Some of the planet's most destructive earthquakes occur where one tectonic plate slides beneath another at a remarkably shallow angle, a setting in which classical Andersonian fault theory predicts insufficient stress for rupture. A team led by Satoshi Ide at the University of Tokyo has now resolved this paradox, analysing a global earthquake catalogue spanning 1976 to 2024 to test theory against half a century of observation.
The researchers focused on the b-value, the statistical ratio of small to large earthquakes, where a low value signals a higher probability of large events. Their analysis showed that fault planes with low dip angles are markedly more likely to exhibit low b-values. Calculations using optimal orientations of the tectonic stress field then demonstrated that enough stress can accumulate on such faults for giant earthquakes, meaning Andersonian theory breaks down at very low angles.
Published in Science Advances, the finding gives a theoretical basis for extending observation campaigns to low-angle fault structures that hazard assessments have tended to overlook, including features where no giant earthquake has yet been recorded. The authors argue that monitoring stress orientation, not merely stress magnitude, in candidate regions could substantially improve probabilistic forecasts of the rare, giant events that dominate seismic risk along subduction zones worldwide.