The influence of faulting style on the size-distribution of global earthquakes

We derive a unifying formulation, reliable at all scales, linking Anderson's faulting theory with the earthquake size-distribution, whose exponent is known as the b-value. Anderson's theory, introduced in 1905, related fault orientation to stress conditions. Independently, laboratory measurements on acoustic emissions have established that the applied differential stress controls their b-value. Our global survey revealed that observed spatial variations of b are controlled by different stress regimes, generally being lower in compressional (subduction trenches and continental collisional systems) and higher in extensional regimes (oceanic ridges). This confirmed previous observations that the b-value depends on the rake angle of focal mechanisms. Using a new plunge/dip-angles-based b-value analysis, we also identified further systematic influences of faulting geometry: steep normal faults (also typical of the oldest subduction zones) experience the highest proportion of smaller events, while low-angle thrust faults (typical of youngest subduction zones) undergo proportionally larger, more hazardous, events, differently from what would be expected by only allowing for rake-angle dependency. To date, however, no physical model has ever been proposed to explain how earthquakes size-distribution, differential stress and faulting styles relate to each other. Here, we propose and analytically derive a unifying formulation for describing how fault orientation and differential stresses determine b-value. Our formulation confirms that b-values decay linearly with increasing differential stress, but it also predicts a different dip-dependent modulation according to the tectonic environment, opening up new ways of assessing a region's seismic hazard.