Characterization of water-vapour retrievals from TOVS/HIRS and SSM/T-2 measurements

This paper studies the retrieval characteristics of water-vapour retrievals from infrared TOVS (TIROS Operational Vertical Sounder) radiances and microwave SSM/T-2 (Special Sensor Microwave/Water-vapour Sounder) radiances. An optimal-estimation retrieval scheme is built around a radiative-transfer model that uses a spectral-band model based on the Malkmus line distribution for the infrared and the MPM92 transmission model for the microwave. Characteristics from synthetic retrievals show that the sensitivity to water vapour in the lower troposphere is low for TOVS and SSM/T-2 over land, but high for SSM/T-2 over ocean. The sensitivity to water vapour in the middle and upper troposphere is high for TOVS, but low for SSM/T-2. The vertical resolution of the retrievals is restricted to a few kilometers for both instruments. The dependence on a priori information is therefore substantial, especially in the lower troposphere. A global retrieval with real data from the TOVS instrument and a priori data from the National Centers for Environmental Prediction re-analysis constrained by an extra total-column constraint from SSM/I(Imager) data shows that the retrieval scheme is able to produce monthly-mean fields similar to other data on monthly-mean precipitable-water fields and layered information about column water vapour. The approach presented allows for a useful error characterization that quantifies the dependence on the a priori information. In general, the best performance of the retrieval scheme is achieved in the tropical upper troposphere.