Karen Mair, Espen Jettestuen, Steffen Abe

Faults, landslides and subglacial till often contain accumulations of granular debris. Their macroscopic motion is at least to some extent determined by the processes operating in this sheared granular material. A valid question in these environments is how the local behaviour at the individual granular contacts actually sums up to influence macroscopic sliding. Laboratory experiments and numerical modelling can potentially help elucidate this. Observations of jamming and unjamming as well as concentrated shear bands that appear to scale with grain size, suggest that a simple continuum description may be insufficient to capture important elements of the behaviour. We therefore seek a measure of the organisation of the granular fabric and the structure of the load bearing skeleton that shows how the individual grain interactions are summing up to influence the macroscopic sliding behaviour we observe. Contact force networks are an expression of this. We choose to investigate variability of the ‘important’ or most connected system spanning strong force networks produced in 3D discrete element models of granular layers under shear. We use percolation measures to identify, compare and track the evolution of these system spanning contact force networks. We show that the geometries and interactions of these load bearing structures, reflected as specific topological measures, are sensitive to size (and likely shape) distribution of the granular material. Importantly, distinct (measurable) changes in the topological structural characteristics of strong force networks with accumulated strain are directly correlated to fluctuations in macroscopic shearing resistance. This illustrates the sensitivity of 3D force networks to details of granular materials (such as size and shape distribution) and also the important bridging role they play between individual grain scale processes and macroscopic sliding behaviour.