Abstract
An earthquake's stress drop is related to the frictional breakdown during sliding and constitutes a fundamental quantity of the rupture process. High‐speed laboratory friction experiments that emulate the rupture process imply stress drop values that greatly exceed those commonly reported for natural earthquakes. We hypothesize that this stress drop discrepancy is due to fault‐surface roughness and strength heterogeneity: an earthquake's moment release and its recurrence probability depend not only on stress drop and rupture dimension but also on the geometric roughness of the ruptured fault and the location of failing strength asperities along it. Using large‐scale numerical simulations for earthquake ruptures under varying roughness and strength conditions, we verify our hypothesis, showing that smoother faults may generate larger earthquakes than rougher faults under identical tectonic loading conditions. We further discuss the potential impact of fault roughness on earthquake recurrence probability. This finding provides important information, also for seismic hazard analysis.
Key Points
The observed discrepancy between laboratory‐ and field‐based stress drop estimates can be linked to fault roughness and strength asperities
Field‐based estimates are biased low, not considering fault roughness. Lab‐based estimates are biased high, representing “only” asperities
Smooth faults require lower stress drops than rough ones to generate the same‐size earthquake, suggesting that the prior are, respectively, weaker