In this paper, we apply the right preconditioned generalized minimal residue (GMRES-RP) method for the analysis of electromagnetic scattering from 1-D ocean surfaces. We first verify that the GMRES-RP method has the desirable features of high accuracy, robustness, and efficiency, and, hence, is well suited to such analysis. We then carry out this analysis systematically by examining the effects of wind speed, incidence angle, ocean spectrum, and the degree of inclusion of the gravity and intermediate waves. With a large surface length of 16 384 λ and a noise floor pushed down below −70 dB, we have found more refined structures than suggested in the literature: At nearnormal incidence angle (θ i = 20°), both VV and HH results remain sensitive to wind speed. Across the incidence angle range, the sensitivity is higher at small wind speed (< 5 m/s) than at medium to high wind speed.
Plant argued that the contribution from intermediate waves separately, not as a part of the large wave spectrum, becomes increasingly important with increasing wind speed and, accordingly, a multiscale model is desired. Our numerical analysis validates this viewpoint by showing that a normalized bistatic scattering coefficient value as large as 4 dB at certain angles for H-pol at L band can be attributed to the intermediate wave and the tilt modulation provided by large waves. The viewpoint is further validated by a comparison of the two-scale model and three-scale model against numerical resultswhere the two-scale model shows appreciable discrepancy in the forward scattering angles whereas the three scale model shows very good agreement with MoM, hence manifesting the necessity of advancing multiscale models beyond the conventional twoscale model. This finding also implies the need for numerical studies to use adequately large surface length, at least to cover the intermediate waves, even for below intermediate incidence cases.
IndexTerms-Electromagnetic scattering, generalized minimal residue (GMRES), intermediate wave, ocean surface, preconditioning.
In this paper we propose an optimized Monte Carlo procedure with the variance reduction technique incorporated. The proposed method makes use of some auxiliary variables which embed the correlation information intrinsic in the it eration process, hence the procedure can be systemized. The use of intrinsic information and the capability to systemize the procedure draw the essential difference of the proposed method with the control variable method in the Monte Carlo literature. The expression for the variance is derived, and an appropriate optimization problem is solved. The proposed method is applied to the analysis of electromagnetic scattering from rough surfaces.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.