Physical retrieval of sea-surface temperature from INSAT-3D imager observations

This paper presents results of physical retrieval of sea-surface temperature (SST) from the INSAT-3D imager observations. Radiance measurements from two thermal infrared imager channels are used for retrieving SST. Prior to their use in the retrieval process, the INSAT-3D radiances are corrected for the biases using a newly developed algorithm. The physical retrieval of SST is performed using an optimal estimation method. The optimisation is performed under two scenarios. In the first scenario, the SST retrieval problem is assumed to be linear while in the second one, the assumption of linearity is relaxed. Further, retrievals are performed by considering the reduced, as well as, the full state vector. Overall, physical retrieval performs much better than the regression-based approach. Moreover, in the physical retrieval, the best performance is obtained in the second scenario with the use of the full state vector. Validation of these retrievals with the iQuam SST resulted in a standard deviation (bias) of similar to 0.65 (0.19) K while those from the regression-based method are found to have standard deviation (bias) of similar to 0.68 (-0.17) K.