Detecting and projecting changes in U.S. precipitation extremes in the GFDL SPEAR large ensemble

Abstract:

Extreme precipitation events are predicted to become heavier and more frequent due to climate change, and these changes are an important factor to consider when updating infrastructure. Currently, official precipitation frequency estimates are available through the NOAA Atlas-14 database, but these historic estimates are based on analysis of point-based observations alone. Observational data across the U.S. is limited as it is only available over a relatively short time period that varies by location. However, this limitation can be overcome by using gridded model data, which also adds the advantage of being able to incorporate climate change scenarios into precipitation risk analysis. Here, a high-resolution global climate model is used to analyze how U.S. 24-hour precipitation extremes at various return periods change over the 1921-2100 time period. Using the 50-km horizontal atmospheric resolution global GFDL SPEAR 30-member ensemble, we quantify extreme precipitation risks across the U.S. and locally under different climate change scenarios (SSP5-8.5, SSP2-4.5, and natural forcings alone). With the large ensemble, we also quantify the uncertainty of characterizing extreme precipitation risk historically and in future projections.