The case of the very old ice from Beacon Valley

Buried glacier ice from Beacon Valley has been studied by Sugden et al. (1995) and claimed to be about 8 million yr old. 40Ar/39Ar analyses of single volcanic crystals within in situ air-fall ash deposits within the drift above the buried ice provide the geochronological control. These results have been criticized (Hindmarsh et al., 1998) on the basis that sublimation over such a long period would have resulted in the complete disappearance of this buried ice. The buried ice is, however, protected from intense sublimation by the protective cover of the drift. Furthermore, this drift is partly composed of ice-cemented sands. The co-isotopic analysis, both in SD and S18O, of this ice cement provides clues about its formation. In a SD-S18O diagram, the samples of the ice cementing the deposit are aligned along a slope of ca. 5, considerably lower than the slope corresponding to the buried glacier ice and modern snow (Fig. 35.4), which is a meteoric water line (slope of ca. 8). The intersection point between this precipitation slope and the line on which the ice that cements the sands is aligned is very close to modern snow samples, suggesting that this snow could be at the origin of the cementing ice. The deuterium excess values of the cementing ice (-85% to -45%) are surprisingly extremely negative and much more negative than in modern snow samples (-4% to +2%). This effect is best explained if the source of the cementing ice has undergone strong evaporation before freezing. Although the deuterium excess of ice can be lowered by partial freezing following melting, the change in deuterium excess never exceeds a few parts per mil. Conversely, evaporation of snow also produces a shift off the precipitation slope but with a larger decrease in the deuterium excess. The simplest scenario in agreement with the isotopic data is that snow has accumulated, has undergone significant evaporation loss and partial summer melting, and has percolated into permeable sand layers where refreezing above the buried glacier ice has occurred. Small amounts of snow melt on sunny summer days have been observed when radiative heating was sufficient to overcome subzero air and ground temperatures. Downward percolation of meltwater is probably facilitated by the depression of the freezing point by the high salt content of the Dry Valleys soils (Bockheim, 1990). Depressions of the ground surface (polygon troughs for example) trap windblown snow and initiate a negative feedback that slows further sublimation of the underlying glacier ice (Marchant et al., 2002). This negative feed-back involves the development of secondary-ice lenses within the covering sands. Such a process acts as an effective protective cap against sublimation for the underlying buried glacier ice, allowing its long time preservation. The highly negative values of deuterium excess, as encountered in this study, are thus in strong support of the proposed scenario.

0 0

Post a comment