Final Report: Development of an Urban Watershed Rehabilitation Method Using Stakeholder Feedback to Direct Investigation and Restoration Planning

EPA Grant Number: R827147
Title: Development of an Urban Watershed Rehabilitation Method Using Stakeholder Feedback to Direct Investigation and Restoration Planning
Investigators: Matlock, Marty D. , Kenimer, Ann L. , Neill, William H. , Peterson, Tarla Rai , Samuelson, Charles D. , Whitten, Guy D.
Institution: Texas A & M University
EPA Project Officer: Hiscock, Michael
Project Period: October 1, 1998 through September 30, 2001 (Extended to March 31, 2003)
Project Amount: $838,767
RFA: Water and Watersheds (1998) RFA Text |  Recipients Lists
Research Category: Water , Water and Watersheds


We developed and tested a method for restoring the ecological integrity of urban watersheds that combined ecology, engineering and social sciences. We tested the following hypotheses:

(1) Risk-based models developed for agricultural nonpoint source (NPS) pollution management can be applied to urban watersheds; and

(2) Stakeholders' understanding of (a) NPS pollution issues, (b) use of systems thinking, (c) the ability to use scientific information about Total Maximum Daily Loads (TMDLs) or rehabilitation options, and (d) communication competence will increase as a result of the Collaborative Learning (CL) intervention.

Baseline data on the San Antonio population were gathered through a survey of a random sample of that population. Stakeholder council members were recruited using both leads developed from the larger San Antonio survey and from face-to-face interviews with residents of the two watersheds, using a snowball sampling technique. Once councils were formed, the CL intervention used iterative input from stakeholders to guide risk-based research and restoration planning. Two integrated eco-indicators —periphyton and bluegill sunfish—were used to evaluate and communicate risk to the stakeholder groups. The effectiveness of the CL process was evaluated using both pre- and post-test surveys of council members and face-to-face interviews with council members at the conclusion of the CL process.

The project also integrated four modeling activities: (1) a geographic information systems-based riparian zone impact zone model; (2) a better assessment science integrating point and NPS- (BASINS) hydrolic simulation program-FORTRAN (HSPF) watershed model; (3) an Ecophys.Fish model; and (4) a STELLA watershed model, linking watershed policy and management decisions to watershed hydrologic response. In addition, a simulation of citizen behavior was developed to integrate human behavioral models with watershed models.

Summary/Accomplishments (Outputs/Outcomes):

Social Science

Collaborative Learning Program Effectiveness. Our discussion of the results focuses on the four social science hypotheses tested in this project. First, with respect to learning/knowledge outcomes, we found support for Hypothesis 2(a) on one of the six objective knowledge survey questions relating to NPS pollution. With respect to Hypothesis 2(b), we found solid support: stakeholders reported greater familiarity and ability to use system thinking principles in ecosystem management at the conclusion of the CL meetings compared to the beginning of the project. Hypothesis 2(c) received strong support on the self-report learning/knowledge items: council members reported that their knowledge of and ability to use scientific information in water quality management improved over the course of the CL program. Hypothesis 2(d), however, was not supported by the statistical analysis of the survey data. In terms of communication competence, stakeholders reported that the quality of communication at the group level (Salado Creek Restoration Council) decreased from pre- to post-test. At the individual level, we found either no change or a change in the opposite direction from Hypothesis 2(d).

With regard to interpersonal relationships, the results were clear: council members believed that relationship quality declined from pre- to post-test survey administration. There also was no change in trust in other council members over the course of the 14-month series of CL meetings. Overall, we found that stakeholder members more positively evaluated the CL format used in the meetings, compared to their previous experiences with other collaborative methods. Participants felt that their effectiveness in negotiating with other members of the Salado Creek Restoration Council declined from pre- to post-test.

Ancillary analyses of water quality beliefs revealed that, overall, participants changed their perceptions of water quality in Salado Creek in a positive direction: water was believed to be more safe for drinking, swimming, fishing, and livestock at post-test compared to at pre-test. This pattern of results was unanticipated based on existing CL theory, but is intriguing. It is possible that the increased knowledge about ecological principles and dynamics gained from the CL program may have stimulated stakeholders to revise their earlier, more negative perceptions of water quality in Salado Creek. Finally, analysis of personal efficacy beliefs showed no changes as a function of participation in the CL program.

In summary, the quantitative survey results suggest that learning about NPS pollution, TMDLs, and rehabilitation options did occur over the course of the CL meetings. There also is clear evidence that participants improved their knowledge about systems thinking in the context of ecosystem management. Thus, in terms of learning/knowledge outcomes, we conclude that the CL intervention was successful in achieving its objectives. Survey data from council members provided little support for the prediction that communication competence would improve with the use of CL. There also is no evidence in the survey results to support the assumptions that interpersonal relationships among stakeholders and negotiation effectiveness improved as a result of the CL program. Finally, it appears that the CL intervention did change beliefs about water quality in the Salado Creek watershed toward greater confidence in the safety of the water for human and animal uses.

The qualitative data from personal interviews with council members suggested a more positive conclusion about the effectiveness of the CL program, compared to the quantitative survey results. Many comments in the post-interviews reinforced the survey-based indicators that significant learning had taken place about NPS pollution, ecological principles and dynamics, and rehabilitation options. The interview data, however, also suggest that participants did improve their communication skills and develop new relationships with others over the course of the program. Agency representatives also were positively impressed with the quality of the CL meetings, and there is evidence from the interview data that the CL approach now is being used as a “model” for other watershed collaborative groups in San Antonio. Many interview comments focused on how the information gained during the project meetings helped the individual to become more active in other community groups and self-governance activities. We conclude that these qualitative results provide complementary, and in some respects, stronger support than the survey results for the original social science hypotheses.

Ecological Science. Ecological investigations included measuring ecological services changes in the San Antonio watersheds (Leon and Salado Creeks and the Upper San Antonio River) over the past 20 years, measuring the ecosystem health of Salado and Leon Creeks using periphyton, measuring the ecosystem health of Salado and Leon Creeks using bluegill sunfish, modeling fecal coliforms in Salado Creek using HSPF, assessing the impact of riparian zone loss on Leon Creek, and assessing the potential for ecosystem restoration using wastewater from the Dos Rios Wastewater Treatment Plant for supplemental flow. The complex suite of measurements employed in this project were effective indicators of ecosystem health, and provided insight into potential remediation strategies.

Ecological Services Change. Based on the estimated size of six land cover categories and Costanza et al.'s ecosystem services values for related biomes, we determined that the total annual ecosystem service values in Bexar County declined from $21.94 million per year to $21.16 million per year from 1976 to 1991. While our study showed that urban sprawl in Bexar County resulted in a decline in the value of ecosystem services delivered by the affected land, it also showed that changes in the value of ecosystem services over time depend on the interaction of changes in various land cover types. Our study suggests that urban spread may not necessarily lead to a large net decline in ecosystem services if there is a concomitant increase in size of other land cover types that provide a greater level of ecosystem services. This is not to say that urban sprawl is beneficial for the delivery of ecological services, but that the negative impacts of the spread of urban and suburban land uses can potentially be offset by other mitigating changes in land cover. This is important because it is unlikely that, with increasing human population pressure, the conversion of land to urban and suburban land use will cease or even dissipate in the near future.

Measuring Ecosystem Health With Periphyton. Eutrophication of streams and rivers is recognized as a growing threat to the health of aquatic ecosystems throughout the United States. Two studies in this project aimed to make periphyton analysis a more effective tool for addressing issues of eutrophication. The first study showed that nutrient delivery via diffusion elicits a greater periphytic algae response than nutrient enrichment via whole-channel dosing. The second study demonstrated the importance of riparian shading on reducing periphytic growth through shading, illustrating the connectivity between stream riparian habitat and aquatic bioprocesses.

Measuring Ecosystem Health With Bluegill Sunfish. This study measured the effects of in-stream environmental factors on bluegill sunfish physiological health and weight gain over 14 days. Metabolic "health" of the fish at the end of the cage trial accounted for 31 percent of variation in median weight change during the cage trial; physicochemical environment, independent of its effects on metabolic health, accounted for an additional 43 percent (74-31 percent); quantity or quality of feed accounted for yet another 15 percent (89-74 percent); leaving only 11 percent (100-89 percent) unaccounted for. Obviously, much of the variation both in median weight change and in median marginal metabolic scope was seasonal, and presumably temperature-related.

Modeling Fecal Coliforms in Salado Creek. The Salado Creek watershed in Bexar County, TX, was modeled using the HSPF model in BASINS. The model was calibrated for hydrology. A sensitivity analysis and first-order approximation were performed to determine the parameters that most influence the in-stream fecal coliform concentration predictions. The specific findings from the study include:

(1) The parameters to which peak in-stream fecal coliform concentrations are most sensitive are those that represent the maximum storage of fecal coliform bacteria over the pervious land segment (SQOLIM), and surface runoff that removes 90 percent of quality constituent from pervious land segment (WSQOP) of PERLND section (corresponding to pervious land segment) of HSPF model.

(2) Other parameters to which in-stream fecal coliform concentration predictions are sensitive are stream water temperature (TWAT), first-order decay rate of quality constituent (FSTDEC), and temperature correction coefficient for the first order decay rate (THFST) of RCHRES section.

These results point out the importance of parameterization in modeling with any complex, process-based watershed model. Small errors in assigning values to the maximum storage of fecal coliform over a given landuse class may result in large errors in predicted coliform counts.

Modeling Leon Creek Riparian Zone Decline. Evaluations of the riparian vegetation of Leon Creek suggest a fragmentation trend from 1987 to 1999 indicated by the Image Difference Calculation, Percent Area Calculation, and the landscape indices. As with most landscape-level assessments, however, the results are not always consistent. There is an inconsistent result in the rural zone, with the percent area showing increases in dense vegetation by 1999, which would be inconsistent with increased fragmentation. The most probable explanation is the conversion of land cover from arid desert vegetation to rural landscaped/irrigated vegetation. Both the percent area calculation, and landscape indices are consistent in the urban zone. Both the image difference calculation, and landscape indices are consistent in both the rural and urban zone, with the image difference calculation showing decreases in vegetation, which would be consistent with fragmentation shown by the landscape indices. When the Leon Watershed was divided into urban and rural zones, an increase in fragmentation level was shown between 1987 and 1999, especially in the urban zone.

Ecosystem Restoration Using Wastewater to Restore Base Flow. This study investigated the effect of increasing dry-period base flow in streams with discharge from a wastewater treatment plant (with pre-defined water quality parameters) on dissolved oxygen (DO) concentration in Leon Creek. An increase in daily mean DO was observed for all treatment scenarios tested. With 95 percent confidence level, a significant change was observed with scenario S3 (one times base flow), that is, doubling the base flow during low-flow periods. This constitutes a risk-based design approach to remedy DO problems by low-flow augmentation.

Modeling Activities

STELLA Watershed Model. A STELLA model was developed for the project as a tool for facilitating group learning among council members. The model was intended to illustrate how changes in watershed condition affect hydrologic response. While limited in its usefulness for hydrologic prediction or for modeling specific management plans, this model did allow stakeholders to view the relationship between selected strategies and runoff generation. It created an effective communication interface between scientists, citizens, and policy makers.

Simulation of Citizen Behavior. Using data from the survey of the general San Antonio population, we: (1) estimated a set of models of watershed-impacting behavior; and (2) produced simulations for the residents in areas adjacent to Salado and Leon Creeks based on what we found to be the most influential predictor variables from step (1). We then developed a statistical model to explain willingness to change that included such variables as environmental attitudes and assessment of local conditions. The simulations suggest that, of the existing attitudes/characteristics of San Antonio citizens, environmentalist views and a sense of being locally empowered are important predictors of willingness to change behavior in ways that would reduce NPS pollution. In terms of public information, the model indicates that if residents were convinced that changes in their behavior would actually improve the quality of the watersheds, they would be more likely to change.

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

Other project views: All 13 publications 7 publications in selected types All 2 journal articles
Type Citation Project Document Sources
Journal Article Grant WE, Peterson TR, Peterson MJ. Quantitative modeling of coupled natural/human systems: simulation of societal constraints on environmental action drawing on Luhmann's social theory. Ecological Modelling 2002;158(1-2):143-165 R827147 (Final)
not available
Journal Article Neill WH. Ecophys.Fish: a simulation model of fish growth in time-varying environmental regimes. Reviews in Fisheries Science 2004;12:233-288. R827147 (Final)
not available

Supplemental Keywords:

ecosystem, stakeholder, complex systems, restoration, collaborative learning., RFA, Scientific Discipline, Water, Geographic Area, Ecosystem Protection/Environmental Exposure & Risk, Water & Watershed, Restoration, State, Aquatic Ecosystem Restoration, Environmental Engineering, EPA Region, Watersheds, Social Science, stakeholder groups, social science research, community involvement, urban watershed rehabilitation method, decision making, community-based research, non-point source pollution, conservation, management alternatives, Region 6, Texas (TX), ecological recovery, integrated assessment, game theory, aquatic ecosystems, environmental rehabilitation, non-point sources, water quality, public policy, ecology assessment models, sociological, watershed restoration, aquatic habitat protection , community values, restoration planning

Progress and Final Reports:

Original Abstract
  • 1999 Progress Report
  • 2000 Progress Report
  • 2001
  • 2002