Potential leakage mechanisms and their relevance … 2-Potential leakage...Potential leakage mechanisms and their relevance to CO 2 storage site risk assessment and safe operations

SUbsurface CO2 storage - Critical Elements and Superior Strategy

Potential leakage mechanisms and their relevance to CO2 storage site risk assessment and safe operations

EditorViktoriya M. Yarushina (IFE)

Contributing authorsViktoriya M. Yarushina (IFE), Reinier van Noort (IFE),

Magnus Wangen (IFE), Elin Skurtveit (NGI), Christian Hermanrud (UiB), Helge Hellevang (UiO),

Alvar Braathen (UiO), Bjørn Kvamme (UiB)

FME SUCCESS Synthesis report Volume 2

2 Synthesis report from FME SUCCESS

Carbon Capture and Storage (CCS) is con-sidered an essential mitigation strategy in order to reduce anthropogenic CO2 emis-sions. To meet the 2°C target set in the Paris Agreement, decarbonization of the global power sector by the 2030s and the heavy industry sector beyond that is critical. CCS is currently the only option for decar-bonizing the steel, chemical and cement industries.

CCS is a proven method (e.g. at Sleipner, Snøhvit, In Salah, Weyburn, Boundary Dam, Quest). There are remaining techni-cal challenges related to upscaling, howev-er, and cost is a critical factor in large-scale deployment of CCS.

In order to stimulate relevant research, the Norwegian Research Council has estab-lished a scheme of Centers for Environ-ment-friendly Energy Research (FME) to develop expertise and promote innovation by focusing on long-term research in select-ed areas of environment-friendly energy, in-cluding CCS.

The FME SUCCESS centerThe SUCCESS center for SUbsurface CO2 storage was awarded FME status in 2009 and was formally inaugurated on 1 January 2010.

Key to public acceptance and successful deployment of CCS, the FME SUCCESS center focuses on effective and safe stor-age of CO2. To meet the regulatory require-ments for Measurement, Monitoring and Verification (MMV), the SUCCESS center seeks to provide a sound scientific base for CO2 injection, storage and monitoring in or-der to fill gaps in strategic knowledge, and to provide a system for learning and devel-opment of new expertise. Such knowledge is vital in order to ensure conformance (con-cordance between observed and predicted behavior), containment (proving storage performance in terms of security of CO2 retention) and contingency (leakage quan-tification and environmental impacts).

The FME SUCCESS center on CO2 storage

The following objectives were defined in the FME SUCCESS application:• To improve our understanding and ability to quantify reactions and flow in carbon storage.• To develop advanced modeling tools for multiphase flow and reaction.• To investigate the integrity of sealing materials, and test their retention capacity.• To improve our understanding and develop new models for the relationship between satu-

ration, flow and geomechanical response.• To improve our understanding and develop new models for geochemical and geome-

chanical interactions.• To improve our understanding and modeling tools for flow and reaction in faults and frac-

tures.• To test, calibrate and develop new monitoring techniques and instrumentation.• To improve the understanding of shallow marine processes and the ecological impact of

CO2 exposure, and develop marine monitoring methods.• To reduce risk and uncertainties in sub-surface CO2 storage.• To facilitate extensive and high-quality education on CO2 storage.

Field excursion Unis CO2 lab workshop, Svalbard 2012

Potential leakage mechanisms and their relevance to CO2 storage 3

The FME SUCCESS center on CO2 storage

One of the strengths of the FME SUCCESS center is its expertise within fundamental, theoretical research, which is internation-ally recognized; the center hence focuses on basic research, interpreting the results of field and laboratory experiments in order to predict the long-term effects of CO2 stor-age. In particular, the center has used the theoretical platform to address critical and relevant scientific issues related to CO2 storage.

Upon inauguration, the SUCCESS center was organized into six scientific work pack-ages and one educational work package

Mid-term evaluationIn 2013, the Norwegian Research Council conducted a mid-term evaluation of the FME centers. The mid-term evaluation of the SUCCESS center concluded that the center needed to undertake major changes in the organization and operational struc-ture to secure integration and industry rel-evance.

Following the recommendations of the mid-term evaluation, the SUCCESS center re-organized the scientific activities into three work packages:

• Work Package 1: Reservoir • Work Package 2: Containment• Work Package 3: Monitoring

An integration Work Package, WP0, was also established for the final two year-pe-riod of the center. WP0 aimed to test and verify new knowledge and methodology developed at the SUCCESS center in con-nection with two case studies. The Skade and Johansen formations were originally chosen as case studies. The Johansen For-mation was later replaced by the Smeaheia project case, which is the selected reservoir candidate for Norwegian full-scale demo project.

Final reportsAs part of the center’s scientific reporting, the center’s partners and board agreed that a set of reports would be written and SUCCESS fall seminar 2015 Norwegian

Petroleum Directorate (NPD)

Brummundal sandstone, SUCCESS fall seminar 2016 field excursion


summarize the major scientific findings and achievements. These reports have been re-ferred to as Long-term Deliverables (LTD).

Knowledge and lessons from the two field pilots, Snøhvit and Sleipner, have been synthesized in separate summary reports (Volume 6 and 7). Lessons from the Long-yearbyen CO2 Lab, which has been an im-portant test site for the SUCCESS center, have been and will be published in dedicat-ed summary volumes of scientific journals.

The case studies on the Smeaheia fault block (deep, confined reservoir) and the Skade Formation (shallow, saline aquifer)

in the North Sea are presented in separate reports in order to demonstrate the value of the results achieved at the SUCCESS cen-ter and associated projects, and determine how they can be applied to better quantify the storage feasibility of untested aquifers. They allow testing of the lessons and knowl-edge from the Snøhvit and Sleipner field pi-lots, and may constrain the range and use of the methods and models developed.

Long-term deliverablesThe LTD reports (5) include the SUCCESS center’s final report on the above-men-tioned deliverables. They aim to synthesize the results and findings of the SUCCESS center, and directly address the objectives of the SUCCESS center (see the figure be-low). The LTD reports cover the following topics:Storage capability (Volume 1)This report summarizes the SUCCESS center’s work on storage capability, which is the ability of a formation to safely store CO2. An important objective of this center has been to identify geological factors and the hydro-geomechanical processes that are most important for determining stor-age capability. The most important factor is whether the storage reservoir is open or closed.

Leakage risks (Volume 2)Summarizing the results from field, experi-mental and theoretical studies of potential leakage mechanisms and their relevance to CO2 storage site risk assessment, this re-port demonstrates that viscous deformation of the shales can play an important role in

their ability to keep CO2 contained and that material properties and their dynamic be-havior in response to the stress introduced by CO2 injection need to be evaluated in order to safeguard operations.

Injectivity (Volume 3)This report presents experimental and com-putational results that have enhanced our understanding of reservoir injectivity, includ-ing a basic understanding of mechanisms, quantification of the expected impact, mod-el calibration and case specific implications. A main outcome is a workflow that includes new computational tools, new geochemical and geomechanical experimental design/data and research-based advice.

Geophysical monitoring (Volume 4)The geophysical monitoring report summa-rizes the SUCCESS center’s work on rock physics related to pore pressure and satu-ration and estimating these two parameters via geophysical monitoring. By estimating their spatiotemporal distribution, we can monitor the migration of injected CO2 and determine whether the containment of stor-age complex is secure.

Marine monitoring (Volume 5)This report synthesizes relevant knowl-edge and data regarding marine monitoring methods and strategies for inorganic car-bon in the water column, based on model-ing and observational work. A cost-effective strategy for a marine monitoring program should optimize the probability of detecting a leak.

The Long term deliverables (LTD) set of reports (5) include the SUCCESS centre final reporting, and they aim at synthesizing results and findings of the SUCCESS centre and relate directly to the objectives of the SUCCESS centre

Christian Hermanrud (SUCCESS funded Prof II position, UiB) with some of his master students


Relevance of workThe collective work of the SUCCESS center addresses various groups of stakeholders and the reporting structure is relevant to dif-ferent communities. The report on storage capability is particularly relevant to storage site selection and Norwegian CO2 storage capacity estimates, based on better con-strained trapping efficiency and immobili-zation potential. The leakage risks report addresses important issues regarding safe operation of CO2 storage and risk manage-ment. The report on injectivity provides valuable knowledge on the planning of CO2 operations and reservoir utilization. Finally, there are two reports on monitoring: the re-port on geophysical monitoring addresses methods for measurement, monitoring and verification (MMV) of the subsurface; while the report on marine monitoring is particu-larly relevant to risk management and miti-gation in the event of leakage to the water column.

Future work and recommendationsCO2 storage has been successfully demon-strated at Million-tonne scale, but needs to be ramped up to Giga-tonne scale in order to achieve global emissions reductions tar-gets. A shown in the report on Large-scale Storage of CO2 on the Norwegian Shelf, there are no technical showstoppers for ramping up CO2 storage (Tangen at al., 2014). However, ramping up to Giga-tonne scale requires 1) better estimate of storage capacity, 2) better pressure management strategies, and 3) smart methods for con-trolling and optimizing CO2 injection (Nøtt-vedt, A., pers. comm. " Mission innovation workshop", 2017).

Better estimate of storage capacity requires more reliable forecasting of CO2 migration and trapping processes, with range of un-certainties. This, in turn, requires improved physics and chemistry-based understand-ing of CO2 flow and transport processes at multiple scales within heterogenous rock media.

Better pressure management strategies im-ply control on pressure limits at both near-well and reservoir scales and quantification of allowable pressurization. Consequently, better understanding of the effects of stress field, pressure history, reservoir/caprock heterogeneities, including faults and frac-tures, is needed.

Smart methods for controlling and optimiz-ing CO2 injection include effective control and handling of transmissivity, near-well geochemical processes, formation dam-age, etc. Well stimulation and next-gener-ation well technologies need to be demon-strated to enable large-scale CO2 injection.Future advances in CO2 storage will likely occur at the interface between industry and academia and be coupled to the execution of ramp-up CO2 storage projects.

Arvid Nøttvedt, Center Manager FME SUCCESS

Ivar Aavatsmark,Scientific Leader FME SUCCESS

Per Aagaard,Scientific Leader FME SUCCESS 2010-2014

Alvar Braathen,Scientific Leader FME SUCCESS 2014-2018


Potential leakage mechanisms and their relevance to CO2 storage

site risk assessment and safe operations

EditorViktoriya M. Yarushina (IFE)

Contributing authorsViktoriya M. Yarushina1, Reinier van Noort1, Magnus Wangen1, Elin Skurtveit2,

Christian Hermanrud3, Helge Hellevang4, Alvar Braathen4, Bjørn Kvamme3

1 Institute for Energy Technology (IFE)2 Norwegian Geotechnical Institute (NGI)

3 University of Bergen (UiB)4 Univeristy of Oslo (UiO)


Photos and illustrations Marit Hommedal, Ole Jørgen Bratland (Statoil), Helge Hansen, Statoil), Harald Pettersen, (Statoil), Daniel Byeers (ISGS),

Tor de Lange (UiB), Olav Olsen (Aftenposten) and photos/illustration from all of the FME SUCCESS Partners.Idea, layout/design: Per Gunnar Lunde and Charlotte Gannefors Krafft, CMR

Table of contents

The FME SUCCESS center on CO2 storage 2

Executive Summary 8

Introduction 9

Leakage mechanisms 12Seal capacity of caprock 12Erosion of well completing cement and corrosion of pipelines as possible leakage mechanisms 16Flow through existing faults and fracture networks 17Induced tensile or shear failure of caprock 19Sand injections and risk of leakage 21Other vertical focused fluid flow structures 21

Conclusion and

recommendations 26

References 28


Executive Summary

Geological carbon capture and storage (CCS) is an emerging technology that has been designed to mitigate the impact of hu-man activities on global warming. In order to achieve public acceptance of this tech-nology, one needs to guarantee its safety for the environment and the population. Potential leakage of CO2 from underground storage sites is one of the risks associated with CCS.

This report summarizes the results from field, experimental and theoretical stud-ies of potential leakage mechanisms and their relevance to CO2 storage site risk as-sessment and safe operations. We evalu-ated both the impact of pre-existing leakage pathways (e.g. faults and fracture networks, abandoned and former wells) and the possi-

bility of triggering the formation of new leak-age structures via hydraulic fracturing or porosity wave propagation during injection. Critical engineering parameters that might lead to reactivation or generation of a new leakage pathway and pressure limits were identified. A new mechanism of CO2 plume movement, which produces localized fluid conduits is suggested. The rates of move-ment and likely breakthrough times are dis-cussed. The flow properties of sealing rocks that are common to the Norwegian offshore reservoirs are summarized based on the results of laboratory measurements of cap-illary entry pressure and permeability. Our laboratory measurements show that shales exhibit a strong coupling between fluid flow and geomechanics. We show that viscous deformation of shales can play an impor-

tant role in their ability to contain CO2. Our results indicate that fluid flow mechanisms through shales are not fully understood and require further experimental and theoretical studies.

The report addresses potential leakage pathways as a knowledge support for site risk assessment. Based on our results, we have made suggestions for further devel-opment of the International Standards on Carbon Dioxide Capture, Transportation and Geological Storage such as ISO/DIS 27914.

"This report summarizes the results from field, experimental and theoretical studies of potential leakage mechanisms and their relevance to CO2 storage site risk assessment and safe operations."


Introduction

The role of CCS as an effective climate miti-gation technology depends on our ability to securely store large volumes of carbon dioxide (CO2) in geological formations for thousands of years. We know that oil and gas have been contained in underground reservoirs for much longer periods of time. Current experience from large-scale injec-tion cases such as Sleipner, Snøhvit, and the CO2-EOR projects in the USA, has confirmed that CO2 can also be stored se-curely. Various trapping mechanisms like

structural and stratigraphic barriers or dis-solution and capillary forces work together in the subsurface to keep it from escaping back into the atmosphere. Nevertheless, as the CO2 injected into geological reser-voirs is lighter than the brine present in the rocks at pressure and temperature condi-tions that are typical of most reservoirs, it may leak through natural or man-made pathways, causing effects on e.g. drinking water and marine ecosystems. The total potential leakage and its extent depends

on a number of parameters, including the injection rate, the proximity of the injection point to the leakage pathway, permeability of the pathway, the depth and thickness of the CO2 storage reservoir, etc. Adequate geological, geophysical and geomechani-cal assessment of a potential CO2 injection site is thus the key to safe operations. EU Directive 2009/31/EC requires that the po-tential for leakage from a storage complex must be characterized during assessment of a future site.

Figure 1: Outline of possible leakage pathways for CO2


A number of research projects funded by the EU and other national schemes, includ-ing SACS, CO2CARE, and ECO2 have worked on identifying and characterizing various leakage mechanisms (see Fig. 1).

It is commonly observed that leakages can occur in abandoned or existing wells if they are poorly cemented or have faults in the casing or cement due to e.g. material cor-rosion and cement alteration (Birkholzer et al., 2011; Humez et al., 2011; Kutchko et al., 2007; Newell and Carey, 2013; Nordbotten et al., 2009). CO2 may also leak due to a lack of a geological seal as a result of re-activation of faults and associated fracture networks (Alikarami et al., 2013; Jeanne et al., 2013; Pruess, 2005; Rutqvist et al., 2007), induced tensile or shear failure of the caprock (Gor et al., 2013; Huang et al., 2015; Rutqvist et al., 2008; Wangen, 2011), diffusion through pore networks (Hou et al., 2012; Huq et al., 2017) or through seal by-pass systems (Cartwright et al., 2007; Cev-atoglu et al., 2015; Hermanrud et al., 2009; Räss et al., 2014; Tasianas et al., 2016). Leakage in the pore network in shales is

the least problematic of these, as the low permeability and relative permeability of the seal would imply a low velocity for the leak-ing fluids – so low that in most cases the CO2 would be dissolved or react by form-ing mineral phases before it reaches the surface. Leakages through faults, fracture networks and wells can be rapid. Leakages in sand injections or other vertical focused fluid flow structures is also a concern where such structures are present.

The most common caprocks on reservoirs targeted for CO2 storage are shales. The properties of shale have been studied in great detail during the past decade, given that shale is also an important source and a reservoir rock in unconventional oil and gas production (e.g. Cassini et al., 2017; Chang and Zoback, 2009; Kwon et al., 2004; Makhnenko et al., 2017; Mondol et al., 2007; Rybacki et al., 2015).

Shale permeability is a key parameter in both of these geo-engineering systems. There has hence been a strong focus on the properties and geochemical/geomechani-

cal interactions that influence shale perme-ability. Burial and cementation can have a significant impact on shale permeability (e.g. Milliken et al., 2012). For example, a high calcite cement content may lead to lower permeabilities, while gas generation due to hydrocarbon maturation (if a shale reaches a temperature that is high enough) can re-open pore pathways, or even create new fractures, hence increasing permeabil-ity (e.g. Prince et al., 2011). The composi-tion of shales does not always impact on its fluid flow properties. Specifically, Al Ismail and Zoback (2016) report a higher perme-ability for shales with a higher total organic content (TOC), but only for shales in which the porosity is mainly located in solid organ-ics and is well connected. In their study of shales consisting mainly of calcite and clay in different ratios, the mineral content did not show a strong correlation with perme-ability, but a higher clay content resulted in a stronger dependence of permeability on effective confining stress.

In contrast, Metwally and Chesnokov (2011) report an increase in permeability with in-creasing quartz (and pyrite) content, a de-cline in permeability with increasing calcite and clay (especially illite) content, and no clear correlation with TOC. Ghanizadeh et al. (2013) studied different shales with vary-ing TOC, quartz, calcite and clay content and reported that the shale composition had no clear effect on the permeability, or on the stress-dependence of the perme-ability. These apparent discrepancies likely reflect the fact that the very low permeability of shales is a result of a complex interplay of many different factors, where small differ-ences can induce large effects.

Shales are strongly anisotropic, with clear

"There has hence been a strong focus on the properties and geochemical/geomechanical interactions that influence shale permeability."

"... very low permeability of shales is a result of a complex interplay of many different factors, where small differences can induce large effects."


layering due to alignment of clay minerals during deposition, which can be further en-hanced by burial and diagenesis. With re-gards to fluid flow, this anisotropy results in a higher permeability for flow parallel to the layering. Metwally and Sondergeld (2011) report an anisotropy coefficient (the ratio of layer-parallel over layer-perpendicular permeability) of roughly 3–5 for tight gas shales. Bhandari et al. (2015) report a coef-ficient of ~40 for Barnett shales. The effec-tive confining stress was not found to influ-ence this value. Al Ismail et al. (2014) report anisotropy coefficients of two and three or-ders of magnitude, using He and CO2 re-spectively, measured on Eagle Ford shales. The stronger anisotropy observed with CO2 was related to the swelling of clay minerals. Ghanizadeh et al. (2014) report anisotropy coefficients in excess of one order of mag-nitude, with greater anisotropy for samples with higher clay contents.

Another important factor that influences shale flow properties is the effective stress, as shale compacts significantly as a result of both elastoplastic and time-dependent creep mechanisms. Such compaction leads to a decrease in pore or crack aperture, reducing the porosity and pore connectiv-ity, and constricting flow. Conversely, when the effective stress decreases, elastic relaxation can lead to a widening of flow

pathways, thus enhancing the permeability. As changes in the pore fluid pressure lead to changes in the effective stress, these changes can have significant effects on the shale permeability, and hence seal integrity. Accurate measurements of the mechanical properties of shales are an important com-ponent for determining how shale caprocks will react to changing fluid pressures. Ex-perimental studies have demonstrated that time-dependent creep can be an important long-term mechanism, in addition to time-independent elastic and plastic deforma-tion (Chang and Zoback, 2009; Räss et al., 2017; Sone and Zoback, 2014).

Accordingly, permeability measurements performed on shales under increasing ef-fective confining stress have indeed shown permeability to be dependent on effective confining stress as a result of compaction (Bhandari et al., 2015; Chalmers et al., 2012; Dong et al., 2010; Ghanizadeh et al., 2014; Heller et al., 2014; Zhang et al., 2015; Zhang et al., 2016). Dong et al. (2010) and Zhang et al. (2015) report exponential cor-relations between shale permeability and effective confining stress, with exponential coefficients in the range of 0.005–0.123. As noted by Chalmers et al. (2012) and Ghanizadeh et al. (2014), mineralogy is an important factor that controls both per-meability and the confining pressure de-

pendence of permeability. Gutierrez et al. (2015) report continuous measurements of the stress-dependent permeability of a compacting Mancos shale sample. They show a two order of magnitude decrease in permeability when the effective mean stress increases from ~2 to ~60 MPa. This is highly consistent with the tenfold decrease in permeability, which Metwally and Sondergeld (2011) report for a 30 MPa increase in effective stress. In addition to this direct correlation between stress and permeability, creep effects on permeability are reported in the work of Chhatre et al. (2015). However, more measurements are required here in order to determine the ef-fects of shale and fluid properties on creep.

In this report, we summarize the work that has been conducted at the FME SUCCESS center on understanding potential leakage processes. The next few chapters explain the main results and findings from field, lab-oratory and theoretical studies that describe major leakage mechanisms, critical factors for their assessment, and challenges asso-ciated with predictive modelling. The report concludes by summarizing our main con-clusions and recommendations for future research and practice.

"Accurate measurements of the mechanical properties of shales are an important component for determining how shale caprocks will react to changing fluid pressures."

"In this report, we summarize the work that has been conducted at the FME SUCCESS center on understanding potential leakage processes."


Leakage mechanisms

When choosing a suitable reservoir, larger structural and partly stratigraphic traps are commonly targeted in hydrocarbon explora-tion and CO2 storage alike. They represent either a lateral change in the depositional system, for instance a delta sandstone shaling out laterally into tight mudstone, or unconformities. The latter can be linked to events of non-deposition or erosion of a possible reservoir/storage formation unit, with overlying units offering a seal. All of these considerations are described in de-tail in textbooks (e.g. Bjørlykke, 2015) and a wealth of articles. Stratigraphic traps are common in sedimentary systems, and rep-resent a challenge in forecasting reservoir behavior and storage efficiency, as sedi-mentary deposits will always offer a degree of heterogeneity. Heterogeneity arises from depositional processes of sedimentation, which obey physical parameters that con-trol body shapes and intrinsic architecture, and thereby related flow properties, on a scale not resolvable with seismic imaging. Reservoir descriptions for a given storage formation will therefore be based on large- scale observations from seismic and drill hole cores/logs, used to identify generic models developed for a given sedimentary environment. Such intrinsic heterogeneities are basically impossible to identify, even in mature fields with a comprehensive history match.

Seal capacity of caprockPhysical trapping of CO2 in the reservoir will occur along the top surface of a given storage formation, underneath the top seal and structural traps. There, the bypassing, mobile CO2 plume will fill small traps, with efficient CO2 capture. In most cases, the numerous intervening stratigraphic traps limit the rate to orders of magnitude less than the rate of seepage/leakage from the

storage reservoir. In order to enter the non-fractured rock, CO2 will have to overcome the capillary entry pressure (pc) in intact water–wet shales, which strongly depends on a pore throat radius (r), surface tension between CO2 and H2O (σ) and the contact angle between H2O and the shale surface in a CO2 atmosphere (θ ):

(1)

Given that typical surface tension is about t 30-40 mN/m (Bachu and Bennion, 2009), the typical contact angle is about 17–45° for supercritical CO2 (Chen et al., 2016) and a dominant pore throat radius in the Draupne shale, which is representative of the shale barriers in the North Sea, is between 2 and 10 nm (Allen, 2014), the theoretically-expected entry pressures based on equa-tion (1) would be between 6 and 30 MPa. Unconsolidated and silty mudstones may have lower entry pressures.

Experimentally-measured CO2 capillary en-try pressure is in the range of 0.1–5 MPa for a selection of what was assumed be intact mudrock samples (Hildenbrand et al., 2004) and 3.5–4.3 MPa for shale from the Draupne Formation in the Troll East

area (Skurtveit et al., 2012). These values are generally lower than expected for the intact samples based on equation (1), and it may be that connected microfractures are responsible for the low values in some of the samples. For instance, indicates a connected pore throat radius of ~600 nm, which is not realistic for any compacted mu-drocks or shales. Harrington et al. (2009) measured the capillary entry pressure for the Nordland shales (the cap for the Utsira CO2 storage site at Sleipner) and came up with a gas breakthrough pressure of 3.1 MPa using nitrogen.

In this experiment, the flow transport mech-anism is suggested as being dominated by flow in pressure-induced pathways with a limited degree of pore water displacement. Angeli et al. (2009) and Skurtveit et al. (2012) recognized the CO2 breakthrough by a marked dilation of the test sample combined with the localized flow, support-ing the idea that microfractures/pores link up to form distinct pathways for CO2 after overcoming the capillary entry pressure in low permeable seal units. The results indicate that advective flow in low perme-able seal units is linked to geomechanical

"Stratigraphic traps are common in sedimentary systems, and represent a challenge in forecasting reservoir behavior and storage efficiency, as sedimentary deposits will always offer a degree of heterogeneity."

2 coscp

rσ θ

=


processes such as microfracturing and/or pore dilation. Results from field scale gas injection testing on bentonite in Äspö Hard Rock Laboratory (Cuss et al., 2014) confirm dilatancy pathways as an important mecha-nism for gas migration also at field scale.

Flow of CO2 through intact caprock greatly depends on the permeability of the rocks, which in turn is very sensitive to stress and pressure changes (David et al., 1994; Dong et al., 2010). Numerical simulations of the hydromechanical response of the reservoir associated with injection of CO2 indicate that injection can reduce the mean effective stress in the reservoir (Rutqvist and Tsang, 2002; Wangen et al., 2016). This will also lead to an increase in pore pressure in the parts of the caprock adjacent to reservoir, causing pore dilation and changes in the permeability of the caprock.

Furthermore, pore dilation will also increase the pore radii and thus reduce the capil-lary entry pressures, facilitating leakage. Shales, poorly lithified sandstones and sand have compressibilities that are very sensitive to stress changes (Skurtveit et al., 2013; Yarushina et al., 2013). This sensitivi-

"... the flow transport mechanism is sug-gested as being dominated by flow in pressure-induced pathways with a limited degree of pore water displacement."

ty is even stronger at low effective confining pressures that are expected to develop dur-ing prolonged injection. In such stress-sen-sitive formations, CO2 flux might increase by orders of magnitude with increasing fluid pressure. We performed permeability measurements in order to evaluate the po-tential CO2 flux through the primary caprock for different stress and pressure conditions, assuming no faulting or fracturing of the caprock and reservoir. First, the effective Darcy permeability for CO2 has been stud-ied in the same experiments as for CO2 en-try pressure. High confining pressure was applied in order to avoid hydrofracturing of the sample and Darcy flow was imposed upon shale samples in the experiments by Hildenbrand et al. (2004), where a slight de-cline in effective CO2 permeability (range of 10-18–10-24 m2) was measured compared to the water permeability (range of 10-19–10-21 m2).

Brine permeability in the range of 3–10×10-19 m2 was measured for the Nord-land shales (Harrington et al., 2009). Skurt-veit et al. (2012) measured effective CO2 permeability in the order of 10-21 m2, within the same order of magnitude as for brine

permeability, and the effective CO2 perme-ability was found to depend on volumetric dilation in the sample.

The effect of stresses on changing per-meability was studied using transient pulse and constant flow techniques on an intact Rurikfjellet shale sample under in situ confining pressure conditions, using argon, supercritical CO2 and water as the metering fluids (van Noort and Yarushina, 2016). The permeabilities measured were plotted against confining pressure in Fig. 2 (see next page). Our initial results, using argon, show a decline in permeability with increasing confining pressure that is largely recovered when the confining pressure is reduced again. We also observe a limited time-dependent decline in permeability at constant pressure.

However, breakthrough effects were ob-served during two subsequent flow-through measurements using CO2, where the per-meability suddenly increased strongly (by at least 1.5 orders of magnitude). Once this flow was stopped, the permeability quickly recovered to near-original values again, but

Longyearbyen CO2 lab, Svalbard


Figure 2: a) All individual permeability measurements performed on the shale core plug. b) First series of measurements, performed at room temperature using argon. The permeability shows a relatively strong dependence on the effective confining pressure. Fur-thermore, some time-dependent healing of the core is observed at a lower confining pressure (@1). c) Argon and CO2 permeabilities at room temperature and at 40°C. During a flow-through test using CO2, there was breakthrough, then the permeability was temporar-ily greatly reduced (@2). d) Permeabilities measured using water and CO2. The water permeability is considerably lower than the CO2 and argon permeabilities. Furthermore, the water permeability shows some time-dependence. The decline in permeability measured using water is not recovered when the effective confining pressure is declined (@3). Flow-through tests were performed again during the CO2 measurements, and similar breakthroughs were observed (@4 and @5). After each breakthrough, the permeability was greatly enhanced. While the permeability shows a recovery to lower values with time, it does not return to the pre-breakthrough values. Finally, during the water permeability measurements, at a constant effective confining pressure of 15 MPa, the permeability declined by a factor two over 13 days (@6). This time-dependent effect is also shown in e).


remained elevated by about 0.5 orders of magnitude for at least several days after-wards. Note that the breakthrough effects observed here do not appear to be related to capillary entry pressure, as flow occurred at lower upstream pressures during pulse measurements. Subsequent series of mea-surements showed a reduced dependence of permeability on confining pressure com-pared to the first series, most likely related to drying effects.

A series of measurements with H2O as the metering fluid showed a more than 1.5 orders of magnitude lower permeability than measurements carried out using ar-gon or CO2. Furthermore, the declines in permeability that were observed when the confining pressure was increased in the presence of water were maintained when the confining pressure was subsequently released, suggesting a permanent effect. At a constant confining pressure of 15 MPa, time-dependent creep resulted in a decline in permeability by a factor two over 13 days (see Fig. 2e). After exposure to water, the permeability of the sample to CO2 was re-duced. Furthermore, breakthroughs were once again observed during constant flow tests, leading to permeability increases of up to 3 orders of magnitude. The observed effect of confining pressure on shale per-meability is the result of a combination of elastic, plastic and viscous deformation of these rocks under confinement (Dong et al., 2010; Sone and Zoback, 2014; van Noort and Yarushina, 2016; Zhang et al., 2016).

More interesting, however, are the addition-al permeability effects of the presence of water, and of CO2 flow during flow-through tests. As argued by Ghanizadeh et al. (2014), the water permeability of shales is probably lower than the permeability to oth-er fluids, due to the flow-inhibiting effects of water adsorption on mineral surfaces. Like-wise, water left behind in the pore network

may inhibit the flow of CO2 by blocking pore throats. During flow-through tests, evapora-tion and/or displacement of this water open the pore throats, thus enhancing perme-ability temporarily. Alternatively, these ef-fects could be the result of the swelling and shrinkage of clay minerals in the shale.

According to Schaef et al. (2012) and Ilton et al. (2012), the potential effects of clay mineral swelling and shrinkage due to ex-posure to (dry) supercritical CO2 on caprock permeability could have significant conse-quences for shale caprocks exposed to CO2 during storage in reservoirs, although this is disputed by e.g. Busch et al. (2016). It is therefore important to better understand these effects and their interplay with con-fining pressure and shale mineral composi-tion. For this purpose, additional permeabil-ity measurements have been performed on confined tablets of pressed smectite clay. Our results show that under static confine-ment (a tablet pressed into a thick steel ring that does not allow this tablet to change its dimensions), exposure to water immediate-ly results in a strong decline in permeability, as clay hydration and swelling block any further flow of CO2.

Under such conditions, if exposure to dry CO2 leads to shrinkage, it would quickly result in the formation of (micro) cracks and fluid pathways, and hence a strong increase in permeability. However, in our

experiments, long-term exposure to CO2 did not result in any such enhancement of permeability. Under dynamic confinement (i.e. a constant effective confining pressure (7.5 MPa) applied through a flexible jacket, allowing the clay tablet to change its outer dimensions), sufficient wetting of the clay tablets likewise resulted in a sample per-meability that declined over a period of sev-eral days. However, a higher overall sample permeability was maintained (~1-2×10-20)m2). Further tests are needed to determine the effects of microstructures and mineral-ogy (clay content) and whether subsequent long-term exposure to CO2 flow can result in an increase in permeability.

The results of experiments depend to a large degree on the sample size and whether or not a sample can be considered as representative. In order to constrain the long-term behavior of a potential CO2 storage site and the rates of fluid seepage through real geological formations on a geological timescale, we studied the history of past fluid communication in the reservoir and caprock shales from the Longyearbyen site (Huq et al., 2017). The Longyearbyen CO2 Lab of Svalbard was established in order to evaluate the potential for geologi-cal carbon sequestration on Spitsbergen. Several monitoring wells were drilled in and around the planned CO2 injection site. The targeted reservoir is a sandstone layer of the De Geerdalen Formation located at ~700–1000 m depth. A thick shale layer, just above the reservoir, was identified as a potential caprock. Near the surface, a thick permafrost layer provided another potential seal (Braathen et al., 2012). We used two tools to investigate fluid commu-nication within and between these entities: 87Sr/86Sr of formation waters extracted from cores using the residual salt analysis (RSA) method, and the δ13C of gases, prin-cipally methane and CO2, degassed from core samples (Huq et al., 2017). The Sr RSA data suggest good lateral communica-tion on a geological timescale. However, there is a distinct barrier to ver-

"... we studied the history of past fluid communication in the reservoir and caprock shales from the Longyearbyen site..."

"A series of measurements with H2O as the metering fluid showed a more than 1.5 orders of magnitude lower permeability than measurements carried out using argon or CO2. "


tical communication within the middle res-ervoir, corresponding to thin but presum-ably laterally extensive (>1.5 km) lagoonal mudrocks. The aquifer is also interpreted as having some degree of vertical internal communication on a geological timescale. The caprock shales show an ongoing mix-ing process of the pore fluid between the caprock and the underlying reservoir. Diffu-sion and flow modelling based on the per-meability values indicated above estimated that the fluid communication across the caprock is extremely slow. The estimated travel time based on these mechanisms is much longer than the age of the rocks (Huq et al., 2017). Thus, at some time in the past, shale fluid transport properties were en-hanced by e.g. the formation of temporary pressure escape features (fractures or fluid conduits) during deep burial and uplift, or during glaciation cycles.

Nonetheless, the smooth compositional gradient in the caprock indicates that fluid mixing has subsequently taken place slow-ly, dominated by diffusion. This interpreta-tion is supported by the gas isotope data, where systematic variations in gas δ13C (-50‰ to -32‰) values also indicate slow and incomplete diffusional fluid mixing. These are positive indicators for caprock efficacy within the shorter timescale of CCS operations.

A substantial reduction in vertical perme-ability in the hydrate forming sections has been observed for reservoirs with condi-tions in which CO2 hydrate can form (Je-mai et al., 2017). Upward CO2 migration is substantially slowed down in these regions, but also horizontal spreading is reduced in these hydrate-forming layers (Jemai et al., 2017; Qorbani et al., 2017; Qorbani et al., 2016).

Erosion of well completing cement and corrosion of pipelines as possible leakage mechanismsInjection wells for carbon dioxide are made of iron casings supported by cement. Ce-ment primarily contains calcium oxide, which transforms into calcium carbonate in reactions involving carbon dioxide. Cement also contains other clay materials similar to calcium carbonate in terms of water struc-turing effects and adsorption thermody-namics. Moreover, iron transforms into iron oxide, due to sodium chloride and the acidic environment.

Pipelines are rusty even before they are installed. While rust generally consists of a mixture of iron oxide, FeO, hematite, Fe2O2 and magnetite Fe3O4 the most important of these is hematite, and it may therefore be used in studies as a model for rust. Aban-doned oil and gas wells in the vicinity of an underground aquifer, plugged with cement, as well as the injection well itself, are sub-ject to further corrosion of the rusty metal surfaces. Erosion of the cement happens due to an acidic water environment contain-ing dissolved (and dissociated) CO2 as well as gas bubbles that are incorporated due to the local hydrodynamics.

There are two reasons for the existence of space in between the rusty pipeline and the cement. First, it is geometrically impossible to achieve total beneficial direct contact between the surfaces of rust and cement due to the distribution of charges on the two surfaces. Second, and even more im-portantly, are the exothermic reactions dur-ing drying of the cement which evaporate water and lead to channel creation. At the SUCCESS center, we have studied erosion and corrosion in an acidic water environ-ment and CO2. We applied the principles

" At the SUCCESS center, we have studied erosion and corrosion in an acidic water environment and CO2. We applied the principles of quantum mechanics in order to characterize charge distributions in cement and rust, and molecular dynamics simulations to evaluate adsorption structures, composition and thermodynamic properties"

of quantum mechanics in order to charac-terize charge distributions in cement and rust, and molecular dynamics simulations to evaluate adsorption structures, composi-tion and thermodynamic properties (Olsen et al., submitted; Olsen et al., 2016). These characteristics are long-term effects, and adsorption is highly non-uniform. Large simulation systems are therefore needed in order to develop a basis for further reaction studies.

In addition to the ions related to CO2 dis-sociation, there are natural ions in the groundwater which are transported into the space between rust and cement. The injec-tion gas can contain monoethylene glycol (MEG), which is sometimes used as a hy-drate inhibitor and corrosion inhibitor during transport. It might also contain some traces of components used in CO2 separation. Our study was limited to the presence of MEG and the salinity of average seawater, as represented by positive sodium ion and negative chlorine ions. Bicarbonate and hy-dronium were balanced according to typical dissociation in relevant conditions.

We found that adsorbed MEG displaced adsorbed water, thus reducing the Coulomb screening of the surfaces, which yielded free energy minima of hydronium and bi-carbonate that were closer to the surfaces (Olsen et al., submitted; Olsen et al., 2016). Ripples in the free energy profiles, originat-ing from the layers of water, were altered due to adsorbed MEG, as adsorbed MEG resulted in free energy minima closer to the surfaces, hence providing more water layers for the ions to traverse. Competition between adsorbed MEG and adsorbed ions


for adsorption sites played a significant role in the changes that appeared in the free energy profiles, due to the adsorbed MEG. The remains of amines from CO2 separa-tion from hydrocarbons is expected to have similar effects on the water and ion charac-teristics on the two surfaces. Although hy-dronium and bicarbonate appeared closer to the surfaces, and as such may be more exposed to reaction sites, the transport as-pects of reaction mechanisms might be af-fected in ways that could slow down erosion and corrosion rates.

Ab initio studies of adsorption structures similar to what was found in this study are needed for greater clarification of the con-sequences of the adsorbed structuring. Ab initio modelling will also provide some of the parameters needed in reaction kinetic modeling of corrosion kinetics, like activa-tion energies for various reactions involved.

Flow through existing faults and fracture networksFaults and associated fracture networks can significantly influence the flow of groundwater, hydrocarbons or CO2. They can act either as barriers to fluid flow or as conduits for fluid circulation. The presence of an intensive fracture network around faults, e.g. in low-porosity formations, can enhance fault permeability, while deforma-tion bands in the fault damage zone might behave as seals for fluid flow. Most studies emphasize the strong influence of fault-zone architecture on fault-zone hydraulic properties. Field studies of natural CO2 reservoirs, which are widespread in sedi-mentary basins worldwide, show that the fracture networks developed in the damage

zone of the faults influence the fluid circula-tion in the subsurface, making them a sig-nificant factor in CO2 storage site assess-ment (Alikarami et al., 2013; Busch et al., 2014; Dockrill and Shipton, 2010). Natural CO2 reservoirs can originate from a num-ber of sources, such as mantle degassing, carbonate rock metamorphism or the deg-radation of organic matter. In many cases evidence of paleo-flow of CO2 consists of the bleaching of sandstones or mineraliza-tion of carbonates.

The data that are most suitable for studies of CO2 leakage in the subsurface come from the oil and gas industry, as the processes that result in leakage of CO2 and hydrocar-bons are the same, although they operate at different timescales. We have investigated the controls of hydrocarbon–water contacts in many Norwegian oil and gas fields, and considered the reasons for leakage from a set of dry structures presented in several MSc theses. We came to the conclusion that leakages have taken place along faults or at fault intersections from many struc-tures, as documented by Hermanrud et al. (2014) for the Barents Sea. The leakages were probably comparatively rapid and short-lived, and the overpressures were not been drained from leaky and overpres-sured compartments. Overburden bright spots are often seen along and/or above

"The presence of an intensive fracture network around faults, e.g. in low-porosity formations, can enhance fault permeability, while deformation bands in the fault damage zone might behave as seals for fluid flow."

leaky faults (Hermanrud and Georgescu, 2014; Simmenes et al., 2017; Simmenes et al., 2015). This means that faults and fault intersections especially could be leakage pathways for injected CO2. Geophysical monitoring is especially important when the CO2 injected meets such features, and a back-up plan needs to be in place to handle leakage though faults. The leakage that we have identified appears to be controlled by shear failure.

At the SUCCESS center, we investigated the fracture corridors of the mid to late Ju-rassic Entrada and the Curtis Formations of the northern Paradox Basin, Utah, which are characterized by discoloration (bleach-ing) due to oxide removal by circulating CO2 and/or hydrocarbon-charged fluids (Ogata et al., 2012). The structures analyzed are located in the footwall of a km-scale, steep normal fault with displacement values in the order of hundreds of meters. The fracture corridors trend roughly perpendicularly and subordinately parallel to the main fault di-rection, and define a systematic network. The fracture corridors pinch and fringe out laterally and vertically into single, continu-ous fractures, following the axial zones of open fold systems related to the evolution of the main fault.

"We have investigated the controls of hydrocarbon–water contacts in many Norwegian oil and gas fields, and considered the reasons for leakage from a set of dry structures presented in several MSc theses. "


Based on the data presented, using the bleaching as a proxy for past fluid migra-tion, we hypothesize that such fracture cor-ridors, which connect localized reservoirs at different stratigraphic levels up towards the surface, represent preferred fluid migra-tion pathways rather than the main faults. The corridors are found (Ogata et al., 2012) (i) in the damage zone of faults; (ii) at fault tips where displacement disappears; and (iii) along the crest of gentle folds oriented perpendicular to faults, reflecting rebound in the footwall proximal to the fault. We also investigated the distribution of deformation features (structures) such as fractures and deformation bands in the Navajo and the Entrada sandstones in the fault core and damage zones of two faults in two locations in southeastern and central Utah – Cache Valley and the San Rafael Swell (Alikarami et al., 2013; Skurtveit et al., 2015).

These two locations had different degrees of calcite cementation and hence are of in-terest in order to study the mechanical and petrophysical properties for identification of the impact of cementation. We have per-formed in situ measurements using a Tiny-Perm II and a Schmidt hammer in order to examine the distribution of permeability and strength/elasticity of rocks within the dam-age zone of these faults.

Microstructural characterization of defor-mation structures for the fault in the San Rafael Swell shows a complex interaction between deformation bands, fractures and the calcite precipitation. The development of deformation bands and their link to frac-turing affected the flow field around the fault from a fault baffle to a conduit (Skurtveit et al., 2015). We have started to look at con-

trolling parameters for fractures within and around faults, but further work is required in order to arrive at a conclusion.

Sediments, faults, fractures and other de-formational features in the Basque Can-tabrian Basin (BCB) were studied as an analogue for an overburden sequence at a leaking CO2 storage site. Extensive seeps, and also some accumulation of hydrocar-bons, make this basin an excellent location for studying the dynamics of large-scale fluid leakage through sediments with a vari-able lithology. We have focused on cement-ed fractures, where the vein minerals hold an extensive record of the fluid flow history through the sediments. Stable isotope and fluid inclusion data, coupled with detailed field observations and petrographic work, have given us a fairly detailed picture of the importance of fractures in a larger migration context. Textures, isotopes and fluid inclu-sion data prove the repeated reactivation of fractures, transmitting large volumes of highly-contrasting fluids at various times. Some samples contain evidence of more than 20 vein mineral generations, where water, gas and oil have migrated.

An interesting comparison between the BCB and the Longyearbyen Basin (LYB) is that the LYB core samples show very few mineralized fractures, while the fre-quency of cemented fractures is very high in the BCB. This suggests that the vertical fluid migration in the LYB has been minimal, which is consistent with the compartment observations reported by Huq et al. (2017).

These results are supported by laboratory tests of fracture permeability on a single shale plug from the Rurikfjellet formation at

the Longyearbyen CO2 storage pilot site in Svalbard, with and without a crack parallel to its axis (and parallel to its bedding) (van Noort and Yarushina, 2016). The measure-ments (which were performed using water as the metering fluid) show that at very low effective confining pressure (2.4 MPa), the permeability of the fractured sample is somewhat enhanced, relative to the same sample before fracturing. However, at high-er effective confining pressures (17.4 MPa), the permeability of the fractured sample is lower than the permeability before fractur-ing. This demonstrates that the pressure and time-dependent compaction of shale caprock in the presence of water can suc-cessfully seal a fracture under in situ PT conditions. In another series of experi-ments, fracture flow properties have been investigated by evaluating the changes in flow and geophysical responses (sonic ve-locity and resistivity) of a naturally-fractured tight sandstone core plug from the De Geer-dalen Formation at the Longyearbyen CO2 storage pilot site (Nooraiepour et al., sub-mitted).

Fracture flow experiments have also dem-onstrated the stress-dependent nature of fracture permeability and pointed out the role of aperture and asperities. The mea-sured fracture permeability is in the order of (0.01–0.1mD) for effective confining pressures ranging from 5–24 MPa, and the permeability seems to be slightly higher for CO2 than for water. Geophysical monitor-ing of CO2 fracture flow has documented a change in acoustic velocity and electrical resistivity when the contribution of matrix flow is minimal. Electrical resistivity is more sensitive to the presence of CO2 in the frac-ture than the sonic velocity measurements.

"We have focused on cemented fractures, where the vein minerals hold an extensive record of the fluid flow history through the sediments. Stable isotope and fluid inclusion data, coupled with detailed field observations and petrographic work, have given us a fairly detailed picture of the importance of fractures in a larger migration context."


which affects the stress state of the rocks. Two common failure scenarios related to the increase of fluid pressure are described in Fig. 3. The most critical shear failure is for cohesionless material, like faults and frac-tures, whereas tensile failure is the most critical for intact material.

There are three main approaches for de-termining fracture pressure: (a) theoreti-cal methods; (b) use of formation well-test data; and (c) analysis of injection data during injection (Bohloli et al., 2017). At the SUCCESS center and in the associ-ated PROTECT project, a direct shear box rig was developed that can determine frictional properties (see Fig. 4) and stress-dependent fracture flow of samples under high stresses representing CO2 storage conditions. This apparatus can be used to measure peak and residual shear stress

Figure 3: Mohr diagram including Mohr-Coulomb failure envelope. Increasing fluid pres-sure shifts the initial stress state closer to the failure envelope.

Induced tensile or shear failure of caprockApart from existing fracture networks and faults, there is a risk of fracture initiation during underground fluid injection. Different fracture modes are possible, depending on the stress state. Tensile fracturing may take place under a weak extensional stress field because of significant stress enhancement at the fracture tips. Shear fracturing may occur if shear stresses reach a failure crite-rion, of which the Mohr-Coulomb criterion is the most common (see Fig. 3). Rocks do not only fracture because of criti-cal far-field stresses, but also because of changes in an internal pore fluid pressure,

Figure 4: Left panel: pre-fractured shale sample from Svalbard before and after testing in the direct shear box. Right panel: shear stress versus shear displacement for four different Svalbard shale samples when subjected to different normal loads, σn.

on pre-fractured or intact samples versus shear displacement. It provides a friction coefficient and cohesion of the material along the tested plane and thus provides us with a more accurate failure envelope (see Fig. 4) for the material and in situ conditions of interest.

At the SUCCESS center, the geomechani-cal models for injection-induced failure on faults have been discussed based on the Snøhvit/Tubåen pilot. Choi et al. (2015) demonstrated a significant difference in seal integrity for drained (unconfined reservoir) versus undrained (confined reservoir) con-ditions in a modelling study. For the case of CO2 injection, the drained condition might be more critical in most cases, compared to the undrained condition. However, the as-sumption of a drained condition would be too conservative for sealing faults, which are often considered impermeable flow barriers. Bohloli et al. (2017) analyzed the injection pressure versus the CO2 injection rate into well F-2H in the Tubåen Formation at Snøhvit and evaluated the methods for determination of fracture pressure.

The results of that work show that at Snøh-vit, the fracture pressure was not exceeded based on the pressure versus rate plot showing a linear increase in the injection pressure with rate. A similar analysis for In Salah indicated injection into a fractured formation (Bohloli et al., 2017). Comparing the data sets from two CO2 injection proj-ects (In Salah and Snøhvit) demonstrates the behavioral range to be expected in fu-ture CO2 injection projects, including near-wellbore effects, matrix flow and fracture flow behavior.


Figure 6: a) Evolution of the well pressure over time during the fracture process. b) The modelled well pressure is compared to the ob-served well pressure in Adventdalen, Svalbard. The well pressure is shown as the difference between the fluid pressure and the initial reservoir pressure.

a) b)

Figure 5: The normal effective stresses in the x and y directions. The effective stress is plotted around a hydraulic fracture that has de-veloped in heterogeneous rock. The heterogeneities in the rock lead to erratic fracture geometries that might be expected in real rocks. The plot shows the tensile stress enhancement at the fracture tips (blue), which are the most likely places for further fracture propaga-tion. The rock along the sides of the fracture becomes compressed by the opening of the fracture (red).

Modelling of the fracture initiation and propagation is a challenging task, even for simple geometries (Charlez, 1997). There are particular challenges associated with the condition for propagation of the fracture and the leak-off of fluid from the fracture into the host rocks (Yarushina and Berco-vici, 2014; Yarushina et al., 2013). A large variety of numerical models have been de-veloped, following the first pioneering ana-lytical models.

They differ in the way they treat the porous rocks, conditions for fracturing, boundary conditions, spatial dimension, numerical methods, size of the computational domain, heterogeneities and fluid flow within the fracture (Adachi et al., 2007; Charlez, 1997; Mack and Warpinski, 2000). The approach used at the SUCCESS center is based

on the Biot model for poroelasticity (Biot, 1941), but it represents the rock strength by bonds (Wangen, 2011). It therefore has similarities with earlier approaches, where the entire porous material is represented by beams (Tzschichholz and Wangen, 1998) or by springs (Flekkoy et al., 2002).

Our formulation models hydraulic fractur-ing of a reservoir rock using the finite ele-ment method (Wangen, 2011) and makes it possible to model the fracturing process without special meshing for the fracture by introducing fracture porosity. The cur-rent formulation deals with tensile (Mode I) fractures. The maximum limit of the strain in each element is used as a fracture cri-terion. The finite element formulation of the displacement field treats the fractured ele-ments by setting Young's modulus in these

elements to zero or a ‘low’ value. The ex-ample below shows hydraulic fracturing of 50m x 50m horizontal part of a reservoir. The bonds in the grid are assigned random bond strength. The fluid is injected at the center of the grid and a fracture propagates away from the well into the reservoir. The presence of the weak bonds creates a branched fracture, as seen from the plots of normal effective stresses in the x and y directions (see Fig. 5).

The random bond strength results in fluc-tuations in the well pressure, which builds up until the first fracture event, followed by a pressure drop (see Fig. 6a). The process of pressure build-up then resumes, generat-ing a new fracturing event. Each fracturing event can be associated with an induced microseismic event.


This model has been calibrated against pressure observations from a well drilled and tested in the Adventdalen valley in Svalbard (Wangen and Yarushina, 2015). The model reasonably replicates observed pressure variations (see Fig. 6b). The dis-crepancy is due to the element size and non-linearity of the model. The simulations assume an average lateral permeability in the interval of measured sandstone perme-abilities. The model of hydraulic fracturing developed can be used to interpret fracture propagation and well pressure behavior, but the non-linearity of the numerical model and physical processes involved require further studies.

The same approach was used for numeri-cal modelling of hydraulic fracturing in 3D (Wangen, 2013). The model has been tested on two cases – a homogeneous rock and a heterogeneous rock. The homoge-neous case yields three orthogonal fracture planes. The fracture fronts corresponding to the tips of the fractures become circular in the heterogeneous formations, where more than three fractures can be initiated simul-taneously. These will expand symmetrically away from the injection point. All fracturing events that can induce microseismicity are at the front in this case, and there are there-fore no events in the interval between the front and the well. The heterogeneous case yields a branched fracture, where there are events that fill the distance between the injection point and the furthermost fracture tips. The elastic energy released in the larg-est events is in the order of 1 MJ, when the well pressure is in the order of 1 MPa. The homogeneous case yields fewer and larger events than the heterogeneous case.

Sand injections and risk of leakage While the presence of sand injections within the reservoir section may result in improved reservoir connectivity and storage capacity in stacked reservoirs (Aavartsmark et al., 2017), sand injection through the caprocks of potential CO2 storage sites may result in CO2 leakage. Karstens and Berndt (2015) documented amplitude anomalies in the overburden above the CO2 injection site at Sleipner. These were attributed to the presence of gas, and it was suggested that they “represent fluid conduits between deep stratigraphic levels and the shallow subsur-face or even the sea floor”.

Such conduits could potentially be sand injections. The presence of sand injections is important for both storage capacity, injec-tivity and leakage risk. The work that has been performed at the SUCCESS center to investigate the trigger mechanisms, and thereby the presence of sand injections, is covered in the report on Storage capability (Elenius et al., 2018).

Repeated (4D) seismic data does not re-veal any amplitude changes in the over-burden above Sleipner after starting CO2

injection. This observation demonstrates that the (potential) fluid conduits between the Utsira Formation and the high amplitude regions in the overburden are not presently significant CO2 flow conduits. One cannot preclude some CO2 leaks through these potential conduits, but the (potential) leak-age rate is so slow that the CO2 would be dissolved in pore water and precipitate as minerals in a shorter time than it would take to leak to the surface. Thus, while leakage may be a general concern in areas with sand injection, such leakage is not a con-cern above the Sleipner field at present.

Other vertical focused fluid flow structuresFocused fluid flow through relatively narrow fluid conduits is reported for both deep and shallow Earth. It characterizes hydrother-mal vent complexes, mud volcanoes, gas venting systems, reservoir leakage, etc. It is often identified using reflection seismics as near-vertical zones of disturbed data in the form of gas chimneys and fluid escape pipes.

Chimney or pipe anomalies are often ob-served at sub-seabed sediments on con-tinental margins. They are found offshore Norway in actively-producing hydrocarbon fields (e.g. Nyegga, Albuskjell, Ekofisk, Eldfisk, Hod, Tommeliten, Vallhall), offshore Nigeria, Namibia, Angola, in the Adriatic and Barents Seas, in the Gulf of Mexico, and even in Arctic regions (Arntsen et al., 2007; Bunz et al., 2012; Gay et al., 2007; Løseth et al., 2011; Moss and Cartwright, 2010; Ostanin et al., 2013; Roy et al., 2014). Chimneys and pipes represent almost verti-

"The presence of sand injections is important for both storage capacity, injectivity and leakage risk. "

"The model of hydraulic fracturing developed can be used to interpret fracture propagation and well pressure behavior, but the non-linearity of the numerical model and physical processes involved require further studies."


Figure 7: Schematic representation of fluid flow focusing due to porosity wave mechanism. High-porosity fluid-filled domains have higher buoyancy than the surrounding rocks, and thus have the ability to flow upwards as a single pocket in deformable rocks with a time-dependent response. Due to the different responses of the rocks to dilation and compaction, elongation of the initial pocket in the vertical direction leads to formation of a vertical fluid conduit.

cal structures that can be traced down to gas/fluid rich reservoir zones. While seismic chimneys in many cases indicate leakage from underlying reservoirs (Løseth et al., 2009), we do not find that structures with overlying chimneys have leaked more than other structures. We find the contrary; chim-neys overlie many hydrocarbon fields, and compartments with overlying chimneys do not have shallower hydrocarbon water con-tacts than neighboring ones without over-lying chimneys. Fluid conduits imaged as seismic chimneys may nonetheless serve as preferential pathways for CO2 migration (Roy et al., 2014; Tasianas et al., 2016). One of the proposed CCS projects in Kwi-nana, Australia was cancelled after the site characterization revealed numerous gas chimneys in the geological strata (https://sequestration.mit.edu/tools/projects/bp_rio_tinto_kwinana.html).

Pockmarks that are directly observable on the seafloor are other indicators of fluid migration. They can often be related to underlying seismic chimneys. There are

three main settings where pockmarks are commonly present: 1) in offshore hydrocar-bon provinces where fluids that leak from reservoirs reach the seafloor; 2) in gas hydrate regions as the result of ongoing or paleo dissociation of clathrates, and 3) in estuarine and delta regions where the con-stantly deposited organic-rich sediments or drowned wetlands trigger the production of shallow gas (Cartwright et al., 2007; Hov-land and Judd, 1988).

Focused fluid flow systems as potential leakage pathways have been the subject of several projects funded by the EU. In situ experiments on CO2 injection into shallow under-consolidated marine sediments in Ardmucknish Bay in Oban, Scotland ac-companied by high-resolution 2D seismic reflection surveys revealed that while CO2

migration was controlled by sediment stra-tigraphy in the early stages of the experi-ment, seismic chimneys developed at an increasing flow rate and the total volume in-jected eventually reached the seabed (Cev-atoglu et al., 2015). Production experience

from Gullfaks and wastewater injection in the Hordaland Group in the Norwegian North Sea also show that injection-induced fluid pressures can result in caprock leak-age (Hermanrud et al., 2013; Løseth et al., 2011).

One of the best-documented CO2 storage operations today is the injection of about one million tons of CO2 per year since 1996 into the Utsira Formation at Sleipner in the Norwegian North Sea. The first repeat seis-mic survey (1999) revealed that the migrat-ing CO2 had spread to nine distinct layers – one of these lying above the 5–6.5 m thick shale. The migrating CO2 appears mainly to have been fed to the different layers from a central vertical feeder located right above the well perforations, which is expressed as a seismic chimney in the seismic data (Boait et al., 2012; Hermanrud et al., 2009).

Well data obtained inside and outside of a gas chimney reveals that chimneys are characterized by increased fluid pressure, compared to the background levels (Løseth et al., 2009). Hydraulic fracturing is one of the possible mechanisms of formation of lo-calized fluid conduits (Arntsen et al., 2007).

Another mechanism that has been studied at the SUCCESS center is the spontane-ous self-localization of porous fluid flow due to mechanical instability called “porosity waves” in the geophysical literature (Con-nolly and Podladchikov, 2007; Omlin et al., 2017a; Revil, 2002; Räss et al., 2014; Ya-rushina et al., 2015). Hydraulic fracturing might be expected in hard brittle rocks while

" This has led to an understanding that viscous deformation plays an important role in reservoirs at timescales that are relevant for CO2 sequestration."


porosity waves form in soft sediments. Thus, rock rheology is an important factor in determining fluid flow focusing mecha-nisms.

Previously, shallow reservoirs were mostly considered to behave elastically and fail as a result of frictional plasticity. Recent advances in experimental rock mechanics have made it possible to perform accurate measurements of time-dependent creep under shallow crustal temperature, pres-sure and stress conditions. This has led to an understanding that viscous deformation plays an important role in reservoirs at tim-escales that are relevant for CO2 sequestra-tion (days to 10 000 years) (Brantut et al., 2013; Chang and Zoback, 2009; Hangx et al., 2010; Räss et al., 2017; van Noort and Yarushina, 2016).

A state-of-the-art model for viscoelasto-plastic deformation coupled with fluid flow (Yarushina and Podladchikov, 2015) was specifically developed at the SUCCESS center in order to address this wide range of rheological responses. The model suc-cessfully reproduces experimental data on both instantaneous elastoplastic deforma-tion and time-dependent creep. Based on this theoretical model, we have developed

new fully-coupled numerical code (i.e. the solid feels the fluid pressure and the fluid flow is affected by solid stresses and defor-mation of the porous rocks), which allows the simulation of CO2 injection and flow in a stressed crust that may deform in a vis-coelastoplastic manner (Räss et al., 2016; Räss et al., 2014).

Rocks are known to be heterogeneous and have a wide porosity distribution, even with-in the same formation or reservoir. This het-erogeneity aids the formation of fluid-filled porous pockets that are lighter than the sur-rounding rocks (see Fig. 7). The buoyancy of porous pockets, combined with the ability of the rock matrix to creep, even under low stresses, thus is the driving force for upward fluid propagation. Slightly higher fluid pres-sures at the top of the porous pocket create deviatoric stresses in the rocks above.

In creeping rocks, even small stresses would cause deformation and lead to pore dilation above the fluid pocket. Fluid pres-sure generated by buoyancy and deviatoric stresses in the rock matrix might be suf-ficiently high to overcome capillary entry pressure for thin intra-reservoir shale layers and caprock. At the same time, the reduced pressures at the bottom of the pocket will

lead to pore contraction, pushing the fluid pocket upwards (Connolly and Podlad-chikov, 2007; Yarushina et al., 2015).

Focusing of the flow from the high-porosity pocket occurs due to non-symmetrical dila-tion/contraction of the pore space; i.e. dila-tion of porosity happens much easier and faster than its contraction. Upward propa-gation of formed fluid conduits is further sustained by creeping rocks. Numerical modelling shows that in soft sed-iments, fluids or gas-filled conduits propa-gate upwards as a self-sustained body, losing the initial connection to the feeding reservoir or an injection point (see Fig. 8). The momentum that a conduit has acquired during propagation is enough for its upward propagation, as long as the permeability and viscosity of the rocks do not vary sig-nificantly. If a conduit meets another layer of even more impermeable but still viscous rocks, it will be halted temporarily, spread-ing the fluid underneath the barrier but will resume its upward propagation on a slower timescale. Conduit formation exhibits self-organizing features, in that the number of conduits and their spatial distribution are controlled by permeability, fluid viscosity and the bulk viscosity of the rocks (McKen-zie, 1987).

"Based on this theoretical model, we have developed new fully-coupled numerical code, which allows the simulation of CO2 injection and flow in a stressed crust that may deform in a viscoelastoplastic manner"

"Recent advances in experimental rock mechanics have made it possible to perform accurate measurements of time-dependent creep under shallow crustal temperature, pressure and stress conditions."


Figure 8: Fluid escape pipes formed in soft sediments as a result of leakage from a fluid-filled reservoir. a) Initial conditions used for numerical modelling: the high-porosity (φ) elliptical reservoir is centered at 1/5 of the domain depth. Inside the reservoir the initial porosity is three times higher than in the rest of the domain (φ0). The gravity force is acting downwards, thus initiating buoyancy-driven fluid extraction. b) Fluid escape pipes began to grow from the initially elliptical reservoir shown on the left. The colors indicate porosity normalized to the homogeneous background value. Note that porosity within the pipe might be higher than in the initial reservoir. c) Fluid-escape pipes in 3D. The colors indicate permeability normalized by fluid viscosity. d) Outline of a typical pipe drawn on the basis of available observations (from (Løseth et al., 2011)).


Figure 9: a) Results from a systematic investigation that shows the dependence of the speed of upward conduit propagation on the bulk viscosity of the porous rocks during compaction, ηC, and dilation, ηD, and the shear viscosity of the rocks, μS. The contoured values are the maximum vertical fluid percolation flux normalized by the background Darcy velocity VDarcy. b) The background Darcy ve-locities (log10) expressed in meter per year as a function of effective permeability k and fluid viscosity, μ. φ0 is the background porosity.

Systematic numerical simulations show that the speed of propagation of conduits can be up to 1 000 times greater than the speed of the background diffusional fluid flow defined by simple Darcy’s law (see Fig. 9) (Omlin et al., 2017b). The Darcy velocity of the flow outside of the conduit mainly depends on the fluid viscosity and permeability (see Fig. 9b).

The flow of CO2 through shales with perme-ability of 10-7 D (10-19 m2) and fluid viscosity of 10-3Pa s results in the Darcy velocity of ≈10-5 ÷ 10-4 m/yr. Fluid-filled conduits can thus propagate with a speed in the order

of ≈1 ÷ 10 cm/yr. It will take 1 000 to 10 000 years to penetrate 100 m of a nearly impermeable shale layer, due to the effects of the time-dependent ability of shales to relax stresses. If the presence of CO2, as measured in the laboratory, shows chemi-cal effects of permeability enhancement (Armitage et al., 2013), the timescale of conduit propagation will be reduced even further, leading to a propagation velocity of meters per year.

Our simulations replicate features observed in CO2 injection operations, such as local-ization of flow into separate conduits and

flow that is locally and periodically faster than predicted by Darcy diffusional flow. The possibility of triggering focused fluid flow via vertical conduits must thus be con-sidered in the risk analysis of the anticipat-ed fluid injection operations during geologi-cal CO2 storage and wastewater injection, and must be added to the list of potential leakage pathways alongside existing faults.

At the same time, accurate creep rate for-mulations based on experimental data need to be included into theoretical models and numerical codes for fluid flow in soft sedi-ments.


A combined experimental, field and nu-merical modelling study of various leakage pathways was undertaken at the FME SUC-CESS center. We showed that oil and gas industry techniques for assessing reservoir compartmentalization may be useful for in-vestigating prospective CO2 storage sites. They may help constrain material param-eters and the past history of fluid commu-nication in the reservoir and caprock, and thus define the quality of the seal. Broader application of caprock characterization techniques, such as geochemical methods, is needed, especially during establishment of new CO2 storage sites.

The possibility of generating new leakage pathways was considered together with leakage through pre-existing structures. The development of leakage pathways dur-ing injection is different in different types of rock. In soft rocks, it occurs due to the formation of focused fluid flow systems as a result of fluid flow instability (i.e. soli-tary porosity waves) and strongly-coupled flow and deformation processes. In hard rocks, hydraulic fractures will be initiated if

Conclusion and recommendations

stresses reach critical levels. Both of these represent preferential leakage pathways. In a sequence of hard/soft rocks, hydrofractur-ing will be halted as soon as it hits the soft rocks. Porosity waves are then responsible for fluid flow through soft layers, as hydrau-lic fracturing of soft rocks is very difficult and unlikely due to the high strain potential of such rocks. While there is a considerable body of work on the generation of hydraulic fractures, much less research has been un-dertaken on the formation of focused fluid flow systems in soft rocks. Further studies on the formation of focused fluid flow sys-tems in soft rocks are thus needed in the future in order to constrain the critical ma-terial and engineering parameters that are responsible for formation of these features and also for predictive modelling at the rel-evant storage sites.

Our work indicates that the flow through fractures and faults depends on the perme-ability of the fault gauge, fracture surface roughness and the degree of confinement, while flow through intact rocks depends on stress and time-dependent permeabil-

ity, elastic bulk moduli and effective bulk viscosity of the rocks. Forecasting flow in rocks with complex material properties is difficult, and more work is needed on cou-pled geomechanics and fluid flow/reservoir simulations. Stress- and time-dependent rock properties such as those measured for shales and a number of other rock types need to be included in numerical models. However, very few commercial simulators and research codes are capable of this, which creates a need for development of a new generation of geomechanical simula-tors coupled with reservoir flow that would account for stress and time dependency of rock properties. Controlling parameters for fractures within and around faults also need better understanding and quantifica-tion. Further research is needed in order to identify actions required to reduce risk.

Based on the results of our work, we sug-gest that the workflow for assessment of fu-ture CO2 injection sites must include studies on two scales. On a large scale, identifica-tion of sealing sequences is needed, which includes all important sealing objects, their position in the sedimentary basin, their interdependence on each other and their function with respect to CO2 retention ca-pacity. On the small scale of an individual object, an evaluation is needed of the mate-rial properties and their dynamic behavior in response to the stress introduced by CO2 injection needs.

The following guidelines and recommenda-tions for environmental practices are based on these experiences.

"Broader application of caprock characterization techniques, such as geochemical methods, is needed, especially during establishment of new CO2 storage sites."


Identifying the ideal reservoirGeological and hydrogeological character-ization of the storage unit should include identification and characterization of fault zones and focused fluid flow features that could affect containment. Conventional techniques for assessing reservoir com-partmentalization in the oil and gas industry may be useful for investigating prospective CO2 storage sites. Isotope geochemistry is an effective tool for understanding the fluid communication in reservoirs and caprock shale on a geological timescale, and thus to help constrain the long-term behavior of a potential CO2 storage site.

Geostatic modellingLarge identified features, including faults, focused fluid flow features, compartmen-talization, etc. need to be incorporated into geological models that should form the ba-sis for flow modeling. The geological mod-els should provide information on poros-ity and permeability distribution, the initial pressure, temperature and stress distribu-tion, elastic moduli and failure parameters already identified by ISO/DIS27914 (2017), but also shear and bulk viscosities of the reservoir and caprock, as they affect flow behavior and the long-term fate of injected CO2.

Predictive flow modellingNumerical modelling must be applied in order to predict the behavior of the stor-age of complex and injected CO2. Oil and gas industry modelling software packages may be used as a first order estimate for

the processes that occur in the reservoir. There is a need for further development of more robust flow simulators, coupled with geomechanical modelling that accounts for realistic stress- and time-dependent proper-ties for reservoir and caprock. Coupled geo-mechanical/flow modelling should be used to identify the injection strategy that would minimize the potential for stress changes, deformation, induced seismicity, forma-tion of fractures and new focused fluid flow pathways.

In addition to the key flow and geomechani-cal modelling parameters outlined by ISO/DIS27914 (2017) (pressure, temperature, fluid saturations; chemical composition of formation fluids; porosity distribution; per-meability distribution; formation geometry; fluid and rock compressibilities; and the thermal properties of fluids and rocks), pres-sure dependence and time dependence of porosity, permeability and rock compress-ibilities must be taken into account. Bulk viscosity of the rock should be included in the list of key material parameters affecting the flow.

Geomechanical site characterizationBaseline geomechanical characterization should include the following rock mechani-cal properties of both the storage unit and overlying seal: strength and deformation properties, according to the observed material behavior of the concerned rock, including stress-dependent elastic moduli (e.g. Poisson’s ratio and Young’s modu-lus); creep rates and stress-dependent bulk viscosity; estimation of the fracture propa-gation pressure and failure parameters, including the cohesion and friction angle. The flow properties of existing faults and focused fluid flow structures also need to be evaluated.

Site closureCoupled geomechanical/flow modelling, which accounts for time-dependent evolu-tion of rock properties (rock weakening, pressure-dependent elastic moduli, rock viscosity) should be used for long-term (post-injection) safety assessment of the injection site, especially for large-scale op-erations.

"Coupled geomechanical/flow modelling should be used to identify the injection strategy that would minimize the potential for stress changes, deformation, induced seismicity, formation of fractures and new focused fluid flow pathways."

"The flow properties of existing faults and focused fluid flow structures also need to be evaluated."


References

Adachi, A., Siebrits, E., Peirce, A. and Desroches, J. (2007), Computer simulation of hydraulic fractures, Int J Rock Mech Min, 44(5), 739-757, doi:10.1016/j.ijrmms.2006.11.006.Al Ismail, M. I., H. S., Reece, J. S. and Zoback, M. D. (2014), The effect of CO2 adsorption on permeability anisotropy in the Eagle Ford shale, paper presented at Unconventional Resources Technology Conference, Denver, Colorado, USA.Al Ismail, M. I. and Zoback, M. D. (2016), Effects of rock mineralogy and pore structure on stress-dependent permeability of shale samples, Phil. Trans. R. Soc. A, 374, 20150428.Alikarami, R., Torabi, A., Kolyukhin, D. and Skurtveit, E. (2013), Geostatistical relationships between mechanical and petrophysical properties of deformed sandstone, Int J Rock Mech Min, 63, 27-38, doi:10.1016/j.ijrmms.2013.06.002.Allen, N. (2014), Pores, porosity and pore size distribution of some Draupne Formation and Colorado Group shales and kerogens, Newcastle University, New-castle University.Angeli, M., Soldal, M., Skurtveit, E. and Aker, E. (2009), Experimental percolation of supercritical CO2 through a caprock, Enrgy Proced, 1(1), 3351-3358, doi:10.1016/j.egypro.2009.02.123.Armitage, P. J., Faulkner, D. R. and Worden, R. H. (2013), Caprock corrosion, Nat Geosci, 6(2), 79-80, doi:10.1038/ngeo1716.Arntsen, B., Wensaas, L., Loseth, H. and Hermanrud, C. (2007), Seismic modeling of gas chimneys, Geophysics, 72(5), Sm251-Sm259, doi:10.1190/1.2749570.Bachu, S. and Bennion, D. B. (2009), Interfacial Tension between CO2, Freshwater, and Brine in the Range of Pressure from (2 to 27) MPa, Temperature from (20 to 125) degrees C, and Water Salinity from (0 to 334 000) mg(.)L(-1), J Chem Eng Data, 54(3), 765-775, doi:10.1021/je800529x.Bhandari, A. R., Flemings, P. B., Polito, P. J., Cronin, M. B. and Bryant, S. L. (2015), Anisotropy and Stress Dependence of Permeability in the Barnett Shale, Transport Porous Med, 108(2), 393-411, doi:10.1007/s11242-015-0482-0.Biot, M. A. (1941), General theory of three-dimensional consolidation, Journal of Applied Physics, 12(2), 155-164.Birkholzer, J. T., Nicot, J.P., Oldenburg, C. M., Zhou, Q. L., Kraemer, S. and Bandilla, K. (2011), Brine flow up a well caused by pressure perturbation from geo-logic carbon sequestration: Static and dynamic evaluations, Int J Greenh Gas Con, 5(4), 850-861, doi:10.1016/j.ijggc.2011.01.003.Bjørlykke, K. (2015), Introduction to petroleum geology, in Petrol Geosci, edited by K. Bjørlykke, pp. 1-29, Springer, Berlin, Heidelberg, doi:10.1007/978-3-642-34132-8_1.Boait, F. C., White, N. J., Bickle, M. J., Chadwick, R. A., Neufeld, J. A. and Huppert, H. E. (2012), Spatial and temporal evolution of injected CO2 at the Sleipner Field, North Sea, J Geophys Res-Sol Ea, 117, doi:10.1029/2011jb008603.Bohloli, B., Ringrose, P., Grande, L. and Nazarian, B. (2017), Determination of the fracture pressure from CO2 injection time-series datasets, Int J Greenh Gas Con, 61, 85–93, doi:10.1016/j.ijggc.2017.03.025.Braathen, A., et al. (2012), The Longyearbyen CO2 Lab of Svalbard, Norway-initial assessment of the geological conditions for CO2 sequestration, Norw J Geol, 92(4), 353-376.Brantut, N., Heap, M. J., Meredith, P. G. and Baud, P. (2013), Time-dependent cracking and brittle creep in crustal rocks: A review, J Struct Geol, 52, 17-43, doi:10.1016/j.jsg.2013.03.007.Bunz, S., Polyanov, S., Vadakkepuliyambatta, S., Consolaro, C. and Mienert, J. (2012), Active gas venting through hydrate-bearing sediments on the Vestnesa Ridge, offshore W-Svalbard, Mar Geol, 332, 189-197, doi:10.1016/j.margeo.2012.09.012.Busch, A., Bertier, P., Gensterblum, Y., Rother, G., Spiers, C. J., Zhang, M. and Wentinck, H. M. (2016), On sorption and swelling of CO2 in clays, Geomech Geophys Geo-, 2(2), 111-130, doi:10.1007/s40948-016-0024-4.Busch, A., et al. (2014), The Green River natural analogue as a field laboratory to study the long-term fate of CO2 in the subsurface, 12th International Confer-ence on Greenhouse Gas Control Technologies, Ghgt-12, 63, 2821-2830, doi:10.1016/j.egypro.2014.11.304.Cartwright, J., Huuse, M., and Aplin, A. (2007), Seal bypass systems, Aapg Bulletin, 91(8), 1141-1166, doi:10.1306/04090705181.Cassini, E., Mylnikov, D. and Makhnenko, R. Y. (2017), Chemical influence of pore pressure on brine flow in clay-rich material, in Advances in Laboratory Testing and Modelling of Soils and Shales (ATMSS), edited by L. L. Ferrari A., Springer, Cham.Cevatoglu, M., Bull, J. M., Vardy, M. E., Gernon, T. M., Wright, I. C. and Long, D. (2015), Gas migration pathways, controlling mechanisms and changes in sediment acoustic properties observed in a controlled sub-seabed CO2 release experiment, Int J Greenh Gas Con, 38, 26-43, doi:10.1016/j.ijggc.2015.03.005.Chalmers, G. R. L., Ross, D. J. K. and Bustin, R. M. (2012), Geological controls on matrix permeability of Devonian Gas Shales in the Horn River and Liard basins, northeastern British Columbia, Canada, Int J Coal Geol, 103, 120-131, doi:10.1016/j.coal.2012.05.006.Chang, C. D., and Zoback, M. D. (2009), Viscous creep in room-dried unconsolidated Gulf of Mexico shale (I): Experimental results, J Petrol Sci Eng, 69(3-4), 239-246, doi:10.1016/j.petrol.2009.08.018.


Charlez, P. A. (1997), Rock mechanics, v. <2 > pp., Technip, Paris.Chen, C., Dong, B., Zhang, N., Li, W. Z. and Song, Y. C. (2016), Pressure and Temperature Dependence of Contact Angles for CO2/Water/Silica Systems Pre-dicted by Molecular Dynamics Simulations, Energ Fuel, 30(6), 5027-5034, doi:10.1021/acs.energyfuels.6b00171.Chhatre, S. S., et al. (2015), Steady-state stress-dependent permeability measurements of tight oil-bearing rocks, Petrophysics, 56, 116-124.Choi, J. C., Skurtveit, E., Bohloli, B. and Grande, L. (2015), Geomechanical assessment of flow barriers observed from 4D time-lapse survey for CO2 storage in Snøhvit field, paper presented at Fourth International Conference on Fault and Top Seal, Almeria, Spain.Connolly, J. A. D., and Podladchikov, Y. Y. (2007), Decompaction weakening and channeling instability in ductile porous media: Implications for asthenospheric melt segregation, J Geophys Res-Sol Ea, 112(B10), B10205, doi:10.1029/2005jb004213.Cuss, R. J., Harrington, J. F., Noy, D. J., Graham, C. C. and Sellin, P. (2014), Evidence of localised gas propagation pathways in a field-scale bentonite engineered barrier system; results from three gas injection tests in the large scale gas injection test (Lasgit), Appl Clay Sci, 102, 81-92, doi:10.1016/j.clay.2014.10.014.David, C., Wong, T. F., Zhu, W. L. and Zhang, J. X. (1994), Laboratory measurement of compaction-induced permeability change in porous rocks - Implications for the generation and maintenance of pore pressure excess in the crust, Pure Appl Geophys, 143(1-3), 425-456, doi:10.1007/Bf00874337.Dockrill, B., and Shipton, Z. K. (2010), Structural controls on leakage from a natural CO2 geologic storage site: Central Utah, USA, J Struct Geol, 32(11), 1768-1782, doi:10.1016/j.jsg.2010.01.007.Dong, J. J., Hsu, J. Y. Wu, W. J., Shimamoto, T., Hung, J. H., Yeh, E. C., Wu, Y. H. and Sone, H. (2010), Stress-dependence of the permeability and porosity of sandstone and shale from TCDP Hole-A, Int J Rock Mech Min, 47(7), 1141-1157, doi:10.1016/j.ijrmms.2010.06.019.Elenius, M. (ed), Sundal, A. (ed), Gasda, S. (ed), Aavatsmark, I., Hellevang, H., Hermanrud, C., Skurtveit, E. and Midttømme, K. (2018), Storage capability, FME SUCCESS synthesis report Vol. 1, pp 32 Rep., CMR. Flekkoy, E. G., Malthe-Sorenssen, A. and Jamtveit, B. (2002), Modeling hydrofracture, J Geophys Res-Sol Ea, 107(B8), doi:10.1029/2000jb000132.Gay, A., Lopez, M., Berndt, C. and Seranne, M. (2007), Geological controls on focused fluid flow associated with seafloor seeps in the Lower Congo Basin, Mar Geol, 244(1-4), 68-92, doi:10.1016/j.margeo.2007.06.003.Ghanizadeh, A., Amann-Hildenbrand, A., Gasparik, M., Gensterblum, Y., Krooss, B. M. and Littke, R. (2014), Experimental study of fluid transport processes in the matrix system of the European organic-rich shales: II. Posidonia Shale (Lower Toarcian, northern Germany), Int J Coal Geol, 123, 20-33, doi:10.1016/j.coal.2013.06.009.Ghanizadeh, A., Gasparik, M., Amann-Hildenbrand, A., Gensterblum, Y. and Krooss, B. M. (2013), Lithological controls on matrix permeability of organic-rich shales: An experimental study, European Geosciences Union General Assembly 2013, Egudivision Energy, Resources & the Environment, Ere, 40, 127-136, doi:10.1016/j.egypro.2013.08.016.Gor, G. Y., Stone, H. A. and Prevost, J. H. (2013), Fracture Propagation Driven by Fluid Outflow from a Low-Permeability Aquifer, Transport Porous Med, 100(1), 69-82, doi:10.1007/s11242-013-0205-3.Gutierrez, M., Katsuki, D. and Tutuncu, A. (2015), Determination of the continuous stress-dependent permeability, compressibility and poroelasticity of shale, Mar Petrol Geol, 68, 614-628, doi:10.1016/j.marpetgeo.2014.12.002.Hangx, S. J. T., Spiers, C. J. and Peach, C. J. (2010), Creep of simulated reservoir sands and coupled chemical-mechanical effects of CO2 injection, J Geophys Res-Sol Ea, 115, doi:Artn B09205 10.1029/2009jb006939.Harrington, J. F., Noy, D. J., Horseman, S. T., Birchall, D. J. and Chadwick, R. A. (2009), Laboratory study of gas and water flow in the Nordland shale, Sleipner, North Sea, in Carbon dioxide sequestration in geological media – State of the science, edited by M. Grobe, J. C. Pashin and R. L. Dodge, pp. 521–543, AAPG Studies in Geology, doi:10.1306/13171259St593394.Heller, R., Vermylen, J. and Zoback, M. (2014), Experimental investigation of matrix permeability of gas shales, Aapg Bulletin, 98(5), 975-995, doi:10.1306/09231313023.Hermanrud, C., Andresen, T., Eiken, O., Hansen, H., Janbu, A., Lippard, J., Bolas, H. N., Simmenes, T. H., Teige, G. M. G. and Ostmo, S. (2009), Storage of CO2 in saline aquifers - lessons learned from 10 years of injection into the Utsira Formation in the Sleipner area, Enrgy Proced, 1(1), 1997-2004, doi:10.1016/j.egypro.2009.01.260.Hermanrud, C., Eiken, O., Hansen, O. R., Nordgård Bolås, H. M., Simmenes, T. H., Teige, G. M. G., Hansen, H. and Johansen, S. (2013), Importance of pres-sure management in CO2 storage, paper presented at Offshore Technology Conference, Houston, Texas, USA.Hermanrud, C., and Georgescu, L. (2014), Relationships between bright amplitudes in overburden rocks and leakage from underlying hydrocarbon reservoirs, paper presented at EAGE Shallow Anomalies Workshop Malta.Hermanrud, C., Halkjelsvik, M. E., Kristiansen, K., Bernal, A. and Stromback, A. C. (2014), Petroleum column-height controls in the western Hammerfest Basin, Barents Sea, Petrol Geosci, 20(3), 227-240, doi:10.1144/petgeo2013-041.Hildenbrand, A., Schlomer, S., Krooss, B. M. and Littke, R. (2004), Gas breakthrough experiments on pelitic rocks: comparative study with N-2, CO2 and CH4, Geofluids, 4(1), 61-80, doi:10.1111/j.1468-8123.2004.00073.x.Hou, Z. S., Rockhold, M. L. and Murray, C. J. (2012), Evaluating the impact of caprock and reservoir properties on potential risk of CO2 leakage after injection, Environ Earth Sci, 66(8), 2403-2415, doi:10.1007/s12665-011-1465-2.Hovland, M., and Judd, A. G. (1988), Seabed pockmarks and seepages: Impact on geology, biology and the marine environment, xii, 293 pp. pp., Graham & Trotman (Kluwer), London, Dordrecht, Boston.Huang, Z. Q., Winterfeld, P. H., Xiong, Y., Wu, Y. S. and Yao, J. (2015), Parallel simulation of fully-coupled thermal-hydro-mechanical processes in CO2 leakage through fluid-driven fracture zones, Int J Greenh Gas Con, 34, 39-51, doi:10.1016/j.ijggc.2014.12.012.Humez, P., Audigane, P., Lions, J., Chiaberge, C. and Bellenfant, G. (2011), Modeling of CO2 Leakage up Through an Abandoned Well from Deep Saline Aquifer to Shallow Fresh Groundwaters, Transport Porous Med, 90(1), 153-181, doi:10.1007/s11242-011-9801-2.Huq, F., S. P.C., Mørkved, P. T., Yarushina, J. L. V. and Johansen, H. (2017), The Longyearbyen CO2 Lab: fluid communication in reservoir and cap-rock, Int J Greenh Gas Con, 63, 59-76.Ilton, E. S., Schaef, H. T., Qafoku, O., Rosso, K. M. and Felmy, A. R. (2012), In Situ X-ray Diffraction Study of Na+ Saturated Montmorillonite Exposed to Variably Wet Super Critical CO2, Environ Sci Technol, 46(7), 4241-4248, doi:10.1021/es300234v.


ISO/DIS27914 (2017), Carbon dioxide capture, transportation and geological storage — Geological storage, edited, pp. 1 - 79, The International Organization for Standardization, Switzerland.Jeanne, P., Guglielmi, Y. and Cappa, F. (2013), Dissimilar properties within a carbonate-reservoir's small fault zone, and their impact on the pressurization and leakage associated with CO2 injection, J Struct Geol, 47, 25-35, doi:10.1016/j.jsg.2012.10.010.Jemai, K., Vafaei, M. T., Kvamme, B. and Chejara, A. (2017), Simulation of CO2 hydrate formation in cold aquifers: nonequilibrium approach, Arab J Geosci, 10(5), doi:Artn 113 10.1007/S12517-017-2888-1.Karstens, J., and Berndt, C. (2015), Seismic chimneys in the Southern Viking Graben - Implications for palaeo fluid migration and overpressure evolution, Earth Planet Sc Lett, 412, 88-100, doi:10.1016/j.epsl.2014.12.017.Kutchko, B. G., Strazisar, B. R., Dzombak, D. A., Lowry, G. V. and Thaulow, N. (2007), Degradation of well cement by CO2 under geologic sequestration condi-tions, Environ Sci Technol, 41(13), 4787-4792, doi:10.1021/es062828c.Kwon, O., Kronenberg, A. K., Gangi, A. F., Johnson, B. and Herbert, B. E. (2004), Permeability of illite-bearing shale: 1. Anisotropy and effects of clay content and loading, J Geophys Res-Sol Ea, 109(B10), doi:Artn B10205 10.1029/2004jb003052.Løseth, H., Gading, M. and Wensaas, L. (2009), Hydrocarbon leakage interpreted on seismic data, Mar Petrol Geol, 26(7), 1304-1319, doi:10.1016/j.marpet-geo.2008.09.008.Løseth, H., Wensaas, L., Arntsen, B., Hanken, N. M., Basire, C. and Graue, K. (2011), 1000 m long gas blow-out pipes, Mar Petrol Geol, 28(5), 1047-1060, doi:10.1016/j.marpetgeo.2010.10.001.Mack, M. G., and Warpinski, N. R. (2000), Mechanics of hydraulic fracturing, in Reservoir stimulation, edited by M. J. Economides and K. G. Nolte, pp. 6.1 - 6.49, Wiley, Chichester, England; New York.Makhnenko, R. Y., Vilarrasa, V., Mylnikova, D. and Laloui, L. (2017), Hydromechanical aspects of CO2 breakthrough into clay-rich caprock, Energy Procedia, 114, 3219 - 3228.McKenzie, D. (1987), The compaction of igneous and sedimentary rocks, J Geol Soc London, 144, 299-307.Metwally, Y. M., and Chesnokov, E. M. (2011), Gas shale; relationships between permeability and intrinsic composition, paper presented at SEG Technical Program Expanded Abstracts.Metwally, Y. M., and Sondergeld, C. H. (2011), Measuring low permeabilities of gas-sands and shales using a pressure transmission technique, Int J Rock Mech Min, 48(7), 1135-1144, doi:10.1016/j.ijrmms.2011.08.004.Milliken, K. L., Esch, W. L., Reed, R. M. and Zhang, T. W. (2012), Grain assemblages and strong diagenetic overprinting in siliceous mudrocks, Barnett Shale (Mississippian), Fort Worth Basin, Texas, Aapg Bulletin, 96(8), 1553-1578, doi:10.1306/12011111129.Mondol, N. H., Bjorlykke, K., Jahren, J. and Hoeg, K. (2007), Experimental mechanical compaction of clay mineral aggregates - Changes in physical properties of mudstones during burial, Mar Petrol Geol, 24(5), 289-311, doi:10.1016/j.marpetgeo.2007.03.006.Moss, J. L., and Cartwright, J. (2010), The spatial and temporal distribution of pipe formation, offshore Namibia, Mar Petrol Geol, 27(6), 1216-1234, doi:DOI 10.1016/j.marpetgeo.2009.12.013.Newell, D. L., and Carey, J. W. (2013), Experimental Evaluation of Wellbore Integrity Along the Cement-rock Boundary, Environ Sci Technol, 47(1), 276-282, doi:10.1021/es3011404.Nooraiepour, M., Bahman, B., Park, J., Sauvin, G., Skurtveit, E., and Mondol, N.H., 2018. Experimental study of brine/CO2 transport through a naturally frac-tured tight sandstones: an integrated fluid flow and geophysical investigation, Geophysics, 83(1), JANUARY-FEBRUARY, 1-12.Nordbotten, J. M., Kavetski, D., Celia, M. A. and Bachu, S. (2009), Model for CO2 Leakage Including Multiple Geological Layers and Multiple Leaky Wells, Environ Sci Technol, 43(3), 743-749, doi:10.1021/es801135v.Ogata, K., Senger, K., Braathen, A., Tveranger, J., Petrie, E. and Evans, J. P. (2012), Fracture-related fluid flow in sandstone reservoirs - Insights from outcrop analogues of South-eastern Utah, in Third EAGE CO2 Geological Storage Workshop, edited, Edinburgh, Scotland, doi:10.3997/2214-4609.20143823 Olsen, R., Leirvik, K. and Kvamme, B. (submitted), Adsorption characteristics of glycols on calcite and hematit, J. Chem. Phys.Olsen, R., Leirvik, K. N., Kvamme, B. and Kuznetsova, T. (2016), Effects of Sodium Chloride on Acidic Nanoscale Pores Between Steel and Cement, J Phys Chem C, 120(51), 29264-29271, doi:10.1021/acs.jpcc.6b10043.Omlin, S., B. Malvoisin, and Podladchikov, Y. Y. (2017a), Pore Fluid Extraction by Reactive Solitary Waves in 3-D, Geophysical Research Letters, 44(18), 9267-9275, doi:10.1002/2017GL074293.Omlin, S., Räss, L. and Podladchikov, Y. Y. (2017b), Simulation of three-dimensional viscoelastic deformation coupled to porous fluid flow, Tectonophysics, doi:10.1016/j.tecto.2017.08.012.Ostanin, I., Anka, Z., di Primio, R. and Bernal, A. (2013), Hydrocarbon plumbing systems above the Snohvit gas field: Structural control and implications for thermogenic methane leakage in the Hammerfest Basin, SW Barents Sea, Mar Petrol Geol, 43, 127-146, doi:10.1016/j.marpetgeo.2013.02.012.Prince, C. M., Steele, D. D., Zelaya, R. and Devier, C. A. (2011), Shale diagenesis and permeability: Examples from the Barnett shale and the Marcellus forma-tion, Search and discovery article #50372, paper presented at AAPG International Conference and Exhibition, Calgary, Alberta.Pruess, K. (2005), Numerical studies of fluid leakage from a geologic disposal reservoir for CO2 show self-limiting feedback between fluid flow and heat transfer, Geophys Res Lett, 32(14), doi:10.1029/2005gl023250.Qorbani, K., Kvamme, B. and Kuznetsova, T. (2017), Simulation of CO2 storage into methane hydrate reservoirs, Non-equilibrium thermodynamic approach, Energy Procedia, 114, 5451 - 5459.Qorbani, K., Kvamme, B. and Olsen, R. (2016), Sensitivity Analysis of CO2 Injection within Saline Aquifers for Storage Purposes in the Form of Hydrate Using a Reactive Transport Simulator, J Chem Eng Data, 61(12), 4148-4156, doi:10.1021/acs.jced.6b00567.Revil, A. (2002), Genesis of mud volcanoes in sedimentary basins: A solitary wave-based mechanism, Geophysical Research Letters, 29(12), 1574, doi:10.1029/2001gl014465.Roy, S., Senger, K., Braathen, A., Noormets, R., Hovland, M. and Olaussen, S. (2014), Fluid migration pathways to seafloor seepage in inner Isfjorden and Adventfjorden, Svalbard, Norw J Geol, 94(2-3), 99-119.


Rutqvist, J., Birkholzer, J., Cappa, F. and Tsang, C. F. (2007), Estimating maximum sustainable injection pressure during geological sequestration of CO2 using coupled fluid flow and geomechanical fault-slip analysis, Energ Convers Manage, 48(6), 1798-1807, doi:10.1016/j.enconman.2007.01.021.Rutqvist, J., Birkholzer, J. T. and Tsang, C. F. (2008), Coupled reservoir-geomechanical analysis of the potential for tensile and shear failure associated with CO2 injection in multilayered reservoir-caprock systems, Int J Rock Mech Min, 45(2), 132-143, doi:10.1016/j.ijrmms.2007.04.006.Rutqvist, J., and Tsang, C. F. (2002), A study of caprock hydromechanical changes associated with CO2-injection into a brine formation, Environ Geol, 42(2-3), 296-305, doi:10.1007/s00254-001-0499-2.Rybacki, E., Reinicke, A., Meier, T., Makasi, M. and Dresen, G. (2015), What controls the mechanical properties of shale rocks? - Part I: Strength and Young's modulus, J Petrol Sci Eng, 135, 702-722, doi:10.1016/j.petrol.2015.10.028.Räss, L., Makhnenko, R. Y., Podladchikov, Y. Y. and Laloui, L. (2017), Quantification of viscous creep influence on storage capacity of caprock, Energy Procedia, 114, 3237 - 3246, doi:10.1016/j.egypro.2017.03.1455.Räss, L., Yarushina, V., Duretz, T. and Podladchikov, Y. (2016), High-resolution numerical modelling to resolve the dynamics of pipe structures in porous media, paper presented at ECMOR XV - 15th European Conference on the Mathematics of Oil Recovery, Amsterdam.Räss, L., Yarushina, V. M., Simon, N. S. C. and Podladchikov, Y. Y. (2014), Chimneys, channels, pathway flow or water conducting features - an explanation from numerical modelling and implications for CO2 storage, Energy Procedia, 63, 3761-3774, doi:10.1016/j.egypro.2014.11.405.Schaef, H. T., Ilton, E. S., Qafoku, O., Martin, P. F., Felmy, A. R. and Rosso, K. M. (2012), In situ XRD study of Ca2+ saturated montmorillonite (STX-1) exposed to anhydrous and wet supercritical carbon dioxide, Int J Greenh Gas Con, 6, 220-229, doi:10.1016/j.ijggc.2011.11.001.Simmenes, T. H., Hermanrud, C., Ersland, R., Georgescu, L. and Sollie, O. C. E. (2017), Relationships between bright amplitudes in overburden rocks and leakage from underlying reservoirs on the Norwegian Continental Shelf, Petrol Geosci, 23(1), 10-16, doi:10.1144/petgeo2016-093.Simmenes, T. H., Hermanrud, C., Georgescu, L., Ersland, R. A. and Sollie, O. C. (2015), Relationships between bright amplitudes in overburden rocks and leakage from underlying hydrocarbon reservoirs in the Barents Sea and western Haltenbanken areas, paper presented at Extended abstract for the 4th EAGE International Conference Fault and top seals, Almenia, Spain.Skurtveit, E., Aker, E., Soldal, M., Angeli, M. and Wang, Z. (2012), Experimental investigation of CO2 breakthrough and flow mechanisms in shale, Petrol Geosci, 18(1), 3-15, doi:10.1144/1354-079311-016.Skurtveit, E., Torabi, A., Alikarami, R. and Braathen, A. (2015), Fault baffle to conduit developments: reactivation and calcite cementation of deformation band fault in aeolian sandstone, Petrol Geosci, 21(1), 3-16, doi:10.1144/petgeo2014-031.Skurtveit, E., Torabi, A., Gabrielsen, R. H. and Zoback, M. D. (2013), Experimental investigation of deformation mechanisms during shear-enhanced compaction in poorly lithified sandstone and sand, J Geophys Res-Sol Ea, 118(8), 4083-4100, doi:10.1002/jgrb.50342.Sone, H., and Zoback, M. D. (2014), Time-dependent deformation of shale gas reservoir rocks and its long-termeffect on the in situ state of stress, Int J Rock Mech Min, 69, 120-132, doi:10.1016/j.ijrmms.2014.04.002.Tangen, G., Lindeberg, E., Nøttvedt, A., Eggen, S. (2014) Large-scale storage of CO2 on the Norwegian shelf Enabling CCS readiness in Europe. Energy Proc.51, 326-333. Tasianas, A., Mahl, L., Darcis, M., Buenz, S. and Class, H. (2016), Simulating seismic chimney structures as potential vertical migration pathways for CO2 in the Snohvit area, SW Barents Sea: model challenges and outcomes, Environ Earth Sci, 75(6), doi:10.1007/S12665-016-5500-1.Tzschichholz, F., and Wangen, M. (1998), Modelling of hydraulic fracturing of porous materials, in Fracture of Rock, edited by M. H. Aliabadi, pp. 227-260, WIT Press, Southhampton.van Noort, R., and Yarushina, V. (2016), Water and CO2 permeability of a shale sample core from Svalbard, Energy Procedia, 97, 67-74, doi:10.1016/j.egy-pro.2016.10.021.Wangen, M. (2011), Finite element modeling of hydraulic fracturing on a reservoir scale in 2D, J Petrol Sci Eng, 77(3-4), 274-285, doi:10.1016/j.petrol.2011.04.001.Wangen, M. (2013), Finite element modeling of hydraulic fracturing in 3D, Computational Geosciences, 17(4), 647-659, doi:10.1007/s10596-013-9346-2.Wangen, M., Gasda, S. and Bjornara, T. (2016), Geomechanical consequences of large-scale fluid storage in the Utsira Formation in the North Sea, Enrgy Proced, 97, 486-493, doi:10.1016/j.egypro.2016.10.056.Wangen, M., and Yarushina, V. M. (2015), A numerical model of hydraulic fracturing observed during injection testing of a well in Adventdalen, Svalbard, in 77th EAGE Conference and Exhibition, edited, Madrid, Spain, doi: 10.3997/2214-4609.201412791 Yarushina, V. M., and Bercovici, D. (2014), Reservoir compaction and fluid leakoff during fluid injection operations, in Fourth EAGE CO2 Geological Storage Workshop edited, Stavanger, Norway, doi:10.3997/2214-4609.20140115 Yarushina, V. M., Bercovici, D. and Oristaglio, M. L. (2013), Rock deformation models and fluid leak-off in hydraulic fracturing, Geophys J Int, 194(3), 1514-1526, doi:10.1093/Gji/Ggt199.Yarushina, V. M., and Podladchikov, Y. Y. (2015), (De)compaction of porous viscoelastoplastic media: Model formulation, Journal of Geophysical Research Solid Earth, 120, doi:10.1002/2014JB011258.Yarushina, V. M., Podladchikov, Y. Y. and Connolly, J. A. D. (2015), (De)compaction of porous viscoelastoplastic media: Solitary porosity waves, Journal of Geophysical Research Solid Earth, 120, doi:10.1002/2014JB011260.Zhang, R., Ning, Z. F., Yang, F.,Wang, X., Zhao, H. W. and Wang, Q. (2015), Impacts of nanopore structure and elastic properties on stress-dependent perme-ability of gas shales, J Nat Gas Sci Eng, 26, 1663-1672, doi:10.1016/j.jngse.2015.02.001.Zhang, R., Ning, Z. F., Yang, F., Zhao, H. W. and Wang, Q. (2016), A laboratory study of the porosity-permeability relationships of shale and sandstone under effective stress, Int J Rock Mech Min, 81, 19-27, doi:10.1016/j.ijrmms.2015.11.006.

Postal AddressFME-SUCCESS

Christian Michelsen Research ASP.O. Box 6031

NO-5892 Bergen, Norway

Visiting AddressChristian Michelsen Research AS

Fantoftvegen 38Bergen, Norway

Contact info [email protected]

[email protected]

www.fme-success.no

Contributing partners 2010 – 2018

This work was funded by the FME SUCCESS center for CO2 storage under grant no. 193825/S60 from the Research Council of Norway.

The FME SUCCESS center is a consortium of partners from industry and science, hosted by Christian Michelsen Research AS.

Documents

Potential leakage mechanisms and their relevance … 2-Potential leakage...Potential leakage mechanisms and their relevance to CO 2 storage site risk assessment and safe operations