TOPOGRAPHIC, GEOPHYSICAL, AND MINERALOGICAL CHARACTERIZATION OF GEOLOGIC STRUCTURES USING A STATISTICAL MODELING APPROACH, UPPER ANIMAS RIVER WATERSHED, COLORADO

Statistical modeling of geologic structures (defined here as faults and veins) in the Animas River watershed study area using GIS enables definition of geophysical, mineralogical, and topographic attributes, which can be portrayed in charts and maps as probabilities. The modeling shows that faults and veins in the watershed have characteristic signatures within discrete value ranges. Topographically, structures are characterized by erosionally resistive and probably silicified ridgelines, or as eroded incised channels and valleys. The planimetric orientations of structures indicated by the trends of calculated geophysical gradients identify east-west, northerly, and northwesterly trends, which are colinear with the trends of mapped geologic structures. Magnetically, the faults and veins are mostly characterized by extremely low magnetization values that are associated with the magnetite-poor mineralogy of silicic vein material. Magnetization boundaries associated with rocks located at depths exceeding 1 kilometer identify surface structures that are likely to provide connectivity between the topographic surface and deep crust, and also reveal buried crustal boundaries. Mapped structures are characterized by moderately high electrical resistivities consistent with the silicic mineralogy of the vein material. The mineralogical character of structures as related to their acid-producing or acid-neutralizing potential suggests no significant association with acid production resulting from the weathering of pyrite. Structures are most likely to occur in propylitically altered areas, and are therefore associated with rocks with the highest acid-neutralizing potential. The predictive models show high probabilities that accurately locate many known structures, but also reveal areas devoid of mapped structures. Many of the permissive areas for unmapped structures, however, lie in places of restricted access, limited outcrop, or thick alluvium, soil, or vegetation cover. Our validation methods show that the models can successfully locate known structures and predict the location of previously unmapped structures. The predictive models presented in this study could be considered in remediation planning, or used as an exploration tool for ground-water or mineral resource applications.