Robert Clark wrote:
It is curious though that the authors argue against their being due to liquid water because of their closeness to the equator.

Alex writes:
The authors seem to buy into the view, as other workers have, that the majority of these flows originate at the surface (e.g., the melting ice or snow models). Note that Nick's CO2 snowpack model is analogous to this as well. I'll admit that these views do have some merit. The observation of flows emanating from dune crests and elevated, isolated peaks would tend to point away from a subsurface origin and, due to the ambient atmospheric pressures and temperatures, would probably preclude liquid H2O as the erosive agent. However, some workers, notabley Michael Hecht of JPL, disagree with this and argue that the metastability of H2O at the martian surface, in combination with the optical geometric characteristics of the gullies, could focus solar illumination to intiate melting.

On the other hand, those who subscribe to the subsurface H2O models (e.g., Malin and Edgett, who originally discovered the Nirgal Vallis gullies), can make liquid H2O work by invoking orbital obliquity excursions, ice damming, geothermal activity, etc.