The best model for the giant floods on Mars
I have to admit, is an aqueous one. The excellent work done over the last 8 or so years in detailed mapping of outburst flood localities has shown an intimate association of volcanism and floods. Clearly, the heat of volcanism (or of deep intrusions that later caused volcanoes) has warmed the ground and melted subsurface ices.
If those ices are largely H2O, then there would indeed be floods of 'water" (albeit a bit or a lot salty or acidic). Despite my modest efforts, the case for CO2 in the giant outburst floods has not strengthened, and indeed has probably weakened as the water exponents make proper detailed models rather than arm-waving.