Sensitivity of convective cell dynamics and microphysics to model resolution for lake-effect shallow convection

Shallow convection over an unfrozen lake (Lake Ontario) during a cold-air outbreak is simulated using the Weather Research and Forecasting model (WRF) with two horizontal grid spacings, 148 m and 1.33 km. The dynamics and microphysics of the simulated convective snow band are compared to radar and aircraft observations. The dynamical and microphysical changes that occur when going from 1.33-km to 148-m grid spacing are explored. Improved representation of the convective dynamics at higher resolution leads to a better representation of the microphysics of the snowband compared to radar and aircraft observations. Stronger updrafts in the high-resolution grid lead to larger ice nucleation rates and produce ice particles that are more heavily rimed and thus faster falling. These changes to the ice particle properties in the high resolution grid limit aggregation rates and result in more realistic radar reflectivity patterns. Graupel, observed at the surface, is produced in the strongest convective updrafts, but only at the higher resolution. Ultimately, the quantitative precipitation forecast is improved at a higher grid resolution. Additionally, the duration of heavy precipitation just onshore, where convection collapses, is better predicted.