Mineral Mapping in Valles Marineris, Mars: a New Approach to Spectral Demixing of Tes

Introduction: We applied multiple endmember demixing [1, 2] to Thermal Emission Spectrometer (TES) data to determine the composition of geologic units in the Valles Marineris (VM), Mars. The goal is to map the compositions of VM dark deposits [3] and layered units of the walls and interior deposits [4-6]. Background: The Valles Marineris (Fig. 1) offer 3-D views into the Martian crust to depths of up to 10 km [7, 8]. The VM are thought to expose the record of geologic history of Mars from early evolution of the crust and possible formation of a megaregolith, through modification by volcanism, possible lacustrine and other hydrologic activity, to tectonic processes that formed and modified the chasmata [9-12]. Determination of the origin of layered deposits in the VM offers insight into geologic history and past climates on Mars and may provide evidence for sites of recent activity that are suitable habitats for life. Current Work: VM geologic units are diverse [13] and are likely to contain minerals (e.g., pyrox-enes, plagioclase, olivine) and weathering products (clay minerals) that have diagnostic signatures in the TES 6-to 40-micron wavelength range. In addition to characterizing the layered units in VM, two other deposits are of interest: young, dark materials within the chasmata floor [3], and possible iron oxides at Ceti Mensa in west Candor Chasma [14, 15]. New software in ISIS was used to extract and process the 143-channel TES data (10 cm-1) prior to demix-ing. TES emissivity data were extracted using the program " vanna " (an ISIS interface to the ASU " vanilla " software) over the VM. These nadir, daytime TES data have surface temperatures between 250° to 350° K, incidence angles between 0 and 80°, and emission angles less than 10°. To reduce noise, up to 6 detector elements were averaged for each orbit and instrument observation, and average emissivities were reported for each detector footprint. These data were then sorted by surface temperature, and the ISIS " lev2raster " program was used to place TES pixels with warmest data on top. The spectral deconvolution is a revision of our previous approach [2] and focuses on the use of multiple endmember spectral mixture analysis [MESMA; 1] to remove atmospheric components and to analyze TES data relative to laboratory spectral measurements of common minerals. This method identifies the minimum number of components required to model the spectrum of each pixel in the spectral …