Sensitivity of Antarctic Marine Ice Sheets to Climate Change: Perspectives from GPR and Sediment Core Analysis of Ross Sea Drift in Taylor Valley, Antarctica

Prentice, M. L.¹, Arcone, S. A.², Delaney, A. J.², Horsman, J. L.³,  Sletten, R. S.♠, and Toner, J.D.♠

¹ Geoscy, LLC, Bloomington, IN; ² Cold Regions Research and Engineering Laboratory, US Army Corps of Engineers, Hanover, NH;    ³ Jencarta, LLC, Woodland Park, CO; ♠ Department of Earth and Space Sciences, University of Washington, Seattle, WA.

Funding: National Science Foundation


Geoscy collaborates on a project in Taylor Valley, Antarctica, a largely ice-free valley in the Transantarctic Mountains that faces the Ross Sea just north of the Ross Ice Shelf (Fig. 1 ). During the Last Glacial Maximum, ~18,000 to ~25,000 years ago, the ice sheets that covered West and East Antarctica were larger than at present and filled the Ross Sea with grounded ice (Fig. 2). The former ice sheet in the Ross Sea, the Ross Sea Ice Sheet, is considered unstable simply by virtue of its base being so far below current sea level (up to 1000s of meters). Because the largest component of the Ross Sea Ice sheet was contributed by ice from West Antarctica, it is often referred to as the West Antarctic Ice Sheet.

Fig. 1a. Map of Antarctica shows the Antarctic Ice Sheet divided by the Transantarctic Mountains into the West Antarctic Ice Sheet (in the western hemisphere) and the East Antarctic Ice Sheet. The Ross Sea is a major marine embayment bisected by the 180º longitude meridian and partially covered by the Ross Ice Shelf.

Fig. 1b. Topographic model of Antarctica shows the immensity of the East Antarctic Ice Sheet and the comparative smallness of the West Antarctic Ice Sheet. But, unlike in the East, the West Antarctic Ice Sheet sits on the seafloor 1000s of meters below sea level which makes it a marine ice sheet and relatively vulnerable to decay .

Fig. 2. Conceptual model of the former marine ice sheet that filled the Ross Sea at the Last Glacial Maximum*. The model shows the directions of ice flow, mainly from West Antarctica, of this Ross Sea Ice Sheet. The model also provides a narrative for the retreat of the ice sheet front (called the grounding line). At the Last Glacial Maximum, the grounding line was far south at the edge of the continental shelf. At 7600 years ago, the front was oriented east-west and located in the McMurdo Sound region (red box). Grounding line positions at subsequent times (6600, 3200, 0 years ago) are also shown. The Ross Ice Shelf is a remnant of the former ice sheet. Our work is focused in Taylor Valley which is located off of McMurdo Sound just north of the current front of the Ross Ice Shelf. Inset shows a satellite image centered on the Ross Ice Shelf.

The  Ross Sea Ice Sheet is important because it is the closest analogue to the modern West Antarctic Ice Sheet. Both are marine ice sheets which are widely regarded as vulnerable to global warming and sea level rise. A recent modeling study concluded that the Antarctic Ice Sheet, principally in West Antarctica, could collapse sufficiently by 2100 due to global warming to raise sea level by 0.5 to 1 m**. This would be catastrophic for humanity. Improved prediction of Antarctic ice response to global warming is therefore a research priority.

A major uncertainty in predictions of West Antarctic Ice Sheet change is how components of the ice sheet behave on timescales of centuries. The only records of marine ice sheet change on timescales this long are geologic records of past fluctuations in such ice sheets. One of the best records of the close-analogue Ross Sea Ice Sheet is found in the McMurdo Sound region of the Ross Sea, particularly Taylor Valley. During the Last Glacial Maximum and for several thousand years since, the Ross Sea Ice Sheet was sufficiently thick in the Ross Sea and McMurdo Sound that it advanced into places like Taylor Valley (Fig. 3).

Fig. 3. Aerial view of the edge of the Ross Sea (sea-ice covered McMurdo Sound) and  lower Taylor Valley. The floor of Taylor Valley is covered with glacial sediment (Ross Sea drift). Blue lines are a conceptual view of the flow of the former Ross Sea Ice Sheet into Taylor Valley and onto Hjorth Hill (far right).

The sediment deposited from the Ross Sea Ice Sheet is referred to as Ross Sea drift. We collected GPR evidence and sediment cores from Ross Sea drift at several locations in Taylor Valley and adjacent coastal areas in order to test some controversial interpretations and reconstruct the history of this ice sheet as it approached peak size and started collapsing. Our results have implications for century-scale marine ice sheet processes that are important to ice sheet collapse and predicting the future stability of Antarctic ice.

One location where we collected GPR profiles is a horse-shoe shaped ridge on Hjorth Hill (Figs. 4, 5-yellow zone). The prevailing interpretation is that local ice, the Wilson Piedmont Glacier, was expanded during the Last Glacial Maximum and met the expanded Ross Sea Ice Sheet along the downhill ridge of the horse-shoe (Figs. 4, 5- adjacent green and red lines).***  Accordingly, this ridge is interpreted as an interlobate moraine with sediment from both ice masses. The crucial implication of this interpretation is that the advances of the Ross Sea Ice Sheet and Wilson Piedmont Glacier to this maximum position were simultaneous. Because the Wilson Piedmont Glacier expands due to increased precipitation, a further implication is that the Ross Sea Ice Sheet advance was partially caused by increased precipitation.

Fig. 4. Aerial view of Hjorth Hill at the mouth of Taylor Valley. Yellow area outlines the horse-shoe shaped ridge at 300 m above sea level. The prevailing interpretation is that Wilson Piedmont ice expanded along the green schematic flowlines to a position marked by the green line.*** Simultaneously, the former Ross Sea Ice Sheet expanded along the blue schematic flowlines uphill to the red line. This interpretation assumes that the ridges are moraines.

Fig. 5. Aerial view of the Hjorth Hill horse-shoe shaped ridge (yellow area). Red line represents the ridge interpreted to mark the maximum uphill extent of Ross Sea drift. Green line represents the maximum extent of Wilson Piedmont drift. The junction between the two limits at 300 m is interpreted as an interlobate moraine.***

The prevailing map (Fig. 6) portrays the junction of the Ross Sea (red on map) and Wilson Piedmont (olive) drifts at 300 m as an interlobate moraine.*** Our GPR transects crisscross this ridge and the adjacent ridge which intersects it, the two together having a very curious horse-shoe shape from an aerial perspective (Fig. 5-inset). To verify our interpretation of the GPR, we recovered two cores (HH0701 and HH0702) from the downhill segment of the horse-shoe ridge.

Fig. 6. Prevailing surficial geologic map of Hjorth Hill.*** Red-colored terrain represents Ross Sea drift. Olive-colored terrain represents Wilson Piedmont drift. Inset shows GPR transects across the horse-shoe shaped ridge, mapped as moraine crests.  A detailed grid of GPR transects was used around the sites of cores HH0701 and HH0702.

We collected GPR profiles at 900-, 400-, 200-, 100-, and 50-MHz across the Hjorth Hill horse-shoe ridge (Figs. 6, 7). We used a helicopter-portable drilling system to collect sediment cores from the crest and flank of the so-called interlobate moraine (Fig. 8).

Fig.7. Aerial view of the Hjorth Hill horseshoe ridge. Red dashed line has been interpreted as interlobate moraine***. Our GPR transects parallel the yellow dashed lines and cover an area of 0.5 km².

Fig. 8. Aerial view along the ridge interpreted as interlobate moraine (dashed red line). This shows the locations of cores HH(07)01 and HH(07)02. Yellow Scott tents below the downhill side of the ridge provide scale.

The GPR and core data show that both ridges are underlain by 20-35 m of discontinuous layers of ice and sediment in roughly equal percentages (Fig. 9). Additionally, several extensive horizontal sheets of ice and lacustrine sediment are continuous under both ridges of the horse-shoe.  There is considerably less ice both within and out side of the horse-shoe. Ice layers are thickened and buckled upward under and so in phase with the ridges.

Fig. 9 100-MHz GPR profile across the Hjorth Hill horse-shoe shaped ridge. Logs of cores HH0701 (8 m long) and HH0702 (15 m long) are shown to scale on the crest and flank of the so-called interlobate moraine. The black line gives our interpretation of the bedrock surface under the ridges. Four horizontal sheets of ice, one with meters of lacustrine sediment, extend continuously under the ridge system. The ridges contain an abundance of discontinuous ice layers, many of which are bowed upward.

Based on the GPR and core data, we think that the ridges are not moraines but, rather, the remnants of the Ross Sea Ice Sheet itself and a mix of glacial sediment, after deformation by Ross ice fluctuations as well as differential ice ablation and sublimation. The glacial ice with subglacial sediment was sheared off the ice sheet edge during cycles of advance and retreat over ice-marginal sediments and probably shallow lake ice. Remarkably, the horseshoe ridge has a stratigraphy as evidenced by the extensive continuous layers that probably represent shallow lakes formed during several retreat episodes.

One preliminary conclusion from our work is that there is no Wilson Piedmont drift in the horse-shoe ridge. Combined with other evidence, we think that Ross Sea drift is more extensive than mapped in Fig. 6 and also overlies Wilson Piedmont drift elsewhere on Hjorth Hill. An implication of this conclusion is that advances of the Ross Sea Ice Sheet into McMurdo Sound and the Wilson Piedmont Glacier were out-of-phase. Because the Wilson Piedmont Glacier contracts mainly because of precipitation deficit, a further implication is that the Ross Sea Ice Sheet advance to the horse-shoe ridge maximum was not related to increased precipitation.  This leaves temperature and sea-level change as drivers that forced changes in the size of the Ross Sea Ice Sheet. These are also the drivers in play in on-going global warming.

* Conway, H., Hall, B. L., Denton, G. H., Gades, A. M., and Waddington, E. D., 1999, Past and future grounding-line retreat of the West Antarctic Ice Sheet: Science, v. 286, p. 280-283.

** DeConto, R. M., and Pollard, D., 2016, Contribution of Antarctica to past and future sea-level rise: Nature, v. 531, no. 7596, p. 591-597.

*** Hall, B. L., Denton, G. H., and Hendy, C. H., 2000, Evidence from Taylor Valley for a grounded ice sheet in the Ross Sea , Antarctica: Geografiska Annaler, v. 82A, no. 2-3, p. 275-303.