Final Report: Developing Methods and Tools for Watershed Restoration: Design, Implementation, and Assessment in the Willamette Basin, Oregon

EPA Grant Number: R827146
Title: Developing Methods and Tools for Watershed Restoration: Design, Implementation, and Assessment in the Willamette Basin, Oregon
Investigators: Bolte, John P. , Jepson, Paul C. , Li, Judith L. , Santelmann, Mary , Smith, Courtland
Institution: Oregon State University
EPA Project Officer: Hiscock, Michael
Project Period: October 1, 1998 through September 30, 2001 (Extended to September 30, 2002)
Project Amount: $809,993
RFA: Water and Watersheds (1998) RFA Text |  Recipients Lists
Research Category: Water , Water and Watersheds


The overall objective of this research project was to integrate models of watershed function and economic characterizations of restoration options with stakeholder priorities and constraints to provide a tool for stakeholders to identify feasible restoration strategies and evaluate the ecological and economic effectiveness of these strategies, while addressing watershed-level function using two case studies. Specific objectives included the following:

· Refine and integrate a set of simplified models relating land use, ecological factors, and watershed hydrology to measures reflecting water quality, habitat, and biodiversity endpoints at a watershed level.

· Coordinate with community-based Watershed Councils to identify and prioritize restoration goals and options for two diverse watersheds.

· Characterize potential restoration strategies from an economic and social perspective.

· Develop a decision-making framework integrating these models and characterizations, coupled with community-based prioritization strategies for generating and prioritizing potential restoration activities.

· Evaluate the impact of using this framework on stakeholder decision-making and transferability of the methodology using two watersheds in the Willamette Basin, OR.

Summary/Accomplishments (Outputs/Outcomes):

The integration of ecological, economic, and social goals is an important element of watershed restoration planning and prioritization. It generally is accepted that for restoration efforts to be successful, each of these goals must be addressed in a manner that reflects stakeholder priorities, objectives, and constraints. Additionally, it is becoming increasingly apparent that restoration strategies based solely on opportunistic, site-scale activities frequently do not accomplish watershed-scale goals. Because watersheds are complex systems involving the integration of human, hydrologic, and ecological processes, it can be difficult to understand the consequences of particular restoration activities on meeting restoration goals. Synthesis tools capturing spatially explicit data are needed to couple human and ecological processes with landscape features to assist in developing effective restoration plans. This research project focused on developing such a decision tool supporting stakeholder groups in determining restoration priorities, given diverse social, economic, and ecological goals.

The general approach developed here employed several key components. First, we used a geographic information system (GIS) to organize and manage spatially explicit data sets relevant to restoration decision-making. Second, a multi-objective decision-making engine was developed to interface with the GIS data sets, identify stakeholder goals, and evaluate a range of restoration alternatives at each site in the landscape to determine optimal plans that address diverse stakeholder goals. Finally, a series of spatially explicit models of watershed function and economic characterizations of restoration options were developed to determine the effectiveness at which the generated restoration plan satisfies stakeholder-determined constraints and priorities. The resulting tool, RESTORE, can assist stakeholders in identifying feasible restoration strategies and evaluate the ecological and economic effectiveness of these strategies at addressing watershed-level ecological, economic, and social function.

The RESTORE tool contains a series of rules that relate specific site-based restoration alternatives, stakeholder goals, and site-specific landscape features to generate feasible restoration plans that reflect stakeholder concerns. These were developed through cooperation with two Watershed Councils representing diverse watershed types and disturbance levels. The analysis framework uses a landscape generator to apply design heuristics that embody ecological, economic, and social constraints and preferences to allocate restoration activities to specific sites based on site features. It also evaluates the resulting landscape options using a series of ecological, social, and economic watershed-scale models.

Sociological analysis of the stakeholder groups represented in the two Watershed Councils has been completed, and we observed that many diverse issues influence local decision-making. We found that using a rule-based framework for capturing qualitative relationships between restoration strategies, stakeholder objectives, and site features is an effective way of representing these relationships in a manner that stakeholders can readily understand. Making restoration recommendations at a site level distributed across a watershed allows for the evaluation of the effectiveness of basin-wide plans at meeting stakeholder goals. Furthermore, the use of multi-objective methodologies provides a stakeholder-accessible method for weighing and balancing competing economic, social, and ecological objectives. It is anticipated that these rules could be readily adaptable to other situations, restoration strategies, and stakeholder goals. Although it is too soon to evaluate the usage of the tool by the Watershed Councils, the initial response has been very positive.

The key findings are summarized below under the major areas of the project.

Decision-Making Framework

The decision-making framework is depicted in Figure 1.

Figure 1. Decision Framework Overview

Major components of the framework include: (1) stakeholder identification and prioritization of restoration alternatives, development of reference condition goals, and identification of criteria for evaluating successful outcomes; (2) a rule-based landscape generator using restoration rules to generate feasible watershed-scale restoration plans using site-level landscape information in conjunction with the goals and objectives from step (1) above; and (3) a watershed-scale plan evaluator employing more detailed ecological and socioeconomic models for evaluating the plan with respect to water quality, water quantity, habitat quality, social, and economic goals.

The decision-making framework employs a set of approximately 400 rules that relate specific stakeholder objectives, restoration alternatives, and site conditions to score the utility of the restoration alternative at meeting the objective/subobjective at specific sites. These rules are the basic mechanism for representing restoration knowledge in the system. The broad classes of objectives considered by the system were developed in consultation with the two Watershed Councils. Simultaneously, we identified a set of possible site-based restoration alternatives for stakeholder and scientist review. Identifying a wide range of alternatives helped to build credibility and public trust throughout the process. When developing the full range of alternatives, a "no action" alternative was included. Restoration alternatives are identified based on: (1) Watershed Council perceptions of feasible restoration strategies, and (2) the utility of the alternative at addressing environmental problems and/or to preserve sensitive areas from potential future environmental pressures.

We characterized the restoration alternatives identified above using rules and constraints that related the efficacy of the restoration alternative at meeting different stakeholder objectives and subobjectives under different site conditions. First, constraints are defined that restrict a restoration alternative based on a site-scale characteristic; for example, a riparian buffer will only be considered if the site in question is adjacent to a stream. Constraints are absolute: if a particular alternative cannot pass the constraints imposed by a particular site, it is eliminated from further consideration.

Rules further relate restoration alternatives to stakeholder objectives. The following is an example rule:

    Rule: Consider a woody riparian buffer for water quality improvement if water quality subgoal includes temperature mitigation and stream order is greater than or equal to 4 (weight [compared to related rules]: 0.50). Confidence in this rule is moderate.

Several items in this rule are noteworthy. First, this rule relates a specific restoration strategy (woody riparian buffer) to a particular objective (water quality improvement) and subobjective (temperature mitigation) through a site-level characteristic (stream order). Second, because this may be one of several rules relating this alternative to this objective, we provided a relative weighting compared to related rules. This is used by the rule engine to accumulate evidence of the utility of this alternative as it evaluates all applicable rules. Third, a confidence indicator is specified for this rule. This allows the rule engine to accumulate an overall confidence value for the utility of a particular restoration alternative based on all applicable rules. For each alternative and each objective, a series of constraints and rules were developed analogous to that as described above. Specific objectives and subobjectives addressed by these rules are provided in Table 1. Rule development is an ongoing task, and as more alternatives are identified or more information relating an alternative to stakeholder objectives and landscape characteristics is provided, rules are added to the system.

Table 1. Objectives and Subobjectives Used in RESTORE

Objective Subobjectives
Social Cost of the Restoration
Private Property Rights
Public Support
Strong Scientific Basis
Low/High-Impact Recreation
Opportunities for Public Education
Water Quality Temperature
Fecal Coliform
Dissolved Oxygen
Water Quantity Low summer flows
Winter flooding
Habitat Grass users
Shrub Users
Conifer Forest Users
Hardwood Forest Users
Coolwater Fish
Aquatic Invertebrates
Terrestrial Invertebrates

A rule engine utilizes these rules to generate restoration plans that meet site-specific and landscape-level ecosystem guidelines as well as being consistent with stakeholder goals and constraints. Because of the unique requirements of landscape generation, a rule engine with specific properties was developed for this effort. This rule engine has the following characteristics: (1) it can handle spatially explicit rules and data sets; (2) it can examine multiple rules and accumulate evidence based on these rules; (3) it can propagate confidence values from multiple rules; and (4) it is computationally efficient. To facilitate the development of rule sets, a high-level rule language specific to the requirement of landscape generation was developed, greatly facilitating the development and maintenance of rule sets. To provide efficient runtime computation, these rule sets are compiled by the rule engine prior to execution.

To generate feasible landscapes, an initial landscape reflecting current conditions is developed by spatially intersecting GIS polygon coverages for land use/land cover, topography, and soils. The resulting polygon coverage provides the base set of cells used as computational units for landscape generation. For the watersheds used in this study, the resulting polygons were typically 0.5 ha to 5 ha in area. These are employed by the landscape generator to generate feasible restoration plans based on site-specific data. Identifying and evaluating these plans is a complex task, but one for which a rich body of knowledge has been developed. From a computational perspective, planning problems of this general class have been widely studied, and generally applicable methodologies are available.

Once a landscape is generated, RESTORE evaluates the restored landscape using a set of evaluative models. All models compare the existing landscape to that resulting from a prescribed level of restoration activities generated from RESTORE. The WET_Hab model evaluates the landscape from habitat perspective, ranking breeding and feeding habitat for a wide range of species and species groups. Hydrology and sediment transport are evaluated using the WET_Hydro model, a spatially explicit, network-based model characterizing surface, subsurface, and instream flow pathways, and WET_Temp, a network-based stream temperature model.

Overall, we were successful in developing an approach and software tool for assisting our stakeholder groups in developing restoration plans. The approach successfully integrated GIS, multi-objective decision-making, rules embodying scientific knowledge of restoration alternatives, and simulation models predicting landscape response to effectively support the stakeholder decision process. The tool was well received by the stakeholder groups for several reasons. First, the fact that it was GIS-based allowed stakeholders to use the tool to visualize their watersheds from numerous perspectives. Second, because stakeholders could easily set specific objective weights and explore the implications of these weights on restoration planning, they felt "in control" of the process. Furthermore, because the scientists in the project focused on providing objective information independent of specific objectives, stakeholders felt that the information provided by the scientist was relatively unbiased. Finally, because the rules that produced the plans were readily available to stakeholders and could be related to specific decisions, the process was perceived to be transparent.

Journal Articles on this Report : 2 Displayed | Download in RIS Format

Other project views: All 32 publications 4 publications in selected types All 2 journal articles
Type Citation Project Document Sources
Journal Article Lamy F, Bolte J, Santelmann M, Smith C. Development and evaluation of multiple-objective decision-making methods for watershed management planning. Journal of the American Water Resources Association 2002;38(2):517-529. R827146 (Final)
R833626 (Final)
  • Full-text: EurekaMag-Full Text PDF
  • Abstract: Wiley InterScience
  • Journal Article Smith CL, Gilden J. Assets to move watershed councils from assessment to action. Journal of the American Water Resources Association 2002;38(3):653-662. R827146 (2000)
    R827146 (Final)
    R825751 (2000)
    R825751 (Final)
  • Full-text: Oregon State University PDF
  • Abstract: Wiley InterScience
  • Supplemental Keywords:

    water, watersheds, ecological effects, ecosystems, restoration, terrestrial, aquatic, habitat, integrated assessment, decision-making, community-based, preferences, sociological, engineering, modeling, Pacific Northwest., RFA, Scientific Discipline, Geographic Area, Water, Ecosystem Protection/Environmental Exposure & Risk, Hydrology, Water & Watershed, Ecosystem/Assessment/Indicators, Ecosystem Protection, Restoration, State, Economics, Monitoring/Modeling, Ecological Effects - Environmental Exposure & Risk, Aquatic Ecosystem Restoration, Watersheds, Ecological Indicators, risk assessment, stakeholder groups, biodiversity, community involvement, Oregon, decision making, conservation, socioeconomics, cost benefit, integrated assessment, ecological recovery, web site development, aquatic ecosystems, GIS, water quality, economic, environmental rehabilitation, ecology assessment models, public policy, socioeconomic, community values, land use, stakeholder feedback, watershed restoration, OR

    Relevant Websites: Exit

    Progress and Final Reports:

    Original Abstract
  • 1999 Progress Report
  • 2000 Progress Report
  • 2001