Contents 1. Introduction
Sea near-surface wind is an important factor for the studies of oil spill propagation, ship detection, marine weather prediction, etc. In particular, wind speed and wind direction are the crucial parameters for the calculation of radar scattering from sea surface, since it influences significantly the formation of wave surface roughness. Among available wind sources, i.e. marine buoy data, numeric weather prediction model, Synthetic Aperture Radar (SAR) is a favorite one to obtain wind parameters, since it can offer the data at a high spatial resolution and in most meteorological conditions. For wind speed retrieval, in spite of many proposed studies in recent years, i.e. Doppler centroid [1], azimuth wavelength cut-off [2], the scatterometry approach [3]–[6] is still one of the most widely used methods, since it can estimate quite accurately wind speed from 2 to 25 m/s. For this approach, radar backscattering or normalized radar cross section (NRCS) is described as a function of wind speed, wind direction, radar incidence angle and radar look, namely Geophysical Model Function (GMF). Once wind direction is determined and the geometry of observation is known from SAR data, wind speed the 10-m reference height can be estimated by inverting the GMF. In C-band, one can find the GMF series, namely CMOD functions [3]–[6], which are developed based on observations from spaceborne microwave scatterometers (ERS-1/2). Therefore, they are also called the empirical (EP) GMF.
