Cold-based glaciation of Pavonis Mons, Mars: evidence for moraine deposition during glacial advance

The three large volcanoes in the Tharsis region of Mars: Arsia, Pavonis, and Ascraeus Montes all have fan-shaped deposits (FSDs) on their northern or western flanks consisting of a combination of parallel ridges, knobby/hummocky terrain, and a smooth, viscous flow-like unit. The FSDs are hypothesized to have formed in the Amazonian during a period of high spin-axis obliquity which redistributed polar ice to the equatorial Tharsis region resulting in thick (> 2 km), flowing ice deposits. Based on previous ice flow simulations and crater surveys, the ridges are interpreted to be recessional drop moraines formed as debris on the ice sheet surface was transported to the ice margin-forming a long ridge sequence over an extended (similar to 100 Myr) period of ice sheet retreat. We test this hypothesis using a high-resolution, thermomechanical ice sheet model assuming a lower ice loss rate (similar to 0.5 mm/year) than prior work based on new experimental results of ice sublimation below a protective debris layer. Our ice flow simulation results, when combined with topographic observations from a long sequence of ridges located interior of the Pavonis FSD, show that the ridged units were more likely deposited during one or more periods of glacial advance (instead of retreat) when repetitive pulses (approx. 120 kyr periodicity) of ice accumulation during high obliquity produced kinematic waves which advected a large volume of surface debris to the ice margin. If ridge deposition does occur during glacial advance, it could explain the cyclic pattern of ridge spacing and would link the dominant, 120 kyr periodicity in obliquity to the time interval between adjacent ridges. By measuring the spacing between these ridges and applying this timescale, we constrain the velocity of glacial margin to be between 0.2 and 4 cm/Earth year-in close agreement with the numerical simulation. This re-interpretation of the FSD ridged unit suggests that the timescale of FSD formation (and perhaps the duration of the Amazonian high obliquity period) was shorter than previously reported.