This presentation focuses on improving our understanding of how the Gridpoint Statistical Interpolation (GSI) data assimilation (DA) scheme handles wind vectors derived from QuikSCAT, and how changes in the analyses modify the predicted structure of tropical cyclones and their environment.

The primary data used in this study are two parallel NCEP GFS analysis-forecast cycles, run during the 2005-8 Atlantic hurricane seasons using the operational versions of the GFS (T382 L64 resolution) and GSI DA scheme. The two cycles are identical except that one includes all QuikSCAT data in the assimilation, while the other excludes all the QuikSCAT data. The influence of assimilating QuikSCAT data on tropical cyclone track forecast errors has so far been found to be small on average, although some interesting individual cases exist. In this presentation, we will elucidate the means by which the NCEP GFS is handling the QuikSCAT data, and will offer recommendations on how TC forecasts may derive benefit from data collected by future scatterometers.