During the time separating two earthquakes, stress accumulates along faults until it reaches a threshold, which is different on every point of a fault. When this threshold is reached, an earthquake starts. We found a way to image the spatial distribution of the rupture threshold along the Japanese subduction faults (Bletery et al., GRL, 2017).

If the rupture threshold is the same everywhere on a large fault portion, this fault portion is more likely to rupture simultaneously (resulting in a large earthquake) than if the rupture threshold is heterogeneously distributed. Below the distribution of the gradient of the rupture threshold on the Japanese subduction faults. Red areas have low gradient and therefore homogeneous rupture threshold (the threshold is pretty much the same from one point of the fault portion to another). Accordingly, the two largest historical earthquakes in Japan occurred on the areas of most homogeneous rupture threshold.

Inferring the spatial variations of the rupture threshold along subduction faults may help assessing the potential magnitude of forthcoming earthquakes on different fault segments.