Abstract Microwave imagers play a key role in numerical weather prediction (NWP), providing information about atmospheric humidity, temperature, cloud, and precipitation, as well as surface information such as skin temperature and sea ice. The atmospheric information has been directly assimilated from radiances for many years, but until recently much of the surface information has come to NWP systems through external retrievals and analyses. The aim now is to extract all available information within the NWP process, directly from radiances, using coupled data assimilation. The scientific benefits should include more complete and consistent use of the surface information in NWP. The practical benefits include the elimination of significant time delays between the microwave observations and when the information reaches the NWP system. The focus of the current work is to infer ocean surface skin temperature from microwave observations using a sink variable approach within the atmospheric data assimilation system, with the future aim of transferring this information to the ocean data assimilation. Improvements in the skin temperature are seen particularly in the region of tropical instability waves, and the updated skin temperature allows an improved simulation of the microwave brightness temperatures.
Scanlon et al. (Mon,) studied this question.