Dynamic rupture propagation with laboratory derived slip-dependent constitutive laws

SUMMARY Accurately modelling earthquake rupture requires constitutive laws that capture how faults weaken during rapid slip. While dynamic simulations commonly employ the Linear Slip-Weakening (LSW) law, laboratory experiments on real fault-hosting rocks show an initial slip-hardening phase. Experiments show that the energy dissipated when stress evolves from the yield up to the peak value is not negligible and the decay from peak to residual stress is not linear. We incorporate the laboratory-derived Ohnaka Slip-dependent (OSD) law based on high-velocity friction data from Carrara marble and gabbro into 2D dynamic rupture simulations and compare rupture behavior with LSW. We compare the results under equivalent conditions of critical slip distance, breakdown work, or energy dissipated during the slip weakening. Our simulations demonstrate how the initial hardening phase affects the energy flux at the crack tip and how the weakening curve shape affects rupture propagation. There is no equivalent condition that allows for LSW rupture behavior to be similar to OSD. Indeed, OSD ruptures display broader stress fronts, smoother slip-rate evolution, larger cohesive zones and systematically lower rupture and peak slip velocities than LSW. Our results also suggest that using LSW with seismologically inferred breakdown work may overestimate fracture energy, rupture speed and high-frequency ground motions.