In about 2 weeks a new paper of mine will be published in "Astrobiology" that develops this theme.

It addresses the near-polar gullies and shows that flows occur each Martian springtime, at temperatures far too cold for any water-based system. (and too cold for liquid CO2, as well as too low a pressure at the surface).

The flows occur in and through drifts of CO2 snow as they defrost.

Therefore I conclude that a solid-vapour phase transition of CO2 is responsible for generating a gas-lubricated flow.

As for the rest of what Stewart and Nimmo say about no CO2 in the subsurface, another paper is in progress on this but will be rather slower to arrive...