Spatiotemporal Characteristics of US Floods: Current Status and Forecast Under a Future Warmer Climate

Abstract

Floods in the US exhibit strong spatiotemporal variability, mainly controlled by precipitation types and catchment attributes. In a future warmer climate, it remains largely unclear how such features change, relative to the current climate. Global Climate Models at coarse resolution cannot resolve convective-scale storms which is one of the major weather systems causing devastating floods in the US. Alternatively, coupled, high-resolution and continental-scale climate and flood simulations can advance our understanding. In this study, we couple a 4-km convection-permitting climate simulation model with a 1-km hydrologic model to simulate the spatiotemporal variability of rainfall and flooding over the contiguous US. In particular, changes in rainfall and flood frequency, spatial scale, and seasonality are explored in major climate divisions. We found: (a) an overall increase in flood frequency (+101.7%) and spatial extent (+44.9%), mainly attributed to more extreme rainfall and variability in the future; (b) weakening rainfall and flood seasonality, resulting in more random and unpredictable events throughout the year; (c) earlier flooding season onsets in the West and snow-dominated regions while delayed onsets in the East driven by drier antecedent soil moisture; (d) correlation between extreme rainfall and flood onsets is becoming stronger in the West yet weaker in the East in the future. Findings in this study can potentially serve as a basis for future flood exposure and risk assessments, as well as more scientific understanding of changing flood-generating mechanisms across the CONUS.