Evaluating Cloud Microphysical Parameterizations in COAMPS-TC for Tropical Cyclone Intensity and Structure Prediction

It is becoming increasingly evident that tropical cyclone (TC) intensity and structure forecasts are very sensitive to representation of microphysical processes in numerical weather prediction models. Development and testing of new and different microphysical parameterizations in COAMPS-TC have been underway at the Naval Research Laboratory (NRL) to improve TC forecasts. In this presentation, we discuss the results from several different versions of microphysics schemes from both the research and operational communities. Two of the microphysical parameterizations that have been tested are: NRL-modified Rutledge and Hobbs scheme currently in COAMPS-TC, and the Thompson scheme (released in WRF V3.3). In a case study of Hurricane Earl (2010), we see large differences (by orders of magnitude) in hydrometeor distribution in the storm inner core region with the different schemes. For example, the Thompson scheme, a partial 2-moment scheme, reduces the upper-level cloud ice by 1-2 orders of magnitude relative to the current version of COAMPS microphysics scheme (control run). This reduction of cloud ice helps mitigate a warm bias at the upper levels associated with the latent heating release during the ice formation and also due to the cloud-radiation interaction. In contrast, the Thompson scheme generates 3-4 times more snow in a layer from 500-200 hPa. The storm intensity as well as size differs significantly in simulations using the Thompson and NRL microphysics, resulting in different pressure-wind relationships. We are also testing a new NRL microphysics scheme that reduces the upper-level cloud ice bias. Additional simulations of a large number of TC cases at a 5-km grid increment are being performed with these and other microphysics schemes for systematic evaluation with the ultimate goal to obtain a better understanding of the microphysical processes and their implication for TC forecasts.