| Understanding the geologic evolution of Mars requires a detailed characterization of Martian surface compositions and their geologic settings. Because much of the Martian surface is covered with regolith or unconsolidated materials, in this work I focused on key locations in which to characterize Martian crustal materials as well as assess the degree of influence from physical sorting processes on measured surface compositions from orbit.;In Chapter 2, I identified and characterized 26 crater central peaks with distinctive spectral signatures from surrounding plains with thermal infrared and visible/near-infrared data, based on a global survey of Martian impact craters between 10-200 km diameter. Some degree of regional clustering is observed, suggesting that subsurface diversity in crust-forming compositions, physical properties or alteration environments are present. The western Noachis Terra area is the only region where multiple THEMIS-defined spectral units are present in individual central peaks, suggesting crustal stratigraphies of igneous units or ejected crustal materials or impact melt from the Argyre Basin.;In Chapter 3, I assessed the degree and occurrence of thermophysical and compositional heterogeneity for 42 dune fields on Mars. Among these, only four exhibit spatial heterogeneity in spectral properties and composition. Two of the four sites show a strong positive relationship between particle size and olivine abundance. The rarity of compositional heterogeneity within dune fields may indicate phenocryst-poor source rocks; alternatively sorting within individual bedforms may be present but is below the resolution of available instruments.;Discoveries of extensive sedimentary rock units on Mars call for a critical assessment of the ability to quantify mineral abundance from lithified sedimentary materials using infrared spectroscopy. In Chapter 4, I characterized the TIR spectral properties of compacted, very fine-grained mineral mixtures and assessed the linearity of spectral combination using non-negative linear least squares minimization. Non-linear relationships between known and modeled abundance are observed. However, using a partial least squares method, accurate mineral abundances can be recovered from TIR spectra of very fine-grained rocks (within +/- 10% of known abundance for 78--90% of mixtures). This method could potentially be applied to landed or orbital TIR observations. |
