Columbia Hills — An exhumed layered igneous intrusion on Mars?

Abstract

The rover Spirit has now performed a series of whole-rock chemical analyses on rock outcrops and boulders along its traverse across the Columbia Hills of Gusev crater. The wide-spread consensus that the Columbia Hills represent a mafic volcanic pyroclastic sequence has problems, however, explaining the compositional variation observed within the Columbia Hills analyses. The positive trend in Fe–Mg space exhibited by the rocks of the Columbia Hills contrasts markedly with volcanic trends controlled by crystal fractionation and no plausible crystal fractionation scheme can explain the co-variation of Fe and Mg, if the analyzed rocks are assumed to represent volcanic compositions defining a liquid line of descent. Positive arrays in Fe–Mg space are, however, common features in whole-rock data collected across cyclically-layered igneous intrusions. The repeated alteration of feldspar-rich versus mafic-rich cumulates in such intrusions results in whole-rock compositions that scatter along mixing lines between feldspar, near the origin, and the mafic cumulate phases, reflecting the magmatic sorting of cumulus crystals as opposed to crystal–liquid equilibria. There is a striking correspondence between the Fe–Mg trends observed in the Columbia Hills data and the cumulate stratigraphy of a number of terrestrial layered intrusions. This raises the possibility that stratification seen at many localities along Spirit's traverse may reflect magmatic sedimentation in a large layered igneous intrusion. According to this interpretation, the most magnesium rocks are harzburgite to olivine norite cumulates without cumulus magnetite. The majority of the rocks of the Columbia Hills, however, appear to be olivine norites to gabbronorites with cumulus magnetite, which are characterized by variable proportions of feldspar versus mafic silicates plus magnetite. Furthermore, phosphorus-rich rocks, which have previously been interpreted to be alkaline basalts, are better explained as leucocratic norites with cumulus apatite. The sequence of crystallization of the magma responsible for the Columbia Hills is therefore interpreted to be: olivine, followed by orthopyroxene and plagioclase, then magnetite, and finally apatite. Ca-rich clinopyroxene appears to have been a late post-cumulus phase. If this interpretation is correct, then the Columbia Hills likely represent a Noachian-age layered intrusion that was exhumed by rebound following the meteorite impact that formed Gusev crater in the Hesperian.