Implementation of a state of the art automated system for the production of cloud/water vapor motion winds from geostationary satellites
The thrust of the proposed effort under this contract is aimed at improving techniques to track water vapor data in sequences of imagery from geostationary satellites. In regards to this task, significant testing, evaluation, and progress was accomplished during this period. Sets of winds derived from Meteosat data were routinely produced during Atlantic hurricane events in the 1993 season. These wind sets were delivered via Internet in real time to the Hurricane Research Division in Miami for their evaluation in a track forecast model. For eighteen cases in which 72-hour forecasts were produced, thirteen resulted in track forecast improvements (some quite significant). In addition, quality-controlled Meteosat water vapor winds produced by NESDIS were validated against rawinsondes, yielding an 8 m/s RMS. This figure is comparable to upper-level cloud drift wind accuracies. Given the complementary horizontal coverage in cloud-free areas, we believe that water vapor vectors can supplement cloud-drift wind information to provide good full-disk coverage of the upper tropospheric flow. The impact of these winds on numerical analysis and forecasts will be tested in the next reporting period.