Insights from Thermo-Mechanically Coupled Modeling of High-Elevation Regions of the Greenland Ice Sheet

Abstract:

As observations become more plentiful through remote sensing and numerical models become increasingly sophisticated, a clear priority of the ice sheet modeling community is to compare model simulations with observations. Temperature and velocity conditions within the Greenland ice sheet and at the bed remain largely unknown with the exception of sparse borehole measurements, but much can be inferred from rigorous thermo-mechanically coupled modeling. Surface velocities on the Greenland ice sheet are well constrained, both from satellite imagery and field observations. We take advantage of the observed surface velocities at the PARCA stakes around the 2,000m elevation contour of the ice sheet as modeling targets that represent a broad range of flow characteristics in different regions. Prescribing ice geometry, we use a two-dimensional thermo-mechanically coupled model to calculate ‘steady-state’ velocity and temperature profiles throughout the depth of the ice along flowlines from the main divide to the 2,000m elevation contour. Vertical velocity calculations are based on first principles of mass conservation, accounting for convergence and divergence of the streamtube width, and the enthalpy-based temperature calculations also incorporate the effects of liquid water content in temperate ice through the flow law parameter. Numerous insights from our simulations are presented for different regions, such as the influence of variable geothermal heat flux, the treatment of basal boundary conditions, and appropriate enhancement factors based on the age of ice. Results indicate that areas of temperate bed do exist in the high-elevation interior in certain sections of Greenland. Also highlighted is the importance of including temperature calculations in ice sheet modeling, particularly in regions with a temperate bed. For example, on the west coast, computations assuming a constant temperature of -5°C result in a 41% underestimation of the surface velocity at the 2,000m elevation location. In contrast, a full thermo-mechanical simulation produces surface velocities that agree well with observations.