| Summary: | 2021 EUMETSAT Meteorological Satellite Conference, 20-24 September 2021, Bucharest, Romania Accurate high-resolution QuikSCAT-derived winds with dense coastal sampling are strategic for both scientific and civil applications. This study presents the implementation of the QuikSCAT spatial response function (SRF) analytical model and its parameterized version by means of a pre-computed look-up-table (LUT), provided by the Brigham Young University (BYU). Their validation and their differences are discussed in detail. In addition, the implementation of the land contribution ratio (LCR) computation and its visual validation are discussed. The slice ?0 noise is also carefully characterized and the measured Kp is compared to that provided in the QuikSCAT level 1b full resolution files (L1B). Different levels of ?0 are considered for the computation of the Kp, ranging from low to moderate and high wind speed regimes. The preliminary results show that the analytical and the LUT-derived SRFs are consistent with each other, even if some non-negligible differences are apparent. In particular, the LUT-derived SRFs have less spatial detail with respect to the analytical ones. Furthermore, analytical SRFs are generally not symmetric with respect to the slice centroid, while the LUT-derived ones are. The validation procedure of the LCR shows that the obtained values are consistent with the coastline. However, the differences between the analytical and the LUT-derived SRFs induce differences in the LCR values of up to a few percent. These differences are expected to impact the coastal wind field retrievals, but the investigation of this aspect is left for future studies. Both measured and L1B Kps decrease with increasing ?0 levels, and horizontally (HH) polarized acquisitions are noisier than those vertically (VV) polarized. The measured Kps show that the farther from the egg centroid, the noisier the slice ?0 , showing a rather symmetric parabolic trend. In contrast, the Kps provided in the QuikSCAT files do not show the same trend, i.e., the Kp is rather constant for slice indices higher than 2. The comparison among the eight available slices shows some inter-calibration issues, especially for HH acquisitions. The more distant the slices from each other, the higher the bias, which can reach 0.8 dB for HH acquisitions. This analysis also shows a non-negligible trend of the bias with respect to the antenna azimuth angle. Once more, the bias is more pronounced for HH acquisitions. Finally, the Kps computed over ocean are lower than those computed over land and sea-ice, where the Kp estimated value is more in particular correlated to the sea-ice fraction
|
|---|