Variational Retrievals Using SSM/I and SSM/T-2 Brightness Temperatures in Clear and Cloudy Situations

Abstract

The direct assimilation of brightness temperatures (Tb?s) from the Defense Meteorological Satellite Program Special Sensor Microwave Imager (SSM/I) and Special Sensor Microwave Water Vapor Profiler (SSM/T-2) in a one-dimensional variational (1D-Var) assimilation system is studied over the oceans. The control variables of the 1D-Var are the natural logarithm of specific humidity (lnq), near-surface wind speed (SWS), and cloud liquid water (CLW) path. Synthetic Tb?s, with and without noise added, were simulated and used to estimate the strength of the 1D-Var assimilation (weight given to the observations) and the information content of the Tb. In clear skies, it is shown that except for very dry profiles (TPW < 5?6 kg m?2) SSM/I Tb?s are superior to the SSM/T-2 Tb?s for the determination of total precipitable water (TPW). In the presence of clouds, the SSM/T-2 TPW retrievals are underestimated and the underestimation increases with CLW. Cloudy profiles should be filtered out. It is also shown that the SSM/I 1D-Var retrievals for low wind speeds are erroneous (<2?3 m s?1) and so are retrievals of low CLW (<0.1 kg m?2). Otherwise, SSM/I Tb retrievals of TPW, SWS, and CLW perform well. 1D-Var analyses using bias-corrected observed Tb?s were also computed for the same case as that using synthetic Tb?s and were compared with retrievals obtained using published regression equations. When both SSM/I and SSM/T-2 Tb?s were assimilated simultaneously (in collocation) in the 1D-Var system, the retrievals of TPW, SWS, and CLW were very similar to those when only SSM/I Tb?s were assimilated. However, the SSM/T-2 Tb also provided specific humidity information in the upper troposphere.