ASCAT Quality Control near Rain

The Advanced Scatterometer (ASCAT) onboard the Metop satellite series is designed to measure the ocean surface wind vectors globally. Generally, ASCAT provides wind products at excellent quality. The quality of the Advanced Scatterometer (ASCAT) derived winds is known to be generally degraded with increasing values of the inversion residual or maximum likelihood estimator (MLE). In the current ASCAT Wind Data Processor (AWDP), an MLE-based Quality control (QC) is adopted to filter poor-quality winds, which has proven to be effective in screening artifacts in the ASCAT winds, associated with increased sub-cell wind variability, notably under rain conditions. However, some poorly verifying winds, which appear in areas with convection, are not screened by the operational QC.
Identification of rain can help to better understand the impact of geophysical effects associated with rain on scatterometer wind quality, and to develop an improved QC approach for scatterometer data processing. In the first part of this report, an image processing method, known as singularity analysis (SA), is used to detect the geophysical effects associated with rain. The performance of SA for rain detection is validated using ASCAT Level-2 data collocated with satellite radiometer rain data. The rain probability as a function of SA-derived singularity exponent (SE) parameter is calculated and compared with other rain-sensitive parameters, such as the MLE. The results indicate that the SA is effective in detecting the presence of rain in ASCAT wind vector cells (WVCs). Moreover, SA is a complementary rain indicator to the MLE parameter, thus showing great potential for an improved scatterometer QC.
In the second part of the report, SA is proposed to complement the current ASCAT QC. The implementation of this new joint QC procedure is investigated, based on a comprehensive analysis of quality-sensitive parameters using the European Centre for Medium-range Weather Forecasts (ECMWF) model winds, the Tropical Rainfall Measuring Mission’s (TRMM) Microwave Imager (TMI) rain data, and tropical buoy wind and precipitation data as reference, taking into account their spatial and temporal representation. The buoy validation results show that the proposed method indeed effectively removes ASCAT winds in spatially variable conditions. It filters three times as many wind vectors as the operational QC, while preserving verification statistics with local buoys. Indeed, rain and wind variability as measured by the ASCAT SE appear well correlated.
Besides rain-induced large wind variability, which is shown to degrade the quality of ASCAT derived winds, no evidence of rain-contamination effects (e.g., rain splashing) have been found. Further analysis is required to exclude rain contamination for ASCAT.
Variable winds are a potential hazard in some applications, such as data assimilation, and the methods developed here may be useful for those applications. For other applications, such as nowcasting and oceanography it may be relevant to keep the flagged wind data since they provide essential information on (highly variable) air-sea interaction processes that cannot be spatially captured by any other wind observing system.