Defense Presentation Title: Computationally Efficient Approaches to Modeling Flood Risk and River Discharge at Scale

Co-Advisors: Dr. Ross Meentemeyer, Director, Center for Geospatial Analytics; Dr. Georgina Sanchez, Research Scholar, Center for Geospatial Analytics

Abstract: River discharge—the volume of water flowing through Earth’s rivers—is highly variable in space and time and at its extremes can cause damaging floods or inadequate water supply. Traditionally, in-situ measurement gauges have been used to understand dynamics in river discharge and anticipate its impacts, but their sparse global distribution leaves extensive data gaps. Most approaches to fill in data gaps are labor intensive and computationally inefficient, limiting their scalability to larger geographic extents and to alternative scenarios of environmental change. Through three computationally novel modeling approaches, my dissertation aims to advance seamless understanding of water distribution at scale by (1) leveraging river modeling to accurately estimate discharge and water storage globally, (2) applying machine learning to estimate flood damage probability across the United States and (3) combining river and terrain modeling to efficiently compute flood hazard under climate change.