Moisture Source Tracking Techniques and Physical Process Analysis Revealing Precipitation Improvement Over North America in CESM2

Abstract The Community Earth System Model Version 2 (CESM2) represents a significant advancement over its predecessor in simulating global precipitation. However, the relative contributions of updated physical parameterizations versus higher model resolution have not been thoroughly investigated. Here, we focus on North America (NA), where CESM2 significantly reduces precipitation simulation biases from May to September during 2010–2019. Through comparative experiments conducted under MERRA‐2 nudging, updated physical processes notably reduce precipitation biases (>45% in regions of severe precipitation biases), which surpasses the contribution of higher resolution (<16%) over NA. Then, using the unique perspectives of atmospheric water tracer (AWT) and physical process analysis, we quantify the key drivers of precipitation improvement over NA in CESM2. The AWT results show that precipitation over NA originating from the northern Atlantic and northeastern Pacific has significantly decreased (by 9.5% and 20.7%, respectively), followed by a 25.0% reduction in local evaporation contribution. The updated cloud‐microphysical processes in CESM2 have weakened convective activity (40.6%) by reducing condensation latent heat release, leading to a decrease in convective precipitation over southern NA. This improvement has significantly reduced the precipitation bias relative to observational data. Further improvement is needed in CESM2's weak simulation of cloud‐macrophysical processes, which leads to a partly underestimated large‐scale precipitation in the mid‐to‐high latitudes of NA. Our findings underscore the critical role of cloud processes in improving simulated precipitation by optimizing moisture redistribution and condensation latent heat release within the nudged experimental framework.