relative importance of processes leading to stress changes in the
TRANSCRIPT
PROCEEDINGS, Thirty-Eighth Workshop on Geothermal Reservoir Engineering
Stanford University, Stanford, California, February 11-13, 2013
SGP-TR-198
RELATIVE IMPORTANCE OF PROCESSES LEADING TO STRESS CHANGES IN THE
GEYSERS GEOTHERMAL AREA
Johannes B. Altmann1, Oliver Heidbach
1, Roland Gritto
2
1Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences,
Section 2.6 Seismic Hazard and Stress Field, Telegrafenberg, 14473 Potsdam, Germany 2Array Information Technology, USA
e-mail: [email protected]
ABSTRACT
Quantification of the different processes that
contribute to the spatio-temporal stress changes at
geothermal sites is critical for the understanding of
induced seismicity. Poroelastic coupling and thermal-
mechanical processes during hot water extraction and
cold water injection are probably the key processes
that contribute to the stress changes in The Geysers
geothermal field. In particular, with the onset of
massive injection of wastewater into the reservoir the
number of induced events with magnitudes M > 4
increased significantly. This raised the question
which stress changing process is the key control of
the induced seismicity rate. In addition to man-made
stress changes the tectonic loading from the Pacific
plate motion relative to the North America plate also
alters the stress state as well as the co-seismic stress
change of the induced events itself.
In order to assess the relative importance of the
various stress changing processes, we built a 3D
thermo-hydro-geomechanical numerical model of
The Geysers geothermal field. The model accounts
for the far-field tectonic loading as well as for the co-
seismic stress changes of the M > 4 events. To solve
the resulting fully coupled partial differential
equations we use the commercial finite element code
Abaqus. After the implementation of an initial stress
state that is calibrated against data of the orientations
of maximum horizontal stress and stress regime from
earthquake focal mechanism solutions, we apply
kinematic far-field boundary conditions derived from
continuous GPS stations. First, preliminary results
reveal that tectonic loading contributes within the
reservoir little compared to the pore-pressure
diffusion on time scales of several years.
Furthermore, the relative importance of stress
changes due to pore pressure diffusion strongly
depends on the spatial and temporal scales that are
analyzed as well as on the uncertainties of the rock
properties in particular the permeability of the rock
matrix.
INTRODUCTION
Fluid injection under high pressure to enhance
permeability of geothermal reservoirs and massive
injection of waste water to maintain reservoir is
common practice (Tester et al., 2007). These
processes are beside the hot water production, i.e.
heat extraction from the underground the key man-
made processes that alter the subsurface stress state.
However, in the past decade the accompanying
induced seismicity became an increasing problem
not only at geothermal reservoirs (Evans et al., 2012;
Majer et al., 2012, 2007) but also related to other
types of underground mining activities such as
hydrocarbon exploitation, as well as potash and coal
mining (Suckale, 2009, 2010).
Whether injection-related seismicity or production-
related seismicity is considered, both types of
seismicity are induced by a stress change. These are
caused by the injected or produced fluid. There are
two components of the fluid, leading to stress
changes. Fluid has a pore pressure component as well
as a temperature component. Fluid injection increases
the pore pressure increase and thus reduces the
normal stress along nearby faults, and thus can cause
seismicity (Byerlee, 1978). Fluid extraction and thus
pore pressure decrease results in subsidence of the
surface around the production well. The heat
extraction and injection of cold water also changes
the stress field by changing the thermal stresses.
Furthermore, depending on the predominant tectonic
regime, in extensional regimes normal faults may be
developed at the flanks of the reservoir, in
compressional regimes thrust faults may be
developed above and below the reservoir (Segall and
Fitzgerald, 1998). Heating up the cold injected fluid
cools down the hot rock and leads to volumetric
contraction. This contraction reduces the friction on
nearby faults (Eberhart-Philips and Oppenheimer,
1984) or even can create new cracks within the rock
volume, both mechanisms resulting in potential
seismicity.
Besides the man-made induced stress changes natural
processes such as plate tectonics and co-seismic
stress transfer due to major seismic events can alter
the stress state substantially. In The Geysers
geothermal field these processes can potentially
contribute significantly to the overall stress state.
To quantify the relative importance of these man-
made and natural processes we built a 3D thermo-
hydro-geomechanical numerical model of The
Geysers geothermal field that describes the stress
changes caused by fluid flow and by temperature
changes as well as the ones from natural tectonic
processes. Our model is a regional approach of the
whole reservoir in contrast to other work that focus
on single fluid injection experiment or 2D effects of
poroelastic effects (Bruel, 2007; Altmann et al.,
2010; Kohl and Megel, 2007; Rutqvist et al., 2006;
2007; Ghassemi and Zhou, 2011; Fakcharoenphol,
2012;)
As we set-up a regional model of The Geysers area,
we also consider fluid injection and production at
different locations. This enables us to model injection
and production scenarios within the entire geothermal
field. As the seismic activity clearly increased with
the start of SEGEP and SRGRP, our main interest of
investigation is on the period starting on 1997 with
SEGEP till now. We also implemented the major
fault system of The Geysers as well as the reservoir
bounding faults in order to assess at a later stage the
reactivation potential of these long fault segments
and the stress changes on these due to reservoir
production and long-term waste water injection in
comparison with tectonic loading.
In the following sections we describe briefly the
overall tectonic setting and the monitored induced
seismicity. We then present the technical and
geometrical setup of our model followed with a
description how we introduce and calibrate the initial
stress field and the loading from plate tectonics.
TECTONIC SETTING AND GEOLOGY
The Geysers geothermal field is located in northern
California, in an active tectonic region. It is
approximately 150 km north of San Francisco, about
60 km east of the San Andreas fault and 20 km south
of the Clear Lake basin.
The Geysers is located east of the Maacama fault and
is bounded to the east by the Collayomi fault.
Maacama as well as Collayomi faults are part of the
broad right-lateral San Andreas transform fault
system, where the Pacific plate adjoins the North
American plate (McLaughlin, 1981; Allis, 1982;
Donnelly et al., 1993; Boyle et al., 2011). Figure 1
shows a map of northern California, in which the
location of The Geysers is marked as solid red line.
The arrows show observed GPS velocity
measurements with respect to a fixed North
American plate. GPS-measurement were taken from
the USGS SFBay area GSP network.
Numerous authors (Donnelly et al., 1993; Bufe et al.,
1981; Oppenheimer 1986; Eberhart-Philips, 1988;
Allis and Shook, 1999) point out that seismic
measurements indicate that the Clear Lake area,
where The Geysers geothermal field is located, is
under extension. The Clear Lake basin is assumed to
be a pull-apart basin (Crowell, 1974; Hearn et al.,
1988; Donnelly et al., 1993) within the broad San
Andreas fault system. The reason for the described
complex tectonic and geological setting in the Clear
Lake area is the subduction of the Farallon plate
beneath the North American plate, which ended
about 3 Ma ago at the latitude of the Clear Lake area
(Donnelly et al., 1993; Boyle et al., 2011). After the
Farallon plate was subducted beneath the North
American plate 3 Ma ago, volcanism commenced in
the Clear Lake area ca. 2.1 Ma ago (Donnelly et al.,
1993). There is evidence that there is a large silicic
magma body beneath the Clear Lake area,
responsible for the heat anomaly of this area (Stimac
et al., 1992; Donnelly et al., 1993; Boyle et al, 2011).
An intrusion of felsic rock into the Meta-Greywacke
about 1.2 Ma ago led to the felsite layer, part of The
Geysers geothermal reservoir (Brikowski 2001). The
felsite layer varies in depth from 0.7 km depth in the
southeast Geysers up to 2.5 km depth in the
northwest Geysers (Brikowski 2001; Boyle et al.,
2011).
Figure 1: Map of The Geysers geothermal field (red
solid line) within the broad San Andreas
fault (SAF) system. Grey solid lines show
the traces of the major faults. Velocities
from continuous GPS sites show horizontal
displacement with respect to a fixed North
American plate in the interseismic phase
where the SAF is locked.
The enormous heat in the underground leads to a
natural geothermal activity in The Geysers area,
Maacama fault
Collayomi fault
SAF
where hot spring and fumaroles appear at the surface.
Due to this obvious thermal activity, already in the
1920’s The Geysers were discovered as potential
energy source and eight wells were drilled during this
time. Commercial energy production started in 1955
and was expanded since then to recently eighteen
active power plants (Lipman et al., 1978; Sanyal and
Enedy, 2011).
OBSERVATION OF INDUCED SEISMICITY
The massive development of the geothermal field as
well as a lack of both natural and artificial recharge,
the reservoir pressure began to decline in the late
1980’s. As a consequence, productivity began to
decrease (Stark et al., 2005; Goyal and Conant, 2010;
Sanyal and Enedy, 2011). In order to prevent the
reservoir pressure to further decline, two injection
projects were set up in order to pump secondary-
treated and tertiary-treated wastewater from nearby
communities into the underground (Goyal and Box,
2004; Stark et al, 2005; Goyal and Conant, 2010;
Sanyal and Enedy, 2011). The Southeast Geysers
Effluent Project (SEGEP) started in 1997, the Santa
Rosa – Geysers Recharge Project (SRGRP) started in
2003. Since then, the reservoir pressure was
stabilized.
Despite of a decreasing production rate since the end
1980’s, seismicity with events of magnitude M > 1.5
increased (Figure 2). A comparison of injection rate,
production rate and observed seismicity (Figure 3)
shows a clear correlation between injection rate and
observed seismic events of M > 1.5. Especially after
the start of the two wastewater injection projects in
1997 and 2003, seismicity increased.
Figure 2: Comparison of injection, production and
observed seismicity in The Geysers. Figure
from Hartline, 2012.
After the start of SRGRP in 2003, also the occurrence
of M > 4 earthquakes increased. But this increase
could temporally not correlated to the wastewater
injection (Majer and Peterson, 2007; Boyle et al.,
2011). Rutqvist and Oldenburg (2008) suggest
injection-related seismicity to occur due to cooling-
induced shear slip along fractures, whereas Denlinger
and Bufe (1982) suggest production-related
seismicity to occur due to temperature and pore
pressure decrease (Boyle et al., 2011).
Figure 3: Correlation between water injection and
observed seismicity (M > 1.5). Figure from
Hartline, 2012.
MODEL GEOMETRY AND DISCRETIZATION
For our geomechanical-numerical model of The
Geysers region we have chosen a rectangular area of
approximately 60 km by 40 km that is oriented in the
direction of the horizontal velocities derived from the
continuous GPS station P195 (Figure 4).
Figure 4: Rectangle shows the location and size of
the geomechanical-numerical model. The
orientation of the model is such that it
aligned along the horizontal GPS velocity
of station P195. Red arrows and triangles
show the kinematic boundary conditions
that will be applied to the model to realize
tectonic loading; triangles denote that the
boundary is fixed.
The size for the geomechanical-numerical model is
chosen at this regional scale in order to model the far
field effect of tectonic loading on the state of stress
within The Geysers and on the reservoir bounding
faults, and that the distance between the model
boundaries and the area of interest (reservoir) is large
enough to avoid any boundary effects on the results.
However, the model is small enough to have a high
resolution in the center of the model to implement
also injection and production points (Figure 5).
Figure 5: Finite element model of The Geysers area.
The reservoir (blue layer) is located in the
model center. Seven major faults included
are: Maacama fault (MF), two branches
of the Geyser Peak fault (GPF), two
branches of the Wight Way fault (WWF),
Cobb Mountain fault (CMF), Collayomi
fault (CF).
The model geometry consists of three horizontal
stratigraphic layers, representing the Greywacke
layer from the surface to a depth of approximately
2.5 km, the Felsite layer down to a depth of 5 km and
the upper crust from 5 km to a depth of 15 km
(Figure 5 and 6).
Figure 6: Geothermal reservoir of The Geysers on
top of the Felsite layer, realized in the
geomechanical-numerical model.
The model geometry also contains seven major faults
of this area, which are implemented as vertical
contact surfaces, intersecting the entire model from
the surface to the bottom in 15 km depth. The
geothermal reservoir itself is located in the center of
the geomechanical-numerical model on top of the
Felsite layer.
In order to solve the fully coupled differential
equations of poroelasticity and thermal diffusion we
discretize the model geometry into 1.2 million
tetrahedral poroelastic elements, to which pore
pressure and temperature can be addressed to. The
resolution increases from 25 m around the injection
and production points to 200 m at the reservoir
boundary up to 2.5 km at the model boundaries.
In order to simulate the effect of fluid injection and
withdrawal on the pore pressure field and on the state
of stress, we included 65 injection and production
points in total. Injection points are located at SEGEP
and SRGRP injection wells, production points around
the power plants. Figure 7 shows a part of the
injection and production points within the reservoir.
Figure 7: Injection and production sites are indicated
as yellow dots, rupture planes of the
induced events with magnitude four and
larger are shown as inclined purple
planes. Faults are represented as large
vertical, colored planes.
Additionally, we included ten rupture planes of
magnitude 4 earthquakes that occurred since the start
of SEGEP in end of 1997 as planar contact surfaces.
For the orientations of the rupture planes we used
strike and dip from earthquake moment tensor
solutions. This gives us the possibility to model the
effect of co-seismic slips on the state of stress.
NUMERICAL MODELING OF STRESS AND
STRESS CHANGES
The model process consists of several consecutive
steps with increasing model complexity. First we
start with the definition of an appropriate initial stress
state of the model area. Then we apply the lateral
kinematic boundary conditions to simulate the
stressing of past tectonic loading. These two initial
phases will deliver the starting stress field before
fluid injection and production are considered.
15 k
m
MF
GPF
WWF
CMF
CF
Initial State Of Stress
The k-ratio as the ratio of the mean horizontal stress
to vertical stress can be used as parameter to describe
the reference stress state of the Earth’s crust. The
compilation of k-ratios presented in Figure 8 shows
that it varies in the first two kilometers significantly.
Measurements from the KTB site (Brudy et al., 1997)
and the SAFOD pilot hole (Hickman and Zoback,
2004) show that the k-ratio is decreasing with depth
and tends to converge to lithostatic (k=1) at greater
depth. As SAFOD is in an tectonic very active region
the k-ratio is much clearly higher due to higher
horizontal stresses compared to the KTB site that is
in central Europe in a tectonically inactive region.
The simplest way to define an initial state of stress is
obtained by applying gravity to a geomechanical-
numerical model with the model bottom and the
model sides fixed in normal directions. However, this
procedure results in an uniaxial stress state that is not
justfied by comparing the stress state with field data
(Figure 8).
In order to define a reference stress state Sheorey
(1994) presents a one-dimensional analytical elasto-
static thermal earth model. He parameterizes the
structure of his earth model in terms of thermal
gradient, thermal expansion coefficient varying with
depth and temperature, gravity and elastic constant
varying with depth. He takes into account the effects
of thermal expansion and gravitational compaction,
and derives an expression for the k-ratio for the
uppermost kilometers of the earth’s crust
(
) (1)
where E is the Young’s modulus in GPa and H depth
in m. This k-ratio is valid for areas with no lateral
density changes and no active tectonics. For
E=90 GPa, the value for the KTB site (Brudy et al.,
1997), this k-ratio fits very well the measured data
from the KTB site (Figure 8). This implies that the
simple uniaxial boundary conditions would not result
in an appropriate initial stress state. Thus, in our
model we thus use the k-ratios that result from the
Sheory model (1994) as it fits the observed data best.
In practical terms of modeling, we build a box around
The Geysers geomechanical-numerical model with
model boundaries inclined to the center of the earth.
In the following, this box is called Sheorey-box. The
inclined model boundaries increases the horizontal
stresses compared to a model with vertical model
boundaries, when the model compacts due to the
applied gravity forces.
The Sheorey-box consists of two layers, an upper
load frame with the same depth like The Geysers
model and a lower layer called compaction layer.
With Equation (1) Sheorey (1994) gives an estimate
of the k-ratio in the upper crust. This formula is valid
for an area without tectonic stresses and with no
lateral density variations. Thus, Equation (1) is used
to calculate an initial state of stress. E is known for
the different layers of The Geysers model, so that we
are able to calculate a k-ratio dependent on H. This
gives us theoretical curves of the k-ratio, one for each
Young’s modulus of the model.
Figure 8: Worldwide measurements of k-ratios
compared to different model of reference
stress states (for details see Hergert &
Heidbach, 2011 and references given
therein).
Figure 9: Geomechanical-numerical model of The
Geysers with round Sheorey-box.
Now, gravity is applied, and from the resulting stress
field k-ratios along vertical paths are calculated,
which are in areas of no lateral density variations.
We use the finite element solver ABAQUS to solve
the resulting poroelastic partial differential equations
in order to perform our modeling studies. We fit this
modeling results to the theoretical curves calculated
after Equation (1). The modeled k-ratios can be
improved by variation of the Poisson’s ratios of The
Geysers model and of the load frame, and by
variation of the Young’s modulus of the compaction
layer.
Figure 10 shows the best fit of modeled and
theoretical k-ratios along two vertical profiles. There
load
frame
compaction
layer
dep
th [
km
]
are three different curves, because of three different
layers in the model, which have different Young's
moduli.
Figure 10: Comparison of k-ratio curves after
Sheorey (1994) (solid lines) with modeled
k-ratios along two vertical paths (location
1, location 2). The difference between the
modeled k-ratios in depths between 2 km
and 3 km is due to different thicknesses of
Greywacke and Felsite layers.
This result is used as the initial stress state of the
model area, the basis which in the following tectonic
loading and fluid injection and production is applied
on. The initial stress state does not consider any
stress changes due to tectonics. As The Geysers is
located in an active tectonic region, tectonic loading
will change the initial stress state. In the following,
we apply boundary conditions to the model,
representing tectonic loading in this area.
Stress Change Due To Tectonic loading
The Pacific plate moves north-northwest with 51.5
mm/a with respect to a fixed North American plate.
In The Geysers area, this movement decreases from
31 mm/a at the Maacama fault to 21 mm/a east of the
Collayomi fault. This means a difference of 10 mm/a
within our modeling area (Figure 4). Therefore, to
model this effect, the southwestern model boundary
is aligned along the measured GPS-vector, so that we
can apply a shear boundary condition of 10 mm/a to
this model boundary. The northeastern model
boundary is fixed, and along the other model
boundaries we apply a linearly decreasing velocity
from 10 mm/a in the west to zero in the east. These
boundary conditions are marked as red arrows and
triangles in Figure 4.
For our modeling approach, we use poroelastic
material properties, given in Table 1.
Table 1: Poroelastic material properties with void
ratio n, hydraulic conductivity kf, Poisson's
ratio, bulk modulus of the solid grains Kg,
bulk modulus of the wetting liquid Kf and
drained bulk modulus Kd. Layer Grey-
wacke
Reservoir Felsite Upper
Crust
n (%) 1.5 1.5 1.5 1.5
kf (m/s) 1.0E-10 5.0E-8 5.0E-10 1.0E-10
0.32 0.35 0.32 0.30
Kg (GPa) 40 50 75 100
Kf (GPa) 2.1 2.1 2.1 2.1
Kd (GPa) 18 26 37 50
After applying these boundary conditions to the
model, the initial state of stress has changed. We use
the regime stress ratio (RSR) after Simpson (1997)
and the orientation of the maximum horizontal stress
to compare the modeling results with observed data.
The RSR has values between 0 and 3 and is a
continuous scale for the stress regime, covering radial
extension (RSR=0), extension or normal faulting
(RSR=0.5), transtension (RSR=1.0), strike slip
(RSR=1.5), transpression (RSR=2.0), compression or
thrust faulting (RSR=2.5) and constriction
(RSR=3.0).
Figure 11 show a RSR contour plot in a depth of 2
km. Values range from 0.65 to 1.3 within The
Geysers. This means, that there is a transtensional
regime with some normal faulting and strike slip
components. This result, we compare to focal
mechanism solutions of earthquakes occurred in The
Geysers (Figure 12). These data show normal
faulting and strike slip regimes. The modeling results,
which also show regime between normal faulting and
strike slip, fit to these observations.
Figure 13 shows a contour plot of the orientation of
the maximum horizontal stress SH in 2 km depth. The
SH orientation varies in the model results between 8
and 18 degrees within The Geysers. These values are
in average about 10 degrees smaller than the
observations (Figure 12), but are within the standard
deviation of mean SH orientation with 27° ± 20°.
To summarize, after applying tectonic boundary
conditions to the geomechanical-numerical model,
the results can explain observed data of tectonic
regimes and SH orientations. In the next step, fluid
0 0.5 1 1.5 2 2.5 3
k-ratio (mean horizontal stress / vertical stress)
14
12
10
8
6
4
2
0
De
pth
in k
m
E = 20 GPa
E = 40 GPa
E = 60 GPa
Model, Location 1
Model, Location 2
Greywacke
Felsite
Upper Crust
injection and production is applied to the model and
stress changes due to these processes are modeled.
Figure 11: RSR (Regime stress ratio) contour plot in
2 km depth. Values within The Geysers
range between 0.65 and 1.3 indicating
transtension to strike slip regime. Note
that The Geysers is in the center of the
model (see Figure 4).
Figure 12: Stress map of The Geysers area. Lines
show the orientation of the maximum
horizontal stress SH derived from
earthquake focal mechanisms solutions
following the World Stress Map quality
ranking scheme (Heidbach et al., 2010);
color indicate the tectonic regime with
NF=normal faulting (red), green strike-
slip (green) and TF=Thrust Faulting
(blue). Rose diagram shows the mean SH
orientation with 27° ± 20°.
Figure 13: Orientation of the maximum horizontal
stress SH in 2 km depth.
Fluid Injection And Production Induced Stress
Changes
Here, we show first results of modeling the pore
pressure and stress change due to fluid injection and
production at The Geysers. For this purpose, we
modeled a scenario of fluid injection and production
starting in end of 1997 with the start of SEGEP,
added in 2003 the injection due to SRGRP and ended
in 2013 with our modeling. Fluid injection due to
SEGEP is represented by fluid injection at 10
injection points, due to SRGRP at further 13 injection
points and due to production at 42 production points.
Injected and produced fluid amounts correspond to
real injected and produced amounts. We have
averaged the injected and produced amounts over the
entire model run time and over all injection and
production points. This means, 28 l/s at each SEGEP
injection point, 30 l/s at each SRGRP injection point
and 48 l/s at each production point.
Figure 14 shows the pore pressure change within The
Geysers. The reservoir pressure is lowered by
approximately 20 to 30 MPa, with higher values in
the vicinity of the production points and with lower
values towards the reservoir boundary. There are
local increases of pore pressure in the vicinity of the
injection points. The overall picture is a decrease of
pore pressure in the reservoir. A decrease in pore
pressure increases the stress components. Figure 15
shows the change in maximum principal stress in the
reservoir. Fluid production and pore pressure
decrease increases the maximum principal stress by
10 to 30 MPa with higher values in the vicinity of the
production points.
Figure 14: Pore pressure change inside The Geysers
reservoir due to injection and production
of fluid.
Figure 15: Change in maximum principal stress
inside The Geysers reservoir due to
injection and production of fluid.
DISCUSSION
As no stress magnitude data are available from The
Geysers, we used the reference state of Sheorey
(1994) that provides k-ratios as well as information
on the SH orientation and the tectonic regime to
calibrate the initial stress state of the model.
The data from the KTB-site can be well explained by
the Sheorey-concept (Figure 8), but not the data from
the SAFOD pilot hole. In principle the Sheorey
reference stress state is only valid for regions with no
active tectonics and no large lateral density
variations. This holds on for the KTB-site, but not for
the SAFOD pilot hole.
Also The Geysers is in an active tectonic region.
Calculating the k-ratio without applying tectonic
boundary conditions, k-ratios and thus RSR values
are too low compared to measured data. Figure 16
shows the RSR in a depth of 2 km before applying
tectonic boundary conditions. The values within The
Geysers range between 0 and 0.3; thus the regime is
highly extensional, and the measured tectonic
regimes could not explained by these results. The
tectonic loading increases the horizontal stresses and
thus the k-ratio and the RSR. Compare the results of
Figure 16 to those of Figure 11, where RSR is plotted
after tectonic loading is applied.
Figure 16: RSR contour plot in 2 km depth, only due
to gravity field (no tectonic loading).
Not only RSR values change by adding tectonic
loading to the modeling process. There is also a
change in the SH orientation. Before applying tectonic
loading to the model, the SH orientation covers the
entire range from 0 to 180 degrees (Figure 17).
The more or less stable measured orientation (Figure
12) cannot be explained by these results. Therefore, it
is essential to add tectonic loading to the modeling
process.
Injection and production of fluid have a clear impact
on the pore pressure field and on the state of stress.
However, the results showing changes in the order of
a few tens of MPa are overestimated; these are
preliminary results that depend significantly on the
assumed permeability of the rock. There are
estimates that the reservoir pressure dropped by a few
MPa (Mossop and Segall, 1997).
Figure 17: Orientations of the maximum horizontal
stress in 2 km depth, only due to gravity
field (no tectonic loading).
CONCLUSION
In order to quantify the relative importance of the
man-made and natural processes that change the
stress state in The Geysers geothermal field a
regional model is needed. Furthermore, we showed
that the implementation of an appropriate initial
stress field of The Geysers is of key importance.
Tectonic regimes and orientations of the maximum
horizontal stress, both derived from focal mechanism
solutions, can be explained by the model results with
the initial stress field and tectonic loading due to the
locked San Andreas fault system.
First results, preliminary results, of long-term
injection and long-term production of fluid show a
decreasing pore pressure and increasing effective
stresses. However, the absolute values seem to be
overestimated given the pore pressure observations
that indicate a smaller stress drop.
For the future, we have developed a tool to make the
permeability of the reservoir dependent on the pore
pressure. When pore pressure exceeds a certain pore
pressure value, permeability increases. This will
increase the fluid flow and lower the pore pressure
changes. Additionally, we intend to investigate stress
changes due to co-seismic slip and in particular the
stress loading history at the reservoir bounding faults.
ACKNOWLEDGEMENT
The work presented was mainly funded by the
Department of Energy Geothermal Technologies
Program under Award Number DE-EE0002756-002
and co-funded by the European Commission within
the FP7 project GEISER, grant agreement no. 24132.
REFERENCES
Allis, R.G. (1982), "Mechanism of induced
seismicity at The Geysers geothermal reservoir,
California", Geophysical Research Letters, 9(6),
629-632.
Allis, R. and Shook, G.M. (1999), "An alternative
mechanism for the formation of The Geysers
vapor-dominated reservoir", Proceedings of the
24th Workshop on Geothermal Reservoir
Engineering, Stanford, SGP-TR-162.
Altmann, J., Müller, T., Müller, B., Tingay, M. and
Heidbach, O. (2010), "Poroelastic Contribution
to the Reservoir Stress Path," International
Journal of Rock Mechanics and Mining Sciences,
47(7), 1104-1113.
Baisch, S., Voros, R., Rothert, E., Stang, H., Jung, R.
and Schellschmidt, R. (2010), "A Numerical
Model for Fluid Injection Induced Seismicity at
Soultz-sous-Forêts," International Journal of
Rock Mechanics and Mining Sciences, 47(3),
405-413.
Boyle, K., Jarpe, S., Hutchings, L., Saltiel, S.,
Peterson, J. and Majer, E. (2011), "Preliminary
investigation of an aseismic 'doughnut hole'
region in the northwest Geysers, California",
Proceedings of the 36th Workshop on
Geothermal Reservoir Engineering, Stanford,
SGP-TR-191.
Brikowski, T.H. (2001), "Deep fluid circulation and
isotopic alteration in The Geysers system: profile
models", Geothermics, 30, 333-347.
Brudy, M., Zoback, M.D., Fuchs, F., Rummel, F. and
Baumgärtner, J. (1997), "Estimation of the
complete stress tensor to 8 km depth in the KTB
scientific drill holes: Implications for crustal
strength", Journal of Geophysical Research, 102,
18453-18457.
Bruel, D. (2007), "Using the migration of induced
seismicity as a constraint for fractured hot dry
rock reservoir modeling," International Journal
of Rock Mechanics and Mining Sciences, 44,
1106-1117.
Bufe, C.G., Marks, S.M., Lester, F.W, Ludwin, R.S.
and Stickney, M.C. (1981), "Seismicity of The
Geysers-Clear lake region", U.S. Geological
Survey Professional Paper, 1141, 129-137.
Byerlee, J. (1978), "Friction of rocks", Pure and
Applied Geophysics, 116, 615-626.
Crowell, J.C. (1974), "Origin of late Cenozoic basins
in southern California", Soc. Econ.
Paleontologists Mineralogists Spec. Pub., 22,
190-204.
Denlinger, R.P. and Bufe, C.G. (1982), "Reservoir
conditions related to induced seismicity at The
Geysers steam reservoir, northern California",
Bulletin of the Seismological Society of America,
72(4), 1317-1327.
Donnelly-Nolan, J.M., Burns, M.G., Goff, F.E.,
Peters, E.K. and Thompson, J.M. (1993) "The
Geysers-Clear lake area, California: Thermal
waters, mineralization, volcanism, and
geothermal potential", Economic Geology, 88,
310-316.
Eberhart-Philips, D. (1988), "Seismicity in the Clear
Lake area, California", Geological Society of
America Special Paper, 214, 195-206.
Eberhart-Philips, D. and Oppenheimer, D.H. (1984),
"Induced seismicity in The Geysers geothermal
area, California", Journal of Geophysical
Research, 89, B2, 1191-1207.
Evans, K.F., Zappone, A., Kraft, T., Deichmann, N.
and Moia, F. (2012), "A survey of the induced
seismic response to fluid injection in geothermal
and CO2 reservoirs in Europe", Geothermics, 41,
doi:10.1016/j.geothermics.2011.1008.1002,
1030-1054.
Fakcharoenphol, P., Hu, L. and Wu, Y.-S. (2012),
"Fully-implicit flow and geomechanics model:
Application for enhanced geothermal reservoir
stimulations", Proceedings of 37th Workshop on
Geothermal Reservoir Engineering, Stanford,
SGP-TR-194.
Goyal, K.P. and Box Jr., W.T. (2004), "Geysers
performance update through 2002", Proceedings
of the 29th Workshop on Geothermal Reservoir
Engineering, Stanford, SGP-TR-175.
Goyal, K.P. and Conant, T.T. (2010), "Performance
history of The Geysers steam field, California,
USA", Geothermics, 39, 321-328.
Ghassemi, A. and Zhou, X. (2011), "A three-
dimensional thermo-poroelastic model for
fracture response to injection/extraction in
enhanced geothermal systems", Geothermics, 40,
39-49.
Hartline, C. (2012), "The Geysers geothermal field,
Fieldwide seismicity analysis, Northwest
Geysers enhanced geothermal system
demonstration project", Energy Institute Fall
Seminar Series.
Hearn Jr., B.C., McLaughlin, R.J. and Donnelly-
Nolan, J.M. (1988), "Tectonic framework of the
Clear lake basin, California", Geological Society
of America Special Paper, 214, 9-20:
Heidbach, O., M. Tingay, A. Barth, J. Reinecker, D.
Kurfeß und B. Müller (2010), "Global crustal
stress pattern based on the World Stress Map
database release 2008", Tectonophysics, 462(1-
4), doi:10.1016/j.tecto.2009.1007.1023.
Hergert, T. und O. Heidbach (2011), "Geomechanical
model of the Marmara Sea region - II. 3-D
contemporary background stress field",
Geophysical Journal International,
doi:10.1111/j.1365-1246X.2011.04992.x,
01090-01102.
Hickman, S. and Zoback, M.D. (2004), "Stress
orientations and magnitudes in the SAFOD pilot
hole", Geophysical Research Letters, 31L15S12,
doi:10.1029/2004GL020043.
Kohl, T. and Megel, T. (2007), "Predictive modeling
of reservoir response to hydraulic stimulations at
the European EGS site Soultz-sous-Forets,"
International Journal of Rock Mechanics and
Mining Sciences, 44(8), 1118-1131.
Lipman, S.C., Strobel, C.J. and Gulati, M.S. (1978),
"Reservoir performance of The Geysers field",
Geothermics, 7, 209-219.
McClure, M. W. and Horne, H. N. (2011),
"Investigation of injection-induced seismicity
using a coupled fluid flow and rate and state
friction model," Geophysics, 76(6), WC183-
WC200.
McLaughlin, R.J. (1981), "Tectonic setting of pre-
Tertiary rocks and its relation to geothermal
resources in the Geysers-Clear Lake area", U.S.
Geological Survey Professional Paper, 1141, 3-
23.
Majer, E., J. Nelson, A. Robertson-Tait, J. Savy, and
I. Wong (2012), Protocoll for Addressing
Induced Seismicity Associated with Enhanced
Geothermal Systems Report, 46 pp, U.S.
Department of Energy.
Majer, E. L., and J. E. Peterson (2007), "The impact
of injection on seismicity at The Geysers,
California geothermal field", International
Journal of Rock Mechanics and Mining Sciences,
44, 1079-1090.
Majer, E. L., R. Baria, M. Stark, S. Oates, J.
Bommer, B. Smith, and H. Asanuma (2007),
Induced seismicity associated with Enhanced
Geothermal Systems, Geothermics, 36, 185-222.
Mossop, A. and Segall, P. (1997), "Subsidence at The
Geysers geothermal field, N. California from a
comparison of GPS and leveling surveys",
Geophysical Research Letters, 24(14), 1839-
1842.
Oppenheimer, D.H. (1986), "Extensional tectonics at
The Geysers geothermal area, California",
Journal of Geophysical Research, 91, B11,
11463-11476.
Rutqvist, J., Birkholzer, J., Cappa, F. and Tsang, C.-
F. (2007), "Estimating maximum sustainable
injection pressure during geological
sequestration of co2 using coupled fluid flow
and geomechanical fault-slip analysis," Energy
Conversion and Management, 48, 1798-1807.
Rutqvist, J., J. Birkenholzer, F. Cappa, C. Oldenburg,
and C.-F. Tsang (2006), "Shear-slip analysis in
multiphase fluid-flow reservoir engineering
application using Tough-Flac, Proceedings of
TOUGH Symposium, LBNL, Berkeley,
California.
Rutqvist, J. and Oldenburg, C. (2008), "Analysis of
cause and mechanism for injection-induced
seismicity at The Geysers geothermal field,
California", Proceedings of the 42nd U.S. Rock
Mechanics Symposium, San Francisco, ARMA,
151.
Sanyal, S.K. and Enedy, S.L. (2011), "Fifty years of
power generation at The Geysers geothermal
field, California - the lessons learned",
Proceedings of the 36th Workshop on
Geothermal Reservoir Engineering, Stanford,
SGP-TR-191.
Segall, P. and Fitzgerald, S.D. (1998), "A note on
induced stress changes in hydrocarbon and
geothermal reservoirs", Tectonophysics, 289,
117-128.
Sheorey, P.R. (1994), "A theory for in situ stresses in
isotropic and transversely isotropic rock",
International Journal of Rock Mechanics and
Mining Sciences, 3(1), 23-34.
Simpson, R. W. (1997), "Quantifying Anderson's
fault types", Journal of Geophysical Research,
102, 17909-17919.
Stark, M.A., Box Jr., W.T., Beall, J.J., Goyal, K.P.
and Pingol, A.S. (2005), "The Santa Rosa -
Geysers Recharge Project, Geysers geothermal
field, California, USA", Proceedings of the
World Geothermal Congress, Antalya, Turkey.
Stimac, J., Goff, F. and Hearn, B.C. (1992),
"Petrologic considerations for hot dry rock
geothermal site selection in the Clear Lake
region, California", Geothermal Resources
Council Meeting.
Suckale, J. (2009), “Induced Seismicity in
Hydrocarbon Fields”, Advances in Geophysics,
51(2), 55-106, doi: 110.1016/S0065-
2687(1009)05107-05103.
Suckale, J. (2010), “Moderate-to-large seismicity
induced hydrocarbon production”, The Leading
Edge, 310-317
Tester, J. W., Anderson, B. J., Batchelor, A. S.,
Blackwell, D. D., DiPippo, R., Drake, E. M.,
Garnish, J. D., Livesay, B., Moore, M. C.,
Nochols, K., Petty, S., Toksoez, M. N. and
Veatch, R. W. J. (2006), "The Future of
Geothermal Energy: Impact of Enhanced
Geothermal Systems (EGS) on the United States
in the 21st Century," Massachusetts Institute of
Technology, Idaho Falls.