Evaluating shallow convection parameterizations with DOE-ARM measurements and high-resolution simulations during a cold-air outbreak

Description

Mixed-phase boundary-layer convective (BLC) clouds occur rather frequently in autumn along the North Slope of Alaska (NSA) during cold-air outbreaks. Unlike the Sc/Cu BL clouds over subtropical oceans, these BL clouds are mixed-phase and produce rather heavy precipitation. For more than a decade the U.S Atmospheric Radiation Measurement (ARM) program has been operating a rich array of profiling and scanning radars, lidars and other sensors at the NSA, to document the vertical structure of the lower troposphere, as well as clouds and precipitation. It is challenging to accurately represent such BLC in global and even regional numerical models, because most shallow convection parameterizations do not represent the sub-grid scale effects of mixed-phase microphysics, cloud-radiation interaction and precipitation. This study will present results from a series of WRF simulations of a BLC episode at the ARM NSA site: a high-resolution (900-m in horizontal) one with convection represented explicitly and three coarser resolution (13.5-km) ones with different shallow convection schemes. Analyses show that the 900-m WRF simulation compares well to the ARM observations in terms of themodynamic and wind profiles, BLC mesoscale organizations, and vertical transects of vertical velocity, reflectivity and liquid water path. This 900-m simulation is then used to evaluate how well shallow convection parameterizations represent BL structure and clouds in coarser resolution simulations. Results show that the three 13.5-km runs with different shallow convection schemes differ significantly in BL and cloud characteristics, due to fundamental differences in their assumptions, approach and implementation in WRF.