The performances of the more complex explicit moisture scheme options in the fifth generation Penn State / National Center for Atmospheric Research Mesoscale Model (MM5) were evaluated in tropical cyclone forecasts using observed microwave brightness temperature observations from Special Sensor Microwave Imager (SSMI) instruments. MM5 forecasts were run at high horizontal resolutions (6 and 2 km horizontal grid spacings) so that cumulus parameterizations were not used. A fast and accurate radiative transfer model (RTM) was used to transform the output of the MM5 into brightness temperatures which could then be compared with SSMI observations. Adding a mixing formula to the RTM in the calculation of the dielectric constant for ice-air mixtures greatly reduced the difference between the lowest brightness temperatures produced by the model and those which were observed, especially at 85 GHz in areas of high precipitating ice concentrations. Furthermore, to increase the consistency between the model and the RTM the slope and intercept parameters of the inverse exponential function used to estimate the drop size distribution of the precipitating particles (rain, snow, and graupel) in the RTM was set to match the values used in the different explicit moisture schemes which produced the input for the RTM.

Preliminary results show that the MM5 tends to produce too much ice (which is evident in cooler model produced brightness temperatures at 85 GHz) no matter which explicit moisture scheme is used when the horizontal grid spacing is set at 6km. A more detailed analysis as well as results from smaller horizontal grid spacings will be presented at the conference.