Seismically imaged buoyant anomalies influence segmentation of the Cascadia subduction zone

Buoyant Asthenosphere Beneath Cascadia Influences Megathrust Segmentation

Miles Bodmer, Doug R. Toomey, Emilie E. E. Hooft, and Brandon Schmandt

 

Abstract:

Great megathrust earthquakes (magnitude 8+) do not typically rupture an entire convergent margin, but rather are limited to one or more along-strike segments.  A fundamental question of subduction zone dynamics and hazard assessment is what physical properties or dynamic processes govern megathrust segmentation. Here we use onshore-offshore teleseismic delay time data to tomographically image the upper mantle seismic structure of the Cascadia subduction zone.  Our results reveal along-strike segmentation in the oceanic asthenosphere beneath the subducting plate with pronounced low-velocity anomalies below regions of increased plate locking and greater occurrence of episodic tremor and slip.   We attribute the anomalous asthenospheric velocities to independent mantle upwellings associated with hotspot-derived material in northern Cascadia and fragmentation processes at a diffuse plate-boundary in southern Cascadia. Based on these relations, we hypothesize that sub-slab buoyancy modulates the plate coupling force at the thrust interface, thereby contributing to the localization of subduction zone segmentation.

 

Plain Language Summary

Interactions between tectonic plates at subduction zones are responsible for generating the world’s largest known earthquakes (magnitudes 8+).  Subduction zones typically do not rupture along their full length, instead they release seismic energy over discrete segments.  What controls the along-strike length and location of these segments is unknown.  Here we present seismic velocity models of the Cascadia subduction zone off the western US, investigating the upper  mantle structure (50-400 km).  We identify two anomalous low-velocity regions beneath the subducting oceanic plate, one below the Olympic peninsula in northern Washington and one beneath northern California.  We infer that these are regions where mantle is rising and, due to the presence of partial melt and possibly elevated temperatures, is more buoyant than the mantle beneath central Oregon.  These buoyant regions correlate well with areas where the plates are more strongly locked together and where greater seismic tremor occurs.  We propose that the buoyant regions beneath Cascadia influence the plate coupling force at the overlaying fault and thus influence where segmentation occurs.

Article:

https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/2018GL078700