Olivine-rich basalt probably argues against long-lived lakes in the canyons, but floods of anything are quick enough not to rot the rocks, especially if they take place against a background of cryogenic temperatures.

So, unfortunately, olivine per se does not adequately support a CO2 flood model.

The distrivution of more mobile and soluble salts such as NaCl, CaSO4 may be more informative, but so far little data.

Odyssey GRS will eventually produce a Cl distribution map, which will be useful. I've looked at the recently released maps for Fe, Si, and Th and they are enigmatic. At the available resolution and overlay on the geology isn't very informative.

I'll make a post later today with my compilation of the published maps...