Spectral Decomposition Method for Large Sea Surface Generation and Radar Backscatter Modeling

Abstract

This paper analyzes different methods to simulate sea surface waves over a large area rapidly and with low computational complexity. Indeed, for wind speed between 1 and 10 m/s, the area of the sea surfaces must range from 10 to 92, 000m 2 to account for all the surface roughness scales which can contribute to the scattering process at microwave frequencies. At frequencies higher than 10 GHz, a sampling rate of one tenth of the wavelength can lead to a prohibitive numerical cost. The impact of these approaches on the surface power spectral density and on the monostatic normalized radar cross section is investigated. The proposed methods consist of splitting the full sea surface height spectrum into sub-spectra of smaller extents. Sub-sea surfaces are generated and combined from different interpolation and recombination techniques. In this paper, an original closed-form expression of the resulting sea surface height spectrum is derived to interpret the simulation results. Finally, the efficiency of the methods in terms of accuracy and memory requirement is analyzed by computing the monostatic normalized radar cross section from sea surfaces with the first-order small slope approximation (SSA1) scattering model.