Retrieval of land surface parameters using passive microwave measurements at 6-18 GHz

An approach is evaluated for retrieval of land surface parameters (soil moisture, vegetation water content, and surface temperature) using satellite microwave radiometer data in the 6-18-GHz frequency range, The approach is applicable to data that will be acquired by the Advanced Microwave Scanning Radiometer (AMSR), planned for launch on the Japanese Advanced Earth Observing Satellite (ADEOS)-II and Earth Observing System (EOS) PM-1 platforms in 1999 and 2000, respectively. The retrieval method is based on a radiative transfer (RT) model for land-surface and atmospheric emission, with model coefficients that can be tuned over specific calibration regions and applied globally. The method uses an iterative, least-squares algorithm, based on six channels of radiometric data. Simulations using this algorithm indicate that, for an assumed sensor noise of 0.3 K in all channels, soil moisture and vegetation water content retrieval accuracies of 0.06 g cm(-3) and 0.15 kg m(-2), respectively, should be achievable in regions of vegetation water content less than approximately 1.5 kg m(-2), A surface temperature accuracy of 2 C should be achievable, except for bare soils, where discrimination between moisture and temperature variability is difficult using this algorithm. These accuracies are for retrievals averaged over the sensor footprint, and they exclude conditions of precipitation, open water, snow cover, frozen ground, or high topographic relief within the footprint. The algorithm has been tested using data from the Nimbus-7 Scanning Multichannel Microwave Radiometer (SMMR) for the years 1982-1985, over the African Sahel, and the retrieval results compared to output from an operational numerical weather prediction model, The retrieval algorithm is shown to discriminate well between soil moisture, vegetation, and temperature variations and provide estimates consistent with the expected accuracies, Future work will investigate possible improvements to the algorithm using alternate formulations, possibly with fewer channels, and extend testing of the algorithm to other regions.