Do coupled megathrusts rupture? A Global Comparison of Megathrust Coupling and Earthquake Slip

To assess seismic hazard along subduction zones, which host Earth’s largest earthquakes, geodesists routinely measure interseismic surface deformation rates and invert them to estimate distributions of slip deficit along the plate interface. The resulting geodetic coupling models highlight portions of the megathrust that are “locked” and accumulating strain, thus identifying  likely candidates for future rupture. However, inherent limitations in these models arise due to poor resolution of offshore observations leading to substantial uncertainty in shallow coupling estimates. Furthermore, recent geodetic records indicate that coupling can change significantly over just a few years, challenging the assumption that it remains stationary over interseismic periods, a critical caveat given that geodetic measurements typically span only a few decades. Beyond these observational challenges, studies have shown that high coupling is not necessarily a prerequisite for dynamic rupture: slip can penetrate creeping regions, and strongly coupled fault segments may act as rupture barriers.

To evaluate whether, and to what extent, geodetically inferred coupling correlates with coseismic slip, we perform a global comparison of slip deficit models and finite-fault slip distributions. We compile the first unified dataset of coupling models including twelve subduction zones and 61 finite-fault models of megathrust earthquakes that ruptured these margins, with a cumulative moment magnitude of 470. We discretize each slip model into a point cloud reflecting its slip distribution, allowing us to quantitatively link slip with coupling values to evaluate their correlation.

Our slip-coupling analysis reveals consistent global patterns: large megathrust earthquakes (Mw ≥ 7.5) preferentially rupture highly coupled regions, whereas smaller events show weaker coupling-slip correlations. Comparison with the null hypothesis in which slip-coupling correlation is completely random highlight that observed slip-coupling correlations are statistically significant. These findings highlight the complex interplay between coupling and rupture behavior, demonstrating that strong coupling alone does not unequivocally predict future earthquake slip patterns.