Image by NASA/JPL/MSSS
The cliff edges, slump structures, tributaries, meanders, and minor deltas are also all in case after case in pristine condition, and not blanketed by deposits from the supposedly different regime operating now.
The slumps at the head of the canyons in particular show the chaotic irregular topography that immediately follows the slumping of large masses of material, and which precedes the settling and surface smoothing that invariably comes with time.
The drainage systems show features that do not fit any example known of liquid water flow. For the apparently relatively recent outflow channel systems, many observers have commented that feeder tributaries are either absent or very poorly developed, ruling out rain as the source of the supposed flowing water.
The channels end abruptly, with insufficient catchments to supply the amount of water modelling has indicated is required to move the missing lithic material, even if melting groundwater is invoked, whether in sudden catastrophic releases, or as trickles over aeons.
Riverbeds, as identifiable continuous courses within the bounds of the valleys and canyons, and down which water is presumed to have flowed, are absent from all of them.
If the transport and erosion system was instead driven by wind operating over huge periods of time, there is no fluid supply problem. The return cycle is clear and the available fluid is almost limitless.
Uphill flow would not be a problem either. Since this system can be seen operating very effectively right now, it would seem best to seriously consider what is, before pondering what might have been.
Wind-blown sand is a very effective abrasive of rock, as a quick consideration of what it has done to the sphinx and the pyramids in a mere few thousand years will demonstrate perfectly well.
I argue that the operative forces are active now, and that much of the action is relatively recent, but as an aside I hasten to add I am a card-carrying geologist and not of the Mt Ararat school of thought. I am perfectly comfortable with a very old planet for a neighbour, and it is central to my case that the processes operating now have been continuously active for billions of years.
Mars now appears to have a much thicker crust than Earth, and shows little evidence of anything like plate tectonics except perhaps in its very early history. Hence isostatic adjustment to erosion or mountain building will be very slow.
If Mars is also dry and without rain, the rate at which erosion would lower land surfaces would be far slower than on Earth. Those factors would act in concert to make drainage channels very stable, long-lived features.
River channels on Earth persist for perhaps hundreds of millions of years at most, in isolated cases where their courses are controlled by plate boundaries. On Mars individual drainage channels may persist for billions of years.