Using TWP-ICE and MC3E observations to expose dynamical and microphysical causes of CRM and LAM simulated deep convection biases

Poster PDF

Description

Ten 3D cloud-resolving model (CRM) simulations and four 3D limited area model (LAM) simulations of an intense mesoscale convective system observed on January 23–24, 2006 during the Tropical Warm Pool-International Cloud Experiment (TWP-ICE) field campaign are microphysically and dynamically compared with each other and observational retrievals in an attempt to explain results in Varble et al. (2011) and other studies showing a high bias in convective radar reflectivity aloft. Whereas several previous studies have focused on redistributing ice mass into different hydrometeor categories or smaller sizes to combat this high bias, our results suggest that it may be as much due to excessively large ice water content as ice category or size. These large ice water contents are a product of large, deep, strong convective updrafts that are nearly undiluted in some cases, allowing large rain water contents to be lofted and frozen.

The highest 10 percent of simulated deep updraft maximum vertical velocities from the surface to ~10 km are comparable in magnitude to the highest dual-Doppler derived updraft vertical velocities at those levels. However, average simulated updrafts from the surface to ~10 km, and all simulated updrafts in the upper troposphere are substantially stronger than the dual-Doppler derived updrafts and other tropical oceanic observations near coastlines in previous studies. We speculate that the large upper tropospheric difference is due to freezing of much more condensate in the simulations than in reality.

Aircraft penetrations into tropical oceanic convection during previous field campaigns between temperatures of 0 to -10°C show a select few vertical velocities of 15–17 m/s and questionable near adiabatic condensate contents on par with those simulated, but at much finer spatial scales than can be resolved by 1-km horizontal grid spacing. While observations are far from extensive, the ones available suggest that high vertical velocities and condensate contents simulated by the CRMs and LAMs may be possible over tropical oceans, but typically at lower frequencies and finer spatial scales than can be resolved by the simulations.

This research is now being extended to LAM simulations of Midlatitude Continental Convective Clouds Experiment (MC3E) cases for which superior observational retrievals exist. Our focus continues to be on the interactions of convective dynamics and different microphysics representations including how these interactions impact stratiform region properties, such as rain rate, which was biased low in TWP-ICE simulations.