Characterizing Ice Shelf Responses to Climate
Change in Antarctica

Ted A. Scambos NSIDC, University of Colorado
Boulder, CO 80309

Christina L. Hulbe Dept. of Geology, Portland State University
Portland, OR 97207

Mark A. Fahnestock ESSIC, University of Maryland
College Park, MD 20742

The 1990s witnessed dramatic changes in the extent of several ice shelves in the northern Antarctic peninsula. As climate in the area warmed by ~2.5°C, a sequence of events was observed, both by satellite and in situ methods, culminating in the rapid disintegration of the shelves. The key factor associated with ice shelf disintegration appears to be the formation of melt ponds on the shelf surface, due to extended melt seasons in summer. Physical models of the ice shelves revealed that even shallow fractures, if filled with meltwater, can propagate through the ice shelf and thus lead to disintegration (an idea discussed previously by several researchers). Recently acquired Landsat 7 image data appear to support this scenario.

The close association of summertime meltwater ponding and ice shelf breakup forces a reassesment of the climate stability limit of ice shelves. The northern limit of Antarctic ice shelves is near the -5°C mean annual temperature isotherm (Mercer, 1978), and this value has been taken as the climate stability limit of all ice shelves. On this basis, many ice shelves were assumed to be quite stable climatically; for example, the Ross Ice Shelf has a mean annual temperature of about -25°C along its northern edge. However, given the importance of summer meltwater generation and ponding, we propose that a more precise climatic limit is a -2.5°C to 0°C mean January temperature. The currently-retreating ice shelves in the Peninsula have mean January temperatures of -1° to +1° C. Associated with this summer warmth is a long melt season. The Larsen "B" and Wilkins ice shelves have melt seasons of 50 to 100 days.

With this revised limit, several ice shelves may be inferred to be much closer to a retreat than previously suspected. We have begun to assess climate and firn parameters related to breakup for other shelf areas using remote sensing techniques. Specifically we are determining : mean January temperatures using in situ AWS data (where available), skin temperature using AVHRR thermal data; melt season duration using AVHRR albedo measurements; SSM/I microwave brightness temperatures; historical records of surface air temperature; and radar backscatter strength associated with the presence of a percolation or saturation zone.

Initial results show that the northeastern Ross Ice Shelf, near Roosevelt Island, has a mean January air temperature of about -5° C, moderate radar backscatter strength, and experiences extensive multiday melt events in many (but not all) years of the last two decades, but no ponding. The Larsen C ice shelf has a mean January temperature of about -1.5 °C and has very high radar backscatter, and experiences a melt season of about 50 days duration on average. Extensive ponding is not currently observed on either shelf.