INVERSIONS OF SIR-C AND AIRSAR DATA FOR THE ROUGHNESS OF GEOLOGICAL SURFACES

The surface roughness of alluvial fans in Death Valley, California, ch anges as the fans age. Because radar backscatter is sensitive to surfa ce roughness, it is possible to map relative surface age using Synthet ic Aperture Radar (SAR) data. We have examined surface roughness estim ates in Death Valley from Shuttle Imaging Radar-C SAR data with the us e of a technique that we call fore ground/background analysis (FBA), w ith the goal of establishing a robust inversion method that is extendi ble to a range of surfaces and conditions. In this method, a foregroun d entity (in this case roughness) is distinguished from complicating b ackground factors (including the distribution of intermediate-scale sl opes, vegetation, and dielectric constant). The inversion for roughnes s is non-unique even when constrained by the use of field measurements . When the range of possible selections is examined it is observed tha t they fall into a small number of domains, each with distinct charact eristics that are probably associated with physical factors such as th e scale of roughness. Solutions were compared with those determined fr om the semiempirical (SEM) and integral equation (IEM) models, and all selections were evaluated with respect to field knowledge. The SEM an d IEM solutions, and those obtained using FBA together with the field measurements, fall into a common domain of solutions that are suscepti ble to contamination by background effects and hence may not be extend ible to other geographic locations. A domain of stable solutions that are more extendible does exist; however, this extendibility is achieve d at the expense of reduced resolution of roughness levels. For Death Valley, we estimate that it is possible to resolve only four levels of roughness, which is far fewer than can be theoretically resolved with existing inversion algorithms. (C)Elsevier Science Inc., 1997.