Using a new biaxial friction apparatus, we conducted experiments of ice-on-rock friction in order to better understand basal sliding of glaciers and ice streams. A series of velocity-stepping and slide–hold–slide tests were conducted to measure friction and healing at temperatures between −20°C and melting. Experimental conditions in this study are comparable to subglacial temperatures, sliding rates and effective pressures of Antarctic ice streams and other glaciers, with load-point velocities ranging from 0.5 to 100 µm s −1 and normal stress σn = 100 kPa. In this range of conditions, temperature dependences of both steady-state friction and frictional healing are considerable. The friction increases linearly with decreasing temperature (temperature weakening) from μ = 0.52 at −20°C to μ = 0.02 at melting. Frictional healing increases and velocity dependence shifts from velocity-strengthening to velocity-weakening behaviour with decreasing temperature. Our results indicate that the strength and stability of glaciers and ice streams may change considerably over the range of temperatures typically found at the ice–bed interface. This article is part of the themed issue ‘Microdynamics of ice’.
In order to explore the effects of dislocations on seismic velocity and attenuation, we conducted a series of forced oscillation and ultrasonic tests on rock analogue samples (polycrystalline borneol) that were predeformed under various differential stress Δσ. Additionally, creep experiments were conducted to determine the steady‐state flow law for borneol. The dominant deformation mechanism of polycrystalline borneol changes from diffusion to dislocation creep at about Δσ=2 MPa. At high stresses, power law creep with a stress exponent of ∼4 was measured. Microstructure of the deformed samples showed wavy grain boundaries due to dislocation‐induced migration and the occasional existence of microcracks. A borneol sample deformed in the dislocation creep regime showed a significant reduction in Young's modulus E and a slight increase in attenuation Q−1 at frequencies lower than 100 Hz, whereas E at ultrasonic frequency (106 Hz) did not reduce. Therefore, a major part of the dislocation creep‐induced anelastic relaxation is a peak with a characteristic frequency between 100 and 106 Hz, which is much higher than the range of grain boundary‐induced anelasticity of this material. Further experiments under higher confining pressure are needed to assess the relative contribution from dislocations and microcracks to this peak.
Observations of glacier slip over till beds, across a range of spatial and temporal scales, show abundant seismicity ranging from microearthquakes and tremor (submeter asperities and millisecond duration) to slow‐slip events ( rupture lengths and durations). A complete understanding of the mechanisms capable of producing seismic signals in these environments represents a strong constraint on bed conditions. In particular, there is a lack of experimental confirmation of velocity‐weakening behavior of ice slipping on till, where friction decreases with increasing velocity—a necessity for nucleating seismic slip. To measure the frictional strength and stability of ice sliding against till, we performed a series of double‐direct‐shear experiments at controlled temperatures slightly above and below the ice melting point. Our results confirm velocity‐strengthening ice–till slip at melting temperatures, as has been found in the few previous studies. We provide best‐fit rate‐and‐state friction parameters and their standard deviations from averaging 13 experiments at equivalent conditions. We find evidence of similar velocity‐strengthening behavior with 50% by volume debris‐laden ice slid against till under the same conditions. In contrast, velocity‐weakening and linear time‐dependent healing of ice–till slip is present at temperatures slightly below the melting point, providing an experimentally supported mechanism for subglacial seismicity on soft‐beds. The stability parameter ( ) decreases with slip velocity, and evolution occurs over large (mm scale) displacements, suggesting that shear heating and melt buildup is responsible for the weakening. These measurements provide insight into subglacial stiffness in which seismicity of this type might be expected. We discuss glaciological circumstances pointing to potential field targets in which to test this frozen seismic asperity hypothesis.
The net rate of snow accumulation b is predicted to increase over large areas of the Antarctic and Greenland ice sheets as the climate warms. Models disagree on how this will affect the thickness of the firn layer – the relatively low-density upper layer of the ice sheets that influences altimetric observations of ice sheet mass change and palaeo-climate reconstructions from ice cores. Here we examine how b influences firn compaction and porosity in a simplified model that accounts for mass conservation, dry firn compaction, grain-size evolution, and the impact of grain size on firn compaction. Treating b as a boundary condition and employing an Eulerian reference frame helps to untangle the factors controlling the b dependence of firn thickness. We present numerical simulations using the model, as well as simplified steady-state approximations to the full model, to demonstrate how the downward advection of porosity and grain size are both affected by b but have opposing impacts on firn thickness. The net result is that firn thickness increases with b and that the strength of this dependence increases with increasing surface grain size. We also quantify the circumstances under which porosity advection and grain-size advection balance exactly, which counterintuitively renders steady-state firn thickness independent of b. These findings are qualitatively independent of the stress-dependence of firn compaction and whether the thickness of the ice sheet is increasing, decreasing, or steady. They do depend on the grain-size dependence of firn compaction. Firn models usually ignore grain-size evolution, but we highlight the complex effect it can have on firn thickness when included in a simplified model. This work motivates future efforts to better observationally constrain the rheological effect of grain size in firn.