S13B-2806
Time-dependent Induced Seismicity Rates Described with an Epidemic Type Aftershock Sequence Model at The Geysers Geothermal Field, California 

Monday, 14 December 2015
Poster Hall (Moscone South)
Eoghan Joseph Totten1, Christopher W Johnson1 and Roland Burgmann2, (1)Berkeley Seismological Lab, Berkeley, CA, United States, (2)University of California Berkeley, Berkeley, CA, United States
Abstract:
To improve understanding of the link between injection/production activity and seismicity, we apply an Epidemic Type Aftershock Sequence (ETAS) model to an earthquake catalog from The Geysers geothermal field (GGF) between 2005-2015 using >140,000 events and Mc 0.8 . We partition the catalog along a northeast-southwest trending divide, which corresponds to regions of high and low levels of enhanced geothermal stimulation (EGS) across the field. The ETAS model is fit to the seismicity data using a 6-month sliding window with a 1-month time step to determine the background seismicity rate. We generate monthly time series of the time-dependent background seismicity rate in 1-km depth intervals from 0-5km. The average wellhead depth is 2-3 km and the background seismicity rates above this depth do not correlate well with field-wide injected masses over the time period of interest. The auto correlation results show a 12-month period for monthly time series proximal to the average wellhead depths (2-3km and 3-4km) for northwest GGF strongly correlates with field-wide fluid injection masses, with a four-month phase shift between the two depth intervals as fluid migrates deeper. This periodicity is not observed for the deeper depth interval of 4-5 km, where monthly background seismicity rates reduce to near zero. Cross-correlation analysis using the monthly time series for background seismicity rate and the field-wide injection, production and net injection (injection minus production) suggest that injection most directly modulates seismicity. Periodicity in the background seismicity is not observed as strongly in the time series for the southeast field. We suggest that the variation in background seismicity rate is a proxy for pore-pressure diffusion of injected fluids at depth. We deduce that the contrast between the background seismicity rates in the northwest and southeast GGF is a result of reduced EGS activity in the southeast region.