A numerical study of the composite surface model for ocean backscattering

A numerical study of 14-GHz backscattering from ocean-like surfaces, described by a Pierson-Moskowitz spectrum, is presented. Surfaces rough in one and two dimensions are investigated, with Monte Carlo simulations performed efficiently through the use of the canonical-grid expansion in an iterative method of moments. Backscattering cross sections are illustrated for perfectly conducting surfaces at angles from 0 to 60° from normal incidence, and the efficiency of the numerical model enables the composite surface theory to be studied in the microwave frequency range for realistic one-dimensional (1D) surface profiles at low wind speeds (3 m/s). Variations with surface spectrum low-frequency cutoff (ranging over spatial lengths from 21.9 to 4.29 cm) are investigated to obtain an assessment of composite surface model accuracy. The 1D surface results show an increase in hh backscatter returns as surface low-frequency content is increased for incidence angles larger than 30°, while νν returns remain relatively constant, all as predicted by the composite surface model. Similar results are obtained for surfaces rough in two dimensions, although the increased computational complexity allows maximum surface sizes of only 1.37 m to be considered. In addition, cross-polarized cross sections are studied in the two-dimensional (2D) surface case and again found to increase as surface low-frequency content is increased. For both 1D and 2D surfaces, backscattering cross sections within 20° of normal incidence are found to be well matched by both Monte Carlo and analytical physical optics (PO) methods for all low-frequency cutoffs considered, and a comparison of analytical PO and geometrical optics (GO) results indicates an appropriate choice of the cutoff wavenumber in the composite surface model to insure an accurate slope variance for use in GO predictions. This choice of cutoff wavenumber is then applied in the composite surface theory for more realistic ocean spectra and compared with available experimental data