Enhancing Extreme Precipitation Predictions With Dynamical Downscaling: A Convection-Permitting Modeling Study in Texas and Oklahoma

Precipitation in the Southern Plains of the United States is relatively well depicted by the Community Earth System Model (CESM). However, despite its ability to capture seasonal mean precipitation anomalies, CESM consistently underestimates extreme pluvial and drought events, rendering it an insufficient tool for extending simulation lead times for exceptional events, such as the abnormally dry May 2011, which helped drive Texas into its worst period of drought in more than a century, and the abnormally wet May 2015, which led to widespread flooding in that state. Ensemble-based regional climate experiments are completed for the two extreme years using Weather Research and Forecasting model (WRF) and downscaled from CESM. WRF simulations are at convection-permitting grid resolution for improved physical representation of simulated precipitation over the Southern Great Plains. By integrating convection-permitting models (CPMs) into each individual member of a CESM ocean data assimilation ensemble, this study demonstrates that high-resolution dynamical downscaling can improve model skillfulness at capturing these two events and is thus a potentially useful tool for forecasting extremely high and extremely low precipitation events at subseasonal or even seasonal lead times. The Community Earth System Model (CESM) has reasonable climatological representation over the Southern Plains region of the United States, but it is still a challenge for CESM to capture the extreme events like severe floods and drought. In this study, we applied a regional climate modeling approach onto CESM at a model resolution capable of resolving convection for selected record-breaking wet and dry years in Texas. Results indicate our modeling approach can improve CESM's capability to simulate both wet and dry extremes and the modeling system skill is reasonable using CESM simulations initialized weeks before the events' occurrence. Different sources of upper atmospheric zonal winds may lead to abnormally wet or dry seasonal precipitation in Texas. Dynamical downscaling could improve the performance of fully coupled low resolution climate model in simulating climate conditions over land Consistent improvements from convective-permitting modeling in the depiction of precipitation for extreme Texas flood and drought years Improved regional circulation pattern depiction, combined with convective-permitting modeling, leads to enhanced precipitation simulation