Polarization Deconvolution and Geophysical Retrieval from a Dual-pol, Cross-track Scanning Microwave Radiometer (AMPR) During OLYMPEX/RADEX

The airborne Advanced Microwave Precipitation Radiometer (AMPR) has been supporting NASA field campaigns for more than two decades. AMPR is a downward-looking, multi-frequency, cross-track scanning instrument that measures the microwave brightness temperatures (TB) of geophysical phenomena like clouds and precipitation, as well as the land and ocean surfaces. In AMPR, the instrument polarization basis rotates with respect to that of the scene as a function of scan angle. Prior to 2011, AMPR was equipped with only one polarization channel per frequency and the unique determination of scene V- and H-pol TB was not possible. The dual-pol upgrade during 2011 made it possible to deconvolve AMPR measurements into scene V- and H-pol TB, thereby enabling improved passive microwave retrievals of precipitation rate (R), cloud liquid water (CLW), columnar water vapor (WV), and ocean surface wind speed (WS). Scene V and H polarization TB biases in dual-pol AMPR data have been observed, and are hypothesized to be separable into absolute and cross-scan asymmetry components. The absolute bias is typically a result of uncertainties in the measured calibration load TB, but the cross-scan asymmetry is a result of inaccurate assumption of the instrument measurement geometry and cross-pol contamination between the polarization channels. We derive a relationship between cross-pol level as well as various instrument and polarization angle errors that seems to explain the observed cross-scan asymmetry. Using a forward model with adjusted instrument geometry, improved polarization deconvolution is achieved. Resulting scene V- and H-pol TB data are input to an empirical retrieval algorithm to retrieve CLW, WV, and WS. The polarization deconvolution and retrieval method described above is applied to OLYMPEX/RADEX observations. Retrieved CLW, WV, and WS are compared to available satellite, airborne remote sensing, and in situ observations. Comparisons between the empirical retrieval, and an independent retrieval approach developed at Colorado State University, are made. The results indicate that AMPR is capable of resolving important physical processes occurring at scales finer than those observed by GMI and other satellite radiometers.