Satellite-derived ocean surface winds provide an unprecedented view of winds over the global oceans. Some of these ocean surface wind data are used operationally (SSMI, QuikSCAT, ASCAT) in global and regional models, but the relative impact of these data sets is still largely unexplored for many data assimilation systems and forecast models, particularly for derived data products like seal-level pressure (SLP) and gradient level winds (GLW). In this study, coordinated assimilation experiments examine the impact of ocean surface wind data on forecasting hurricanes and mid-latitude weather systems by both global and regional models (GEOS5 and WRF, respectively).

Cycling data assimilation experiments using combinations of ocean surface wind data sets with GEOS5 are used to supply first guess fields and lateral boundary conditions for regional assimilation and forecast experiments with the WRF. A series of NCEP global analyses with and without the assimilation of QuikSCAT data are also used to supply ICs and BCs for the WRF to assess the impact of QuikSCAT data on regional scales from assimilation in an operational global model. Specifically, we have examined impacts on forecasts of hurricanes Katrina and Ophelia (2005). We will compare and contrast results from the global experiments (GEOS5) and the regional experiments (WRF), using forecast hurricane track and intensity errors as the primary metrics to assess the relative data impacts. Traditional RMS error and anomaly correlation scores are used to measure mid-latitude impacts. QuikSCAT-derived SLP and GLW data have shown a great deal of promise for hurricane initialization and our study explores this with tandem GEOS5/WRF experiments.