micro-seismicity and seismic moment release within the ... › ere › pdf › igastandard › sgw...

10
PROCEEDINGS, Thirty-Ninth Workshop on Geothermal Reservoir Engineering Stanford University, Stanford, California, February 24-26, 2014 SGP-TR-202 Micro-seismicity and seismic moment release within the Coso Geothermal Field, California J. Ole Kaven a , Stephen H. Hickman a and Nicholas C. Davatzes b a USGS Earthquake Science Center, Menlo Park, CA 94025 b Earth and Environmental Science, Temple University, Philadelphia, PA 19122 [email protected] Keywords: Microseismicity, Coso, relocation, moment release. ABSTRACT We relocate 16 years of seismicity in the Coso Geothermal Field (CGF) using differential travel times and simultaneously invert for seismic velocities to improve our knowledge of the subsurface geologic and hydrologic structure. We expand on our previous results by doubling the number of relocated events from April 1996 through May 2012 using a new field-wide 3-D velocity model. Relocated micro- seismicity sharpens in many portions of the active geothermal reservoir, likely defining large-scale fault zones and fluid pressure compartment boundaries. However, a significant fraction of seismicity remains diffuse and does not cluster into sharply defined structures, suggesting that permeability is maintained within the reservoir through distributed brittle failure. The seismic velocity structure reveals heterogeneous distributions of compressional (Vp) and shear (Vs) wave speed, with Vs generally higher in the Main Field and East Flank and Vp remaining relatively uniform across the CGF, but with significant local variations. The Vp/Vs ratio appears to outline the two main producing compartments of the reservoir at depths below mean ground level of approximately 1 to 2.5 km, with a ridge of relatively high Vp/Vs separating the Main Field from the East Flank. Detailed analyses of spatial and temporal variations in earthquake relocations and cumulative seismic moment release in the East Flank reveal three regions with persistently high rates of seismic activity. Two of these regions exhibit sharp, stationary boundaries at the margins of the East Flank that likely represent barriers to fluid flow and advective heat transport. However, seismicity and moment release in a third region at the northern end of the East Flank spread over time to form an elongated NE to SW structure, roughly parallel both to an elongated cluster of seismicity at the southern end of the East Flank and to regional fault traces mapped at the surface. Our results indicate that high-precision relocations of micro-seismicity and simultaneous velocity inversions in conjunction with mapping of seismic moment release can provide useful insights into subsurface structural features and hydrologic compartmentalization within the Coso Geothermal Field. 1. INTRODUCTION Geothermal reservoirs derive their capacity for fluid and heat transport in large part from faults and fractures. Micro-seismicity generated on such faults and fractures can be used to identify larger fault structures as well as fractures that provide access to hot rock and the fluid storage and recharge capacity necessary to have a sustainable geothermal resource. Additionally, inversion for seismic velocities using micro-seismicity permits imaging of regions subject to the combined effects of variations in fracture density, liquid/steam content, lithology and other factors. The Coso Geothermal Field (CGF), located east of the Sierra Nevada batholith, is situated in a tectonically active region that features strike-slip and normal faults as well as numerous magmatic intrusions evident at the surface such as rhyolite domes (Duffield et al., 1980; Manley and Bacon, 2000). Two groups of major faults can be distinguished at the surface based on their orientation and style of faulting (Fig. 1): northwest trending faults with dextral strike-slip that form prominent lineaments with uncertain ages (Duffield et al., 1980; Unruh and Hauksson, 2006; Davatzes and Hickman, 2006) and north to northeast trending normal faults that dip both west and east and may have been active in the Quaternary (Hulen, 1978; Unruh and Hauksson, 2006; Davatzes and Hickman, 2006). These faults appear to divide the reservoir into two distinct hydrologic compartments: the Main Field and the East Flank (Fig.1), as shown by analysis of temperature logs (Kaven et al., 2011) and by clustered microseismicity (see below). These two compartments constitute the primary, productive geothermal resource within the CGF. Seismicity within the CGF occurs both tectonically and as a consequence of injection and production within the field (Monastero et al., 2005). The seismicity is recorded by the Navy Geothermal Program Office (GPO) using a combination of down-hole and surface seismometers. Initial locations by the GPO generally reveal diffuse clouds of seismicity that occur predominantly in the Main Field and the East Flank (see Kaven et al., 2013). These “clouds” do not clearly coincide with the mapped fault traces at the surface. Subsurface imaging of fault structure from reflection seismic (Monastero et al., 2005) or magnetotelluric imaging (Newman et al., 2008) lacks sufficient resolution; thus, the relationship of these earthquake locations to reservoir-scale fault structure at depth remains poorly constrained. Several studies have used data recorded by the GPO and the regional Southern California Seismic Network to analyze earthquakes within and adjacent to the CGF, improve hypocentral locations, and invert for the velocity structure of the greater Coso area (e.g. Lees, 1998; Wu and Lees, 1999; Hauksson and Unruh, 2007; Seher et al., 2011). Other studies have focused on improving hypocentral locations and obtaining moment tensors for microseismicity associated with individual hydraulic fracturing events (Foulger et al. 2008; Julian et al., 2010). In a regional study of subsurface variations in compressional (Vp) and shear (Vs) velocities from July 1993 to June 1995, Wu and Lees (1999) found low Vp/Vs ratios in the Main Field at geothermal production depths. They suggested that this anomaly might represent a hot, fluid-depleted zone. The goal of our investigation is to better constrain the subsurface geometry of faults that may act as either

Upload: others

Post on 04-Jul-2020

7 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

PROCEEDINGS, Thirty-Ninth Workshop on Geothermal Reservoir Engineering

Stanford University, Stanford, California, February 24-26, 2014

SGP-TR-202

Micro-seismicity and seismic moment release within the Coso Geothermal Field, California

J. Ole Kavena, Stephen H. Hickman

a and Nicholas C. Davatzes

b

aUSGS Earthquake Science Center, Menlo Park, CA 94025

bEarth and Environmental Science, Temple University, Philadelphia, PA 19122

[email protected]

Keywords: Microseismicity, Coso, relocation, moment release.

ABSTRACT

We relocate 16 years of seismicity in the Coso Geothermal Field (CGF) using differential travel times and simultaneously invert for

seismic velocities to improve our knowledge of the subsurface geologic and hydrologic structure. We expand on our previous results by

doubling the number of relocated events from April 1996 through May 2012 using a new field-wide 3-D velocity model. Relocated micro-

seismicity sharpens in many portions of the active geothermal reservoir, likely defining large-scale fault zones and fluid pressure

compartment boundaries. However, a significant fraction of seismicity remains diffuse and does not cluster into sharply defined structures,

suggesting that permeability is maintained within the reservoir through distributed brittle failure. The seismic velocity structure reveals

heterogeneous distributions of compressional (Vp) and shear (Vs) wave speed, with Vs generally higher in the Main Field and East Flank

and Vp remaining relatively uniform across the CGF, but with significant local variations. The Vp/Vs ratio appears to outline the two main

producing compartments of the reservoir at depths below mean ground level of approximately 1 to 2.5 km, with a ridge of relatively high

Vp/Vs separating the Main Field from the East Flank. Detailed analyses of spatial and temporal variations in earthquake relocations and

cumulative seismic moment release in the East Flank reveal three regions with persistently high rates of seismic activity. Two of these

regions exhibit sharp, stationary boundaries at the margins of the East Flank that likely represent barriers to fluid flow and advective heat

transport. However, seismicity and moment release in a third region at the northern end of the East Flank spread over time to form an

elongated NE to SW structure, roughly parallel both to an elongated cluster of seismicity at the southern end of the East Flank and to

regional fault traces mapped at the surface. Our results indicate that high-precision relocations of micro-seismicity and simultaneous

velocity inversions in conjunction with mapping of seismic moment release can provide useful insights into subsurface structural features

and hydrologic compartmentalization within the Coso Geothermal Field.

1. INTRODUCTION

Geothermal reservoirs derive their capacity for fluid and heat transport in large part from faults and fractures. Micro-seismicity generated

on such faults and fractures can be used to identify larger fault structures as well as fractures that provide access to hot rock and the fluid

storage and recharge capacity necessary to have a sustainable geothermal resource. Additionally, inversion for seismic velocities using

micro-seismicity permits imaging of regions subject to the combined effects of variations in fracture density, liquid/steam content,

lithology and other factors.

The Coso Geothermal Field (CGF), located east of the Sierra Nevada batholith, is situated in a tectonically active region that features

strike-slip and normal faults as well as numerous magmatic intrusions evident at the surface such as rhyolite domes (Duffield et al., 1980;

Manley and Bacon, 2000). Two groups of major faults can be distinguished at the surface based on their orientation and style of faulting

(Fig. 1): northwest trending faults with dextral strike-slip that form prominent lineaments with uncertain ages (Duffield et al., 1980; Unruh

and Hauksson, 2006; Davatzes and Hickman, 2006) and north to northeast trending normal faults that dip both west and east and may have

been active in the Quaternary (Hulen, 1978; Unruh and Hauksson, 2006; Davatzes and Hickman, 2006). These faults appear to divide the

reservoir into two distinct hydrologic compartments: the Main Field and the East Flank (Fig.1), as shown by analysis of temperature logs

(Kaven et al., 2011) and by clustered microseismicity (see below). These two compartments constitute the primary, productive geothermal

resource within the CGF.

Seismicity within the CGF occurs both tectonically and as a consequence of injection and production within the field (Monastero et al.,

2005). The seismicity is recorded by the Navy Geothermal Program Office (GPO) using a combination of down-hole and surface

seismometers. Initial locations by the GPO generally reveal diffuse clouds of seismicity that occur predominantly in the Main Field and the

East Flank (see Kaven et al., 2013). These “clouds” do not clearly coincide with the mapped fault traces at the surface. Subsurface imaging

of fault structure from reflection seismic (Monastero et al., 2005) or magnetotelluric imaging (Newman et al., 2008) lacks sufficient

resolution; thus, the relationship of these earthquake locations to reservoir-scale fault structure at depth remains poorly constrained.

Several studies have used data recorded by the GPO and the regional Southern California Seismic Network to analyze earthquakes within

and adjacent to the CGF, improve hypocentral locations, and invert for the velocity structure of the greater Coso area (e.g. Lees, 1998; Wu

and Lees, 1999; Hauksson and Unruh, 2007; Seher et al., 2011). Other studies have focused on improving hypocentral locations and

obtaining moment tensors for microseismicity associated with individual hydraulic fracturing events (Foulger et al. 2008; Julian et al.,

2010). In a regional study of subsurface variations in compressional (Vp) and shear (Vs) velocities from July 1993 to June 1995, Wu and

Lees (1999) found low Vp/Vs ratios in the Main Field at geothermal production depths. They suggested that this anomaly might represent a

hot, fluid-depleted zone. The goal of our investigation is to better constrain the subsurface geometry of faults that may act as either

Page 2: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

conduits or barriers to fluid migration and to define the nature and extent of hydrologic compartmentalization within the CGF. We also

seek evidence for the relative roles of discrete fault zones versus diffuse fracture networks in controlling heat and fluid transport within the

reservoir. We build on results from Kaven et al. (2011, 2012, 2013) with a more extensive analysis of the seismic data recorded by the

GPO to improve absolute and relative earthquake locations, develop a newly refined velocity structure, and calculate seismic moment for

events within the CGF.

Figure 1: Topographic map of Coso Geothermal Field (CGF), showing Navy GPO seismic network stations, geothermal injection

and production well head locations, and regional fault traces mapped by Jayko (2009). The approximate boundaries of the

two main producing reservoir compartments (the Main Field and East Flank) are shown as black polygons.

2. DATA & METHODS

We used seismic data recorded on a triggered event basis by Navy GPO from April 1996 to May 2012 at 20 permanent stations and 30

temporarily deployed stations. Most stations are three-component borehole geophones (Litton 10-11, 4.5 Hz geophones, with a few surface

stations (L22 three-component short-period seismometers). While the geophones remained identical throughout the time span of interest,

the digitizers were changed in February of 2002 from Nanometrics RD1 to Nanometrics HDR24 instruments.

As described in more detail by Kaven et al. (2013), we began our analysis by revising the Navy GPO catalog using GPO network data to

determine absolute locations and origin times for events within and adjacent to the CGF. The full GPO catalog from April 1996 to May

2012 consists of over 140,000 events, including regional events and teleseismic events. However, the GPO catalog for the region

immediately adjacent to the geothermal reservoir contains more than 80,000 events. Owing to changes in database access protocols, Navy

GPO phase picks were available only from April 1996 through October 2009; thus, phases for more recent events had to be picked

manually. In total we inspected waveforms and picked P- and S-wave arrival times for approximately 25,800 triggered events recorded by

the GPO network from November 2009 through May 2012. We located over 60% of these triggered events, resulting in roughly the same

Page 3: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

number of located events per month as are contained in the prior Navy GPO picked catalog. Combining these manually picked events with

our event locations using GPO phase picks, we obtained absolute locations for a total of 83,790 events within the CGF over from April

1996 to May 2012. To ensure that hypocentral parameters (absolute locations and origin times) are of comparable quality throughout the

period of interest (see Kissling et al., 1994), we used a consistent and reliable one-dimensional velocity model derived from bootstrapping

velocity inversions to locate all events. We tracked travel-time RMS horizontal and vertical errors, and assigned quality grades in our

absolute (single-event) locations. We found that all measures of the hypocentral parameter quality remained constant throughout the time

span considered, with quality of event locations skewed toward higher grades (A,B, which correspond to low RMS, low location errors and

small azimuthal gaps) in the years that we manually picked.

As in prior studies (Kaven et al., 2012, 2013), we obtained finer-scale images of subsurface structure and hydrologic compartmentalization

by employing high-precision relative relocation techniques. We used the revised event catalog and phase picks to derive differential travel

times for events with a minimum of 12 phase picks and offset at most 3 km from one another to relatively relocate 31,600 events and

simultaneously invert for the three-dimensional P- and S-wave velocity structure. We utilized the code tomoDD to carry out this relocation

and simultaneous velocity inversion, starting with the one-dimensional P- and S-wave velocity model from previous work (Zhang and

Thurber, 2003; Kaven et al., 2012). Although Kaven et al. (2013) used both differential travel time and waveform cross-correlation

relocation techniques for selected areas of the CGF, in the present study only differential travel times were used. Roughly 80% of all events

were retained during this relocation; rejected events did not have sufficiently similar events nearby (as determined by location offsets or

small RMS differential travel times) in sequentially more stringent relocation iteration and convergence criteria.

We calculated seismic moment (M0) and moment magnitude (Mw) using Navy GPO recordings for the newest recording epoch starting in

February 2002. We use nominal instrument responses for the geophones and HDR24 digitizers to simulate the instrument response,

integrate the velocity traces to displacement, and then integrate the first pulse following either the P- or S-wave arrival (Boatwright, 1980).

Seismic moment is calculated from the pulse using:

M0 = 4prv3r udtò (1)

following Prejean and Ellsworth (2002). We use the P- and S-wave velocity near the average source depth in the reservoir for seismic

velocity. We take the L1-norm of all stations per event to calculate event M0 and event Mw. using the relationship developed in Eaton

(1992).

3. RESULTS AND DISCUSSION

3.1 Event locations

Both absolute locations and relative relocations demonstrate that seismicity in the CGF is concentrated in the two major compartments, the

Main Field to the southwest and the East Flank to the northeast (Figs. 1, 2). Nearly all seismicity recorded within the CGF is located within

these two compartments, which are separated by a region nearly void of any seismicity through the time period covered by the revised

catalog. This lack of seismicity between the two major compartments likely indicates that both compartments are confined and fluid and

heat exchange between the two compartments may be limited. Further evidence for the separation of the two compartments is provided by

temperature data from the CGF (Fig. 4), which indicates substantially lower temperatures between these two main compartments than

within them (Kaven et al., 2011).

The results of our differential travel time relocations for the entire field over the time period April 1996 to May 2012 highlight the

refinement in hypocentral parameters made possible by the relative relocation technique (Fig. 2). Initial locations are significantly more

diffuse (Fig.2a). The relocation utilizing P- and S-wave differential travel times from phase picks significantly sharpens the distribution of

earthquakes throughout the reservoir, resulting in more distinct clustering (Fig.2b). Seismicity in the Main Field is shallower when

compared to the East Flank (Fig. 2c). Despite significant improvement in relative location precision and structural delineation when

utilizing this more comprehensive data set (compare Fig. 2b with results in Kaven et al., 2011, 2012), major through-going features remain

evasive when viewing all seismicity, even in cross-section. This indicates that a large amount of seismicity occurs on pervasive, small-

scale faults and fractures that are distributed throughout the two main CGF reservoir compartments.

Relocated seismicity in the East Flank (Fig. 3) shows this diffuse seismicity in more detail and indicates that most seismicity occurs at

depths of active injection and production (Fig. 3b). Sharp boundaries that persist over time are visible along some edges of the East Flank

compartment, especially to the south, southeast and east, past which little seismicity is observed. However, seismicity along the northern

and western edges of the East Flank is much more diffuse, and exhibits temporal migration to the northwest, especially starting around

2008. This migration presumably reflects changes in injection or production in this portion of the field. Although the migration of

microseismicity within the East Flank over time (Fig. 3) offers important clues to the geometry and extent of fluid pressure pathways and

barriers, major structural features, such as through-going faults, are still not clearly delineated in this close-up view of the East Flank.

Page 4: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

Figure 2: Earthquake locations within the Coso Geothermal Field as derived in this study over the time period April 1996 through

May 2012. a) Single-event (absolute) locations obtained from hypoinverse locations of ~32,000 events. These events were

chosen from the complete revised catalog of over 140,000 events by selecting events with a minimum of 12 phase picks (see

text). b) Relative (double-difference) relocation results for the events shown in Fig. 2a, using only events with low RMS

residuals during the inversion c) Vertical cross-section of relocated seismicity in Fig. 2b, looking to the northwest (viewing

direction shown as arrow in Fig. 2b). Depths indicated are relative to mean sea level, which is 1 km below average ground

elevation. No vertical exaggeration.

Page 5: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

Figure 3: a) Relocated seismicity in the East Flank of the CGF, color-coded by date of occurrence (oldest events in dark green,

youngest events in yellow). Also shown are locations of well pads for injection and production wells along with horizontal

projections of these wells. Portions of wells that are in hydraulic communication with the reservoir (either through an

open-hole interval or slotted-liners) are indicated by thicker lines. The approximate boundary of the East Flank

compartment is shown as a black polygon. b) Vertical cross-section of relocated seismicity and wells shown in Fig. 3a,

looking to the southwest (viewing direction shown as arrow in Fig. 3a). Depths indicated are relative to mean sea level and

there is no vertical exaggeration.

3.2 Velocity Structure

To ensure that the three-dimensional velocity structure obtained during simultaneous inversion for relative hypocentral parameters and

velocity structure is robust, we enforced a minimum ray-path crossing criteria appropriate to the grid spacing used in our model. We find

that ray path coverage, indicated by the derivative weighted sum (DWS) of ray lengths for both P- and S-waves passing through each grid

point are sufficient at most depths in the Main Field and the East Flank to allow for an inversion grid with a 200m horizontal and 500m

vertical point spacing (Fig. 4, first and third columns). Although not shown here, we also find that large-scale features, particularly in Vs,

are resolved similarly when using a larger horizontal grid spacing of 400m in the inversion and enforcing the same ray-crossing conditions.

Page 6: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

Figure 4: Velocity model of the CGF as obtained from simultaneous earthquake relocations and field wide 3D velocity inversion.

Depths of each image (relative to sea level) are indicated along the right-hand margin. First column: summed P-wave path

lengths (DWS) at inversion grid points. DWS, or derivative weighted sum refers to the sampling density at each grid points

and serves as an approximation of model resolution (Zhang and Thurber, 2007). Second column: Vp. Third column:

weighted, summed S-wave path lengths at inversion grid points. Fourth column: Vs. Fifth column: Vp/Vs ratio. Results for

Vp, Vs and Vp/Vs are not shown (white spaces) in regions of low P- or S-wave ray coverage (i.e., DWS<50).

P-wave velocity (Fig.3, second column) at all depths is highly heterogeneous, but with relatively high velocities in the region between the

Main Field and East Flank at depths relative to sea level (RSL) of less than 0.5 km (Fig.3, second column). However, at greater depths

(≥1.0 km RSL), relatively high P-wave velocities are found throughout the East Flank and to the northwest of the Main Field. The S-wave

velocity inversion indicates relatively high velocities within these two reservoir compartments (Fig. 3, fourth column), at least where

sufficient ray coverage exists above a depth of 1 km RSL. Julian et al. (2008) suggested that regions of high Vs within the shallow Main

Field may reflect decreases in fluid pressure relative to those in adjacent portions of the field. This is consistent with our observation that

relatively high Vs within the Main Field and East Flank coincides in plan view and in depth with seismicity associated with ongoing

geothermal production (e.g., Fig. 3b).

In the CGF, the ratio of P- and S-wave speeds (Vp/Vs) reflects both small-scale heterogeneity of Vp and large-scale variations in Vs

(Fig.4). The Vp/Vs ratio is commonly used to infer sub-surface variations in rock and fluid properties, with Vp/Vs being most sensitive to

changes in temperature, effective stress (normal stress minus pore pressure), porosity/pore geometry (including fractures), and fluid

saturation state (e.g., Wu and Lees, 1999). In low porosity crystalline rock, such as the granitic rocks comprising the CGF, decreases in

Page 7: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

Vp/Vs ratio are generally attributed to increases in temperature or decreases in porosity, water saturation or fluid pressure (O'Connell and

Budiansky, 1974; Chatterjee et al., 1985).

Figure 5: Comparison of Vs, Vp/Vs, and temperature (T) at a depth below sea level of 0.5 km. The first two panels are from Figure

4 (row 3) whereas the third panel is a smoothed interpolation of borehole temperature logs, normalized to the maximum

temperature recorded (after Kaven et al., 2011). Note the positive correlation between relatively high Vs, low Vp/Vs and

temperature in the Main Field and East Flank compartments.

Both the East Flank and the Main Field exhibit relatively low Vp/Vs at depths of 0 to 1 km RSL (Figs. 4 and 5), which coincides with the

densest clusters of seismicity (c.f. Fig. 2b). Seismicity associated with geothermal production is present in both these portions of the CGF

at comparable depths of 0 to 2 km RSL (Fig. 2c). Thus, we postulate that these localized decreases in Vp/Vs may result from relatively

low water saturation or fluid pressure due to ongoing geothermal production or may result from naturally elevated temperatures within the

main field and east flank. Although it is difficult to differentiate between the effects of fluid saturation (or pressure) and temperature, high

Vs in the East Flank relative to the Main Field (where temperatures are higher; Fig. 5) suggests that Vp/Vs in the CGF may be more

influenced by saturation state than by temperature. Also note the increased Vp/Vs between the East Flank and Main Field at intermediate

depths (0 to 1 km RSL). This region of high Vp/Vs is also relatively aseismic (Fig. 2), suggesting that the effects of geothermal production

on fracture porosity and/or saturation state in the region between the Main Field and East Flank may be minimal.

3.3 Seismic Moment

Seismic moments (M0) were calculated as described above for events in the revised CGF catalog, covering February 2002 through May

2012. We find that moment magnitude (Mw) scales approximately linearly with the duration magnitude (Md) calculated by the Navy GPO

prior to 2009, although they start to diverge for Mw greater than about 2. We currently are using the nominal seismometer and digitizer

responses in our Mw calculations and have only calibrated our results to a few events that are also in the Southern California Seismic

Network (SCSN) catalog. Thus, all results should be considered preliminary pending additional calibration.

We sum the moment of all events with Mw>0.5 for a particular year in 100 m squares in the East Flank (Fig. 6) to highlight regions in

which the relatively largest seismic moment release occurs. We choose Mw>0.5 to ensure that we include seismicity above the field-wide

magnitude of completeness determined by Kaven et al. (2013), thereby reducing systematic errors in moment release comparisons between

years. The resulting pattern of cumulative seismic moment reveals several regions of high cumulative moment release within the East

Flank compartment. A region of large moment release is in the southern East Flank, exhibiting a southwest to northeast trend that is most

prominent from 2004 to 2007 and again in 2009. In the central region of the East Flank a region of high cumulative seismic moment

release is prevalent in all years, but in more detail, with varying magnitude and location. Finally, starting in 2009 the northern part of the

East Flank experiences the highest cumulative seismic moment release, also trending southwest to northeast.

The consistent southwest to northeast trend of the southern and northern regions of high cumulative seismic moment release in the East

Flank likely indicate larger-scale fault and fracture zones that respond to changes in injection and production. These fault and fractures

zones are probably pre-existing features that aid in transport of fluid and heat, facilitating the growth of seismicity beyond the previous

margins of the East Flank (especially along its northern margin, see Fig 3) but may also serve as boundaries to fluid flow across them.

Page 8: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

Figure 6: Cumulative seismic moment per year in 100 m squares within the East Flank for the years shown. Horizontal projections

of injection and production wells are shown in blue and red, respectively. Grey polygon indicates the approximate

boundary of the East Flank compartment.

4. CONCLUSIONS

Careful relocation of 16 years of seismic data from the Coso Geothermal Field (CGF) provides a significantly improved catalog of absolute

and relative earthquake locations. These results produce a significant sharpening of previously more diffuse clouds of seismicity, revealing

distinct fluid pressure compartments and persistent boundaries along some sides of the productive geothermal reservoir. A significant

portion of the seismicity, however, remains diffuse even after relocation. This suggests that seismicity throughout the reservoir on

pervasive small-scale faults and fractures plays a critical role in creating and maintaining the permeability needed for viable energy

production within the CGF.

Simultaneous seismic velocity inversions and earthquake relocations in the CGF reveal heterogeneous distributions of Vp, while Vs

appears to be relatively uniform and high within the two main reservoir compartments: the Main Field and East Flank. The ratio of Vp and

Vs, Vp/Vs, is relatively low within the Main Field and East Flank, but relatively high in an elongated region separating these two

compartments. Although it is difficult to differentiate between the effects of temperature and fluid saturation (or pressure) on Vp/Vs, the

relatively low Vp/Vs observed in the two main compartments – especially in the East Flank, where Vs is higher but temperatures are lower

than in the Main Field –may reflect variations in fluid saturation state due to ongoing geothermal production.

Yearly comparisons of cumulative moment release in the East Flank compartment reveal regions of high moment release that border the

East Flank to the south and north and one that is in its interior. These regions exhibit temporal and spatial variations in moment release rate

that may correspond to changes in geothermal injection and production, as reflected for example by migration of microseismicity beyond

the East Flank’s previous northern boundary. The general northeast to southwest trend of the two regions of high moment release at the

East Flank boundaries agrees well with the regional structural fabric, suggesting that these regions may represent large-scale active faults

Page 9: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

or fracture zones. These fault and fracture zones likely aid in transporting fluid and heat to the producing regions of the reservoir and in

promoting reservoir growth, while also likely serving as boundaries to fluid and heat flow across them.

ACKNOWLEDGEMENTS

We thank the Navy GPO and Terra-Gen for permission to use their seismic data. This work was funded by the USGS Mendenhall Post-

Doctoral Fellowship program and the DOE Geothermal Technologies Program under Interagency Agreement DEEE0001501, with

additional support provided by the USGS Energy Resources and Earthquake Hazards Programs. Any use of trade, firm, or product names is

for descriptive purposes only and does not imply endorsement by the U.S. Government.

REFERENCES

Boatwright, (1980), “A spectral theory for circular seismic sources: Simple estimates of source dimension, dynamic stress drop, and

radiated seismic energy”, Bull. Seism. Soc. Am., 70, 1–28.

Chatterjee, S.N., Pitt, A.M., and Iyer, H.M. (1985), "Vp/Vs ratios in the Yellowstone National Park Region, Wyoming”, Journal of

Volcanology and Geothermal Research, 26, 213-230.

Davatzes, N. C. and Hickman, S. H.. (2006), "Stress and faulting in the Coso geothermal field: update and recent results from the East

Flank and Coso Wash”, Proceedings 31st Workshop on geothermal Reservoir Engineering, Stanford University, Stanford, California,

January 30 – February 1. 2006, SGP-TR-179, 12p.

Duffield, W. A., Bacon, C. R., and Dalrymple, G. B. (1980), "Late Cenozoic volcanism, geochronology and structure of the Coso Range,

Inyo County, California”, Journal of Geophysical Research, 85, 2381-2404.

Eaton, J. (1992), “Determination of amplitude and duration magnitudes and residuals from short-period seismograms in Northern

California”, Bull. Seism. Soc. Am., 82, 533–579.

Foulger, G.R., Julian, B.R. and Monastero, F. C. (2008), “Seismic monitoring of EGS tests at the Coso Geothermal area, California, using

accurate MEQ locations and full moment tensors”, Proceedings 33rd Workshop on geothermal Reservoir Engineering, Stanford

University, Stanford, California, January 28 – 30, 2008, SGP-TR-179, 8p.

Hauksson, E. and Unruh, J. (2007), "Regional tectonics of the Coso geothermal area along the intracontinental plate boundary in central

eastern California: Three-dimensional Vp and Vp/Vs models, spatio-temporal seismicity patterns, and seismogenic deformation”,

Journal of Geophysical Research, 112, B06309.

Hulen, J. B. (1978), "Geology and alteration of the Coso geothermal area, Inyo County, California”, U.S. Dept. of Energy: Geothermal

Energy.

Julian, B.R., Foulger, G.R. and Monastero, F. C. (2008), “Time-dependent seismic tomography and its application to the Coso Geothermal

Area, 1996-2006”, Proceedings 33rd Workshop on geothermal Reservoir Engineering, Stanford University, Stanford, California,

January 28 – 30, 2008, SGP-TR-185, 4p.

Julian, B.R., G.R. Foulger, F.C. Monastero and S. Bjornstad (2010), “Imaging Hydraulic Fractures in a Geothermal Reservoir”, Geophys.

Res. Lett., 37, L07305, 10.1029/2009GL040933

Kaven, J.O., Hickman, S.H. and Davatzes, N.C. (2011), ”Micro-Seismicity, Fault Structure, and Hydrologic Compartmentalization Within

the Coso Geothermal Field, California”, Proceedings 36rd Workshop on geothermal Reservoir Engineering, Stanford University,

Stanford, California, January 31 – February 2, 2011, SGP-TR-191, 8p

Kaven , J.O., Hickman, S.H. and Davatzes, N.C. (2012), ”Mapping of Fluid Compartments with Micro-Seismicity and Seismic Velocities

Within the Coso Geothermal Field, California”, Proceedings 37rd Workshop on geothermal Reservoir Engineering, Stanford

University, Stanford, California, January 30 – February 1, 2012, SGP-TR-194, 8p

Kaven J. O. , J.O., Hickman, S.H. and Davatzes, N.C. (2013), ”Micro-seismicity within the Coso Geothermal Field, California, from 1996

to 2012”, Proceedings 38rd Workshop on geothermal Reservoir Engineering, Stanford University, Stanford, California, February 11

– 13, 2011, SGP-TR-198, 8p

Jayko, A. S. (2009), “Surficial Geologic Map of the Darwin Hills 30’ x 60’ Quadrangle, Inyo County, California”, U.S. Geological Survey

Scientific Investigations Map 3040, scale 1: 100,000. 38

Kissling, E., Ellsworth, W.L., Eberhart-Phillips, D. and Kradolfer, U. (1994), “Initial reference models in local earthquake tomography”,

Journal of Geophysical Research, 99, no. B10, 19635-19646

Lees, J. M. (1998), "Multiplet Analyses at Coso Geothermal”, Bulletin of the Seism. Soc. America, 88, no.5, 1127-1143.

Manley, C.R. and Bacon, C.R. (2000), “Rhyolite thermobarometry and the shallowing of the magma reservoir, Coso volcanic field,

California”, Journal of Petrology, 41, p. 149-174.

Monastero, F.C., Katzenstein, A.M., Miller, S.J., Unruh, J.R., Adams, M.C. and Richards-Dinger, K. (2005), “The Coso geothermal field:

A nascent metamorphic core complex”. GSA Bull., 177(11/12), 1534-1553.

Page 10: Micro-Seismicity and Seismic Moment Release Within the ... › ERE › pdf › IGAstandard › SGW › 2014 › K… · Micro-seismicity and seismic moment release within the Coso

Kaven, Hickman and Davatzes

10

Newman, G.A., Gasperikova, E., Hoversten, G.M., and Wannamaker, P.E. (2008), “Three-dimensional magnetotelluric characterization

of the Coso geothermal field”, Geothermics, 37, p. 369-399.

Prejean, S.G. and Ellsworth, W.L. (2001), “Observations of Earthquake Source Parameters at 2km Depth in the Long Valley Caldera,

Eastern California”, Bull. of the Seism. Soc. of Am., 91, 2, 165-177

O'Connell, R.J., and Budiansky, B. (1974), "Seismic velocities in dry and saturated cracked solids”, Journal of Geophysical Research,

79, no. 35 5412-5426.

Schaff, D.P., Bokelmann, G.H.R., Beroza, G.C., Waldhauser, F., and Ellsworth, W.L. (2002) “High resolution image of the Calaveras

Fault seismicity”, Journal of Geophysical Research, 107, no.B9, 2186, doi 10.1029/2001JB000633.

Schaff, D.P., Bokelmann, G.H.R., Ellsworth, W.L., Zanzerkia, E., Waldhauser, F., and Beroza, G.C. (2004), "Optimizing correlation

techniques for improved earthquake location”, Bulletin of the Seism. Soc. America, 94, no.2, 705-7021.

Seher, T., Zhang, H. Fehler, M., Yu, H., Soukhovitskaya, V., Commer, M. and Newman, G. (2011), “Temporal Velocity Variation

beneath the Coso Geothermal Field Observed using Seismic Double Difference Tomography of Compressional and Shear Wave

Arrival Times”, GRC Transactions, 90, p. 1743-1747.