Precipitation profile retrievals using temperature-sounding microwave observations

[1] A Bayesian retrieval algorithm was developed to demonstrate the potential of microwave temperature-sounding channels for precipitation profile retrieval from spaceborne observations. The algorithm uses a database from combined cloud-radiative transfer model simulations of Hurricane Bonnie which was observed during the field campaign Convection and Moisture Experiment 3 (CAMEX-3). Sounding channels from two oxygen absorption complexes at 50 - 57 GHz and 118.75 GHz were combined to make use of their differential response to absorption and scattering by hydrometeors. The retrieval method was applied to airborne observations with the National Polar-Orbiting Operational Environmental Satellite System (NPOESS) Aircraft Sounder Testbed-Microwave (NAST-M) radiometer on board the ER 2 aircraft. The retrieved profiles were used to simulate radar reflectivities to be compared to ER 2 Doppler observations from the same aircraft. The results lead to the following conclusions: ( 1) The sounding channels provide the unique potential of cloud slicing because the channel-dependent variation of clear-air absorption allows the adjustment of maximum sensitivity to different altitudes. ( 2) The validation with radar data indicates a good performance of the algorithm; however, at least four channels in each absorption complex are required to constrain the retrieval well enough with the observations. ( 3) The validation also shows that even though a Hurricane Bonnie cloud model simulation was used, the simulated database does not represent the observations very well. ( 4) The database may bias the retrievals by the underlying assumptions on temperature, humidity, and hydrometeor distributions. Sounding channels are more sensitive to a possible temperature bias in the database. While this bias can be corrected, the retrievals forward the database bias to the retrieved products independent of sensor. For global applications, more flexible retrieval approaches are required that are capable of constraining the algorithm according to the local situation.