摘要 : Reconstructions of the orbital parameters of Mars spanning the last similar to 20 Myr, combined with global circulation models, predict multiple cycles of accumulation and degradation of an ice-rich mantle in the mid-latitudes, dr... 展开 Reconstructions of the orbital parameters of Mars spanning the last similar to 20 Myr, combined with global circulation models, predict multiple cycles of accumulation and degradation of an ice-rich mantle in the mid-latitudes, driven primarily by insolation at the poles during periods when obliquity was more than ten degrees greater than it is today (i.e., >similar to 35 degrees). While evidence of an ice-rich "latitude dependent mantle" (LDM) consistent with these predictions is abundant, features indicative of cycles of emplacement and degradation of this unit are isolated and rare. In addition, fundamental physical properties of the LDM, such as paleo-thickness maxima, have not been determined. Gullies, which are sinuous channels found on steep slopes in mid- and high-latitudes, interact with the LDM and provide a stratigraphic feature useful for documenting both cyclical emplacement/removal and thickness estimates in past climate regimes. In the southern hemisphere, where gullies are most common, we present extensive evidence of (1) cyclical degradation and removal of gullies in the lower mid-latitudes (30-40 degrees S), and (2) burial and exhumation of inverted gully channels in the transitional latitude band between dissected and preserved LDM (40-50 degrees S), which can only be accounted for if an additional tens of meters of LDM were present at these locations during channel formation. These relationships support a model in which end-to-end gully evolution is controlled by the behavior of the LDM: at lower latitudes, gullies incise an ice-rich substrate and are removed when that ice becomes unstable, and at higher latitudes gullies are buried by successive emplacement of LDM where ice remains stable near the surface. Further, the presence of dormant buried gullies implies that present-day activity within gullies, likely to be controlled by the behavior of CO2 frost, is insufficient to explain the entire gully population, and that conditions conducive to increased gully activity preceded the most recent phase of LDM emplacement. (C) 2015 Elsevier Inc. All rights reserved. 收起