A better understanding of ice water content and liquid water content is important to improve our knowledge of cloud and precipitation processes and to validate microphysical schemes in numerical simulation models. In the former work by the authors, by investigating the variability of ice and liquid precipitation contents estimated by using some existing algorithms (radar only or radiometer only) in hurricane eyewalls and rainbands for a 1-year Tropical Rainfall Measuring Mission (TRMM) tropical cyclone database, a large variable range of estimations of total ice and liquid water mass was shown among those algorithms, with a high estimation by the TRMM 2A12 algorithm (radiometer only, Kummerow et al. 1996) and an extremely low estimation by algorithms developed from TRMM 2A25 (radar only, Iguchi et al. 1998) output.

This paper describes an iterative retrieval algorithm that combines the use of active and passive observations to estimate the vertical profiles of hydrometeors and parameters of exponential particle size distributions (PSDs). By using radar-derived hydrometeor profiles as the input to a microwave forward radiative transfer model, the differences between observed brightness temperatures and simulated ones can be minimized by adjusting the PSD interception coefficients N0r, N0s, N0g (for rain, snow, and graupel respectively). The rain, snow, and graupel fraction profiles for different rain types are assumed based on recent microphysical measurements in field programs. Airborne Doppler radar on the NASA ER-2 (EDOP) and Advanced Microwave Precipitation Radiometer (AMPR, 10, 19, 37, and 85 GHz) observations from the 4th Convection and Moisture Experiment (CAMEX-4) are used for the first test. High resolution observations are degraded into lowest resolution (AMPR 10 GHz). Nadir profiles are estimated for convective and stratiform regions associated with Hurricane Humberto (2001). Convergence can be reached quickly for any strong rain region (convective or stratiform), especially for the 10 and 85 GHz channels. For anvil and warm rain regions, there are some difficulties to get convergence for 19 and 37 GHz brightness temperatures. As compared with profiles calculated from existing algorithms, a good agreement between the new algorithm and Black (1990)’s Ze-IWC relationship can be seen in ice regions, but in rain regions the new algorithm is close to TRMM 2A12’s statistical profiles. Future work will focus on refining the cloud classification technique by adding Doppler velocity criteria, improving the assumption of rain, snow, and graupel fraction profiles by combining observations and 3D cloud model simulation results, and applying the new algorithm into TRMM satellite data.