Context and objectivesIn this study, a physically based two-scale model was used to compute the radar responses to the sea surface in non-fully developed state. The "non-fully developed state" term refers to the surface waves system. It is indeed usually assumed that waves are generated by wind, and that the wind only,' is sufficient to determine completely the sea surface characteristics. When this is the case, the sea state is said to be fully developed. However, it is easy to understand that other waves sources can exist, needing then to take their contribution into account in the wave model. Therefore the first objective is to study the relevant sea surface statistical information in term of the sea spectrum.
This spectrum is then to be used as input for the scattering model that was analysed.
The classical development leading to the expression of the backscattering coefficient is recalled and then, several improvements are to be brought to this model. The method was further validated. For nadir incidence, we used TOPEX C and Ku Bands altimeter (nadir looking radar system) measurements. ERSI scatterometer (off-nadir looking radar) data, were used to validate the two-scale method under large ranges of incidence angles and wind speeds and directions.
Finally, consequences for a synergetic inversion procedure combining TOPEX and ERS-1 data were next inferred. An inversion algorithm is implemented, allowing the retrieval of the wind vector, as usual empirical methods, but also of another very interesting ocean surface parameter, the significant slope. The inversion procedure was run for the different sea spectra analysed in the first chapter. Several comparisons were made and inverted values of the wind vector are confronted to assimilated ground truth data set.
Expected scientific results
A surface model has been developed for non-fully developed state. This model converges to a fully developed one for limit conditions. The electromagnetic model used has also been completely re-analysed, criticised and improved whenever possible. The method has further been validated for nadir and off-nadir incidence. Finally, an inversion algorithm has been developed to retrieve both the wind vector and the significant slope. The inversion procedure has been run and the results confronted to ground truth data. The method has been showed globally efficient.