Stochastic Modeling and Risk Management for Water Resources Systems Under Changing Climatic Conditions

Abstract

Water resources are indispensable for the sustainable development of the human society. A variety of hydrological modeling and water resources management tools based on simulation and optimization have been developed to address the current water issues worldwide. However, there are many challenges arising from climate change, human disturbances and enormous uncertainties and complexities. Thus, there is a global need for advanced methodologies that can support the modeling and management of water resources systems in an effective and efficient way. In this dissertation research, a spectrum of methods have been developed to deal with the stochastic modeling and risk-based management problems for water resources systems. These methods include: (i) a Stepwise Clustered Hydrological Inference (SCHI) model that can establish the complex nonlinear relationships between climatic conditions and streamflow for hydrological forecasting; (ii) a flexible and effective hydro-climatic modeling framework based on the Providing Regional Climates for Impacts Studies (PRECIS) modeling system and stepwise cluster analysis for hydrological modeling under the changing climatic conditions; (iii) a Stepwise-cluster-analysis-based Probabilistic Collocation Expansion (SPCE) method for the stochastic simulation and forecast of hydrologic time series; (iv) a hydrologic frequency analysis framework based on change point analysis and Bayesian parameter estimation to deal with the nonstationarity and uncertainties in hydrological risk analysis; (v) an Interval-parameter Two-stage Fuzzy Stochastic Integer Programming (ITFSIP) model for risk-based flood diversion management under multiple uncertainties. The proposed methods have been applied to the Xiangxi River Watershed in China and the Grand River Watershed in Canada, in order to demonstrate their capabilities and performances in precipitation-runoff modeling, climate change impact analysis, uncertainty quantification, frequency analysis, and systematic water resources and risk management. The major contribution of this research lies in the development of innovative approaches for tackling various uncertainties and complexities in the hydrological cycle and water resources systems. This research can provide scientific and practical bases for robust hydrological modeling and reliable water resources management.