Recent changes to Antarctic Peninsula ice shelves: What lessons have been learned?
Keywords: Antarctica, climate change, global sea level, global warming, ice flow, Larsen Ice Shelf, melting, South Pole.
C. L. HULBE Note 1Department of Geophysical Sciences, University of Chicago, 5734 S Ellis Avenue, Chicago, IL 60637, USA, firstname.lastname@example.org
Received April 8, 1997, published April 11, 1997
Summary: The past half-century has seen a dramatic increase in air temperature and the retreat of small fringing ice shelves around the Antarctic Peninsula. These changes will have little impact on the mass balance of Antarctic ice as a whole or on global sea level but do offer insight into the processes of ice shelf flow. No similar changes have been observed south of the Peninsula. Moreover, model simulations of the large Antarctic ice sheets predict an increase, not a decrease, in ice volume in future warming scenarios.
Recent changes to ice shelves around the northern Antarctic Peninsula have inspired various environmentally-minded groups to warn that Antarctic ice is about to become a victim of "global climate warming." There is clearly a connection between warming around the Antarctic Peninsula and the collapse of peninsular ice shelves. Profound ecological changes are also occurring in response to local climate change. However, temperatures in the interior of the continent have remained fairly constant (Mosely-Thompson 1992) and it is not yet known whether the observed warming is part of a global trend or is simply a normal fluctuation in local climate. Moreover, warming may actually increase the volume of ice stored in the large Antarctic ice sheets. Dramatic as the retreat of peninsular ice has been, that ice is less than 1% of the total Antarctic ice volume (Swithinbank 1988) and its maximum possible contribution to global sea level is less than 50 cm. It seems that, in the rush to demonstrate the perils of human-induced environmental degradation, much of what scientists have learned about ice in Antarctica is being ignored.
Ice shelves around the northern coasts of the Antarctic Peninsula have been retreating for the last few decades. The most recent events to gain attention occurred on the Larsen Ice Shelf, a series of small shelves fringing the eastern coast of the Peninsula from about 71 to 64 degrees S latitude (Figure 1).
The seaward front of the northernmost Larsen Ice Shelf (Larsen-A) began a gradual retreat in the late 1940’s that ended dramatically in January of 1995, when almost 2000 square km of ice disintegrated into hundreds of small icebergs during a storm (Rott et al. 1996). At the same time, a 70 km by 25 km iceberg broke off the ice front of Larsen-B, between the Jason Peninsula and Robertson Island (Figure 2). The settings and styles in which this and other peninsular ice shelves (for example, the Wordie and Müller Ice Shelves on the western side of the Peninsula; Doake and Vaughn 1991, Vaughn and Doake 1996) have retreated are different but the events all coincide with an observed 2.5 degree Celsius warming around the Antarctic Peninsula over the last 50 years.
Ice shelves are thick (hundreds of meters) platforms of floating ice that today comprise about 2% of the volume of Antarctic ice. They form where inland glaciers and ice sheets discharge into the ocean. Once afloat, the ice flows by gravity-driven horizontal spreading. Resistance to flow is provided by lateral shear where the shelf flows past bay walls and islands, and by compression upstream of sea-floor rises and islands. The rises and islands are often called "pinning points," a name which belies their importance to the stability of an ice shelf. Ice shelves gain mass by flow from inland ice, by snow accumulation on the upper surface of the shelf, and in some areas by freezing of seawater onto the lower surface. The relative importance of those contributions depends on the size and location of the shelf. Small peninsular shelves are composed almost entirely of meteoric (snowfall-derived) ice, whereas the bulk of larger shelves, such as the Ross Ice Shelf in West Antarctica, derives from inland ice. Mass is lost primarily by iceberg calving at the seaward ice cliff and secondarily by melting at the lower surface. Except in the northern Peninsula, surface melting makes a trivial contribution to mass loss. Recent estimates of the proportion of ice loss due to calving for all of Antarctica range from 75% to 90% (van der Veen 1991). The uncertainty is due to the difficulty of measuring melting rates beneath ice shelves. Small iceberg calving (less than 28 square km) is common, whereas larger events occur less frequently. The shelf front may advance for tens of years before an iceberg the size of the 1995 Larsen-B iceberg breaks off. Indeed, the current Larsen-B front position is similar to its 1960 position on the American Geographical Society map of Antarctica (1981). In the absence of external forcing, the cycle of ice advance and calving maintains a stable shelf front position over time.
Changes in local climate affect ice shelf mass balance (the accounting between the mass of ice gained and the mass of ice lost by the shelf) and thus, the size of the shelf and the location of its calving front. An increase in snowfall, either on the shelf itself or on the inland ice which feeds it, will cause the ice shelf to gain mass. A decrease in snowfall will have the opposite effect. Variations in snow accumulation on the shelf itself are felt quickly, whereas changes in the inland ice may take decades or centuries to be felt. An increase in atmospheric temperature, if large enough to push summer temperatures above the freezing point, will increase mass loss directly by increasing melting at the upper surface. Warmer sea surface temperature (SST) that may accompany warmer air temperature could also increase the rate of ice shelf melting. The indirect effects of warmer air temperature are important as well. Meltwater collecting in surface crevasses (wedge-shaped cracks in the ice that normally close by about 50 meters depth) allows the cracks to open to greater depth because water exerts a larger outward pressure on crevasse walls than does air. Water-filled crevasses may penetrate to the bottom of the ice, possibly weakening the ice shelf and hastening its decay. Warmer air and sea surface temperatures could produce a positive effect on mass balance by promoting evaporation, which in turn increases snow accumulation over inland ice and the mass of ice flowing into the shelf. Simple comparison between mean summer air temperatures and the locations of extant ice shelves led Mercer (1978) to propose a climatic limit for ice shelf stability: North of the -4 degree Celsius mean annual isotherm, ice shelves should be unstable. The present-day warming and retreat of ice shelves around the northern Antarctic Peninsula seem to confirm that hypothesis.
The effects of CO2-induced climate warming on Antarctica have been studied using numerical models that simulate ice flow and changes in ice sheet and ice shelf size over time (Huybrechts and Oerlemans 1990, Budd et al. 1994). The model predictions can be summarized in three main points. First, the glaciers and ice shelves of the Peninsula are lost. Second, the large Ross and Ronne-Filchner Ice Shelves, into which the West Antarctic Ice Sheet drains, are lost to basal melting assumed to be caused by increased SST. Third, the volume of ice in both the West Antarctic and East Antarctic ice sheets increases as a result of increased snow accumulation. This somewhat surprising result arises because CO2-induced climate warming increases SST, which decreases sea ice cover around Antarctica. As sea-ice coverage decreases, available moisture, and in turn, precipitation, increase. In one 2 x ambient CO2-induced climate warming scenario, net precipitation over Antarctica nearly doubles. Thus, while some spatial redistribution of ice occurs, there is no dramatic change in total ice volume.
Despite its small role in the mass balance of Antarctic ice as a whole, the recent collapse of Larsen-A and the ongoing changes on Larsen-B and other peninsular ice shelves offer an exciting opportunity to learn about poorly-understood ice shelf processes such as iceberg calving, ice shelf rift (crevasses that penetrate the full ice thickness) evolution, and pinning-point dynamics.
Although large iceberg calving events are routine for ice shelves, sudden disintegration is not. The unusual breakup of Larsen-A may have been the consequence of weakening caused by extreme surface melting during several consecutive warm summer seasons in the 1990’s. Water-filled crevasses could have penetrated the entire thickness of the shelf, creating a network of wounds, held together by remaining bridges between crevasses and by frozen sea water within the crevasses. The importance of sea-ice crevasse-filling was first suggested by a study of shelf-front dynamics at the Hemmen Ice Rise, near the NE corner of the Ronne-Fichner Ice Shelf (MacAyeal et al., in preparation). The study used a combination of satellite radar interferometry and numerical models of ice shelf flow. By comparing observed and model-derived interferograms, we found that the layer of sea ice and in-blown snow that fills many shelf-front rifts acts as a structurally cohesive unit, capable of transmitting stress across the rift, essentially gluing the rift walls together. However, rift-filling "glue" may be more vulnerable to warming and more likely to fail during large storm events than the thicker ice shelf ice.
The dynamical effects of the 1995 disintegration and calving events were studied using a numerical model that simulates pre-and post-calving ice shelf flow (Hulbe, unpublished report). Neither event changed significantly the stress regime or seaward flow of the remaining ice shelf. That result is due to two circumstances. First, the ice in most of the pre-calving Larsen-B iceberg was already spreading near the rate predicted for a freely-floating iceberg of equivalent thickness (Weertman 1957). Thus, to the interior of the ice shelf, the ice extending beyond the post-calving shelf front was dynamically equivalent to the sea water that replaced it. Second, the shelf maintained contact with the chain of islands along its northern boundary (the Seal Nunataks and Robertson Island). The resistance to flow provided by those pinning points stabilized the shelf front position. In a model experiment where the shelf maintained contact with Robertson Island but not with the Seal Nunataks, flow across the northern edge of the ice shelf increased by about 50%, on average, and the ice shelf thinned rapidly. The time required for complete decay cannot be computed because the volume of ice flowing into the shelf from peninsular glaciers is not well known. However, it is clear that the support provided by a small group of nunataks is key to the survival of the remaining ice.
Ice at the boundary where the shelf flows past a pinning point may be vulnerable to climate warming, although little is currently known about this subject. The ice in such shear margins is typically crevassed, is often broken into rubble, and in some instances, contains large sea-water filled voids (MacAyeal et al. 1986). Warmer sea water in the boundary voids will warm the ice internally, softening it and increasing the rate at which it can flow past the pinning point. Crevasses filled with meltwater may penetrate to the base of the shelf, thereby weakening contact between the shelf and the small islands that stabilize it. New rifts discovered during the recent Greenpeace (1997) cruise around the Antarctic Peninsula may be such sources of instability. These proposals are speculative at best and require studies, like the combined interferometry-modelling work described above, to be confirmed or denied.
What does the future hold for Antarctic ice? Unless there is a change in the observed warming trend, further retreats of fringing ice shelves along the Antarctic Peninsula are likely. This seems dramatic on the human scale but is less so on the geologic scale. The present incarnations of Antarctic Peninsula ice shelves have existed for only the last several thousand years and have, in that time, experienced cycles of advance and retreat (Clapperton 1990, Domack et al. 1995). The present glaciological events may be part of a normal long-term cycle. How ongoing changes to global climate affect the interior of Antarctica remains to be seen. The most sophisticated models available predict an increase, not a decrease, in the volume of ice in the West and East Antarctic ice sheets in future warming scenarios.