Krypton and Xenon as Indicators of Convective Zone Thickness in Firn at Megadunes, Antarctica

Air trapped in ice on the Antarctic plateau (Vostok, Dome Fuji and Dome C) during glacial periods exhibits unexpectedly low gravitational fractionation, as inferred from δ15N and δ40Ar. The deduced thickness of the ``diffusive zone'', in which molecular diffusion dominates as gas transport mechanism, is thinner by 30-40 m than the total firn thickness estimated with densification models. It has been hypothesized that this apparent inconsistency is due to a thickened near-surface ``convective zone'', in which convective mixing overwhelms molecular diffusion and thus prevents isotopic fractionation. However, there is currently no means to estimate the convective zone thickness in the past. In studies of today's firn, a significant convective zone has been found at low accumulation sites (Megadunes, 20 m; Vostok, 13m; Dome Fuji, 9 m) and at a very windy site (YM85, 14 m). The deep convection at the Megadunes site is probably attributable to the very low accumulation rate, which leads to very high permeability by making large firn grains and deep surface cracks. In this study, we measured 40Ar/36Ar, Kr/Ar and Xe/Ar ratios of the Megadunes firn air, in order to test the effect of deep convection on gravitational fractionation of heavy noble gases. In theory, the dominance of convection over molecular diffusion is larger for heavier gases than for lighter gases due to their low diffusivity. As a result, the gravitational fractionation of heavier gases should be smaller than that of lighter gases (i.e. the diffusive zone is thinner for heavier gases) if a deep convective zone exists. Our result indeed shows smaller gravitational fractionation of Kr/Ar and Xe/Ar than of 40Ar/36Ar when normalized by the mass differences for each gas pair. A preliminary measurement of 86Kr/82Kr also suggests smaller gravitational fractionation. Possible future application includes the reconstruction of the thickness of past convective zones by measuring deep ice core noble gases. This topic is important for validating densification models for the glacial periods, which may help constrain the phase relationship between climate and greenhouse gas variations.