Modeling of Precipitation-Impacted Microwave Radiances for Application to Assimilation in Numerical Weather Prediction Models

NWP models routinely assimilate satellite microwave sensor observations in non-precipitating environment.  Recently, in part due to the availability of combined active/passive space-based measurements (TRMM PR/TMI, GPM DPT/GMI) much effort has been focused on attempting to extract useful information from observations made in precipitation affected areas.

In this work, we set requirements on fast microwave forward models for use in retrieval and assimilation applications by assessing the impact on modeled brightness temperatures of uncertainties in assumed cloud optical properties, types of hydrometeors included in the calculations (solid ice, rain and melted particles) and vertical and horizontal inhomogeneities in rain and ice distributions within a microwave sensor FOV and by examining the effects of forward model errors on the quality of retrieval products. This assessment is performed within the framework of a One-Dimensional Variational Analysis first applied to MetOp-A AMSU/MHS measurements simulated using realistic hydrometeor distributions derived from co-located TMI/PR observations augmented with MetOp-A AVHRR-derived information about vertical and horizontal extent of non-precipitating ice and liquid clouds (to which radars are little sensitive). Ways of including radar information to constrain the inversion of the passive measurements are proposed.  Verification of the findings with real data is discussed.