CO₂-SO₂ clathrate hydrate formation on early Mars

International audienceIt is generally agreed that a dense CO2-dominant atmosphere was necessary in order to keep early Mars warm and wet. However, current models have not been able to produce surface temperature higher than the freezing point of water. Most sulfate minerals discovered on Mars are dated no earlier than the Hesperian, despite likely much stronger volcanic activities and more substantial release of sulfur-bearing gases into martian atmosphere during the Noachian. Here we show, using a 1-D radiative-convective-photochemical model, that clathrate formation during the Noachian would have buffered the atmospheric CO2 pressure of early Mars at ∼2 bar and maintained a global average surface temperature ∼230 K. Because clathrates trap SO2 more favorably than CO2, all volcanically outgassed sulfur would have been trapped in Noachian Mars cryosphere, preventing a significant formation of sulfate minerals during the Noachian and inhibiting carbonates from forming at the surface in acidic water resulting from the local melting of the SO2-rich cryosphere. The massive formation of sulfate minerals at the surface of Mars during the Hesperian could be the consequence of a drop of the CO2 pressure below a 2-bar threshold value at the late Noachian-Hesperian transition, which would have released sulfur gases into the atmosphere from both the Noachian sulfur-rich cryosphere and still active Tharsis volcanism. A lower value of the pressure threshold, down to ∼0.5 bar, could have been sufficient to maintain middle and high latitude regions below the clathrate formation temperature during the Noachian and to make the trapping of SO2 in clathrates efficient. Our hypothesis could allow to explain the formation of chaotic terrains and outflow channels, and the occurrence of episodic warm episodes facilitated by the release of SO2 to the atmosphere. These episodes could explain the formation of valley networks and the degradation of impact craters, but remain to be confirmed by further modeling