Glacial geology

Diatoms in subglacial sediments yield clues regarding west antarctic ice-sheet history and ice-stream processes

REED SCHERER, Institute of Earth Sciences-Quaternary Geology, Uppsala University, Uppsala, Sweden

SLAWEK TULACZYK, Division of Earth and Planetary Sciences, California Institute of Technology, Pasadena, California 91125

Alley et al. (1987) suggested that rapid flow of ice stream B, West Antarctica, can be explained by active deformation of a layer of unconsolidated, water-saturated till, about 6 to 10 meters thick. Such a layer is recognized seismically (Blankenship et al. 1987). Hot-water boreholes at Upstream B have provided access to the bed for sampling of these sediments by piston coring and other methods. The sediments recovered are consistent with predicted physical properties, but active deformation of this layer has not been established. Engelhardt and Kamb (in press) report an in situ test that suggests that basal sliding accounts for between 69 and 83 percent of ice-stream motion at this site and that till deformation may be restricted to the upper few centimeters, rather than the predicted thickness of as much as 10 meters.

Here we present "quantitative microstratigraphic" analysis of sediments recovered from various boreholes in the Upstream B area. We use absolute abundance of diatom fragments as sedimentary tracers. Diatoms provide excellent tracers for several reasons.

They have known initial conditions (size, shape, approximate age, and environment of origin).
Fragments as small as 1 or 2 micrometers can be recognized as being derived from diatoms (although fragments generally cannot be identified to the species level).
New data on diatom comminution places constraints on physical conditions in the subglacial environment (R. Scherer, N. Iverson, and T. Hooyer unpublished data, manuscript in preparation).

Upstream B diatoms are too rare to perform accurate estimates of whole fossils, but abundance estimates of small diatom fragments (<5 micrometers) in the <250-micrometer fraction of these sediments are highly reproducible, using the method of Scherer (1994), despite the fact that the number of fragments generated from a single diatom can vary from fewer than 5 to more than several hundred.

The data demonstrate a distinct difference between stratigraphic sediments recovered by piston coring and sediment samples that include the particles in closest proximity to the ice. The uppermost sediments contain a significantly higher concentration of diatoms and diatom fragments than those below. Core samples contain a mean concentration of 1.4x105 fragments per gram dry sediment (fr/g) (n=14), whereas samples that approximate the top of the sediment column, contain a mean of 1.8x10⁶ fr/g (n=11), an order of magnitude higher (figure). These values are several orders of magnitude lower than "typical" Ross Sea glacial marine sediments, including subglacial sediments from Crary Ice Rise (Scherer et al. 1988) and Ross Ice Shelf Project site J-9 (Harwood, Scherer, and Webb 1989).

These results suggest that the till column, recovered by piston coring, is distinct from the sediments in direct contact with ice and the basal water system. Better preservation of diatoms in the uppermost sediments suggests that it has undergone less shearing than sediments that characterize the underlying weak but cohesive till. This inferred uppermost layer has never been recovered undisturbed, probably because of high water content and lack of cohesion, as well as disturbance due to drilling.

Miocene fossils strongly dominate the diatom assemblage of the upper unit, but several sediment samples have been found to contain rare Pliocene and Quaternary marine diatoms as well (including Scherer 1991; new data). These are found exclusively in the uppermost subglacial sediments, never in the diatom-poor core samples. Post-Miocene diatoms in these sediments have been interpreted as direct evidence of previous ice-sheet collapse (Scherer 1991, 1993).

Several lines of evidence argue for a subglacial origin, rather than an eolian source, for the post-Miocene diatoms. For the sake of brevity, we list three.

A strong dominance of Miocene diatoms in the uppermost sediments, which are clearly of subglacial provenance (including diatomaceous clasts up to 0.5 centimeters), despite only trace occurrences of diatoms in the underlying till deposit, indicates that diatoms are transported from upstream source beds in a thin basal layer. This finding demonstrates that whole diatoms can be transported from upstream deposits.
Large-size Quaternary diatoms (many >60 micrometers) are as common as small diatoms, which are more easily transported by wind.
No modern nonmarine diatom specimens have been found in Upstream B sediments, despite the fact that extant nonmarine diatoms greatly outnumber marine diatoms in filtered antarctic ice (Kellogg and Kellogg 1996; R. Scherer and S. Tulaczyk, unpublished data from ice stream B).

The above observations cannot be reconciled with a purely eolian source for the Quaternary diatoms.

Diatom data indicate that sediments in contact with the ice represent a distinctly different sedimentary unit than the underlying meters-thick diamicton recovered by piston coring. The uppermost sediments beneath Upstream B are either a very thin, waterlain, and actively deforming sediment layer, the favored view, or suspended sediment-rich water. We tentatively conclude, based on diatom and other data, that the (centimeters-thick) upper sedimentary unit is mobile, probably actively deforming, and the lower, thicker (meters-thick) unit is relict and not undergoing active deep deformation as described by Alley et al. (1987). The contact between the two sedimentary units is unconformable and probably an active erosion surface. The upper unit, therefore, may be analogous to the mobile drift layer described by Alley et al. (1987), but deformation is currently restricted to a very thin layer. The lower unit may have been formed by deep deformation, certainly the diatom data suggest significant shearing, but the deposit was probably formed under a different glacial regime than present conditions.

Samples were collected with National Science Foundation funding to Kamb and Engelhardt (grant number OPP 93-19018). Funding for the analyses reported here came from Swedish Natural Sciences Research Council awards to Scherer and from Uppsala University.

References

Alley, R.B., D.D. Blankenship, C.R. Bentley, and S.T. Rooney. 1987. Till deformation beneath ice stream B. 3. Till deformation: Evidence and implications. Journal of Geophysical Research , 92(B9), 8921-8929.

Blankenship, D.D., C.R. Bentley, S.T. Rooney, and R.B. Alley. 1987. Till deformation beneath ice stream B. 1. Properties derived from seismic travel times. Journal of Geophysical Research , 92(B9), 8903-8911.

Engelhardt, H.F., and B. Kamb. In press. Basal sliding of ice stream B. Journal of Glaciology.

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.

Kellogg D., and T. Kellogg. 1996. Diatoms in South Pole ice. Geology , 24(2), 115-118.

Scherer, R.P. 1991. Quaternary and Tertiary microfossils from beneath ice stream B: Evidence for a dynamic west antarctic ice sheet history. Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planetary Change Section) , 4(4), 395-412.

Scherer, R.P. 1993. There is direct evidence for a late Quaternary collapse of the west antarctic ice sheet. Journal of Glaciology , 39(133), 716-722.

Scherer, R.P. 1994. A new method for the determination of absolute abundance of diatoms and other silt-sized sedimentary particles. Journal of Paleolimnology , 12(2), 171-180.

Scherer, R.P., D.M. Harwood, S.I. Ishman, and P.-N. Webb. 1988. Micropaleontological analysis of sediments from the Crary Ice Rise, Ross Ice Shelf. Antarctic Journal of the U.S. , 23(5), 34-36.