Assimilation of clear sky water vapor information from AIRS data

Data assimilation system (DAS) of the NASA model GEOS-5 uses the GSI analysis solver developed at NCEP. For assimilation of the satellite brightness temperature, background information is interpolated to the observation location and radiative transfer code (CRTM) generates a pseudo-observation of brightness temperature needed for minimization. Current quality control of AIRS data discards satellite brightness temperature measurements from regions with clouds. The GOES-5 model often predicts clouds in locations that the satellite measurements consider cloud-free and therefore the calculated water vapor information is not correctly represented at such locations. In order to mitigate this problem, we adjust the background specific humidity, which represents water vapor information in the GOES-5 DAS model, by removing cloud humidity contribution. The new variable, clear sky specific humidity, is then used for assimilation of the AIRS data and impact of the correction is discussed.