Simulating Mixed-Phase Clouds Over Coastal Antarctica During a Significant Snowfall Event in a High-Resolution Regional Model

Abstract

Global climate models and reanalysis products have revealed large, persistent downwelling shortwave radiation biases over the Southern Ocean and coastal Antarctica, likely caused by the incapability of models to accurately simulate frequent low-level mixed-phase clouds in these regions. In this study, we use the ground-based observations collected at Davis, Antarctica during the Precipitation over Land and The Southern Ocean field campaign in austral summer of 2019 to assess the capability of the high-resolution regional Unified Model (UM) to reproduce precipitating clouds off coastal Antarctica. We test the new UM RAL3 (Regional Atmosphere and Land 3) configuration with double-moment Cloud AeroSol Interacting Microphysics scheme and bimodal cloud fraction scheme, running at the spatial resolution of 1.5-km. We compare it to the previous RA2M configuration with a single-moment cloud microphysics scheme and unimodal cloud fraction scheme. The RAL3 exhibits marginally degraded meteorological conditions relative to RA2M compared with observations. For cloud properties, the UM regional models can generally simulate the phase, vertical structure and timing of events during the sublimation and precipitation periods. Nevertheless, overestimated ice water path and potentially underestimated liquid water path (LWP) contribute to positive surface shortwave biases and negative longwave biases. The RA2M simulates more LWP, though we suggest for the wrong reasons due to its ice nucleating parameterization. Our results suggest that the new double-moment cloud microphysics combined with bimodal cloud fraction parameterizations, while having reduced performance in some respects, has large potential to better represent low-level mixed phase clouds for this region.