Using TRMM observations to improve numerical simulations of precipitation within tropical cyclones

The potential of numerical weather prediction models for short-term forecast of tropical cyclones can be increased and monitored using microwave satellite observations.Improvements in cloud model microphysical parameterizations, as well as in radiative transfer models, is a necessary step to overcome some of the limitations of coupling satellite observations with forecast models, and of microwave precipitation retrieval algorithms. Our present goal is to find an optimal microphysics representation that can be used for future tropical cyclone simulations and forecasts.

We have completed a large cloud resolving forecast simulation of hurricane Bonnie using the University of Wisconsin-Nonhydrostatic Modeling System (UW-NMS). The simulation reproduced the cyclone remarkably well, the peak intensity was very close, and the storm made landfall within 100 km of the observations. However, we encountered major differences when comparing the precipitation structure to that implied by the Tropical Rainfall Measuring Mission (TRMM) overpasses. The comparison between the simulated upwelling brightness temperatures and reflectivies with the measurements of the TRMM microwave Imager (TMI) and Precipitation Radar (PR) revealed biases that suggested needs to alter the bulk microphysical formulations of the model. The constant-slope negative exponential size distribution parameterization assumed in the model seemed to be the major reason for the observed discrepancies.

Sensitivity studies of the simulated microwave appearance of the tropical cyclone and its precipitation structure to the bulk microphysics parameterization in the model have been carried out. The results have exposed several key issues concerning the model's representation of the sizes and densities of the hydrometeors, and of their implied microphysical processes. These issues include:

- the relevance of the correlations between ice particles and dynamics (such as updraft graupel correlations), and the impact ofthese factors on the mass of rain at the lower levels of the cloud and on the surface rain rates;

- the importance of the vertical sorting between large and small graupel particles within the clouds, due to their different terminal velocity, and its effect on the precipitation structure of the tropicalcyclone and on the simulated brightness temperatures;

- the use of a single category of graupel to represent both graupel developing from rimed ice and graupel growing from frozen raindrops;

- the use of a constant slope size distribution for rain;

Results of the implementation of alternative microphysical approaches designed to resolve these issues will be reported. Comparison to ground truth microphysical structure derived from in situ reconnaissance aircraft measurements of the CAMEX and TEFLUN-B field experiments will be discussed.