Alteration Hypotheses
» Greater connectivity of riparian (i.e., stream-side) vegetation, or forested riparian
vegetation, is associated with greater diversity and abundance of mussels, other
macroinvertebrates, and native fish.
* Watershed areas dominated by agricultural, urban, or mining land uses are associated
with poorer physical habitat quality and biological diversity than are forested or naturally
vegetated areas.
Water Quality Hypothesis
* Proximity to nonpoint-source runoff (from agricultural activities and urban areas) and
point-source discharges (including coal mining discharges) results in detrimental
structural changes to native mussel and fish populations.
Available data sets for sub watersheds of the Clinch Valley were examined and an
analysis plan was developed. Because of data limitations, it was decided to undertake a
preliminary analysis in a subwatershed, Copper Creek (Figure 5-1), to determine the appropriate
spatial scale for analysis of riparian vegetation and land uses, and to identify appropriate
5-11
-------
biological measures as surrogates for the assessment endpoints. It was also decided that TVA
would organize the available information in a geographic information system (GIS).
5.2.3 Risk analysis
5.2.3.1 Methods
Analyses were based on data collected at many locations in the watershed over several
years. Monitoring programs that provided key data for this risk assessment included TVA's
Clinch-Powell River Action Team Survey and the Cumberlandian Mollusc Conservation
Program. Land cover data used in this risk assessment were derived from LANDS AT Thematic
Mapper imagery, classified into 17 discrete categories including several different forest types,
urban and developed land, pasture and cropland. All terrain data (e.g. elevation and slope) were
derived from a mosaic of USGS digital elevation models (DEM) at 30-m resolution. USEPA's
River Reach File 3 provided stream network data. Locational data were also available for mines,
coal preparation plants, major transportation corridors, urban centers, and biological sites in the
basin. Several measures of instream habitat quality, including bottom substrate characteristics,
bank stability, riparian vegetation integrity, channel morphology and instream cover, were used
to characterize habitat condition. A multimetric habitat quality index (similar to QHEI; see
Appendix 2-B) was also used. However, water quality data were insufficient to allow
determinations either of land-use effects on water quality or water-quality effects on the
assessment endpoints. Therefore, it was necessary to directly examine the relationships between
land uses, instream habitat quality and the assessment endpoints, without reference to water
quality per se.
Since data directly matching the assessment endpoints were not available, surrogate
measures were used. For example, few data were available on native threatened, endangered or
5-12
-------
rare fish species. However, the Index of Biotic Integrity (IBI), a multimetric index describing
the status of the fish community, had been determined by TVA at a number of locations
throughout the watershed and was considered to be a reasonable measure for the second
assessment endpoint (for more information on the IBI see Appendix 2-B). Data on mussel
species richness and abundances were also limited, but preliminary study in the Copper Creek
subwatershed showed a reasonable correlation between IBI score and mussel species richness,
and therefore IBI values were used to supplement the mussel species data. For benthic
macro invertebrates, the EPT index, consisting of the number of taxonomic families present from
the orders Ephemoptera (mayfly), Plecoptera (stonefly) and Tricoptera (caddisfly), had been
determined in some locations. These orders are known to be sensitive to adverse water quality
and are replaced by other macroinvertebrates as water quality diminishes.
Forward stepwise multiple regression analyses and/or univariate statistical analyses of
data within a GIS were used to test stressor-response associations. GIS maps were produced that
examined each risk hypothesis. In many cases it was necessary to reduce the underlying
variability by truncating the elevation range of sites included in order to detect source-response
or stress-response relationships.
5,2.3.2 Copper Creek pilot study
Copper Creek was chosen for pilot analysis because it was a comparatively data-rich
subwatershed, and it presented a simpler case in that agricultural uses were the major sources of
stressors. Findings, which were used to structure the analysis of the entire Clinch Valley
watershed, included the following:
« Agricultural uses in the riparian zone had more of an influence on instream habitat
quality and fish community integrity (IBI) than did upland agricultural land use
5-13
-------
* Effects of human activity in the riparian zone could be observed in native fish and
mussels as much as 1500 m downstream of the activity (e.g., Figure 5-4)
» IBI score was correlated with mussel species richness
Land use in the riparian corridor had a stronger effect on IBI than did an overall index of
habitat quality, although particular habitat parameters - such as instream cover score, and
degree to which stream substrates were free from embedding fine sediments
(clean substrate score8) did correlate well to IBI and EPT
* After analyzing riparian corridor data at widths of 50, 100, 200 and 500 m and at varying
lengths, a riparian corridor zone measuring 200 m across (100 m to either side of the
stream) and extending 500 to 1500 m upstream was found to be the appropriate spatial
area in which to analyze land-use effects on fish and mussels,
5.2.3.3 Clinch Valley
The most successful analytical approaches in the Copper Creek pilot study, noted above,
were applied to the entire Clinch Valley watershed. Because other parts of the watershed are
subjected to stressors from the coal industry and urbanization, the riparian land cover
analyses were expanded to include land uses other than agriculture. Land use analyses included
the following:
« Proximity to different types of mining activities
Proximity to urban/industrial areas
* The percentage of land use in the area that was forested, pasture, cropland, or urban
a TVA defines this parameter as "substrate embeddedness," To make the directionality of the score (1 = poorest, 4=
best) more intuitive, it is here renamed "clean substrate score.
5-14
-------
0
10 22 40 78 94 100
% Agricultural Use
FIGURE 5-4
Fish community integrity as a function of agricultural land in a riparian corridor of 200 m width
and 1500 m length in Copper Creek (from Diamond et al.8)
5-15
-------
» Proximity to three classes of roads, including major U.S. highways, State roads, and
county roads.
5.2.3.3.1 Effects of land use on habitat quality
Some effects of riparian-corridor land use upon instream habitat quality could be
discerned when variability was reduced by limiting sites analyzed to those occurring between
350 and 450 m elevation. Forty-two percent of among-site variability in the habitat quality index
(N = 85) could be explained by riparian land use. Stream sedimentation was lower where
cropland was 3% of total land use. Riparian integrity was better in areas in which pasture or
herbaceous land was < 50% of the total land use. Instream cover was poor if urban use was 20%
of the surrounding area upstream. Instream cover and clean substrate scores were affected by
both percent pasture/herbaceous cover and percent urban area nearby. The relationships between
land use and habitat quality suggest that instream habitat will have the highest probability of
being satisfactory for aquatic life if agricultural land use is relatively low and urban influences
are small.
5.2.3.3.2 Relationships between land use and biological measures of
effect
Among sites of 350 - 450 m elevation, riparian land uses explained 55% of variability in
IBI scores (N = 38) and 29% in EPT scores (N = 34). Percent pasture area was positively related
to IBI while proximities to mining, crops and urban areas were negatively related. The
apparently positive effect of pasture land on IBI was unexpected based on the pilot results for
Copper Creek and the negative relationship between pasture area and riparian integrity observed
at these sties. A likely explanation is that IBI may respond positively to moderate nutrient
enrichment and that negative effects of mining and urban development are comparatively much
5-16
-------
worse. The number of native mussel species was inversely related to several land uses including
(in order of significance): percent urban area; proximity to mining; and percent cropland. In the
multiple regression model these factors accounted for 26% of the observed variation in mussel
species richness. Collectively, the analyses demonstrated that mining and urban areas are more
detrimental than pasture areas to aquatic fauna in this watershed.
5.2.3.3.3 Relationships between habitat quality and biological measures
of effect
In stepwise regression analyses of sites 350 - 500 m in elevation, habitat measures proved
less effective than land uses at explaining variance in biological measures. Regression models
explained 29% of the variance in IBI (N = 81) and 23% in EPT (N = 65). However, in univariate
analyses where IBI was categorized as either poor or good based on TVA's criteria, both
instream cover and clean substrate scores were clearly related to fish IBI: sites with either low
instream cover or highly embedded substrates had a >90% chance of having poor fish
community integrity (Figure 5-5). The low overall explanatory power indicates either that both
of these biological measures were responding primarily to non-habitat related factors or that the
habitat quality measures used were not sufficiently sensitive indicators of physical stressors in
this basin.
5.2.3.3.4 Cumulative source index for each site
A cumulative source index for each site was computed, based on how many of four
stress-causing land uses (sources of stressors) were present within 2 km upstream of the site.
The four sources were: active coal mining or processing; major transportation corridors; > 10%
urban area; and > 10% cropland area. IBI was inversely related to the cumulative number of
sources present (Figure 5-6A) and was consistently "poor" or "very poor" (TVA rating) at sites
5-17
-------
Relationship Between Clean Stream Sediments and Fish IB I
(T-test, p = 0.02, N =80)
s H-u
0 }£
W 32
e "
» o «
E 2-8
"g 2,4
*2.0
03
4) -1 R
o 1'°
I I
Pr
1 -1
1 1
tnr fin
1 1 :
nrt
(
HI ±Std. Dev,
< ' ±Std. Err.
o Mean
Relationship Between Instream Cover and Fish 181
(T-test, p< 0.001, N = 80)
s
O A f\
O ^'u
1 3-6
1 3.2
O
| 2.4
I i
!
i 1
Poor
Good
Id ±Std. Dev.
CH ±Std. Err.
D Mean
FIGURE 5-5
Relationship between two instream physical habitat parameters, clean sediment (substrate
embeddedness) and instream cover, and IBI score, where IBI is categorized as either poor
(impaired) or good (unimpaired) based on TVA's criteria; fish community impairment is
associated with poorer habitat quality as measured by these two parameters
(from Diamond et al. ).
5-18
-------
60
50
40
ffl
30
20
10
B
IBI impairmef
t threshold
123
Cumulative Stressors
Mean+SD
Mean-SD
Mean+SE
Mean-SE
a Mean
1 2
Number of Stressors
FIGURE 5-6
Fish IBI (A) and maximum number of mussel species (B) in the Clinch/Powell basin as a
function of the number of stressors (from Diamond and Serveiss6)
5-19
-------
having all four sources present. In nearly all of these cases (88%), the proximal sources were
urban areas and mining. Similar results were found for the maximum number of mussel species
present at a site (Figure 5-6B). Sites having 2 or more proximal sources had a >90% probability
of having fewer than 2 mussel species present. Sites with one or no sources of stress had
between 4 and 18 species, which is still far less than the historical number of species reported
(>35 species at many sites ).
5.2.3.3.5 Potential effects of toxic chemicals
The risk analysis was hampered by the lack of water quality data sufficient for
examining correlations between water quality parameters, including toxic chemical
concentrations, and biological effects. The significant amount of variance in biological indices
that was unexplained by land use and habitat quality data suggests that other factors were at play.
Toxic chemicals may be released in municipal or industrial effluents, from coal mining or
processing activities, or transportation accidents. While macroinvertebrates can recolonize an
area within a relatively brief period following an episodic release, recolonization by fish and
especially molluscs may require years or decades, depending on distance and barriers to other
colonized areas. Figure 5-7 illustrates effects observed after catastrophic spills at Westmoreland
Coal Company and the APCO power plant on the Powell and Clinch rivers, respectively. In
1998, a large coal slurry impoundment on the upper Powell River failed, resulting in a massive
fish kill and substantial mortality of native mussels for a distance of more than 20 miles
downstream. A 1999 truck accident on the upper Clinch River in the Cedar Creek area resulted
in substantial loss of mussels, including more than 300 threatened and endangered mussels.12
5-20
-------
70
60
CA 50
0>
"5
g.40
CO
(fl
W
3
20
10
0
APCO spills
Clinch River, RM 266.1
1967 and 1970
Westmoreland Coal oil spill
Powell River, RM 177
1987
_L
Year
1898 1950 1965 1972 1978 1985 1988 1990 1991
FIGURE 5-7
Number of mussel species recorded over time at two sites in Clinch/Powell watershed affected
by large toxic point-source discharge events (from Diamond et al.6)
5-21
-------
5.2.4 Risk characterization
Risk analysis examined the available data on land use, instream habitat parameters and
biological assemblages and produced a limited set of statistical associations. The risk
characterization step interpreted these associations to suggest what the primary sources of risks
are and to explain observed trends in stream faunal diversity. It also described uncertainties and
presented management recommendations.
5.2.4.1 Ecological risks
Analyses indicated that up to 55% of the variability in stream fauna could be explained
by land uses, with mining and urban land uses exerting the most adverse effects. Key factors
appeared to be sedimentation and other forms of habitat degradation stemming from urban and
agricultural land uses and toxics from coal and urban areas. Riparian areas with more forested
land cover and less cropland, urban, or mining activity tended to be associated with less
sedimentation, more instream cover for aquatic fauna, cleaner substrates, and higher fish and
native mussel species richness. Our results suggest that if agricultural or urban use upstream is
great enough within the riparian zone, sedimentation effects and subsequent loss of habitat will
ensue for some distance downstream (1-2 km). These effects are accentuated in higher-gradient,
headwater areas.
Although riparian vegetation can reduce deleterious land use effects on water quality, it
is not clear that improvement of the riparian corridor alone in this watershed will necessarily
result in recovery of native mussel and fish populations. Little or no recovery of threatened or
endangered mussel or fish species has been observed in this basin despite improved water
quality.1 hi fact, results of this study suggest that the risk of native species extirpation is likely to
increase as more sources of potential stress co-occur. Of 10 remaining mussel concentration
5-22
-------
sites studied, only half appeared to be reasonably isolated from major roads, urban areas, mines,
and agricultural areas. This information suggests that native mussel populations are relatively
vulnerable to likely sources of stress in this watershed and that further extinctions or extirpations
are probable unless additional resource protection measures are taken.
Native fish and mussels have a high risk of extirpation due to endemism (i.e., restriction
to a very limited geographic area) and habitat fragmentation, resulting in populations that are too
inbred, small in size, and more susceptible to stressors. Populations are now more widely
separated than they were historically,3 which could lead to reduced recruitment success and
declining populations, especially in the presence of stressors. Therefore, it may be most useful to
further protect those populations that appear vulnerable due to proximity to mining, urban areas,
or transportation corridors. Protection and/or enhancement of the riparian corridor at these sites,
as well as protection from toxic spills and discharges, is probably as important for sustaining
endemic species as stocking new or historically important areas. If stream habitat as well as
water quality can be maintained or improved, present mussel and fish populations might be able
to expand into nearby areas, thus increasing the distribution and abundance of these species.
5.2.4.2 Uncertainties
Several uncertainties limited our ability to discern associations between causes and
effects in the upper Clinch Valley. First and foremost, as has just been noted, the available
biological information was only infrequently coincident in time and place with relevant instream
chemical measurements. Second, physical habitat assessment data were fairly qualitative and
relatively infrequent. Given the observed importance of physical stressors such as sedimentation
on valued resources in this water body, resource managers should use more robust habitat
assessment techniques that provide more quantitative data on impairments. Third, the
5-23
-------
macroinvertebrate measure EPT relies on family-level taxonomy, reducing its ability to
discriminate changes in the benthic community; a generic- or specific-level index probably
would provide better information. Fish IBI appeared to be a more sensitive index to stressofs,
probably because the metrics in this index have been demonstrated to be sensitive in a number of
other watersheds. Fourth, the apparent relationship between fish IBI and mussel species richness
or abundance, observed in the Copper Creek subwatershed, needs to be explored in more detail.
IBI is composed of a number of metrics, such as native species richness, that were potentially
more explanatory of mussel assemblages, but the unaggregated data were not available to this
analysis. It must be noted, however, that any comparisons between native mussel and fish or
macroinvertebrate data will be limited by the lack of overlap in sampling locations between
TVA's monitoring programs. Only eight sites in the entire watershed had data on mussels and
either IBI or EPT. Because of the paucity of mussel species occurrence data, the risks to mussel
species in the watershed could be over- or understated,
5.2.4.3 Management recommendations
The risk assessment has helped lend further credence to what many resource managers
had long conjectured were problems within the watershed, thereby providing more scientific
support to take actions to address problems. Based on the assessment findings, the USFWS and
TNC are considering the following types of management actions: riparian buffer protection;
building spill prevention devices along transportation corridors near streams and restricting the
type of materials transported over certain bridges; limited access of livestock to streams; better
monitoring and control of mine discharges to streams; maintaining existing natural vegetation;
BMPs for pasture and agricultural land to reduce sediment loading; and better treatment of
wastewater discharges.
5-24
-------
5.3 ECONOMIC ANALYSIS
The overarching goal for integrated ecological and economic analysis was to utilize the
findings of ERA in an economic analysis that would be relevant to environmental management
decisions in the watershed. The economists' team chose to focus on values held by valley
residents as important for determining how local decision-makers would act. The economic
analysis therefore addressed the task of valuing potential changes in biological diversity and
other ecological services at risk in the upper Clinch River Valley in Virginia and Tennessee, as
expressed by Valley residents. This task presented two major challenges. First, credible
measures of economic value needed to be integrated with the ecological assessment endpoints
such that the results would be useful in analyzing risk-relevant management and development
scenarios. Second, the techniques used in the study needed to be consistent with economic
principles of individual welfare maximization and to minimize biases associated with the
measurement process.
Ecologists, such as those conducting the W-ERA, and Clinch Valley residents were
thought to view the ecological assessment endpoints differently. Ecologists believe that
biodiversity is important for a number of reasons, including its contribution to ecosystem
resilience, i.e., the ability to withstand perturbations (such as from natural or human-caused
stress) without shifting to a different kind of ecological state.14 As stated earlier, however,
Valley residents had rated "preserving our rare plant and animal species" lower than five other
environmental concerns listed, and therefore might be unlikely to attach much value to the
diversity of the Valley's mussel fauna, for example. However, mussel health is a good indicator
of water quality, which residents had rated as most important. Because mussels are very
5-25
-------
sensitive to pollution, poor water quality will tend to impact mussels before other species in the
river, and before human health. The economists expected that Valley residents would value the
service provided by mussels as water quality indicators. Their approach was to design a survey
that would interpret the results of ERA in terms most likely to be meaningful to Valley residents.
This section is organized as follows: In Section 5.3.1, choice modeling is explored as a
potential tool for solving this difficult valuation problem. Section 5.3.2 presents a methodology
for integrating a choice modeling approach with ERA in the upper Clinch Valley, and Section
5.3.3 discusses the choice model results.
5.3.1 Methods for valuing biodiversity and environmental quality
5.3.1.1 Conjoint analysis vs. contingent valuation
Current approaches for assessing the value of environmental change, including changes
in biodiversity, involve predicting an outcome associated with the change and then using a
method such as the contingent valuation method (CVM,, see Section 2.2.2 and Appendix 2-A) to
estimate individuals' willingness to pay (WTP), for a beneficial change, or willingness to accept
(WTA) for a change that is detrimental.3 For example, Rubin et al.15 estimate the value of
preserving spotted owls in order to determine the benefits of preserving old growth forests, and
Stevens et al.16 calculate WTP for various levels of preservation of Atlantic salmon and bald
eagles. However, CVM tends to focus on losing or gaining the whole good, whereas
management decisions tend to address changing characteristics of the goods.17 For example, a
typical CVM question might be worded as follows:
a The use of WTP or WTA is a function of the perceived property right as well. See Freeman27 for a discussion.
This question was contrived for demonstration purposes only. A high-quality CVM survey would convey much
more information before the valuation questioa was posed.
5-26
-------
The upper Clinch/Powell watershed, which lies in southwestern Virginia and
northeastern Tennessee, is threatened by water quality insults from agricultural operations,
coal processing facilities, and urban runoff. The watershed is important habitat of many plants
and animals, including eleven endangered mussels that are found only in the Clinch River. The
river and adjacent areas are also used for recreational fishing, canoeing, picnicking, hunting,
and to a lesser extent, commercial fishing.
A nonprofit organization is seeking voluntary donations to purchase land and
conservation easements to protect water quality in the Clinch/Powell watershed. These lands,
which in total would comprise 2,200 acres and would help ensure the protection of 15 miles of
stream habitat, would then be managed by state land management agencies as preserved land.
Would you be willing to contribute $X to aid in the purchase of the land and conservation
easements?
In theory, CVM can measure both use and nonuse components of economic value (see
Section 2.2.2); however, all these components would be lumped together in the WTP estimate.
By contrast, conjoint analysis (CA) asks individuals to make choices about which state of the
world they would prefer, given that different states have differing levels of certain definable
attributes. The choice model, a variant of CA, elicits individuals' preferences by asking them to
consider a series of trade-offs. In contrast to CVM, which asks individuals to explicitly state
their WTP for a proposed change in environmental quality, choice models ask individuals to
choose from a series of possible outcomes (choice sets). This allows the researcher to obtain the
trade-offs that an individual is willing to make between any attributes presented in the choice
sets, as well as to estimate WTP.
5-27
-------
Choice models ask questions that may be more familiar to individuals. Individuals are
asked to choose among bundles of goods according to the level of attributes of each bundle. For
example, individuals routinely make choices among goods that have multiple attributes, such as
among five automobiles having different colors, engines, interiors, etc. A typical choice task
might ask the subject to choose the most preferred of the five, each having different
characteristics, including price. In contrast to CVM, which would ask the individual to assign a
price to each of the cars, the choice model task is more representative of the choices that
individuals regularly face in making transactions. CA relies less on the information contained in
I fi
the description of the scenario and more on the description of the attributes of each alternative.
The family of CA models, of which the choice model is a member, is receiving
increasing attention in the economics literature as well as in policy circles. Its use has been
legitimized by National Oceanic and Atmospheric Administration's (NOAA) proposed Habitat
Equivalency ruling, which arose in part due to criticisms of CVM during the Exxon Valdez
damage assessment case (60 FR 39816).a hi particular, NOAA recommended CA as a tool to
measure in-kind compensation for damaged natural assets.
Regional development problems and multiple use management are perhaps the ideal tests
of the usefulness of the choice model. With proper survey construction, the researcher can
measure many characteristics including use and nonuse values, as well as indirect use values
such as ecological services (see Section 2.2.2 for definitions of these values). Conjoint models
are particularly useful for disentangling likely complementarities between attributes. For
a Habitat equivalency argues that the appropriate measure of natural resource damages due to, say, an oil spill, is
provision (or augmentation) of ecological services that substitute for the services lost (e.g., improvement of wetlands
in other areas).
5-28
-------
example, changes in water quality could be positively correlated with endangered fishes, sport
fishing, and water-based recreation; with choice models, the effects of each of the attributes on
welfare can be estimated independently.
5.3.1.2 Choice modeling framework
To explain individuals* preferences for alternative states of the Clinch River Valley,, this
effort used a random utility model (RUM) framework, which is widely used in dichotomous-
choice CVM and travel cost modeling, as well as in CA. RUMs rely on choice behavior and
assume that individuals will choose the alternative that gives them the highest level of utility;
i.e., RUMs estimate the probability that an individual will make a selection based on the
attributes and levels of each possible choice. The RUM is directly estimable from choice models
(see Appendix 5-B for technical detail of the RUM framework).
5.3.2 Integrating the choice model with the ecological risk assessment
The task of integrating the measurement endpoints from the upper Clinch Valley ERA
(especially, IBI and mussel species richness) with indicators of social value proved a formidable
challenge, since they were not the type of endpoint the ordinary citizen is likely to think about in
his day-to-day life. Meetings were held in Abington, VA and Norris, TN between the
economists, ecological risk assessors and other individuals who had shown interest in biological
resource management in the Clinch Valley. The decision was made to approach the problem of
lack of familiarity with the ecological endpoints in two ways. First, succinct wording was
developed to express the relationship of these ecological endpoints to quality of life. After
several iterations, a survey was drafted, presented to focus groups, revised and then pilot tested.
Second, socially meaningful endpoints were included that were complementary to the
ERA measurement endpoints but outside of the ERA's original scope. For example, increased
5-29
-------
forestation of the riparian corridor would not only help protect mussel and fish biodiversity but
also increase the diversity and abundance of terrestrial fauna and birds and improve the quality
of smallmouth bass fishing. Since these endpoints are jointly produced, it was important that
they be jointly valued. Their inclusion expanded the choice sets to more fully describe the state
of the Clinch Valley environment and the auxiliary benefits of management policies aimed at
preserving biodiversity.
5.3.2.1 Choice model design
Choice model surveys are complex by nature. Each possible choice comprises bundles of
attributes, with each attribute having different levels. Because the potential for
miscommunication between the researcher and the survey recipient via the survey instrument is
great, two formal focus groups of 6 and 11 subjects and three informal focus groups were
conducted to inform our survey design. The first informal group was conducted in September
2000 using staff and students of the University of Tennessee. The second and more formal focus
group was conducted by an expert facilitator in St. Paul, VA in November 2000. The third and
fourth focus groups were conducted at the University of Tennessee in January and February
2001. The final focus group was conducted in Oak Ridge, TN in February 2001 using residents
of Anderson County, TN, the westernmost county in our study.
The focus groups allowed the participants to home in on those attributes correlated with
management changes that are likely to be important to the residents of the Clinch River Valley.
Six attributes were identified, with the number of levels per attribute varying from 2 to 6 (Table
5-3); see Table 5-4 for an example choice set from the survey. The "cost to household" attribute
allowed the estimation of conventional WTP measures. Interaction with the "Agricultural
5-30
-------
TABLE 5-3
Attributes and attribute levels used in survey questionnaire. Attribute levels making up options
A and B in a given choice set varied among those listed; attribute levels for option C were the
same in all choice sets." Corresponding model variable names are given in parentheses.
Attribute
Agriculture-
free zone
Aquatic Life
Sportfish
Songbirds
Agricultural
income
Cost to
Household
($ per year)
Attribute Levels for Options A & B
25 yards Clinch 11 0
yards tributaries
(BIGZONE)
full recovery
(FULLRECOV)
increase
(SPORTING)
10 yards Clinch /5
yards tributaries
(SMALLZONE)
partial recovery
(PARTRECOV)
no change
increase population
(SONGINC)
no change
$100
$75
$50
none
continued decline
decrease
(SPORTDECL)
no change
$1 million/yr decrease
(AGDECL)
$25
$10
$5
(COST)
Option C: No New
Action
none
continued decline
no change
no change
no change
no change
a The choice sets are designed to allow for the efficient estimation of the parameters of all of the attributes. While
SMALLZONE and BIGZONE are our policy variables, they arc varied independently of the other variables. For
example, it is possible to have choice sets that include the 25yard/10yard agriculture exclusion (BIGZONE), but
have SPORTDECL or have CONTINUED DECLINE for the level of aquatic life. Individuals would be expected to
focus on the outcomes and not the policy attribute.
See Appendix 5-A for explanatory text that was provided in the survey
5-31
-------
TABLE 5-4
Sample question and choice set from survey questionnaire
Which option for the future of agriculture and the environment in the Clinch Valley
do you prefer the most, Option A, Option B, or Option C? Option C is the status quo,
or what is currently happening and will continue to happen with no further environmental
or agricultural policies. Note that some of these options might not seem completely
realistic in real life. We ask that you do your best to assume that each option is
possible and then choose your most preferred option.
Option SSJ
Option B
ft;',Oj I - ',:!;:,';>
No New Action
Agriculture-free
zone
| yards Giinch/5 yards
:- ;," '':=: ;: .
10 yards Clinch/5 yards
tributaries
none
Aquatic Life
"full recov
partial recovery
v tii ill I il
Sportfish
no change
increase
no chanae
Songbirds
increase
no ch<
Agricultural
income
Cost to
Household
($ per year)
ck
no change
$50
Please check the option that you would choose:
Option A Option B
Option C
5-32
-------
Income" attribute allowed investigation of whether individuals think society as a whole, or
farmers and ranchers alone, should bear the burden of increased environmental quality.
Choice model variables were specified based on these attribute levels, and a priori
predictions of their signs were made (Table 5-5). The variables that represent the attributes
agriculture-free zone, aquatic life, and sportfish were each decomposed into two separate,
effects-coded variables to control for the three levels that each of these variables can take (see
Louviere et al.19 for a full discussion). Effects codes are an alternative to dummy variable codes
and are useful when interpreting the coefficients of a choice model.18'19 SMALLZONE and
BIGZONE represent the size of the agriculture-free zone;2 these are expected to be positive,
albeit weakly. PARTRECOV and FULLRECOV should be positive as individuals should be
more willing to choose options that lead to higher levels of recovery for aquatic life, other factors
being equal. SPORTDECL should be negative as individuals should be less likely to choose
options that represent decreases in sportfish populations, whereas SPORTING should be positive
by similar reasoning. SONGINC is expected to be positive, since many people value the
presence of songbirds. AGDECL is expected to be weakly negative, since income declines are
detrimental to the regional economy but not all respondents expect to be affected directly. COST
is expected to be negative; individuals are less willing to choose options that have higher costs
associated with them. Alternative-specific constants corresponding to options A and B
(ASCA, ASCB) are included to incorporate any variation in the dependent variable that is not
a An omitted third variable for the status quo, NOZONE, is implicit in the model; its coefficient can be determined
based on the coefficients of the included variables,
5-33
-------
TABLE 5-5
Choice model variables and expected sign
Variable3
CHOICE"
SMALLZONE0
BIGZONEC
PARTRECOVC
FULLRECOVC
SPORTDECLC
SPORTING0
SONGINCC
AGDECLC
COST
EDUC
AGE
RTVERVIS
MOSTMPO
FISHLIC
ENVORG
ASCAd
ASCBd
Expected Influence of Variable
NA
+
+
+
+
'
+
+
-
-
+
?
+
+
+
+
7
7
a Variable names are explained in Table 5-3 or in text
b Dependent variable
c Effects-coded variable
d Alternative-specific constant
5-34
-------
explained by the choice set attributes or respondent characteristics; there was no a priori
expectation as to their signs.
Selected socioeconomic information thought to be important was also included in the
choice model (Table 5-5). For example, it is common (though not universal) in the literature to
see more support for measures to improve environmental quality as the level of education
increases,20 so EDUC is expected to be positive. RJVERVIS, which is equal to 1 if the subject
visited the Clinch within the last year, is expected to be positive, since individuals are expected
to choose outcomes that improve the quality of their visits to the river. Likewise, MOSTIMP
(which equals 1 if the individual believes either that recreation is the most important use, or that
environmental quality is the biggest issue in the Clinch Valley) is expected to have a positive
sign. FISHLIC, which equals 1 if the individual holds a fishing license, should be positive;
individuals who fish should be more likely to choose options 1 and 2 that generally include better
environmental quality. ENVORG, which equals 1 if the individual belongs to an environmental
organization, should be positive. There was no a priori expectation about the effect of AGE on
choice.
Having defined these parameters, a RUM-based choice model (Appendix 5-B) is
developed as follows:
CHOICE = or, ASCA + «2ASCB + #SMALLZONE + &BIGZONE
+/?3PARTRECOV + /?4FULLRECOV + remaining 5-1
attributes and socioecono mic parameters + e
where the remaining attributes and socioeconomic parameters are all of the remaining terms in
Table 5-5.
5-35
-------
5,3.2.2 Survey implementation
Final language to describe the choice attributes to respondents was developed (Appendix
5-A). Respondents were asked to answer eight choice sets.a An example choice set is found in
Table 5-4.
Surveys were mailed to a random sample of 400 households in the Clinch River Valley,
with the majority being distributed in the Virginia portion of the valley, b'c Principles from
Dillman's Total Design Method21 were followed. Approximately two to three weeks after the
survey mailing, a reminder postcard was mailed to thank participants and encourage non-
respondents to return their surveys.
5.3.3 Results of economic analysis
Ninety one subjects completed the choice study (response rate was 23%); 76 provided
complete responses for all eight choice sets, generating 1824 acceptable observations for analysis
(see Table 5-6 for summary statistics).
A fractional factorial design was employed to develop a survey based on this choice model, A fall factorial design
would have required 648 (= 33*22*6') different choice sets. The %MKTDES macro in SAS was used to choose 16
choice sets that are meaningful and will still allow the main and interaction effects to be estimated. These 16 choice
sets were then blocked into two blocks of eight choice comparisons. One outcome of the focus group process was
that subjects indicated that the 16 choice sets that they had initially evaluated were too many.
The delivery envelope for the survey was personalized and included a cover letter, the survey, supporting
documents, and a stamped return envelope. Surveys were printed on legal size (8.5"xl4w) paper folded as a 20-page
booklet and stapled along the spine. The supporting documents were printed on letter size paper.
This survey was distributed as part of a larger study employing four different survey versions. The other surveys
allowed the examination of the trade-offs of strictly environmental attributes such that a preference-based index can
be constructed; a version where mussel protection implies trade-offs in employment in several sectors of the
economy; and a version designed to test the similarities between choice and contingent valuation models. Results of
the other surveys are still pending.
5-36
-------
TABLE 5-6
Summary statistics
Variable
EDUC
AGE
RIVERVIS
MOSTMPO
CHOICE
SMALLZONE
BIGZONE
PARTRECOV
FULLRECOV
SPORTDECL
SPORTING
SONGINC
AGDECL
COST
FISHLIC
ENVORG
Mean
13.409
45.855
0.592
0.627
0.333
-0.249
-0.236
-0.101
-0.287
-0.476
-0.328
-0.157
-0.358
24.391
0.453
0.200
Std. Dev
1.426
14.723
0.492
0.484
0.472
0.8239
0.836
0.902
0.746
0.707
0.875
0.987
0.933
31.810
0.498
0.400
Min
6
18
0
0
0
-1
-1
-1
-1
-1
-1
-1
-1
0
0
0
Max
16
81 |
1
1
1
1
1
1
1
1
1
1
1
100
1
1
Observations8
1800
1824
1824
1800
1824
1824
1824
1824
1824
1824
1824
1824
1765
1824
1800
1800
a There are 1824 possible observations representing 3 possible choices on 8 choice occasions for
each of 76 subjects. Ninety one subjects completed this Version of the choice study, but only 76
have complete responses for all eight choice sets.
5-37
-------
5,3.3.1 Results of choice model estimation
The interpretation of the coefficients in conditional logit models suggests how utility or
satisfaction changes given a change in the attribute. The parameters also reveal how the
probability that an alternative is chosen changes as the level of the attribute changes.
Parameter values obtained for the discrete choice model generally show the expected
signs and the joint power of the model is very good, as evidenced by a McFadden's R2 of 0.27
(Table 5-7). The signs of the coefficients on the attribute variables are consistent with the priors.
Both small and large agriculture-free zones serve to increase the probability that an alternative is
chosen, but the small zone has a stronger effect than was anticipated. Full recovery for aquatic
life and increases in sportfish are also positive, whereas decreases in sportfish have a negative
effect of the probability of choice. AGDECL is negative and significant, indicating that
individuals are less likely to choose alternatives if they know that agriculturalists have to pay
part of the costs of recovery efforts. COST is negative and significant, indicating a decreased
likelihood of choosing an alternative as the tax price increases.
In this model, each subject generates 24 observations (i.e., 3 possible choices on 8 choice
occasions) in the data set; thus, socioeconomic characteristics are invariant across choice sets.
The only way to control for socioeconomic effects is through interactions with the alternative
specific constants or interaction with the attributes. The decision was made to interact education,
age, gender, fishing license, and membership in environmental organizations with the alternative
specific constants. The interpretation of these interactions is complicated as well. For example,
ASCA*MALE and ASCB*MALE both are negative and significant (Table 5-7), indicating that
the probability of choosing Option A or B rather than the status quo in any of the eight choice
sets is lower for men than for women.
5-38
-------
TABLE 5-7
Results for conditional logit with CHOICE as dependent variable
Variable
SMALLZONE
BIGZONE
PARTRECOV
FULLRECOV
SPORTDECL
SPORTING
SONGINC
AGDECL
COST
ASCA
ASCAxEDUC
ASCAxAGE
ASCAxMALE
ASCAXMOSTIMPO
ASCAxFISHLIC
ASCAxENVORG
ASCB
ASCBxEDUC
ASCBxAGE
ASCBxMALE
ASCBxMOSTIMPO
ASCBxFISHLIC
ASCBxENVORG
Number of Observations3
Log-Likelihood
Log-Likelihood(O)
McFadden's Rho-square
Coeff
0.697
0.306
0.084
0.831
-0.727
0.593
0.079
-0,157
-0.033
-0.771
0.010
-0.013
-0.624
0.790
0.256
0.492
-1.505
0.044
-0.018
-0.696
0.561
0.700
0.246
Std. Error
0.155
0.158
0.148
0.150
0.179
0.127
0.120
0.069
0.004
1.288
0.088
0.008
0.266
0.264
0.250
0.308
1.465
0.099
0.011
0.313
0.310
0.302
0.379
T-statistic
4.497
1.936
0.564
5.541
-4.054
4.679
0.657
-2.271
-8.654
-0.599
0.119
-1.508
-2.345
2.993
1.024
1.597
-1.027
0.448
-1.671
-2.223
1.806
2.318
0.648
P-value
0.000
0.053
0.573
0.000
0.000
0.000
0.511
0.023
0.000
0.549
0.905
0.132
0.019
0.003
0.306
0.110
0.304
0.654
0.095
0.026
0.071
0.020
0.517
526
-423.759
-577.870
0.267
a There are 608 choice occasions in the data set, but only 526 observations have complete responses for the variables
in the regression. A choice occasion represents a set of three alternatives: one outcome is selected as the preferred
option by the individual, the other two are not.
5-39
-------
5.3.3.2 Calculation of part-worths
Using the coefficients from Table 5-7, implicit prices (with respect to the COST variable)
were obtained for each of the choice variables (Table 5-8). These are typically called the part-
worths in the conjoint/choice model literature.3 While in theory the calculation can be made in
terms of any one attribute for any other, the most intuitive trade-offs are those between dollars
and the other attributes. We can estimate the part-worths by dividing the coefficient on one of
the attribute variables by the coefficient on the COST variable and multiplying that result by
negative 1, For example, the part-worth on full recovery of aquatic life is
Dollar value of full recovery of aquatic life = - 5-2
\Ps)
where p$ is the coefficient on the variable COST. Respondents were willing to pay substantially
more for a small than for a large agriculture-free zone, suggesting perhaps that (a) the idea of an
agriculture free zone is attractive in and of itself, independent of any benefits expressed in the
other attributes, but that (b) such land use restrictions are most attractive when kept to a
minimum. The dollar-valued part-worth for partial recovery of aquatic life was insubstantial in
comparison to full recovery, and that for an increase in songbirds was similarly insubstantial in
comparison to that for improved sport fishing, or to the negative part-worth associated with a
decline in sport fishing.
a This is the marginal rate of substitution concept in economics upon which indifference curves are based. Simply, it
gives the trade-offs that an individual is willing to make between bundles of goods while holding utility constant.
5-40
-------
TABLE 5-8
Implicit prices, or implied willingness to pay for a given attribute level as compared with
the status quo
Attribute
SMALLZONE
BIGZONE
PARTRECOV
FULLRECOV
SPORTDECL
SPORTING
SONGINC
AGDECL
Implicit price ($)a
21.12
9.27b
2.55C
25.18
-22.03"
17.97
2.39°
-4.76
a Since the payment vehicle described in the survey was a change in tax rate
(see Appendix 5-A), values should be assumed to represent annual amounts.
b Coefficient for BIGZONE was marginally significant (see Table 5-7)
c Not significantly different from zero
5-41
-------
5.3.3.3 Calculating the value of a biodiversity management program
Economists are often interested in calculating the change in welfare, or well-being
(Section 2.2.1), due to a change in public policy. The p estimates allow the calculation of
compensating variation (CV), or total WTP, associated with any policy definable in terms of the
attributes. First, the utility of the status quo is calculated by substituting the appropriate variable
values defining the status quo attribute levels into Equation 5.1. Next, the utility of
the policy is calculated using the values corresponding to the attribute levels that define the
policy. Then, CV is given by
CV = (status quo utility utility of new policy) 5-3
fi$
Following these techniques for obtaining CV2 and using the coefficients in Table 5-7, the choice
model allows valuation of the multi-attribute change to be evaluated (e.g., in the case where
management actions lead to simultaneous improvements [or declines] in the various facets of the
ecosystem). If, for example, the status quo utility were taken as zero and a change in agricultural
practices were to improve habitat for mussel populations, sportfish, and songbirdsand farmers'
income were unaffected by the program-the welfare for the representative individual would
increase by $54.81 (i.e., the average respondent would be willing to pay $54.81 annually to move
from the status quo to the state of the world having the new agricultural practices). It is this
ability to derive multiple welfare measures for complex ecosystem changes that sets choice
models apart from CVM studies that allow calculation of the value for only a single policy
change.
5-42
-------
5.4 DISCUSSION
This section evaluates the cumulative outcome of the W-ERA and economic analysis
conducted in the upper Clinch Valley by comparison to the generalized conceptual approach for
ERA-economic integration developed in Chapter 3 (see Figure 3-1). As explained in Section
1.5.1, the Clinch Valley analyses were undertaken prior to the development of this conceptual
approach, and the economic analysis was initiated following completion of the W-ERA. For
these reasons, the studies conducted in the Clinch Valley should not be viewed as integrated in
any ideal sense. However, the conceptual approach for integration can be used to examine these
efforts in the larger context of watershed decision-making and management and to gain insights
as to ways that integration can be improved. The following discussion compares specific
components of the conceptual approach with work carried out in the Clinch Valley.
5.4.1 Consultation with extended peer community
The conceptual approach for integration has defined the "extended peer community" as
consisting of interested and affected parties, decision-makers, and scientific peers and has
argued, in agreement with the National Research Council23 and others,24'26 that these parties
should be actively engaged throughout assessment processes (see Sections 2.1.1.5, 3.2 and
3.3.5). The ERA for the upper Clinch Valley was undertaken by a diverse, interdisciplinary and
multiagency workgroup that included both government and nongovernment representatives, and
the risk characterization was conducted with scientific consultation (a workshop held by
USEPA) and formal peer review. The result was a creative, state-of-the-art analysis, the findings
of which have helped to identify potential management actions by workgroup member
organizations.
5-43
-------
Decisions were made at an early stage to conduct the W-ERA without an open process
for broader, public involvement. Through an informal survey, and long experience working in
the region, analysts had indications that community residents valued biodiversity less highly than
water quality, on one hand, and economic development opportunities, on the other. Therefore,
the management goal on which the ERA was based, which focused on biological integrity,
reflected the values of the technical specialists and environmental managers who composed the
interagency workgroup, rather than a broader stakeholder consensus as in other W-ERAs. This
decision undoubtedly allowed the workgroup to tackle the difficult problems of data gathering
and analysis more expeditiously; arguably, it may also have limited the development of broader
community awareness of biodiversity issues and mutual understanding of necessary trade-offs
for environmental protection.
The economic analysis team benefited from several consultations with members of the
ERA workgroup and selected stakeholder group representatives, in which ERA findings were
explained and regional economic development goals were discussed. Informal and formal
consultations (focus groups) with watershed residents were held to avoid miscommunication
between analysts and the public. The resulting survey instrument may be thought of as a
structured form of consultation with the public, in which aspects of ecological risk were
presented and feedback, in the form of choices between alternative states, was elicited.
Interestingly, certain results of the economic analysis ran counter to expectation about residents'
values. Survey respondents appeared willing to trade-off a portion of regional agricultural
income in order to obtain full recovery of aquatic life, and willing to accepteven to help
fundmeasures that would limit agricultural use of the riparian zone to improve habitat.
5-44
-------
5.4.2 Baseline risk assessment
The conceptual approach for integration defines baseline risk assessment as the
assessment of risks currently and into the future if no new management action is taken (Section
3.3.3). The upper Clinch Valley ERA used existing data to characterize the risks (and
uncertainties) affecting the assessment endpoints according to current conditions and trends. It
identified the impacts of multiple sources and stressors, and pointed to the future likelihood of
continued extirpations of species if stressors are not more effectively managed. It provided
models (in this case, empirical relationships) that could be used to assess the impacts of
management policies, including spatial relationships of riparian zone land use and in-stream
biological response and the impacts of multiple stressors. It did not attempt to evaluate any
management alternatives, however.
5.4.3 Formulation, characterization and comparison of alternatives
According to the conceptual approach, economic analysis of environmental problems
usually requires the evaluation of some action or policy to determine who would be affected,
how they would be affected, and to what extent. Therefore, it includes the steps in which
alternatives are formulated (Section 3.3.4), analyzed and characterized (Section 3.3.6) and then
compared to one another (Section 3.3.7). In the Clinch Valley case study, the economic analysis
had to examine management alternatives, even though the ERA had not done so. The economic
V
analysis specified two hypothetical agricultural policies (in addition to a status quo alternative)
for use in choice model construction. The apparent coherency of the choice model results
suggests that respondents understood the proposed policies and choice sets and that the model is
valid. However, it should be understood that the model per se does not characterize a specific
alternative. Rather, it is a flexible, albeit semiquantitative, tool that could be useful for
5-45
-------
comparison of specific policies after they had been analyzed and characterized, as Figure 5-8
illustrates.
Figure 5-8 compares the analytic processes used in two steps, analysis and
characterization of alternatives and comparison of alternatives, with those of a hypothetical
example that was presented in Figure 3-3. In the hypothetical example, the ecological risks,
economic effects, and health or other (sociocultural) effects of the management alternatives were
analyzed quantitatively to the extent feasible. Endpoint changes that could not be quantified
were expressed qualitatively. A stated preference study was used to value the nonmarket welfare
effects of the alternatives and improve the estimation of their net social benefits (Section 3.4.2).
Methods used in this case study comprise a subset of those described in the example.
Although the Clinch Valley W-ERA quantified relationships between land uses and ecological
endpoints, the endpoint changes expected to result from the two riparian management policies
introduced in the economic study were not quantified. Similarly, the financial costs and other
economic effects of implementing the policies were not analyzed. Equity issues were not
examined, and human health or other effects were not considered relevant to this case study. The
stated preference survey used qualitative language to describe expected ecological
improvements, whereas both the cost attribute and the attribute describing potential regional
impacts on agriculture were numerical (Table 5-4 and Appendix 5-A).
As a result, the choice model derived from the stated preference study would be capable
of comparing the benefits of these or other policies only after additional work was done. The
analysis and characterization of real alternatives would require the following additional steps:
5-46
-------
ANALYSIS & CHARACTERIZATION OF ALTERNATIVES
Ecological Risk Economics Health or Other
; H point
laii , i >ajK : I coats
'iPT1 :;".. !->3l-!:,g?pri
i
Qualitatively
describe
other changes
Qualitatively analyze
equity, economic impact
Express primary changes
in common language
Express equity effects, impacts
in common language
I
j; : , -
: Jl p "
Stated preference
study
COMPARISON OF ALTERNATIVES
T'lll'.' '_' '
-.'nati '/.'fi
ililfc
I'M ,'. I-.'.', .
- -
FIGURE 5-8
Techniques used for analysis, characterization, and comparison of management alternatives in
the Clinch Valley watershed, as compared to the example shown in Figure 3-3. White boxes and
bold type show features included in this analysis.
5-47
-------
determination of the decision context, including who could make the decision to
implement a given alternative, how they would decide, and who would stand to gain or
lose as a result (as part of planning, see Section 3.3.1)
detailed formulation of the alternatives, including design of structural (e.g., fencing) and
nonstructural (e.g., institutional) implementation measures (see Section 3.3.4)
determination of the ecological outcomes (efficacy, in terms of instream biological
response), economic outcomes (costs, including opportunity costs) and uncertainties of
the policy (see Section 3.3.6).
Using the choice model as a comparison tool would present several additional challenges.
Since the actual efficacy of a given exclusion zone for enhancing aquatic life can be estimated
only with substantial uncertainty, it would be difficult to determine how a given, best estimate of
increase in IBI should be evaluated in the choice model if the available choices are "partial" and
"full" recovery. Respondents ascribed statistically significant value only to "full" recovery. Yet
even a substantial, predicted increase in IBI would not necessarily signal a recovery of extirpated
species (and certainly not of extinct species), and implementation of an exclusion zone would not
reduce the very substantial risks from, e.g., transportation spills; therefore it would be hard to
rate any agricultural policy as leading to "full" recovery. Similar problems would be
encountered in coding the effects of an actual policy on sportfish and songbirds. Ultimately
there would be heavy reliance on expert judgment to interpret the ecological data and to apply
the choice model.
Nonetheless, the apparently successful development of this choice model suggests that
models of this type can be used for comparative welfare analysis of watershed management
5-48
-------
policies. What remains unanswered, however, is the important question of whether welfare
estimates are useful to decision-makers in a given case. Whereas large water resource
development projects may require welfare estimates, other kinds of decisions may not. For
example, if biodiversity protection in the upper Clinch Valley will continue to depend largely on
success by organizations such as The Nature Conservancy at acquiring federal grants for
voluntary riparian protection programs, and private funds for land acquisition, as is presently the
case, it is not clear that welfare estimates are needed. For any other protection mechanism under
consideration, the decision context specific to that mechanism would need to be examined to
determine what information is needed for decision support.
5.4.4 Adaptive implementation
The conceptual approach for integration suggests that when uncertainties are great,
management decisions should be implemented in an adaptive fashion, with continual
Devaluation of effectiveness and, as necessary, redesign (Section 3.3,9). The nature and
magnitude of biological response that may result from any program of riparian zone protection
are uncertain. However, programs can be adaptively designed in such a way that early stages of
implementation will yield the information needed to resolve specific questions and improve the
effectiveness of later stages. Riparian dimensional analysis indicated that the instream impacts
of riparian land use were most observable over a downstream distance of 500-1500 m (see
Section 5.2.3.2). This suggests that stream reaches of appropriate lengths in different
subdrainages could be pre-selected as treated and untreated replicates, with protection efforts
targeted accordingly. Such an approach could yield valuable information on the amount of
investment required to meet voluntary or regulatory goals for stream quality improvement in the
upper Clinch Valley and other, similar watersheds.
5-49
-------
5.5 REFERENCES
1. Neves RJ., Mollusks, in Virginia's Endangered Species, Terwilliger, K. Ed., McDonald and
Woodward Publishing Company, Blacksburg, VA, 1991,251.
2. Neves, RJ. et al., An Evaluation of Endangered Mollusks in Virginia, Virginia Commission
of Game and Inland Fisheries, 1980,149.
3. Stein, B., Kutner, L., and Adams, J., The Status of Biodiversity in the United States, The
Nature Conservancy, Oxford University Press, NY, 2000.
4. Jones, J. et al., Survey to Evaluate the Status of Freshwater Mussel Populations in the Upper
Clinch River, VA, Final Report, U.S. Fish and Wildlife Service, Abingdon, VA, 2000.
5. Walters, T., Small dams as barriers to freshwater mussels (Bivalvia, Unionidae) and their
hosts, Biological Conservation, 75, 79,1996.
6. Diamond, J.M. et al., Clinch and Powell Valley Watershed Ecological Risk Assessment,
EPA/600/R-01/050, U.S. Environmental Protection Agency, Office of Research and
Development, National Center for Environmental Assessment, Washington, DC, 2002.
7. Diamond, J.M. and Serveiss, V.B., Identifying sources of stress to native aquatic fauna using
a watershed ecological risk assessment framework, Environmental Science and Technology,
35,4711,2001.
8. Serveiss, V.B., Applying ecological risk principles to watershed assessment and
management, Environmental Management, 29,145, 2002.
9. USEPA, Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F, Risk Assessment
Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
5-50
-------
10. USEPA, Clinch Valley Watershed Ecological Risk Assessment Planning and Problem
Formulation - Draft, EPA/630/R-96/005A, U.S. Environmental Protection Agency, Risk
Assessment Forum, Washington DC, 1996.
11. Ortmann, A.E., The nayades (freshwater mussels) of the upper Tennessee drainage with
notes on synonomy and distribution, Proceedings of the American Philosophical Society, 52,
1918.
12. Hylton, R., Setback Hinders Endangered Mussel Recovery, Triannual Unionid Report, 16,
25, 2002.
13. Allen, J.D., Stream Ecology, Structure and Function of Running Waters, Chapman and Hall,
New York, NY, 1995.
14. Peterson, G.D., Allen, C.R., and Holling, C.S., Ecological resilience, biodiversity and scale,
Ecosystems, 1, 6, 1998.
15. Rubin, J., Helfand, G., and Loomis, J., A benefit-cost analysis of the northern spotted owl:
results from a contingent valuation survey, Journal of Forestry, 89, 25,1991.
16. Stevens, T.H. et al., Measuring the existence value of wildlife: what do CVM estimates really
show?, Land Economics , 67, 390, 1991.
17. Hanley, N., Wright, R.E., and Adamowicz, V., Using choice experiments to value the
environment, Environmental and Resource Economics, 11(3-4), 413, 1998.
18. Boxall, P.C. et al., A comparison of stated preference methods for environmental valuation,
Ecological Economics, 18, 243, 1996.
5-51
-------
19, Louviere, J.J., Hensher, D.A., and Swait, J.D., Stated Choice Methods: Analysis and
Application, Cambridge University Press, Cambridge, UK, 2000.
20. Sanders, L., Walsh, R., and Loomis, J., Toward empirical estimation of the total value of
protecting rivers, Water Resources Research, 26, 1345,1990.
21. Dillman, D.A., Mail and Telephone Surveys, the Total Design Method, Wiley, New York,
1978.
22. Cameron, T.A., A new paradigm for valuing non-market goods using referendum data:
maximum likehood estimation by censored logistic regression, Journal of Environmental
Economics and Management, 15, 355,1988.
23. NRC, Understanding Risk: Informing Decisions in a Democratic Society, Washington, DC,
1996.
24. Funtowicz, S.O. and Ravetz, J.R., A new scientific methodology for global environmental
issues, in Ecological Economics: The Science and Management ofSustainability, Costanza,
R. Ed., 1991,10,137.
25. Scheraga, J.D. and Furlow, J., From assessment to policy: lessons learned from the U.S.
National Assessment, Human and Ecological Risk Assessment, 7,1227,2002.
26. PCCRARM, Framework for Environmental Health Risk Management,
Presidential/Congressional Commission on Risk Assessment and Risk Management,
Washington, DC, 1997.
27. Freeman III, A.M., The Measurement of Environmental and Resource Values: Theories and
Methods, Resources for the Future, Washington, DC, 1993.
5-52
-------
APPENDIX 5-A
Excerpt from Survey Administered by the University of Tennessee: Explanation of
Hypothetical Agricultural Policies and their Potential Impacts
Background Information on the Clinch River Valley
The upper Clinch and Powell Rivers represent some of the last free-flowing river
segments in the Tennessee River system. Together, they drain approximately 3800 square miles
of land area. The Clinch and Powell Valley has one of the most diverse concentrations of
freshwater mussels and fish species of any river in North America. Many of the valley's mussel
and fish species are on the decline. Twenty-two mussels and eleven fish species are listed as
endangered or threatened. Moreover, the Clinch River Valley has many species that are found
nowhere else. Of the 50 mussel species that are listed by the U.S. Fish and Wildlife Service as
"Threatened" or "Endangered", 16 are found in the Clinch River Valley.
Ecologists believe that biodiversity is important for a number of reasons, including its
contribution to the health of the ecosystem (diverse ecosystems can better withstand and recover
from stressors such as drought). Mussel species are good indicators of the health of the
ecosystem. Because mussels are very sensitive to pollution, poor water quality will often affect
mussels before it has an impact on other species in the river and before it has a direct impact on
human health.
Although employment in the region is increasingly migrating to the manufacturing,
service, and tourism sectors, the economy of the valley has historically been based on coal
mining and agriculture. More than 40% of coal production in Virginia occurs within the
Clinch/Powell Valley and much of the discharge of pollutants in the region is not regulated.
The combined effects of raising livestock, pesticide runoff and soil erosion from farming,
forest clearing for development, coal mining and processing, discharge from sewage treatment
facilities and septic tanks, chemical spills, runoff from roads, parking lots, and chemically treated
lawns decrease water quality and reduce mussel and fish abundance and diversity.
Evaluating Changes in Agriculture to Protect the Environment
One cause of reduced water quality in the river is that livestock get into the river,
crashing mussels, eroding river banks, and muddying the water. Intensive cultivation of crops
near the river allows fertilizers, pesticides, soil and other substances to contaminate the river as
well.
5-53
-------
These problems could be lessened by the development of an "agricultural free zone" in
the immediate proximity of the river. This zone, where crop planting and grazing would be
restricted, could be of different widths. In our study, we ask you to compare the present case of
no agriculture free zone with two alternative zone sizes: a zone 10 yards wide on the Clinch and
5 yards wide on tributaries or a zone that is 25 yards wide on the Clinch and 10 yards wide on
tributaries.
Farmers who keep cattle would need to construct fences to keep the livestock out of the
exclusion zones. Fences would keep the cattle from trampling the mussels, reduce erosion and
sedimentation of the river. Trees would shade the river water, reducing its summertime
temperature and increasing the dissolved oxygen level, which would benefit aquatic life. As the
pastures revert to more naturally occurring types of vegetation, songbird and wildlife populations
could increase. The construction offences and substitute watering facilities for the cattle, and the
loss of the use of the land are costly for farmers. Farmers who grow crops would not be able to
plant in the zones, which may be among their most fertile (and flattest) land holdings.
However, the farmers need not bear the full cost of the policy. A pilot project has been
underway where non-profit organizations such as The Nature Conservancy have been
compensating farmers who construct fences and take lands near the river out of production. This
type of project could be expanded and funded through a small increase in taxes for everyone in
the Clinch Valley. The questions below ask respondents to compare possible alternative policies.
One primary difference among the policies is the extent to which the farmers or the taxpayers
bear the costs. Farmers could be fully or partially compensated for their losses. Another set of
differences involve the levels of the environmental characteristics. These changes in agricultural
practices may have effects on aquatic life, sportfish, and songbirds. The ranges of these effects
that we would like you to consider are as follows:
Aquatic life: includes all non-game fish and mussels. Changes are in terms of diversity,
abundance and distribution throughout the watershed.
Continued Decline = continued decreases in diversity, abundance and distribution in the
Clinch River and its tributaries.
Partial Recovery = some improvement in the Clinch River, but no improvement in
tributaries
Full Recovery = improvement in the Clinch River and its tributaries
Sportfish: Includes smallmouth bass, trout, etc. Changes are in terms of number and average
size.
No change = current numbers and distribution of sizes
Increase = 20% increase in Clinch and tributaries
Decrease = 20% decrease in Clinch and tributaries
5-54
-------
Songbirds: Changes are in terms of variety of species and number of birds found in the Clinch
River Valley
No change = current numbers of birds and varieties of species in the valley
Increase = 20% increase in numbers of birds in the valley
Agricultural income: Changes are in terms of lost income in the agricultural sector of the
Clinch River Valley economy. These losses would accrue to farmers in the 21 counties that are
part of the valley as a result of decreased production.
No change = no change in agricultural income
Small decrease = $1 million/year total decrease in production. This represents less than 1
percent of total farm income for the valley.
Cost to household: One way of financing improvements to the quality of the Clinch River is to
ask residents of the valley to share in the costs of protection. If you live in the Virginia portion of
the valley, this could be implemented through small changes in state income taxes. If you live in
the Tennessee portion of the valley, this protection could be paid for through small changes in
local property taxes.
5-55
-------
APPENDIX 5-B
RANDOM UTILITY MODEL
Using random utility theory, one can model discrete choices assuming that individuals
make choices that maximize their utility, or well-being. If the utility of alternative / is greater
than the utility of alternative j, the individual will choose i. Utility is composed of both
deterministic (environmental quality, income, etc.) components and random, individual-specific,
components that are unobservable to the researcher. The random utility model (RUM)
framework is directly estimable from conjoint rankings and choice models.
Following Roe et al.2 and Stevens et al.3, the utility of a management program i is given
by
'UV,*> 5-B-l
where the utility (U) of program i for the individual is a function of the attributes (q) of i and
where z represents individual characteristics. While utility is an interesting measure of
preferences, it is not particularly valuable because it does not reflect the trade-offs, financial or
otherwise, that individuals must make in order to consume a bundle of goods. Thus one typically
considers the indirect utility function, which expresses utility as a function of income and prices:
Ui=vi(pi,qi,m,z) + £i 5,B_2
where v is indirect utility and/? and m represent price of the state of the world i and income of
the individual, respectively.
Then the standard RUM can be estimated from the discrete choice conjoint data using
conditional logit:
5-56
-------
The probability that the program having attributes i is chosen is the probability that the indirect
utility of program i plus a random, unobservable error is greater than the indirect utility of
program 0 and its error term.
Then v is estimated using a linear functional form of the indirect utility function, by
means of the conditional logit model specified generally as:
v = const + ^Attributes + fySodoeconomic 5-B-4
The stylized model in Equation 5-B-4 generates the probability of choosing a particular option
given the levels of attributes of the option and the individual's (socioeconomic) characteristics.
The P's generated from the above equation are the coefficients associated with each of the
attributes in the choice model.
To estimate the welfare impacts, or willingness to pay, for a change from the status quo
state of the world to the chosen state one calculates:
where CV (compensating variation) is the income adjustment necessary to leave the individual as
well off with bundle i as they were with bundle 0, Additionally, the p's from Equation 5-B-4 can
be used to calculate implicit prices, or part- worths, for each variable with respect to all of the
other variables in the model (see Section 5.3.3,2).2
a This is the marginal rate of substitution concept in economics upon which indifference curves are based. Simply, it
gives the trade-offs that an individual is willing to make between bundles of goods while holding utility constant.
5-57
-------
REFERENCES
1. Boxall, P.C. et al., A comparison of stated preference methods for environmental valuation,
Ecological Economics, 18, 243,1996.
2. Roe, B., Boyle, K.J., and Teisl, M.F., Using conjoint analysis to derive estimates of
compensating variation, Journal of Environmental Economics and Management, 31,145,
1996.
3. Stevens, T.H., Barret, C., and Willis, C., Conjoint analysis of groundwater protection
programs, Agriculture and Resource Economics Review, October, 229, 1997.
5-58
-------
6. SEEKING SOLUTIONS FOR AN INTERSTATE CONFLICT OVER WATER AND
ENDANGERED SPECIES: PLATTE RIVER WATERSHED
6.1 WATERSHED DESCRIPTION
6.1.1 Watershed resources and impacts of development
The central Platte River floodplain in Nebraska, which includes the 130 km of river
known as the "Big Bend Reach," is rich in biodiversity and ecologically complex. The reach
extends from near Lexington, NE on the west to immediately below Grand Island on the east.
Nested within the Platte River watershed (Figure 6-1), which encompasses 223,000 km2 (86,000
mi2) in Colorado, Wyoming and Nebraska, the central floodplain occupies 13,280 km2 (5130
mi2) and hosts a diverse assemblage of ecosystems, plants and animals. Approximately 50
species of mammals and several hundred species of terrestrial birds use the cottonwood-willow
forests and wet meadow grasslands near the river for breeding or stopover habitat during
migration,1 Nearly one-half million sandhill cranes {Grus canadensis) and several million ducks
and geese use the Platte River during their annual migration.2 In addition, the central Platte
River floodplain supports nine species of plants and animals that are listed as threatened or
endangered, including the interior least tern {Sterna antillarum athalassos), the piping plover
{Chamdrius melodus) and the whooping crane {Grus americana), and another 12 species that are
candidates for federal listing.3 The high levels of biodiversity found in this reach are at risk,
however, due to the cascading effects of reduced water flows and development on ecosystem
structure and function.
Irrigation water from the Platte River and adjacent aquifers has made the Platte Valley a
highly productive agricultural region, providing irrigation water to over one million acres. Water
storage reservoirs such as Lake McConaughy and Johnson Reservoir have provided increased
6-1
-------
Lexington Grand Island
Kearney
FIGURE 6-1
The watershed of the North Platte, South Platte and Big Bend Reach of the Platte River in the
Great Plains of the USA. Towns and reservoirs mentioned in the text are indicated.
6-2
-------
recreational and sportfishing opportunities, contributing to the more than two million recreational
visitor days per year provided by the river. Platte River hydropower stations help meet regional
energy demand by supplying 300 MW of hydroelectric power. As a result, the natural
hydrologic regime has been influenced by more than 200 upstream diversions as well as by 15
dams and reservoirs on the North and South Platte Rivers, all but one of which are in Colorado
and Wyoming.4 This elaborate network of dams, diversions, and irrigation canals has resulted in
a 70% decline in peak discharge,5
From a hydrogeomorphological perspective, the Platte River is braided stream whereby
the main channel contains a network of smaller channels separated by small islands called braid
bars. Braided rivers are also characterized by highly erodible banks and an abundance of
sediment. In a braided system that is unregulated, the number and location of the channels and
braid bars may change quickly as a function of stream discharge and sediment load. In turn, the
dynamic nature of braided rivers creates a mosaic of habitats such as shifting sandbars, side-arm
channels, backwaters, and temporally inundated floodplains. Combined, this rich array of
habitats supports high levels of floral and faunal biodiversity. Critically, however, flood-pulsed
hydrology6 is needed to sustain this diversity of habitats and species. These flood pulses
typically occur in spring as a function of snow melting in the stream's headwaters with river
disturbance scouring established habitats and creating new ones. The flood pulse also maintains
an important seasonal connection of the river channel to the floodplain, which distributes energy
and nutrients between the river and the land, and supports ecosystem functions such as
production, decomposition, and consumption.6"8 On the Platte and other rivers, these water
fluctuations also drive patterns of vegetation succession.9"11
6-3
-------
In contrast to unregulated river systems, damming and other alterations to the natural
flow regime alters the nature of the pulse transmitted to the Platte floodplain. As a result, the
Platte has experienced reduced channel movement and environmental heterogeneity. In addition,
in regulated Rivers such as the Platte, sediments become trapped behind dams, so downcutting
and erosion occur in the downstream channel, further isolating the channel from the
floodplain.12-14
Channel width on the Platte has been reduced 85-90% over the last century or so.11'15
Establishment ofPopulus dominated forests has followed narrowing of the main channel and
stabilization of river braids. Approximately half of the active channel present in the middle
Platte in the 1930s had succeeded to woodlands by the 1960s due to the combined effects of
irrigation, streamflow regulation and drought.11 In total, some 9500 ha ofPopulus woodland are
established in the Big Bend reach.
The significant alteration-of the natural flow regime notwithstanding, high levels of
faunal biodiversity are associated with the present channel structure. Two species of particular
concern are Platte River populations of the least tern and piping plover listed as endangered
and threatened, respectively by the U.S. Fish and Wildlife Service. Terns and plovers nest on
large, high-elevation, barren sandbars. Historically, spring flooding during ice pack breakup
would scour vegetation off of midstream sandbars, leaving the necessary open nesting substrate.
Establishment of riparian forest has significantly reduced available habitat. Sandhill cranes,
perhaps the flagship species of the Platte, are also highly dependent upon open channel habitat.
Approximately 80% of the continental population of cranes spend about six weeks in spring
staging on the central Platte River. Sandhill cranes roost in open channels and forage for
invertebrates in nearby wet meadows and for waste corn in nearby farm fields. * Much has
6-4
-------
been written about the preferences of roosting cranes for open channel habitat. In general, cranes
prefer roosting in shallow water and with channel widths of 500 feet or more and rarely inhabit
those that are less than 150 feet. They may roost in concentrations of 20,000 per mile. Roosting
on the river protects them from their predators. The issue is complex, however, because many
factors are involved in the selection of roosting sites including availability of and distance to off-
channel food (wet meadows and corn fields), weather, water depth, stream flow, and distance of
roosts to tall vegetation.2'18"23 Crane use has declined in the upper Platte River coincident with
dramatic channel narrowing between 1930 and 1957, and has since increased farther downstream
where channels have narrowed less.24 However, large populations of cranes roost in the
relatively narrow channels of the North Plate River or roost away from the river in wet
meadows.22 The effects, if any, of such displacement are unknown.
The channels are also important to a wider variety of migratory water-birds including
whooping cranes and a variety of ducks and geese.16'17525"27 Waterfowl population estimates
during migration range from 5 to 9 million individuals in spring.28'29 Most of the migration
population consists of snow geese, Canada geese, greater white-fronted geese, mallard and
northern pintail.
Wet meadows that flank the Platte River support a rich assemblage of migratory and
breeding grassland birds.30 Of principal concern to this avian community are the effects of lower
water tables on habitat structure and forage and particularly habitat fragmentation.30'31 An
important conservation objective is the maintenance of sufficiently large habitat patches for core-
grassland (no-edge) species including upland sandpiper, bobolink, grasshopper sparrow,
dickscissel and meadowlark.30
6-5
-------
While alteration of the natural hydrological regime poses significant risk to many species,
the establishment and evolution of the riparian Populus forest has created significant ecological
opportunity for other species, principally those which use riparian forests. For example, based
on a two year study of 72 woodland patches, Colt32 showed that these forests support some 50
species of breeding birds, including 32 neotropical migrant species, a guild of birds that includes
several species with populations at risk. Further, Colt and Jelinski (unpubl data) have
preliminary findings that suggest nest success is high, and that some species are not rendered as
vulnerable to deleterious edge effects (e.g., predation and nest site parasitism) found elsewhere
on the Great Plains. The resulting increase in avian biodiversity as a result of altered flows
broadens the number of stakeholders to include those concerned about off-channel species.
The seeming bonanza of forest bird species may substantially change, however. In less
than a century, and barring a catastrophic major disturbance, the Populus dominated forests will
almost be completely replaced via succession by equilibrium forests dominated by Fraxinus
(ash) as Johnson10 has predicted for the Missouri River floodplain forests. A profound
biodiversity decline may result because a large proportion of flora and fauna is restricted to, or
strongly associated with, Populus communities (Jelinski and Colt, unpubl paper). It is well
established that maximum diversity of trees, birds, and small mammals occurs in older Populus
forests midway along the sere.9'33'34
In summary, the flood-pulse system6'8 that is characteristic of the central Platte River
floodplain links hydrology with biological communities and ecosystem processes in complex
ways.9'11>31 Alteration of the natural flow regime for hydropower, food production, and
recreation has changed the dynamic nature of the river and places some species and habitats at
6-6
-------
risk. At the same time, hydrological alteration of the Platte has created ecological windows of
opportunity for a number of other species.
The effect of altered flows, habitat fragmentation, and agrochemical runoff on riparian
vegetation in the central Platte River floodplain have been extensively studied, whereas effects
'y
on some avian communities have barely been investigated, and science is only in the early
stages of predicting impacts on fish and other wildlife communities. ]'36
6.1.2 Watershed management efforts
A long history of efforts to protect the resources of the central Platte River floodplain
forms the backdrop for the ecological and economic analyses discussed in this chapter.
Conservation organizations and governmental agencies have worked to improve avian habitat
along the Big Bend Reach, while Federal and State agencies and various stakeholders have
sought ways to resolve enmeshed conflicts between economic demands for water withdrawal and
environmental needs for increased, and seasonally varying, instream flows as determined under
the Endangered Species Act (ESA). Over the past 25 years, a number of management initiatives,
often backed by technical analyses, have been tried.
To improve habitat suitability for cranes, waterfowl and native grassland birds, the
National Audubon Society, Platte River Whooping Crane Maintenance Trust, and The Nature
Conservancy have acquired tracts of wet meadow and river channel. They have eliminated
roads, fences and buildings and have consolidated land units to reduce disturbance and habitat
fragmentation. The Natural Resource Conservation Service of the U.S. Department of
Agriculture (NRCS) and the U.S. Department of Interior's Fish and Wildlife Service (USFWS)
have cooperated to restore wet meadow and open-channel roost habitat for cranes by removing
woody vegetation from sandbars in the river channel. These actions have not been without
6-7
-------
controversy, however, as the mechanical removal of some tracts of late serai vegetation to
recreate early-successional habitats has favored the requirements of certain wildlife species while
destroying established habitat for others. There is also scientific disagreement over the extent to
which riparian land management can effectively substitute over the long term for restoration of
stream flow. 23'37
Concerning required flows, some scientists contend that high stream flows are needed
periodically to prevent vegetative growth on sandbars and sustain the wide and shallow riverine
habitat preferred by whooping and sandhill cranes,38'39 whereas others contend that such scouring
flows are of little value and may actually be harmful in the case of fish, because scouring flows
lead to lower reservoir levels and higher water temperatures.40 The terms of the legal debate
over stream flow are defined by ESA provisions that prohibit any Federal action jeopardizing the
continued existence of a species designated as threatened or endangered, and provide that
USFWS determine species' requirements based on the best available scientific information. The
USFWS has determined that an additional 417,000 acre-feet (514 hm3) per year of water is
needed to meet endangered species needs for the Big Bend Reach in a wet-to-average year.3'
Absent any agreement as to how to make up that deficit, this determination is sufficient to
preclude any major water consuming action that constitutes a federal nexus. In other words, the
U.S. Forest Service (USFS) water leases in Colorado cannot be easily renewed; Wyoming cannot
pursue additional, federally permitted upstream water storage projects that would increase
consumptive use; and the public power districts in Nebraska cannot be assured of getting a long-
a This annual volume does not include less frequent flow recommendations such as a 5-year peak flow of 16,000 cfs
for channel maintenance.
6-8
-------
term hydropower license from Federal Energy Regulatory Commission (FERC) unless some
accommodation of the competing demands can be made.
Stakeholder groups have been actively involved in management discussions that have
occurred in the context of water right litigation, power plant licensing hearings, legislative
debates and other venues (FERC, 1998), Environmental interests in all political jurisdictions
(Colorado, Wyoming, Nebraska and the USFWS) tend to agree on the need for increased and re-
regulated stream flow and management of riparian lands for endangered species protection.
Irrigation interests are much more parochial both between and within states. Upstream surface
water irrigators have sought the right to continue irrigating and, in some instances, the right to
develop additional acreage. Downstream surface water irrigators want their water supply
protected against additional depletions from upstream irrigation or environmental demands.
Groundwater irrigators in all locations have sought the right to pump at will, irrespective of
stream flow considerations. Hydropower interests want high reservoirs to maximize feet of head
and would like to make reservoir releases during the summer months when electricity is worth
the most. Coal fired electric utilities want assured cooling water supplies and expansion
opportunities. Finally, recreation interests have mixed demands, including moderate reservoir
storage levels, stream flows that sustain fishing and waterfowl hunting, and easy access to the
river and to bird watching opportunities.
Since 1976 the Nebraska Department of Water Resources (DWR) has held over 400 days
of public hearings to address proposed diversions or requested instream uses of Platte River
water. From 1983-1997, the public power districts in Nebraska were in negotiations with the
FERC over the relicensing of Lake McConaughy. hi addition, from 1986 - 2001 the states of
Wyoming and Nebraska were in litigation over the interstate allocation of Platte River water.
6-9
-------
The straggle to manage the Platte system has led to several attempts to facilitate resource
management decisions, including empirical modeling with and without stakeholder input, several
negotiation formats, multi-state litigation and, most recently, a tristate-federal Cooperative
Agreement41 that takes an interim, adaptive management approach to the problem. One of the
first organized attempts to reach a compromise solution was an adaptive environmental
assessment process which began in 1983. Called the Platte River Forum, this approach involved
identifying a group of experts and stakeholders and assembling them in a single location for one
week. This group first identified the relevant impact variables and policy options. Then, with
the help of experts, the associated technical relationships were described in mathematical terms
and computerized. The idea was that stakeholder participation and input would lead to a widely
supported simulation model and agreement regarding the consequences of management
options.4 This expectation proved to be invalid. Not only did participants fail to agree on all the
facts, but even when there was general agreement on how the natural system worked, differing
value judgments and varying objectives prevented completing a model that was very useful for
determining how the water should be used.43
The Platte River Forum was responsible in part for the formation of a small research
group to develop a multi-objective model of the Platte. This model was built by a group of
university professors without stakeholder involvement.44 Whereas the Platte River Forum
focused on the physical aspects of the river system and considered only a small set of
alternatives, the multi-objective model focused on the delineation of trade-off curves for
numerous alternatives. The intent was to improve on the Platte River Forum by producing
additional information for decision-making and to do so without the inefficiencies and biases of a
committee of 30, many of whom represented stakeholder interests rather than areas of expertise.
6-10
-------
The outcome of the multi-objective modeling approach can best be characterized as good science
that was unused and ineffective. The scientists involved, operating independent of political
pressure, were able to produce a credible operational model, but the results were not embraced
by any interest group or decision-maker.
A third attempt to resolve the water management problem involved relicensing of
hydropower plants. From 1986 until a provisional hydropower license was issued in 1997, the
Central Nebraska Public Power and Irrigation District and the Nebraska Public Power District
were involved in an intensive effort to get the FERC to relicense their Platte River hydropower
facilities. The central issue was protection of threatened and endangered species, but NEPA
requirements associated with licensing a public resource also meant that broader fish and wildlife
issues, including sandhill crane habitat, had to be addressed. The major hydropower facility
involved is part of the Kingsley Dam which creates Lake McConaughy.
Lake McConaughy is the largest reservoir on the Platte River and the closest one to the
endangered species habitat. Historically, Lake McConaughy has been used to directly irrigate
over 200,000 acres (77,000 ha) and to enhance the groundwater supply for an additional 300,000
acres (112,000 ha).45 It has also been managed as a fishery in cooperation with the Nebraska
Game and Parks Commission and is a significant recreational resource drawing over 600,000
annual visitors per year. For nearly 50 years, however, the water entering Lake McConaughy
was managed in a serial dictatorship with irrigation receiving first priority for the water,
followed by hydropower and recreation. Endangered species were not considered. This all
changed when the original hydropower license expired in 1987. FERC required the Districts to
address wildlife habitat maintenance and enhancement, which led to extensive study by the
Districts and by environmental interest groups, and eventually to intensive negotiations between
6-11
-------
the Districts, environmental interests and FERC. However, the parties were unable to agree on
how to balance endangered species with other needs. Licenses were nevertheless issued
provisionally, with a requirement that the districts' operations be coordinated with the proposed
Cooperative Agreement.
As the pressures for reallocating water to meet endangered species needs mounted,
Nebraska interests sought to broaden the responsibility for meeting these needs to include
Colorado and Wyoming. Of the two million acres irrigated with surface water within the Platte
Basin, Colorado has 56 percent, Wyoming 12 percent and Nebraska 32 percent. It seemed unfair
to Nebraska water interests that they should have to meet endangered species needs without
appropriate contributions from Colorado and Wyoming.45 At the same time Colorado was facing
endangered species problems with Forest Service water rights and with potential irrigation
projects, while the threat of subjecting U.S. Bureau of Reclamation projects to consultations
under the ESA had eastern Wyoming and western Nebraska irrigators nervous.45 All three states
found that cooperation was in their mutual interest and negotiated the Cooperative Agreement,
initiated in 1994 and signed on July 1,1997.
The Cooperative Agreement constituted a multistate-federal effort to protect Platte River
endangered species without unduly constraining the availability of water for other uses. It
established a preliminary agreement to increase instream flow by an average of 130,000-150,000
acre-feet (160-185 hm3) and to acquire an initial 10,000 acres (3,900 ha) of an eventual 29,000
acres (11,200 ha) of riparian habitat, but did not set forth where all of the water would come
from nor what land would be acquired. The participants had three years to study alternatives and
to agree on sources of water and land, including a distribution of the costs. (As of this writing,
however, progress has been slow and the period for reaching agreement has been extended to
6-12
-------
June, 2003 and may be extended further.) If agreement is reached, the plan is to be put in place
and monitored for 10-13 years to determine how well the program is meeting endangered species
needs. If an agreement is not reached, the public power districts in Nebraska may lose their
provisional hydropower licenses, holders of water right leases on Forest Service lands will find
renewal very difficult, new surface water development in all states will be difficult, if not
impossible, and actions to protect endangered species will be further delayed.
Whether the Cooperative Agreement is successful or not remains to be seen, but thus far
none of the management approaches used have led to a comprehensive resource management
plan that addresses the conflicting demands of competing interest groups.
6,2 ECOLOGICAL RISK ASSESSMENT
6.2.1 Planning
Concern over threats to the valued biodiversity of the central Platte River floodplain,
coupled with evidence that various agencies and stakeholders would be willing participants
(Table 6-1), motivated the U.S. Environmental Protection Agency (USEPA) in 1993 to establish
an interdisciplinary workgroup to begin a watershed ecological risk assessment (W-ERA). The
goal was to obtain a better understanding of how the central Platte River landscape and
associated flora and fauna are being impacted by water withdrawal and other stressors. The
workgroup was composed of individuals with disparate interests and responsibilities and many
years of experience working in the central Platte River watershed. The planning process
included face-to-face dialogue between assessors and resource managers, a group tour of the
watershed, symposia, public meetings, focus group meetings and teleconferences.
Recognizing that any protective management actions would have to be weighed against
the need for human uses, the workgroup developed the following management goal for the
6-13
-------
TABLE 6-1
Participants in planning for the central Platte River floodplain W-ERA
Central Nebraska Public Power and Irrigation District
Nebraska Public Power District
Nebraska Department of Environmental Quality
Nebraska Natural Resources Commission
Central Platte Natural Resources Districts
Nebraska Game and Parks Commission
Tri-Basin Natural Resources Districts
Nebraska Department of Agriculture
The Nature Conservancy
Prairie Plains Resources Institute
Platte River Whooping Crane Maintenance Trust (PRWCMT)
University of Nebraska Lincoln and Kearney
US Fish and Wildlife Service
US Geological Survey
US Department of Agriculture
6-14
-------
watershed: protect, maintain and, where feasible, restore biodiversity and ecological processes
in the central Platte River floodplain, to sustain and balance ecological resources with human
uses. The management goal is a qualitative statement that addresses concerns expressed by
various agencies and management organizations as well as the floodplain residents and other
stakeholders.
6.2.2 Problem formulation
This section summarizes the problem formulation exercise conducted for the central
Platte. The intricacies of that process, and the limitations of the resulting analyses presented in
the following section, illustrate the difficulty of narrowing a broad management goal for a large
and complex system to a tractable set of risk assessment problems.
The management goal was interpreted by representatives from USEPA's Region VII and
Office of Water, the USFWS, the U.S. Geological Survey and Nebraska officials (listed in Table
6-1) into potentially implementable environmental management objectives (Table 6-2). A more
detailed description of the watershed than that presented in Section 6.1 was developed, along
with a description of the environmental problems in the watershed. The environmental problems
emanate from a combination of physical and chemical stressors. Of the many human-caused
stressors thought to be interfering with attainment of the goal, eight principal stressors were
selected by the workgroup (Table 6-3), using a Delphi ranking technique4 that documents
iterative group input and helps groups reach consensus. Nine ecological assessment endpoints,
representing three spatial scales, were selected (Table 6-4) that met the criteria of (a) relevance
to environmental management objectives, (b) ecological relevance and (c) susceptibility to
stressors (see Section 2.1.1.2).
6-15
-------
TABLE 6-2
Eleven environmental management objectives that are implicit in and required to achieve the
management goal
Affected Area
Channel
Riparian Forest
Backwaters
Floodplain
Landscape
Environmental Management Objective
1
2
3
4
5
6
7
8
9
10
11
Restore and maintain stream channel dynamic equilibrium
Maintain sufficient flows to prevent high temperatures detrimental
to native fish populations
Maintain range of successional stages of forest vegetation
Maintain and reestablish backwater ecosystems
Maintain and restore hydrologie connectivity between river
channels through surface flows
Maintain hydrologie connectivity between river channels and wet
meadow ecosystems
Maintain and reestablish natural diversity in wet meadow systems
Maintain and reestablish natural diversity in native upland systems
Protect and where feasible reestablish the mosaic of habitats in the
central Platte River floodplain to support key ecological functions
and native biodiversity.
Maintain diversity of water-dependent wildlife including migratory
and nesting birds, mammals, amphibians, reptiles and invertebrates.
Prevent toxic levels of contamination in water consistent with state
water quality standards
6-16
-------
TABLE 6-3
Principal stressors (and their primary sources) in the central Platte River floodplain
Altered surface water regime (dams and diversions)
Truncated sediment supply (dams and diversions)
Altered ground water regime (dams, diversions, groundwater withdrawal and irrigation)
Physical alteration of habitat (land conversion to agriculture, including drainage of wet
meadows, and clearing of vegetation for wildlife management)
Nutrients (fertilizer use)
Toxic chemicals (agricultural biocide use)
Harvest pressure (fishing, seining, waterfowl hunting)
Direct disturbance (roads, off-road vehicles, bird watching)
6-17
-------
TABLE 6-4
Ecological assessment endpoints for the central Platte River floodplain W-ERA.
Landscape scale
Habitat scale
Organism/Population
level
Floodplain landscape mosaic structure, function and change
Open channel configuration and distribution for migratory birds
Side channel and backwater area and connectivity to main channels
Riparian vegetation successional stage, areal extent and dispersion
Wet meadow composition and abundance
Sandhill crane and waterfowl diversity, abundance and dispersion
Core grassland breeding bird diversity and abundance
Amphibian survival and reproduction
Riverine and backwater fish and invertebrate survival and reproduction
6-18
-------
As in other risk assessments discussed previously, detailed conceptual models, developed
for each endpoint, were used to hypothetically attribute stressors to their sources and to explain
their impact on the assessment endpoints. Three of the nine assessment endpoints, or
representative elements of them, subsequently were selected as priorities for detailed quantitative
analysis. Those endpoints, and the corresponding risk hypotheses that were derived from the
conceptual models, are presented in Table 6-5. These three were selected because they capture
the predominant concerns regarding birds and unique habitat in the floodplain and because they
crystallize water and riparian management conflicts. All three are linked to the fact that lower
rates of flow reduce channel habitat for species such as sandhill cranes, piping plovers and least
terns '17>1 '4 and reduce shallow groundwater levels, thereby desiccating wet meadows and
reducing habitat diversity.48 However, lower flows promote the establishment of riparian forests
favored by other avian species.
The embattled nature of the Platte River management problem was evident during the
problem formulation process. An initial draft of the planning and problem formulation report
was presented to, and amended by, the stakeholder group in February of 1996. Subsequently, the
draft was further revised by the risk assessment team, in accordance with USEPA's concurrently-
developing ERA guidance. Upon release of the revised draft,49 some of the stakeholders
considered the revised draft overly environmentalist in tone and a breach of group process, and
they formally complained to USEPA by way of their Congressional representatives. To some
extent, this disagreement reflects a divergence in values and objectives between the larger
environmental community and those who live in the region. As such, it is characteristic of the
problems encountered when the benefits of environmental improvements accrue to a broad
community, while most of the costs are incurred locally.
6-19
-------
TABLE 6-5
Selected assessment endpoints and stressors and the associated risk hypotheses developed during problem formulation for the
central Platte River floodplain W-ERA.
Priority
Assessment
Endpoints
Riparian
vegetation
successional
stage, areal extent
and dispersion
Core grassland
breeding bird
diversity and
abundance
Sandhill crane
abundance and
distribution
Principal Stressors
Altered surface water
regime
Truncated sediment supply
Physical alteration of habitat
Toxic chemicals
Altered ground water
regime
Physical alteration of habitat
Altered surface water
regime
Truncated sediment supply
Physical alteration of habitat
Direct disturbance
Altered ground water
regime
Risk Hypotheses
1. Lower flows have led to reduced reworking of channels, greater cottonwood regeneration, less
heterogeneity of riparian vegetation.
2. Reductions in sediment may alter development of river braids by lowering river bed elevation,
decreasing sediment deposition on floodplain, increasing stability and reducing riparian heterogeneity.
3. Removal of riparian woodland vegetation by mowing and cutting reduces patch size and diversity of
riparian vegetation.
4. Herbicide drift and runoff from agricultural fields have caused physiological stress and perhaps
increased mortality in riparian vegetation.
5. Lowered water table reduces diversity of wet meadow vegetation and renders adults, eggs and young
more susceptible to predation.
6. Loss of habitat, reduction of patch size, and fragmentation of habitat may lead to decline of species
requiring large wet meadows.
7. Lower flows lead to additional woody plant establishment, channel narrowing and deepening, and
roosting habitat fragmentation. These changes reduce roost suitability, increase crowding and may
increase susceptibility to disease or other catastrophic events.
8, Reductions in sediment supply reduce channel braiding and thus open-channel roosting habitat.
9. Wet meadow conversion to crops has fragmented crane foraging, loafing and resting habitat;
channelization has reduced roosting habitat suitability.
10. Auto and rail traffic and crane-based tourism disturb migrating cranes.
1 1 . Lowered water tables reduce the production of wetland invertebrates, tubers and seeds that provide
forage for migrating cranes.
Source: Jelinski
-------
6.2.3 Analysis
Because of reassignments and shifting priorities, only a portion of the quantitative
exposure and stress-response analyses that were contemplated could be completed, even for the
reduced list of three assessment endpoints. This section presents those partial analyses,
6,2.3.1 Riparian vegetation snccessional stage, areal extent and dispersion
The risk hypotheses attributed fragmentation and loss of heterogeneity of riparian
vegetation to reductions in instream flow and sediment supply, as well as to riparian habitat
management measures., including mowing to create crane roosting habitat. It was also
hypothesized that agricultural herbicide use may pose additional stress. Reductions in mean
annual flow, peak flow and sediments in the central Platte River during the period of regulation
are well documented, as are reductions of active (unvegetated) channel area, increases in wooded
area and decreases in wet meadow area since the onset of regulation. ' '' " ' Therefore, the
veracity of hypotheses 1 and 2 (Table 6-5) is not much questioned, but efforts to develop
quantitative relationships between these variables, to enable estimates of risk, were not
completed. Analysis of herbicide impacts on riparian vegetation was not undertaken, nor was
there an analysis of riparian management effects on patch dimensions.
6.2.3.2 Core grassland breeding bird diversity and abundance
Risk hypotheses postulated that lowered ground water levels and habitat destruction and
fragmentation reduced habitat suitability for, and survival of, several grassland nesting species.
Therefore, an analysis of habitat use data was performed. Helzer and Jelinski30 surveyed 45 and
52 grassland patches, in 1995 and 1996 respectively, in the central Platte River valley. Patch
size ranged from 0.12 to 347 ha; roughly half of these meadows were used for grazing, the others
for haying. In each patch, four randomly selected, 100-m transects (4 ha total area) were
6-21
-------
surveyed twice between May 17 and July 5; and species that are exclusively grassland nesters
were censused. Where intended sampling area exceeded patch size, patches of similar size
characteristics were combined. Patch area and perimeter were determined using aerial
photographs and digital planimeter. Thirteen wet meadow breeding species were found during
the two field seasons; the six most common were used in species occurrence models, and all 13
were used in species richness analysis.
Occurrences of all six common species and species richness were most strongly (and
inversely) correlated to perimeter-area ratio, indicating that habitat use by wet-meadow nesting
species is maximized in patches that provide the most abundant interior area, free from edge
effects. These findings directly supported hypothesis 6 (Table 6-5). Since wetness or
vegetational diversity within these patches was not measured, hypothesis 5 was not evaluated.
An analysis of diversity and abundance of 50 woodland breeding bird species was also
carried out but was not completed (Colt and Jelinski, unpublished data). During the 1995 and
1996 breeding seasons, birds were censused in 72 woodland habitat patches ranging in size from
0.02-44 ha and were analyzed in relation to five spatial variables (related to patch size and shape)
and 15 habitat structural variables (e.g., tree species richness, average tree basal area, canopy
height, tree density, percent area flooded). In preliminary findings (May 2000 communication
by D. Jelinski to V. Serveiss and R. Fenemore), both richness models and occurrence models
(the latter were significant for 24 species) tended to indicate that although structural variables
(including canopy cover, shrub stem density and percent area flooded) were significant for some
species, spatial variables related to patch size were more important in general. These findings
suggest that a statement similar to hypothesis 6 can be made for woodland avifauna.
6-22
-------
6.2.3.3 Sandhill crane abundance and distribution
Over a six-week period during spring migration, approximately 500,000 sandhill cranes
stage in the central Platte River floodplain, with an individual staying about 2-4 weeks to rest
and accumulate fat reserves. Cranes are known to roost in the evening in broad, shallow
segments of the river channel. They prefer channels at least 150 m wide and 10-15 cm deep,
with unobstructed views. Though they will roost in channels less than 150 m wide, they avoid
those less than 50 m in width.23*51 Faanes and LeValley24 evaluated population changes among
four staging areas and found that a west-to-east shift had occurred. This shift was attributed to
loss of roost habitat in some of the western river segments and to scouring river flows and human
removal of woody vegetation providing more desirable roost site in some eastern segments.
Controversy exists, however, as to whether the river channel is now in a state of equilibrium with
respect to suitability for crane roost habitat, or in a state of decline.11'39
Risk hypotheses attributed reductions in roost suitability to reduced river flows, reduced
sediment supply, reduced acreage (and wetness) of wet meadows, channelization, and direct
c*y
disturbance (Table 6-5). The Cadmus Group attempted to evaluate relationships between
sandhill crane distribution and habitat and to develop a model capable of predicting future
changes in crane use of staging habitat in the central Platte River valley. Using habitat data
determined in 1982,29 coupled with USFWS annual, one-day crane census data for the flanking
years 1980 -1984, evaluations were performed by bridge-to-bridge river segment (N = 15), by
river reach (N = 10) and by crane staging area (N = 4). Associations by bridge segment were
weak, most likely because bridge segments are not ecologically meaningful. On the river reach
scale, mean unobstructed channel width showed the best relationship to crane density (r 0.45;
p<0.05), while the density of wet meadows (ha of wet meadows per river kilometer) showed a
6-23
-------
rather weak relationship to crane density. When data are aggregated by staging area, the
relationships improve, and crane density is a function of both mean channel width and the
density of wet meadows, in a two-step relationship. First, if mean channel width is less than
about 50m, cranes will not be present. For staging areas with mean channel widths greater than
50 m (i.e., Kearney to Chapman, Lexington to Kearney, Sutherland to North Platte), the
following best-fit regression model was obtained:
ABUND = 318 + 3.74 MEADOW -139 ALFALFA 6-1
where ABUND is crane density (numbers/km of river) and MEADOW and ALFALFA are
density (ha/km of river) of wet meadows and alfalfa fields, respectively. For this model, the
adjusted r2 was 0.754 and p was 0.0002; the standard errors of the intercept, MEADOW, and
ALFALFA were 147,1.28, and 0.67, respectively. The regression of crane abundance versus
density of wet meadows alone was also significant (p = 0.0002; adjusted r2 of 0.665); the best fit
equation for this model was:
ABUND = 39.9 + 5.49 MEADOW 6-2
in which the standard errors for the intercept and Meadow were 69 and 1.08, respectively. These
findings are generally consistent with aspects of hypotheses 7-9 and 11 (Table 6-5). They
demonstrate that there is an apparent threshold for acceptable channel width, above which the
availability of forage habitat (especially wet meadows, and to a more limited extent alfalfa) is
most important. However, data on channel widths and areas of wet meadows and alfalfa fields
more recent than 1982 were unavailable to test the model, limiting its confidence and reliability.3
3 The PRWCMT has collected additional data on crane use between 1998 and 2002, but as of this writing not all of it
has been converted to a useable form.
6-24
-------
Furthermore, the relationship between the primary stressors - i.e., reductions of flow, sediment
and ground water level - and either the habitat variables or crane abundance could not be
investigated by this approach, and thus the analysis was not directly applicable to decisions
related to water management. Data on direct disturbance (hypothesis 10) were not available for
analysis.
6.2.4 Risk characterization
As mentioned above, risk analyses for the central Platte River floodplain were not
completed,, and therefore risk characterization, or the translation of exposure and response
analyses into meaningful and, where possible, quantitative - statements about risk, could not
be carried out. Nonetheless, the W-ERA served to summarize existing knowledge about risks to
a set of valued ecological endpoints in the region, to focus information needs on a set of risk
hypotheses and to provide new data and quantitative relationships for several of these endpoints.
These findings are potentially valuable because factual disagreements underlie some of the
ongoing resource management disagreements discussed in Section 6,1.2. Whereas the questions
currently driving policy are specific to the water and habitat needs of federally threatened and
endangered species, the ecological risk problem was formulated more broadly to examine the
ecological integrity of the region as a whole. These results do not directly address the question
of target flows in the Big Bend Reach, but they do speak directly to the importance of
maintaining broad, active river channels and a diverse riparian landscape mosaic - i.e., one that
includes wet meadow patches with large interior dimensions and forested patches of varying
serai stage - as the means to protect regional biodiversity, particularly of birds.
6-25
-------
6.3. ECONOMIC ANALYSIS
Environmental economics often approaches environmental management problems as
budget-constrained, social-utility maximization problems, in which a key role for analysis is the
quantification of policy-relevant costs and benefits, including those related to nonmarket goods
(see Section 2.2), so that a socially optimal policy can be found. Ecological economics often
takes a similar approach - while adding a sustainability or other biophysical constraint.
Experience with Platte River decision-making, however, suggests that technical analysis alone
does not lead to a resource management equilibrium, either optimal or suboptimal. First,
information asymmetries create principal-agent problems (see Sections 2.2.1 and 2.2.5). For
example, states have an incentive to overstate their political compensation costs for providing
environmental water. Second, the presence of multiple objectives and stakeholder groups means
that the optimal management plan is different for each stakeholder group and that a global social
optimum cannot be achieved without weighting the relative importance of each. Such weights
are never explicitly assigned, but instead are implied by the decisions that are taken. A resource
management equilibrium is reached only when each stakeholder group believes that the cost of
further negotiations or political action exceeds the value of the expected change in outcome, a
condition which closely approximates the classic Nash equilibrium in economic game theory.53
All participants in the dispute over environmental management in the central Platte River
floodplain have a strong incentive to reach a solution. Without a negotiated solution, the federal
government will have greater difficulty meeting its ESA obligations; agriculturalists could face
federal imposition of very high instream flow requirements; environmentalists would encounter
further delays before instream flows are increased; and the states face continued uncertainty,
hampering their individual water management and economic development programs and
6-26
-------
threatening higher costs if a settlement is imposed. In spite of these incentives, the parties have
been unable to reach an agreement. A case in point is the need to follow-up the general
agreement reached in the Cooperative Agreement - i.e., to increase Platte River flows by
130,000-150,000 acre-feet for 10 years and to monitor the results with a specific agreement as
to how the state and Federal parties will provide and pay for the water. All stakeholder groups
continue to argue over technical issues and to take strategic positions designed to improve the
resource management outcome from their point of view.
Recent developments suggest that selected game theory techniques (see Section 2.2.5)
may be useful in resolving this conflict. Game theory occasionally has been applied to water
resource management problems during the last decade. Becker and Easter54 used game theory to
analyze the dependency among eight states and two provinces concerning water diversions from
the Great Lakes. Diversion decisions were modeled under different scenarios with different
restrictions on the lakes where diversions could occur. The results suggested that states do not
necessarily divert water because they stand to gain relative to the status quo, but because they
may lose more if they follow an alternative future strategy. In a case similar to the central Platte,
Adams et al.55 proposed game theoretic models in the form of computer simulations to
investigate the likely outcome of negotiations among agricultural water users, environmental
groups and municipal water users in California. Their results indicate that the outcome of the
negotiation process depends crucially on the institutional structure of the game, the input each
group has in the decision-making process, the coalitions of groups that can implement proposals,
the scope of negotiations and the outcome if parties fail to reach agreement.
The principal appeal of game theory to the central Platte bargaining problem is that it
offers the potential of inverting the problem from a case where stakeholder representatives
-------
propose solutions to each other to one where stakeholders respond to solutions suggested by
game models. This increases the possibility that an equilibrium solution will be found, because
all bargaining strategies are simultaneously considered and because mathematical manipulation
is likely to reveal solutions that may not emerge in a round table bargaining process. Although a
realistic game model for this situation is unlikely to have a solution that meets all constraints,
and it will certainly not have a unique solution, the game theory approach may still have
considerable merit. It forces the participants to consider the role of incentives and strategic
behavior in bargaining and, if nothing else, increases the likelihood that individual stakeholder
groups will pursue policy options that are attractive enough to all participants to have a
reasonable potential for successful implementation.
The decision to focus the economic analysis on the Cooperative Agreement process, and
to use game theory, was made by the economic research team of the University of Nebraska -
Lincoln (UN-L) in their application for a USEPA grant. Some team members had a longstanding
involvement with the instream-flow negotiations. After the grant was awarded, and prior to the
start of work, an informational meeting was held in 1999 involving USEPA, the UN-L research
team, a representative of the Nebraska Natural Resources Commission (familiar with stakeholder
concerns and the Cooperative Agreement), the Platte Watershed Program Coordinator of the
UN-L Cooperative Extension Service (familiar with habitat management efforts), and the lead
researcher for the W-ERA. Participants were informed regarding the status of the W-ERA, the
status of the Cooperative Agreement, and the proposed economic research approach.
For this analysis the central Platte management problem was defined in terms of two
game models: Model I, which addresses who should provide and pay for environmental water
(i.e., water reallocated to instream flow for purposes of maintaining or enhancing biodiversity),
6-28
-------
and Model II, which addresses how much water should be so reallocated. Data for Model I were
obtained from available reports, whereas Model II required a survey of households in Colorado,
Nebraska and Wyoming. The next sections present the methods and results of each model in
turn.
6.3.1 Model I: Determining who should provide and pay for environmental water
The parties to the Cooperative Agreement (initiated in 1994 and signed in 1997 by
Colorado, Wyoming and Nebraska) have agreed as to an incremental amount of instream water
(i.e., 140,000 acre-feet) that would constitute a first step in the adaptive implementation of
measures to protect threatened and endangered species in the central Platte River floodplain.
However, they have not been able to fully agree on the source of the water or who would pay for
it (as well as a number of other administrative details). This study hypothesized that an auction
approach capable of addressing information asymmetries would lead to an agreement in
circumstances where other negotiating strategies may break down. After examining auction
techniques (see Klemperer56 for a comprehensive review), the approach selected was a second-
price, sealed-bid sequential procurement auction with descending bidding and predetermined
cost shares. In a sequential procurement auction, one unit (in this case, a given quantity of
water) is auctioned at a time, and a single buyer receives bids from several sellers. In a
descending-bid (or English) procurement auction, price falls incrementally until only one seller
remains. If the auction is of the second-price (or Vickrey) variety, the winning seller receives the
second-lowest bid, which eliminates the incentive for a seller to bid higher than his minimum
price. Most of the auction literature deals with auctions where a single unit is sold at a time.
Sequential versions of each standard auction type exist, although their use is not well
researched.56
6-29
-------
The only players in this game are the three states. Environmental or agricultural interest
groups are not players because their primary concern is assumed to be the amount of
reallocation., not who pays and at what price. The federal government's only role is to commit to
a given cost share at the beginning of the game. The predetermined cost shares define how much
each state and the U.S. Department of Interior (DOI) contribute to the cash pool for purchasing
environmental water. The state with the winning bid incurs an obligation to supply the water in
return for a payment from the cash pool. Although the use of predetermined cost shares may be
unusual or unexpected, it is consistent with the terms of the Cooperative Agreement mentioned
in Section 6.1.2.
It is well-known that for a sealed-bid, second-price auction it is a dominant strategy for
each player to announce costs truthfully.56 The descending English auction design does not
necessarily result in truthful revelation of all costs, but it does result in a dominant strategy
equilibrium that minimizes welfare costs. All players bid until only the two lowest cost players
remain; then the agent with the second lowest cost stops at his cost and the lowest cost player
wins the auction with a bid equal to (or slightly below) the second lowest cost. Mathematical
details and proof that the strategies result in a Nash equilibrium have been reported elsewhere.57
6.3.1.1 Data sources
The data needs for this model consisted of acquisition costs, third party costs and political
compensation costs. Acquisition costs represent what each state will need to spend to acquire the
water for reallocation to environmental uses, such as for acquiring water rights, for providing
additional storage, or other costs depending on the water source. Acquisition costs were
Co
compiled from a recent report prepared for use by the states and the DOI in resolving the
central Platte management problem. Third party costs were assumed to be 10 percent of
6-30
-------
acquisition costs based on historical levels of unemployment and underemployment and on
regional input-output model results for the central Platte region59 and the states of Nebraska,
Colorado and Wyoming,
Political compensation costs are the payments above expected opportunity costs (i.e.,
foregone economic benefits) that the states may demand as compensation for the political
turmoil and economic uncertainties associated with agreeing to supply a given quantity of water.
These values can be inferred from game results if the game is actually played, rather than
simulated as in this study. For purposes of this analysis, three different levels of political
compensation were defined which, based on the investigators' observation of the Cooperative
Agreement Governance Committee's discussions on this issue, were expected to bound the
problem: no compensation, moderate and high. Political compensation for the moderate case,
expressed as a multiple of the real cost, started at near zero for the first blocks of water supplied
by a single state and increased exponentially to 20 percent of real cost at 50,000 acre-feet and to
57 percent at 140,000 acre-feet of water supplied. Corresponding points on the political
compensation function for the high compensation case were 40 percent of real costs at 50,000
acre feet and 113 percent at 140,000 acre feet.
Simulations assumed a cost-share policy consisting of Colorado 0.2, Nebraska 0.2,
Wyoming 0.1 and the DOI 0.5. These shares are based on the initial cost allocations that were
incorporated in the 1997 Cooperative Agreement between the states and the DOI. Water was
procured in blocks of 10,000 acre-feet with minimum bid increments of $0.50 per acre-foot.
Results were computed for water supply quantities ranging from 10,000 to 420,000 acre-feet per
year (i.e., the total increment recommended by USFWS), but all welfare comparisons were
6-31
-------
calculated for a quantity of 140,000 acre-feet, the target quantity adopted under the Cooperative
Agreement,
6.3.1.2 Model I results
Water supply costs under three different political compensation policies are depicted in
Figures 6-2a to 6-2c. In Figure 6-2a, the observed difference between marginal cost and bid
price is the second-price gain, whereas in Figures 6-2b and 6-2c that difference includes political
compensation costs as well. Under a no-political-compensation policy (Figure 6-2a) the costs
are lowest, but Nebraska would need to supply 110,000 out of 140,000 acre-feet, or 79 percent of
the water. This finding reflects the fact that most of the low-cost water is in Nebraska,6 but
results of preliminary multi-state negotiations to develop a water supply plan suggest that a cost
minimization approach is not likely to be politically acceptable. Under these circumstances one
would expect Nebraska to bid high in order to either get adequate political compensation or
induce another player to supply the water, whichever comes first. Under the simulated effect of
political compensation (Figures 6-2b and 6-2c), exponential increases in Nebraska's bid price,
and a corresponding increase in cumulative budget costs, provide incentives that cause supply by
Wyoming and Colorado to increase. However, net welfare costs (Table 6-6) increase less than
budget costs, because the budget increase is largely in the form of political compensation
transfers among the parties. The second price effect increases as political compensation
increases, with second price gains going to those who supply the water. Most of the increased
welfare costs accrue to the federal share, because they supply no water and therefore receive no
second price gains or political compensation transfers.
In summary, these findings present a scenario in which a mutual supply agreement,
unachievable up to this point, could be reached for a modest increase in total welfare costs (when
6-32
-------
ouu -
400
onn
JUU -
onn
1 nn
luu -
n -
a. No political
compensation
* _ ----"
5K Nebraska marginal cost
Wyoming marginal cost
A Colorado marginal cost
Bid price received
Cumulative budget cost
^^piTfr^Hi********'
_
....--"""
^.-^t- * *"
**-* A
iuu
80
fin
An
i
- n
£"
JO
CL
onn
o
to
O * nf\
Q lUU
Rnn
onn
zuu -
mn
n -
b. Moderate political
compensation
^^KXJK*^
-.^^^XXX^^JK.*
c. High political
compensation
^eC^C^*
* ..--""'
0 50 100
h-
>
~*
I
1 40 x 1 03 acre-feet , .-t 't ' ^
^^^^^
^j^^^^^^A^^
^^
r^'
150 200 250 300 350 4
t/>
In
en o
bU Q
"S
. An D)
3
CO
3
o
»- inn
an
ou
-in
- 4U
'
«
on
A
DO
Cumulative Water Amount Supplied (103 acre-feet/yr)
FIGURE 6-2
Price of 10,000-acre-foot increments of environmental water, and cumulative cost, assuming
different levels of political compensation
6-33
-------
TABLE 6-6
Welfare effects from supplying 140,000 acre-feet of environmental water.
Level of Political
Compensation
None
Water Supplied (AF/yr)
Budget Cost ($/yr)
Second Price Gain
Political Compensation
Net Welfare3
Moderate
Water Supplied (AF/yr)
Budget Cost ($/yr)
Second Price Gain
Political Compensation
Net Welfare
High
Water Supplied (AF/yr)
Budget Cost ($/yr)
Second Price Gain
Political Compensation
Net Welfare
Welfare Costs3
Colorado
0
-3,057,823
0
0
-3,057,823
0
-3,552,900
0
0
-3,552,900
20,000
-3,960,740
+111,191
+372,900
-3,476,649
Nebraska
110,000
-3,057,803
+1,772,443
0
-1,285,360
100,000
-3,552,900
+1,362,207
+2,201,000
+10,307
80,000
-3,960,740
+867,424
+2,881,100
-212,216
Wyoming
30,000
-1,528,912
+360,652
0
-1,168,260
40,000
-1,776,450
+404,612
+442,700
-929,138
40,000
-1,980,370
+469,561
+889,500
-621,309
Federal
0
-7,644,560
0
0
-7,644,560
0
-8,882,250
0
0
-8,882,250
0
-9,901,850
0
0
-9,901,850
Total
140,000
-15,289,120
+2,133,095
0
-13,156,003
140,000
-17,764,500
+1,776,819
+2,643,700
-13,353,981
140,000
-19,803,700
+1,448,176
+4,143,500
-14,212,024
'Welfare costs represent the real cost of the water to all parties combined. Net welfare is equal to the budget cost less
that part of the budget cost which represents transfer payments. Both second-price gain and political compensation
payments affect the distribution of welfare among the parties but not total welfare, because the loss to the paying party
equals the gain to the receiving party.
6-34
-------
compared to a least-cost scenario). The auction approach would resolve principal-agent
problems (see Section 2.5.5) by creating incentives for each state to incrementally reveal its true
political compensation costs. The resulting agreement is likely to benefit all the parties because
each can choose between supplying the water at an acceptable minimum price or paying
someone else to supply it.
6.3.2 Model II: Determining how much water to allocate to environmental use
Whereas Model I examined only the provision of water and constrained the problem to a
negotiation among three States, Model II casts the negotiation problem more broadly. Questions
(policy attributes) examined in this model were the following:
1. What method or approach should be used for meeting endangered species' needs
in the central Platte River floodplain (Method attribute)?
2. What is the appropriate level of investment in meeting species' needs (Cost
attribute)?
3. Who should make that investment (Whopays attribute)?
The players included the federal government and environmental and agricultural interest
groups, as well as the states. Because all parties stand to gain if agreement is reached, the
decision process was modeled as a cooperative multilateral bargaining game. Policy options
were defined as a combination of the three policy attributes. For each attribute there were five
choices or levels .i.e., five methods, five cost alternatives and five payment policies, which
produced a potential for 125 different policies (53 = 125). Policy evaluation criteria were based
on the utility of (i.e., relative preference for) each policy on the part of each of the game
participants. Utility was also expected to vary not only by group but also according to the level
of knowledge about ecological risks and the likely regional impacts of environmental policies.
6-35
-------
To develop this game, it was necessary to conduct a survey of preferences in Nebraska, Colorado
and Wyoming. The following subsections will discuss, respectively, the survey approach, the
mathematical definition of Model II, and the results of Model II simulations.
6,3.2.1 Household survey of environmental preferences
In November 2000, a total of 4,150 households in Colorado, Nebraska and Wyoming
were randomly selected from lists compiled by Experian (Costa Mesa, CA), a private company
specializing in the compilation of mailing lists. Survey procedures consisted of a first mailing,
followed by a reminder postcard about 10 days later; then those who had not responded within
10 days following the postcard were sent a second copy of the survey.
The survey consisted of four parts. In Parts 1-3, respondents were posed a series of
statements and asked to indicate whether they agreed or disagreed (Parts 1 and 2) or opposed or
supported (Part 3) each statement, on a five-point scale.51 Part 1 assessed general attitudes
regarding water and threatened and endangered species policy in the three Platte River states, but
because these responses did not figure directly in model construction they are not discussed in
detail here. Part 2 examined technical beliefs and the responses were used to assess the effect of
respondent level of knowledge on policy preferences. Part 3 examined policy attributes and
options, and these responses were used to compute respondent and interest group preferences for
various policy attributes. Part 4 asked questions about demographics, and this information was
used to identify respondents with particular bargaining groups (state of residence, and
agricultural or environmental interest group) to be represented in the model.
6.3.2.1.1 Level of knowledge
The 10 statements posed to respondents in Part 2 (Table 6-7) were similar in form to risk
hypotheses, which postulate a causal relationship between a source or a stressor and an endpoint.
6-36
-------
U)
TABLE 6-7
Statements used in the household preferences survey to assess respondent level of knowledge; answers regarded by researchers as correct;
and basis. Respondents were asked to rate agreement/disagreement on a five point scale.
Technical statement appearing in Part 2 of household preference survey
a. Maintaining a wider Platte River channel is not necessary for sustaining a
large and healthy Sandhill Crane population.
b. Increased stream flow will help maintain a wide Platte River channel for
use by cranes and other wildlife.
c. Increased wet meadow acreage is needed to meet the food needs of cranes
and other wildlife in the Central Platte Valley.
d. Increased instream flows would significantly increase the quantity and
quality of wet meadows.
e. The changes in regional income and employment that result from
reallocating up to 420,000 acre-feet of water from agriculture to
endangered species are likely to be so small that they will go unnoticed by
most of the people living in the Platte Valley region.
f. Policies to maintain or increase the current flows in the Platte River will
lead to increased water costs for people living in communities located near
the river.
g. Ground water irrigation has lowered the water table in some parts of the
Central Platte Valley,
h. Ground water irrigation has adversely affected wet meadows in some parts
of the Central Platte Valley.
i. Improved habitat will result in an increased number of Sandhill Cranes using
the Platte River.
j. An increased number of Sandhill Cranes will result in increased tourism in
the Central Platte region.
Correct answer
False
True
True
True
unknown
unknown
True
True
unknown
unknown
Basis for statement/answer (and relationship to risk hypotheses as
numbered in Table 6.1)
Cadmus Group;52 Currier & Ziewitz29 (Risk Hypotheses 1 & 7)
Sidle et al..;15 McDonald & Sidle, 62 (Risk Hypotheses 1 & 7)
Cadmus Group;52 Faanes & Le Valley;24 Currier & Ziewitz29 (Risk
Hypotheses 9 & 1 1)
Hurr;63 The Ground-water Atlas of Nebraska64 (Risk Hypotheses 5 &
11)
The Ground-water Atlas of Nebraska
The Ground-water Atlas of Nebraska64
-------
Whereas risk hypotheses generally refer to existing relationships, however, these statements
tended to be in the form of inferences about the future, to emphasize their relevance to policy.
Six of the 10 statements are regarded to have correct answers; the other four were of interest
because they are often claimed, but their veracity is uncertain. Seven of the 10 statements
pertained to ecological endpoints, including the shallow water table, wet meadows, cranes and
other wildlife. These statements roughly corresponded to several of the risk hypotheses (Tables
6-5 and 6-7).a Three of the seven ecological statements dealing with the habitat needs of cranes
were based on the expert opinion of the researchers. A simple sum of responses to the six
verifiable statements constituted the knowledge index, KL, in Model II, after appropriate
transformations so that a higher value meant more knowledge in all cases.
6.3.2.1.2 Utility of policy attributes
In Part 3, each of the three policy attributes, Method, Cost and Who pays, was described,
and five different levels were defined for each (Table 6-8). Respondents were asked to rate their
support of each of these 15 attribute levels individually. Next, seven policy options (each
consisting of one Method level, one Cost level and one Who pays level) were selected out of the
total of 125 possible combinations that would capture the range of potential responses over each
attribute. Utility ratings for these options were used to derive attribute weights in Model II as
further described below. These attribute weights were multiplied times utility scores for each
attribute and summed across the three attributes that define a policy to determine the utility
scores for all 125 policy options.
3 The fact that many of those hypotheses were not evaluated in the W-ERA does not mean that these statements are
not scientifically supported; in many cases the hypothesis is regarded as supported but the underlying relationship
needs to be better quantified.
6-38
-------
TABLE 6-8
Descriptions of the three policy attributes and their respective levels, a-e, that were evaluated in
part 3 of the household preferences survey
Method:
Five different methods for meeting threatened and endangered species needs on the central Platte are described
below.
a. Meet all endangered species needs using least cost methods of water conservation, water reallocation and
riparian land management, even if this means purchasing or leasing substantial quantities of water from
agriculture.
b. Meet all endangered species needs using a combination of water conservation, water reallocation and
riparian land management programs, but minimize the purchase or leasing of water from agriculture, even
if this increases the cost of meeting these needs.
c. Meet as many endangered species needs as possible using riparian land management and water
conservation programs to provide for endangered species, but do not purchase or lease any additional
water from agriculture, even if this means that the continued existence of the species involved may be at
risk.
d. Use a combination of water conservation, water reallocation and riparian land management implemented
on a trial basis over several years to make certain that the program is necessary and effective before
making large public investments, even if this means there is a potential for continued risk to threatened
and endangered species.
e. Invest in all endangered species protection methods as long as the economic benefits from such
investments are greater than the costs, even if this means continued risk to threatened or endangered
species.
Cost:
To provide for threatened and endangered species on the Platte River, the cost to federal taxpayers throughout the
U.S. and state taxpayers in Colorado, Nebraska and Wyoming could range from zero to $40,000,000 per year. The
amount will depend on what priority we choose to attach to species protection; on the level of risk to species
extinction that we choose to accept; and on the species protection methods that we choose to use. Five different
investment policies for meeting threatened and endangered species needs on the central Platte are described below.
a. Invest nothing to protect Whooping Cranes, Least Tems and Piping Plovers.
b. Invest whatever the U.S. Fish and Wildlife Service (USFWS) says is needed for the species to return to
non-threatened status (currently estimated to cost as much as $40,000,000 per year).
c. Invest about 25 percent of what the USFWS says is needed, or $10,000,000 per year.
d. Invest about 50 percent of what the USFWS says is needed, or $20,000,000 per year.
e. Invest about 75 percent of what the USFWS says is needed, or $30,000,000 per year.
Who
pays:
Another important policy dimension concerns the question of who should pay for species protection. Should it be
the federal government, the states involved in using the resources, private environmental interests, or some
combination? The following five potential policies reflect these choices.
a. All costs paid by the federal government.
b. Federal government pays 50 percent and private environmental interests pay the remaining 50 percent.
c. Federal government pays 50 percent and the states of Colorado, Nebraska and Wyoming pay equal shares
of the remaining 50 percent.
d. Federal government pays 50 percent and the states of Colorado, Nebraska and Wyoming pay the
remaining 50 percent in proportion to the amount of Platte River water consumed in each state (Colorado
20%, Nebraska 20% and Wyoming 10%).
e. Federal government pays one-third, private environmental interests pay one-third and the states of
Colorado, Nebraska and Wyoming split the remaining one-third in proportion to the amount of Platte
River water consumed in each state (Colorado 13%, Nebraska 13% and Wyoming 7%).
6-39
-------
The utility of a given environmental policy, within a particular interest group, was
defined as the adjusted sum of preference scores for the attributes of that policy, as follows:
Uij-WnMjj + WeQj + WoPy + KAFy ' 6-3
where:
Uy = utility or preference score for interest group i, policy optionj;
My = attribute score by interest group / for Method, policy j, on a 1 to 5 scale;
Cij = attribute score by interest group i for Cost, policy/, on a 1 to 5 scale;
Pij = attribute score by interest group / for Who pays, policy/, on a 1 to 5 scale;
KAFy = knowledge adjustment factor for interest group z, policyy', as described
below;
and Wn, W,2, and WB are attribute weights. The knowledge adjustment factor (KAF) was
defined as the difference between the mean Ujj for those in interest group i whose knowledge
level KLj, as defined in Section 6.3.2.1.1, was one standard deviation or more above the mean
and the mean Uy for the entire interest group /'. However, KAF was set to zero unless the
participants in the game chose to invest in education as one method of reaching agreement, or
chose to ignore the preferences of those in each interest group who were not technically
knowledgeable.
The attribute weights Wn, Wi2 and Wj3 would be unnecessary if Method, Cost and Who
pays were of equal importance to respondents within a given interest group. If this were the
case, then the overall utility Uy of a policy option (after adjusting to equivalent scales) would be
similar whether it was derived by summing a group's mean utility scores for the individual
attribute levels that composed the policy or using that group's utility scores for the policy
evaluated as a whole. Because this was not the case, raw attribute weights Bj, 82 and 63 were
6-40
-------
determined for each interest group i by regressing raw utilities RUj for the seven whole policies
over the scores of the three individual attributes to obtain the following equation for each group:
RUi = Bio + BjiMj + Bi2Ci + Bi3Pj + 3 6-4
where M, C and P are the 1 to 5 scores for the three policy attributes and E is an error term. The
regression coefficients were then normalized across attributes to get a total value of 1.0 by
dividing each non-normalized "Bj" value by the quantity (Bn + Bi2 + Bj3), such that for each
group the normalized weights become:
Wii+Wj2 + Wi3 =1.0 6-5
These normalized weights were then used to adjust the individual attribute scores for all 125
policy alternatives as shown in Equation 6-3 .a
6.3.2.2 Bargaining theory and model solutions
The previous subsection defined utility for each policy by bargaining group. Here the
problem of combining those utilities to identify the most globally preferred policies is addressed.
The primary objective of the bargaining process is to find the policy option, defined as a
combination of policy attributes, that maximizes total utility and is acceptable to all groups. In
the bargaining literature and the broader literature of social and public choice, certain solution
concepts seem to prevail. This section will introduce three of the most commonly used solution
concepts for the bargaining model at hand: the utilitarian, Nash and egalitarian solutions. Each
of these solutions will later be applied to the data obtained from the survey to determine if there
are policy options that emerge repeatedly. An option chosen by different bargaining processes,
a The concept of utility as used here is simply a preference rating. It depends on how important the consequences of
a policy choice are to the respondent and also on what he or she believes the consequences will be. Knowledge can
influence utility by changing the respondents* beliefs regarding consequences.
6-41
-------
which represent different social judgments, is most likely to be the policy option that would
emerge from a real bargaining game. If, on the other hand, the policy options chosen by
different bargaining solutions are very different, then one has to investigate the conditions of the
bargaining process and the background for the social judgment much more carefully. If there is
no attribute-level combination that is minimally acceptable to all groups, the players have four
options: 1) negotiate a lower level of minimally acceptable utility; 2) change the water supply
costs by negotiating a reduction in the political compensation factor in Model I; 3) change the
preference functions of participants by providing improved biological and/or economic
information; or 4) declare an infeasible solution.
Let X denote the set of available alternatives. In our case, X equals the set of 125 policy
options that could be chosen. Let N denote the set of agents. Later on, three different sets of
agents will be considered:
N={ Agricultural Interest, Environmental Interest) ^{AgjErj}, and
N={AgColorado, AgNebraska, AgWyoming, EnColorado, EnNebraska, En Wyoming}
={AgC, AgN, AgW, EnC, EnN, EnW}.
To model the theory applied to these agents, a generic set of agents N= { 1 , . . . ,n} and a
generic agent i are denoted. Similarly, there are a generic set of alternatives X and generic
alternatives x and y. Next, it is assumed that each agent associates a cardinal utility Uj (x) with
each policy option x, estimated as u,-j in Equation 6-3. (Alternatively, the ordinal ranks of
alternatives are taken as utility information, ignoring intensities of utility across alternatives and
across agents.) Since now each policy option x induces a vector (uj (x),. . ., un (x) ), the decision
of choosing a policy option boils down to deciding which vector of utilities is acceptable for all
6-42
-------
agents. In order to preserve efficiency of bargaining outcomes, only bargaining solutions that are
Pareto efficient are considered; that is, for any policy option chosen by the bargaining
solution, there does not exist another policy option such that all agents are weakly better off and
at least one agent is strictly better off (see Section 2.2.1).
6.3.2.2.1 Utilitarian solution
The utilitarian solution is the policy option which maximizes the sum of all agents'
utilities and can be depicted as
max
*eA r f
1=1 6-6
where: u, is the cardinal or ordinal utility for agent i, for some vector of policy options x.
6.3.2.2.2 Nash solution
The Nash solution is the policy option which maximizes the product of all agents*
utilities and can be depicted as
max u, (x)
X
-------
where: Uj is the cardinal or ordinal utility for agent i, and Uj is that for all other agents, for some
vector of policy attributes x.
In terms of social policy, the utilitarian solution represents that set of decision rales
where there is no concern for the relative utility of agents. Any gain in total utility is considered
an improvement irrespective of how the total is distributed across agents. The Nash solution
essentially incorporates the concept of diminishing marginal utility, while the egalitarian solution
takes the potential concern for equity or fairness one step further. Let us demonstrate with a
simple example. Suppose there are two agents and three policy options. Option A produces 1
unit of utility for agent 1 and 10 units for agent 2; option B produces 4 units for each agent; and
option C produces 6 units for agent 1 and 3 units for agent 2. In this case, the utilitarian solution
would favor option A (1+10 > 6+3 > 4+4), whereas the Nash solution would favor option C (6 x
3>4x4>lxlO), and the egalitarian solution would favor option B (4-4 < 6-3 < 10-1). The
respective solutions can also be referred to as the sum, product and equity solutions.
6,3.2.3 Survey results
This section summarizes the survey findings with an emphasis on their application to
model calculations; tabularized responses to survey questions are presented in Supalla et al. A
total of 1,187 useable surveys were returned, for an overall response rate of 26 percent. The
response rate for Nebraska residents was highest at 32 percent, followed by Wyoming at 24
percent and Colorado at 22 percent. These relatively low response rates suggest a likelihood of
response bias, although there were no particular indications of response biases within or between
interest groups. One would generally expect, however, that those who were better educated and
most interested in the problem would be the most likely to respond.
6-44
-------
6.3.2.3.1 Demographics
Demographic responses showed that respondents were in fact somewhat older and better
educated than the general population. The average age of respondents was 53, over 38 percent
had a Bachelor's degree or better education, and less than 5 percent had not graduated from high
school. The age distribution was essentially the same for each state, but the Colorado
respondents were significantly better educated than those from Nebraska or Wyoming.
Approximately 14 percent of respondents were farmers or ranchers, over 18 percent were self
employed in other ways, about 13 percent worked for state or local government and the
remainder were either employed by other types of organizations or retired. The employment
distribution was very similar for each state, except for agriculture. Very few of the Colorado
respondents were farmers or ranchers (7%), compared to 12 percent for Wyoming and 19 percent
for Nebraska. Differences in the proportion of state respondents who were farmers or ranchers
reflect in part, actual differences in the proportion of each state's population that is engaged in
agriculture, but these differences may also reflect a self-selection bias. Farmers in Nebraska,
especially central Platte irrigators, are more likely to be directly impacted by central Platte
programs and, thus, more likely to take the time to respond to the survey.
A relatively large number of respondents were affiliated with agricultural or
environmental interest groups. In total, about 17 percent of respondents were affiliated with
agricultural groups and 31 percent with environmental groups. The three states were quite
similar, except that only 8 percent of Colorado respondents were affiliated with agriculture, and
only 19 percent of Nebraska respondents were affiliated with environmental groups compared to
48 percent in Colorado. This suggests that interest groups may be a major source of information
6-45
-------
on central Platte issues for that part of the population that was interested enough in the issues to
respond to the survey.
6.3.2.3.2 Attitudes regarding environmental policy
About one-third of respondents agreed that society should ensure species protection
regardless of cost. There was very strong support for having the federal government and private
environmental organizations pay for species protection rather than the states. Two-thirds of
respondents agreed that the federal government rather than the states should pay for most of the
cost and 80 percent agreed that private environmental organizations should also contribute.
There was also strong support for the idea that the economic base provided by irrigated
agriculture should be protected. Over 70 percent would be willing to pay more for species
protection to protect the economic base, and over 50 percent were willing to protect the
economic base even if it meant increased risk to endangered species. Surprisingly, 55 percent
would support paving twice as much for environmental water as an alternative to reducing
irrigation.
There were few significant attitudinal differences between the states. Colorado residents
were much more likely than Wyoming or Nebraska residents to agree that society should ensure
environmental integrity regardless of the cost. Wyoming respondents were not supportive of
each state supplying one-third of the environmental water, while Nebraska respondents
supported this alternative. This probably reflects a concern among Nebraska residents that the
state may be asked to provide more than a one-third share and a belief by Wyoming residents
that their equitable share is less than one-third.
6-46
-------
6.3.2.3.3 Technical beliefs regarding central Platte River
environmental problems
There was considerable disagreement and/or lack of knowledge concerning physical
environmental attributes. Only 24 percent of Colorado residents, 29 percent of Wyoming
residents and 41 percent of Nebraska residents were aware, defined as agreed or strongly agreed,
that a wide river channel is important to cranes. Less than 50 percent of the respondents in all
states recognized that increased stream flow would help maintain a wide river channel. There
was greater recognition of the environmental importance of wet meadows and of the link
between groundwater irrigation and wet meadow production, but the number of correct
responses was still below 50 percent in nearly all cases. Respondents in all states also expressed
considerable uncertainty with respect to the economic effects from management alternatives.
Nearly an equal number of people agreed as disagreed with statements concerning the effects of
changes in the amount of irrigation or tourism on the regional economy.
Differences between the states may suggest some reasons for the technical beliefs that are
held. Over 21 percent of Nebraska respondents disagreed with the statement that groundwater
irrigation adversely affects wet meadows, compared to 11 percent for Colorado and 12 percent
for Wyoming. Similarly, 22 percent of Nebraska respondents disagree with the contention that
improved habitat will increase the number of cranes, compared to 10 and 13 percent for
Colorado and Wyoming, respectively. These differences suggest that there may be an inclination
on the part of some respondents to deny recognition of technical relationships that do not support
their policy position and/or that imply some responsibility for an adverse impact. The Nebraska
sample contains a relatively large proportion of irrigators, many of whom may be reluctant to
accept scientific claims about how their activities may affect the Middle Platte ecosystem.
6-47
-------
6.3.2.3.4 Level of support for policy attributes
Data on the level of public support for each of five different levels of each of three policy
attributes were used in game models to find bargained policy solutions. Preferences were
analyzed by state and for each of two interest-related bargaining groups, agricultural and
environmental (Table 6-9). Respondents were classified as agricultural if they indicated that
they were self-employed as a farmer or rancher, employed by an agricultural interest group or
affiliated with the Farm Bureau, the Farmers Union or an irrigation district. Respondents were
classified as environmental if they indicated that they were employed by an environmental
interest group; affiliated with the Sierra Club, The Nature Conservancy or the Audubon Society;
or agreed or strongly agreed with the statement that "Society should ensure that the needs of
threatened and endangered species are met regardless of economic cost." Respondents who
qualified as agricultural based on employment or interest group affiliation, but who also agreed
that society should meet the needs of endangered species irrespective of economic cost, were
considered as both agricultural and environmental. Those respondents who either could not be
classified as exclusively agricultural nor exclusively environmental were classed as "other" and
were included in state totals but were not analyzed as a separate bargaining group.
For the Method attribute, the level receiving the strongest support from all states as
measured by the average score for all residents was adaptive management (Appendix 6-A).
Colorado's second best choice was to meet all needs while minimizing water, but the second best
option preferred by Nebraska and Wyoming respondents was to do the best possible job of
meeting endangered species needs with no reallocation of irrigation water. Agricultural interests
in all states strongly preferred either an adaptive management approach or a program that
produced as much endangered species protection as possible without reallocating any water from
6-48
-------
TABLE 6-9
Respondent classification into bargaining groups, by state. Based on type of employment,
interest-group affiliation, and attitude regarding endangered species, a respondent could be
classified as either agriculture, environmental, both, or neither.
Bargaining Group
Agriculture
Environmental
Other
Total
Colorado
Nebraska
Wyoming
All States
Numbers of Respondents
24
143
132
299
105
86
257
"448
55
110
166
331
184
339
555
1,078
6-49
-------
agriculture (Appendix 6-A). They were most strongly opposed to the idea of meeting all needs
irrespective of the costs. Environmental interests preferred to meet all needs, although they also
expressed considerable support for an adaptive management approach.
Expressed support for different levels of investment (Cost attribute) was somewhat
mixed, but the strongest support in all states was for a $10M annual investment, which is about
25 percent of what many observers believe it would take to fully implement USFWS
recommendations. However, 32 percent of all Colorado respondents expressed strong support
for investing whatever it took to meet USFWS recommendations. Agricultural interests
preferred to invest nothing, or perhaps $10M per year, but there was very little support among
agriculturalists in all states for spending more than $10M per year.
The payment policy results (Who pays attribute) were especially interesting. All states
preferred that private environmental groups pay a significant part of the cost, which is contrary to
current proposals to address the problem. The reasons for preferring private contributions are
unknown, but the leading hypothesis is that respondents believe those who get the most utility
from environmental improvements should also pay the most. The first choice of all states was a
payment policy consisting of one-third federal, one-third private and one-third state, with the
state one-third being distributed between the three states in proportion to current water use.
Wyoming respondents objected strongly to each state paying a equal share of the aggregate state
share, but there were no other significant differences between the states. The strongest support
for some private contribution to the cost of meeting endangered species needs came from
agricultural interests, but surprisingly there was also substantial support from environmental
interests for requiring some private cost sharing. This may reflect a belief that the benefits from
6-50
-------
endangered species protection accrue disproportionately to environmental interests and, thus, the
entire burden should not fall to general taxpayers.
6.3.2.4 Model II results
6.3.2.4.1 Weights for policy attributes
Responses to a sampling of 7 of the 125 policies were used as described in Equations 6-4
and 6-5 to derive attribute weights for each of the three states and for agricultural and
environmental bargaining groups within each state (Appendix 6-A). Except for the
environmental interest group in Wyoming, the most heavily weighted policy attribute was
payment policy and the least important was the method of meeting endangered species needs.
Environmental interests generally placed more weight on method and less on payment policy,
compared to agricultural interest groups.
6.3.2.4.2 Policy preferences
Weighted utility scores were computed for all 125 policy options for each bargaining
group using Equation 6-3; these are cardinal utilities (not presented). To facilitate comparisons,
utility scores for each group were ranked from 1 to 125, where the best option is ranked 125 and
the poorest has a ranked score of one; these are ordinal utilities. The full array of 125 policy
options was then reduced to 17 by eliminating those which were not Pareto efficient (Tables
6-10, 6-11 and 6-12). An option was considered Pareto inefficient if it was possible to improve
the level of total utility across groups without making one or more groups worse off. The level
of support for the more efficient options was considered in more detail.
Surprisingly, the highest ranked option in each state was the same, option N, which
consists of an adaptive management program using both riparian land management and improved
stream flow to protect endangered species, at an investment level of $10M per year, with the
6-51
-------
TABLE 6-10
Definition of Pareto efficient policy options: attribute levels corresponding to each policy.
Policy
Option
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
Q
R
Attribute Level3
Method
d. Adaptive
Management
d. Adaptive
Management
d. Adaptive
Management
a. All Needs, Least
Cost
d. Adaptive
Management
a. All Needs, Least
Cost
e. Benefit-Cost
Approach
d. Adaptive
Management
a. All Needs, Least
Cost
b. All Needs,
Minimum Water
d. Adaptive
Management
a. All Needs, Least
Cost
b. All Needs,
Minimum Water
d. Adaptive
Management
a. All Needs, Least
Cost
b. All Needs,
Minimum Water
b. All Needs,
Minimum Water
Cost
c. Invest $10M,
25% of Need
a. Invest Nothing
b. Invest $40M,
perUSFWS
c. Invest $10M,
25% of Need
c. Invest $10M,
25% of Need
d. Invest $20M,
50% of Need
c. Invest $10M,
25% of Need
a. Invest Nothing
b. Invest $40M,
perUSFWS
b. Invest $40M,
perUSFWS
b. Invest $40M,
perUSFWS
c. Invest $10M,
25% of Need
c. Invest $10M,
25% of Need
c. Invest $10M,
25% of Need
d. Invest $20M,
50% of Need
d. Invest $20M,
50% of Need
e. Invest $30M,
75% of Need
Who pays
a. All Costs Paid by Feds
b. Feds 50%, Private 50%
b. Feds 50%, Private 50%
b. Feds 50%, Private 50%
b. Feds 50%, Private 50%
b. Feds 50%, Private 50%
d. Feds 50%, States 50%
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
' A full description of each policy attribute and level is found in Table 6-8,
6-52
-------
TABLE 6-11
Pareto efficient policy preferences,3 by state
Policy Option
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
0
R
Ranked Utility Scores
Colorado
78
80
93
70
98
65
69
117
113
119
123
116
121
125
1.15
120
118
Nebraska
104
89
84
94
117
69
82
119
92
99
114
120
122
125
105
113
102
Wyoming
89
110
103
91
119
77
60
123
92
104
121
114
120
125
96
113
101
8 Policy options are ranked from 1 to 125 with 125 being the highest or best option.
See Table 6-10 for a description of each policy option.
6-53
-------
TABLE 6- 12
Pareto efficient policy preferences,2 by bargaining group and state
Policy
Option11
A
B
C
D
E
F
G
H
1
J
K
L
M
N
P
0
R
Colorado
Ag
Utility
Rank
115
116
53
113
124
63
90
117
33
44
52
114
120
125
64
75
58
Envl,
Utility
Rank
42
8
82
46
43
62
58
54
123
125
122
109
117
105
115
121
119
Nebraska
Ag
Utility
Rank
94
125
106
114
122
103
47
112
51
60
72
84
93
108
71
77
61
Envl.
Utility
Rank
76
28
55
39
59
30
83
113
115
120
122
117
124
125
114
118
121
Wyoming
Ag
Utility
Rank
111
125
106
113
123
93
46
107
42
47
65
73
82
97
51
60
49
Envl.
Utility
Rank
46
27
73
68
55
69
85
94
125
124
123
121
119
115
122
120
117
All
Ag
Utility
Rank
105
125
98
114
123
92
50
115
45
52
72
84
94
110
65
74
56
All Envl.
Utility
Rank
42
9
71
47
45
53
79
91
124
125
123
115
120
114
117
121
122
' Policy options are ranked from 1 to 125 with 125 being the highest or best option.
' See Table 6-10 for a description of each policy option.
6-54
-------
federal government paying one-third, the states one-third and private environmental groups.one-
third (Table 6-11). Under this option the states' share is split proportionally between the states
according to historical water use. The lowest ranked options in all states were generally those
which called for investing nothing.
Policy preferences of interest groups within a state were much more varied (Table 6-12).
The first choice of agricultural interests in both Nebraska and Wyoming was option B, which
consists of adaptive management at a very low level of investment, with all costs paid by the
federal government and private environmental interests. Agricultural interests in Colorado
preferred option N, which is surprisingly consistent with the preferences of all citizens in each of
the three states. Environmental interests in Colorado and Wyoming preferred meeting all
endangered species needs, while reallocating as little water as possible, with expenditures up to
$40M per year, with costs shared equally by the federal government, the states and private
interests.
6.3.2.4.3 Bargaining solutions
The bargaining challenge, therefore, lies in finding a solution to differences of opinion
within, rather than between, states. The magnitude of this challenge can be seen by analyzing
how acceptable a given group's preferred option is to competing bargaining groups (Table 6-13).
For example, examining the seventh row of Table 6-13, all agriculture prefers an adaptive
management plan with minimal water reallocation and minimal investment, with 50 percent of
the costs paid by private environmental groups and 50 percent by the federal government (option
B). Moving to the end of the seventh row, environmental interests aggregated across states rank
option B as their ninth poorest option, which places it in the bottom 10 percent of the 125
choices being considered. Environmental interests (last row) prefer option J, which would meet
6-55
-------
TABLE 6- 13
Comparison of preferred policy options between competing interest groups
j
Group
CO
NE
WY
COAg
NEAg
WYAg
AllAg
CO Envl
NE Envl
WY Envl
All Envl
Preferred
Option
N
N
N
N ,
B
B
B
J
N
I
J
CO
NE
WY
CO
Ag
NE
Ag
WY
Ag
All
Ag
CO
Envl
NE
Envl
WY
Envl
All
Envl
Rank of Preferred Option3
125
125
125
125
80
80
80
119
125
113
123
125
125
125
125
89
89
89
99
125
92
40
125
125
125
125
110
110
no
104
125
92
108
125
125
125
125
116
116
116
44
125
33
12
108
108
108
108
125
125
125
60
108
51
49
97
97
97
97
125
125
125
47
97
42
104
110
110
110
110
125
125
125
52
110
45
3
105
105
105
105
8
8
8
125
105
123
125
125
125
125
125
28
28
28
120
125
115
125
115
115
115
115
27
27
27
124
115
125
92
114
114
114
114
9
9
9
125
114'
124
125
' Policy options are ranked from 1 to 125 with 125 being the highest or best option
-------
all endangered species needs at a cost of up to $40M per year, with costs shared equally between
the federal government, the states and private environmental interests. Agricultural interests
rank
option J as their third poorest option. These comparisons suggest that a bargaining process is
needed to find an acceptable middle ground that lies somewhere between, at one extreme, a
program that meets all endangered species needs (as determined by USFWS), involves a major
reallocation of water from agriculture, and costs up to $40M per year; and at the other extreme, a
program that reduces the reallocation of water to an absolute minimum, costs much less, but
exposes endangered species to significant risk.
Three solutions to a multilateral bargaining game were computed in a search for the
policy options most likely to be acceptable to all of the principal interest groups (Table 6-14).
Policy N is both the utilitarian and Nash solution (Equations 6-6 and 6-7), whether using cardinal
or ordinal utility. However, the egalitarian solution (Equation 6-8) is policy option D when
using cardinal utility and option A when using ordinal utility.
These results suggest that if the bargaining agents were not concerned about equity
between groups they would adopt policy N, which is an adaptive management approach meeting
only some of the endangered species needs, spending $10M per year, with the costs split evenly
between the federal government, the states and private environmental groups. However, if
equity was more of a concern, the solution would involve a similar approach with about the same
level of investment, but with no state contribution to program costs.
If policy option N is selected, environmental groups are likely to be reasonably satisfied,
because a reasonable amount of endangered species protection will be provided and the costs
will be widely shared. However, at least part of the agricultural community is likely to be
6-57
-------
TABLE 6-14
Results of bargaining models, all bargaining groups
Pareto
Efficient
Options
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
0
R
Cardinal Utility
Utilitarian
Nash
Egalitarian
Ordinal Utility
Utilitarian
Nash
Egalitarian
Rank of Policy Option a
101
102
87
105
1.16
81
72
122
103
111
115
123
124
125
110
117
109
104
103
87
109
117
82
69
122
91
105
112
123
124
125
108
115
106
123
87
73
125
118
90
63
95
8
9
20
49
32
70
24
29
19
98
85
92
102
110
82
76
121
99
107
115
123
124
125
112
117
109
95
56
98
96
105
80
81
119
90
103
114
123
124
125
111
117
109
121
106
105
u_ 12°
94
40
88
25
2
98
11
44
1
91
4
78
101
' Options are marked from 1 to 125, with 125 being most preferred.
6-58
-------
uncomfortable with a program that reallocates water away from agriculture in ways they believe
may not be justified on a cost-benefit basis, especially when the states are paying a significant
share of the cost.
6.3,2.4.4 Potential impact of education on policy preferences
An important policy issue concerns the extent to which education might reduce the level
of disagreement between bargaining groups. Two questions would need to be answered. First,
does the tendency for groups to disagree appear to be related to the level of technical knowledge
within the groups? If the answer is yes, then would education improve the level of technical
knowledge and the level of agreement? While the second question was beyond the scope of the
current project, the first question was analyzed by comparing the policy preferences of more and
less knowledgeable survey respondents.3
Knowledgeable respondents were defined as those whose knowledge index score, as
defined in Section 6.3.2.1.1, was at least one standard deviation above the mean in each state.
Average utility scores for the knowledgeable and non-knowledgeable classes were computed and
compared for the 17 Pareto efficient policy options. An aggressive education program was
arbitrarily assumed to be able to change the level of support for the Pareto efficient policies by
non-knowledgeable citizens by an amount equal to one-half the average difference between the
knowledgeable and non-knowledgeable classes. Hence, the appropriate adjustments were made
to the non-knowledgeable scores and a new interest group average calculated for each Pareto
efficient policy option. Rank orderings of the 17 options with and without the assumed education
a The effect of knowledge on policy preferences was also addressed with a logit model which analyzed the effect of
knowledge on the probability that an individual would support environmentally intense policies. This analysis found
a strong statistical relationship between knowledge and level of policy support.
6-59
-------
effect were then compared to determine if there was any appreciable effect on what option was
most preferred by each interest group and, most importantly, to determine if the knowledge
effect brought the interest groups closer to an agreement on the best policy option. 5
In all states the effect of improved knowledge was to bring the agricultural and
environmental interest groups closer to agreement. In Nebraska, the effect was primarily on the
agricultural interest group. Nebraska agriculture's first choice went from option B, which calls
for investing nothing in endangered species protection, to option N, which was the first choice of
Nebraska environmental interests before the effect of improved knowledge. With improved
knowledge the first choice of Nebraska environmental interests became option J, which is similar
to option N, but calls for a higher level of investment. For Wyoming, the effect of improved
knowledge was also to make environmentally strong options more acceptable to agricultural
interests. Both Wyoming agricultural and Wyoming environmental interests preferred option I
after the knowledge effect was imposed, whereas previously, Wyoming agricultural interests
preferred a much lower level of investment in endangered species protection. For Colorado,
there was no significant knowledge effect on environmental interests, but agricultural
preferences changed from preferring adaptive management option N to preferring to meet all
needs, option L.
6.3.2.4.5 Policy implications of Model II
The results from Model II suggest that the most important differences of opinion
regarding central Platte management policies exist between agricultural and environmental
interest groups within each state, rather than between states. At the aggregate level, all three
states preferred a policy which called for an adaptive management approach that minimized the
reallocation of water from agriculture and involved a modest level of investment, with the costs
6-60
-------
shared equally between the federal government, the states and private environmental interests.
Within Nebraska and Colorado, however, agricultural interests preferred to invest nothing, with
everything paid for by the states and private environmental interests, while environmental
interests preferred a much more aggressive program to ensure endangered species protection,
with costs split evenly between the federal government, the states and private environmental
interests. Colorado agricultural interests were more supportive of environmental objectives, but
still preferred less endangered species protection than did Colorado environmental interests.
An analysis of policy attributes found that the dominant attribute in nearly all cases was
payment policy (i.e., Who pays; see Appendix 6-A, Table 6-A-5). Private environmental
interests showed a surprising willingness to support private contribution to the costs of central
Platte management programs, and agricultural interests were much more willing to endorse a
significant endangered species protection program, if the state cost share was minimized and
there was a substantial private contribution. All interest groups were quite receptive to an
adaptive management approach that is quite similar to the programs now being pursued by the
states and the DOI under the terms of the Cooperative Agreement,
Application of three different sets of bargaining rules all resulted in solutions which
called for an adaptive management approach that minimized the reallocation of water, with an
equal sharing of the costs between federal, state and private entities. The egalitarian solution,
however, suggested that if the agents were more concerned about equity, they should pursue a
somewhat more aggressive program of endangered species protection with less of a state
contribution to the total cost.
An analysis of the impact of technical knowledge on policy preferences found that well
informed people had much stronger environmental preferences compared to those who were less
6-61
-------
well informed. It was found that much of the disagreement between agricultural and
environmental interest groups would cease to exist if both groups had technical beliefs that were
similar to those held by well informed individuals. This finding suggests that ecological risk
information might have a role in changing public opinion, leading to reduced conflict and
perhaps improved resource management. However, there is also a possibility that some
respondents knowingly answered technical questions incorrectly, in cases where an incorrect
answer supported their strongly held values and policy positions. It is also possible that
individuals may reject as biased any new information that did not support such values. Before
definitive conclusions can be drawn, further research is needed regarding the effectiveness of
ecological-risk education in changing technical beliefs and policy preferences.
6.4 DISCUSSION
Chapter 3 put forward a conceptual approach for the integration of ERA and economic
analysis for watershed management (Figure 3-1). In that ideal approach, integration occurs in all
stages of assessment. Because economists' involvement began late in the assessment process for
the central Platte River floodplain, the process depicted in Chapter 3 was not followed in several
respects. That ideal process nonetheless provides a useful framework for evaluating the methods
used and degree of ecological-economic integration achieved in this case study.
6.4.1 Assessment planning and problem formulation
The conceptual approach calls for the interrelated steps of assessment planning and
problem formulation to be carried out in advance of analysis (Figure 3-1). In this case, a formal
planning process that included stakeholders was conducted at the outset of the W-ERA. Planners
discussed watershed values and challenges and crafted a very broad management goal - i.e., "to
protect, maintain, and where feasible restore biodiversity and ecological processes..." - and a list
6-62
-------
of eleven management objectives (Section 6.2.1). The W-ERA assessment team then worked to
distill those objectives and existing knowledge of the watershed into assessment endpoints,
conceptual models and risk hypotheses (Section 6.2.2).
The economic research effort was not yet conceived at this stage, and economists were
not involved in this process. The economic study was initiated later, with a coordination meeting
that had minimal stakeholder representation and occurred after most of the W-ERA work had
already been completed. Therefore, ecological risk assessors did not have the benefit of
considering economic concepts, research approaches or management insights, and while
economists heard a brief report of the W-ERA approach, they did not benefit from a close
collaboration with that effort, nor did they engage a broad range of stakeholder groups in their
work. This limited degree of coordination resulted in a divergence of analytic objectives and
perspectives. The ecological analysis studied habitat requirements of dozens of riparian-
dependent avian species whereas the economic analysis addressed only the needs of endangered
species.
6.4.2 Formulating alternatives, and baseline ecological risk assessment
Whereas ERA alone does not necessarily require the formulation of management
alternatives, economic analysis usually is concerned with alternatives, so their formulation
usually is a condition for integrated study (Figure 3-1). The Platte River W-ERA sought only to
characterize baseline risk, i.e., risks that exist now or are likely to occur if no new management is
undertaken. The risk models that were developed (i.e., models describing floodplain segment
use by sandhill cranes, and meadow or woodland patch use by nesting birds) dealt with a subset
of the ecological assessment endpoints. They are potentially applicable to management
questions but were not developed with a specific decision context or set of alternatives in mind.
6-63
-------
The economic analysis, on the other hand, formulated two sets of management
alternatives. Model II focused on finding a compromise solution from among 125 different
options (later narrowed to 17 Pareto efficient policies) for floodplain management, especially
dealing with instream flow amount and payment. Model I provided a tool, an auction market, for
use by stakeholders in deciding who would provide alternative levels of environmental water.
The economic analysis thus focused directly on resource management choices that were linked to
the dominant issue in the basin, rather than addressing a broader, yet less pragmatically focused,
array of baseline ecological risks. Had the economists been part of the W-ERA planning
discussions, there would have been an opportunity to discuss these alternatives and thus better
harmonize the ecological and economic analyses. Discussion of management alternatives during
assessment planning might also have narrowed the scope of the W-ERA, limiting the number of
management objectives and risk hypotheses, and sharpening its analytic focus.
6.4.3 Analysis and characterization of alternatives, and comparison of alternatives
The analysis and characterization of management alternatives and the comparison of the
alternatives are two closely related steps in the conceptual approach (Figure 3-1). Each
management alternative is to be examined in the light of both ecological risks and economic
outcomes and, as applicable, other analyses (e.g., health or quality of life). Diagrammatic
examples of a variety of approaches to these two steps were given in Figures 3-2, 3-3 and 3-4.
The approach employed in this case study is illustrated in Figure 6-3, which is a
modification of Figure 3-3. The likely ecological and economic outcomes of various watershed
management policy attributes were described in a survey of preferences, and survey results were
used to evaluate specific policies (i.e., attribute combinations).
6-64
-------
ANALYSIS & CHARACTERIZATION OF ALTERNATIVES
Ecological Risk
;.!.:. i IpoJFii
c ii?"irv-r Vjfl
'
Qualitatively
describe
other changes
Express primary changes
in common language
Economics
Quantify financial costs
and market-based
economic effects
Qualitatively analyze
equity, economic impact
Express equity effects, impacts
in common language
J
"Survey of preferences"
Evaluate policy options,
find bargaining solutions
Other Analyses
:.r'.l;>. ',i.
'"orr]1- '..'i^r'.
fea^r'lc
' '
-I-:- ,
Jlher ctia'Fi
COMPARISON OF ALTERNATIVES
.... chc
DP I .:. V-.'
FIGURE 6-3
Techniques used for analysis, characterization and comparison of management alternatives in the
central Platte River floodplain, as compared to the example shown in Figure 3-3. White boxes
and bold type show features included in this analysis.
6-65
-------
The ecological point of departure for the economic study was a determination by USFWS
that a given increment of instream flow and restoration of wet meadow acreage are needed to
ensure protection of endangered species.3 This level of provision was described qualitatively in
the survey as "meeting the needs" of endangered species; lesser levels of provision were
described as placing the species "at risk" (Table 6-8). Annual costs to fund the USFWS program
were described in dollar terms. The market-based economic effects of the program (such as the
impacts of foregoing water diversion or pumping, or removing land from production) were not
estimated or described. However, equity and economic impact concerns were implicit in the
wording of policy options that minimized the purchase of water from agriculture or that
discussed different cost-sharing options.
In Figure 6-3 the term "survey of preferences" is substituted for "stated preference
survey," because the latter usually refers to methods that ask individuals to place a value on
specific changes to the environment, whereas in this case the results will not provide estimates of
value either directly or indirectly. Analysis of survey results yielded policy-specific estimates of
utility for each of several bargaining groups. A subsequent step used the utilitarian, Nash or
egalitarian approaches to rank-order the policies. Estimates of the net social benefit of policies
could not be derived, in part because market-based economic effects of the policy options were
not determined, but also because the survey of preferences did not estimate willingness to pay.
Ecological economics stresses that economic analyses should account for the biophysical
constraints that exist in the ecological systems that support all human activity (see Section 2.2.6).
The W-ERA for the Platte River did not formulate or evaluate any management alternatives. It
is important, therefore, to examine the degree to which the economic models were informed or
constrained by information on ecological risks, hi general, the economic analysis regarded
6-66
-------
ecological risk as technical information which could influence the preferences of stakeholders.
Ecological risk was constraining only to the extent that stakeholders regarded risk reduction an
important objective relative to the trade-offs involved. With this approach no answer is regarded
as scientifically correct; all that science does is provide trade-off and preference information to
facilitate public decision-making. Model I, the auction model, did several things: (1) it provided
a tool for efficiently "negotiating" who will supply a given quantity of water and at what price;
(2) it provided a method of estimating the budgetary supply costs associated with different
quantities of environmental water; and (3) it provided an indication of the price that stakeholders
would pay in the form of welfare and budget costs for using the negotiating efficiencies inherent
in a second-price auction instead of a direct negotiation first price approach. Providing these
functions required no ecological risk information.
Model II used preference information for policy options that ranged from providing
"whatever the USFWS says is needed..." to providing nothing. The Model II bargaining
solutions were based on utility and not constrained by conditions ensuring species' survival,
beyond respondents' preference for doing so. If respondents preferred policies that were lower-
cost or involved less reallocation of water, it is not clear whether they were accepting as valid the
biological opinion of the USFWS and voting against full support for maintaining the species,
whether they did not believe that water reallocation would be helpful to the species; or were
uncertain about key technical relationships and therefore preferred an incremental, try-it-and-see
approach. An analysis of the impact of technical information on policy preferences suggested
that facts were a very important determinant of policy preferences. Policy preferences changed
markedly and the differences between interest groups narrowed substantially if one assumed that
with education the less well informed stakeholders would develop preferences similar to those of
6-67
-------
their better informed colleagues. If this assumption were substantiated, it would raise the
possibility that an effective program of educational outreach, carried out in conjunction with a
bargaining process, could provide an effective biophysical constraint. However, this study did
not investigate the actual effectiveness of education in a situation of longstanding conflict, and
therefore it cannot be concluded that the bargaining approach, per se, is effectively constrained.
It is possible that the process of adaptive implementation, such as that envisioned by the
Cooperative Agreement, would afford constraints ensuring species survival, but much depends
on the view one takes of adaptive implementation as a management and political strategy. If it
serves as a reliable feedback mechanism, whereby stakeholders' preferences are updated by new
information, then biophysical constraints may be effective, even when not explicit in a
preference-based model. An adaptive management approach that is politically feasible may
reach desired ecological goals at a slower pace than some would prefer, but it may still be the
most effective approach if full and immediate implementation is not politically feasible.
6.4.4 Consultation with extended peer community
USEPA's Ecological Risk Assessment Guidelines recommend fully involving
stakeholders in planning but maintaining strict separation of science from policy in subsequent
steps, whereas others have emphasized the limitations of science and the importance of ongoing
consultation, throughout the analysis, with an extended peer community (see Sections 2.1.1.5
and 3.3.5). The W-ERA formally established a stakeholder panel for participation in planning,
but problem formulation was conducted by a more limited technical team. Near the end of
problem formulation, consultations with stakeholders were held and a draft was reviewed, but
subsequent changes made by the technical team alienated at least one stakeholder group. The
economic analysis was not constrained by a formal requirement for stakeholder involvement and
6-68
-------
used more limited and informal mechanisms. Lacking their strong involvement, however, it is
not yet clear whether the parties to the Cooperative Agreement will make use of the game theory
results.
6.4.5 Decisions and adaptive implementation
Even if ecological risk, economics and other information are well integrated and well
tuned to the decision context, it is normal for any high-stakes decision to require negotiation after
the analyses are completed. The game theory models developed here may be well-suited to the
support of an ongoing negotiation because they can respond quickly to changes in negotiating
position and suggest new solutions. The approach may also be useful over a longer period of
adaptive implementation, in which system modification and feedback result in new learning, and
a new set of policy solutions is sought.
Adaptive implementation is important not only for its merit as a management approach
but also as an aid to difficult negotiations. When disagreements about the true behavior of the
system prevent the parties from agreeing on costly remedies, an adaptive approach can present
an attractive compromise in that it holds out the promise of improved knowledge about the
system. But care must be taken to distinguish between a policy that is truly adaptive and one that
is simply incremental. Walters66 argues that incrementalism (making small improvements
without taking large risks) is not effective as an information-generating strategy. "Such policies
result in strongly correlated inputs, and in state variables being correlated with inputs,... so the
effects of each cannot be distinguished." An ideal strategy from an informational standpoint
would consist of repetitive sequences moving from one extreme to the other, each of sufficient
duration to allow observation of responses of key variables. Managers tend to be risk-averse,
however, and under substantial pressure to avoid extremes. An actively adaptive policy,
6-69
-------
therefore, must somehow establish a balance between learning (via policies designed to
maximize probative value) and short-term performance (maintaining the system nearest its status
quo).66
A key question, therefore, about the value of the Cooperative Agreement as an
informative policy is whether the initial increment of 140,000 acre-feet, and evaluation period of
10-13 years, will be sufficient, in light of natural hydrologic variability and the slowness of
successional processes, to induce unambiguous changes in key variables such as area of active
channel. Since only an unambiguous response would be likely to promote agreement about
subsequent actions, the prospects for reducing conflict over the long term through this game
theoretic approach are closely tied to adaptive implementation's effectiveness.
6.5 REFERENCES
1. Johnsgard, P. A., Birds of the Great Plains: Breeding Species and Their Distribution,
University of Nebraska Press, Lincoln, Nebraska, 1979.
2. Sidle, J.G. et al, Aerial thermal infrared imaging of Sandhill Cranes on the Platte River,
Nebraska, Remote Sensing of Environment, 43, 333, 1993.
3. Sidle, J.G. and Faanes, C.A., Platte River ecosystem resources and management, with
emphasis on the Big Bend Reach in Nebraska, U.S.Fish and Wildlife Service, Grand
Island, NE and Northern Prairies Wildlife Research Center, 1997, Available from
http://www.npsc.nbs.gov/resource/othrdata/platte2/platte2.htmfcontents.
6-70
-------
4, FERC, Draft environmental impact statement: Kingsley Dam and North Platte/Keyston
diversion dam projects, FERC/DEIS-0063, Federal Energy Regulatory Commission, Office
of Hydropower Relicensing, Washington, DC, 1992.
5, Eschner, T.R., Hadley, R.F., and Crowley, K.D., Hydrologic and morphologic changes in
channels of the Platte River Basin in Colorado, Wyoming and Nebraska: A historical
perspective, 1277-A, 1983, 1.
6. Junk, W.J., Bayley, P.B., and Sparks, R.E., The flood pulse concept in river-floodplain
systems, in Proceedings of the International Large River Symposium(LARS), Dodge, D. P.
Ed., Canadian Special Publication of Fisheries and Aquatic Sciences, Ottawa, Canada,
1989,110.
7, Sparks, R.E. et al., Disturbance and recovery of large floodplain rivers, Environmental
Management, 14, 699, 1990.
8. Sparks, R.E., Need for ecosystem management of large floodplain rivers and their
floodplains, BioScience, 45, 168, 1995.
9. Currier, P.J., The floodplain vegetation of the Platte River: Phytosociological forest
development and seedling establishment., Dissertation, Iowa State University, Ames, Iowa,
1982.
6-71
-------
10. Johnson, W.C., Dams and riparian forests: Case study from the upper Missouri River,
Rivers, 3, 229,1992.
11. Johnson, W.C., Woodland expansion in the Platte River, Nebraska: Patterns and causes,
Ecological Monographs, 64,45,1994.
12. Petts, G.E. and Lewin, J., Physical effects of reservoirs on river systems, in Man's Impact
on the Hydrological Cycle in the United Kingdom, Hollis, G. E. Ed., Geo Abstracts Ltd.,
Norwich, U.K., 1979, 79.
13. Hickin, E.J., River channel changes: Retrospect and prospect, in Modern and Ancient
Fluvial Systems, Collinson, J. D. and Lewin, J. Eds., Blackwell Scientific Publications,
Oxford, U.K., 1983,61.
14. Petts, G.E., Impounded Rivers: Perspectives for Ecological Management, John Wiley &
Sons, Chichester, U.K., 1984.
15. Sidle, J.G., Currier, P.J., and Miller, E.D., Changing habitats in the Platte River Valley of
Nebraska, Prairie Naturalist, 21, 91,1989.
16. Krapu, G.L., Reineche, K.J., and Frith, C.R., Sandhill cranes and the Platte River,
Transactions of the 47th North American Wildlife and Natural Resources Conference, 542.
6-72
-------
17. Faanes, C.A., Aspects of the nesting ecology of least terns and piping plovers in central
Nebraska, Prairie Naturalist, 15, 145,1983.
18. Krapu, G.L. et al.5 Habitat use by migrant Sandhill Cranes in Nebraska, Journal of Wildlife
Management, 48, 407, 1984.
19, Lingle, G.R., Strom, K.J., and Ziewitz, J.W., Whooping crane roost site characteristics on
the Platte River, Buffalo County, Nebraska., Nebraska Bird Review, 54, 36, 1986.
20. Iverson, G.C., Vohs, P.A., and Tacha, T.C., Habitat use by mid-continent sandhill cranes
during spring migration, Journal of Wildlife Management, 51,8, 1987.
21-r Norling, B.S., Anderson, S.H., and Hubert, W.A^, Nocturnal behaviour of Sandhill Cranes
roosting in the Platte River, Nebraska., Naturalist, 23, 17,1991.
22. Folk, M.J. and Tacha, T.C., Sandhill crane roost site characteristics in the North Platte
River Valley, Journal of Wildlife Management, 54, 480, 1990.
23. Davis, C.A., Sandhill crane migration through the central Great Plains: A contemporary
perspective, Proc. Great Plains Migration Symposium, Lincoln, NE, Mar. 7,2 A.D.
24. Faanes, C.A. and LeV alley, MJ., Is the distribution of Sandhill Cranes on the Platte River
changing?, Great Plains Research, 3,297,1993.
6-73
-------
25. Sharpe, R.S., The origins of spring migratory staging by sandhill cranes and white-fronted
geese., Transactions of the Nebraska Academy of Sciences, 6, 141, 1978.
26. Ducey, L, Breeding of the least tern and piping plover on the lower Platte River, Nebraska,
Nebraska Bird Review, 49, 45, 1981.
27. Jorde, D.G.H. et al., Effects of weather on habitat selection and behavior of mallards
wintering in Nebraska., Condor, 86, 258, 1984.
28. USFWS, The Platte River Ecology Study, Special Research Report, Northern Prairie
Wildlife Research Center, Jamestown, North Dakota, 1981, 187.
29. Currier, P.J. and Ziewitz, J.W., Application of a sandhill crane model to the management of
habitat along the Platte River, Proceedings of the 1985 Crane Workshop, 315.
30. Helzer, C.J. and Jelinski, D.E., The relative importance of patch area and perimeter-area
ratio to grassland breeding birds, Ecological Applications, 9, 1448,1999.
31. Jelinski, D.E., Middle Platte River floodplain ecological risk assessment planning and
problem formulation, Completed under EPA Assistance Agreement CR 826077, School of
Environmental Studies, Queens University, Kingston, Ontario, 1999.
32. Colt, C.J., Breeding bird use of riparian forests along the Central Platte River: A spatial
analysis, M.S. thesis, University of Nebraska, 1997.
6-74
-------
33. Keammerer, W.R., Johnson, W.C., and Burgess, R.L., Floristie analysis of the Missouri
River bottomland forests in North Dakota, Canadian Field Naturalist, 89, 5, 1975.
34. Hibbard, E.A., Vertebrate ecology and zoogeography of the Missouri River valley in North
Dakota, PhD thesis, North Dakota State University, 1972.
35. Johnson, F.R. and Desvousges, W.H., Estimating stated preferences with rated-pair data:
environmental, health and employment effects of energy programs, Journal of
Environmental Economics and Management, 34, 79, 1997.
36. Strange, E.M., Fausch, K.D., and Covich, A.P., Sustaining ecosystem services in human-
dominated watersheds: Biohydrology and ecosystem process in the South Platte River
Basin, Environmental Management, 24, 39, 1999.
37. Habi Tech, Inc., Hydrologic components influencing the conditions of wet meadows along
the Central Platte River, Nebraska, Lincoln, Nebraska, 1-31-1993.
38. Johnson, W.C., Channel Equilibrium in the Platte River, 1986-1995, Department of
Horticulture, Forestry, Landscape, and Parks. South Dakota State University, Brookings,
South Dakota, 1996.
39. Currier, P. J., Woody Vegetation Expansion and Continuing Declines in Open Channel and
Habitat on the Platte River in Nebraska, The Platte River Whooping Crane Critical Habitat
Maintenance Trust, Grand Island, Nebraska, 1995.
6-75
-------
40. Chadwick and Associates, Forage fish monitoring study, Central Platte River, Nebraska,
1993,1994.
41. PRESP, Cooperative Agreement for the Platte River Research and Other Efforts Relating to
Endangered Species Habitat Along the Central Platte River, Nebraska, Platte River
Endangered Species Partnership, 1997, Available from
http://www.platteriver.org/librarv/CooperativeAgreement/index.htm.
42. Gilliland, M.W. et al., Simulation and decision making: The Platte River Basin in
Nebraska, Water Resources Bulletin,, 21,1985.
43. Bleed, A. et al., Decision making on the Danube and the Platte, Water Resources Bulletin,
26, 1990.
44. Razavian, D. et al., Multistage screening process for River Basin planning, Journal of
Water Resources Planning and Management, 116, 323,1990.
45. Aiken, J.D., Balancing endangered species protection and irrigation water: The Platte
River Cooperative Agreement, Great Plains Natural Resource Journal, 3,119,1999.
46. Mitchell, B., Resource and Environmental Management, Longman, London, 1997.
47. Kirsch, E. M., Habitat selection and productivity of least terns on the lower Platte River,
Nebraska., Wildlife Monographs, 132, 1996, 48.
6-76
-------
48. Wesehe, T.A., Skinner, Q.D., and Henzey, R.J., Platte River wetland hydrology study,
University of Wyoming, Laramie, 1994.
49. USEPA, Middle Platte River floodplain ecological risk assessment planning and problem
formulation, Draft, EPA 630/R-96/007a, Risk Assessment Forum, U.S. Environmental
Protection Agency, Washington, DC, 1996.
50. Johnson, W.C., Adjustment of riparian vegetation to river regulation in the Great Plains,
USA, Wetlands, 18, 608, 1998.
51. Armbruster, M.J. and Farmer, A.H., Draft Sandhill Crane Habitat Suitability Model,
Proceedings from the 1981 Crane Workshop, 136.
52. Cadmus Group, Ecological risk assessment for watersheds: Data analysis for the Middle
Platte River, EPA Contract 68-C7-002, Work Assignment B-02, Cadmus Group, Laramie,
Wyoming, 1998.
53. .Gibbons, R., Game Theory for Applied Economists, Princeton University Press, Princeton,
NJ, 1992.
54. Becker, N. and Easter, K.W., Water diversions in the Great Lakes Basin analyzed in game
theory framework, Water Resources Management, 9, 221, 1995.
6-77
-------
55. Adams, G., Rausser, G., and Simon, L., Modeling multilateral negotiations: an application
to California water policy, Journal of Economic and Behavior and Organization, 97,1996.
56. Klemperer, P., Auction Theory: A Guide to the Literature, Journal of Economic Surveys,
13, 1999.
57. Supalla, R. et al, A game theory approach to deciding who will supply instream flow
water, Journal of the American Water Resources Association, 38, 959,2002,
58. Boyle Engineering Corp., Platte River water conservation/supply reconnaissance study,
1999.
59. Jenkins, A. and Konecny, R., The Middle Platte Socioeconomic Baseline, Plate River
Studies, 1999.
60. Boyle Engineering Corp, Reconnaissance - Level Water Action Plan, Prepared for
Governance Committee of the Cooperative Agreement for Platte River Research, Boyle
Engineering Corp, Lakewood, CO, Sept. 14, 2000.
61. Babbie, E.R., Index and scale construction, in The Practice of Social Research,
Wadsworth Publishing Company, Belmont, CA, 1979,15.
62. McDonald, P.M. and Sidle, J.G., Habitat changes above and below water projects on the
North Platte and South Platte Rivers in Nebraska., Prairie Naturalist, 24, 149, 1992.
6-78
-------
63. Hurr, R.T., Groundwater hydrology of the Mormon Island Crane Meadows Wildlife Area .
near Grand Island, Hall County, Nebraska, U.S. Geological Survey Professional Paper
1277, U.S. Geological Survey, 1983.
64. Anonymous, The Groundwater Atlas of Nebraska, Conservation and Survey Division,
Institute of Agriculture and Natural Resources, University of Nebraska, Lincoln, Nebraska,
1998.
65. Supalla, R. et al., Game theory approach as a watershed management tool: A case study of
the Middle Platte ecosystem, Project Completion Report for U.S. EPA Assistance
Agreement R 82698701, Department of Agricultural Economics, University of Nebraska,
Lincoln, NE, 2002.
66. Walters, C, J., Adaptive Management of Renewable Resources, Macmillan, New York,
1986.
6-79
-------
APPENDIX 6-A
SUMMARY OF SURVEY RESPONSE INFORMATION USED TO CALCULATE
UTILITY OF ENVIRONMENTAL MANAGEMENT POLICY OPTIONS FOR THE
CENTRAL PLATTE RIVER FLOODPLAIN
Table 6-8 describes three environmental policy attributes (Method, Cost and Who pays),
each having five levels, by which 125 policy options (i.e., 53 attribute level combinations) for
addressing the central Platte River environmental management problem are defined. Bargaining
groups with respect to that environmental problem are determined as a combination of state
residency and interest group membership, as defined in Section 6.3.2.3.4 and Table 6-9.
Equations 6-3, 6-4 and 6-5 define the methods by which survey response data for several
bargaining groups are used to derive each group's utility scores for each policy option. This
Appendix summarizes certain information used in the calculation of utility. First, the degree of
support for individual policy attribute levels is presented by State (Table 6-A-l) and interest
group (Tables 6-A-2, 6-A-3 and 6-A-4). Next, the results of regression analyses conducted to
establish the relative weights of the attributes are presented (Table 6-A-5).
6-80
-------
TABLE 6-A-l
Degree of support for policy attributes, by state
Policy Attribute and Level"
Do Support"
CO
ME
WY
Don't Support0
CO
NE WY
Percent of all Respondents
Method
a. All Needs, Least Cost
b. All Needs, Minimum Water
c. Best Possible, No Ag Water
d. Adaptive Management
e. Benefit-Cost Approach
41
52.6
36.7
64.1
26.9
24.8
37.6
43.2
63
38.7
29.2
37.9
45.1
63.9
37.9
39.7
25.8
46.7
21.6
47.6
49.5
33.6
31.6
17
31
52.7
35.7
34.4
17.1
36.6
Cost
a. Invest Nothing
b. Invest $40M, per USFWS
c. Invest $10M, 25% of Need
d. Invest $20M, 50% of Need
e. Invest $30M, 75% of Need
15.7"
31.8
36.2
33.6
23.8
19.6
16.6
39
22.4
13.2
23.3
19.7
30.6
18.9
11.7
73.1
51.7
39.4
42.4
48.4
62.2
63.2
37.6
51.5
57.6
59.7
63.9
43.8
54.3
60.3
Who pays
a. All Costs Paid by Feds
b. Feds 50%, Private 50%
c. Feds 50%, States 50% Equal
d. Feds 50%, States 50% Prop.
e. Feds 1/3, Pvt.1/3, States 1/3
Proportional to Use
32.2
39.9
27.8
43.5
61.6
34.9
39.6
26.2
29.4
51.3
33
49.2
17,5
34.4
53.2
57
44.5
53.9
37.9
25
48.7
39.8
54.7
46.6
29.6
51.7
35
64.3
45.1
31.1
aA full description of each policy attribute and level is found in Table 6-8.
blncludes responses of "strongly support" and "support."
'Includes responses of "strongly oppose" and "oppose."
6-81
-------
TABLE 6-A-2
Degree of support for policy attribute levels in Colorado, by interest
Policy Attribute and Level3
Do Support*3
Ag
Envl.
No Opinion
Ag
EnvL
group
Don't Support0
Ag | Envl.
Percent of Classified Respondents
Method
a. All Needs, Least Cost
b. All Needs, Minimum Water
c. Best Possible, No Ag Water
d. Adaptive Management
e. Benefit-Cost Approach
26.1
60.9
73.9
87
43:5
56.7
69.3
22.9
55.3
17.7
4.3
13
. 8.7
8.7
26.1
Cost
a. Invest Nothing
b. Invest $40M, per USFWS
c. Invest $10M, 25% of Need
d. Invest $20M, 50% of Need
e. Invest $30M, 75% of Need
^_ 49.9
0
56.5
18.2
4.5
4.4
58
30.8
40
38.1
22.7
22.1
15.7
12.8
27
4.5
9.1
8.7
9.1
13.6
69.6
26.1
17.4
4.3
30.4
20.6
8.6
61.4
31.9
55.3
7.3
13.8
27.1
24.4
28.4
54.5
90.9
34.8
72.7
81.8
88.3
' 28.3
42.1
35.6
33.6
Who pays
a. All Costs Paid by Feds
b. Feds 50%, Private 50%
c. Feds 50%, States 50% Equal
d. Feds 50%, States 50% Prop.
e. Feds 1/3, Pvt.1/3, States 1/3 Proportional
to Use
29.4
58.8
5.9
5.9
47.1
37.7
39.7
43.8
58.7
18.4
4.3
8.7
0
0
4.2
9.4
17.6
22.6
23.2
18.4
66.5
39.1
87
87
45.8
52.9
42.6
33.6
18.1
13.2
"A full description of each policy attribute and level is found in Table 6-8.
Includes responses of "strongly support" and "support."
Includes responses of "strongly oppose" and "oppose."
6-82
-------
TABLE 6-A-3
Degree of support for policy attribute levels in Nebraska, by interest group
Policy Attribute and Level3
Do Support
Ag
Envl.
No Opinion
Ag
Envl.
Don't
Support0
Ag
Envl.
Percent of Classified Respondents
Method
a. All Needs, Least Cost
b. All Needs, Minimum Water
c. Best Possible, No Ag Water
d. Adaptive Management
e. Benefit-Cost Approach
19.6
35.6
57.3
69.9
47,1
38.8
52.9
27.1
56.5
22.4
14.7
19.8
17.5
12.6
29.4
24.7
28.2
21.2
21.2
23.5
65.7
44.6
25.2
17.5
23.5
36.5
18.8
51.8
22.4
54.1
Cost
a. Invest Nothing
b. Invest $40M, per USFWS
c. Invest $10M, 25% of Need
d. Invest $20M, 50% of Need
e. Invest $30M, 75% of Need
35.6
9.8
35
21.6
5.9
7.1
32.1
44.4
30.5
33.3
16.8
13.7
16
17.6
18.8
8.2
23.8
18.5
22
25
47.5
76.5
49
60.8
75.2
84.7
44
37
47.6
41.7
Who pays
a. All Costs Paid by Feds
b. Feds 50%, Private 50%
c. Feds 50%, States 50% Equal
d. Feds 50%, States 50% Prop.
e. Feds 1/3, Pvt.1/3, States 1/3 Proportional
to Use
38
50
19
24.8
41
37.3
36.1
34.5
44.6
54.8
12
15
15
15.8
16
20.5
20.5
20.2
24.1
23.8
50
35
66
59.4
43
42.2
42.4
45.2
31.3
21.4
aA fell description of each policy attribute and level is found in Table 6-8.
Includes responses of "strongly support" and "support."
Includes responses of "strongly oppose" and "oppose."
6-83
-------
TABLE 6-A-4
Degree of support for policy attribute levels in Wyoming, by interest group
Policy Attribute and Level3
Do Support1"
Ag
Envl.
No Opinion
Ag
Envl.
Don't
Support0
Ag
Envl.
Percent of Classified Respondents
Method
a. All Needs, Least Cost
b. All Needs, Minimum Water
c. Best Possible, No Ag Water
d. Adaptive Management
e. Benefit-Cost Approaqh
13.2
34.6
71.2
83.3
50.9
61.1
52.8
18.7
48.1
23.6
11.3
13.5
15.4
9.3
22.6
13.9
28.7
15.9
19.4
19.8
75.5
51.9
13.5
7.4
26.4
25
18.5
65.4
32.4
56.6
Cost
a. Invest Nothing
b. Invest $40M, per USFWS
c. Invest $10M, 25% of Need
d. Invest $20M, 50% of Need
e. Invest $30M, 75% of Need
42.3
3.8
37.7
13.7
0
6.5
50
27.8
27.5
25.9
19.2
13.5
18.9
19.6
19.6
5.6
10.2
28.7
26.6
28.7
38.5
86.7
43.4
66.7
80.4
88
39.8
43.5
45.9
45.4
Who pays
a. All Costs Paid by Feds
b. Feds 50%, Private 50%
c. Feds 50%, States 50% Equal
d. Feds 50%, States 50% Prop.
e. Feds 1/3, Pvt.1/3, States 1/3 Proportional
to Use
37
60.4
5.9
13.5
47.1
34.9
41.1
36.2
61.1
57.5
7.4
7.5
5.9
15.4
3.9
12.8
14
15.2
14.8
12.3
55.6
32.1
88.2
71.2
49
52.3
44.9
48.6
24.1
30.2
"A foil description of each policy attribute and level is found in Table 6-8.
^Includes responses of "strongly support" and "support."
"Includes responses of "strongly oppose" and "oppose."
6-84
-------
TABLE 6-A-5
Policy attribute weights by bargaining group"
Interest Group
Intercept
Method, M
Cost, C
Who pays, P
Colorado, State, N = 994
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0.772
0.211
0.021
0.28
0.119
0.020
0.16
0.413
0.020
0.56
Colorado Agricultural, N = 154
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0.855
0.068
0.070
0.10
0.382
0.081
0.55
0.242
0.087
0.35
Colorado Environmental, N = 840
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0.873
0.191
0.030
0.27
0.192
0.030
0.27
0.321
0.030
0.46
Nebraska State, N= 1,179
Reg, Coefficients, B
Standard Error
Normalized Weights, W
0.900
0.093
0.017
0.14
0.204
0.018
0.30
0.387
0.017
0.57
Nebraska Agricultural, N = 674
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0,628
0.056
0.031
0.07
0.198
0.036
0.25
0.524
0.035
0.67
Nebraska Environmental, N = 505
Reg. Coefficients, B
Standard Error
Normalized Weights, W
1.729
0.055
0.043
0.13
0.026
0.043
0.06
0.332
0.043
0.81
Wyoming State, N = 999
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0.663
0.129
0.018
0.17
0.198
0.019
0.26
0.420
0.018
0.56
Wyoming Agricultural, N = 646
Reg. Coefficients, B
Standard Error
Normalized Weights, W
0.840
0.078
0.043
0.12
0.177
0.054
0.27
0.396
0.050
0.61
Wyoming Environmental, N = 353
Reg. Coefficients, B
Standard Error
Normalized Weights, W
1.035
0.154
0.031
0.24
0.117
0.032
0.18
0.370
0.030
0.58
"See Equations 6-4 and 6-5 for explanation of variables and attribute weights
6-85
-------
-------
7. CONCLUSIONS
This document has introduced fundamental concepts and methods in ecological risk
assessment (ERA) and economic analysis of environmental problems, especially as applied to
watersheds (see Chapters 1 and 2), and it has developed a conceptual approach for their
integration in watershed management (see Chapter 3, and especially Figure 3-1). It has
described and evaluated case studies of three U.S. watersheds in which watershed ERA
(W-ERA) was conducted, followed by economic analysis that utilized the W-ERA findings
(Chapters 4-6). This closing chapter draws general conclusions from this research effort. For
the most part, it leaves aside issues that are particular either to ERA itself or to economic
analysis and focuses on the problem of their integration.
These conclusions do not constitute a comprehensive list of recommendations for
integrating ERA and economic analysis. The conceptual approach for integration presented in
Chapter 3 is more complete in that regard. Rather, they are a set of important observations
drawn from an overview of these three case studies. The conclusions provide further insight on
certain topics raised by the conceptual approach, but additional studies are 'still needed to explore
that approach more fully.
7.1 ACHIEVING ECOLOGICAL-ECONOMIC INTEGRATION REQUIRES A
COHERENT STRATEGY
The central conclusion arising from evaluation of the case studies is that watershed
problems should be approached with a coherent strategy for assessment and management. If
decision-makers need to consider both ecological risks and economic factors (and perhaps other
factors), a strategy that guides their integration is necessary. The conceptual approach described
in Chapter 3 provides such a strategy. The approach is based on the U.S. Environmental
7-1
-------
Protection Agency (USEPA) Framework for Ecological Risk Assessment?'2 and it modifies or
augments that framework as needed to accommodate economic analysis, and to address a
broader management context. Its elements are similar to those of other frameworks that have
been used in environmental management (see Table 3-4 and Appendix 3-A). Although this
document presents the conceptual approach before the case studies, to serve as a guide to their
evaluation, it was developed following their completion and should be considered the main
outcome of this body of investigation.
The case studies help illustrate the need for the conceptual approach. The W-ERA
studies were not undertaken with economic integration as a goal. The economic studies did have
such a goal, but used only a limited set of guiding principles; i.e., each economic analysis was to
address the same system, problems and ecological assessment endpoints analyzed by the
W-ERA, and it was to be relevant to decision-making. The approaches used were novel and the
results are potentially useful, but in each case their usefulness could have been improved by a
more comprehensive approach, as is detailed in the following sections. For example, the lack of
an interdisciplinary assessment planning and problem formulation process contributed in one
case to divergent views of goals and endpoints. In two cases (Clinch and Platte), management
alternatives were formulated for economic analysis, but the likely ecological effects of those
alternatives were not quantitatively assessed, limiting the scope of the conclusions. Also, in two
cases (Darby and Clinch) the economic analysis tools chosen were not clearly aligned to the
relevant decision context; that is, it was not shown that they were developed with a set of
decisions and decision-makers in mind. Use of the conceptual approach for integration
theoretically could have helped avoid these limitations.
7-2
-------
It is unlikely however, that ideal conditions will often exist in which ecologists,
economists, other specialists and stakeholders can make a clean start to define a problem
together, using an inclusive, analytic process. More likely is the kind of situation described in
these case studies, where some baseline of study and stakeholder involvement has been
established and an effort is made later to inject additional elements. Although it may be
infeasible to restart the entire process, it is nonetheless advisable to revisit key portions of the
early steps of assessment, with stakeholder involvement, so as to harmonize management
objectives, decision context, management alternatives and assessment endpoints to the extent
possible. It is also important to use the guiding considerations presented in Section 3.2 (see
Table 3-2) to identify ways to make ongoing efforts more integrated in character.
7.2 INTEGRATION REQUIRES ASSESSMENT PLANNING AND PROBLEM
FORMULATION TO BE INTERDISCIPLINARY
The conceptual approach emphasizes the need for eeologists and economists (and other
specialists as required) to participate together in the steps of assessment planning and problem
formulation. The fact that the ERA and economic analysis were done sequentially in these case
studies, rather than in a more integrated fashion, limited their value for management. In the Big
Darby Creek watershed of Ohio, a team of ecologists and economists from Miami University
built upon a W-ERA that had been initiated several years earlier by USEPA, Ohio EPA and a
number of other partners.3"5 Economic analysis in the Clinch Valley of Virginia and Tennessee
was by an interdisciplinary team headed by the University of Tennessee-Knoxville (UT-K) and
used the results of a W-ERA previously conducted by USEPA, U.S. Fish and Wildlife Service
(USFWS), Tennessee Valley Authority and other partners.6^8 hi the central reach of the Platte
River, a study by economists from the University of Nebraska-Lincoln (UN-L) built upon the
7-3
-------
foundations of a W-ERA that had been initiated by USEPA, UN-L, USFWS, and U.S.
Geological Survey, with participation by a host of local stakeholder groups.9"11
In the W-ERA efforts, planning and problem formulation were systematic and
painstaking, but economists were not involved. When the economic studies were initiated,
informational meetings were held with members of the W-ERA teams, but these did not reopen
fundamental questions about the management problems, so views about management goals and
objectives were not necessarily the same.
The lack of a common view was most pronounced in the Platte River case study. The
W-ERA team viewed the vegetative diversity and dynamic character of the braided-river-channel
landscape mosaic as an endpoint in itself, as well as the diversity of fauna using its various
habitats. The economic team focused more narrowly on current efforts among the three Platte
River states and the federal government to reach agreement on provisions to meet the needs of
three endangered species, the interior least tern (Sterna antillarum athalassos), the piping plover
(Charadrius melodus) and the whooping crane (Grus americana). The W-ERA analyzed
conditions affecting the use of river segments by sandhill cranes (Grus canadensis), whose needs
overlap substantially with the endangered species', but they also analyzed the effects of
landscape patch size on grassland and woodland breeding birds, whose needs are less relevant to,
and in some cases conflict with, those of the endangered species. In the other two case studies
there was not a significant divergence of views; however, the lack of a joint assessment planning
exercise may have contributed to the failure to identify a decision context for the economic
assessment, as well as certain other weaknesses discussed below.
7-4
-------
7.3 RESEARCH IS NEEDED ON THE DEVELOPMENT AND USE OF
INTEGRATED CONCEPTUAL MODELS
A conceptual model is a graphical depiction, typically a box-and-arrow diagram, of the
hypothesized relationships between human activities, ecological stressors, and ecological
assessment endpoints (refer to Section 2.1.1.2 for explanation, and Figure 5-3 for an example).
According to the conceptual approach for integration, an interdisciplinary problem formulation
process should include the development of extended conceptual models (see Section 3.3.2). In
extended models, risk hypotheses show how sources and stressors affect economic endpoints, or
services, as well as ecological assessment endpoints. An extended model also includes risk
management hypotheses, which we have defined as explanations of how management
alternatives are expected to affect sources, exposures, effects and services. Their development
should involve environmental program managers, if the management actions are in the form of
programs or policies, and environmental engineers or restoration specialists, if the actions
involve structural changes to ecosystems. Their development also requires the involvement of
land owners and other stakeholders whose active cooperation may be instrumental in solving the
environmental problem. Extended models were not developed in these case studies, and at
present we are not aware of examples of the use of these extended models in a risk assessment
context. The National Center for Environmental Assessment of USEPA's Office of Research
and Development is presently initiating work to gain experience with their development and use.
7.4 CLEARLY FORMULATED MANAGEMENT ALTERNATIVES FACILITATE
INTEGRATED ANALYSIS
Describing management alternatives is an important way to frame the integration
problem. Any given alternative will entail a unique set, or bundle, of ecological, economic and
other kinds of changes. Some of those changes may be judged beneficial and some detrimental,
7-5
-------
some to a larger and others to a lesser degree. The heart of the integration problem is to
somehow evaluate the signs and magnitudes of all these changes collectively, on a common
scale, to determine if one alternative can be clearly preferred over another.
In the Big Darby watershed, three possible land use scenarios (low-density ranchettes,
low-density cluster, and maintaining agriculture) and a most-likely base case (high density
residential) were described in some detail, and their respective ecological, economic and quality-
of-life impacts were determined. Using the contingent valuation method (CVM), the researchers
were able to jointly value the different impacts. Although each respondent was posed only one
of the three possible choices, mean willingness to pay (WTP) serves as a kind of referendum on
these three alternatives.
In the Clinch Valley study, two hypothetical policies for establishing voluntary
agriculture-free riparian zones (i.e., a narrow zone and a wider zone), compensated by property
or income tax revenues, were employed in a conjoint survey. The choice sets included in the
survey were generated as random combinations of these policies and other attributes describing
potential ecological outcomes and individual payments, and therefore the choices did not
correspond to specific policy scenarios. However, the resulting choice model could be used to
generate a mean WTP for obtaining any policy scenario that could be described from those
attributes (as compared to the status quo), and such a value would have an interpretation similar
to the Big Darby result. As mentioned above, however, the expected ecological outcomes of
such a policy were not estimated.
The Platte study differed in that the problem of determining a preferred policy was
viewed not as one of determining mean WTP but rather as determining what policy certain
competing factions were most likely to find mutually acceptable. Like the other two, it elicited
7-6
-------
responses to preference questions that combined ecological and economic dimensions, but unlike
them it used this information to model a negotiation process. Using principles from game theory
(the study of interacting decision-makers),12 the model analyzed 125 hypothetical policies for
meeting endangered species needs, where a given policy described the method of meeting those
needs, its cost and who would pay. As in the Clinch case study, the expected ecological
outcomes of the policies were not analyzed. Since management alternatives are important for
economic analysis and for decision-making, their formulation should receive careful attention
from all parties involved in the assessment, and their ecological outcomes should be estimated.
7.5 CAREFUL EFFORT IS REQUIRED TO RELATE ECOLOGICAL ENDPOINTS
TO ECONOMIC VALUE
An important step in the problem-formulation phase of ERA is the selection of ecological
assessment endpoints. Assessment endpoints are chosen that are considered ecologically
relevant, susceptible to the stressors of concern and relevant to the environmental management
objectives (see Section 2.1.1.2). The likelihood of adverse effects on these endpoints is
described in the risk-characterization phase (see Section 2.1.1.4). The challenge of ERA -
economic integration includes determining economic value (see Section 2.2.2) associated with
those changes as well as characterizing other linkages between the ecological system,
management actions and economic value (see Section 3,3.2).
In these case studies, endpoints chosen for ERA because of their ecological importance
sometimes posed a challenge for economic analysis. Whereas the freshwater mussel faunas of
the Big Darby and Clinch systems are considered ecologically significant, members of the
general public who are unaware of their diversity and threatened status may be unconcerned
about their survival. To counter this problem, in the Clinch study the survey text mentioned
7-7
-------
mussels ten times in its brief introductory paragraphs, explaining their unusual degree of
diversity in the Clinch Valley, their usefulness as an indicator of water quality, and their
sensitivity to pollutants and susceptibility to crashing by the hooves of cattle, before posing the
choice sets. In the central reach of the Platte River, where management concerns have centered
on endangered or conspicuous migratory waterfowl, the landscape-ecological viewpoint
employed in the W-ERA treated landscape diversity, and several less conspicuous bird species,
as additional endpoints. These endpoints did not factor in the economic (i.e., game theoretic)
analysis, but if efforts had been made to value these endpoints, similar problems would have
been faced.
Another complication occurred when the measurement methods that were used to express
the ecological endpoints, or were a surrogate for the endpoints, were not readily understandable
to the public. For example, even if the public considers a diverse stream fauna to be important,
they may have difficulty determining what they would be willing to give up in order to obtain,
e.g., a 3-point or 10-point improvement in a multimetric ecological index. Since these indices
may be composites often or so individual measures, it is impossible to make a scientifically
precise statement about the meaning of any such change. Yet because these indices are
becoming widely relied upon to indicate the presence or absence of biological impairment
(Section 2.3.1 and Appendix 2-B), they are likely to be a critical part of the available knowledge
base about ecological risk in a watershed, and ways must be found to adequately communicate
their meaning if individuals are to determine how such changes affect their welfare. In the Big
Darby CVM study, the index of biotic integrity (IBI, a fish assemblage indicator) and
invertebrate community index (ICI, a stream-bottom community indicator) were used as risk
assessment endpoints. CVM survey respondents were shown a table (see Table 4-1) in which
7-8
-------
each of the four land use scenarios was rated from "low" to "high" for each of four stressors
(nutrients, sediments, toxins and flow pattern) and were told that a "high" level posed a "risk to
stream integrity." In the Clinch Valley study, where the study of risks relied heavily on IBI,
respondents were presented with choice sets in which one of six attributes of the choice was
"aquatic life," and the possible levels were "full recovery," "partial recovery" or "continued
decline" (see Tables 5-3 and 5-4). The supporting text (see Appendix 5-A) explained that
"partial recovery" meant "some improvement" in the Clinch River but not in its tributaries,
whereas "full recovery" meant "improvement" in both the Clinch River and its tributaries. Both
surveys avoided direct presentation of the indices, using instead qualitative description. While
the descriptors for the Darby study could be related back to the results of scenario impact
analysis, those for the Clinch were not as easily related to a given physical change.
Where environmental management may have particular objectivesfor example, the
protection of water quality and stream biological integritythe results of management actions
can affect additional endpoints as well. Therefore, the ecological information set needed for
economic analysis may be broader than that envisioned in the ERA (if problem formulation for
the ERA did not include consideration of management alternatives). In the Clinch Valley, for
example, management actions examined in the economic study included hypothetical policies to
compensate farmers for voluntarily restricting agriculture from a riparian buffer area. Besides
improvements in diversity of native mussels and fish, which were the ERA endpoints, the
economists expected that such policies would improve sport fishing and enhance the presence of
songbirds, which were not included in the ERA. Consideration of songbirds turned out to be
unimportant in this case, since respondents did not appear to value them significantly (see Table
5-8), but sport fishing was important. Full analysis of the economic benefits of these policies
7-9
-------
therefore would have required analysis of sport fish response. Other potential economic benefits
of riparian zone restoration may result from enhanced nonavian wildlife habitat, reduced nutrient
export, increased sequestration of carbon and improved value of river-corridor recreation such as
canoeing. Had an attempt been made to capture these values as well, additional ecological and
economic endpoints only tangentially related to the original management goal would have been
required.
Estimating the benefits of a given change in the ecological condition of water bodies
requires better procedures. In 1986, Mitchell and Carson13 reported on a national survey of U.S.
households to quantify water quality-related WTP. They used a "water quality ladder" that
established progressively increasing use levels (i.e., boating, game fishing, swimming, drinking)
for surface waters. These levels equated to points on a cardinal scale determined as a combined
index of five conventional water quality parameters: fecal coliforms, dissolved oxygen,
biological oxygen demand, turbidity and acidity (pH). Thus, the benefit of any change in those
parameters could be associated with WTP, but the index reflected only a narrow set of pollutants
and did not include any direct measurements of stream biological communities.
Since then, substantial progress has been made in the development of state programs for
biological monitoring and the use of indices such as IBI and ICI in water quality standards
(WQS). These programs have not required a detailed understanding on the public's part of the
measures that underlie the indices or a feel for their numerical scales. Work must be done to
expand the scientific basis of the water quality ladder to include a broader set of ecological
measures, or in some cases to replace exposure (pollutant) measures with response (biological)
measures. In addition, the uses, or rungs, that were originally examined need to be expanded to
better reflect the full variety of uses that have been designated in state WQS programs (personal
7-10
-------
communication with John Powers, USEPA Office of Water) as well as other levels of quality to
which the public may attribute value. For example, respondents in the Big Darby and Clinch
Valley studies probably recognized freshwater mussels as valued components of those aquatic
communities, yet a level of water quality sufficient to support game fish (the highest rung of the
ladder) may not be sufficient to promote "full recovery" of mussels.
Thus, informed decisions (i.e., ones where decision-makers understand the inherent trade-
offs) require techniques that link the kinds of indices ecologists currently measure to values held
by the public. Part of the challenge, therefore, is translating indicators into common language.
By the same token, ecological measures may require adjustment. For example, if the public
values the response of the instream biological community to stream corridor restoration, then
ecological measures of the efficacy of such projects should not be limited to modeled changes in
water quality parameters. Similarly, if sport fishing and bird watching are among the values the
public places on such projects, then measuring aquatic community integrity alone is not
sufficient. Further, since ecological measurements are highly variable, and model predictions
highly uncertain, research needs to include methods to enable the public to understand and
account for ecological uncertainty in their preferences.
7.6 THE APPROPRIATE TOOLS FOR ANALYSIS AND COMPARISON OF
ALTERNATIVES DEPEND ON THE DECISION CONTEXT
To weigh management alternatives, analysts should select comparison methods that fit
the decision context. If decision criteria are constrained by statute or regulation, then the
comparison procedure must include any required information and be capable of segregating any
precluded information. For example, regulatory impact analyses conducted by USEPA (see
Section 2.3.2) may require an analysis in which all costs and benefits are monetized to the
7-11
-------
greatest extent feasible. By contrast, U.S. Army Corps of Engineers project evaluation
procedures maintain separate accounts for changes in the national output of goods and services
(expressed in monetary units) and changes in the net quantity and/or quality of desired ecosystem
resources (expressed in physical units).15 These decision contexts imply particular comparison
procedures, whereas in other contexts procedures can be subjective or ad hoc. Other important
differences in context may be as follows:
one entity has clear authority to decide vs. many parties will negotiate
one decision will be made affecting a large area vs. many small decisions will be made,
each affecting only one land parcel, stream segment or political jurisdiction
decision-makers expect to reach a decision point once analysts have presented all
information vs. decision-makers expect to examine data, construct alternatives and
engage in an active decision process.
To ensure successful integration researchers need to categorize environmental management
situations based on the decision context and evaluate the full complement of comparison
procedures available, in order to identify compatibilities between context and procedure.
All three of the case studies surveyed watershed residents, and in some cases other
members of the public, and used information about respondents' preferences to produce tools
that integrated ecological and economic factors. These tools are potentially useful in decision-
making and management. In the Big Darby case, the Miami team developed a broadly framed,
contingent valuation method (CVM) approach for comparing economic, ecological and quality-
of-life outcomes among four alternative futures. Preferences were expressed via a WTP measure
that is consistent with standard economic theory and therefore useful in a variety of decision
contexts.
7-12
-------
The UT-K study developed a choice model to measure valley residents' preferences
regarding hypothetical riparian management policies in the upper Clinch and Powell Rivers.
Since the model's parameters correspond to a set of attributes of the choice, the model is
sufficiently flexible that it could be used (in conjunction with expert judgment) to refine the
design of an actual policy so as to maximize its value to Clinch Valley residents.
In the central reach of the Platte River, the UN-L economists modeled the utility of a
large set of potential policies from the perspectives of different groups. They also investigated
an auction method whereby the states of Colorado, Wyoming and Nebraska could more readily
agree on water price and supplier.
The tool developed by each team has potential for application to management problems
in the watershed studied, and the methods involved could be adapted to other environmental
settings. The CVM approach taken in the Darby presented concrete choices in readily
understandable terms. Although the choices between development scenarios were hypothetical,
the visual impact of photographic examples of each kind of development as it occurs in the
watershed made the choice very realistic. On the other hand, the attribute-driven models
developed as part of the Clinch and Platte analyses afforded greater analytic flexibility, which
could have substantial value in later phases of management. Since negotiation between affected
parties can play an important role in decision-making, an analytic approach that can respond
quickly to changes in design (i.e., without the requirement of a new survey) may be very useful.
If an incremental or adaptive (learn-as-you-go) implementation approach is to be used, a flexible
model would be preferred that could be revised and rerun using newly acquired information
about the effectiveness of the management approach. However, if the public is becoming more
informed in the process, then a new survey may be required in any case.
7-13
-------
The game theoretic approach employed in the Platte River case was carefully selected to
fit a specific decision context, i.e., a tristate negotiation to meet Endangered Species Act
requirements. On the other hand, the tools developed for the Darby and Clinch watersheds
provided information about regional development preferences but were not directed at a
particular set of decision-makers. Thus, although the latter tools are potentially useful, it is less
clear that they are the best tools for management of those watersheds.
7.7 RESEARCH IS NEEDED ON TRANSFERRING THE VALUE OF
ECOLOGICAL ENDPOINT CHANGES
Environmental management problems tend to be highly unique, complicating the direct
transfer of economic findings from one situation to another. A given watershed under study is
likely to differ in some key characteristic from another where a similar problem may already
have been studied. The novel methods developed in each of these case studies undoubtedly
could be adapted for use in other systems. Given the expense and time of conducting surveys,
however, analysts need to understand whether there are dimensions of value that are less variable
across systems. The Big Darby results suggest that one might be able to improve the
comparability of WTP estimates by using the numerical IBI change and area affected (or perhaps
stream miles affected) as normalizing factors. Work in that case is still ongoing to determine if
ecological value can be estimated as a fraction of WTP. The Clinch Valley case study
decomposed WTP according to a set of attributes, determining part-worths for each. We might
hypothesize that such a partial value, if normalized for magnitude and extent of stream
improvement, would be less variable across situations than would a more bundled estimate.
These assumptions require validation, however.
7-14
-------
7.8 THE ROLE OF ECOLOGICAL RISK INFORMATION IN THE
MEASUREMENT OF PREFERENCES REQUIRES FURTHER RESEARCH
Individuals' preferences about uncertain outcomes reflect their expectations about those
outcomes, and expectations depend on beliefs.16 When individuals know little about an
environmental management problem, the information provided in a survey will have an
important influence on the construction of beliefs and the statement of preferences.17 The
purpose of ERA is to develop accurate information about the nature, magnitude and certainty of
adverse effects to ecological resources, given present circumstances and sometimes under
different prospective management regimes. The challenge of integration therefore goes beyond
determining how to associate preferences with risk outcomes; it includes determining the
appropriate use of risk information to inform (or even construct) preferences.
The treatment of information and belief used in the Clinch Valley study was the most
conventional, in that the mail-out questionnaire included some introductory and explanatory text
(including discussion about mussels, as pointed out earlier) to help respondents understand the
questions, and it asked questions about respondent age, income, education, environmental beliefs
and affiliations to help characterize the respondent population and determine the factors that
underlie preference. The Platte River study similarly employed a mail-out survey with an
informative preamble and demographic questions, but it took the additional step of asking
respondents' agreement or disagreement with a series of statements about the environmental
management problem. These were intended to determine not only attitudes but knowledge, since
the statements were considered to have known, correct answers,3 and responses were used to
a In reality there was some ambiguity about this distinction, since not all of the answers could be clearly established
by documentation.
7-15
-------
score respondents' knowledge level (see Section 6.3.2.1.2). This information was then used to
speculate about the potential effects of better information on negotiation outcomes. By contrast,
the Big Darby survey used an in-person presentation approach, with a detailed script and a
computer-based slide show including many photographs, to clearly illustrate each of the
development scenarios and their anticipated outcomes. Risk assessors have often recognized risk
communication as an important field for research and development of practical techniques. The
differing approaches used in these surveys highlight the importance of defining best practices
and exploring novel techniques for risk communication in survey design and in other stages of
decision-making,
7.9 FINAL WORD
Because watershed boundaries often encompass areas that are ecologically and socially
complex, assessment and management of watershed problems can be complex as well.
Processes to support watershed decision-making need to be flexible and adaptable to a given
context, and multidisciplinary analyses are often required. Differences in methodology between
the disciplines, especially between the natural and social sciences, can complicate the decision-
making task, but as these case studies have shown they also provide fertile ground for the
development of unique approaches. The conceptual approach for integration of ERA and
economic analysis presented in this document offers a set of principles and procedures that can
help ensure that analyses are constructively focused and mutually supportive. It also offers a
coherent framework within which other novel, analytical approaches should be explored.
7.10 REFERENCES
1. USEPA, Guidelines for ecological risk assessment, EPA/630/R-95/002F, Risk Assessment
Forum, U.S. Environmental Protection Agency, Washington, DC, 1998.
7-16
-------
2. USEPA, Framework for ecological risk assessment, EPA/630/R-92/001, Risk Assessment
Forum, U. S. Environmental Protection Agency, Washington, DC, 1992.
3. Cormier, S.M. et al., Assessing ecological risk in watersheds: a case study of problem
formulation in the Big Darby Creek watershed, Ohio, USA., Environmental Toxicology
and Chemistry, 19, 1082, 2000.
4. Schubauer-Berigan, M.K. et al., Using historical biological data to evaluate status and
trends in the Big Darby Creek watershed (Ohio, USA), Environmental Toxicology and
Chemistry, 19, 1097, 2000.
5. Gordon, S.I. and Majumder, S., Empirical stressor-response relationships for prospective
risk analysis, Environmental Toxicology and Chemistry, 19, 1106, 2000.
6. Diamond, J.M., Bressler, D.W., and Serveiss, V.B., Diagnosing causes of native fish and
mussel species decline in the Clinch and Powell River watershed, Virginia, USA,
Environmental Toxicology and Chemistry, 21, 1147, 2002.
7. Diamond, J.M. and Serveiss, V.B., Identifying sources of stress to native aquatic fauna
using a watershed ecological risk assessment framework, Environmental Science and
Technology, 35, 4711, 2001.
7-17
-------
8. Diamond, J.M. et al., Clinch and Powell Valley watershed ecological risk assessment,
EPA/600/R-01/050, U.S. Environmental Protection Agency, Office of Research and
Development, National Center for Environmental Assessment, Washington, DC, 2002.
9. Colt, C.J., Breeding bird use of riparian forests along the Central Platte River: A spatial
analysis, M.S. thesis, University of Nebraska, 1997.
10. Cadmus Group, Ecological risk assessment for watersheds: Data analysis for the Middle
Platte River, EPA Contract 68-C7-002, Work Assignment B-02, Cadmus Group, Laramie,
Wyoming, 1998.
11. Jelinski, D.E., Middle Platte River floodplain ecological risk assessment planning and
problem formulation, Completed under EPA Assistance Agreement CR 826077, School of
Environmental Studies, Queens University, Kingston, Ontario, 1999.
12. Varian, H., Microeconomic Analysis, W.W. Norton and Company, NY, 1992.
13. Mitchell, R.C. and Carson, R.T., The Use of Contingent Valuation Data for Benefit/Cost
Analysis in Water Pollution Control, Final Report, EPA Assistance Agreement # CR
810224-02, Resources for the Future, Washington, DC, 1986.
14. Schiller, A. et al., Communicating ecological indicators to decision-makers and the public,
Conservation Ecology, 5, 19 [online], 2001.
7-18
-------
15. USAGE, Planning Guidance Notebook, ER 1105-2-100, U.S. Army Corps of Engineers,
Washington, DC, 2000.
16. Diamond, P.A. and Hausman, J.A., Contingent value: Is some number better than no
number?, Journal of Economic Perspectives, 8, 45,1994.
17. Gregory, R., Liehtenstein, S., and Slovic, P., Valuing environmental resources: A
constructive approach, Journal of Risk and Uncertainty, 7, 177, 1993.
7-19
-------
?/EPA
United States
Environmental Protection
Agency
Office of Research and Development
National Center for Environmental
Assessment
Cincinnati, OH 45268
Official Business
Penalty for Private Use
S300
EPA/600/R-03/140R
September 2003
www.epa.gov
------- |