Gamma-ray constraints on the chemical composition of the martian surface in the Tharsis region: A signature of partial melting of the mantle?

The Gamma-Ray Spectrometer (GRS) on board Mars Odyssey is the first instrument to provide quantitative estimates of the composition of the martian surface at the global scale. At the present time published concentration maps are available across a broad equatorial swath for Ca, Cl, Fe, H, K, Si and Th. However, it remains to be determined to what extent the observed chemical signatures are the result of primary magmatic processes or secondary alteration at the surface. With this question in mind, we have considered in detail the GRS signature on the western part of the Tharsis region including Olympus Mons. This region was chosen as it is clearly of volcanic origin, has had a prolonged magmatic history, and is of spatial extent sufficiently large that potential contributions of surrounding regions to the GRS signature are limited. It also has a characteristic composition, being notably low in silica relative to the average composition of the martian mid-latitudes. The major element chemistry of the surface in this region (SiO 2, FeO, CaO, K 2O) is compared with predictions of thermodynamic calculations of liquids produced by partial melting of the martian mantle composition proposed by Dreibus and Wänke [Dreibus, G., Wänke, H., 1985. Mars, a volatile-rich planet. Meteoritics 20 (No. 2), 267–381] over a range of pressures and temperatures. For all four oxides, the Tharsis compositions are consistent with liquids produced at pressures from 15 to 20 kbar, and degrees of partial melting between 3 and 10%. This pressure of melting implies a magma source depth in the range ∼ 125–170 km, consistent with current estimates of the thickness of the lithosphere in this region. The idea that the surface chemistry in the Tharsis region is dominated by the signature of primary magmatism would therefore appear to be a reasonable working hypothesis, although as more data become available, more stringent and detailed tests of this simple model will be possible.