Extreme rainfall, flood scaling and flood policy options in the United States

Abstract

River flood and rainfall have been shown to exhibit scale invariance behavior over a range of space and time scales. Although various approaches have been taken to investigate and model the various scaling aspects of rainfall and floods, little theoretical work has been done on the relation between the scaling of rainfall and flood. If available, such a theory would provide frequency estimate for extreme rainfall and floods outside the range of observations and could also be used to estimate floods at ungaged basins. The relationship between rainfall and flood scaling is the main focus of this thesis. We use a two step approach to investigate the relationship between exponent of peak flows and the scaling of rain. First, we use data analysis to verify existing theories that relate the multi scaling behavior of rainfall to the simple scaling behavior of the IDFs. Second, we use a model to relate the scaling of the IDFs to the scaling of peak flows with basin area. We find that, although temporal rainfall shows multiscaling, the IDFs exhibit simple scaling and peak floods show simple or mild multiscaling. We validate our findings by using U.S. peak annual flow data and rainfall from a few New England stations. Extreme floods damage mitigation requires sound and integrated policy making. We review the flood disaster mitigation situation in the U.S., carry out policy analysis and recommend options for a successful and sustainable flood disaster policy in the U.S.