A CLEAN-based iterative nonlinear statistical rain rate retrieval algorithm

An iterative nonlinear statistical retrieval algorithm for estimating rain rate from multispectral and multiresolution passive microwave brightness temperature imagery of a convective oceanic storm is described. Both SHF and EHF radiometric channels are considered. The retrieval utilizes a combination of a non-iterative nonlinear statistical retrieval algorithm and the iterative CLEAN algorithm to "deconvolve" rain rates directly from simulated diffraction-limited satellite-based brightness temperature observations. Features of the CLEAN based retrieval algorithm are (1) that it allows the use of high spatial resolution brightness imagery to guide the retrieval of rain rate (which is most correlated with the spectral features found in the lower frequency and hence lower resolution channels) and (2) that the computationally intensive portion of the iterative process is constrained to a single brightness temperature calculation. The brightness temperatures are obtained using a planar-stratified scattering-based forward radiative transfer calculation for the channels and design characteristics of the proposed Multispectral Imaging Microwave Radiometer (MIMR). This radiative transfer model allows for five hydrometeor phases: suspended cloud water, rain, suspended cloud ice, dry snow, and dry graupel. The sensor observation simulation convolves antenna gain patterns appropriate for a sensor at a 705 km orbital altitude with a 1.6-m aperture with channels at 6.0, 10.69, 18.7, 23.8, 36.5, and 89.0 GHz.