MAPPING PLAYA EVAPORITE MINERALS WITH AVIRIS DATA - A 1ST REPORT FROM DEATH-VALLEY, CALIFORNIA

Efflorescent salt crusts in Death Valley, California, were mapped by using Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data and a recently developed least-squares spectral band-fitting algorithm. Eight different saline minerals were remotely identified, including three borates, hydroboracite, pinnoite, and rivadavite, that have not been previously reported from the Death Valley efflorescent crusts. The three borates are locally important phases in the crusts, and at least one Of the minerals, rivadavite, appears to be forming directly from brine. Borates and other evaporite minerals provide a basis for making remote chemical measurements of desert hydrologic systems. For example, in the Eagle Borax Spring area, the AVIRIS mineral maps pointed to elevated magnesium and boron levels in the ground waters, and to the action of chemical divides causing subsurface fractionation of calcium. Many other chemical aspects of playa brines should have an expression in the associated evaporite assemblages. Certain anhydrous evaporites, including anhydrite, glauberite, and thenardite, lack absorption bands in the visible and near-infrared wavelength range, and crusts composed of these minerals could not be characterized by using AVIRIS. In these situations, thermal-infrared remote sensing data may complement visible and near-infrared data for mapping evaporites. Another problem occurred in wet areas of Death Valley, where water absorption caused low signal levels in the 2.0-2.5 mum wavelength region that obscured any spectral features of evaporite minerals. Despite these difficulties, the results of this study demonstrate the potential for using AVIRIS and other imaging spectrometer data to study playa chemistry. Such data can be useful for understanding chemical linkages between evaporites and ground waters, and will facilitate studies of how desert ground-water regimes change through time in response to climatic and other variables.