Estimation of fault fracture energy and shear strength drop in large earthquakes: Implications for fluid pressure and tectonic regime

Earthquakes represent the most visible manifestation of tectonic stress accumulation and release within the Earth’s lithosphere. Despite their fundamental importance, the absolute stress levels that drive earthquake rupture remain poorly constrained. In particular, fault shear strength is strongly influenced by fluid pressure at depth, yet its magnitude and variability, especially within the locked zones of megathrusts and across different tectonic regimes, are still not well understood. In this study, I estimate the shear strength of the lithosphere by quantifying the total energy budget of large earthquake ruptures. The total released energy is partitioned into radiated energy and energy dissipated during fault slip, commonly referred to as breakdown (or fracture) energy (G). Radiated energy can be robustly estimated from seismic waveforms or earthquake source time functions. In contrast, reliable estimates of fracture energy are more challenging. To address this, I employ a finite-width slip-pulse model for steady-state dynamic rupture propagation (Rice et al., 2005), combined with heterogeneous kinematic rupture models of large earthquakes. The analysis is based on 208 finite-fault rupture models from the NEIC database, spanning Mw ≥ 7 earthquakes between 1990 and 2025. The adoption of a self-healing pulse rupture framework is motivated by the widely observed property that earthquake rise times are approximately an order of magnitude shorter than total rupture durations, on average about one-seventh of the rupture time in my database.

My results indicate that fracture energy (G) is strongly controlled by the heterogeneous distribution of rupture velocity, rise time, and slip during earthquake rupture. Fracture energy is underestimated by approximately a factor of five when heterogeneity in the rupture process of large earthquakes is neglected, and by nearly an order of magnitude when estimates are based on a classical crack rupture model. Assuming that megathrust earthquakes undergo an almost complete strength drop during rupture, as observed for the 2011 Tohoku earthquake, our estimates represent lower bounds on fault shear strength across global subduction zones and the oceanic lithosphere. The results reveal pronounced fault weakening during megathrust ruptures, with a global average shear strength of approximately 6 MPa. Tsunami earthquakes correspond to the weakest faults, with shear strengths on the order of ~2 MPa, implying that fluid pressures are extremely elevated across most subduction interfaces worldwide. Using these shear strength estimates, I infer a global average pore-fluid pressure ratio (λ = P_f / σ_lith) of approximately 0.9 for subduction megathrusts. In contrast, the oceanic lithosphere at mid-ocean ridges, transform faults, and fracture zones is nearly an order of magnitude stronger, indicating fluid pressures close to hydrostatic conditions. These pronounced contrasts demonstrate that fluid pressure may play a first-order role in controlling the strength of the Earth’s lithosphere.