Behavior of an Inversion-Based precipitation Retrieval Algorithm with High-Resolution AMPR Measurements Including a Low-Frequency 10.7-GHz Channel

Abstract

A microwave-based, profile-type precipitation retrieval algorithm has been used to analyze high-resolution passive microwave measurements over an ocean background, obtained by the Advanced Microwave Precipitation Radiometer(AMPR) flown on ANASA ER-2 aircraft. The analysis is designed to first determine the improvements that can be gained by adding brightness temperature information from the AMPR low-frequency channel (10.7 GHz) to a multispectral retrieval algorithm nominally run with satellite information at 19, 37, and 85 GHZ. The impact of spatial resolution degradation of the high-resolution brightness temperature information on the retrieved rain/cloud liquid water contents and ice water contents is then quantified in order to assess the possible biases inherent to satellite-bawd retrieval. The tests are conducted on a dataset obtained during a preliminary flight experiment that took place on 18 October 1990 over a Gulf of Mexico squall line that developed south of the Florida Panhandle. Careful inspection of the high-resolution aircraft dataset reveals five distinctive brightness temperature features associated with cloud structure and scattering effects that are not generally detectable in current passive microwave satellite measurements. Recovery of such high-resolution information by satellites would generally be expected to improve precipitation retrieval, but these improvements have never been quantified and thus are addressed in this study. Results suggest that the inclusion of 10.7-GHz information overcomes two basic problems associated with three-channel retrieval. First, unresolved rainfall gradients in the lower cloud layers due to 19-GHz blackbody saturation effects are recovered when the 10.7-GHz channel data are included. Second, unrealistic oscillations in the retrieved rain liquid water contents that arise from the highly variable scattering signatures at 19, 37, and 85 GHz are eliminated by virtue of the 10.7-GHz Rayleigh frequency probing into the lower cloud containing the bulk of the liquid water. Intercomparisons of retrievals carried out at high-resolution and then averaged to a characteristic satellite spatial scale to the corresponding retrievals in which the brightness temperatures are first convolved down to the satellite scale suggest that with the addition of the 10.7-GHz channel, the rain liquid water contents will not be negatively impacted by special resolution degradation. That is not the case with the ice water contents as they appear to be quite sensitive to the imposed scale, the implication being that as spatial resolution is reduced, ice water contents will become increasingly underestimated. The overall implications of this study in the context of the upcoming United States?Japan Tropical Rainfall Measuring Mission are that the inclusion of a 10.7-GHz frequency on the passive microwave radiometer and the relatively higher spatial resolution of the low and intermediate frequencies at 10.7, 19, and 35 GHz resulting from the relatively low orbit (? 350 km) will lead to significantly improved microwave-based rainfall measurements over what are currently available today.