Classification of transient triggering mechanisms of aftershocks in the post-seismic phase of the 2017 Pohang earthquake, South Korea

Abstract

SUMMARY The 2017 Mw 5.5 Pohang earthquake occurred near an enhanced geothermal system site and generated thousands of aftershocks, the largest of which, a Mw 4.6 earthquake, occurred 87 d after the mainshock. Redistribution of the groundwater pressure perturbed by the mainshock has been suggested as a cause of the post-seismic stress changes triggering several aftershocks, including the time-delayed event. However, to date, possible effects of variations in pore pressure on the aftershock occurrence have not been quantified in this region. Therefore, we conducted poroelastic modelling to evaluate this contribution to spatiotemporal distribution of the aftershocks, including the delayed event, using a fully coupled hydromechanical code. To construct a poroelastic model, a segmented fault geometry and a layered lithological structure were used. In addition, we utilized a kinematic slip model, a split-node algorithm and in-situ properties to simulate reliable coseismic and post-seismic behaviour. Our reference model successfully reproduced coseismic surface deformation in a line-of-sight direction, comparable to the corresponding observation from interferometric synthetic aperture radar, and was calibrated using groundwater measurement in a well. In addition to constructing the reference model, a series of numerical simulations were conducted to explore the effects and sensitivities of various hydraulic conductivities. Finally, the modelled Coulomb stress changes and spatiotemporal distribution of the aftershocks were analysed to elucidate the transient triggering mechanisms based on conditional statements to classify the mechanisms into several subsets. The classification showed that the poroelastic effect driven by depth/conductivity-dependent fluid diffusion is more critical to aftershock occurrence than the diffusion in the entire simulation time, and we propose that the delayed earthquake of Mw 4.6 could be correlated with poroelastic triggering rather than diffusion triggering. Furthermore, we inferred that this poroelastic effect could contribute to decay of aftershocks, particularly relatively small-magnitude aftershocks, as well as slow this decay in bedrocks. However, the proposed model does not explain all of the observed aftershocks, and other driving forces or triggering mechanisms need to be considered.