To allow cloud initialization of numerical weather prediction models over a wider geographical region, design and testing is underway for direct assimilation of satellite imagery (starting with GOES) within the hourly updating Rapid Refresh (RR) to improve its short-range cloud forecasts for aviation, severe weather and other users.

This effort builds upon the generalized cloud analysis package developed for the current hourly updated Rapid Update Cycle (RUC) and extended for the Rapid Refresh regional assimilation system (Weygandt et al. AMS Sat Conf). Within the operational RUC 3DVAR assimilation, the cloud analysis package utilizes NESDIS GOES satellite derived cloud top pressure and temperature data, as well as surface METAR multi-level cloud and visibility information. Within the Rapid Refresh system (currently in real-time testing at NOAA/ESRL/GSD as a planned replacement for RUC in 2010), the cloud analysis has been integrated into the Gridpoint Statistical Interpolation (GSI) analysis software. In addition to use of satellite products in the cloud analysis, the GSI analysis includes a full satellite radiance assimilation procedure (see related presentation by Devenyi et al).

The generalized RUC/RR cloud analysis already remaps GOES cloud-top temperature to a cloud-top pressure using the background 1-h RUC/RR forecast of 3-d temperature, water vapor mixing ratio, and hydrometeors (cloud water, rain, snow, ice, graupel) and 1-h surface forecast of skin temperature. This design is applicable to use of IR and near-IR imagery data as part of the same forecast/satellite/METAR cloud analysis. Initial plans and development in this new project sponsored by the JCSDA will be described in this paper.