Correction to "Backscattering from a Gaussian-distributed perfectly conducting rough surface"

In an earlier analysis of backscattering from a perfectly conducting Gaussian-distributed composite surface, shadowing was properly included in the formal results. However, due to the use of an incorrect form for the shadowing function, its effect was erroneously evaluated. The correct shadowing function is incorporated into the earlier analysis and its impact on the scattering near grazing incidence is determined. Two consequences of shadowing are found to significantly reduce a near grazing incidence. These effects result from the nonzero variance of the large-scale slopes and the total shadowing (with probability one) of points on the surface with unfavorably oriented slopes. Of the two effects, the former is the most significant since it causes to go to zero at grazing incidence. Particular attention is given to the physical reasoning behind the causes and effects of shadowing.