As part of the Pacific Landfalling Jets (PACJET) Experiment, rapid-scan satellite-derived cloud motion vectors are being assimilated in real-time into a 20-km version of the Rapid-Update Cycle (RUC) weather prediction model. These hourly wind-vectors are created from special 7.5 min visible and infrared GOES-10 satellite imagery as part of the GOES rapid-scan Winds Experiment (GWINDEX). The assimilation of these data provide a unique opportunity to assess their potential impact on high-resolution numerical predictions of landfalling winter storms emerging from the data-sparse Pacific Ocean region.

The cloud motion vectors are obtained by tracking cloud brightness features between subsequent satellite images. The enhanced temporal resolution provided by the GWINDEX rapid-scan strategy should lead to significant gains in both vector quality and quantity. Special automated postprocessing steps are used to ensure optimal observation height assignments and provide objective data quality control. Within the RUC PACJET domain, which extends approximately 1500 km offshore from the U.S. West Coast, a few thousand GWINDEX wind vectors are typically available each hour. These wind data, along with rawinsonde, profiler (404 and 915 MHz), RASS, VAD, ACARS, and surface data, are assimilated into the RUC model using a 3DVAR analysis procedure.

We will present a comparison of results for predictions made with and without the rapid-scan winds. Overall statistical comparisons will be augmented by analysis of specific cases, which highlight the forecast differences.