Climate change. Why is it hard to predict the future of ice sheets?

(IPCC) (1) highlights the improved accuracy of measurements of current sea-level rise, as well as greater certainty in the projected impacts of global warming on non-polar glaciers and thermal expansion of the oceans. These advances heighten confidence in projections of the most predictable components of sea-level rise, but the IPCC's projections specifically exclude the contribution that could arise from rapidly changing flow in ice sheets, especially in Greenland and West Antarctica. Why does so much uncertainty surround the future of ice sheets and their impact on sea-level rise? Compared with the coupled ocean-atmosphere climate system, an ice sheet might seem a rather simple system to model numerically. Ice sheets are composed of a single, largely homogeneous material. Their viscous flow is governed by the Navier-Stokes equation formulated in the mid-19th century. They move so slowly that turbulence, Coriolis, and other inertial effects can be ignored. Stresses within the ice are handled well in the latest generation of ice sheet models (2). It is in specifying the stress boundary conditions on two of the ice sheet interfaces—its base and its seaward margin—that the difficulty arises. At the base of the ice sheet, the stress resisting ice flow can vary by orders of magnitude , depending on the pressure of subglacial meltwater and the slipperiness of sediments. The transience and complexity of water flow beneath ice streams is only now becoming apparent (3). At the basal boundary , interactions among water flow, friction, sediment deformation, and heat flow become so intertwined that calculating the resistive stress from first principles tests the ingenuity of glaciologists. Nor is it certain that the basal boundary condition will remain constant on the decadal to centennial time scales that are of interest to the IPCC, especially in Greenland, where meltwater can flood through crevasses to lubricate the base of the ice sheet (4). At the margin of the ice sheet, the ice begins to float, interacts with the ocean, and eventually calves into icebergs. This boundary controls whether the ice sheet is stable to perturbations, induced perhaps by warmer oceans or atmosphere. Early theories suggested that the location of the margin might be unstable enough that a small perturbation could trigger runaway retreat inland (see the figure) (5). Since then, glaciologists have debated whether such extreme behavior could really occur. A new boundary-layer theory for coastal ice shows the way forward (6). This theory still needs to …