Abstract
We investigate the temporal and spatial dependence of source parameters
of micro-earthquakes (<M3) before and during water injection at the
Northwest Geysers between 2005 and 2010. Our objective is to understand
the relation among injection, production and source mechanisms of
micro-earthquakes. We examine a small area that surrounds an injection
well (Prati 9) that extends into the deep zone. We utilize three
approaches to determine the source parameters of the micro-earthquakes;
the Empirical Green's Function (EGF) method (Viegas et al., 2010),
NetMoment method (Hutchings, 2002), and moment tensor inversion (Minson
and Dreger, 2008). We first compare the source parameters of 30 earthquakes
determined using the three approaches for validation purposes, and
then we determine the source parameters for all the earthquakes located
within a small volume around the well head before and during injection.
We find a good correlation coefficient of 91\% between the monthly
water injection-rate and the number of induced micro-earthquakes
located inside the small volume, with a zero time lag, indicating
that the seismic response to water injection is less than a month
time. We find the b-value in the Gutenberg-Richter law, which equates
the proportion of small earthquakes to large ones, increased from
1.3 to 1.6 with the start of water injection, indicating an increase
of the number of small earthquakes relative to larger earthquakes
due to reservoir stimulation. Our results indicate that micro-earthquakes
at the Northwest Geysers have on average stress drops (mean of 11
MPa) comparable to the ones of natural occurring tectonic earthquakes
in the region (around 17 MPa). We notice that the shape of the earthquake
cloud is slightly elongated in the SW-NE direction, consistent with
the preferential alignment direction of microcracks found in anisotropic
studies, indicating that slip is being facilitated in pre-existing
cracks. The study of micro-seismicity is a useful tool in reservoir
exploration management, as it can be used to track the release of
strain and the injected fluid flow paths, and to characterize the
permeability of the reservoir. The source information has implications
for understanding the physics of faulting and the principal mechanisms
involved in induced seismicity.
Users
Please
log in to take part in the discussion (add own reviews or comments).