LANDSAT MAPPING OF TEMPORAL CHANGES IN MINERALOGY OF EVAPORITE BASINS

Modern evaporite basins are dynamic landscapes where surface mineralogy responds rapidly to changes in hydrology, solute saturation, land use, and surface processes. The Bonneville basin (BB) in northwestern Utah was established as a large saline pan as Lake Bonneville desiccated, accumulating carbonate-clay mixtures, gypsum, and halite. Like many other great saline pans, the BB has become an important site for industrial salt extraction and other anthropogenic uses. Monitoring the spatial distribution of the BB mineralogy is becoming increasingly important, as climate change and anthropogenic influences are modifying the landscape. Field-based mapping of the BB, or any saline pan, prove problematic and inefficient due to inaccessibility, rapid change in surface conditions, and varying spatial distribution of mineralogy caused by seasonal cycles and weather events. Here we establish and apply a new method for mapping variations in surface mineralogy, utilizing the multispectral Landsat Thematic Mapper (TM) and Operational Land Imager (OLI) sensors, at the BB and other evaporite basins. Exploitation of the contrasting spectral shapes within the existing mineralogy is achieved through band-math algorithms that effectively highlight specific evaporite minerals. Band-math algorithms were attained through comparison of downsampled field and lab-based spectra to identify exploitable spectral differences. The algorithms follow (Red-SWIR1)/(Red+SWIR1) for halite, ((NIR-SWIR1)-(Red-SWIR1))/(NIR+RED+SWIR1) for gypsum, and ((SWIR1-NIR)+(SWIR1-SWIR2))/(NIR+SWIR1+SWIR2) for carbonate-clay differentiation. Each scene mapped is atmospherically corrected, subset to the study area, and isolated to surface deposits by masking of bedrock, vegetation, and standing water. Resulting maps show substantial differences in the spatial distribution of mineralogy within each decade from 1986-2016 associated with changes in land use and climate. The applicability of this approach is examined by applying the algorithms to other evaporite basins, including the Salar de Atacama and Salar de Uyuni, with variable success. Some of these differences may be attributed to variation in evaporite mineralogy, mixture composition, evaporite basin maturity, topography, and other factors.