Shadowing by non-Gaussian random surfaces

The approach developed by Smith for estimating the shadowing properties of a Gaussian rough surface is applied to an arbitrarily distributed surface. Under the assumption that the surface height is statistically independent of the surface slopes, a very simple result is obtained. In particular, it is shown that the general bistatic shadowing function is completely determined by a single integration over the marginal probability density function for the surface slopes in the plane of incidence. From this general result it is immediately obvious that for backscatter the shadowing function is unity at normal incidence and zero at grazing incidence. The backscattering shadowing function is computed for an exponential slope density and compared to the Gaussian case.