Samples are instrumented with displacement markers and thermocouples. Deformation can therefore be treated as simple shear. Models are unconfined on their downslope ends, rendering longitudinal stresses negligible. Models are confined laterally by the Perspex strongbox walls, preventing lateral extension within the sample during testing. Thus we were able to model the deformation of the lower 14 m of an ice mass approximately 50 m thick. Steel plate was placed on top of the ice model, replicating an overburden of approximately 36 m of ice (at 80 g). In the following paper we describe the physical modelling of two (laboratory-prepared) isotropic, polycrystalline ice models (0.75 × 0.25 × 0.18 m) at enhanced gravity levels (80 g) in a geotechnical beam centrifuge. Obviously the best method for investigating ice in the laboratory would be if stress gradients similar to those found in Nature could be replicated. Detailed experimental analyses of ice samples in the laboratory have until now only been undertaken using uniform stress fields in uniaxial or triaxial tests. ice thickness, gravity and ice density) which produces a stress gradient within the ice column. In Nature, ice deformation is the response of the glacier/ice sheet to the applied self-weight stress of the ice mass (i.e. Glen’s flow law is a well-established general law for steady-state glacier ice deformation, and many laboratory tests and field measurements have been undertaken which have shown the generality of the law to be correct.

Strongbox thermal expansion