A comparison of inverse methods and basal sliding laws applied to a hindcast model of Jakobshavn Isbræ from 2009 to 2018.

Jakobshavn Isbræ (JI), on the West coast of Greenland, is one of the fastest flowing outlet glaciers of the Greenland Ice Sheet, draining 7% of the ice sheet area. Since the late 1990s it has dramatically accelerated, thinned and retreated in a series of phases alternating with periods of quiescence. It exhibits strong seasonal variations in flow speed and calving front position. Between 2012 and 2015, JI attained its point of furthest retreat and flow speeds in excess of 17 km/yr. Since 2016 it has modestly thickened, concurrent with deceleration and readvance of the calving front.

The very fast flow and strong annual and interannual variability present significant challenges for ice sheet modellers. We model the evolution of JI between 2009 and 2018 using the BISICLES ice sheet model. The standard modelling technique of assimilating surface velocity observations to infer a power law basal friction coefficient for a snapshot in time fails to account for rapidly changing basal conditions, underestimating the annual variability. We implement a time-series inverse method in which regular velocity observations are assimilated throughout the study period to produce a time-evolving basal friction coefficient. This method is able to reproduce the large annual variations in flow speed much more accurately than the static method.

This reliance on regular observations to drive the model poses a problem for future projections. We compare a range of sliding laws applied with the normal snapshot inverse method. A modern regularized Coulomb friction sliding law is better able to reproduce JI’s annual variations in flow speed due to its ability to modulate the basal friction in response to movement of the grounding line. As a result, it may be a more appropriate choice of sliding law for modelling the future evolution of fast-flowing outlet glaciers.