Sea surface temperature estimation using active and passive microwave remote sensing

Although sea surface temperature (SST) is often estimated from satellite infrared observations, passive microwave techniques have the advantage of being applicable in cloudy regions. The goal of this research is to combine active and passive microwave measurements to improve the accuracy of the SST retrievals. A non-linear iterative inversion method for the retrieval of surface wind (U) and sea surface temperature (SST) using space-borne microwave scatterometer and radiometer measurements is developed and applied to data collected by the SeaSat satellite. This retrieval algorithm is physically based, using an integrated microwave model for normalized radar cross section ({dollar}sigmasb0){dollar} and surface emissivity and a nonlinear least squares estimation technique. The {dollar}sigmasb0{dollar} model is empirical, based on a SeaSat data set; While the emissivity model is physically-based. In the emissivity model the microwave reflectance R is estimated using the ocean wave spectrum of Bjerkaas and Reidel and a bistatic two-scale scattering model, and the microwave emissivity e is calculated as {dollar}e=1-R{dollar} where R is microwave reflectivity. The validity of this active and passive microwave algorithm is evaluated for the case of an eastern pacific test area employing 14.6 GHz SeaSat scatterometer (SASS) and 6.6 GHz V-pol radiometer (SMMR) data. The results show that SST can be reliably estimated from SeaSat SASS/SMMR measurements for an open ocean area. Two sets of SMMR cells were analyzed. The averaged test-site correlation coefficient between the reference SST (measured by ships) and the SeaSat-based estimates of SST is 0.8916 and the corresponding standard errors are smaller than 0.7{dollar}spcirc{dollar}C. These results suggest that a physically based SST retrieval scheme can be useful in improving the accuracy of SST measurements made with passive microwave remote sensing.