A multi criteria decision support system for watershed management under uncertain conditions

Abstract

Nonpoint source (NPS) pollution is the primary cause of impaired water bodies in the United States and around the world. Elevated nutrient, sediment, and pesticide loads to waterways may negatively impact human health and aquatic ecosystems, increasing costs of pollutant mitigation and water treatment. Control of nonpoint source pollution is achievable through implementation of conservation practices, also known as Best Management Practices (BMPs). Watershed-scale NPS pollution control plans aim at minimizing the potential for water pollution and environmental degradation at minimum cost. Simulation models of the environment play a central role in successful implementation of watershed management programs by providing the means to assess the relative contribution of different sources to the impairment and water quality impact of conservation practices. While significant shifts in climatic patterns are evident worldwide, many natural processes, including precipitation and temperature, are affected. With projected changes in climatic conditions, significant changes in diffusive transport of nonpoint source pollutants, assimilative capacity of water bodies, and landscape positions of critical areas that should be targeted for implementation of conservation practices are also expected. The amount of investment on NPS pollution control programs makes it all but vital to assure the conservation benefits of practices will be sustained under the shifting climatic paradigms and challenges for adoption of the plans. Coupling of watershed models with regional climate projections can potentially provide answers to a variety of questions on the dynamic linkage between climate and ecologic health of water resources. The overarching goal of this dissertation is to develop a new analysis framework for the development of optimal NPS pollution control strategy at the regional scale under projected future climate conditions. Proposed frameworks were applied to a 24,800 ha watershed in the Eagle Creek Watershed in central Indiana. First, a computational framework was developed for incorporation of disparate information from observed hydrologic responses at multiple locations into the calibration of watershed models. This study highlighted the use of multiobjective approaches for proper calibration of watershed models that are used for pollutant source identification and watershed management. Second, an integrated simulation-optimization approach for targeted implementation of agricultural conservation practices was presented. A multiobjective genetic algorithm (NSGA-II) with mixed discrete-continuous decision variables was used to identify optimal types and locations of conservation practices for nutrient and pesticide control. This study showed that mixed discrete-continuous optimization method identifies better solutions than commonly used binary optimization methods. Third, the conclusion from application of NSGA-II optimization followed by development of a multi criteria decision analysis framework to identify near-optimal NPS pollution control plan using a priori knowledge about the system. The results suggested that the multi criteria decision analysis framework can be an effective and efficient substitute for optimization frameworks. Fourth, the hydrologic and water quality simulations driven by an extensive ensemble of climate projections were analyzed for their respective changes in basin average temperature and precipitation. The results revealed that the water yield and pollutants transport are likely to change substantially under different climatic paradigms. And finally, impact of projected climate change on performance of conservation practice and shifts in their optimal types and locations were analyzed. The results showed that performance of NPS control plans under different climatic projections will alter substantially; however, the optimal types and locations of conservation practices remained relatively unchanged.