Image by NASA/JPL/MSSS
Very rapid sublimation of vast quantities of carbon dioxide, combined with seismic shaking, would produce and expel to the surface extensive and dense layers of this gas at many locations, which would at many points flow downhill katabatically and catastrophically, carrying with it vast quantities of erosive material.
Hence the misleading appearance that there were huge and rapid water floods acting like the great Washington State water deluge and producing erosion and deposition features similar to the Washington Scablands. Hence also all the fresh channels, the teardrop islands and all the other puzzling erosion features. All done in a flash by gas, or if you prefer, a belch following a punch in the belly.
There is another puzzle. If there was a lot of liquid water in the past, and there is a lot of water ice now, as some believe, where are the tracks of the obligatory water-ice glaciers?
The planet is littered with convenient slopes, as mentioned. There should be terminal, lateral and median moraines and eskers aplenty - long ridges of jumbled rocks.
Some have been claimed to be present, in particular in the Hellas crater, but they should be very common particularly around the poles, and are not.
Also, even if the postulated Martian water ice is now too rigid, given the gravity field and the present atmosphere, for plastic deformation to take place, it should have flowed that way at some time during the supposed climatic transition.
The giant catherine wheels that are the two polar ice caps show that whatever type of ice is there now, it does deform in response to planetary rotation at least.
But curiously, if there is water ice at the poles, there does not seem to be glaciers on the ice-cap margins. They should be there, even though downhill is towards the cap in the north. If the ice is frozen CO2, this is less curious, as the liquid phase, probably necessary to lubricate flow, would be missing.
Glaciers do not show the tight sinusoidal meanders seen in valley after valley, so those at least are not glacial forms, and would require special explanation, even if the straighter ones were glacial.
So far there seems to be no evidence that pebbles or rocks are moved by anything other than gravity, meteorite impacts or volcanism. If they were, the inevitable wind winnowing should have exposed many instances of well-sorted and rounded pebble, cobble and boulder beds along the rivers, if liquid water was the transport medium.
Boulders up to tens of metres across can be seen tumbled down many canyon slopes. If the scree slopes are made up largely of fine material, most of the rocks would sink into the sediment as it moved downslope, as happens with rocks on sandy beaches here.
This would occur with or without water being involved, as a consequence of mass transport of different sized materials and the action of gravity.
The dark patches on the floors of several canyons may be re-exposures of this debris. If a probe were to be landed there, the nature of the sorting and whether or not the rocks are river-rounded from long-distance movement, or only wind-and-dust-abraded from short journeys, should be evident on close-up photos, and should help resolve the question.