Signatures of Fluid-Pressure Triggering, Natural and Induced: Comparing Migrating Earthquake Swarms in Long Valley Caldera, California and Azle, Texas

Friday, 19 December 2014: 2:40 PM
David R Shelly1, William L Ellsworth1, Emily K. Montgomery-Brown1, David P Hill1, Stephanie G Prejean2 and Margaret Mangan1, (1)US Geological Survey, Menlo Park, CA, United States, (2)Alaska Volcano Observatory Anchorage, Anchorage, AK, United States
Earthquake swarms are common in volcanic regions where fluids exsolved from magma bodies may trigger a sequence of earthquakes as they episodically transit from ductile surroundings into the neighboring brittle crust. Such swarms do not fit mainshock-aftershock patterns commonly seen with tectonic seismicity. Similar swarms are increasingly being observed in association with industrial fluid injection. Here, we investigate and compare earthquake swarms at Long Valley Caldera (a volcanic center) and Azle, Texas (near active industrial production and wastewater-disposal wells). We perform waveform-correlation-based event detection coupled with double-difference relative location, using catalog events as waveform templates to examine their space-time evolution. This greatly enhances the earthquake catalog, providing locations for ~4 times as many events as in the routine catalog, with location precision often better than 10 m. Despite differences in depth, fluid chemistry, and tectonic environment, we find pronounced spatiotemporal migration of earthquake hypocenters in both of these swarms.

Both swarms occurred on well-defined fault structures. The Long Valley south moat swarm followed an inactive, steeply dipping fault zone. Earthquakes began ~7.5 km beneath the surface on 7 July 2014, propagated dominantly upward and northward ~600 m in the first 12 hours, and continued sporadically into late July 2014. The Azle swarm of 28-29 January 2014 activated an ancient normal fault at shallow depths of ~3 km, just below the injection and extraction horizon, and propagated ~400 m northeastward along strike and bilaterally along dip over ~12 hours. For both swarms, we hypothesize that the migration of hypocenters reflects corresponding diffusion of a fluid pressure front within pre-existing fault zones. In some cases we observe larger earthquakes followed by rapid (but not instantaneous) propagation of events into previously inactive zones, perhaps reflecting enhanced permeability and accelerated fluid pressure diffusion via the newly ruptured fault patch. By integrating observations from natural and induced earthquake swarms, we aim to understand factors controlling the combined physics of faulting and fluid pressure diffusion in the crust.