Short-term probabilistic earthquake risk assessment considering time-dependent b values

Laboratory experiments highlight a systematic b value decrease during the stress increase period before failure, and some large natural events are known to show a precursory decrease in the b value. However, short-term forecast models currently consider only the generic probability that an event can trigger subsequent seismicity in the near field. While the probability increase over a stationary Poissonian background is substantial, selected case studies have shown through cost-benefit analysis that the absolute main shock probability remains too low to warrant significant mitigation actions. We analyze the probabilities considering both changes in the seismicity rates and temporal changes in the b value. The precursory b value decrease in the 2009 L'Aquila case results in an additional fiftyfold probability increase for a M6.3 event. Translated into time-varying hazard and risk, these changes surpass the cost-benefit threshold for short-term evacuation.