Retrieval of Atmospheric Water Vapor and Temperature Profiles over Antarctica and Greenland

The current atmospheric water vapor and temperature profile retrievals are particularly challenging over land surfaces due to uncertainties in the estimation of the background surface emission. We present the iterative atmospheric retrieval model that resolves the surface emissivity and results in more accurate atmospheric profiles. We show retrieved atmospheric water vapor and temperature profiles utilizing Ka- to G- bands observations onboard AMSU and ATMS instruments on Aqua, SNPP, and JPSS.

The Greenland and Antarctica ice sheets are major and increasingly important contributors to the global sea level rise through the melting of the ice masses. Thus, monitoring and understanding melt events are more important than ever. However, the fact that the emissivity of land surfaces is higher than the emissivity of sea surfaces makes atmospheric retrieval more challenging over land. Over the polar ice sheets, the problem is further compounded by the generally dry atmosphere, which means the atmospheric contributions are low. In contrast, the background emission of the ice sheet surface is relatively high.

We coupled the atmospheric retrieval algorithm with the background emission algorithm in an iterative loop to improve the accuracy of the atmospheric profiles. The atmospheric temperature and humidity profiles were retrieved by optimal estimation using the sounding channels (Ka- to G- band) available on AMSU and ATMS. The algorithm used time and space co-incident analysis fields from MERRA-2 (Modern-Era Retrospective analysis for Research and Applications, Version 2) as the prior for the pressure, temperature, water vapor, and liquid water profiles. We estimated the atmospheric temperature from the ~50 GHz channels and moisture from the ~183 GHz (high water vapor absorbing channel) channels. Upon completion of the atmospheric retrieval algorithm, the atmosphere profiles are updated, which will be used by the surface retrieval in the next loop. The looping continues until the atmospheric retrieval converges. The novelty of the algorithm is the coupling of the surface retrieval with the atmospheric retrieval allowing enhanced accuracy. We used radiosonde data to assess the performance of the retrieval.