Participants:	Erik Bängtsson (Scientific Computing, Uppsala University)
	Björn Lund (Earth Sciences, Uppsala University)
	Maya Neytcheva (Scientific Computing, Uppsala University)

Summary: In recent years the demand in modelling and numerical simulation of various (very) large real-life problems has much increased. This research project addresses the above necessity and targets a particular physical phenomenon - the simulation of the response of the solid Earth to glacial advance and recession, and the resulting crustal stress field. Ongoing glacioisostatic recovery, or postglacial rebound, is registered in several regions in the world, particularly, in Northern Scandinavia. All evidence at hand indicate that glacial rebound must be taken into account in the context of other problems, such as safety assessment of long-term radioactive waste repositories, where safety should be predicted over a period of 100000 years.

The only feasible way to study the effects of post-glacial rebound on crustal dynamics is to perform numerical simulations on a computer using efficient and robust numerical techniques. The computationally heaviest part of such simulations is the solution of large linear systems of equations. Within this project we will develop and analyse novel preconditioning techniques to solve the discrete system of partial differential equations as arising in finite element modelling of glacial rebound problems, and will implement those on high performance computers, enabling in this way fast overall simulation time and studying problem details on a scale currently unavailable.

Image borrowed from: Department of Geography and Earth Sciences at Brunel University.