Tanaka K. L. (2000) Dust and ice deposition in the Martian geologic record. Icarus 144(2), 254-266.
This paper's abstract opens with the statement:
The polar layered deposits of Mars demonstrate that thick accumulations of dust and ice deposits can develop on the planet if environmental conditions are favorable. These deposits appear to be hundreds of millions of years old, and other deposits of similar size but of greater age in nonpolar regions may have formed by similar processes. Possible relict dust deposits include, from oldest to youngest: Noachian intercrater materials, including Arabia mantle deposits, Noachian to Early Hesperian south polar pitted deposits, Early Hesperian Hellas and Argyre basin deposits, Late Hesperian Electris deposits, and the Amazonian Medusae Fossae Formation.
I think Tanaka lays out a convincing argument throughout this paper tying together age relationships, photo-geologic observations, and a big picture view of Mars as possibly only he can. He lays out a potential history of Mars with the Noachian being marked by heavy deposition of dust/ice material created by pyroclastic eruptions, fluvial activity, impacts, and weathering. This is followed by a cooling of the climate where only episodic deposition occurs in only favorable places at the end of the Noachian/Hesperian. This period and subsequent periods are then marked primarily by deep erosion and redeposition of this material at the poles and in the Argyre and Hellas basins.
There is also an extensive discussion critical of the polar wander idea of Shultz and Lutz (1988). He cites problems with age-dating potential paleo-polar materials, and inconsistencies of the story with regard to the timing and growth of Tharsis. The conclusion being that any significant polar wander occurred very early in martian history and it's record has been destroyed.
It is amazing to note that this paper was written at the cusp of the revolution of scientific data from Mars, with only the preliminary returns from the MOC and MOLA to inspire it. I do not possess enough familiarity with the data to know how much things have changed, but to my eye many of the observations used to back up the arguments in this paper are still valid today. In fact much of the data returned has strengthened this argument.
One topic that I think is not treated well in this paper is the role of obliquity variations in driving dust/ice deposition and erosion processes. Given this driver, it may have been possible to have dust/ice deposition throughout martian history -- with gradually depleted ice reservoirs as increasing amounts of ice finds stable locations within the crust which are immune to climate changes caused by obliquity variations. This water could also be lost to space.
Finally in the conclusions Tanaka provides a means for discriminating between ice/dust deposits and fluvial deposits:
Such deposits [dust/ice deposits] should be fine grained and thus have low thermal inertia. They may be layered, reflecting climate-induced cyclic variations in composition. Airfall deposits may be draped over preexisting topography and may erode by mass wasting and (or) eolian erosion. In contrast, fine-grained fluvial deposits should have flat to gently sloping surfaces, be restricted to depressions, and have feeder channels and a local source region.
Of course the main idea behind the source of these deposits today is that they are the result of groundwater upwelling and evaporation similar to what has been proposed for Meridiani. I'm not clear how this changes these criteria or not. I think that the draping should be a defining characteristic distinguishing between these two depositional mechanisms, but aeolian reworking can always be invoked to explain draping no matter what mechanism is proposed.