The Geomechanics of CO2 Storage in Deep Sedimentary Formations
- 7.4k Downloads
This paper provides a review of the geomechanics and modeling of geomechanics associated with geologic carbon storage (GCS), focusing on storage in deep sedimentary formations, in particular saline aquifers. The paper first introduces the concept of storage in deep sedimentary formations, the geomechanical processes and issues related with such an operation, and the relevant geomechanical modeling tools. This is followed by a more detailed review of geomechanical aspects, including reservoir stress-strain and microseismicity, well integrity, caprock sealing performance, and the potential for fault reactivation and notable (felt) seismic events. Geomechanical observations at current GCS field deployments, mainly at the In Salah CO2 storage project in Algeria, are also integrated into the review. The In Salah project, with its injection into a relatively thin, low-permeability sandstone is an excellent analogue to the saline aquifers that might be used for large scale GCS in parts of Northwest Europe, the U.S. Midwest, and China. Some of the lessons learned at In Salah related to geomechanics are discussed, including how monitoring of geomechanical responses is used for detecting subsurface geomechanical changes and tracking fluid movements, and how such monitoring and geomechanical analyses have led to preventative changes in the injection parameters. Recently, the importance of geomechanics has become more widely recognized among GCS stakeholders, especially with respect to the potential for triggering notable (felt) seismic events and how such events could impact the long-term integrity of a CO2 repository (as well as how it could impact the public perception of GCS). As described in the paper, to date, no notable seismic event has been reported from any of the current CO2 storage projects, although some unfelt microseismic activities have been detected by geophones. However, potential future commercial GCS operations from large power plants will require injection at a much larger scale. For such large-scale injections, a staged, learn-as-you-go approach is recommended, involving a gradual increase of injection rates combined with continuous monitoring of geomechanical changes, as well as siting beneath a multiple layered overburden for multiple flow barrier protection, should an unexpected deep fault reactivation occur.
KeywordsCO2 storage Geomechanics Caprock Fault reactivation Seismicity Modeling Saline formations
A growing concern that increasing levels of greenhouse gases in the atmosphere are contributing to global climate change has led to a search for economical and environmentally sound ways to reduce carbon dioxide (CO2) emissions. One promising approach is CO2 capture and storage (CCS) in deep geologic formations, such as depleted oil and gas reservoirs, unminable coal seams, and deep saline aquifers (Bachu 2008; Benson and Cole 2008; Bickle 2009; Plasynski et al. 2011). Geological media suitable for CO2 storage must have sufficient capacity and injectivity, and must confine, or at least sufficiently delay the lateral or vertical migration of CO2 to the surface. Such geological media are mainly deep sedimentary formations, including oil and gas reservoirs and deep saline aquifers (Bachu 2008; Benson and Cole 2008; Bickle 2009).
Recent experience from CO2 injection at a number of pilot projects, as well as a few ongoing commercial projects demonstrates that CO2 geological storage in deep sedimentary formations is technologically feasible (Michael et al. 2010). However, if CCS is implemented on the scale needed to make noticeable reductions in atmospheric CO2, a billion metric tons or more must be stored annually, with the largest injection operations in regions with the highest CO2 emissions (Benson and Cole 2008). Securing such a large volume will require a solid scientific foundation, and that defines the coupled hydrologic–geochemical–geomechanical processes that govern the long-term fate of CO2 in the subsurface. Also needed are methods to characterize and select sequestration sites, subsurface engineering to optimize performance and cost, approaches for ensuring safe operation, monitoring technology, remediation methods, regulatory overview, and an institutional approach for managing long-term liability (Benson and Cole 2008).
Geomechanical processes associated with geologic carbon storage (GCS) received some initial attention in the late 1990s. A few published generic modeling studies and discussions indicated that geomechanics would likely play a central role should GCS be applied at a large scale (Rutqvist and Tsang 2002; Saripalli and McGrail 2002; Li et al. 2002; Streit and Hillis. 2004; Yamamoto and Takahashi 2004; Hawkes et al. 2005). More recently, studies of actual CO2 injection sites, such as the In Salah CO2 storage project in Algeria, have shown that significant geomechanical changes may indeed occur, depending on the injection pressure and site specific geomechanical conditions (e.g. Rutqvist et al. 2010; Bissell et al. 2011; Verdon et al. 2011; Zhou et al. 2010). The importance of geomechanics has recently become more widely recognized as the possible magnitude and extent of pressure disturbance associated with large-scale CO2 storage operations have become more apparent (Zoback 2010). There are concerns related to the potential for triggering notable (felt) seismic events and how such events could impact the long-term integrity of a CO2 repository, as well as how it could impact the public perception of GCS (Zoback 2010). The shut-down of a geothermal project in Basel after a 3.4 magnitude seismic event (Häring et al. 2008) is a recent well-known example, in which ground motion felt by nearby communities was a show stopper. Recently, some GCS projects (e.g., in the Netherlands and Germany) have been put on hold because of public concerns about potential earthquakes, leakage, and impact on property values. Ultimately, it might be public perception that will determine whether CCS is implemented at a large scale (Benson and Cole 2008). Thus, large-scale GCS has to be developed with great care, and geomechanics will play a key role in site-specific GCS risk analyses to ensure safe operation.
This paper provides a review of the geomechanics associated with CO2 storage in deep sedimentary formations, in particular saline aquifers, the most promising GCS option in terms of storage capacity, but still challenging in terms of site characterization and assessment of containment. The paper first introduces the concept of storage in deep sedimentary formations, the geomechanical processes and issues related to such an operation, and the relevant geomechanical modeling tools. This is followed by a more detailed review of geomechanical aspects, including reservoir stress-strain and microseismicity, well integrity, caprock sealing performance, and the potential for fault reactivation and notable seismic events. Geomechanical observations at some of the current GCS field deployments are also integrated into the review. To date, the most important and outstanding GCS deployment from a geomechanical perspective is the In Salah CO2 storage project (Mathieson et al. 2010, 2011). At In Salah, CO2 is injected at about a 1.8–1.9 km depth into a relatively tight and thin sandstone formation at an injection overpressure substantially above initial formation pressure. At this site, injection-induced geomechanical changes have been observed and these are important components of the field monitoring program (Mathieson et al. 2010, 2011). Some geomechanical changes have also been observed at other CO2 storage projects, such as Weyburn, Canada (Chalaturnyk 2007; White 2009) and Arneth, Utah (Zhou et al. 2010), which are both associated with enhanced oil recovery (EOR). At the North-Sea-based Sleipner CO2 storage project, on the other hand, no significant geomechanical changes have been detected, which is not surprising considering the minimal injection overpressure at that site (Eiken et al. 2011). The In Salah CO2 storage project, with its relatively low-permeability injection zone, is an excellent analogue for deep saline formations in large parts of Northwest Europe, the U.S. Midwest, and China, where large scale GSC will be an attractive option for mitigating CO2 emissions from large power plants (Wright et al. 2009). Lessons learned from the In Salah project will be particularly important for what we can learn about managing significant geomechanical changes occurring in the field and how such mechanical changes can be utilized for field monitoring and verification of the underground storage performance.
2 GCS in Deep Sedimentary Formations
As mentioned, geologic media suitable for GCS consist mainly of oil and gas reservoirs and deep saline aquifers that are found in sedimentary basins (Bachu 2008; Benson and Cole 2008; Bickle 2009). Storage of gases, including CO2, in these media has been demonstrated on a commercial scale by past EOR operations, natural gas storage, and acid gas disposal. Some of the risks associated with GCS are similar to these industrial activities for which extensive safety and regulatory frameworks are in place. Specific risks are related to the operational (injection) phase and to the post-operational phase, where the risks of most concern are those posed by the potential for acute or chronic CO2 leakage from the storage site (Bachu 2008).
Capacity, to accept the intended volume of CO2.
Injectivity, to take in CO2 at the rate that it is supplied from the CO2 emitter(s).
Confinement, to prevent the migration and leakage of the buoyant and mobile CO2 from the storage space to other places in the subsurface, to shallow potable groundwater or to the surface.
Deep sedimentary formations can possess these characteristics, because generally only sandstone and carbonate rocks have the porosity needed to provide storage capacity and the permeability required for injectivity. At the same time, confining low-permeability shales and evaporites, such as salt beds and anhydrites, provide a primary physical barrier to CO2 migration out of the intended storage unit (Bachu 2008). In such formations, injection could take place at a depth below 800 m so that the CO2 would be within the temperature and pressure range for it to be a supercritical fluid. As a supercritical fluid, the CO2 behaves like a gas with low viscosity, but has a liquid-like density of 200–900 kg/m3, depending on pressure and temperature. Thus, large volumes of CO2 gas would be compressed to a higher density supercritical fluid, and stored in a much smaller volume.
In deep sedimentary formations, CO2 can be stored through a variety of physical and chemical trapping mechanisms (Gunter et al. 2004; Bachu 2008; Benson and Cole 2008). Physical trapping includes: (1) static trapping of mobile CO2 in stratigraphic and structural traps, as well as (2) residual-gas trapping in the pore space at irreducible gas saturation, in which case CO2 is immobile because of the interfacial tension between CO2 and formation water. Chemical trapping occurs when CO2 dissolves in subsurface fluids (solubility and ionic trapping) and may then be involved in chemical reactions with the rock matrix (mineral trapping). The CO2 will first be trapped by primary mechanisms, namely static and hydrodynamic trapping below the caprock in oil and gas reservoirs and deep saline aquifers. Over time, a series of secondary (chemical) trapping mechanisms start operating, mechanisms that do not necessarily increase the CO2 storage capacity, but do increase the storage security (i.e., diminish the potential for leakage and the amount of CO2 that may migrate or leak) (IPCC 2005; Bachu 2008; Benson and Cole 2008). As will be discussed below, while geomechanics is important both for the short and long-term performance of a CO2 repository, it is perhaps most important related to physical trapping during active CO2 injection.
3 Geomechanical Processes and Key Technical Issues
If reservoir pressure becomes sufficiently high, more substantial, irreversible mechanical changes could occur, e.g., creating new fractures, straining the well assembly, or reactivating larger faults within the reservoir, in the caprock or overburden (Fig. 1, bottom right). This could possibly open new flow paths through otherwise low-permeability capping formations and thereby enable enhanced, buoyancy-driven, upward migration of the injected CO2. Moreover, reactivation of a fault could potentially result in a notable seismic event (e.g. magnitude >3) that, if felt, may raise concern in local communities. Notable seismic events might also be induced by reactivating minor faults (e.g., faults a few hundred meters long) that might intersect the injection zone but might not be detectable by seismic surveys. For this reason, a multilayered system of several low permeability layers in the overburden (as illustrated in Fig. 1, top) will be important for impeding upward migration of the CO2, even in the event of fault reactivation at the depth of the injection zone.
Considering these geomechanical processes and key technical issues, it is obvious that geomechanics plays a central role in determining the suitability of an injection site, including the potential for triggering notable seismic events and other risks. A risk assessment will be required at any future commercial GCS site and might be performed at the early stages of a project to help in site selection, communicating project goals and procedures to the public, and aiding regulators in permitting for the project (Stauffer et al. 2009; Oldenburg et al. 2009). In such a risk analysis, key technical risk features, events, and processes related to geomechanics is not limited to notable seismic events; but also include the risk for CO2 release to shallow potable-groundwater aquifers or to the surface, e.g., via leaky wells or faults.
4 Coupled Flow and Geomechanical Models for GCS
In recent years, a number of coupled fluid flow and geomechanical numerical models have emerged as suitable for the analyzing various geomechanical issues associated with GCS. Desirable capabilities for such a numerical simulator includ the capture of non-isothermal multiphase flow and transport for supercritical CO2 and brine coupled with geomechanical processes. Examples of the numerical simulators applied to study the geomechanical aspects of GCS, include TOUGH-FLAC (Rutqvist et al. 2002; Rutqvist 2011), FEMH (Bower and Zyvoloski 1997; Deng et al. 2011), OpenGeoSys (Wang and Kolditz 2007; Goerke et al. 2011), CODE-BRIGHT (Olivella et al. 1994; Vilarrasaa et al. 2010), ECLIPSE-VISAGE (Ouellet et al. 2011), STARS (CMG 2003; Bissell et al. 2011), NUFT-SYNEF (Morris et al. 2011a, b), DYNAFLOW (Preisig and Prévost 2011), as well as other simulators in which multiphase flow codes such as TOUGH2, ECLIPSE, and GEM have been linked with geomechanical codes (e.g. Rohmer and Seyedi 2010; Ferronato et al. 2010; Tran et al. 2010). Some coupled fluid flow and geomechanical simulators can also include geochemistry of various levels of sophistication, from nonreactive to fully reactive solute transport, e.g., linking TOUGHREACT to FLAC3D (Rutqvist et al. 2002; Taron et al. 2009), or Retraso to CodeBright (Kvamme and Liu 2009). Such capabilities can be used to assess long-term geochemical and geomechanical changes in reservoir and caprock when exposed to CO2. These coupled models require a large number of input parameters that might not be readily available, such as parameters for geomechanical-geochemical interactions. In fact, simplified models might be sufficient for studying subsets of coupled processes. For example, when studying large-scale geomechanical changes and the potential for fault reactivation, driven by large-scale reservoir pressure changes, a single phase fluid flow model coupled with geomechanics might be sufficient (e.g. Chiaramonte et al. 2011a). Moreover, analytical and semi-analytical models (e.g. Streit and Hillis 2004; Soltanzadeh and Hawkes 2009; Selvadurai 2009; Mathias et al. 2010; Rohmer and Olivier 2010) or numerical multiphase flow models linked with analytical geomechanical models (e.g. Lucier et al. 2006; Chiaramonte et al. 2008; Vidal-Gilbert et al. 2010) can also be useful for first-order analysis and quick assessment of CO2 injection-site suitability.
5 Injection-Induced Strain, Stress and Microseismicity
As soon as fluid injection starts, changes in reservoir stresses and strains can quickly propagate laterally within the injection zone, along with an expanding fluid pressure. The pressurization causes vertical expansion of the reservoir and changes in the stress field. These induced changes are, in general, proportional to the magnitude of the pressure increase, ΔP, and depend on the geometry and geomechanical properties (such as compressibility) of the reservoir and surrounding sediments.
The stress evolution within and around the reservoir is important as a source of potential mechanical inelastic responses, including shear reactivation of existing fractures that could result in seismic events. During injection, the vertical stress will be approximately constant, equal to the weight of the overburden. The horizontal stress, on the other hand, will change with the injection as a result of poro-elastic stress, a phenomenon commonly observed during oil and gas production (Zoback 2007). The injection-induced changes in horizontal stress will depend on the geometry and poro-elastic properties of the reservoir-caprock system (Rutqvist et al. 2007, 2008).
Again, a simple analytical analysis may be useful, this time to estimate the potential for induced seismicity and compare with the numerical simulation results. Using the stress magnitudes shown in Fig. 5, and an initial hydrostatic fluid pressure of 18.1 MPa, the strength-to-stress margin before injection can be estimated to σ′mci = 3σ′h − σ′H = 3(Sh − Pi) − (SH − Pi) = 3(28.6 − 18.1) − (45.5 − 18.1) = 4.1 MPa. For a maximum fluid pressure of about 28.1 MPa (or an overpressure ΔP = 10 MPa), the strength-to-stress margin can be estimated as σ′mc = 3σ′h − σ′H = 3(Sh + Δσh − Pi − ΔP) − (SH + Δσh − Pi − ΔP) = σ′mci + 2(Δσh − ΔP) = 4.1 + 2(7.5 − 10) = −0.9 MPa. To this, one can also add an estimate of stress changes caused by the cooling shrinkage of the rock around the wellbore. For the volume around the wellbore at the scale of tens of meters, a maximum cooling of about ΔT = −5°C was calculated from the numerical simulation (Fig. 4). One can then estimate the thermal stress from ΔσT = αT × ΔT × E/(1 − 2υ) (assuming a fully constrained sediment) where αT is the coefficient of linear thermal expansion. For αT = 1 × 10−5 °C−1, ΔT = −5°C, E = 6 GPa, ν = 0.2, we arrive at a cooling-induced stress of −0.5 MPa. The strength-to-stress margin can then be estimated as σ′mc = σ′mci + 2(Δσp +ΔσT − ΔP) = 4.1 + 2(7.5 − 0.5 − 10) = −1.7 MPa. This value is consistent with the numerically calculated values close to the injection well in Fig. 5 and therefore provides confidence in the numerically calculated distribution.
The observations from both Weyburn and Aneth, illustrate the fact that geomechanical changes can occur outside the injection zone and are not necessarily confined to areas where the actual pressure changes occurs (Rutqvist et al. 2008). Microseismic activities in the overburden have also been observed during reservoir pressure depletion associated with hydrocarbon production at some fields (Segall 1989; Odonne et al. 1999). Note, however that all the injection induced microseismc events recorded at In Salah, Otway, Weyburn and Aneth to date are so small that they are only detectable using geophones. For example, the biggest events of magnitude 0 or less as recorded at Aneth and Otway could be associated with shear slip along fractures with radius on the order of a meter or less. The clustered microseismic activities along structures such as observed at Aneth have also been observed at other underground injection operations, including the 1970s fluid injection experiments at Matsushiro, Japan (Ohtake 1974; Cappa et al. 2009), and Rangeley, Colorado (Raleigh et al. 1976), as well as at the Soultz geothermal project in France (Baisch et al. 2010). This indicates the existence of a fracture network, for example associated with the damage zone around a fault or within fractured rock zones where smaller (meter scale) fractures can be reactivated or mechanically coalesced. As will be discussed in more detail in Sect. 9, larger events that can be felt would require much larger source dimension. For example, a magnitude 4 seismic event would require a km-sized fault-rupture area (Cappa and Rutqvist 2011b).
Finally, reservoir stress and strain evolution might also be impacted by chemically mediated compaction, especially in carbonate rich reservoirs, where CO2 exposure would promote dissolution of minerals. It is well known from CO2 flooding related to EOR that dissolution of carbonates in rock-forming minerals may enhance porosity and permeability, leading to worm holes and fingering (Ross et al. 1982). Moreover, enhanced mineral dissolution may also lead to mechanical weakening as well as enhanced compaction creep and subsidence (Nguyen et al. 2011; Spiers et al. 2010; Le Guen et al. 2007; Espinoza et al. 2011). A number of laboratory studies have investigated the effects of supercritical CO2 on rock samples, showing the importance of intergranular pressure solution, especially in limestone with rock-forming carbonates, but also in sandstones with cementing carbonates. In general, experiments have shown that CO2 exposure leads to acceleration of creep by up to several orders of magnitude in carbonate-rich rocks (Le Guen et al. 2007; Spiers et al. 2010). In a field setting, such increased compaction strain will not only impact the stress evolution within the reservoir, but (through stress transfer) the surrounding rock and overburden as well, and lead to some irreversible subsidence. Very few laboratory data exist and no field studies related to CO2 injection have been focused on the potential for chemically mediated compaction so it is difficult to quantify this effect. Le Guen et al. (2007) observed about 1% compaction strain over a year of exposure to CO2-rich fluid and to estimate long term effects such data could perhaps be extrapolated to tens or hundreds of years using mechanical creep models.
6 GCS Injectivity and Limits on Injection Pressure
For a site with a given injectivity, the injection rate will be limited by the maximum overpressure that can safely be applied without causing any unwanted damaging geomechanical changes that could jeopardize the confinement of the injected CO2. Future regulations related to underground CO2 injection might limit the admissible injection pressure to the fracturing gradient evaluated by leak-off tests in the caprock, just above the intended storage zone. For example, it is the current practice for deep injection of liquid waste in the U.S regulatory standard that maximum injection pressure should be less than the measured fracture-closure pressure. Below the fracture-closure pressure, any existing fractures cannot open and no new fractures can form, implying no enhanced migration of waste fluids out of the injection intervals (USEPA 1994). The regional guidance for implementation is that the maximum injection pressure can be determined either by a site-specific fracture-closure pressure derived from direct or indirect testing, or by formation-specific default values for the fracture-closure pressure gradients. For example, a default value of 0.0129 MPa/m (130% of the hydrostatic pressure gradient) is given for the Mt. Simon formation in Illinois; 0.0181 MPa/m (181% of the hydrostatic pressure gradient) is reported for the Dundee Limestone in the Michigan Basin (USEPA 1994; Zhou et al. 2008). In the case of acid gas injections in western Canada, the operating license limits the maximum bottom-hole pressure to no higher than 90% of the fracturing pressure, to ensure that fracturing of the caprock does not occur, but in many cases actual bottom-hole pressure is much lower when injection takes place in depleted oil or gas reservoirs (IAE 2003). However, in the case of potential future commercial CO2 injections from an emitting power plant, the injection will take place at a much larger scale, and therefore other issues, such as potential for fault reactivation and the long-term sealing performance, may be limiting factors for the admissible overpressure (Streit and Hillis 2004; Hawkes et al. 2005; Lucier et al. 2006; Li et al. 2006; Rutqvist et al. 2007, 2008).
Lucier et al. (2006) performed one of the first site-specific geomechanics and injectivity studies, related to carbon capture and storage at the American Electric Power’s (AEP) 1.3-GW Mountaineer Power Plant in New Haven, West Virginia. Lucier et al. (2006) completed a coupled fluid flow and geomechanical analysis of the Rose Run Sandstone, a potential injection zone located at a depth of about 2,500 m. They found that injectivity would be quite limited as a result of limited thickness and low to moderate permeability. They estimated that for a single injection well, the formation could take only a small fraction of the Mountaineer emissions (7 Mt/year), even when applying a relatively high injection pressure (limited to a magnitude just below the fracturing gradient). Moreover, the analysis indicated that induced seismicity in the caprock could occur with a slight increase in pressure if optimally oriented faults would be present (Lucier et al. 2006). However, the site is overlain by multiple low-permeability layers of shales and permeable strata. Consequently, even if a minor fault were reactivated in the caprock just above the injection zone, and if such reactivation would open a new flow path, the overlying multiple shale layers could provide sufficient protection from upward migration of the CO2 (Gupta et al. 2011). Since 2009, some CO2 injection has commenced within a 20 MW CO2 capture and storage Product Validation Facility (PVF). The results showed that the Rose Run Sandstone indeed has limited permeability thickness and suffers from a relatively low fracturing gradient. However, another rock unit, the Copper Ridge Dolomite, with thin vuggy dolomite interbed, had about 100 times higher permeability-thickness (Gupta et al. 2011). It was concluded that the vuggy zones in the Copper Ridge Dolomite have significant local storage potential and likely also to offer regional-scale storage. Nevertheless, in July 2011, the Mountaineer project was placed on hold by AEP, because of uncertain economic and policy conditions.
The In Salah CO2 storage project is also a prime example with regard to injectivity management. As mentioned, the injection takes place at a depth of about 1.8–1.9 km, into a relatively low-permeability, 20 m thick, water-filled Carboniferous, fractured sandstone (Wright 2011; Mathieson et al. 2011). To ensure adequate injectivity and CO2 flow rates across the low-permeability sand face, it was decided to use long-reach (about 1–1.5 km) horizontal injection wells. Moreover, the horizontal injection wells were oriented parallel to the direction of the minimum principal horizontal stress and perpendicular to the orientation of the dominant fracture set, which is vertical and strikes roughly NW-SE (Fig. 2). In addition, the three CO2 injection wells, KB-501, KB-502 and KB-503, were strategically located in areas of high porosity-thickness, as evaluated from a 3D seismic survey (Riddiford et al. 2004; Foster et al. 2004). Reservoir simulations of the CO2 injection at In Salah indicate dominant fracture flow at least at injection well KB-502, where the effective permeability exceeds that of the intact sandstone matrix (Bissell et al. 2011).
At In Salah, the maximum wellhead pressure for CO2 injection was initially set to 18 MPa, in order for the flowing bottom-hole pressure to be below the fracturing pressure of the caprock above the injection zone. This was based on the best estimate of the minimum principal stress depth gradient, which was evaluated from leak-off tests to correspond to a mud weight of 1.7 g/cc (Darling 2006). At 1,800 m depth, this will correspond to about 30 MPa. Using wellbore simulations tools, Bissell et al. (2011) and Rutqvist et al. (2011b) showed that downhole pressure might at times have been higher than 30 MPa, which according Bissell et al. (2011) would exceed the fracturing pressure within the reservoir (but not in the caprock above the injection zone). However, this fracturing pressure refers to the internal wellbore pressure that could fracture the wall of a vertical well considering the rock tensile strength. The most important parameter for maximum sustainable injection pressure is the minimum compressive in situ stress, which for the depth of the injection zone and the caprock above the injection zone can be estimated to about 28.5–30 MPa (Darling 2006; Rutqvist et al. 2011a).
Injection data at one of the injection wells at In Salah (KB-502) have shown a pressure-dependent injectivity (Bissell et al. 2011). In fact, reservoir simulations have shown that it is not possible to match the injection data without considering some type of increasing injectivity with pressure. Bissell et al. (2011) indicated that the injection data showed an abrupt injectivity increase at a bottomhole pressure of 28.6 MPa, which would indicate a sudden fracture opening, with complete loss of contacts between fracture surfaces. The observed permeability changes at KB-502 might also be related to an apparent opening of a deep fractured zone as observed by InSAR ground surface deformation monitoring. The potential for opening of a deep-seated fractured zone will be discussed in more detail in Sect. 9. However, in terms of injectivity management, the maximum well head pressure has now been lowered at KB-502, below the limit of the apparent fracture opening. This is an example of monitoring coupled fluid flow and geomechanical processes leading to important observations regarding the subsurface reservoir performance, and preventative actions.
7 Wellbore Integrity
The well design of a deep CO2 injection well will typically involve at least two strings of casing, including surface casing cemented to the land surface to isolate the well from the shallower aquifers of drinking water (Tsang et al. 2008). The second casing will extend all the way to the injection zone and is cemented to ensure that there is no cross flow between adjacent brine formations. Furthermore, injection tubing is set into a packer, which is a mechanical device set in the well to isolate the injection zone, to ensure that injection via the tubing is emplaced in the target injection zone (Tsang et al. 2008).
Geomechanical processes can impact well integrity both during drilling and completion, as well as during the actual CO2 injection. Drilling and completion of wells through layers of shale and mudstone requires special attention, because rock failure and deformation associated with wellbore instability may damage wellbores, leading to increased permeability and potential leakage paths (Orlic 2009). Pore-fluid chemistry affects chemically sensitive rocks, particularly shales, so that their mechanical properties may change substantially (Dusseault 2011). Radial deformation of the cement sheath is commonly caused by shrinkage during cement hydration, which might result in cracking of the cement sheath or debonding at the rock/cement or cement/casing interface, allowing for radial and vertical migration of fluids (Orlic 2009). During CO2 injection, changes in reservoir stress and deformations acting on the well assembly will impact the hydraulic aperture of rock-cement and casing-cement interfaces, the most likely pathways for potential CO2 leakage around abandoned wells (Bachu and Bennion 2009; Tau et al. 2011). Fractures, if formed in the concrete, will have a rock-fracture-like stress-dependent aperture that could be significantly impacted by coupled chemical degradation once exposed to CO2 (Huerta et al. 2009; Wigand et al. 2009). Analysis of well geomechanics may require a detailed model including all the mechanical components of the well assembly, linked to reservoir-scale modeling of reservoir pressure, temperature, and stress evolution (Rutqvist et al. 2011b). However, such detailed geomechanical modeling would not be conducted for all the wells at a site. From a geomechanics perspective, the first analysis would be to look at the potential compaction strains and shear strains in the geological layers—and their interfaces—and then evaluate their potential impact on the well assembly. Analysis of potential CO2 leakage along wells may then be conducted using separate multiphase flow analysis such as exemplified in the work by Nordbotten et al. (2009), including large number abandoned wells. The analysis of Nordbotten et al. (2009), shows the benefits of multiple caprock layers to mitigate upward leakage rates.
At In Salah, the injection is conducted from state-of-the-art CO2 injection wells, cased and cemented down to the injection zone. As in any major drilling campaign, a wellbore stability study was performed before the start of drilling. However, severe instability problems were encountered in the first well (KB-501), requiring a major redesign of the program (Foster et al. 2004). Despite this, the well successfully reached a horizontal length of 1,300 m (Foster et al. 2004). The horizontal well sections are 6 inch open hole completions with a pre-slotted liner, oriented optimally with respect to the in situ stress field to avoid borehole instability. CO2 leakage is unlikely along these state-of-the art injectors. However, there has been one instance of very small CO2 leakage along an old appraisal well that was drilled in the 1980s and not cemented across the interval of the CO2 injection zone (Mathieson et al. 2011; Ringrose et al. 2009; Oldenburg et al. 2009). This highlights the risk with abandoned wells, although in this case the leakage was related to a faulty wellhead and not to geomechanical or geochemical changes caused by CO2 injection. Nevertheless, the well has since been fully decommissioned, and surface flux and soil gas monitoring will continue around the well for the foreseeable future (Mathieson et al. 2011).
8 Caprock Sealing Performace
In oil and gas reservoirs, the ability to contain buoyant gas has been proven, but in the case of saline aquifer storage, the sealing properties of the overlying barrier layers may not be well known, or cannot be that well constrained (Orlic 2009). However, experience from both industrial and natural analogues indicates that aquifer storage is indeed a viable option in the right type of settings, contingent upon having an adequate caprock as a top seal. For example, underground gas storage in aquifers is an industrial analogue that like GCS also requires gas injection at an overpressure above ambient hydrostatic water pressure (Katz and Tek 1981). Among the few reported instances of gas leaks from such facilities, the majority have been problems related to the effectiveness of the caprock (Evans 2009). For example, at the Leroy gas storage facility in Wyoming, observations of gas bubbles at the ground surface suggested a pressure-triggered leakage mechanism involving hydraulic seal changes in the reservoir caprock at an overpressure of a few MPa (Araktingi et al. 1984; Nelson et al. 2005).
If the CO2 overpressure exceeds the capillary entry pressure, the caprock will act as a permeability barrier in the sense of allowing for Darcy type flow, as shown in Fig. 8, upper right (Ingram et al. 1997). The capillary retention properties are still important, because relative permeability effects could reduce the CO2 mobility substantially and thereby further reduce the upward buoyancy migration of the supercritical CO2. A high injection pressure may cause fracturing or opening of fractures and minor faults that might exist in the caprock, causing a fractured seal as shown in Fig. 8, lower left. Such fracturing can increase the connectivity of permeable interbeds, and if it could propagate across all the overlying sealing layers, a breach of the top seal may occur (Ingram et al. 1997).
The importance of existing faults and fractures on caprock sealing performance have also been observed related to geological containment of gases in naturally overpressured sediments and gas reservoirs (Poston and Berg 1997; Sibson 2003). In such formations, re-shear of existing cohesionless faults and fractures favorably oriented for frictional reactivation provides the lower bound to overpressures, whereas drainage of conduits by hydraulic-extension fracturing is important only in the case of intact caprock under low differential stress (Sibson 2003). The mechanical properties of the top seal units are also critical in determining how shear reactivation could change the sealing capacity of a caprock unit (Fig. 8). The shear strength of mudrock and shale could be very low, with friction angle as low as 10 degrees (Ingram and Urai 1999; Hajdarwish 2006). However, despite low strength, a caprock unit consisting of soft plastic (ductile) rock can deform in a plastic but self-sealing way, and might therefore be more favorable than a stiff, brittle and dilatant rock (Nygård et al 2006: Ingram and Urai 1999). Laboratory tests have shown that permeability of fractures in shale may not necessarily increase with shear, but could actually decrease depending on the stress normal to the fractures (Gutierreza et al. 2000).
Minor faults intersecting a shale caprock might have more complex architecture. For example, faults with offsets on the order of a few meters may have a low permeability fault core of a few centimeters and a more permeable damage zone of several tens of centimeters on each side (Constantin et al. 2004). A fracture network in the damage zone might initially be mineral filled and tight and sealing, but with sufficiently high pressure, fluid may propagate through the fracture network, breaking mineral sealing and thereby opening up a new flow path along the fault’s damage zone. Again, geomechanical monitoring of surface deformations and microseismicity is crucial for early detection of unwanted and potentially damaging geomechanical changes, such as reactivation of a fault intersecting the caprock. An example of such monitoring is provided by the In Salah project, in which InSAR monitoring provided indications of a fractured zone opening at the lowermost parts of the caprock; as a result, preventative actions regarding the maximum injection pressure were taken (Mathieson et al. 2011; Bissell et al. 2011). The apparent opening of this deep-seated fractured zone, and geomechanical analysis of that event will be discussed in the next section.
The potential impact of chemical-mechanical coupled processes on caprock sealing performance has been investigated by several research groups, and these investigations have indicated that they would have a minor impact on shales or mudstone caprocks. Studies generally show that for a low-permeability caprock, over thousand of years of exposure, only the lowest few meters of the caprock may be affected by the CO2-rich fluid, and the effect on hydraulic and mechanical properties appears to be very small (Gaus et al. 2005; Ojala 2011; Fleury et al. 2010; Hangx et al. 2010; Bildstein et al. 2010). The potential for CO2 replacing interlayer water leading to shrinkage of the clay minerals is an open issue (Gaus 2010). If such CO2-induced shrinkage would occur in the lower parts of a caprock, the horizontal stress might be significantly reduced, and shrinkage cracks could occur. These processes have been mentioned associated with upwelling of CO2 across thin shale layers at Sleipner (Chadwick et al. 2004). Similar shrinkage effects as a result of mineral dissolution in carbonate-rich rock are also possible (Espinoza et al. 2011). However, most of the important coupled chemical-mechanical effects are expected to occur along faults and fractures that might exist or were created by high CO2 injection pressure (Johnson et al. 2005; Gherardi et al. 2007; Andreani et al. 2008; Min et al. 2009; Gaus 2010).
9 Potential Fault Reactivation and Notable Seismic Events
The potential for injection-induced fault reactivation and seismicity is an important issue from safety, storage security, and public acceptance perspectives. Since the well known Denver Basin earthquakes associated with underground injection in the 1960s, more cases of injection-induced seismicity have been documented (Nicholson and Wesson 1992; Sminchak and Gupta 2003; Nicol et al. 2011). In general, where injection-induced seismicity has been triggered and monitored, the rates and maximum magnitudes of induced events increase with rising reservoir pressures, total fluid volumes, and injection/extraction rates (Nicol et al. 2011). The similarity between water injection and CO2 injection in terms of injection induced seismicity at the smaller scale has been indicated by Verdon et al. (2010). Note, though, that of the thousands of injection wells in the world, only a few have induced notable seismic events (Nicholson and Wesson 1992). Thus, while some unfelt microseismic events are to be expected at many sites, notable seismic events may only have the potential to occur at some specific sites, related to local structural geology, stress conditions, rock-mass properties (e.g., soft and ductile or hard and brittle, or fractured), and depending on injection parameters.
Injection-induced seismic events result from the release of stored energy triggered by the increase in fluid pressure from the injection. This is not just limited to seismically active areas, but could also occur within the seismically quiet intraplate crust, such as in the midwestern United States (Zoback 2010). The intraplate brittle crust may often be under a compressive stress state of frictional failure equilibrium and have stored strain energy that can be released by fluid injection, but would otherwise remain stored (Zoback and Zoback 1989; Zoback and Harjes 1997). The reason for these areas being seismically quiet is not the stress field, but its being subjected to a low crustal strain rate. A large-scale CO2 operation would likely induce strain rates far exceeding the natural background strain rates within and around the area pressurized by the injection. This is not only of concern in the areas occupied by the stored CO2 around the injection well, but up to several tens of kilometers away, where significant pressure buildup and interference with other injection wells could occur (Birkholzer and Zhou 2009). Over such a vast area, some significant faults may not be avoidable, and the likelihood and potential consequences of reactivation of such faults needs to be evaluated at any injection site (Morris et al. 2011b).
There are many recent examples of site-specific modeling studies related to predicting the potential for fault reactivation and other geomechanical processes during CO2 injection. These include the Mountaineer project, West Virginia (Lucier et al. 2006), the Teapot Dome Pilot Field, Wyoming (Chiaramonte et al. 2008, 2011b), the Dogger Carbonates of the Paris Basin (Vidal-Gilbert et al. 2009), a petroleum onshore field, Brazil (Mendes et al. 2010), the In Salah CO2 storage project, Algeria (Rutqvist et al. 2010; Bissell et al. 2011; Morris et al. 2011a, b: Baroni et al. 2011; Fokker et al. 2011), the Po River plain, Northern Italy (Ferronato et al. 2010), an offshore gas field in the Dutch sector of the North Sea (Orlic et al. 2011), Snøhvit, Barents Sea (Chiaramonte et al. 2011a); Otway, Australia (Vidal-Gilbert et al. 2010), and Ketzin, Germany (Ouellet et al. 2011). The majority of these analyses show that CO2 injection may induce fault reactivation depending on the applied injection pressure, but there are great uncertainties in the in situ stress field as well as in the assumed strength of fault properties. There might be an estimate of the depth-dependent stress field based on the word stress map, site-specific leak-off tests, and borehole break-out data (Zoback 2007). However, across faults, the local stress field might be significantly different than the average crustal stress field. Also, the fault architecture and shear strength along a fault could be strongly heterogeneous. Therefore, it is very difficult to predict when and exactly at what injection pressure a fault might reactivate. On the other hand, coupled fluid flow and geomechanical modeling may reasonably indicate how the stress field changes as a result of injection, and in which areas the potential for fault reactivation will increase or decrease with the injection. Thus, it can be used to identify areas and individual known faults with the highest potential for reactivation, and thereby guide the field monitoring. Once field data are available from the actual injection, such data can then be used to recalibrate the coupled models, and then forward modeling can be used for testing new injection parameters.
Another concern is that the reactivation of a minor fault might open up a new flow path for upward migration of CO2, as discussed above in Sect. 7. In fact, natural larger-scale releases of CO2, such as in volcanic areas, frequently occur along faults, usually along channels or intersections of faults (e.g. Annunziatellis et al. 2008; Cappa et al. 2009). Gas leakage though faults from hydrocarbon reservoirs or underground gas storage facilities has also been inferred from seismic surveys as a “push down” of the seismic signature (Zoback and Zinke 2002; Juhlin et al. 2007). For example, Zoback and Zinke (2002) correlated hydrocarbon leakage through faults observed from seismic signatures to areas where gas production would tend to cause the stress field to move towards failure. Moreover, natural gas seeps are also structurally controlled along faults, such as evident from offshore fluid seepage in the Santa Barbara Basin, California (Eichhubl et al. 2000). These are examples where gas seepage out of a reservoir has been observed, whereas there are thousands of gas fields proven to contain gas for millions of years, with no sudden gas release, even for a large number of gas-producing field areas near active faults, such as at Ventura Basin, California. Thus, it is very difficult to predict how much fault permeability will change in the case of an injection-induced fault reaction. However, fault permeability could be subject of sensitivity studies and potential permeability changes could perhaps be bounded using fault permeability models (e.g., Cappa and Rutqvist 2011a; Seyedi et al. 2011) and to investigate potential consequence, should a fault be reactivated.
Finally, note that the injection zone at In Salah is intersected by a number of faults that theoretically could be close to being reactivated in strike-slip shear (Morris et al. 2011a). These minor faults may extend into immediate caprock horizon, but there are no faults mapped extending from the injection zone through the caprock and into the shallow potable water aquifer. The faults appear to have some effect on the fluid flow and the ground surface uplift pattern (Ringrose et al. 2011, Morris et al. 2011a). However, to date, the CO2 injection operation at In Salah has not resulted in any felt seismic event and InSAR surface deformations do not indicate substantial strike-slip shear movements. For example, if a fault did reactivate in a strike-slip shear movement of a few centimeters or more, it should be visible on the InSAR surface deformation data. The fact that faults have not been substantially reactivated in strike-slip could indicate that the reservoir pressure is not sufficiently high to cause such strike-slip fault reactivation. However, even if the reservoir pressure were sufficiently high, the faults may be prevented to slip as a result of mechanical resistance from the surrounding rock. Note that only a small (20 m thick) part of the faults intersecting the injection zone might be pressurized, whereas parts of the fault planes above and below the injection zone might not be pressurized. The faults could thereby be mechanically restricted from substantial strike-slip movements. Nevertheless, the In Salah is another example of an injection operation in which no felt seismic events occur, and for the development of GCS at a larger scale, there is a need to further develop methods for identifying and distinguishing such sites from those that would be more prone to injection-induced seismicity.
10 Concluding Remarks and Recommendations
The importance of the geomechanical processes and modeling associated with geologic CO2 storage can hardly be overemphasized. In particular, recent concerns about the potential for injection-induced fault reactivation and notable (felt) seismic events are key geomechanical issues that need to be addressed before commercial scale GCS operations can be deployed at a large scale. Although no felt seismic event has been reported from any of the current CO2 storage projects, it should be recognized that potential future commercial GCS operations from large power plants will require injection at a much larger scale. Geomechanical modeling can certainly be used to guide the site selection and evaluate the potential for injection-induced fault reactivation and seismic events, but as mentioned, it is difficult to predict such an event and its magnitude. The main uncertainties are the in situ stress field and large-scale geomechanical properties and their distributions. It is the large-scale pressure buildup, associated crustal straining, and potential undetected faults that might be of greatest concern. The risk is generally expected to increase with injection volume, since this will increase the possibility that the expanding reservoir pressure reaches critically stressed faults of larger dimensions. To deal with this issue, a best-practice framework is needed for the site investigations. This could involve documentation of the historical natural seismicity, assessment of the potential for induced seismicity, and recommended steps for mitigation of the risk of the induced seismicity, as well as addressing the human element (Myer and Daley 2011).
Notwithstanding the potential for triggering notable (felt) seismic events, the potential for buoyancy-driven CO2 to reach potable ground water and the ground surface is an important issue from safety and storage-efficiency perspectives. Much depends on the caprock and overburden fluid flow properties and also how these might change as a result of geomechanical changes. For large-scale CO2 storage in deep saline aquifers, it appears that substantial overpressure may be required—lower than the fracturing pressure of the caprock, but perhaps exceeding the caprock’s capillary entry pressure and the caprock would then act as a permeability barrier. We know from aforementioned natural and industrial analogues that fractures and faults in the caprock are important, and pressure-induced reshear of such fractures and faults could give rise to increased gas seepage sufficient to reduce the storage pressure over geologic time (Sibson 2003). This might be a small change in permeability and seepage when related to CO2 storage, taken over a time period of 100 or 1,000 years. Still, such changes and potentially increased CO2 seepage could reduce the effectiveness of a CO2 storage operation.
Considering potential issues related to geomechanical changes associated with large-scale CO2 injection operations, a staged, learn-as-you-go approach, in a multibarrier repository would be recommended. Such an approach would include the following components: (1) selecting a site with an overburden of multiple low-permeability layers, providing a multiple barrier system, (2) carefully designing and regulating the injection pressure based on estimated in situ stress to avoid damaging geomechanical changes (3) conducting longer-term injection tests using water or CO2 combined with monitoring of geomechanical changes (such as ground surface deformations and microseismicity) to evaluate the potential for induced seismicity and to calibrate geomechanical models, (4) redesigning injection parameters and gradually ramping up the CO2 injection rate under continuous monitoring of geomechanical changes, (5) conducting long-term operational injection with continuous monitoring of geomechanical changes for early detection of potentially damaging geomechanical changes and for tracking of underground fluid movements.
In this staged, learn-as-you-go approach, the monitoring of geomechanical changes (such as ground-surface deformations and induced microseismicity) is vital for tracking underground fluid pressure evolution, possibly detecting emerging fault reactivations, and providing an early detection of potential migration out of the storage complex. As mentioned above, the use of such monitoring and management of significant geomechanical changes has been demonstrated at the In Salah CO2 storage project. A staged approach with an initially longer-term injection, accompanied by seismic and surface-deformation monitoring can be used to determine how prone the site is to triggering notable (felt) seismicity, and can provide the necessary data on site-specific seismicity (if any) to make predictions for long-term operational conditions. For example, at In Salah, it appears that many minor faults within the reservoir could be close to being critically stressed for strike-slip reactivation, but monitoring during injection shows only some microseismic events, while no felt events has occurred. During CO2 operation, geomechanical monitoring, with early detection of unwanted geomechanical changes, enables preventative actions, e.g., reducing injection pressure. This has also been demonstrated in practice at In Salah. However, at injection operations where significant induced seismicity occurs, it is recommended to make such changes in the injection parameters gradually, avoiding abrupt changes, because many of the largest seismic events at other types of underground injections (such as the aforementioned Basel project) have occurred after shutting down the injection.
Finally, GCS beneath a system of multilayered overburden can be very beneficial from storage security perspective, because it provides multiple flow barriers against upward buoyancy-driven CO2 migrations, which may be crucial in the event of unexpected reactivation of deep faults that might cross the storage unit. Thus, even if a deep fault is reactivated, and if such reactivation would lead to a local flow path out of the intended storage unit, the overlying low permeability layers could still provide necessary protection by impeding upward migration of the CO2.
This work was jointly supported by the Assistant Secretary for Fossil Energy, Office of Natural Gas and Petroleum Technology, through the National Energy Technology Laboratory, and the In Salah JIP and their partners BP, Statoil, and Sonatrach under the U.S. Department of Energy Contract No. DE-AC02-05CH11231. Technical review comments by Allan Mathieson and Rob Bissell of the In Salah JIP, U.K.; Bogdan Orlic at TNO, Utrecht, The Netherlands; Frederic Cappa, at the Berkeley Lab; Pengzhi Pan at the Chinese Academy of Sciences, Wuhan, China; Shinsuke Nakao and Takashiro Funatsu at AIST, Tsukuba, Japan; Ki-Bok Min and Jaewon Lee at the Seoul National University, Korea; as well as editorial review by Dan Hawkes at the Berkeley Lab are all greatly appreciated.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Abbot A (2004) Plans resurrected to raise Venice above the encroachment sea. Nature 427:184Google Scholar
- Araktingi RE, Benefield ME, Bessenyei Z, Coats KH, Tek MR (1984) Leroy storage facility, Uinta County, Wyoming—a case history of attempted gas migration control. J Petrol Technol 36(1):132–140Google Scholar
- Bachu S, Bennion DB (2009) Experimental assessment of brine and/or CO2 leakage through well cements at reservoir conditions. Int J Greenh Gas Control 3:494–501Google Scholar
- Baroni A, Estublier E, Deflandre JP, Daniel JM (2011) Modelling surface displacements associated with CO2 re-injection at Krechba. In: Proceedings of 45th US Rock mechanics/geomechanics symposium, San Francisco, CA, June 26–29Google Scholar
- Bickle MJ (2009) Geological carbon storage. Nature Geosci 2(December 2009):815–819Google Scholar
- Bildstein O, Kervévan C, Lagneau V, Delaplace P, Crédoz A, Audigane P, Perfetti E, Jacquemet N, Jullien M (2010) Integrative modeling of caprock integrity in the context of CO2 storage: evolution of transport and geochemical properties and impact on performance and safety assessment. Oil Gas Sci Technol Rev IFP 65:485–502Google Scholar
- Bissell RC, Vasco DW, Atbi M, Hamdani M, Okwelegbe M, Goldwater MH (2011) A full field simulation of the In Salah gas production and CO2 storage project using a coupled geo-mechanical and thermal fluid Flow simulator. Energy Procedia 4:3290–3297Google Scholar
- Burch DN, Daniels J, Gillard M, Underhill W, Exler VA, Favoretti L, Le Calves J, Lecerf B, Potapenko D, Maschio L, Morales JA, Samuelson M, Weimann MI (2009) Live hydraulic fracturing monitoring and diversion. Oilfield Rev 21:18–31Google Scholar
- Cappa F, Rutqvist J (2011b) Impact of CO2 geological sequestration on the nucleation of earthquakes. Geophys Res Lett 38:L17313Google Scholar
- Carles P, Bachaud P, Lasseur E, Berne P, Bretonnier P (2010) Confining properties of carbonated Dogger caprocks (Parisian Basin) for CO2 storage purpose. Oil Gas Sci Technol Rev IFP 65:461–472Google Scholar
- Chadwick A, Arts R, Bernstone C, May F, Thibeau S, Zweigel P (2008) Best practice for the storage of CO2 in saline aquifers: observations and guidelines from the SACS and CO2STORE projects. British Geological Survey, KeyworthGoogle Scholar
- Chalaturnyk RJ (2007) Geomechanical characterization of the Weyburn field for geological storage of CO2. In: Proceedings of 1st Canada-US rock mechanics symposium, Vancouver, Canada, 27–31 May 2007Google Scholar
- Chiaramonte L, White JA, Johnson S (2011a) Preliminary geomechanical analysis of CO2 injection at Snøhvit, Norway. In: Proceedings of 45th US Rock mechanics/geomechanics symposium, San Francisco, CA, June 26–29Google Scholar
- CMG (2003) User’s guide STARS. Computer Modelling Group Ltd., CalgaryGoogle Scholar
- Constantin J, Peyaud JB, Vergély P, Pagel M, Cabrera J (2004) Evolution of structural fault permeability in argillaceous rocks in a polyphased tectonic context. Phys Chem Earth 29:25–41Google Scholar
- Darling T (2006) Krechba overburden review. In: Sala JIP report, April 2006Google Scholar
- Deng H, Dai Z, Jiao Z, Stauffer PH, Surdam RC (2011) Simulation of CO2 sequestration at Rock Spring uplift, Wyoming: heterogeneity and uncertainties in storage capacity, injectivity and leakage. Int J Greenh Gas Control (submitted April 2011)Google Scholar
- Fjær E, Holt RM, Horsrud P, Raaen AM, Risnes R (2008) Petroleum related rock mechanics, 2nd edn. Elsevier, AmsterdamGoogle Scholar
- Fleury M, Pironon J, Le Nindre YM, Bildstein O, Berne P, Lagneau V, Broseta D, Pichery T, Fillacier S, Lescanne M, Vidal O (2010) Evaluating sealing efficiency of caprocks for CO2 storage: an overview of the geocarbone-integrity program and results. Oil Gas Sci Technol Rev IFP 65:435–444Google Scholar
- Fokker PA, Orlic B, Van der Meer LGH, Geel CR (2011) Geomechanical modelling of surface uplift around well KB-502 at the In Salah CO2 storage site. SPE—73rd European association of geoscientists and engineers conference and exhibition 2011, pp 5380–5384Google Scholar
- Foster L, Aidoo A, Tourqui A, Castle-Smith A, Mezdour J, Taylor M (2004) Multi-disciplinary teamwork delivers outstanding Krechba Horizontal well performance. The 6th Journees Scientifiques et techniques, 20 to 22 June 2004, Algiers, AlgeriaGoogle Scholar
- Geertsma J (1973) Land subsidence above compacting oil and gas reservoirs. J Petrol Tech 25:734–744Google Scholar
- Gibson-Poole CM, Raikes S (2010) Enhanced understanding of CO2 storage at Krechba from 3D seismic. In: Proceedings of 9th annual conference on carbon capture and sequestration, Pittsburgh, Pennsylvania, May 10–13, 2010Google Scholar
- Goerke UJ, Park CH, Wang W, Singh AK, Kolditz O (2011) Numerical simulation of multiphase hydromechanical processes induced by CO2 injection into deep saline aquifers. Oil Gas Sci Technol Rev IFP 66:105–118Google Scholar
- Gunter WD, Bachu S, Benson SM (2004) The role of hydrogeological and geochemical trapping in sedimentary basins for secure geological storage for carbon dioxide. In: Baines SJ, Worden RH (eds) Geological storage of carbon dioxide. Geol Soc Spec Publ 233:129–145Google Scholar
- Gupta N, Gupka NRO, Spitznogle G, Bollinger R, Bhattacharya I (2011) Geologic storage and monitoring at American Electric Power’s Mountaineer plant—what have we learned after one year of injection. In: Proceedings of 10th annual conference on carbon capture and sequestration, Pittsburgh, Pennsylvania, May 2–5, 2011Google Scholar
- Hajdarwish AM (2006) Geologic controls of shear strength behavior of mudrocks. PhD Thesis, Kent State University, Department of Geology, Ohio, USAGoogle Scholar
- Hawkes CD, McLellan PJ, Bachu S (2005) Geomechanical factors affecting geological storage of CO2 in depleted oil and gas reservoirs. J Can Pet Technol 44:52–61Google Scholar
- Hofstee C, Seeberge F, Orlic B, Mulders F, Van Bergen F, Bisschop R (2008) The feasibility of effective and safe carbon dioxide storage in the De Lier gas field. First Break 26:53–57Google Scholar
- Iding M, Ringrose P (2010) Evaluating the impact of fractures on the performance of the In Salah CO2 storage site Int J Greenhouse Gas. Control 4:242–248Google Scholar
- IEA (2003) Acid gas injection: a study of existing operations. IAE Greenhouse Gas R&D Programme Report No PH4/18, April 2003Google Scholar
- Ingram GM, Urai JL, Naylor MA (1997) Sealing processes and top seal assessment. In: Moller-Pedersen P, Koestler AG (eds) Hydrocarbon seals: importance for exploration and production. Norwegian Petroleum Society (NPF) Special Publication 7:165–175Google Scholar
- IPCC (Intergovernmental Panel on Climate Change) (2005) In: Metz B, Davidson O, de Coninck HC, Loos M, Mayer LA (eds) Special report on carbon dioxide capture and storage. Cambridge University Press, CambridgeGoogle Scholar
- Johnson JW, Nitao JJ, Morris JP (2005) Reactive transport modeling of cap rock integrity during natural and engineered CO2 storage. In: Benson S (ed) CO2 capture project summary, vol 2. Elsevier, AmsterdamGoogle Scholar
- Kanamori H, Anderson DL (1975) Theoretical basis of some empirical relations in seismology. Bull Seism Soc Am 65:1073–1095Google Scholar
- Kanamori H, Brodsky EE (2001) The physics of earthquakes. Physics Today 54:34–40Google Scholar
- Katz DL, Tek MR (1981) Overview on underground storage of natural gas. J Pet Technol 33:943–951Google Scholar
- Li Q, Wu ZS, Li XC, Ohsumi T, Koide H (2002) Numerical simulation on crust deformation due to CO2 sequestration in deep aquifers. J Appl Mech JSCE 5:591–600Google Scholar
- Mathieson A, Midgley J, Dodds K, Wright I, Ringrose P, Saoul N (2010) CO2 sequestration monitoring and verification technologies applied at Krechba, Algeria. Lead Edge 29:216–222Google Scholar
- Mendes RA, Costa AM, Sousa Jr LC, Pereira LC (2010) Risks and mitigation problems in a CO2 injection project for a petroleum onshore field in Brazil. In: Proceedings of 44th US Rock mechanics symposium, Salt Lake City, Utah, USA, June 27–June 30, 2010Google Scholar
- Nelson CR, Evans JM, Sorensen JA, Steadman EN, Harju JA (2005) Factors affecting the potential for CO2 leakage from geologic sinks. Plains CO2 reduction (PCOR) partnership. Energy & environment research center, October 2005, University of North Dakota, North Dakota, USAGoogle Scholar
- Nguyen MT, Bemer E, Dormieux L (2011) Micromechanical modeling of carbonate geomechanical properties evolution during acid gas injection. In: Proceedings of 45th US Rock mechanics/geomechanics symposium, San Francisco, California, USA, June 27–June 30, 2011Google Scholar
- Orlic B, ter Heege J, Wassing BBT (2011) Assessing the short-term and long-term integrity of top seals in feasibility studies of geological CO2 storage. Proc 45th US Rock mechanics/geomechanics symposium, San Francisco, California, June 26–29Google Scholar
- Poston SW, Berg RR (1997) Overpressured gas reservoirs. Society of Petroleum Engineers, RichardsonGoogle Scholar
- Riddiford F, Wright I, Bishop C, Espie T, Tourqui A (2004) Monitoring geological storage: the In Salah gas CO2 storage project. In: Proceedings of 7th international conference on greenhouse gas control technologies (GHGT-7), Vancouver, Canada, September 5–9, 2004Google Scholar
- Ringrose P, Atbi M, Mason D, Espinassous M, Myhrer Ø, Iding M, Mathieson A, Wright I (2009) Plume development around well KB-502 at the In Salah CO2 storage site. First Break 27:85–89Google Scholar
- Rohmer J, Seyedi DM (2010) Coupled large scale hydromechanical modelling for caprock failure risk assessment of CO2 storage in deep saline aquifers. Oil Gas Sci Technol Rev IFP 65:503–517Google Scholar
- Rutledge J, Zhou R, Huang L, McPherson B (2009) Microseismic monitoring of CO2 injection in the Aneth Oil Field, San Juan County, Utah. The Society of Exploratory Geophysicists (SEG) Summer workshop in Banff, Canada, August 23–27, 2009Google Scholar
- Rutqvist J, Liu H-H, Vasco DW, Pan L, Kappler K, Majer E (2011a) Coupled non-isothermal, multiphase fluid flow, and geomechanical modeling of ground surface deformations and potential for induced micro-seismicity at the In Salah CO2 storage operation. Energy Procedia 4:3542–3549CrossRefGoogle Scholar
- Rutqvist J, Moridis GJ, Grover T, Silpngarmlert, Cottett TS, Holdich SA (2011b) Coupled multiphase fluid flow and wellbore stability analysis associated with gas production from oceanic hydrate-bearing sediments (in review)Google Scholar
- Seyedi DM, Ducellier A, Vu MN (2011) A two-scale model for simulating the hydromechanical behavior of faults during CO2 geological storage operation. Proc 45th US Rock mechanics/geomechanics symposium, San Francisco, California, June 26–29, 2011Google Scholar
- Sibson RH (2003) Brittle-failure controls on maximum sustainable overpressure in different tectonic stress regimes. Bull Am Assoc Petrol Geol 87:901–908Google Scholar
- Spiers CJ, Liteanu E, Zhang X (2010) Compaction creep of granular calcite under upper crustal conditions. Effects of aqueous pore fluids and supercritical CO2. Geophysical research abstracts, 12, EGU2010-6816-1, EGU General Assembly 2010Google Scholar
- Ross GD, Todd, AC, Tweedie JA, Will AGS (1982) The dissolution effects of CO2-brine systems on the permeability of U.K. and North Sea calcareous sandstones. Society of petroleum engineers, SPE 10685-MS, SPE enhanced oil recovery symposium, 4–7 April 1982, Tulsa, OklahomaGoogle Scholar
- Tran D, Nghiem L, Shrivastava V, Kohse B (2010) Study of geomechanical effects in deep aquifer CO2 storage. In: Proceedings 44th US Rock mechanics symposium, Salt Lake City, Utah, USA, June 27–June 30, 2010Google Scholar
- USEPA (U.S. Environmental Protection Agency) (1994) Determination of maximum injection pressure for class I wells, United States environmental protection agency region 5—underground injection control section regional guidance #7. EPA, Washington, DCGoogle Scholar
- Vasco DW, Ferretti A, Novali F (2008b) Reservoir monitoring and characterization using satellite geodetic data: interferometric synthetic aperture radar observations from the Krechba field, Algeria. Geophysics 73(6):WA113–WA122Google Scholar
- Verdon JP, Kendall JM, Maxwell SC (2010) A comparison of passive seismic monitoring of fracture stimulation due to water versus CO2 injection. Geophysics 75:MA1–MA7Google Scholar
- Wright I (2011) In Salah CO2 storage JIP lessons learned. 10th annual conference on carbon capture and sequestration, Pittsburgh, Pennsylvania, May 2–5, 2011Google Scholar
- Wright I, Ringrose P, Mathieson A, Eiken O (2009) An overview of active large-scale CO2 storage projects. Society of petroleum engineers, SPE 127096, SPE international conference on CO2 capture, storage, and utilization, 2–4 November 2009, San Diego, California, USAGoogle Scholar
- Yamamoto K, Takahashi K (2004) Importance of the geomechanics for the safety of CO2 geologic sequestration. In: Proceedings of 3rd Asian Rock mechanics symposium, Nov 30–Dec 2, 2004, Kyoto, JapanGoogle Scholar
- Zhou R, Huang L, Rutledge J (2010) Microseismic event location for monitoring CO2 injection using double-difference tomography. Lead Edge 29:201–214Google Scholar
- Zoback MD (2010) The potential for triggered seismicity associated with geologic sequestration of CO2 in saline aquifers. American Geophysical Union (AGU), EOS Trans. AGU, 91(52), Fall Meeting, Suppl., Abstract NH11C-01Google Scholar
- Zoback MD, Harjes HP (1997) Injection-induced earthquakes and crustal stress at 9 km depth at the KTB deep drilling site, Germany. J Geophys Res Solid Earth 102:18477–18491Google Scholar
- Zoback MD, Zoback ML (1989) Tectonic stress field of the conterminous United States. Geol Soc Am Mem 172:523–539Google Scholar