On the Quality Control of HY-2 Scatterometer High Winds

Abstract

The operational wind processor for the Ku-band scatterometers onboard HY-2 satellite series uses a quality control (QC) scheme based on the maximum likelihood estimator (MLE). Since it is difficult to discriminate rain contamination from “true” high winds, the MLE-based wind QC is set in a conservative way, which rejects up to ~35% of high winds (w ≥ 20 m/s) in HY-2 scatterometers (HSCATs). In this paper, the sensitivity of MLE and its spatially averaged value (i.e., MLEm) to wind quality and rain is reconsidered by analyzing the collocated HSCAT observations and buoy data, as well as rain data from the global precipitation measurement satellite’s microwave imagers. It shows that MLEm is more effective than MLE in terms of flagging rain data. More interestingly, the HSCAT high winds are much less strongly affected by rain, compared to the prior Ku-band pencil-beam scatterometers (e.g., RapidScat). Consequently, a MLEm-based approach is proposed to improve the HSCAT wind QC, particularly for high winds. The new QC method results in ~8% rejections at 20 m/s and above. Compared to the collocated buoy winds, the HSCAT high winds preserved by the new QC (but rejected by the operational QC) are of fairly good quality.