Global Sensitivity of Precipitation Extremes to Cloud Microphysics
A superparameterized (SP) climate model contains within each grid column a high-resolution model capable of simulating individual clouds. Similar to a previous study by the authors focusing on the contiguous United States (CONUS), the representation of cloud microphysics within such a model is found to impact precipitation extremes globally. It does this by altering large-scale atmospheric motions that then feedback onto rain rates on long time scales (~1 year). A new result, however, is that the choice of microphysics can directly impact local rainfall extremes on short time scales (~5 days) in the tropics.
Precipitation extremes are important for agriculture, infrastructure, and water storage. Despite this, it is difficult for climate models to simulate their correct magnitude as compared with observations. SP models have shown improvement by being able to better simulate clouds. But, there are still choices to be made on how to represent the small-scale processes that govern the lifecycle of cloud droplets and ice crystals. It is shown that this choice leads to statistically significant differences in extreme precipitation. Regardless, the modeled rain rates lie between those from two sets of observations.
As climate models with resolution high enough to represent clouds become more commonplace, it is natural to wonder if this is enough to accurately simulate precipitation extremes. Microphysics processes, which include the formation and breakup of liquid droplets and ice crystals, remain unable to be directly simulated by nature of their small spatial scales. Thus, their net effect on clouds must be modeled indirectly. We investigate multiple ways of doing this within the SP version of the Community Atmosphere Model (CAM) and the resultant effect on rainfall extremes. While feedbacks onto large-scale weather patterns are important globally as they were in CONUS, significant local effects manifest in the tropics. We determine the cause to be stronger vertical air motions using a 2-moment microphysics scheme as compared to a 1-moment one. Finally, we compare SP-CAM's output to two commonly used datasets from the Global Precipitation Climatological Project (GPCP) and the Tropical Rainfall Measuring Mission (TRMM). We find that, despite the statistically significant differences documented above, both microphysics schemes lead to precipitation extremes that fall in between those from the two sets of observations.