A comparative study of several composite surface models used to produce a wind estimation algorithm using NSCAT scatterometer data

A theoretical model function for the normalized radar backscattering cross section (NRCS) of the ocean surface used in a wind retrieval algorithm using NSCAT data is presented. The model is based on the two scale roughness model proposed by G.R. Valenzuela [1978]. The NRCS varies with both wind speed and direction when measured at moderate incidence angles. Dependence of the NRCS and wind field is incorporated into the model via an ocean wavenumber spectrum. The present authors have tested the performance of the model using four ocean wave spectra proposed in the literature: M. A. Donelan et al. [1987], T. Elfouhaily et al. [1997], J. R. Apel [1994] and R. Romeiser et al. [1997]. They have found that the overall performance of all four spectra is within rms experimental error for VV polarization. The mathematical complexity of the models proved to be of great importance in the computational calculation