REED SCHERER and A. ALDAHAN, Institute of Earth Sciences-Quaternary Geology, Uppsala University, Uppsala, Sweden
GÖRAN POSSNERT, Tandem Laboratory, Uppsala University, Uppsala, Sweden
Sediments beneath the west antarctic ice sheet contain invaluable information regarding ice-sheet history and glacial processes. Samples of sediments have been recovered from beneath the west antarctic ice sheet at Upstream B and the nearby shear marginal zone and, most recently, from beneath Upstream C. Additional samples have been recovered from beneath Crary Ice Rise, Ross Ice Shelf Project site J-9, and basal debris from the ice core at Byrd Station (figure). We are exploring the potential of the cosmogenic isotope beryllium-10 (10Be) in these subglacial sediments as a tracer of ice-sheet history and glacial processes. In this article, we outline our initial working hypotheses and assumptions.
The first researchers to use 10Be in sediments from the interior of West Antarctica were Yiou and Raisbeck (1981), who studied samples from beneath the southern part of the floating Ross Ice Shelf. They found low 10Be concentrations in core samples (less than 107 atoms per gram) and interpreted this finding as indicative of pre-Quaternary sediment in agreement with Harwood, Scherer, and Webb (1989). The surface layer, in contact with sub-ice shelf water, had a measurable concentration (approximately 108 atoms per gram), but even this was an order of magnitude lower than the concentrations in Quaternary marine sediment.
The cosmogenic isotope 10Be (halflife 1.5 million years) is produced both in the atmosphere and in situ at the Earth's surface. In situ production is not significant in fine-grained submarine or subglacial sediments, so all 10Be detected in the clay-sized fraction of subglacial sediments is assumed to be originally from atmospheric fallout. The 10Be atoms become part of the sedimentary record by adsorbing onto available clay particles.
We discuss three potential sources of 10Be to the west antarctic ice-sheet subglacial environment:
10Be release by basal melting is insignificant where the ice sheet is frozen to the bed but could become significant where the basal melt rate is high, particularly in regions of high geothermal flux. We consider advection of 10Be beneath a floating ice shelf at Upstream B and vicinity, deep within the west antarctic interior, to provide at best a very minor source of 10Be because ice shelves this far south would likely be unstable and short-lived. We also consider the possibility that tidal pumping might introduce a small amount of 10Be from sub-ice shelf water, across a broad coupling zone, although sub-ice shelf waters are likely depleted in 10Be (Aldahan et al. in press). Sediments older than late Miocene will have experienced significant 10Be decay, thus high residual 10Be concentration in subglacial sediments must reflect contribution from post-Miocene deposition in an open marine environment, implying disintegration of the ice sheet, as reported by Scherer (1991).
To apply 10Be successfully as a tracer for addressing questions of basal dynamics or ice-sheet history, it is important to understand the relative contributions of the three 10Be sources described above. To realize this goal, we must further our understanding of glacier bed conditions, especially with regard to basal water production and flow. If a water conduit system or sheet flow removes excess water from the bed, then some fraction of the 10Be released from basal melt will likely be flushed to the grounding line. Rapid basal flow seems to be the case at Upstream B, where borehole experiments demonstrate that water at the bed flows, at least locally, significantly faster than the ice (which itself flows at approximately 400 meters per year) (Engelhardt and Kamb 1997). In a basal sliding scenario, 10Be might accumulate only if there is net clay deposition and no erosion.
Ice flow directly linked with till deformation in a several meters thick till layer would provide a mechanism for continuously adding and storing ice-derived 10Be in the sediments. Continuous mixing would distribute 10Be throughout the active till layer. Consequently, a vertical 10Be profile through the sediment column would reveal no stratigraphic variability. Detecting a basal melt signal would, however, require a very significant 10Be addition, because the signal would become diluted through the sediment column. In fact, rapid till mixing by deformation processes would result in an absence of stratigraphic variability in 10Be content, whether derived from relict sediments or basal ice melt.
We have calculated the anticipated input of 10Be by basal melt, assuming a 1-millimeter-per-year basal melt rate and 105 10Be atoms per gram of melted ice, over the entire Holocene (last 10,000 years). This scenario yields a total 10Be inventory of 108 atoms per square centimeter. With this inventory and continuous 10Be accumulation for 10,000 years with no erosion of the surface sediment, the 10Be concentration is expected to be about 107 atoms per gram of sediment in contact with the ice. If this amount is diluted through a 10-meter-thick deforming till devoid of residual 10Be, then the average concentration should be less than 106 atoms per gram, or near the detection limit. These parameters would result in undetectable levels even if mixed through a 1-meter-thick till, but a signal should be detectable in a deformed layer of less than 1 meter.
The above calculations assume no erosion of the sediment column. Obviously, erosion at the bed and mixing with older sediments would significantly reduce the ice-derived 10Be content of the sediments, just as higher basal melt rates would increase the 10Be input. One may be tempted, for these calculations, to extend the sub-ice release time well beyond the Holocene, which in a closed system would lead to higher 10Be content. Clear evidence has been found, however, of very significant erosion and removal of older subglacial debris in West Antarctica during Quaternary and earlier glaciations.
From the above calculations, we tentatively conclude that significant 10Be concentration in sediments beneath the ice sheet is easier to explain by marine deposition in the west antarctic interior than by basal ice melt, although basal melt and flow conditions must be better constrained. An absence of 10Be throughout the sediment column would reflect old sediment, derived from Miocene and older strata, with little contribution from younger deposits, and an undetectable contribution from basal ice melt or other sources. We will continue to develop these hypotheses as we analyze our 10Be data from ice stream B, ice stream C, the Crary Ice Rise, the Ross Ice Shelf Project site J-9, and Ross Sea diamictons, in hopes of testing theories regarding ice-sheet history and glacial processes.
Financial support for this study was provided by the Swedish Natural Sciences Research Council, Uppsala University, and the Knut and Alice Wallenberg Foundation. We especially thank Barclay Kamb and Hermann Engelhardt of California Institute of Technology for providing samples for analysis. This study was also supported by National Science Foundation grant OPP 93-19018.
Aldahan, A., Shi-Ning, G. Possnert, J. Backman, and K. Boström. In press. 10Be records from sediments of the Arctic Ocean covering the last 350ka. Marine Geology.
Engelhardt, H., and B. Kamb. 1997. Basal hydrolic system of a west antarctic ice stream: Constraints from borehole observations. Journal of Glaciology , 43(144), 207-230.
Harwood, D.M., R.P. Scherer, and P.-N. Webb. 1989. Multiple Miocene marine productivity events in West Antarctica as recorded in upper Miocene sediments beneath the Ross Ice Shelf (site J-9). Marine Micropaleontology , 5, 91-115.
Scherer, R.P. 1991. Quaternary and Tertiary microfossils from beneath ice stream B: Evidence for a dynamic west antarctic ice sheet history. Global and Planetary Change , 4(4), 395-412.
Yiou, F., and G.M. Raisbeck. 1981. The age of sediments beneath the Ross Ice Shelf as implied by cosmogenic 10Be concentrations. EOS, Transactions of the American Geophysical Union , 62(17), 297. [abstract]