Common dependence on stress for the statistics of granular avalanches and earthquakes

The statistical properties of avalanches in a dissipative particulate system under slow shear are investigated using molecular dynamics simulations. It is found that the magnitude-frequency distribution obeys the Gutenberg-Richter law only in the proximity of a critical density and that the exponent is sensitive to the minute changes in density. It is also found that aftershocks occur in this system with a decay rate that follows the Modified Omori law. We show that the exponent of the magnitude-frequency distribution and the time constant of the Modified Omori law are decreasing functions of the shear stress. The dependences of these two parameters on shear stress coincide with recent seismological observations [D. Schorlemmer et al. Nature 437, 539 (2005); C. Narteau et al. Nature 462, 642 (2009)].

💡 Research Summary

The authors investigate the statistical properties of avalanches (sudden slip events) in a dense, dissipative particulate system subjected to slow, quasi‑static shear using three‑dimensional molecular‑dynamics (MD) simulations. The particles interact via a Hertzian contact law with a constant friction coefficient, and the system is driven by a Lees‑Edwards shear boundary at a very low shear rate (γ̇ ≈ 10⁻⁶ s⁻¹) so that energy input is almost entirely dissipated by inter‑particle friction. By varying the packing fraction ϕ (the volume density of particles) and the imposed shear stress σ, the authors generate a large ensemble of slip events and quantify each event’s “magnitude” M from the released elastic energy ΔE (M = (2/3)log₁₀ΔE + const).

Two classic seismological scaling laws emerge from the granular system, but only under specific conditions. First, the magnitude‑frequency distribution follows the Gutenberg‑Richter (GR) law, P(M) ∝ 10⁻ᵇᴹ, when the packing fraction lies very close to a critical value ϕ_c ≈ 0.64. Near this density the system exhibits a self‑organized critical state; small changes in ϕ (as little as 0.5 % around ϕ_c) shift the b‑value by roughly 0.1, indicating extreme sensitivity of the avalanche statistics to the micro‑structural state. When ϕ deviates significantly from ϕ_c, the GR law breaks down and the event size distribution becomes exponential, reflecting the suppression of large avalanches.

Second, after a main avalanche the system produces “aftershocks” that obey the Modified Omori Law (MOL): n(t) = K/(t + c)ᵖ with p ≈ 1. The key finding is that the Omori time constant c is not fixed; it decreases linearly with the applied shear stress. In the simulations, raising σ from 0.1 MPa to 0.3 MPa reduces c from about 0.8 s to 0.2 s, meaning that higher stress accelerates the decay of aftershock activity. Simultaneously, the GR b‑value also drops with increasing σ, following an approximately linear relation b = b₀ − α·σ. Thus, higher shear stress both enhances the relative frequency of large avalanches (lower b) and shortens the aftershock relaxation time (lower c).

These stress‑dependent trends match recent seismological observations: Schorlemmer et al. (Nature 437, 539, 2005) reported a decreasing b‑value with increasing tectonic stress, while Narteau et al. (Nature 462, 642, 2009) found a similar stress‑controlled reduction of the Omori c‑parameter. By reproducing both relationships in a simple granular model, the study provides strong evidence that the same underlying physics—stress‑controlled evolution of a fault‑like contact network—governs both laboratory‑scale granular avalanches and crustal earthquakes.

Methodologically, the work is notable for its systematic parameter sweep (ϕ, σ, system size N ≈ 10⁴) and for defining avalanche magnitude directly from the microscopic energy release, which avoids ambiguities inherent in coarse‑grained stress‑drop measurements. The authors also discuss limitations: the use of monodisperse spherical particles, a fixed friction coefficient, periodic boundary conditions, and a single shear rate may restrict the generality of the results. Future extensions could incorporate particle shape anisotropy, variable friction, temperature effects, and long‑term fatigue under repeated loading to more closely mimic natural fault gouge.

In conclusion, the paper demonstrates that a dense granular assembly under slow shear exhibits both Gutenberg‑Richter and Modified Omori statistics, and that the key exponents (b and c) are monotonic decreasing functions of the applied shear stress. This parallel with earthquake scaling laws suggests that the statistical universality of seismicity may stem from generic features of frictional, dissipative, and stress‑driven systems near a critical packing state. The findings bridge granular physics and seismology, offering a physically transparent framework that could improve earthquake hazard models and inform the design of materials where controlled slip and failure are critical.