摘要：Extreme precipitation has widespread impacts over the western US (WUS), which are expected to be exacerbated in the future given thermodynamically driven increases in atmospheric moisture and shifts in atmospheric circulation patterns. However, internal climate variability complicates how these factors affect future projections of precipitation changes. In this study, we investigate late 21st-century changes in mean-state responses over the North Pacific Ocean and associated WUS precipitation extremes using the Multi-Model Large Ensemble Archive, which is a recent and valuable resource to disentangle the effects of model structural differences from internal variability. We found that inter-model differences in precipitation extremes over the WUS are large: from south to north (here, from California to Washington), models show different levels of increases for both the frequency and intensity of heavy precipitation. In general, the Pacific Northwest sees a more consistent signal than California. We show that wet-season mean circulation influences extreme precipitation, by using a decomposition of the relative contributions of thermodynamic (atmospheric humidity) and dynamic (large-scale circulation) driven changes. The thermodynamic contribution dominates, however, the dynamic contribution varies with latitude and differs substantially between model ensembles. Inter-model differences affect the spread in not only the magnitude, but also the sign of the dynamic contribution. This implies that mean circulation changes contribute to uncertainty in quantifying the regional structure of concentrated moisture flux and future extreme events. This result highlights the need to more accurately constrain projected North Pacific coupled circulation changes to narrow the spread in projections of future precipitation extremes. Plain Language Abstract In this study, we investigate how future warming regulates the mean-state changes of precipitation over the North Pacific Ocean and the associated precipitation extremes over western US (WUS) coastal regions. We use the recent Multi-Model Large Ensemble Archive, a valuable resource that provides the ability to account for both inter-model physical differences and internal climate variability in driving precipitation changes, which is especially important when considering relatively rare events such as precipitation extremes. This study contrasts multiple large ensembles to quantify changes to regional precipitation extremes and their relationship with the mean state. We revealed that, unlike the mean-state changes, future predictions of precipitation extremes over the WUS are associated with large uncertainty. When tracing the changes of moisture flux transport, it is found that models show enhanced moisture flux with a northward shift mostly and increased onshore flow along the WUS coast. We further evidenced that an improved understanding of the regional structure of circulation changes can have a substantial impact on the overall changes to extremes, which will be crucial for advancing risk assessments and estimating the potential risks of future extreme rainfall and flooding events regionally.
