Evaluation of cryo-hydrologic warming as a cause for increased ice velocities in the lower percolation zone of Sermeq Avannarleq, Southwest Greenland

Thomas Phillips, Harihar Rajaram, William Colgan, 

The area of West Greenland experiencing surface melt is increasing at a rate of
~3.9%/year, in response to a > 200m increase in equilibrium line altitude (ELA) between 1990 and 2000. Recent observations indicate that meltwater is retained in the englacial and subglacial cryo-hydrologic systems (CHS) through multiple years in this region.  Latent heat transfer from this retained meltwater has the potential to warm ice relatively rapidly (decadal time scales), a process we have termed "cryo-hydrologic warming" (CHW).  Warmer ice temperatures lead to reduced ice viscosity and potentially higher ice velocities. Interferometric Synthetic Aperture Radar (InSAR) derived ice surface velocity data from Southwest Greenland in wintertime 2001/02 and 2007/08 suggest a broad band of increased ice velocity just upstream of the mean 2000/10 equilibrium line altitude (ELA).  We incorporated CHW into a flowline thermo-mechanical model of Sermeq Avannarleq (SA), west Greenland, to evaluate CHW as a possible mechanism to explain this increase in ice velocity.  Our model also considers the influence of softer Wisconsin ice at greater depths in the ice, and basal sliding.  The dependence of the flow law parameter on depth and the local temperature was explicitly represented.  We calculate mutually consistent temperature and velocity fields accounting for subtle thermo-mechanical feedbacks such as the enhancement of horizontal advection of cold ice by basal sliding.  Our model uses measured ice surface and bedrock elevations, thus avoiding potential errors from calculating ice thickness based on mass balance.  The model incorporates CHW through the dual-column parameterization of Phillips et al. [2010]. We compare model simulations with (i) CHW active over the entire ice thickness in the lower percolation and ablation zones ("full-depth CHW"), (ii) CHW active only in the surface 80 m of the ice sheet ("surface CHW"), and (iii) "no CHW" to represent a traditional thermo-mechanical model. The spatial extent of CHW is prescribed based on the 2001 and 2007 ELA positions, and thus incorporates the upstream expansion of the ice sheet area experiencing melt over the past decade. The ice surface velocities produced by the "full-depth CHW" simulations are in good agreement with observed velocities, reproducing the increase in inland ice velocity between 2001 and 2007. The "no CHW" and "surface CHW" simulations, however, significantly underestimate observed ice surface velocities both in 2001 and 2007. The higher ice velocities in the "full-depth CHW" simulations are attributable to decreased ice viscosities associated with increased ice temperatures, as well as an increase in the extent of temperate bed conditions which permit basal sliding. The extent of temperate bed conditions predicted by the "full-depth CHW" simulations is consistent with the locations of well-documented observations of basal sliding at Sermeq Avannarleq. In contrast, the "no CHW" and "surface CHW" simulations predict temperate bed conditions over a smaller region upstream from the terminus than inferred by basal sliding observations. We also evaluate several alternative explanations for the increase in lower percolation zone ice velocity. Our results and interpretation suggest that CHW significantly contributed to increased inland ice velocities in response to an upward migration of the ELA.