PNNL

Abstract

Regional climate simulations over the continental U.S, were conducted for the 2011 warm season using the Weather Research and Forecasting model at convection-permitting resolution (4km) with two commonly used microphysics parameterizations. (Thompson and Morrison). Sensitivities of the simulated MCS properties and feedbacks to large-scale environments are systematically examined against high-resolution geostationary satellite and 3-D mosaic radar observations. MCS precipitation including precipitation amount, diurnal cycle, and distribution of hourly precipitation intensity are reasonably captured by the two simulations despite significant differences in their simulated MCS properties. In general, the Thompson simulation produces better agreement with observations in MCS upper level cloud shield and precipitation area, convective feature horizontal and vertical extents, and partitioning between convective and stratiform precipitation. More importantly, Thompson simulates more stratiform rainfall, which agrees better with observations than does Morrison and results in top-heavier heating profiles from robust MCSs. A stronger dynamical feedback to the large-scale environment is therefore seen in Thompson, wherein an enhanced mesoscale vortex behind the MCS strengthens the synoptic-scale trough and promotes advection of cool and dry air into the rear of the MCS region. The latter prolongs the MCS lifetimes in the Thompson relative to the Morrison simulations. Hence different treatment of cloud microphysics not only alters MCSs' convective-scale dynamics, but also has significant impacts on their macrophysical properties such as lifetime and precipitation. As long-lived MCSs produced 2-3 times the amount of rainfall compared to short-lived ones, cloud microphysics parameterizations have profound impact in simulating extreme precipitation and the hydrologic cycle.