Type 2 Terrain: Compositional Constraints on the Martian Lowlands

Introduction: The Mars Global Surveyor Thermal Emission Spectrometer (TES) team has identified two distinct Martian terrains, Type 1 and Type 2 [1,2]. Type 1 is interpreted as basaltic. It is found throughout the southern highlands, particularly in the Syrtis Major region. Type 2, is interpreted as a basaltic andesite and is most strongly concentrated in the Amazonian age northern lowlands. Possible andesites on Mars have been discussed previously in the literature. The Pathfinder team identified their “sulfur-free rock” as having a basaltic andesite composition [3]. Even as far back as Viking, terrestrial andesite was used as a spectral analog in modeling the surface rocks near the Viking 1 lander [4]. There have also been basaltic andesite compositions suggested from flow rheology [5,6]. The interpretation of basaltic andesite has caused considerable discussion since it was put forth as an explanation for the TES data. Andesites on Earth are relatively rare, found mostly in subduction zones. If Mars does not have plate tectonics, then Martian andesites would have to form by a different mechanism. More silicic magma compositions can be obtained through fractional crystallization. This requires extracting small amounts of melt, a difficult way to produce such large amounts of andesite. However, adding even small amounts of water to a melt can produce andesitic magmas. Minitti and Rutherford [7] have found that an andesite of the “sulfur-free rock" composition can easily be created from a SNC parental magma with the addition of just 1 wt % H2O. Interpretation of TES Type 2 terrain: Deconvolving TES spectra is not an easy task. Global dust and atmospheric components must first be estimated [8]. The TES team uses a spectral deconvolution program to determine the modal composition of their residual surface spectra [9]. This process involves fitting the spectra with endmembers from a spectral library [10]. This deconvolution method has proven itself to be rather successful for determining modal mineralogy of known samples to within ~10% error for major minerals, though it does not appear to be reliable for minor minerals present at <10% [9,11]. The method however, is highly dependent on several assumptions: 1) The minerals present occur in exposed rocks or as large particles [12,13]. 2) Mixtures behave as a linear combination of individual components [13,14]. 3) The correct endmembers are present in the spectral library [11]. Glass Issues: The TES team’s spectral deconvolution of Type 1 terrain gives a solution of 50% plagioclase, 25% clinopyroxene, and 15% sheet silicates (concentrations of 15% or less are considered to be at or below the detection limit [9]). Type 2 terrain gives 35% plagioclase, 10% clinopyroxene, 15% sheet silicates, and 25% obsidian-like (high silica) glass [1]. However, the TES spectral deconvolution library used contained no pigeonite and one glass sample, an obsidian-like highsilica glass. The TES team recognizes that these missing library endmembers may affect interpretations [15,16]. Due to the highly disordered nature of glass, the spectral features of all glasses are generally broad and amorphous. Spectral variation of glasses are shown in figure 1 for a wide range of compositions. In fitting the TES spectra, broad or glassy spectra tend to act as convenient “fillers” as the program tries to minimize the error in its fit. Thus, if glass is indeed a major component of terrain Type 2, its derived silica content depends directly on the glasses available as endmembers in the spectral library.