Assessing Hydrological and Water Supply Drought Interactions Across the Contiguous U.S. Through Integrated Hydrological and Reservoir Operation Modeling
Advisor: Professor Ximing Cai
ABSTRACT
Drought, which can be as serious as a billion-dollar natural disaster, propagates from meteorological to hydrological drought, ultimately ending with severe socioeconomic consequences. In the Anthropocene era, drought propagation and impact have become increasingly complex due to intensive interactions between hydrological and human systems. This dissertation addresses this complexity by examining the relationship among hydrological drought, reservoir storage condition (a major human interference to streamflow), and water supply drought in river basins with significant storage regulation, employing an integrated hydrological and water management modeling approach.
The dissertation is structured in three interconnected parts. The first part develops a novel machine learning model to extract real-world reservoir operation rules, which provides a realistic reservoir representation for large-scale hydrological simulation at the river basin scale. A transparent model structure uncovers real-world water management patterns for major reservoirs across the contiguous United States (CONUS). Building on this data-driven reservoir operation model, the second part establishes an integrated modeling framework to simulate the role of reservoir operation in drought propagation from hydrological drought to water supply drought across CONUS river basins. Key findings include spatially heterogeneous trade-off impacts in term of drought duration versus intensity, caused by reservoir operation on hydrological drought, reflecting various roles of reservoir operation in drought management. Based on these insights, a new water supply drought index is developed, incorporating both streamflow and reservoir storage levels via a power-low relation. The third part digs into individual drought events, examining the dynamic interaction between hydrological drought and reservoir storage drought. This analysis reveals some typical patterns of the interactions between and in/out phases of hydrological drought and reservoir drought, and their spatial distribution across the CONUS, reflecting the various influences of regional specific climate, reservoir capacity, and water requirement factors.
With improved human dimension representation in drought propagation through realistic representation of reservoir in large-scale hydrologic modeling, this dissertation bridges a critical gap in our understanding of how anthropogenic factors modulate natural drought processes. The research establishes a coupled hydrological and water management modeling framework, fusing data-driven reservoir operation modeling with process-based hydrological modeling. This integrated approach enhances our ability to understand, predict, and manage water supply droughts in the context of complex human-natural systems.