Modelling the dynamics and boundary processes of Svalbard glaciers

Abstract

The focus of this thesis is on improving our understanding of surface and basal processes in the context of glaciers in Svalbard. At the surface, interactions with the atmosphere and underlying snow determine the surface mass balance. A coupled model is applied to Nordenskiöldbreen, a tidewater glacier in central Svalbard, to quantify the surface mass balance and the significance of refreezing of percolating melt water. It is found that for 1989-2010, the annual mass balance was negative (-0.39 m w.e. yr-1) and that refreezing contributed substantially to the mass budget (0.27 m w.e. yr-1). Climate sensitivity experiments reveal that seasonal inhomogeneity in future warming leads to a relatively low sensitivity of the mass balance in a changing climate. Substantial uncertainty in mass balance modeling, as well as in the interpretation of local mass balance observations, stems from the lack of detailed knowledge of how snow accumulation varies in space and time. A novel inverse approach is presented to extract accumulation from radar data. The method involves iteratively running a coupled model to simulate the firn evolution, while calibrating accumulation. Resulting annual accumulation patterns for 2007-2012 along transect on Nordenskiöldbreen reveal strong variability both in space and time. Preferential snow deposition occurs on steep slopes in the lee of terrain undulations. Small-scale accumulation variability is shown to have a negative impact on the surface mass balance. The second part of the thesis focuses on subglacial conditions. At the ice-bed interface, the rate of basal motion depends on the interplay of stresses, thermodynamics and hydrology. Regardless of variability in the external climate forcing, internal interactions may lead to feedbacks inducing periodic changes, such as have been observed in surges on Svalbard. Cyclic behavior is studied in synthetic experiments using a 3-D ice-flow model (PISM). High- and low-frequency oscillations are identified and differ in terms of volume fluctuations and changes in the polythermal basal structure. The crucial role of parameters controlling basal sliding is discussed. In a next chapter, an inverse modeling approach is used to reconstruct distributed bed topography from surface height data. Detailed knowledge of basal topography is relevant both for estimating ice volume contained in ice masses as well as for accurate time-dependent dynamical modelling. Synthetic experiments illustrate robustness of the method, whereas application of the approach to Nordenskiöldbreen demonstrates applicability of the inverse approach in a real glacier setting. In a final chapter, ongoing work towards implementation of a new hydrology model in PISM is presented and discussed. The principle aim of a hydrology model is to compute water pressures, as sliding laws used in ice dynamical models commonly relate the degree of basal friction to the effective pressure of the ice. A linked-cavity drainage model is connected to englacial storage, which enables simulating spatio-temporal water pressure variations, sub-glacial water transport and an evolving drainage system capacity. We test model behavior and discuss the potential for coupling the hydrology model to ice dynamics and application at a wide range of spatial and temporal scales.