An Investigation of the Accuracy of Coulomb Stress Changes Inferred From Geodetic Observations Following Subduction Zone Earthquakes

Earthquake clustering along plate boundaries suggests that earthquakes may interact, and static Coulomb stress change (CSC) is commonly invoked as one possible mechanism for stress transfer between earthquakes and nearby faults. Previous work has addressed the precision of CSC predictions that are influenced by observational noise, inversion regularization, and simplified modeling assumptions. Here we address the accuracy of CSC predictions informed by geodetic observations in subduction zones where inversion model resolution is poor. We conduct synthetic tests to quantify the degree to which the sign and magnitude of CSC can be reliably inferred from slip distributions inverted from various geodetic observations (interferometric synthetic aperture radar (InSAR), GPS, and seafloor observations). We find that in an idealized subduction zone, CSC can only be confidently inferred for receiver faults far (tens of kilometers) from the earthquake source, though this distance shortens with the addition of synthetic seafloor observations. We apply this methodology to the 2010 M(w)8.8 Maule earthquake and identify 13 aftershocks from a population of 475 documented events for which we can confidently resolve coseismic stress changes. These results demonstrate that the low model resolution of fault slip inversions in subduction zones limits our ability to address fundamental questions about earthquake sources and stress interactions. Nonetheless, our results highlight that stress change predictions are considerably more accurate after the introduction of seafloor geodetic observations. Additionally, we show that InSAR observations are not required to substantially improve stress change approximations in regions where GPS may be the only viable observation, such as in island arcs settings.