Crystal orientation fabric anisotropy causes directional hardening of the Northeast Greenland Ice Stream

Abstract

The dynamic mass loss of ice sheets constitutes one of the biggest uncertainties in projections of ice-sheet evolution. One central, understudied aspect of ice flow is how the bulk orientation of the crystal orientation fabric translates to the mechanical anisotropy of ice. Here we show the spatial distribution of the depth-averaged horizontal anisotropy and corresponding directional flow-enhancement factors covering a large area of the Northeast Greenland Ice Stream onset. Our results are based on airborne and ground-based radar surveys, ice-core observations, and numerical ice-flow modelling. They show a strong spatial variability of the horizontal anisotropy and a rapid crystal reorganisation on the order of hundreds of years coinciding with the ice-stream geometry. Compared to isotropic ice, parts of the ice stream are found to be more than one order of magnitude harder for along-flow extension/compression while the shear margins are potentially softened by a factor of two for horizontal-shear deformation.

Fig. 1. Overview of the study area and radar methods.

a Surface ice-flow velocities of the Greenland ice sheet with the locations of deep ice-core drill sites and the outline of the study area. b Onset region of the Northeast Greenland ice stream (NEGIS) and collected data. c Example of an airborne radargram crossing the ice stream near the EGRIP (East Greenland ice-core project) ice coring site, showing internal birefringence-induced extinction node lines particularly pronounced near the shear margins. As illustrated in panel d, the birefringence power extinction nodes arise in horizontally anisotropic ice through the interference of two orthogonal radar wave components travelling at slightly different wave speeds. e Two intersecting radar profiles, and f schematic example of travel-time differences of internal reflections.

Fig. 2. Spatial distribution of horizontal crystal orientation fabric (COF) anisotropy.

Depth-averaged difference in horizontal eigenvalues (Δλ) inferred by a radar crosspoint travel-time analysis, b radar beat-signature analysis, c phase-sensitive radio-echo-sounding (pRES) travel-time analysis, and modelled by a d COF evolution model implemented in Elmer/Ice, and e Specfab COF evolution model along a flow line. Both the travel-time analysis of airborne and pRES radar as well as the beat-signature approach, are sensitive to the COF orientation relative to the antenna orientation and should thus be regarded as lower-bound limits, while the results obtained from COF evolution models can be regarded as absolute values. The background shows the satellite-based surface flow velocities.

Fig. 3. Estimated flow-enhancement factors for radar-derived and modelled crystal orientation fabric (COF) compared to isotropy.

Flow-enhancement factors for along-flow pure-shear compression/extension (a–c) and horizontal shear deformation (d–f). Both compressional/extensional and shear enhancement factors are calculated from the COF estimates obtained from travel-time differences (a, d), beat-signature analysis (b, e), and Elmer/Ice-flow modelling (c, f). Note that the enhancement factors displayed in this figure are calculated in the eigenframe (COF coordinate system), since the true orientation relative to the flow direction is unknown, although it can be simulated with Elmer/Ice (see Supplementary Information 2.1 and 3). The term ‘along-flow’ therefore, refers to the direction of the smallest horizontal eigenvalue. Background in all panels shows satellite-based surface velocities.

Fig. 4. Summary of flow mechanics and evolution of the crystal orientation fabric (COF) in the ice-stream onset region.

The COF distribution in the Northeast Greenland ice stream (NEGIS) is a result of the deformation history. The dominant vertical compression outside the ice stream leads to a vertical single maximum, which rotates into the horizontal plane towards the ice-stream margins, where horizontal shear is the dominant deformation mechanism. In the upstream part of the NEGIS, ice-flow channelling and along-flow acceleration create a vertical girdle with a superimposed horizontal single maximum. Downstream of the East Greenland ice-core project (EGRIP), divergent flow and stagnant flow velocities lead to a reversed deformation, so c-axes rotate back into vertical symmetry. By the downstream end of the survey region, stagnant ice-stream width and increased flow velocities again cause along-flow extension and the transition into a girdle-type COF. The COF affects the ice viscosity, e.g. leading to considerable stiffening for pure-shear deformation in parts of the ice stream.