Satellite microwave humidity sensors such as the Advanced Microwave Sounding Unit (AMSU) on the NOAA polar orbiting satellites have great potential to improve the global analysis of atmospheric moisture profiles. A great challenge is to retrieve moisture over land, where the variable surface emissivity hinders the retrieval of the atmosphere. An optimal estimation, simultaneous retrieval of water vapor profile, temperature profile, cloud liquid water and surface emissivity from AMSU observations has been developed at CIRA to attack this problem. The system is called the CIRA 1-Dimensional Variational Optimal Estimator (C1DOE).

A global matchup radiosonde and AMSU dataset has been created and used to evaluate the retrieval performance. Preliminary comparisons of retrieved moisture profiles to the radiosonde profiles indicate that the retrieval provides improved fields of water vapor, especially above the 700 hPa surface. A comparison of retrieved total precipitable water (TPW) to the highly accurate Global Positioning System (GPS) TPW fields over CONUS show that C1DOE resembles the spatial structure of the GPS TPW. The retrievals will be demonstrated in clear sky and the impact of cloud liquid water will be discussed.

The sensitivity of this retrieval to formulations in the land surface emissivity first guess, the number of retrieved variables, and covariance and error of the retrieved variables is presented. Comparison of the simulated radiances to the AMSU data show that for water vapor the specification of levels in the model affects the error, especially at the critical 183 GHz channels.

A discussion of the potential of this work to lead into assimilation of moisture into forecast models will be presented.