| description abstract | Experimental wind datasets were derived for two time periods (13?20 July and 24 August?10 September 1995) from GOES-8 observations processed at the University of Wisconsin Cooperative Institute for Meteorological Satellite Studies (UW CIMSS). The first dataset was focused on Tropical Storm Chantal, and the second dataset was focused on the multiple-storm environment that included Hurricanes Humberto, Iris, and Luis. Both datasets feature a processing and quality control strategy designed to optimize the quantity and content of geostationary satellite-derived winds in the vicinity of tropical cyclones. Specifically, the winds were extracted from high-density targets obtained from multispectral imagery, which included three water vapor bands (6.7, 7.0, and 7.3 ?m), infrared, and visible. The Navy Operational Global Atmospheric Prediction System (NOGAPS) was used as the vehicle to determine the impact of these winds upon tropical cyclone track forecasts. During the 1995 Atlantic hurricane season the NOGAPS forecasts were found to be quite skillful, displaying relative improvement of tropical cyclone position error with respect to CLIPER (climate and persistence) of 20% at 24 h, 35% at 48 h, and 33% at 72 h. The NOGAPS data assimilation system was run with and without the high-density GOES-8 winds for the two aforementioned time periods. The assimilation of these winds resulted in significant improvements in the NOGAPS forecasts for Tropical Storm Chantal and Hurricane Iris and mixed results for Hurricanes Humberto and Luis. Overall, for all four cyclones, the NOGAPS forecasts made with the use of the UW CIMSS winds displayed relative improvement of forecast position error with respect to those made without the use of the UW CIMSS winds of 14% at 24 h, and 12% at both 48 and 72 h. | |
