Physically Consistent Modeling of Dike-Induced Deformation and Seismicity: Application to the 2014 Bárðarbunga Dike, Iceland

Dike intrusions are often associated with surface deformation and propagating swarms of earthquakes. These are understood to be manifestations of the same underlying physical process, although rarely modeled as such. We construct a physicsâ€based model of the 2014 Bárðarbunga dike, by far the best...

Full description

Bibliographic Details
Published in:Journal of Geophysical Research: Solid Earth
Main Authors: Heimisson, Elías R., Segall, Paul
Format: Article in Journal/Newspaper
Language:unknown
Published: American Geophysical Union 2020
Subjects:
Online Access:https://doi.org/10.1029/2019jb018141
Description
Summary:Dike intrusions are often associated with surface deformation and propagating swarms of earthquakes. These are understood to be manifestations of the same underlying physical process, although rarely modeled as such. We construct a physicsâ€based model of the 2014 Bárðarbunga dike, by far the best observed large dike ( >0.5 km³) to date. We constrain the background stress state by the total dike deformation, the timeâ€dependent dike pressure from continuous GPS and the extent of the seismic swarm, and the spatial dependence of frictional properties via the spaceâ€time evolution of seismicity. We find that the geodetic and earthquake data can be reconciled with a selfâ€consistent set of parameters. The complex spatial and temporal evolution of the Bárðarbunga seismicity can be explained by dikeâ€induced elastic stress changes on preexisting faults, constrained by observed focal mechanisms. In particular, the model captures the segmentation of seismicity, where only the newest dike segment is seismically active. Our results indicate that many features of the seismicity result from the interplay between timeâ€dependent magma pressure within the dike and stress memory effects. The spatial variability in seismicity requires heterogeneity in frictional properties and/or local initial stresses. Modeling suggests that the dike pressure drops during rapid advances and increases during pauses, which primarily causes the segmentation of the seismicity. Joint analysis of multiple data types could potentially lead to improved, physicsâ€based eruption forecasts. © 2020 American Geophysical Union. Received 1 JUN 2019; Accepted 22 JAN 2020; Accepted article online 26 JAN 2020. We thank Tim Greenfield, Bob White, and Thorbjörg Ãgustsdóttir for providing access to earthquake locations and magnitude estimates prior to publication. We also thank Sigrún Hreinsdóttir for providing the 8 hr GPS time series and Andy Hooper for the processed and downsampled interferograms. We thank Jeanâ€Luc Got and an anonymous reviewer ...