Sea Wind Extraction From RADARSAT-2 and Scatterometer Data Over the Gulf of Mexico

• Published in: IEEE geoscience and remote sensing letters 2017-07, Vol.14 (7), p.1007-1011
• Main Authors: Fraga Rodrigues, Douglas, Landau, Luiz, Rebelo Torres Junior, Audalio, Lanzillotti, Regina, Pellon de Miranda, Fernando
• Format: Article
• Language: English
• Subjects: Anemometers; Bootstrap method; CMOD-IFR2; Coefficients; Correlation; Correlation coefficient; Correlation coefficients; Data; Extraction; Extreme values; Fields; Fluid flow measurement; Gulf of Mexico; Mathematical models; Modes; Oils; Outliers (statistics); Radar measurements; Radarsat; RADARSAT-2; Resampling; Resolution; Scatterometers; Sea measurements; Sea surface; Spatial resolution; Synthetic aperture radar; Wind direction; Wind estimation; Wind fields
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2017.2692044

A marine wind field extraction methodology that produced results with 1.5 × 1.5 km of spatial resolution using RADARSAT-2 wide and ScanSAR narrow acquisition modes was applied over the Gulf of Mexico. Bootstrap method with median values was applied on the spatial resampling step and showed to be efficient to eliminate extreme σ 0 values from double bounce by targets on the water, which are undesirable for wind estimation, without the need of postprocess to exclude outliers. An empirical CMOD-IFR2 model was used with wind direction information to initialize the model obtained from QuikSCAT scatterometer. The estimated values were compared with measurements from four anemometers. Modeled wind fields were resampled at the same QuikSCAT spatial resolution with the bootstrap method, and correlation coefficients were found to be greater than 0.75 for all ScanSAR narrow images. The initial spatial resolution probably plays an important issue on the results, as the correlation coefficients for all wide images were almost zero. This hypothesis must be further investigated.