Numerical scattering from 3D randomly rough surfaces using FVTD

A numerical technique for calculating the scattering from 3D randomly rough surfaces has been developed. A formulation for a total-field/scattered-field source for planar multi-layered media was implemented in a parallelized finite-volume time-domain (FVTD) computational engine. Methods for generating rough surfaces with Gaussian statistics were used to create computational geometries. Monte Carlo simulations for the scattering from the rough surfaces were performed using a scattered-field formulation of the FVTD method, which represents the rough surface as contrast sources. These contrast sources generate the scattered fields and are more easily absorbed by the absorbing boundary conditions (ABCs), improving efficiency and accuracy. The scattered near-fields were then transformed to the far-field, coherently averaged, and subsequently normalized to obtain a normalized radar cross-section. The validity of the technique is demonstrated by showing favorable comparisons with the small perturbation model (SPM).