Resolution-Dependent Impact of Extratropical Cyclones on Winter U.S. Precipitation Bias in the GFDL SPEAR Model

Extratropical cyclones (ETCs) are the primary drivers of winter precipitation across the United States, accounting for up to 85% of total precipitation. This study uses the GFDL SPEAR models at atmospheric resolutions of 100 km, 50 km, and 25 km to examine how ETC dynamics impact precipitation patterns and biases across the United States. Higher-resolution models reduce ETC-related precipitation biases in the Southwest and Midwest but increase biases in coastal regions, including the West Coast and the Eastern United States. To understand these biases, we decompose ETC-related precipitation biases into those driven by precipitation frequency and intensity. Coastal precipitation biases are mainly due to overestimations of both the occurrence and intensity of precipitation, which are related to ETC frequency and intensity, respectively. In inland areas, biases are largely driven by occurrence bias associated with ETC frequency. Notably, higher-resolution models simulate amplified ETC frequency and intensity biases in coastal regions, while showing a decrease in ETC frequency bias in inland regions. This increase is especially linked to the overestimation of small-scale ETCs, which considerably inflate frequency-driven precipitation bias. Additionally, improvements in AMIP runs suggest that these biases are partly connected to SST bias. These findings emphasize the sensitivity of precipitation representation to ETC dynamics and underscore the importance of addressing resolution-dependent and SST related biases to improve midlatitude precipitation simulations in climate models.