Variational Assimilation of Humidity Using SSM/I and SSM/T-2 Brightness Temperatures

Recently, a number of NWP centers have studied and/or implemented the variational assimilation of SSM/I (Special Sensor Microwave/Imager) retrieved total precipitable water (TPW) in 3D and 4D-var systems (4D-Var=4D variational assimilation). The assimilated TPW was obtained using either a 1D-Var system or a regression equation applied to the brightness temperatures. In both cases, the vertical distribution of the humidity increments is determined by the background error covariances. It is thus important to also assimilate brightness temperatures from humidity sounders such as the DMSP SSM/T-2 and/or the NOAA AMSU-B. Both instruments sample water vapor absorption in the 183 GHz line.

A new set of more accurate predictors for the water vapor optical depth at the 183 GHz frequencies was found for the radiative transfer model RTTOV. The 1D-Var code developed by ECMWF (based on RTTOV) for the SSM/I was adapted to handle (1) more than one set of predictors for the water vapor absorption, (2) mixed polarization, (3) a variable scan angle and (4) instruments with different view angles. A comparison of 1D-var humidity retrievals over the oceans from synthetic and observed F14 SSM/I and SSM/T-2 brightness temperatures in clear-sky and cloudy conditions will be presented. Biases of the modeled versus observed brightness temperatures as a function of scan position will also be discussed.