Putting faults into motion by remote dynamic triggering: local ground-motion orientations seem to eclipse strain in the northern Chilean subduction forearc crust

Abstract

Transient stress perturbations caused by passing waves of distant earthquakes have been observed to activate fault slip. Observations of remotely triggered earthquakes at distances greater than ~2-3 mainshock fault lengths suggest that certain conditions promote fault activation, including large-amplitude shaking at periods below ~ 10 seconds within a geothermal setting and extensional and/or transtensional tectonics. Yet, it is still unclear if remote dynamic triggering is ubiquitous. An additional complication to determining the prevalence of triggering is that there are likely many small-magnitude earthquakes within sparsely instrumented regions that are uncataloged. Additionally, large mainshock signals can mask smaller local events that also are missing from the local catalogs. As a result, possible triggering mechanism(s) remain enigmatic. Bounding the necessary physical conditions for remote triggering, such as determining the upper or lower bounds of stress or strain amplitudes, the orientation of the seismic wave’s traversal with respect to the local stress field or fault geometry, or the geologic properties conducive to triggering can help provide clues about the physics of remote triggering.              The northern Chilean subduction margin provides an ideal setting to study remote dynamic triggering. Its dense instrumentation provides a long history of both seismic and aseismic deformation in both the subduction system and forearc faults, including the Atacama fault system. Our investigation combines a new, detailed regional earthquake catalog (2007-2021) from Sippl et al., (2023) and documented cases of triggered aseismic slip in the Atacama fault system (Victor et al., 2018). We use a twofold approach to determine the prevalence of earthquake triggering by candidate mainshocks that produce strains at our target location ranging from 1 to ~140 microstrain. The approach uses 1) a difference-of-means test of cataloged seismicity outside of the mainshock cluster (including foreshocks and aftershocks), and 2) a waveform-based approach to look for earthquake triggering at seismic stations located close to creepmeters that recorded triggered aseismic slip events.  We find a lack of evidence of persistent, statistically significant seismicity increases outside of the mainshock cluster associated with any of the candidate mainshocks.  Notably, there is an absence of significant seismicity changes outside of clustered foreshock or aftershock seismicity associated with the series of 11 M6.2-8.2 earthquakes that produced high-strain-rate events during the 2014 Iquique sequence. Seismic recordings of the 2011 M9.1 Japan earthquake at stations CX.PB01-CX.PB14 located near creep meter stations CAR3 and CH01 on the Mejillones peninsula near the Chomache fault reveal evidence of remote triggering. We observe local, uncataloged earthquakes that are only visible after applying a high pass filter that removes the mainshock signal that otherwise overprinted and swamped the local signals.  The uniformity of particle motions (circular or oblong) generated by local earthquakes on multiple stations (N=9) is lacking in most non-triggering mainshocks. This uniformity suggests that the orientation of transient stress perturbations imparted by the mainshock waveforms, in relation to the local fault orientations, may play a role in the triggering process.