An inversion technique for estimating the wind-dependent short wave spectral density from the CMOD4 and composite surface models

Calculations of the normalized radar cross-section (NRCS) from the CMOD4 model are inverted using the composite surface model (CSM) to estimate the short wave spectral density in the C-band region of the short wave spectrum. Simulations of the composite surface model are used to train an artificial neural network. Short wave spectral densities from a variety of short wave models are used with the CSM. The simulated NRCS from the CSM is used as the input parameter and the short wave spectral density is used as the output parameter. The properly trained neural network is then used with the CMOD4 model to generate the corresponding short wave spectra at different wind speeds. These wave spectra are compared with predictions of existing empirical wave spectral models and appear to agree well in shape. However, the actual spectral levels were quite different