WET OR DRY? A PETROLOGICAL STUDY OF THE MARTIAN MANTLE

Most martian meteorites (Shergottite – Nakhlite – Chassignite, SNC) support a dry mantle on Mars (H₂O < 60 ppm) [1]. SNC are yet < 2.4 Ga and from < 15 different ejection sites located in the same region, which suggest they are not representative of the whole mantle. The discovery of igneous rocks of evolved compositions (SiO₂ > 53 wt.%) scattered in Noachian excavated terrains [2], in Gale crater [3], and in the martian meteorite Northwest Africa 7533 [4] yet suggest potential widespread evolved magmatism that might have required water-rich parental magmas. This study investigates the amount of water contained in Mars’ mantle and the resulting evolution of magmas through time.

The viscosity at which the mantle convects is related to both the temperature and volatile contents, which we consider here as water. Using Mars’ thermal history [5], inversion models backed out an average H₂O content in the mantle through time through the calculation of the evolution of mantle viscosity that matches the temperature and elastic thickness of Mars’ lithosphere [6]. Using the same thermal history and the deduced H₂O content, pMELTS isentropic models estimated the evolution of the composition of primary magmas and the residual mineralogy. Fractional crystallization of melts produced at 4.0 and 1.0 Ga were then run using rhyoliteMELTS models at crustal depth.

The inversion models show that H₂O ~ 600 ppm were contained in the mantle 4.5 Ga and decreased down to ~300 ppm today. Such values are in agreement with initial H₂O contents inherited from accretion and Mars’ water inventory. Mars’ cooling led to favoring sub-alkaline residual compositions in early Mars with abundant opx, while recent Mars would display alkaline residual melts with abundant cpx and olivine. The occurrence of water favors opx over cpx and delayed feldspar crystallization, which led to felsic igneous products at lower fractionation stage (< 80wt.%) than in a dry scenario. Hence, a hydrous mantle on Mars would have favored the formation of evolved crustal components on Mars that was suggested by the low crustal densities estimated by topographic-geoid ratio and seismic data.

[1] Udry et al. (2020) JGR, 125(12)

[2] Sautter et al. (2015) Nat. Geosc. 8(8)

[3] Humayun et al. (2013) Nat. Geosc. 503(7477)

[4] Wray et al. (2013) Nat. Geosc., 6(12)

[5] Filiberto (2017) Chem. Geol., 466

[6] Thiriet et al. (2018) GRL, 123