We report on 475 measurements of depth to ice-cemented ground in four high-elevation valleys of the Quartermain Mountains, McMurdo Dry Valleys, Antarctica. These valleys have pervasive ice-cemented ground, and the depth to ice-cemented ground and the ice composition may be indicators of climate change. In University Valley, the measured depth to ice-cemented ground ranges from 0–98 cm. There is an overall trend of increasing depth to ice-cemented ground with distance from a small glacier at the head of the valley, with a slope of 32 cm depth per kilometre along the valley floor. For Farnell Valley, the depth to ice-cemented ground is roughly constant (c. 30 cm) in the upper and central parts of the valley, but increases sharply as the valley descends into Beacon Valley. The two valleys north of University Valley also have extensive ice-cemented ground, with depths of 20–40 cm, but exhibit no clear patterns of ice depth with location. For all valleys there is a tendency for the variability in depth to ice-cemented ground at a site to increase with increasing depth to ice. Snow recurrence, solar insolation, and surface albedo may all be factors that cause site to site variations in these valleys.

The McMurdo Dry Valleys are among the driest, coldest environments on Earth and are excellent analogues for the Martian northern plains. In preparation for the 2008 Phoenix Mars mission, we conducted an interdisciplinary investigation comparing the biological, mineralogical, chemical, and physical properties of wetter lower Taylor Valley (TV) soils to colder, drier University Valley (UV) soils. Our analyses were performed for each horizon from the surface to the ice table. In TV, clay-sized particle distribution and less abundant soluble salts both suggested vertical and possible horizontal transport by water, and microbial biomass was higher. Alteration of mica to short-order phyllosilicates suggested aqueous weathering. In UV, salts, clay-sized materials, and biomass were more abundant near the surface, suggesting minimal downward translocation by water. The presence of microorganisms in each horizon was established for the first time in an ultraxerous zone. Higher biomass numbers were seen near the surface and ice table, perhaps representing locally more clement environments. Currently, water activity is too low to support metabolism at the Phoenix site, but obliquity changes may produce higher temperatures and sufficient water activity to permit microbial growth, if the populations could survive long dormancy periods (∼106 years).

Dry permafrost on Earth is unique to the Antarctic and is found in the upper elevations of the McMurdo Dry Valleys. Despite its widespread presence in the Dry Valleys, the factors that control the distribution of dry permafrost and the ice-cemented ground below it are poorly understood. Here I show, by a combination of theoretical analysis and field observations, that the recurrence of snow can explain the depth of dry permafrost and the location of ice-cemented ground in Antarctica. For data from Linnaeus Terrace at 1600–1650 m elevation in Upper Wright Valley a recurrence intervals of about two years explains the presence of ground ice at 25 cm depth, under 12.5 cm of dry permafrost. Snow recurrence periods longer than 10 years would create only dry permafrost at this site. The snow gradient in University Valley resulting from the windblown snow from the polar plateau creates a corresponding gradient in the depth to ice-cemented ground. On the floor of Beacon Valley, the presence of dry permafrost without underlying ice-cemented ground indicates snow recurrence intervals of more than 10 years and implies that the ancient massive ice in this valley is not stable. Snow recurrence may also set the depth to ground ice on Mars.