A Modified Semi-Empirical Radar Scattering Model for Weathered Rock Surfaces

Abstract

This study presents a modified semi-empirical radar scattering model for weathered rough rock surfaces. Weathered rocks generally have dry surfaces except for a few hours after heavy rain due to their rapid drainage compared to bare soils. We find that the dielectric properties of the rocks themselves and the moisture content of a marginal amount of soil patches in and around the rock surfaces have a negligible effect on radar backscattering. Therefore, radar backscattering from rock surfaces is mainly controlled by surface roughness. Here, we modified the cross-polarization ratio model developed by Oh, which is formulated by only surface roughness parameters with no dependence on soil moisture. Based on LiDAR surface topography data collected from weathered rock surfaces in the Canadian Arctic and corresponding quad-polarimetric RADARSAT-2 synthetic aperture radar (SAR) data, we determined the best fit radar backscattering model for weathered rock surfaces. The modified model was successfully applied to estimate the surface roughness of weathered rock surfaces up to approximately ks = 7 where k is the radar wavenumber (= 2π/λ) and s is the root mean square (RMS) height. This approach avoids the rapid saturation feature observed at ks > 3 in other models.