CONFERENCE PROCEEDINGS

Analysis of In-Situ Stress Regime in the

Alberta Basin, Canada, for Performance
Assessment of CO2 Geological
Sequestration Sites

Christopher D. Hawkes, University of Saskatchewan,

Stefan Bachu and Kristine Haug, Alberta Energy and Utilities Board, and
Adam W. Thompson, University of Saskatchewan, Canada

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
 CONFERENCE PROCEEDINGS

ABSTRACT

Oil and gas reservoirs and deep saline aquifers are primary candidates for long-term geological
sequestration of greenhouse and acid gases. Risk assessment for sequestration projects must include
predictions of sequestration zone performance. These performance assessments will guide the selection of
sequestration sites and/or operating parameters, such as injection pressure and rate, that mitigate
leakage risks. If natural fractures or faults are present, then bottom-hole injection pressures higher than
the minimum in-situ stress may open these fractures. Pressures higher than the fracture breakdown
pressure will fracture the reservoir and/or caprock. In both cases, CO2 or acid gas will leak from the
sequestration unit. Thus, it is essential to properly estimate the minimum stress and fracture breakdown
pressure and devise injection strategies that will maintain pressures below these at all times.

An extensive database of micro- and mini-fracture, leak-off and fracture breakdown pressures measured
by industry in the Alberta Basin, Canada, and data from close to fifty acid-gas injection operations in the
basin has been used to develop a methodology for estimating the minimum in-situ stress both at regional
and local scales throughout the basin on the basis of these data. Minimum horizontal stress gradients
close to 17 kPa/m have been estimated for much of the basin from leak-off tests conducted over depths
from ~250 to ~1000 m. Gradients close to 13 kPa/m have been estimated for depths ranging from ~1000
to ~2500 m from fracture breakdown pressures that are believed to be affected by reservoir production
and pressure depletion. Approved maximum bottom-hole injection pressures at the acid-gas injection
operations in western Canada are safely below the minimum horizontal stress, thus avoiding re-opening
pre-existing fractures, if any exist. Since minimum horizontal stresses are lower by 12.3% on average
than rock fracturing pressures, new fractures will not be induced as a result of injection. Using similar
operating requirements for CO2 injection will ensure the mechanical safety of CO2 geological
sequestration operations.

1. INTRODUCTION

Implementation of technologies to capture carbon dioxide (CO2) and sequester it in geological formations
will be necessary in order to achieve significant reductions in atmospheric emissions of anthropogenic
greenhouse gases. Oil and gas reservoirs and deep saline aquifers (collectively referred to as sequestration
reservoirs in this paper) are primary candidates for long-term geological sequestration. By ratifying the
Kyoto protocol, Canada has committed to reduce by 2012 its greenhouse gas emissions to 6% below the
1990 level, although currently emissions are approximately 20% above that level. Carbon dioxide
emissions in the province of Alberta represent approximately 30% of Canada’s emissions. Unlike in other
provinces, two thirds of these emissions are from large stationary sources, such as coal-fired power plants,
refineries and oil sands plants, that are conducive to CO2 capture and sequestration in the Alberta Basin
that underlies the province (Bachu and Stewart, 2002).

If the CO2 injected into sequestration reservoirs leaks to the atmosphere in significant quantities, then the
operation will have failed to mitigate the global effect of global warming, its very purpose of being
implemented. Locally, leaked CO2 may contaminate other energy and mineral resources in the
sedimentary succession, and potable groundwater resources if it reaches shallow zones. In extreme cases,
where high-rate leakage occurs through a localized conduit to the surface, CO2 may endanger life. Thus, it
is essential to assess the potential for CO2 leakage and to implement site selection criteria and operating
parameters that would mitigate the possibility of CO2 escaping from the sequestration reservoir. Risk
analysis is an integral component of this assessment. It involves an evaluation of the types of events that
may result in leakage, the likelihood of these events, and their potential consequences. Risk analysis for

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
2
 CONFERENCE PROCEEDINGS

greenhouse gas sequestration is currently in its early stages, with most of the emphasis being placed on
performance assessment (i.e., identifying events and their likelihoods), and less on consequences (e.g.,
Chalaturnyk et al., 2004). Similarly, this paper deals with selected, important aspects of performance
assessment, more specifically, the identification of possible leakage events that are controlled by the
geomechanical integrity of the low-permeability caprock that seals the sequestration reservoir, and
methods for mitigating the likelihood of these events. The potential for leakage through wellbores due to
factors controlled by geomechanical parameters is also discussed. Other, important aspects of leakage
through wellbores (e.g., well abandonment procedures, chemical interaction between CO2 and well
casings and cements) are not considered here.

The geomechanical integrity of the sequestration reservoir is controlled by the stress regime at the site and
by operating parameters such as bottom-hole injection pressure. Stresses are tensorial in nature and are
characterized by the three principal components, σ1, σ2 and σ3, which are orthogonal, and by their
orientations. By definition, one of the principal stresses must intercept free surfaces, such as the ground
surface, at right angles. Thus, in the case of divergent sedimentary basins (e.g., foreland, intra-cratonic,
platform-margin) with gentle surface relief, a basic assumption is that the ground surface is semi-planar
and sub-horizontal; hence the principal stress directions are approximately vertical and horizontal. In this
case, the principal stresses are denoted by σV for the vertical stress, and σHmin and σHmax for the smaller
and larger horizontal stresses, respectively. This assumption is not valid in the case of convergent
sedimentary basins (e.g., in subduction areas and intramontane belts), in the vicinity of major fault zones
that deflect stresses, around salt domes or in overthrust and deformed zones. The assumption of vertical
and horizontal principal stresses applies to the Alberta Basin in Canada east of the Rocky Mountain
Deformation Front. In most sedimentary basins σHmin is less than σV, and indeed previously published
stress magnitude measurements in the Alberta Basin indicate that the smallest principal stress, σ3, is
horizontal (i.e., σ3 = σHmin) except possibly in parts of the foothills and in the shallow part of the basin in
northeastern Alberta near the edge of the basin at the Canadian Precambrian Shield (Bell and Babcock,
1986; Bell et al., 1994; Bell and Bachu, 2003). This means that natural and induced fractures will be
vertical and oriented in a direction perpendicular to σHmin (i.e., along the trajectory of σHmax). Only at
shallow depth, less than 300 m, σV may be less than σHmin, in which case fractures will be horizontal.

2. LEAKAGE MECHANISMS CONTROLLED BY IN-SITU STRESSES

Following is a brief summary of leakage mechanisms that are relevant to this paper, illustrated in Figure 1.
For a more complete discussion, the reader is referred to Hawkes et al., (2004). Presumably, the initial
selection of a sequestration reservoir would include an assessment of the presence of faults or natural
fractures. Ideally, sequestration sites without any of these features would be selected. If faults or fractures
are present, however, the initial investigation would also need to demonstrate that they have low
permeability and/or they are closed. For example, a considerable amount of effort has been devoted to the
development of procedures for assessing fault seal capacity in hydrocarbon reservoirs (Yielding, 2002).

2.1 Poor Cement Emplacement in Enlarged Boreholes

During drilling operations, shear yielding and detachment of yielded rock from the borehole wall may
result in severely enlarged and rugose boreholes. The extent of this enlargement is a direct function of in-
situ stress magnitudes, rock compressive strength and drilling mud density (e.g., Hawkes and McLellan,
1999). Caprocks, which most often are shales, tend to be relatively weak and prone to yielding and
enlargement. Casing-cement-formation bond quality is adversely affected by enlarged hole conditions. As
such, determination of in-situ stress magnitudes can be used to design better drilling programs to avoid

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
3
 CONFERENCE PROCEEDINGS

Poor cement/
enlarged hole
Natural fractures

Hydraulic
fracture

CO2
Fault

Figure 1: Possible CO2 leakage pathways affected by in-situ stresses.

these problems in new wells, or to back-analyze yielding around existing wells to identify those with the
greatest risk of enlargement.

2.2 Re-Activation (Slip) and Re-Opening of Faults and Natural Fractures

The permeability of faults or natural fractures that initially act as seals may increase if the shear stresses
acting on these weak interfaces reach critical levels. As described in Hawkes et al. (2004), an increase in
pore pressure within a fault or fracture plane will increase the tendency to slip. In cases where elevated
pore pressures do not migrate into the weak interfaces, but rather induce a global stress change by
uniformly pervading the entire sequestration reservoir, the slip tendency of faults or fractures can either
increase or decrease. The factors controlling this include in-situ stress magnitudes and orientations, and
the orientation and friction angle of these weak interfaces. The most critical orientation for slip of faults
or fractures is in a plane that is 45° from the maximum in-situ stress (σ1), dipping in the direction of the
minimum in-situ principal stress (σ3).

Even if fractures or faults have low permeability, the stress changes induced in the reservoir during
injection may re-open them. For example, if elevated, near-well pore pressures migrate even part way into
a fault or fracture surface, the weak interface will be forced open once the pore pressure exceeds the in-
situ stress acting normal to its surface. Once this process has initiated, and the weak interface is propped
open at its mouth, continued injection will enable the propagation of the re-opening process deeper into
the caprock.

The most critical orientation for re-opening of fractures and faults is in a plane normal to the minimum in-
situ stress (σ3) component. These weak interfaces can re-open once the injection pressure exceeds σ3.
Similarly, the least critical orientation of fractures and faults is in a plane normal to the maximum in-situ

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
4
 CONFERENCE PROCEEDINGS

stress (σ1). These weak interfaces will re-open only once the injection pressure exceeds σ1. All other
fracture orientations will re-open at injection pressures between σ3 and σ1.

2.3 Hydraulic Fracturing

During fluid injection, high bottom-hole pressures may induce a tensile stress state around the wellbore. If
the induced tensile stress magnitude exceeds the rock’s tensile strength, then a tensile fracture will
develop. Assuming that the injection fluid is clear (i.e., contains no particulate matter that might plug the
induced fracture), this tensile fracture will propagate. If it propagates upwards out of the injection zone, it
can create a leakage pathway through the caprock.

For an intact, linear elastic rock and a non-penetrating fluid (i.e., within the time-frame of the fluid
injection period, no pore pressure increase occurs in the rock matrix near the borehole wall), the
theoretical value for the fracture breakdown pressure pb is (Hubbert and Willis, 1957):

p b = 3σ H min − σ H max − α p res + TS (1)

where: σHmin = minimum horizontal in-situ stress

σHmax = maximum horizontal in-situ stress
pres = reservoir pore pressure
K bulk
α = Biot’s coefficient, equal to 1 −
K grain
Kbulk = static bulk modulus of porous rock
Kgrain = static bulk modulus of constituent mineral grains
TS = rock tensile strength

Although based on a number of idealized, simplifying assumptions, equation (1) shows that fracture
breakdown pressures should be greater than the minimum horizontal in-situ stress magnitude.

Factors that can cause fracture breakdown pressures to deviate from the value predicted by equation (1)
include near-well pore pressure increase during injection, non-linear rock properties, thermally-induced
stresses, and the presence of natural fractures in the injection zone. In the case of the latter, so-called
fracture breakdown may actually occur as the re-opening of these existing fractures, as discussed in the
preceding section.

Figure 2 summarizes the relationship of injection pressures to the leakage mechanisms described above. If
natural fractures or faults are present, then bottom-hole injection pressures higher than the minimum in-
situ stress may open these fractures. Pressures higher than the fracture breakdown pressure will fracture
the reservoir and/or caprock. In both instances, the injected fluid, in this case CO2 or acid gas, will leak
from the sequestration reservoir. Thus, it is essential to properly estimate the minimum stress and fracture
breakdown pressure and devise injection strategies that will maintain pressures below these at all times.
Further, in settings where the in-situ stress state is highly anisotropic and critically-oriented faults or
fractures are present, there is a need to determine the stress magnitudes and orientations in order to assess
the threshold pressure at which a leakage event due to shear failure might occur.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
5
 CONFERENCE PROCEEDINGS

Fault Re-opening of Hydraulic

Reactivation natural fracturing
Bottomhole Injection Pressure fractures Alberta Basin

σ1 = σV or σHmax

σ3 = σHmin

Only in tectonically
active settings

pres

Figure 2: Important injection pressure thresholds that must be identified. The regions with the
darkest shading denote pressures where the likelihood of inducing a leakage event is greatest.

3. IN-SITU STRESS DATA SOURCES AND TRADITIONAL INTERPRETATION

METHODOLOGIES

3.1 Vertical Stress Magnitude

The vertical stress (σV) at any point in a basin is equivalent to the weight of the overburden. Integration of
a bulk density log taken after drilling will provide the vertical stress at depth (D), according to:

D
σ V = ∫ ρ b dz (2)
0

where ρb is the bulk density of the fluid-saturated rock. Density logs that are used for estimating σV should
be as complete as possible and record little hole enlargement. In practically all wells there is an upper
unlogged interval, and an average density for rock and near-surface deposits has to be assumed.

3.2 Minimum Horizontal Stress Magnitude from Fracture Closure Pressures

The minimum horizontal stress (σHmin) can be estimated through various means. One way is using micro-
and mini-fracture tests, extended leak-off tests and massive hydraulic fracture records to interpret the
fracture closure pressure, which should correspond to the minimum stress magnitude (Figure 3). The
accuracy of the estimates decreases from micro-fracture tests to massive hydraulic fractures. Micro-
fracture tests involve initiating a hydraulic fracture within a short packed-off interval by slowly injecting

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
6
 CONFERENCE PROCEEDINGS

pb
pi pi = fracture initiation
Bottomhole pressure, BHP; (leakoff) pressure
Injection rate, Q FPP pb = fracture breakdown
ISIP pressure ( ≥ pi)

BHP
FPP = fracture propagation
σ3 pressure
FCP ISIP = instantaneous shut-
in pressure
FCP = fracture closure
Q pressure

Time
Figure 3: Typical pressure-time record for a micro- or mini-fracture test. Similar pressure-time
records may be obtained from extended leak-off tests, although these test types are rare (none are
available for the Alberta Basin). Fracture initiation pressure and fracture breakdown pressure may
occur at the same point, especially in the case of a clear injection fluid (which has no particulate
matter to plug the fracture once it initiates).

a small volume (~1 m3) of a low-viscosity fluid (e.g., water), and opening and closing the fracture several
times until a consistent closure pressure is obtained. If the fracture closure pressure is less than the
overburden load (σV), then this pressure is equated with the smallest principal stress, i.e. σHmin (Gronseth
and Kry, 1983). Mini-fracture tests are similar to micro-fracture tests and typically involve a relatively
high-rate injection of viscous fluids in excess of 10 m3 (e.g., McLellan, 1988). The magnitude of σHmin
can be equated with fracture closure pressures interpreted from pressure versus time records, but these
tests measure the average in-situ stress over a larger rock volume than the micro-fracture tests. Previous
work in western Canada has shown that closure pressures from micro- and minifracture tests are
practically indistinguishable (Woodland and Bell, 1989).

Table 1 lists the data that were used for this investigation. These data were extracted from a larger dataset,
from which pressures that obviously had no relationship to stress magnitudes (e.g., unreasonably high or
low) were filtered out. Of note is the fact that no fracture closure pressures were available for the 757
leak-off tests (because these were not extended leak-off tests), nor for the 583 massive hydraulic fracture
stimulation treatments. The novel methods used to estimate σHmin magnitudes from the leak-off and
fracture breakdown pressures recorded for these tests will be discussed subsequently.

Table 1: Data available for minimum horizontal stress analysis in the Alberta Basin.
Parameter / Test Type Number
Fracture closure pressure / Micro-fracture 15
Fracture closure pressure / Mini-fracture 91
Leak-off pressure / Leak-off test 757
Fracture breakdown pressure / Hydraulic fracture 583
Borehole breakout orientation / Oriented caliper log 248

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
7
 CONFERENCE PROCEEDINGS

3.3 Minimum Horizontal Stress Orientation

Stress orientations are important because they indicate the most likely direction of propagation for
induced fractures, and they may provide clues as to the orientation of natural fractures. The most
commonly used information for estimating stress orientations is derived from borehole breakouts (Bell,
2003). These breakouts are intervals in a well where caving has occurred on opposite sides of a borehole,
so that it is laterally elongated, and are diagnostic of anisotropic compression around the borehole (i.e.,
σHmin ≠σHmax). In quasi-vertical wells (<5° from vertical) through transversely isotropic rocks, breakout
caving elongates the wellbore parallel to σHmin (Bell, 2003; Zoback et al., 2003). Breakouts are best
displayed on borehole imaging logs, but logging tools, such as dipmeters, are also suitable for
documenting breakouts (Bell, 2003).

4. MINIMUM HORIZONTAL STRESS ESTIMATION FROM LEAK-OFF AND FRACTURE

BREAKDOWN PRESSURES IN THE ALBERTA BASIN

Conventional leak-off tests are terminated once leak-off occurs. In these tests, leak-off is defined as the
first point of deviation from a straight line on a plot of injection pressure versus injected volume. As such,
it is not possible to observe fracture propagation, shut-in response and fracture closure. Figure 4 illustrates
a typical leak-off pressure record, and lists the various mechanisms that may result in leak-off. Of greatest
interest is the fact that leak-off will occur at pressures very close to σHmin for cases where natural fractures
oriented sub-normal to the σHmin direction are present, or perhaps if fractures induced during a previous
leak-off test are present (Figure 5). Consequently, if a large number of leak-off pressures have been
recorded in a given region, a “lower bound” to these data should provide a reasonable estimate of σHmin.
The term “lower bound” is used loosely in this case, as there may be a few data points that fall at lower
pressures, either due to a leak in the casing or cement, or into large pores or open natural fractures that
cannot be plugged by the drilling mud solids. In general, it should be possible to differentiate these data
points from the stress-related leak-off pressures, and draw a “lower bound” envelope that ignores them.

The underlying principle of the lower bound method may be visualized with the help of Figure 6.
Successful application of the lower bound method has been demonstrated in the central North Sea by
Edwards et al. (1998) and on the North West Shelf of Australia by Addis et al. (1998). To the authors’
knowledge, this technique has not been previously applied in the Alberta Basin.

4.1 Leak-off Pressure Analysis

In order to apply the lower bound method to the leak-off pressures (LOP) available for the Alberta Basin,
it was first necessary to sub-divide the dataset geographically. Figure 7 shows the regional break-up that
was ultimately used. For regions 2 through 4, the boundaries were drawn so as to encompass localized
concentrations or clusters of data. Regions 1, 5, 6 and 7 were drawn in regions with lower data density,
but in such a way as to group data points obtained from rocks of similar geologic age. Region 8 was
drawn to encompass all remaining data points that are scattered throughout the eastern portion of the
basin, but, because of its large areal spread, it is expected that the results would be less conclusive and
representative.

Review of the leak-off pressures indicated that the majority of these measurements were made at
relatively shallow depths. The greatest concentration occurred from 250 to 500 m, with most of the
remaining data falling between 500 and 1000 m. As such, greatest confidence in the results obtained from
leak-off tests is obtained only for these relatively shallow depths.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
8
 CONFERENCE PROCEEDINGS

LOP

Leak-off could indicate:

Bottomhole pressure - Initiation of a tensile fracture (LOP > σHmin)
- Re-opening of an existing natural fracture that is
not normal to σHmin (LOP > σHmin)
- Re-opening of an existing natural fracture that is
normal to σHmin OR a fracture that w as induced
in a previous injection period (LOP ~ σHmin)
- Leakage into casing-borehole annulus due to
bad cement job (LOP has no correlation to
σHmin , but most likely is low er)
- Lost circulation into large pores or natural
fractures that are not effectively plugged by
drilling mud solids (LOP ~ pore pressure)
pumping
Often, this is the end of injection

Volume of fluid pumped and/or Time

Figure 4: Pressure behavior for injection until just after leak-off occurs.

LOP

"Class 2" behaviour (Edw ards et al., 1998)

Bottomhole pressure

Pressure curve becomes relatively flat once

leak-off occurs.

LOP is close to (slightly greater than) σHmin

pumping

Volume of fluid pumped and/or Time

Figure 5: “Class 2” leak-off behavior observed if fracture(s) sub-normal to σHmin are already
present.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
9
 CONFERENCE PROCEEDINGS

Leak-off into
natural fractures
~normal to σHmin
Fracture breakdown
Premature leak-off pressure for intact,
(casing integrity, linear elastic rock at

Lower bound
Frequency

rock with vugular high injection rates

porosity or open
natural fractures)

Leak-off into other

natural fractures;
breakdown of intact,
non-linear elastic rock

pres σHmin pb LOP

Figure 6: Conceptual illustration of the distribution of leak-off pressures with respect to in-situ
stress magnitudes and reservoir pressure. A “Lower Bound” to the distribution of a large number
of leak-off pressures is a reasonable estimate of σHmin.

Figure 8 shows a plot of leak-off pressures versus depth for Region 3. This plot shows that a well-defined
lower bound envelope exists, corresponding to a gradient of 17 kPa/m. Similarly well-defined lower
bound envelopes, and similar stress gradients, were found for Regions 2 and 4. The quality of the lower
bound envelopes ranged from moderate to poor for the other regions. A summary of the results for all
regions, including comments on the confidence of the interpreted σHmin values, is given in Table 2.

4.2 Fracture Breakdown Pressure Analysis

The mechanics of leak-off during a leak-off test and fracture breakdown during a hydraulic fracture
treatment are essentially the same. In the case of an intact rock, leak-off and fracture breakdown pressures
correspond to the pressure required to initiate a tensile fracture. This assessment is not always recognized;
i.e., when leak-off tests are continued to the point of fracture breakdown, this breakdown pressure is often
regarded as the fracture initiation pressure. However, given that the solids present in drilling muds tend to
plug fractures when they initially form, fracture breakdown only occurs once a fracture has been initiated
and has been opened to an aperture too wide to be plugged by the mud solids. When natural fractures are
present, leak-off pressures and hydraulic “fracture breakdown” pressures actually correspond to the
pressure at which the existing fractures are forced open. Consequently, a lower bound analysis of fracture
breakdown pressures should also provide an estimate of the minimum stress.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
10
 CONFERENCE PROCEEDINGS

Figure 7: Sub-division of the Alberta Basin dataset into regions suitable for lower bound analysis of
σHmin.

Before proceeding to describe the analysis of fracture breakdown pressures, it is necessary to identify one
significant difference between leak-off and fracture breakdown tests. Leak-off tests are typically run
immediately below casing shoes, in non-reservoir rocks. As such, the pore pressures in these rocks are
unaltered from their initial, in-situ values. Fracture breakdown pressures, on the other hand, are measured
in reservoir formations that are about to be fracture stimulated. In many cases, these reservoirs have been
produced for extended periods of time before the stimulation treatment. As such, pore pressures have
often decreased significantly. This clarification is important because it has been demonstrated both
theoretically and using field data that pressure depletion tends to lower σHmin (e.g., Addis, 1997).

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
11
 CONFERENCE PROCEEDINGS

Leak-off Pressure (MPa)

0 5 10 15 20
250
Grad σHmin = 17 kPa/m

350
Depth (m)

450

550

650

Figure 8: Depth profile showing leak-off pressures measured in Tertiary and Upper Cretaceous-age
formations in Region 3 of the Alberta Basin. The lower bound to these data corresponds to a
gradient the minimum horizontal stress of 17 kPa/m.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
12
 CONFERENCE PROCEEDINGS

Stress, Pressure (MPa)

0 20 40 60 80 100
250

+ LOP
MF
pb
750

Grad σHmin = 13 kPa/m

(depleted)
1250
Depth (m)

1750

2250

2750

3250

Figure 9: Depth profile showing leak-off pressures (LOP’s), fracture breakdown pressures (pb) and
fracture closure pressures from mini- and micro-fracture tests (MF) in Region 3 of the Alberta
Basin. The different geologic ages included in the dataset shown include Tertiary, Upper
Cretaceous, Lower Cretaceous, Carboniferous through Jurassic, and Devonian. The lower bound to
the fracture breakdown pressures below 800 m corresponds to a gradient of 13 kPa/m. The fracture
closure pressures that fall on this lower bound envelope were likely measured in pressure-depleted
reservoirs, whereas the closure pressures above the envelope were likely measured in reservoirs
that have experienced little or no depletion.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
13
 CONFERENCE PROCEEDINGS

Table 2: Summary of SHmin gradients interpreted from lower bound analysis of leak-off and fracture breakdown pressures in the Alberta
Basin.
Region Leak-Off Pressures (LOP) Fracture Breakdown Pressures Comment
(pb)
Depth and Age SHmin gradient Depth and Age SHmin Gradient
300-600 m 18.0 kPa/m 600-2300 m 14.0 kPa/m Shallow: Sparse data, but well-defined envelope.
1
Tertiary – Upper U. Cretaceous - Deep: Well-defined envelope; sparse data below 1500 m
Cretaceous Devonian
300-700 m 17.2 kPa/m 700-3000 m 13.0 kPa/m Shallow: Well-defined envelope; sparse data below 450
2
Tertiary – Upper U. Cretaceous - m.
Cretaceous Devonian Deep: Well-defined envelope.
250-1000 m 17.0 kPa/m 1000-3000 m 12.9 kPa/m Shallow: Well-defined envelope; sparse data below 500
3
Tertiary – Upper U. Cretaceous – m.
Cretaceous Devonian Deep: Well-defined envelope.
300-700 m 17.0 kPa/m 700-3200 m 13.0 kPa/m Shallow: Well-defined envelope; sparse data below 500
4
Tertiary – U. Cretaceous - m.
Lower Devonian Deep: Well-defined envelope; sparse data below 2700
Cretaceous m.
500-1100 12.8 kPa/m* 1100-2400 12.0 increasing *
Shallow: Based on fracture breakdown pressures. Well-
5
Jurassic-Carb. Jurassic-Carb. to 19.0 kPa/m defined envelope.
Deep: Sparse data, but well-defined envelope.
300-800 15.0 kPa/m* 800-2300 m 14.0 kPa/m *
Shallow: Based on fracture breakdown pressures.
6
Lower L. Cretaceous – Sparse data. LOP’s affected by premature leak-off.
Cretaceous Carboniferous Deep: Sparse data.
600-1500 m 17.0 kPa/m* 1500-2400 m 13.0 kPa/m *
Shallow: Based on LOP’s and pb’s. Sparse data.
7
Devonian* Devonian Deep: Sparse data.
200-1100 m 17.0 kPa/m* 1100-2100 m 12.0 increasing *
Shallow: Sparse data. Based on LOP’s, pb’s and mini-
8
Cretaceous L. Cretaceous- to 16.5 kPa/m frac closure pressures.
Devonian L. Cretaceous - Deep: Sparse data. Based on pb’s and mini-frac closure
Carboniferous pressures.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
14
 CONFERENCE PROCEEDINGS

Based on the arguments presented above, a lower bound analysis was attempted on fracture breakdown
pressures measured in the Alberta Basin. Figure 9 shows for illustration the results obtained for Region 3.
Similar to the results obtained at shallower depths for leak-off pressures, a well-defined lower bound
envelope can be defined . In this case, the envelope corresponds to a gradient of 13 kPa/m. Given that this
gradient is significantly lower than the 17 kPa/m measured at shallower depths, it is felt that the 13 kPa/m
gradient is only valid for reservoirs that have been pressure-depleted. Similarly well-defined envelopes
for breakdown pressures were obtained for Regions 1, 2, 4 and 5. Results ranging from moderate to poor
were obtained for regions 6, 7 and 8.

Regarding Figure 9, it is worth noting that a common envelope seems to exist for rocks of various
lithologies and all ages in Region 3, and this has been observed in other regions of the basin as well. This
is significant because the analyzed sedimentary succession includes carbonate-dominated Devonian and
Carboniferous strata deposited during the platform-margin stage of evolution of the Alberta Basin, as well
as Jurassic to Tertiary-age siliciclastics deposited during the foreland stage. Apparently, any contrasts in
horizontal stresses due to the different mechanical properties of these rocks are too small to affect the
lower bound method.

As it currently stands, the regional lower bound envelopes derived from fracture breakdown pressures
would lead to upper limits on injection pressures that are conservative (low) for non-depleted CO2 storage
sites. Efforts to characterize the stress-depletion response of reservoirs in the Alberta Basin are currently
ongoing. With further analysis of each region, it may become possible to generate a set of lower bound
envelopes that are differentiated based on pore pressure, enabling more appropriate minimum horizontal
in-situ stress estimates at specific sites. Alternatively, as described in the following section on acid-gas
injection sites, it is possible to apply the lower bound methodology in a localized sense, to obtain
minimum horizontal in-situ stress estimates that are site-specific.

5. STRESS REGIMES FOR ACID-GAS INJECTION SITES IN THE ALBERTA BASIN

Acid-gas, a mixture of CO2 and H2S that is produced from sour gas reservoirs in western Canada, has
been injected into deep geological formations for close to 15 years with a good safety record. Injection
currently takes place at 44 sites into depleted oil and gas reservoirs, and deep saline aquifers (Figure 10).
Injection at 5 sites has been rescinded, and at 2 other sites it has been suspended by the regulatory agency
because the injection reservoir has been repressurized above the approved limit. One operation, although
approved, has never been implemented. In a few cases there are two and even three sites at the same
geographic location because acid gas is injected into different sequestration reservoirs. The average
injection depth varies between 824 and 3432 m. At roughly 60% of the sites injection takes or took place
in carbonate rocks, and the balance is or was injected into siliciclastics. In most cases shales and shaly
siliciclastics constitute the overlying caprock. The remainder of the injection zones are confined by tight
limestones, evaporites and anhydrites. Close to 5 Mt of acid-gas have been injected by the end of 2003 at
a cumulative annual rate of ~1 Mt/year. In many ways, the acid-gas injection operations in western
Canada constitute a commercial-scale analogue for CO2 geological sequestration, and a more detailed
history and description of these operations, including regulatory requirements, can be found in Bachu and
Gunter (2004, 2005) and Bachu and Haug (2005).

Of particular importance in the permitting process and the operation of acid gas injection sites is
avoidance of acid gas migration from the injection zone to other formations, shallow groundwater and/or
the surface. To avoid migration through fractures, the injection zone should be free of natural fractures
and the bottom hole injection pressure must be at all times below a certain threshold to ensure that

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
15
 CONFERENCE PROCEEDINGS

Figure 10: Location and status of acid-gas injection operations in the Alberta Basin at the end of
2004.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
16
 CONFERENCE PROCEEDINGS

fracturing is not induced. In the case of acid gas injection into deep saline aquifers, the maximum bottom
hole injection pressure is usually set at 90% of the rock “fracture pressure” (which is equivalent to the
fracture breakdown pressure, pb, used elsewhere in this paper). In the case of injection into depleted oil or
gas reservoirs, lately regulatory agencies are setting the maximum bottom hole injection pressure to the
initial reservoir pressure to avoid possible reservoir damage during production.

5.1 In-Situ Stress Regime

The methods described earlier in this paper were used to interpret the stress regime at each of the acid-gas
injection sites, although they were applied at a more localized scale around each injection site. For
example, to estimate σHmin at a given site, the leak-off or breakdown pressure with the lowest gradient was
selected from nearby wells. This essentially corresponds to the lower bound method described earlier.
Vertical stresses were obtained by interpolation between values calculated from the nearest wells for
which bulk density logs were available, including sometimes the injection well itself.

Minimum horizontal stresses, varying between 10.6 MPa at 705 m depth and 58.9 MPa at 3386 m depth,
increase with depth with a basin-wide average gradient of 16.6 kPa/m (R2=0.935; Figure 11), although
stress gradients vary locally between 13.6 and 19.5 kPa/m. Maximum vertical stresses vary between 15.2
MPa at 705 m depth and 79.1 MPa at 3386 m depth, with local vertical stress gradients that vary between
21.2 and 25.2 kPa/m. Vertical stresses increase with depth with a basin-wide gradient of 23.8 kPa/m
(Figure 11). The maximum horizontal stress (σHmax), which cannot be estimated with confidence form the
available data, has a value between these σHmin and σV.

Figure 11: Variation with depth of minimum horizontal stresses at acid-gas injection sites in
western Canada.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
17
 CONFERENCE PROCEEDINGS

Borehole breakouts in wells across the basin were used to map stress trajectories at the basin level, as
previously done for the entire Western Canada Sedimentary Basin and for the coal-bearing Cretaceous –
Tertiary strata in southern and central Alberta (Bell et al., 1994; Bell and Bachu, 2003). The direction of
the maximum horizontal stress (σHmax) in the Alberta Basin is generally normal to the Rocky Mountain
Deformation Front (hence the direction of σHmin is parallel to it). This suggests that a certain level of
tectonic stress, caused by past orogenic processes, is present in the basin. The directions of the minimum
and maximum horizontal stresses at the acid-gas injection sites were obtained using this dataset through
interpolation or extrapolation, and vary between 22oN and 60oN. The practical value of the interpreted
orientations is that, if tensile fractures occur, they will develop in the plane normal to σHmin; i.e., in the
direction of σHmax perpendicular to the Rocky Mountains.

5.2 Comparison of Injection and Fracture Pressures

As shown in Figure 12, minimum horizontal stresses are greater than the maximum bottom hole injection
pressure (BHIP) licensed by the regulatory agency in all but one case. Operators usually inject at lower
pressures than the maximum approved one, but occasionally may reach this value. Only at one site,
rescinded in June 1999, the σHmin value determined in this investigation is less than the maximum BHIP;
however, given the modest accuracy of these stress estimates, it is quite possible that σHmin is in reality
greater than the maximum BHIP. Furthermore, the average wellhead injection pressure at this site was
less by 1,000 kPa than the maximum approved wellhead injection pressure (WHIP). The data indicate that,
at these acid-gas injection sites, at no time was there any danger of opening existing fractures, if any are
present, as a result of injection.

Figure 12: Relation between the minimum horizontal stress (σHmin) and maximum bottomhole
injection pressure (BHIP) at acid-gas injection sites in western Canada.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
18
 CONFERENCE PROCEEDINGS

In the approval process, regulatory agencies in western Canada require operators to indicate the rock
fracturing pressure in the sequestration reservoir and its confining caprock, and operators have used a
variety of means to estimate it. In eleven cases the operator performed an injectivity test in the injection
well itself. In four other cases the operators estimated the fracturing pressure on the basis of tests run in
the corresponding sequestration reservoir in neighboring wells (mini-fracture tests, hydraulic and acid
fracture stimulation jobs). In 30 cases the fracturing pressure was estimated on the basis of various types
of calculated estimates for fracture gradients, and in seven cases no value or estimate was provided.

Figure 13 shows the variation with depth of the fracturing pressure as provided by operators in their
applications to the regulatory agencies, showing an increase with depth in the range from 14,000 to
~61,000 kPa, with an average gradient of 19 kPa/m. In all cases minimum horizontal stresses are less than
the fracturing pressure (Figure 13), on average by 12.3%, although they vary from as low as 58% to as
high as 98% of the fracture pressure. The wide variability in the ratio between these parameters is partly
due to the fact that most of the values were estimated rather than measured, and also due to the variable
nature of fracturing (i.e., fracture breakdown) mechanisms, as discussed previously (see Figure 6). The
character of the data shown in Figure 14 also confirms that authors’ belief that the lower bound to many
fracture breakdown pressures should be close to σHmin. It should be pointed out, however, that in several
cases the proximity of fracture pressure to σHmin is likely due to the fact that, unless fracture pressure is
measured, the operators prefer to be conservative and underestimate its value in their calculations, thus
ensuring the safety of their operations.

Figure 13: Variation with depth of fracturing pressure at acid-gas injection sites in the Alberta
Basin, as indicated by operators.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
19
 CONFERENCE PROCEEDINGS

Figure 14: Relation between the minimum horizontal stress (σHmin) and fracturing pressure at acid-
gas injection sites in western Canada.

6. CONCLUSIONS

• In-situ stress magnitudes have a significant impact on the performance of CO2 geological
sequestration sites. They control the likelihood of such potential leakage events as hydraulic
fracturing, re-opening or shearing of natural fractures and faults, and leakage up wellbores due to
enlarged hole conditions.

• A relatively novel “lower bound” methodology has been used in the Alberta Basin to estimate
minimum horizontal stress magnitudes from leak-off and fracture breakdown pressures obtained by
industry drilling and fracture stimulation operations, respectively.

• Minimum horizontal stress gradients close to 17 kPa/m have been estimated for much of the Alberta
Basin from leak-off tests conducted over depths from ~250 to ~1000 m in rocks that have not been
affected by pressure depletion caused by reservoir production.

• Minimum horizontal stresses, with an estimated gradient close to 13 kPa/m for depths ranging from
~1000 to ~2500 m, were interpreted at a regional scale on the basis of fracture breakdown pressures,
and are believed to be affected by reservoir depletion.

• In-situ stresses have been estimated using conventional and “lower bound” techniques at
approximately 50 sites in western Canada where injection has successfully disposed of more than 5

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
20
 CONFERENCE PROCEEDINGS

Mt of acid gas over the past 15 years, and a basin-wide average gradient of 16.6 kPa/m has been
determined for these sites. This result is consistent with the minimum horizontal stress gradient of 17
kPa/m found from leak-off tests.

• Maximum bottom hole injection pressures at the acid-gas injection operations in western Canada are
safely below the minimum horizontal stress. Thus, there has been no driving mechanism to re-
opening pre-existing fractures, if any exist.

• Minimum horizontal stresses are lower by 12.3% on average than rock fracturing pressures, thus there
has been no driving force to induce new fractures.

• Minimum horizontal stresses are 30% less on average than the vertical stress at the injection sites,
confirming that fractures in the Alberta Basin are vertical, except in some shallow zones (a few
hundred meters deep) in the northeast at the basin edge.

• The orientation of induced fractures would be perpendicular to the Rocky Mountain Deformation
Front, varying between ~22 and 60 degrees North at the acid-gas injection sites.

• This investigation has demonstrated techniques for estimating minimum stresses using commonly
available industry data. However, to avoid overly conservative limits on injection pressures, it is
suggested that hydraulic tests should be conducted at each potential sequestration site to determine
the minimum stress magnitude (σHmin) and further to assess if the fracture pressure significantly
exceeds σHmin.

REFERENCES

Addis, M.A. (1997): The Stress-Depletion Response of Reservoirs, SPE Paper 38720, In Proceedings of
the SPE Annual Technical Conference and Exhibit, San Antonio, Texas, October 5-8, 1997, pp. 55-65.
Addis, M.A., Hanssen, T.H., Yassir, N., Willoughby, D.R. and Enever, J. (1998): A Comparison of Leak-
Off Test and Extended Leak-Off Test Data for Stress Estimation. SPE Paper 47235. In Proceedings of
SPE/ISRM Eurock ’98, Vol. 2, Trondheim, Norway, July 8-10, 1998, pp.131-140.
Bachu, S. and Gunter, W.D. (2005): Acid gas injection in the Alberta basin, Canada: a CO2 storage
experience. In Geological Storage of Carbon Dioxide (S.J. Baines and R.H. Worden, eds.).
Geological Society Special Publication, UK, SP 233, p. 225-234, 2004.
Bachu, S. and Gunter, W.D. (2005): Overview of acid-gas injection operations in western Canada. In
Proceedings of 7th International Conference on Greenhouse Gas Control Technologies. Volume 1:
Peer-Reviewed Papers and Plenary Presentations (E.S.Rubin, D.W.Keith and C.F.Gilboy, eds.), IEA
Greenhouse Gas Programme, Cheltenham, UK, In Press.
Bachu, S. and Haug, K. (2005): In-Situ Characteristics of Acid-Gas Injection Operations in the Alberta
Basin, Western Canada: Demonstration of CO2 Geological Storage. In The CO2 Capture and Storage
Project (CCP) for Carbon Dioxide Storage in Deep Geologic Formations For Climate Change
Mitigation, Volume II “Geologic Storage of Carbon Dioxide with Monitoring and Verification” (S.
M. Benson, ed.). Elsevier, London, U.K., p. 867-876, 2005.
Bachu, S. and Stewart, S. (2002): Geological sequestration of anthropogenic carbon dioxide in the
Western Canada Sedimentary Basin. Journal of Canadian Petroleum Technology, v. 41, no. 2, p. 32-
40, 2002.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
21
 CONFERENCE PROCEEDINGS

Bell, J.S. (2003): Practical Methods for Estimating In Situ Stresses for Borehole Stability Applications in
Sedimentary Basins, Journal of Petroleum Science and Engineering, v. 38, pp. 111-119.
Bell, J.S. and Babcock, E.A. (1986): The Stress Regime in the Western Canadian Sedimentary Basin and
Implications for Hydrocarbon Production, Bulletin of Canadian Petroleum Geology, v. 34, pp. 364-
378.
Bell, J.S. and Bachu, S. (2003): In Situ Stress Magnitude and Orientation Estimates for Cretaceous Coal-
Bearing Strata Beneath the Plains Area of Central and Southern Alberta. Bulletin of Canadian
Petroleum Geology, v. 51, no. 1, pp. 1-28.
Bell, J.S., Price, P.R. and McLellan, P.J. (1994): In situ stresses in the Western Canada Sedimentary
Basin. In: Geological Atlas of the Western Canada Sedimentary Basin, G. Mossop and I. Shetsen
(comp.). Canadian Society of Petroleum Geologists and Alberta Research Council, Calgary, AB, pp.
439-446.
Chalaturnyk, R., Zhou, W., Stenhouse, M., Sheppard, M. and Walton, F. (2004): Long-Term Risk
Assessment of the Weyburn Storage Site. In IEA GHG Weyburn CO2 Monitoring and Storage Project,
Summary Report 2000-2004 (M. Wilson and M. Monea, eds.), Petroleum Technology Research
Centre, Regina, Saskatchewan, Canada, pp. 211-268.
Edwards, S.T., Meredith, P.G. and Murrell, S.A.F. (1998): An Investigation of Leak-Off Test Data for
Estimating In Situ Stress Magnitudes: Application to a Basin-Wide Study in the North Sea. SPE
Paper 47272, In Proceedings of SPE/ISRM Rock Mechanics in Petroleum Engineering Eurock ’98,
Vol. 1, Trondheim, Norway, July 8-10, 1998, pp. 357-365.
Gronseth, J.M. and Kry, P.R. (1983): Instantaneous Shut-in Pressure and its Relationship to the Minimum
In Situ Stress. In Hydraulic Fracturing Stress Measurements (M.D. Zoback and B.C. Haimson, eds.),
National Academy Press, Washington, D.C., pp. 55-63.
Hawkes, C.D. and McLellan, P.J. (1999): A New Model for Predicting Time-dependent Failure of Shales:
Theory and Application, Journal of Canadian Petroleum Technology, vol. 38, no. 12, pp. 49-55.
Hawkes, C.D., McLellan, P.J., Zimmer, U. and Bachu, S. (2004): Geomechanical Factors Affecting
Geological Storage of CO2 in Depleted Oil and Gas Reservoirs. In Proceedings of the CIM Petroleum
Society’s 5th Canadian International Petroleum Conference, Calgary, Alberta, Canada, June 8 – 10,
2004, 13 p.
Hubbert, M.K. and Willis, D.G. (1957): Mechanics of Hydraulic Fracturing, Transaction of the American
Institute of Mining, Metallurgical, and Petroleum Engineers, v. 210, pp. 153-166.
McLellan, P.J. (1988): In Situ Stress Prediction and Measurement by Hydraulic Fracturing, Wapiti,
Alberta. Journal of Canadian Petroleum Technology, v. 27, no. 2, pp. 85-95.
Woodland, D.C. and Bell, J.S. (1989): In Situ Stress Magnitudes from Mini-frac Records in Western
Canada. Journal of Canadian Petroleum Technology, v. 28, no. 5, pp. 22-31.
Yielding G. (2002): Shale Gouge Ratio-Calibration by Geohistory. In Hydrocarbon Seal Quantification,
NPF Special Publications 11, pp. 1-15.
Zoback, M.D., Barton, C.A., Brudy, M., Castillo, D.A., Finkbeiner, T., Grollimund, B.R., Moos, D.B.,
Ward, C.D. and Wiprut, D.J. (2003): Determination of Stress Orientation and Magnitude in Deep
Wells, International Journal of Rock Mechanics & Mining Sciences, v. 40, pp. 1049-1076.

FOURTH ANNUAL CONFERENCE ON CARBON CAPTURE AND SEQUESTRATION DOE/NETL

May 2-5, 2005
22