ASCAT wind quality under high subcell wind variability conditions

Abstract

The assessment and validation of the quality of satellite scatterometer vector winds is challenging under increased subcell wind variability conditions, since reference wind sources such as buoy winds or model output represent very different spatial scales from those resolved by scatterometers (i.e., increased representativeness error). In this paper, moored buoy wind time series are used to assess the correlation between subcell wind variability and several Advanced Scatterometer (ASCAT)‐derived parameters, such as the wind‐inversion residual, the backscatter measurement variability factor, and the singularity exponents derived from an image processing technique, called singularity analysis. It is proven that all three ASCAT parameters are sensitive to the subcell wind variability and complementary in flagging the most variable winds, which is useful for further application. A triple collocation (TC) analysis of ASCAT, buoy, and the European Centre for Medium‐range Weather Forecasting (ECMWF) model output is then performed to assess the quality of each wind data source under different variability conditions. A novel approach is used to compute the representativeness errors, a key ingredient for the TC analysis. The experimental results show that the estimated errors of each wind source increase as the subcell wind variability increases. When temporally averaged buoy winds are used instead of 10 min buoy winds, the TC analysis results in smaller buoy wind errors (notably at increased wind variability conditions) while ASCAT and ECMWF errors do not significantly change, further validating the proposed TC approach. It is concluded that at 25 km resolution, ASCAT provides the best quality winds in general.

Evaluation of HY-2A Scatterometer Wind Vectors Using Data from Buoys, ERA-Interim and ASCAT during 2012–2014

“…Regarding the wind direction at low wind speeds, H2SCAT (similar to QuikSCAT) showed a decentralized distribution in the residual analysis, which is attributed to the incompleteness of the ambiguity removal algorithm using the median filter technique [34], together with errors in the geophysical model at low wind speeds [13]. Many factors including rain, increased local wind variability, confused sea state (waves), and land/ice contamination might induce poor quality of wind retrievals [44]. The H2SCAT showed its strongest performance at moderate speeds, especially in range 4-10.8 m¨s´1 range (see Table 6).…”

Toward an Improved Wind Quality Control for RapidScat

“…NWP model (e.g., ECMWF) wind output can be collocated at each WVC by interpolating the model forecasts both spatially and temporally to the scatterometer acquisition location and time, ensuring sufficient collocations over a short period. However, the lowresolution NWP model does not well resolve the wind field at scatterometer scales, i.e., about 25 km, particularly in the presence of rain or high wind variability [7], [11]. Buoy winds are generally accurate.…”

“…Buoy winds are generally accurate. Under high subcell wind variability conditions, the mean buoy wind vectors derived from a series of 10-min discrete buoy measurements have proven to be the most accurate wind source at 25-km scales [11]. However, buoy data are too scarce to assess the mentioned indicator sensitivity to rain and data quality.…”

“…The singularity exponents (SEs) derived from SA also provide quantitative information about the sub-and/or inter-WVC variability, and is a complementary QC indicator to MLE. The combination of SE and MLE has been proposed to improve the ASCAT wind QC in [10], and to classify and quantify the ASCAT wind uncertainty as a function of the wind variability, based on a triple collocation (TC) analysis of ASCAT, buoy, and the European Centre for Medium-Range Weather Forecasting (ECMWF) model output [11]. For Ku-band scatterometers, rain is the most important source of wind quality degradation, and rain detection is thus very relevant [12]- [14].…”

Evaluation of Marine Surface Wind Speed Observations From AMSR2 on GCOM-W Satellite

“…However, this assumption is not always valid. Triple collocation analysis [25], [26] has been utilized to explicitly evaluate the errors in the reference data. By using the collocated datasets analyzed in the previous sections, triple collocation was applied to quantify the random errors in the AMSR2 and …”