Bob Clark wrote:
In this LPSC report they appear to be suggesting the gullies would have
to be formed in periods of high angle of tilt in Mars obliquity cycle,
because Mars would be too cold at the surface at those latitudes
currently.

Alex writes:
In their JGR-Planets paper, Mellon and Phillips evaluated two possible
mechanisms: (1) melting of near-surface ground ice by solar heating and
conduction from the surface; and (2) melting (and refreezing) of water
at the base of the permafrost by geothermal heating. They found (1) to
be implausible because "...near-surface ground ice does not reach the
melting point of water under a range of conditions of soil thermophysical
properties, latitudes, obliquities, and surface slopes. The atmosphere
remains too dry for the ground ice to melt, even at high obliquity;
instead, ice sublimates before reaching melting temperatures." Mellon
and Phillips favor the geothermal model in (2) because it is consistent
with the spatial distribution of the gullies, though they caution under this
scenario that near surface liquid water would have to have contained
dissolved salts in concentrations of 15–40% to adequately depress the
melting point in the cryosphere. Hyperconcentrated brines deriving from
salt-forming compounds may be inferred from Viking and Mars Pathfinder
surface data, but they have not been directly observed. Moreover,
orbital spectroscopy has not resolved the putative salt deposits at the
gully sites, something that should be observed in model two. In fairness,
though, MGS TES simply does not have the required spatial resolution.
The results from 2001 Mars Odyssey's THEMIS should help considerably.
However, such salt deposits should be high albedo and discernible to the
MOC, which has found no evidence for their existence to date.

Alex R. Blackwell
University of Hawaii