A generalized simulation capability for rotating-beam scatterometers

Rotating-beam wind scatterometers exist in twotypes: rotating fan-beam and rotating pencil-beam. In ourstudy, a generic simulation frame is established and verifiedto assess the wind retrieval skill of the three different scat-terometers: SCAT on CFOSAT (China France Oceanogra-phy SATellite), WindRad (Chinese Wind Radar) on FY-3E,and SeaWinds on QuikSCAT. Besides the comparison of theso-called first rank solution retrieval skill of the input windfield, other figures of merit (FoMs) are applied to statisti-cally characterize the associated wind retrieval performancefrom three aspects: wind vector root mean square error, am-biguity susceptibility, and wind biases. The evaluation showsthat, overall, the wind retrieval quality of the three instru-ments can be ranked from high to low as WindRad, SCAT,and SeaWinds, where the wind retrieval quality strongly de-pends on the wind vector cell (WVC) location across theswath. Usually, the higher the number of views, the betterthe wind retrieval, but the effect of increasing the number ofviews reaches saturation, considering the fact that the windretrieval quality at the nadir and sweet swath parts stays rela-tively similar for SCAT and WindRad. On the other hand, thewind retrieval performance in the outer swath of WindRad isimproved substantially as compared to SCAT due to the in-creased number of views. The results may be generally ex-plained by the different incidence angle ranges of SCAT andWindRad, mainly affecting azimuth diversity around nadirand number of views in the outer swath. This simulationframe can be used for optimizing the Bayesian wind retrievalalgorithm, in particular to avoid biases around nadir but alsoto investigate resolution and accuracy through incorporatingand analyzing the spatial response functions of the simulatedLevel-1B data for each WVC.