Using remote dynamic earthquake triggering as a stress-meter of the Lower Rhine Embayment fault system in western Germany

Abstract

Faults zones in the Earth’s crust alter permeability architecture relative to country rock and can function as fluid conduits. Documented cases of long-distance earthquake interactions suggest that pore-pressure gradients resulting from conduit flow can activate seismic slip where receiver faults might be sensitive to external forcing. When external stress forcing can be quantified, for example, in the form of ground motions that can be converted to stress, it provides an opportunity to measure the stress perturbation required to nucleate slip in cases where fault activation is triggered.     In this study, we investigate the stress state of faults in the Lower Rhine Embayment (LRE), western Germany. We do so by quantifying the occurrence of remote dynamic triggering by transient stresses imparted by passing waves of distant mainshocks. The LRE hosts a system of normal faults with mean estimated slip rates of 0.1 mm/yr and moderate seismicity. We use the continuous Bensberg catalog starting in 1990 to estimate the statistical significance of seismicity rate changes surrounding teleseismic mainshocks identified as triggering candidates. We identify 21 teleseismic mainshocks with ML > 7 (1990 – 2015) and ML > 6 (2016 – present) that generate a theoretical peak-ground velocity (PGV) >0.02 cm/s within the study area. Two mainshocks associate with statistically significant seismicity-rate increases following the passing of their surface waves: the 1992 Roermond, and the 2021 M8.2 Chignik, Alaska earthquakes. Both mainshocks generated PGV values > 0.017 cm/s at 30s and have back-azimuths that are roughly parallel to the dominant strike of LRE faults. We observe a migrating sequence of earthquakes in the 10 days following the Roermond earthquake, where roughly half occur outside of the classical aftershock zone of ~2-3 fault lengths. We infer dynamic triggering to play a role in the generation of the migrating sequence, as migration outpaces diffusion time scales assuming realistic crustal diffusivity values of up to 3 m2/s. The July 2021 Alaska earthquake likely triggered a sequence of ~16 locatable earthquakes. The observed surface PGV values of the Alaska and Roermond earthquakes correspond to peak dynamic stress values of 1.4 kPa and < 30 kPa, respectively. Thus, stress values at the hypocentral depth of the triggered sequence of ~16 events inferred from 30s Rayleigh waves of the Alaska earthquake would correspond to 50-66% of the observed surface value.   Using remote dynamic triggering as a stress-meter to estimate stress thresholds that can potentially activate faults has important implications for earthquake physics, as well as for society. The LRE is being targeted for geothermal energy production. Prior work documents a series of 14 earthquakes of Mw > 5.0 since the 14th century, including the 1992 Mw 5.3 Roermond earthquake. Therefore, quantifying the triggerability of faults at a future energy production site prior to operation should be a key step in assessing the potential for fault reactivation.