Governments around the world are seeking to address the economic and humanitarian consequences of climate change. One of the most graphic indications of warming temperatures is the melting of the large ice caps in Greenland and Antarctica. This is a litmus test for climate change, since ice loss may contribute more than a metre to sea-level rise over the next century, and the fresh water that is dumped into the ocean will most likely affect the ocean circulation that regulates our temperature.
Melting ice is not itself a sign of climate warming, because the ice sheet is constantly being replenished by new snow falling on its surface. The net amount of ice loss in fact results from a quite delicate imbalance between the addition of new snow and the discharge of ice to the ocean in the form of icebergs. Understanding this imbalance requires an understanding of how fast the ice moves.
Oxford Mathematician Ian Hewitt has been addressing this question using fluid-dynamical models. “We model glacial ice as a non-Newtonian viscous fluid. The central difficulty in computing the ice flow is the boundary condition at the base. In conventional fluid mechanics a no-slip condition would apply, but the presence of melt water that has penetrated through cracks in the ice acts as a lubricant and effectively allows slip - often a very significant amount of slip.”
In a recent study Ian has combined a model of the ice flow with a model for the water drainage underneath the ice to account for this varying degree of lubrication. "This new model is able to explain seasonal variations in the flow of the ice that have been observed using GPS instruments - it moves faster during spring and early summer, and slows down slightly in autumn. Intriguingly, the net effect on the ice motion over the course of the year can be both positive and negative, depending on which of these - the acceleration or subsequent deceleration - dominates. This depends on the amount of melt water that is produced on the surface, which has almost doubled over the last decade.”
Ongoing work is attempting to combine the modelled behaviour with satellite observations to better constrain what might happen in the future. Ian Hewitt talks about this research in an Oxford Sparks podcast.