Validation of QuikSCAT wind retrievals in tropical cyclone environments

QuikSCAT wind speed retrievals in over a dozen tropical cyclones are validated against Stepped Frequency Microwave Radiometer (SFMR), wind speed analyses (H*Wind), and global positioning satellite (GPS) dropwindsonde observations. Four QuikSCAT retrieval products are evaluated: 25.0 km resolution Near Real Time (NRT), 12.5 km NRT, 2.5 km Ultra Hi-Resolution, and an Ultra Hi-Res product empirically adjusted for hurricane environments.

Tropical cyclones (TCs) present a difficult challenge for active scatterometry due to heavy rain and wind speeds beyond the upper design limits of retrieval algorithms. Rain effects have qualitatively been well documented but little validation work has been performed due to difficulty in obtaining reliable surface truth measurements in TCs. This study exploits SFMR and GPS dropwindsonde observations taken from National Oceanic and Atmospheric Administration (NOAA) research and Air Force Reconaissance aircraft from within tropical cyclones. In addition, comparisons are also made to H*Wind analyses. H*Wind is a data assimiliation and analysis tool that produces a gridded (~6 km resolution) surface wind speed snapshot. Recent Atlantic Basin tropical cyclone seasons were searched for collocations of QuikSCAT passes with available SFMR and GPS Dropwindsonde data.

Results will be presented that show a consistent low bias in the QuikSCAT retrievals near the tropical cyclone core. This suggests that attenuation of the signal by rain dominates over any enhanced backscatter. Retrievals outside of the tropical cyclone core are remarkably accurate. A statistical comparison of the four retrieval algorithms shows that the low resolution NRT product has approximately a 40% lower root mean square error overall. However, there are large differences in algorithm performance from case to case.