Snow slab avalanches, characterized by a distinct, broad fracture line, are released following anticrack propagation in highly porous weak snow layers buried below cohesive slabs. The anticrack mechanism is driven by the volumetric collapse of the weak layer, which leads to the closure of crack faces and to the onset of frictional contact. Here, on the basis of snow fracture experiments, full-scale avalanche measurements and numerical simulations, we report the existence of a transition from sub-Rayleigh anticrack to supershear crack propagation. This transition follows the Burridge–Andrews mechanism, in which a supershear daughter crack nucleates ahead of the main fracture front and eventually propagates faster than the shear wave speed. Furthermore, we show that the supershear propagation regime can exist even if the shear-to-normal stress ratio is lower than the static friction coefficient as a result of the loss of frictional resistance during collapse. This finding shows that snow slab avalanches have fundamental similarities with strike-slip earthquakes.

Abstract. Dynamic crack propagation in snow is of key importance for avalanche release. Nevertheless, it has received very little experimental attention. With the introduction of the propagation saw test (PST) in the mid-2000s, a number of studies have used particle tracking analysis of high-speed video recordings of PST experiments to study crack propagation processes in snow. However, due to methodological limitations, these studies have provided limited insight into dynamical processes such as the evolution of crack speed within a PST or the touchdown distance, i.e. the length from the crack tip to the trailing point where the slab comes to rest on the crushed weak layer. To study such dynamical effects, we recorded PST experiments using a portable high-speed camera with a horizontal resolution of 1280 pixels at rates of up to 20 000 frames s−1. We then used digital image correlation (DIC) to derive high-resolution displacement and strain fields in the slab, weak layer and substrate. The high frame rates enabled us to calculate time derivatives to obtain velocity and acceleration fields. We demonstrate the versatility and accuracy of the DIC method by showing measurements from three PST experiments, resulting in slab fracture, crack arrest and full propagation. We also present a methodology to determine relevant characteristics of crack propagation, namely the crack speed (20–30 m s−1), its temporal evolution along the column and touchdown distance (2.7 m) within a PST, and the specific fracture energy of the weak layer (0.3–1.7 J m−2). To estimate the effective elastic modulus of the slab and weak layer as well as the weak layer specific fracture energy, we used a recently proposed mechanical model. A comparison to already-established methods showed good agreement. Furthermore, our methodology provides insight into the three different propagation results found with the PST and reveals intricate dynamics that are otherwise not accessible.

Dry-snow slab avalanches result from crack propagation in a highly porous weak layer buried within a stratified and metastable snowpack. While our understanding of slab avalanche mechanisms improved with recent experimental and numerical advances, fundamental micro-mechanical processes remain poorly understood due to a lack of non-invasive monitoring techniques. Using a novel discrete micro-mechanical model, we reproduced crack propagation dynamics observed in field experiments, which employ the propagation saw test. The detailed microscopic analysis of weak layer stresses and bond breaking allowed us to define the crack tip location of closing crack faces, analyze its spatio-temporal characteristics and monitor the evolution of stress concentrations and the fracture process zone both in transient and steady-state regimes. Results highlight the occurrence of a steady state in crack speed and stress conditions for sufficiently long crack propagation distances (> 4 m). Crack propagation without external driving force except gravity is possible due to the local mixed-mode shear-compression stress nature at the crack tip induced by slab bending and weak layer volumetric collapse. Our result shed light into the microscopic origin of dynamic crack propagation in snow slab avalanche release that eventually will improve the evaluation of avalanche release sizes and thus hazard management and forecasting in mountainous regions.

Abstract. Dry-snow slab avalanches start with the formation of a local failure in a highly porous weak layer underlying a cohesive snow slab. If followed by rapid crack propagation within the weak layer and finally a tensile fracture through the slab, a slab avalanche releases. While the basic concepts of avalanche release are relatively well understood, performing fracture experiments in the laboratory or in the field can be difficult due to the fragile nature of weak snow layers. Numerical simulations are a valuable tool for the study of micromechanical processes that lead to failure in snow. We used a three-dimensional discrete element method (3-D DEM) to simulate and analyze failure processes in snow. Cohesive and cohesionless ballistic deposition allowed us to reproduce porous weak layers and dense cohesive snow slabs, respectively. To analyze the micromechanical behavior at the scale of the snowpack (∼1 m), the particle size was chosen as a compromise between low computational costs and detailed representation of important micromechanical processes. The 3-D-DEM snow model allowed reproduction of the macroscopic behavior observed during compression and mixed-mode loading of dry-snow slab and the weak snow layer. To be able to reproduce the range of snow behavior (elastic modulus, strength), relations between DEM particle and contact parameters and macroscopic behavior were established. Numerical load-controlled failure experiments were performed on small samples and compared to results from load-controlled laboratory tests. Overall, our results show that the discrete element method allows us to realistically simulate snow failure processes. Furthermore, the presented snow model seems appropriate for comprehensively studying how the mechanical properties of the slab and weak layer influence crack propagation preceding avalanche release.

For the release of a slab avalanche, crack propagation within a weak snowpack layer below a cohesive snow slab is required. As crack speed measurements can give insight into underlying processes, we analysed three crack propagation events that occurred in similar snowpacks and covered all scales relevant for avalanche release. For the largest scale, up to 400 m, we estimated crack speed from an avalanche movie; for scales between 5 and 25 m, we used accelerometers placed on the snow surface and for scales below 5 m, we performed a propagation saw test. The mean crack speeds ranged from 36 ± 6 to 49 ± 5 m s−1, and did not exhibit scale dependence. Using the discrete element method and the material point method, we reproduced the measured crack speeds reasonably well, in particular the terminal crack speed observed at smaller scales. Finally, we used a finite element model to assess the speed of different elastic waves in a layered snowpack. Results suggest that the observed cracks propagated as mixed mode closing cracks and that the flexural wave of the slab is responsible for the energy transfer to the crack tip.

Abstract. Dry-snow slab avalanches start with the formation of a local failure in a highly porous weak layer underlying a cohesive snow slab. If followed by rapid crack propagation within the weak layer and finally a tensile fracture through the slab appears, a slab avalanche releases. While the basic concepts of avalanche release are relatively well understood, performing fracture experiments in the lab or in the field can be difficult due to the fragile nature of weak snow layers. Numerical simulations are a valuable tool for the study of micromechanical processes that lead to failure in snow. We used a three-dimensional discrete element method (3D-DEM) to simulate and analyze failure processes in snow. Cohesive and cohesionless ballistic deposition allowed us to reproduce porous weak layers and dense cohesive snow slabs, respectively. To analyze the micromechanical behavior at the scale of the snowpack (~ 1 m), the particle size was chosen as a compromise between a low computational cost and a detailed representation of important micromechanical processes. The 3D-DEM snow model allowed reproducing the macroscopic behavior observed during compression and mixed-modes loading of dry snow slab and weak snow layer. To be able to reproduce the range of snow behavior (elastic modulus, strength), relations between DEM particle/contact parameters and macroscopic behavior were established. Numerical load-controlled failure experiments were performed on small samples and compared to results from load-controlled laboratory tests. Overall, our results show that the discrete element method allows to realistically simulate snow failure processes. Furthermore, the presented snow model seems appropriate for comprehensively studying how the mechanical properties of slab and weak layer influence crack propagation preceding avalanche release.

Abstract. For a slab avalanche to release, we need sustained crack propagation in a weak snow layer beneath a cohesive snow slab – a process we call dynamic crack propagation. Field measurements on crack propagation are very scarce. We therefore performed a series of crack propagation experiments, up to 9 m long, over a period of 10 weeks and analysed these using digital image correlation techniques. We derived the elastic modulus of the slab (0.5 to 50 MPa), the elastic modulus of the weak layer (50 kPa to 1 MPa) and the specific fracture energy of the weak layer (0.1 to 1.5 J m−2) with a homogeneous and a layered-slab model. During crack propagation, we measured crack speed, touchdown distance, and the energy dissipation due to compaction and dynamic fracture (5 mJ m−2 to 0.43 J m−2). Crack speeds were highest for experiments resulting in full propagation, and crack arrest lengths were always shorter than touchdown lengths. Based on these findings, an index for self-sustained crack propagation is proposed. Our data set provides unique insight and valuable data to validate models.

Snow slab avalanches are released following anticrack propagation in highly porous weak snow layers buried below cohesive slabs. The volumetric collapse of the weak layer leads to the closure of crack faces followed by the onset of frictional contact. Here on the basis of snow fracture experiments, full-scale avalanche measurements, and numerical simulations, we report the existence of a transition from sub-Rayleigh anticrack to supershear crack propagation involving the Burridge-Andrews mechanism. Remarkably, this transition occurs even if the shear-to-normal stress ratio is lower than the static friction coefficient as a result of the loss of frictional resistance during collapse. This finding represents a new paradigm in our understanding of snow slab avalanches presenting fundamental similarities with strike-slip earthquakes.