Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

A review of CO2 storage in geological formations emphasizing modeling, monitoring and capacity estimation approaches

A review of CO2 storage in geological formations emphasizing modeling, monitoring and capacity... The merits of C O capture and storage to the environmental stability of our world should not be underestimated as emissions of greenhouse gases cause serious problems. It represents the only technology that might rid our atmosphere of the main anthropogenic gas while allowing for the continuous use of the fossil fuels which still power today’s world. Underground storage of C O involves the injection of C O into suitable geological formations and the monitoring of the injected plume 2 2 over time, to ensure containment. Over the last two or three decades, attention has been paid to technology developments of carbon capture and sequestration. Therefore, it is high time to look at the research done so far. In this regard, a high-level review article is required to provide an overview of the status of carbon capture and sequestration research. This article presents a review of C O storage technologies which includes a background of essential concepts in storage, the physical processes involved, modeling procedures and simulators used, capacity estimation, measuring monitoring and verification techniques, risks and challenges involved and field-/pilot-scale projects. It is expected that the present review paper will help the researchers to gain a quick knowledge of CO sequestration for future research in this field. Keywords CO storage · Geological formation · Modeling for CO storage · Mechanism of CO storage · CO storage 2 2 2 2 projects 1 Introduction are greener technologies such as nuclear energy and wind energy which reduce the combustion of fossil fuels associ- The global warming scourge is threatening to ravage human- ated with emission sources and energy efficiency. The con- ity. Rising sea levels, increases in average global air and sea tinued need for fossil fuels across the world and the rela- surface temperatures, widespread snow and ice melting are tively slow pace of renewable energy development suggests notable effects of global warming (IPCC 2007). The implica- that the amount of undesired different gases being emitted tion of these indicators in the long run on health, nutrition into the atmosphere will remain on the increase. It is impera- and the economy can be ill-afforded and therefore has been tive, therefore, the ways should be developed in which these the subject of a great deal of research to date. Numerous harmful gases can be expunged from the atmosphere. strategies have been employed or are under intense scru- Greenhouse gases, a term for the climate-unfriendly gases tiny as a means of tackling climate change, some of which emitted into the atmosphere, provide a threat to our ecosys- tem with CO accounting for 82% of greenhouse gases in the atmosphere. Though the global warming potential (GWP) of Edited by Yan-Hua Sun CO is less than other greenhouse gases (US Environmental * Achinta Bera Protection Agency 2014), the sheer amount of CO being achintachm@gmail.com emitted into the atmosphere makes it the most significant of all greenhouse gases for efficient climate control. Department of Petroleum Engineering, Khalifa University The advent, development and implementation of carbon of Science and Technology, Sas Al Nakhl Campus, P.O. Box 2533, Abu Dhabi, UAE dioxide capture, utilization and storage (CCUS) technology promises to reduce the amount of greenhouse gases enter- Drilling, Cementing, and Stimulation Research Center, School of Petroleum Technology, Pandit Deendayal ing the atmosphere. CCUS encompasses the capture of car- Petroleum University, Raisan, Gandhinagar, Gujarat 382007, bon dioxide and its associated compounds from producing India Vol:.(1234567890) 1 3 Petroleum Science (2019) 16:1028–1063 1029 sources, compression, transportation and the utilization of The technology is currently at the research stage without the captured C O for processes such as injection into deep any existing pilot tests. underground geological formations for permanent storage C. Geological sequestration is the most widely used seques- and injection into existing oil fields for additional recovery tration technology. In this process, CO is stored in geo- of hydrocarbons. logical underground structures such as saline aquifers, Some previous review articles summarized the different depleted oil and gas reservoirs and unmineable coal beds physicochemical methods responsible for suitable CO stor- (IPCC 2007; Kaldi et al. 2009; Metz 2005; Pashin and age and the difficulties in different aspects (Riaz and Cinar Dodge 2010). A short description of all storage sites is 2014; Belhaj and Bera 2017; Aminu et al. 2017; Thakur given below: et al. 2018). The main motivation of this review paper is to 1. Saline aquifer formations: Saline aquifer forma- present all aspects of CCUS projects worldwide along with tions represent the best salted sink for storage of the technologies, modeling issues and physicochemical pro- CO among all geological options due to their enor- cesses occurred during the C O sequestration within geolog- mous storage capacity (Grobe et al. 2009). Recently, ical formation. This review will serve as a single handbook estimates of the order of 103 Gt C O have been for understanding CCUS and to provide researchers the facts made for the Alberta deep saline basin by account- about CCUS in the oil industry. C O flooding for enhanced ing for the solubility trapping mechanism (Bachu oil recovery is one of the effective methods in additional oil and Adams 2003). Another example is the injection recovery. The injected carbon dioxide can be stored in the of the produced CO into the Utsira aquifer in the formation of the reservoir. Therefore, it is important to know North Sea (Korbøl and Kaddour 1995; Torp and the rock capacity and power to store the carbon dioxide for Gale 2004). It is required that the aquifer be saline a long time. because this already makes it unsuitable for indus- Storage of CO has been employed in different parts of trial, agricultural and human purposes (Aydin et al. the world. The modes of storage can be broadly classified 2010; Metz et al. 2005). into natural and man-made modes of storage. Natural modes Other storage modes which have been employed include terrestrial sequestration, while man-made storage for the storage of CO include basalts (Gislason includes storage in geologic formations. Several modes for and Oelkers 2014) and mineral carbonation (Oelk- utilizing and storing CO have been explored as follows: ers et al. 2008). Among all geologic sequestration mechanisms, deep saline aquifers represent the A. Terrestrial sequestration is the capture of C O from ones exhibiting highest sequestering capability, as the atmosphere and storing it into soils and vegetation. against those provided by depleted oil and gas res- Removal of CO from the atmosphere through photo- ervoirs and unmineable coal beds (IPCC 2007; Torp synthesis and prevention of the emission of C O from and Gale 2004; Kaldi et al. 2009; Parson and Keith terrestrial sources are the mechanisms for terrestrial 1998). storage. It has been postulated to provide an important 2. Depleted oil and gas reservoirs: Previously produc- mechanism for the storage of carbon dioxide (Litynski ing oil and gas fields which have been considered et al. 2006; Thomson et al. 2008). uneconomical for further production of hydrocar- B. Ocean sequestration qualifies as the largest possible sink bons are suitable candidates for geological seques- for carbon dioxide storage with an estimated potential tration. Characteristics required for a storage site are storage of 40,000 gigatonnes (Gt) of CO (Herzog et al. present in such formations and have been employed 1997, 2000; Lal 2008) and the possibility of storing over for geologic sequestration. An important advan- 90% of current CO emissions. It involves the injection tage is that they have been adequately character- and deposition of CO into the water body at depths ized previously. Additionally, the safe and secure below 1 km either from moving ships, fixed pipelines nature of these formations which have been able to or offshore platforms. At this depth, water has a lower store oil and gas over a long period of time makes density than the injected CO and the latter is expected them prime candidates. Existing numerical com- to dissolve and disperse into the water body (Metz et al. puter models of such formations which have been 2005). However, there are huge concerns over the envi- history-matched provide improved confidence in ronmental impact of CO on marine life from the acidity the formations. Infrastructures and wells used in the of seawater near the injection point (Seibel and Walsh development of these fields are also available for 2001). The scalability of experiments involved in ocean CO injection. Storage capacity available in depleted sequestration is also very difficult, thus requiring expen- reservoirs is significantly lower due to the need to sive field experiments (Adams et al. 1998a, b; Auerbach avoid exceeding pressures that can damage the cap et al. 1997; Herzog et al. 1997; Seibel and Walsh 2003). 1 3 1030 Petroleum Science (2019) 16:1028–1063 rock and the significant leakage threat posed by the China (Liang et al. 2009) and different parts of the abandoned wells. (A potential for leaks exists behind world for simultaneous EOR and storage processes well casings.) (Ghomian et al. 2008; Gozalpour et al. 2005; Liu 3. Deep unmineable coal beds: CO has been employed et al. 2013; Moritis 2000; Narinesingh et al. 2014). for the recovery of methane from coal seams during the enhanced coal bed methane (ECBM) recovery This integrated review will discuss storage of CO in various process (Busch and Gensterblum 2011; Mukherjee geological formations with a focus on saline aquifers. Sec- and Misra 2018; Pan et al. 2018b). Produced meth- tion 1 contains the introductory part of the review. Section 2 ane from this source can be utilized as an energy discusses the properties of the gas which favors storage as source. Coal beds have very large fracture networks well as trapping mechanisms and the physical processes through which gas molecules can diffuse into the involved in the storage process. Section 3 gives a summary matrix and desorb tightly adsorbed methane. C O of the pilot- and commercial-scale projects which are in the has been proven to raise methane recovery to about planning phase, in operation or have been abandoned. In 90% from 50% when conventional methods are Sect. 4, we discuss the modeling strategies for CO which applied. Injected C O is stored in the formations have been applied in the literature. Section 5 covers the esti- after methane has been recovered. Storage in coal mation methods for storage capacities. In Sect. 6, an over- beds can take place at shallower depths than other view of the measuring, monitoring and verification tools and formation types and as such relies on CO adsorp- challenges is provided. Section 7 reports the risks and chal- tion on the coal surface. However, the technical fea- lenges that may be present before commercial application of sibility of this storage process strongly depends on field-scale projects. Finally, conclusions and recommenda- the coal’s permeability as a result of its depth vari- tions are provided in Sect. 8. It is expected that the entire ation with the influence of effective stress on coal manuscript will provide an overview of CCUS issues of past, fractures (Metz et al. 2005). present and future challenges for newcomers in this field. The laboratory and field testing feasibility of commercial CO injection into coal beds and seams has been reported in the San Juan Basin, which 2 CO storage in saline aquifers is the world’s first ECBM project (Reeves 2001). Other enhanced coal bed methane recovery projects 2.1 Conditions required for storage sites reported in the world for laboratory and field testing include the Sydney Basin in Australia (Saghafi et al. The selection of a geological site for storage must be done 2007) and deep coalbed methane in Alberta Canada to meet three main conditions: capacity, injectivity and (Gunter et al. 1997). containment. The requirement of the capacity of a storage 4. CO storage during enhanced oil recovery: CO is site ensures that the selected site possesses adequate pore 2 2 used for enhanced oil recovery (EOR) from mature volumes to store large amounts of C O . Typical conditions fields. CO for EOR operations has been employed would mean that the site should contain significant porosity in the miscible and immiscible states. When injected and/or occupy a very large area. Injectivity of C O is assured into oil, CO has the capability to swell the oil, if the candidate formation possesses high permeability reduce its viscosity and reduce interfacial tension ensuring that lower wellhead pressures can be used to main- and in some cases become miscible with the oil tain desired injection rates. Competent cap rocks and sealing allowing for single-phase flow. Of the two miscible faults (if present) are necessary to ensure that the injected states for EOR via CO injection, miscibility of CO CO does not escape to the surface or leak into groundwa- 2 2 2 in oil usually provides higher recoveries. The abil- ter due to the lower density of the C O gas compared with ity of CO to become miscible in oil is determined resident brine. For successful storage of carbon dioxide, it by the minimum miscibility pressure (MMP). At is required that C O be stored in a supercritical phase, the and above this pressure, CO is miscible in oil and state in which C O exists when it is compressed to higher 2 2 below, it is immiscible. Though C O injection in pressures and temperatures (about 89 °F and 7.4 MPa). In this process is done primarily for EOR, it comes this phase, C O possesses properties of a liquid but flows as with the added benefit of storage of CO contribut- a gas. Essentially, CO is required to be stored at this state 2 2 ing to minimizing the global warming scourge. Over due to its higher density, reducing the buoyancy differential the last decade, CO has been used in over 70 EOR between CO and in situ fluids (Grobe et al. 2009; Kane 2 2 operations around the world with over 40 reported and Klein 2002; Koide et al. 1992). Though the density of in West Texas (Moritis 2000), Weyburn Field in CO is higher when injected underground, it remains signifi- Canada (Malik and Islam 2000), Shengli Oilfield in cantly lower than the density of in situ brine which lies in 1 3 Petroleum Science (2019) 16:1028–1063 1031 the region of 1200–2000 kg/m depending on the salinity of underground in the long term is essential. During the injec- the brine. The implication of this density differential is the tion process in the targeted formation, viscous forces are the buoyant movement of CO when injected underground and dominant forces for the migration of C O. CO is then stored 2 2 2 thus demanding the presence of low-permeability cap rocks in either the supercritical or the gas phase as a function of which overlay the aquifer. depth at the associated pressure and temperature. Once the injection stops, the supercritical CO tends to migrate 2.2 Trapping mechanisms upward through the porous and permeable rock as a result of the buoyancy effect created by its density difference com- The storage capacity, containment and injectivity of CO pared to other reservoir fluids and laterally via preferential are dependent on the geological and petrophysical proper- pathways until a cap rock, fault or other sealed discontinuity ties of the target formation. The injected supercritical CO is reached (Han 2008). This will prevent further migration is securely trapped underground via two major trapping of the CO as shown in Fig. 2. In depleted oil and gas fields, mechanisms (physical trapping and geochemical trapping) the movement of the CO can also be halted by abandoned (Fig. 1). The effectiveness of the storage process is governed wells sealed with solid cement plugs. The risk associated by a combination of both trapping mechanisms to ensure with such trapping is leakages behind casing or through long-term storage (Coninck et al. 2005). the mentioned plugs. Thus, many studies have been con- ducted on the leakage of CO through geological structures 2.2.1 Physical trapping and existing wells (Ambrose et  al. 2017; Eke et  al. 2011; Lewicki et al. 2007; Scherer et al. 2015; Shipton et al. 2004, Physical trapping is the process where C O maintains its 2006; Temitope and Gupta 2019; Zakrisson et al. 2008). physical nature after injection into an aquifer. It can be sub- divided into structural (hydrostratigraphic) and residual 2.2.1.2 Residual/capillary trapping As supercritical C O (capillary) trapping. Generally, the time period for physical percolates through storage formations, reservoir fluids are trapping is believed to be less than a century (Juanes et al. displaced. The movement of the C O occurs in two direc- 2006). tions: upward as a result of density differences and later - ally due to viscous forces. Reservoir fluid fills the spots left. 2.2.1.1 Structural trapping Structural trapping is usually However, some of the CO is left behind as disconnected/ the first form of trapping encountered during geological residual droplets in the pore spaces as displayed in Fig. 3. sequestration, and a similar mechanism has kept oil and Surface tension between CO and brine acts to halt the gas securely stored underground for millennia. Geologi- CO movement, thereby causing higher capillary entry pres- cal structures such as anticlines covered with cap rocks sure than the average rock pressure as suggested by Saadat- (an ultra-low-permeability layer), stratigraphic traps with/ poor et al. (2010). At this point, C O becomes immobilized without sealed faults are employed for the storage of C O in the pores at residual gas saturation. It is usually observed as a mobile phase or supercritical fluid. Maximization of in rocks with small-scale capillary heterogeneities. Recent this storage mechanism to ensure that C O injected remains studies have revealed that capillary trapping appears to be Structural (hydro stratigraphic) trapping Physical trapping Residual (capillary) trapping < 100 years Sorption trapping CO trapping mechanisms Slow diffusion in aqueous phase Solubility trapping Convective processes Geochemical trapping Mineral trapping Reaction with minerals Fig. 1 Different CO trapping mechanisms during the geological storage process 1 3 1032 Petroleum Science (2019) 16:1028–1063 Buoyant CO plume trapped by the seal (cap rock) Injection well Cap rock Porous media (aquifer) CO makes its way to the top of the aquifer Buoyant CO plume trapped by sealing fault Sealing fault Fig. 2 Physical trapping of injected C O as a result of the formation structure Cap rock2.2.2 Geochemical trapping CO gets trapped in the pore throats of the Geochemical trapping occurs when C O changes its physical porous media, as it makes and chemical nature by undergoing series of geochemical its way to the cap rock reactions with the formation brine and the rock and ceases to remain in the mobile or immobile phase. This interaction ensures the disappearance of CO as a separate phase and further increases storage capacity, making this an appropri- Grain ate feature of long-term storage. 2.2.2.1 Solubility trapping In a similar manner by which Porosity filled with water sugar dissolves in tea, C O dissolves in other fluids in either the supercritical or gaseous phase. Solubility trapping occurs as a result of the dissolution of the CO in the brine, Fig. 3 Residual trapping of injected C O as a result of the formation leading to dense CO -saturated brine. At this point, it ceases pore structure. Arrows in the diagram indicate the movement of the to remain a separate phase which eliminates any buoyancy CO plume effect. Over time, CO -saturated brine becomes denser than the surrounding reservoir fluids and falls to the bottom of a more efficient mechanism to trap CO underground in the the formation over time, culminating in more secure C O 2 2 short term compared to other short-term trapping mecha- trapping (Fig. 4). nisms (Burnside and Naylor 2014; Lamy et al. 2010). Its The dissolution of CO in the aqueous phase leads to the efficiency is due to exhibition of higher capillary forces to formation of weak carbonic acid which decomposes over + − buoyant forces, causing CO to appear as pore-scale bubbles time into H and HCO ions (Eq. 1). It can also react with 2 3 rather than being retained by a somewhat compromised cap other cations in the formation brines to form insoluble ionic rock. Furthermore, it provides an advantage of no risk of species as highlighted in Eqs. 1–4. CO solubility in forma- major failure associated with structural traps over a short tion water decreases as temperature and salinity increase. time scale (Jalil et al. 2012). + − CO + H O ↔ H + HCO 2 aq 2 (1) ( ) 1 3 Petroleum Science (2019) 16:1028–1063 1033 Cap rock Brine saturated with CO Convection Convection CO drops into the aquifer Convection Porous media (aquifer) Fig. 4 Pictorial representation of solubility trapping via convective mixing, one of the mechanisms for the dissolution of CO into aquifers modeling of these reactions is critical to the success of 2+ + CO sequestration predictions. This trapping mechanism Ca + CO + H O ↔ H + CaHCO (2) 2(aq) 2 3(aq) 2 is dependent on the rock minerals, the pressure of the gas, + + temperature and porosity and has been found to produce Na + CO + H O ↔ H + NaHCO (3) 2(aq) 2 3(aq) significant changes in the rock permeability and porosity (Benson and Cole 2008; Kampman et al. 2014). Perkins 2+ + Mg + 2CO + 2H O ↔ 2H + Mg(HCO ) . (4) 2(aq) 2 3 2(aq) et al. (2004) predicted from a simulation study that all the CO injected into the Weyburn Oil Field will be converted 2.2.2.2 Mineral trapping Mineral trapping occurs as a to carbon dioxide minerals after 5000 years. They reported greater mineralization capacity for the cap rock and overly- result of the conversion of CO into calcite due to reactions with solid minerals. This trapping is believed to be relatively ing formation rock, which is quite significant for leakage risk assessment. The capacity is estimated based on the amount slow since it occurs during/after solubility trapping and con- sidered as the most permanent form of storage. C O in the of minerals available for carbon dioxide precipitation and the quantity of C O used in the reaction processes. The most aqueous phase forms a weak acid which reacts with rock minerals to form bicarbonate ions with different cations striking advantage of mineral trapping mechanism over the other mechanisms is that it prevents CO from existing as a depending on the mineralogy of the formation. An example of such reaction with potassium basic silicate (Eq.  5) and separate phase, thus ensuring that its upward movement is halted and also enhances the formation of stable precipitates calcium (Eq. 6) is shown below: (Xu et al. 2001, 2003, 2004). 3K-feldspar + 2CO + 2H O ↔ Muscovite There are multiple mechanisms responsible for the stor- 2(aq) 2 (5) + − age operating simultaneously and on different time scales + 6Quartz + 2K + 2HCO which influence the storage capacity estimate. The interac- tion between various mechanisms is quite complex, evolves 2+ + Ca + CO + H O ↔ Calcite + 2H 2(aq) 2 (6) with time and depends highly on local conditions. An exam- ple of time scale evolution of different mechanisms at play Precipitation of carbon dioxide minerals is invariably in a deep saline formation is as shown in Fig. 5. induced by reactions with the rock formations depending on the mineralogy of these formations. Hence, geochemical 1 3 1034 Petroleum Science (2019) 16:1028–1063 The drainage and imbibition-like processes during the 2.3 Physical processes during  CO storage injection and post-injection stages of CO storage lead to hysteresis, a process where the capillary pressure and rela- A number of physical processes are involved in the injection and post-injection phases of carbon dioxide. CO trapping in tive permeability curves change pathways. This phenomenon has been described as being very critical to the successful aquifers is aided by three physical processes buoyancy (grav- ity), viscous forces and capillary forces (Kong et al. 2013). modeling of C O trapping processes (Ghomian et al. 2008; Juanes et al. 2006; Spiteri et al. 2005). This is because as the During the injection period of C O into aquifers, viscous forces are the dominant forces for the vertical and lateral CO migrates upward after the injection phase, the remain- ing CO plume gets disconnected due to water displacing migration of C O due to pressure gradients created by the 2 2 injection processes. The injected fluid (CO ) displaces the CO at the trailing edge and becomes a series of blobs. C O 2 2 is trapped in these blobs, and the mechanism is termed resid- formation fluid (brine) in a drainage-like process. In the post-injection phase, a combination of buoyancy ual or capillary trapping mechanism, which over time results in the dissolution of the CO in the formation brine. and capillary forces are responsible for the trapping of CO . 2 2 Buoyancy forces are usually greater than capillary forces and Heterogeneity and wettability of the aquifer are also key considerations in this mechanism. Heterogeneity has been viscous forces after injection in deep saline aquifers, leading to upward migration of C O . Buoyancy results from density subdivided into the small and large scales (Gershenzon et al. 2014; Lasseter et al. 1986; Li and Benson 2014). Viscous and differences between the injected CO and the aquifer brine causing the CO to migrate upward after injection displacing capillary forces dominate the flow, while gravity forces are generally regarded as unimportant when small-scale hetero- water in an imbibition-like process. The upward migration leads to gravity segregation, and geneities are considered. When large-scale heterogeneity is considered, the formation possesses variable pore throat sizes, further migration to the surface is prevented by the ultra- low permeable seal at the formation top. Once reaching which are likened to different capillary tubes sizes. As a result, a variable amount of entry capillary pressure is required to the top of the formation, the vertical migration is halted, while the lateral migration continues until a sealing fault displace the formation fluid. This leads to more CO being trapped as the entry pressure is overcome. Wettability and or formation boundary is reached. Thorough geomechani- cal analysis has to be made to ensure that leakage of C O interfacial tension changes have been proven to alter the capil- lary pressures in a porous medium (Bennion and Bachu, 2006; does not occur when the buoyant CO reaches the seal. One means of leakage is when the pressure of the C O is high Chiquet et al. 2007; Jung and Wan 2012; Park et al. 2015; Yang et al. 2005). The basic definition of capillary pressure enough to overcome the entry pressure of the seal (Hesse et al. 2006). Others could be due to the cap rock fractures, (Eqs. 7 and 8) and Young–Laplace equation (9) can be shown as follows in terms of mathematical forms: thermal stresses in the caprock as a result of temperature variation between the injected CO and aquifer and the pres- P = P − P (7) c nw w ence of open faults, fractures and abandoned wells (Chiquet et al. 2007; Goodarzi et al. 2013). Geomechanical considera- P = = (8) tions involving cap rock integrity are one of the factors that d d affect the sequestering capacity of the overlying seal. 2 cos w,CO P = P − P = (9) c CO w where d is diameter; R is the pore throat radius; P is defined Injection as the capillary pressure; P and P are the pressures of nw w Trap filling the non-wetting and wetting phases, respectively; P is CO Physical trapping the pressure of C O ;  is the water surface tension;  is the 2 w interfacial tension;  is the interfacial tension between Dissolution w,CO water and CO , and θ is the contact angle between the wet- Residual CO trapping ting medium and the rock surface. Mineralization In a typical CO –water system, CO is usually described 2 2 Adsorption as the non-wetting phase, while water is the wetting phase; 1 2 3 4 5 6 however, it has been proven that during the C O upward 110 10 10 10 10 10 migration, this wetting state can be changed (Broseta et al. Time, years 2012; Chiquet et al. 2007; Marckmann et al. 2003; Siemons et al. 2006; Yang et al. 2005). Equation 9 shows that the cap- Fig. 5 Time frame for trapping mechanisms in deep saline formations during and after injection (IPCC 2007; Metz et al. 2005) illary pressure is dependent on the pore throat radius, R, the 1 3 Petroleum Science (2019) 16:1028–1063 1035 interfacial tensions (  ) and the contact angles (θ) between storage and on the effective monitoring tools which could be the wetting medium and the rock surface. Therefore, the used for large-scale injections. These projects can be broadly interfacial tensions and wettability have a significant impact classified according to the storage location of the different on the sequestration capabilities of aquifer rocks. projects (saline, EOR, depleted gas reservoirs, ECBM), During the residence time of trapped CO in the blobs based on the mode of capture of the carbon dioxide (power and ganglia, C O dissolves into brine and this dissolution plants CCS projects and non-power plant CCS projects) and has been proven to occur by three principal mechanisms. based on the current status of the projects (planned, ongoing They are (a) diffusion of CO within the aqueous phase, and completed CCS projects). (b) reactions with the host minerals (classified as mineral The Sleipner project in Norway is the first case of large- trapping) and (c) convective mixing driven by slight density scale commercial C O storage in the world (Torp and Gale differences between the water saturated with CO and the 2004). The project began in 1996 and injected about a mil- unsaturated water (Ennis-King and Paterson 2003; Hassan- lion tons of CO into the sands of the Utsira Formation zadeh et al. 2007). Ennis-King and Paterson (2003) stated which is about 900 m below the bottom of the North Sea. that the dominant mechanism for long-term dissolution of The major incentive behind the commencement of the Sleip- CO in the formation brine is convective mixing rather than ner project was the need for minimization of taxes placed on pure diffusion as it is in orders of magnitude faster than dif- the direct emission of C O into the atmosphere (Christian- fusion and chemical reaction with the host mineral. sen 2001; Global CCS 2017; Kongsjorden et al. 1998). The The disproportionate dissolution of CO in brine leads to companies involved were faced with the options of paying gravitational instabilities which could further aid in solubil- heavy taxes for atmospheric emissions or injecting the CO ity trapping. Several researchers have worked on trying to into saline aquifers. Injection of C O into saline aquifers determine the onset time of convective mixing and the influ- provided a beneficial means for cost reduction by the par - encing factors (Bestehorn and Firoozabadi 2012; Ennis-King ties involved. Policies such as carbon dioxide pricing which and Paterson 2003; Hassanzadeh et al. 2007; Rasmusson would coerce companies with high CO emissions into con- et al. 2015; Riaz et al. 2006; Xu et al. 2006b). Ennis-King sidering the need for C O storage are major ways to ensure and Paterson (2003) used a linear stability analysis technique emissions into the atmosphere are significantly reduced. to provide an estimate of the time required for convective Another incentive for CO storage is the low cost of captur- instability to begin. They predicted the time to be typically ing; this has especially been noticed in the current field-scale up to tens of years, and this method has been used by sev- projects where CO injected was obtained from the separa- eral other researchers (Hassanzadeh et al. 2006; Hesse et al. tion of CO from produced gases, thus reducing the need 2006; Riaz et al. 2006). Riaz et al. (2006) determined the for capturing from coal plants which have not undergone critical time and wavelength or the onset of convective mix- separation and would cost more to capture from such plants. ing using the method of linear stability. It was determined The high cost of capturing CO from combustion processes that the critical time varies between 2000 years and 10 days has triggered the idea of carbon dioxide capture utilization and the critical wavelength varies between 200 and 0.3 m and storage (CCUS) where the CO could also be used for for a permeability variation of 1–3000 mD. Rasmusson et al. enhanced oil recovery and revenue derived from the pro- (2015) applied the Rayleigh number (Ra) in determining duced oil could be used to offset the cost of capturing and the onset of gravity-driven instabilities. They predicted that injecting into formations. The success of the Sleipner project a prerequisite for Ra, which must be greater than a critical elicited the increased field deployments on CO storage. Ra, is required for the onset of density-driven instabilities. Several pilot-scale projects have also been implemented Finally, as C O remains dissolved in the brine, it forms weak across the world. These projects typically inject small acids which react with the host minerals to form precipitates amounts of CO into identified formations for a small period (Gunter et al. 2000; Kumar et al. 2005; Xu et al. 2001). of time. These projects provide answers to questions of inter- est to the investigators. The first pilot-scale project in the USA was the Frio Project where about 1600 tons of CO was 3 Field‑scale projects on  CO storage injected at a depth of about 1500 m below the surface for a period of 10 days (Hovorka et al. 2006). The Frio Project CO sequestration projects are currently ongoing or in the provided information about the movement of CO plume and 2 2 planning stage across the world. Notable among these are the was able to validate numerical models developed to analyze Sleipner project in Norway, the Weyburn Project in Canada subsurface CO migration. Other notable pilot-scale projects and the In Salah Project in Algeria. Tables 1, 2, 3 and 4 pre- are the Cranfield Project (Hosseini et al. 2013; Hovorka et al. sent the lists of most of the projects. These field-scale injec - 2013), Decatur Project (Finley 2014; Senel et al. 2014), Ket- tions of CO into candidate formations have provided more zin site in Germany (Kempka and Kuhn 2013; Martens et al. insight into the physics of the processes involved in geologic 1 3 1036 Petroleum Science (2019) 16:1028–1063 1 3 Table 1 Storage projects across the world: saline aquifer projects Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Captain UK Captain Clean Energy 3.8 Mt/year Power Pipeline to offshore deep Planning James (2013) Limited (CCEL) owned by saline formations Summit Power and CO Deep Store Don Valley UK 2Co Energy Ltd, Samsung 4.9 Mt/year Power Pipeline for sequestration Planning Construction & Trading, in offshore deep saline BOC formations Killing Holme UK C.GEN NV and National 2.5 Mt/year Power Offshore pipeline for storage Planning Grid in deep saline aquifers Korea CCS Korea Korea Carbon Capture and 1 Mt/year Power Ulleung deep saline forma- Planning Lee et al. 2012 Sequestration R&D Center tion or the Gorae gas (KCRC) reservoir Lianyungang China Summit Power, National 1 Mt/year Power Binhai for saline aquifers or Planning Pang et al. (2012) and Qiao Grid, CO Deep Store North Jiangsu oilfields for et al. (2012) EOR Longyearbyen NorwayUNIS CO Lab-AS Not available Power Onshore storage in a saline Planning Braathen et al. (2012) aquifer White Rose UK Capture Power Limited, the 2 Mt/year Power Pipeline to offshore storage Planning Verdon (2014) consortium of Alstom UK in a saline aquifer Limited, Drax Power Lim- ited and National Grid plc Cranfield USA SECARB 1–1.5 Mt/year Non-power Saline reservoir, Tuscaloosa In operation since 2009 Lu et al. (2012) Sandstone Formation, down dip of the mature Cranfield Oil Field Citronelle USA SECARB, Denbury, South- 0.25 Mt/year Non-power The southern flank of the In operation since 2011 Haghighat et al. (2013) ern Energy Citronelle dome Decatur USA Archer Daniels Midland, 1 Mt/year Non-power Sequestration in Mount In operation since 2011 Zhou et al. (2010) MGSC (Led by Illinois Simon sandstone State Geological Survey), Schlumberger Carbon Services and Richland Community College Kevin Dome  USA Big Sky Partnership, 0.125 Mt/year Non-power The Duperow Formation Planning Riding and Rochelle (2005) Schlumberger Carbon (3900 ft) Services, Vecta Oil & Gas Ltd, Lawrence Berkeley National Lab, Los Alamos National Lab Petroleum Science (2019) 16:1028–1063 1037 1 3 Table 1 (continued) Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Wasatch Plateau USA  South West Partnership 1 Mt/year Non-power The Jurassic Entrada Forma- Planning Parry et al. (2007) (SWP), New Mexico tion and Navajo sandstone Institute of Mining and Technology, University of Utah, Schlumberger and Los Alamos National Laboratory Fort Nelson CanadaPlains CO Reduction Part- 2.2 Mt/year Non-power Middle Devonian carbonate Planning nership (PCOR), Spectra rock Energy, British Columbia Ministry of Energy, Mines and Petroleum Resources Quest Canada Athabasca Oil Sands Project: 1.1 Mt/year Non-power Injection into the Cambrian Launched in November Brydie et al. (2014) Shell Canada, Chevron Basal Sands 2015 Canada, and Marathon Oil Sands Sleipner Norway StatOil 0.9 Mt/year Non-power Utsira Formation In operation since 1996 Chadwick et al. (2004) Ketzin Germany GFZ German Research 0.06 Mt/year Non-power Stuttgart sandstone reservoir In operation since 2008 Schilling et al. (2009) Centre for Geosciences and Ketzin partners Snohvit Norway Statoil ASA, Petoro AS 0.7 Mt/year Non-power Saline Tubaen sandstone In operation since 2007 Hansen et al. (2013) (Norwegian state direct formation reservoirs interest), Total E&P Norge AS, GDF Suez E&P Norge AS, Norsk Hydro, Hess Norge ULCOS Florange France ArcelorMittal and ULCOS 0.7–1.2 Mt/year Non-power Onshore deep saline forma- On hold Global CCS (2017) (Ultra-Low-CO2-Steel) tions Ordos China Shenhua Group 1 Mt/year Non-power EOR/saline aquifer In operation since 2010 Li et al. (2016) Gorgon Australia Gorgon Joint Venture (Chev- 3.4–4.0 Mt/year Non-power Dupuy Formation 2.5 km Under construction Flett et al. (2009) ron Australia, ExxonMobil, below Barrow Island Shell, Tokyo Gas, Osaka Gas and Chubu Electric) Yulin China Shenhua Group, Dow 2–3 Mt/year Non-power Onshore deep saline aquifers Planning Li-ping et al. (2015) Chemicals Minami-Nagaka Japan EnCana, IEA 0.015  Mt/year Non-power Haizume Formation In operation since 2002 Zwingmann et al. (2005) Frio USA Bureau of Economic Geol- 177 t/day Non-power Frio Formation In operation since 2004 Hovorka et al. (2006) ogy of the University of Texas Teapot Dome USA Rocky Mountain Oilfield 170 t/day Non-power Tensleep and Red Peak In operation since 2006 Friedmann and Stamp (2006) Testing Center (RMOTC) Formation 1038 Petroleum Science (2019) 16:1028–1063 1 3 Table 2 Storage projects across the world: C O EOR/storage projects Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Santos Basin Brazil Petrobras, BG E&P 1 Mt/year Non-power Lula and Sapinhoá oil In operation since 2013 Saini (2017) Brasil Ltda, Petrogal fields Brasil Boundary Dam Canada SaskPower 1 Mt/year Power EOR in Weyburn field, October 2014-date Stéphenne (2014) excess to be used at the Aquistore project Bow City Canada Bow City Power, Can- 1 Mt/year Power Pipeline for EOR Planning Global CCS (2017) solv (Subsidiary of Shell), Luscar, Fluor Pembina Canada Penn West 50 t/day Non-power Cardium Formation In operation since 2005 1 Mt/year Non-power EOR in 2 carbonate In operation since 2000 White (2009) Weyburn-Midale Canada Cenovus Energy, fields: Weyburn field Apache Canada, PTRC (Petroleum (6500 t/day) and Technology Research Midale field (1200 t/ Center) day) Zama Canada PCOR, Apache Canada 0.067 Mt/year Non-power EOR in Zama Keg In operation since 2006 Smith et al. (2009) Ltd River oil field Alberta Carbon Trunk Canada Enhance Energy Inc. 14.7 Mt/year Non-power Injection into the Clive Planning Cole and Itani (2013) Line oil reservoir Daqing China Alstom and China 1 Mt/year Power EOR in nearby fields Planning Xiuzhang (2014) Datang Corporation Dongguan China Dongguan Taiyangzhou 1 Mt/year Power EOR in Shangdong Planning Liu et al. (2016) Power Corporation, Province Xinxing Group, Nan- jing Harbin Turbine Co, KBR, Southern Company, GreenGen China GreenGen 2 Mt/year Power Onshore EOR Planning Haszeldine (2009) Shengli China Sinopec 1 Mt/year Power EOR in Shangdong Planning Liang et al. (2009) Province Uthmaniyah Saudi Arabia Saudi Aramco 0.8 Mt/year. Power Pipeline for onshore In operation Liu et al. (2012) EOR Uthmaniyah Saudi Arabia Saudi Aramco 0.8 Mt/year. Non-power Pipeline for onshore In operation since 2015 Liu et al. (2012) EOR ESI CCS Project United Arab Emirates Abu Dhabi Future 0.8 Mt/year Non-power EOR, Rumaitha Zone- Started in 2017 Temitope et al. (2016) (UAE) Energy Company B, and Bab Zone-B (Masdar) and Abu Dhabi National Oil Company (ADNOC) Petroleum Science (2019) 16:1028–1063 1039 1 3 Table 2 (continued) Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Taweelah United Arab Emirates Abu Dhabi Future 2 Mt/year Power Injection for EOR Planning (UAE) Energy Company (Masdar) and Taweelah Asia Power Company (TAPCO) and Emirates Alu- minium (EMAL) Bell Creek USA PCOR, Denbury 1 Mt/year Non-power EOR at Bell Creek oil Planning Gorecki et al. (2012) field, Montana Hydrogen Energy USA SCS Energy  3 Mt of CO captured Power Pipeline to onshore Planning Global CCS (2017) California Project annually EOR in Occidental’s (HECA) Elk Hills oil field Kemper County USA Mississippi Power, 3.5 Mt of CO annually Power Pipeline for onshore In construction Global CCS (2017) Southern Energy, EOR KBR 1 Mt/year Non-power EOR in West Hasting’s In operation since 2013 Global CCS (2017) Port Arthur USA Air Products and Chemicals, Den- and Oyster Bayou oil bury Onshore LLC, fields, Texas University of Texas Bureau of Economic Geology and Valero Energy Corporation Texas Clean Energy USA Summit Power Group 2–3 Mt/year captured Power EOR in the Permian Planning Global CCS (2017) Project (TCEP) Inc, Siemens, Fluor, Basin Linde, R.W. Beck, Blue Source and Texas Bureau of Eco- nomic Geology WA arish Petra Nova USA Petra Nova Holdings: 1.4 Mt of CO captured Power Pipeline for onshore In construction Global CCS (2017) a 50/50 partnership annually EOR in the West between NRG Energy Ranch Oil Field in and JX Nippon Oil Jackson County, & Gas Exploration Texas Corp. 1040 Petroleum Science (2019) 16:1028–1063 Table 3 Storage projects across the world: depleted reservoir projects Project name Country Company opera- Total planned Capture mode CO fate Status of project References tors storage In Salah Algeria BP, Sontrach, 1.2 Mt/year Non-power The Krechba 2004–2011 Ringrose et al. and Statoil Formation suspended (2009) Otway Australia CO CRC 0.065 Mt/year Non-power The Waarre In operation Underschultz et al. (Coopera- Formation since 2008 (2011) tive Research Center for Greenhouse Gas Technolo- gies) *ROAD (Rot- Netherlands E.ON Benelux, 1.1 Mt/year Power Pipeline for stor- Planning Global CCS terdam Opslag Electrabel, age in depleted (2017) en Afvang GDF Suez and reservoirs Demonstrate Alstom project) K12B Netherlands Gaz de France 0.365 Mt/year Non-power Rotleigendes In operation van der Meer et al. since 2004 (2006) Peterhead UK Scottish and 1 Mt/year Power Pipeline to off- Planning Southern shore depleted Energy (SSE) Goldeneye gas and Shell reservoir Northern Reef USA Midwest 0.365 Mt/year Non-power Depleted oil field In operation Trend Regional Car- in the Northern since 2013 bon Sequestra- Reef Trend tion Partner- (carbonate ship (MRCSP). reservoir) DTE Energy, Core Energy and Batelle Table 4 Storage projects across the world: C O ECBM projects Project name Country Company opera- Total planned Capture mode CO fate Status of project References tors storage Fenn Big Valley Canada Alberta Research 50 t/day Non-power Mannville group In operation since Gunter et al. (2005) Council 1998 CSEMP Canada Suncor Energy 50 t/day Non-power Ardley Formation In operation since Shi and Durucan 2005 (2005) Qinshui Basin China Alberta Research 30 t/day Non-power Shanxi Formation In operation since Wong et al. (2007) Council 2003 Yubari Japan Japanese Ministry 0.004 Mt/year Non-power Yubari Formation In operation since Shi et al. (2008) of Economy, 2004 Trade and Industry Recopol Poland TNO-NITG 1 t/day Non-power Silesian Basin In operation since van Bergen et al. (Netherlands) 2003 (2003) 2013) and the Otway Project in Australia (Etheridge et al. (2015) noted that the embryonic stage of technology on 2011; Underschultz et al. 2011). CO capture would mean high costs of capture from power Even though carbon dioxide capture is outside the scope plants for early movers. Early movers need to be encour- of this review, it is obvious that the deployment of many aged by governments through subsidies. Successful cases carbon dioxide storage projects would be dependent on the of subsidies by the government can be seen in the Bound- cost and success of carbon capture processes. Celia et al. ary Dam Project by SaskPower and the Quest Project by Shell both in Canada. 1 3 Petroleum Science (2019) 16:1028–1063 1041 solved the equations using the streamlined methodology. 4 Modeling strategies employed for  CO Though their model was able to solve the pressure-driven storage flow in complex flow fields, it was limited by the assump- tions of a simple geochemical model and incompressible Numerical modeling is typically carried out before the flow. Qi et al. (2009) used the model developed by Obi and commencement of injection projects. They are used for Blunt (2006) to postulate a design strategy for injection predictions and optimizations. The flow path of the injected of CO which would render a large percentage of the CO CO needs to be predicted prior to injection. Furthermore, 2 2 immobile on the pore scale. As their work was focused on the optimization of well location needs to be properly maximizing the gas trapped via the residual gas trapping assessed during the planning phase. Several authors have mechanism, they modified the existing model by assigning attempted to model the plume movement of injected C O in relative permeabilities on a block by block basis. All in all, saline formations. Modeling of CO storage in saline aqui- these papers have been able to demonstrate the feasibility fers is usually performed using either analytical or numeri- of modeling storage of CO in saline aquifers by employing cal models. The choice of modeling technique employed is the streamlined methodology. Streamline simulations are, dependent on the aims of the researchers, the nature of the however, best suited to processes where limited pressure problem and the data available. Analytical models have the changes are expected to occur. Given that only injection is advantage of providing a quick insight into the suitability usually modeled in CO storage in saline aquifers thus lead- of a formation for storage. Zhou et al. (2008) employed an ing to significant pressure changes, streamline simulations analytical model to determine the storage capacity in saline have found limited applications in C O storage modeling. aquifers and expected pressure buildup during storage Vertical equilibrium models work by discretizing the simu- operations. Mathias et al. (2009a) developed approximate lation domain only in the horizontal direction leaving one solutions for pressure buildup in aquifers assuming vertical layer in the vertical direction. Two forms of the vertical pressure equilibrium and accounting for the Forchheimer equilibrium model exist: vertically integrated numerical flow of CO and brine. Solutions from the study were sub- models which include capillary forces and analytical mod- sequently applied in the screening of potential C O storage els including a sharp interface where the capillary pres- sites (Mathias et al. 2009b). Analytical models have been sure zone is thin with homogeneous formation parameters. used for plume migration studies. Nordbotten et al. (2005a) The technique capitalizes on the strong density differential also developed approximate solutions for the prediction of between supercritical C O and the in situ brine which leads the plume migration path in a CO storage site. The model to a marked upward increase in the CO . Particularly, on was validated with the commercial simulator ECLIPSE short time scales, the density differential could lead to a with very good accuracy. The underlying assumptions of strong buoyancy segregation of the two fluids. The idea analytical models are, however, too simplistic and cannot behind this technique is to derive a better understanding account for reservoir property and model geometry het- of the lateral plume spread and the segregation between erogeneities. More so, the complex geochemical reactions the different fluid phases. Its limitation is in its inability expected in CO storage cannot be reliably captured by to model heterogeneity in the vertical direction. The tech- analytical models. Streamline simulations, vertical equilib- nique has, however, been applied (Gasda et al. 2009, 2011) rium models and regular, conventional grid-based numeri- in modeling of CO storage. Another modeling technique cal models are forms of numerical modeling techniques which has been applied to the simulation of C O in aquifers which have been applied for the modeling of C O storage is the inversion percolation technique. In this approach, (Cavanagh and Haszeldine 2014; Gasda 2010; Jiang 2011; viscous forces are ignored; therefore, the only forces that Li et al. 2012; Obi and Blunt 2006; Pruess 2008; Saadawi dominate the flow are the capillary and gravity forces. et al. 2011; Wheeler et al. 2008). Streamline simulations Consequently, this technique is most suitable in systems work by splitting the simulation domain into small grid with low fluxes. Inversion percolation is employed when sizes and determining the pressure in each grid block using the capillary number (ratio of viscous forces to capillary a finite difference technique. The resulting pressure field force) is less than 0.0001. High-resolution inversion perco- is applied in tracing the streamlines which determine the lation models are noted for their simplicity and the speed expected flow fields. As opposed to other forms of numeri- of their numerical solutions. Limitations of this approach cal modeling, streamline simulations are faster and com- are, however, found when flow rates are high and capillary putationally efficient as flow equations are reduced to one- heterogeneity is not pronounced. Notably, this approach has dimensional equations along the streamlines. Obi and Blunt been employed in the modeling of the In Salah Field Pro- (2006) and Qi et al. (2009) have applied streamline simula- ject and the Sleipner storage with a high degree of accuracy tions in the modeling of CO storage. In their model, Obi (Cavanagh and Ringrose 2011; Cavanagh and Haszeldine and Blunt (2006) coupled transport and flow equations and 2014). Conventional 3D simulations making use of highly 1 3 1042 Petroleum Science (2019) 16:1028–1063 developed numerical discretization techniques have been k k k ⎛ ⎞ xx xy xz used to overcome the shortcomings of the other techniques ⎜ ⎟ k = k k k (12) yx yy yz ⎜ ⎟ by incorporating all relevant physics such as expected pres- k k k ⎝ ⎠ zx zy zz sure increases and heterogeneities in both the vertical and horizontal directions. Typically, they employ finite differ - Conservation of energy can also be solved for by equat- ence/element/volume techniques to solve transport and flow ing the summation of the time rate of change of the energy equations. In addition, they are able to couple other related term, advection and conduction terms to the source term as physical phenomena such as geochemistry, geomechanics shown below: and thermal changes. As a result of the detailed modeling of inherent physics, the regular 3D grid-based numerical modeling techniques are more computationally costly than  ( s U )+(1 − ) C T s s the other techniques. Most commercial simulators which (13) have been employed for modeling of CO storage issues + ∇ ⋅ ( q H ) −∇ ⋅ (∇T)= S have full modeling capabilities (Class et al. 2009; Nghiem et al. 2009). where U represents the specific internal energy, H is the Modeling of C O storage is a multi-component, multi- specific enthalpy, T is the temperature, C is the specific heat phase process with the two fluid phases as the brine and capacity, and all other symbols have definitions as described a CO -rich phase and the components like C O, H O, dis- 2 2 2 earlier. Subscript s represents the solid phase. solved salts in the brine and rock minerals. It should be noted These equations (Eqs. 10, 11 and 13) represent the fun- that the number of components modeled can be different damental equations for the modeling of storage of CO in depending on the problem to which it is applied. The funda- porous media (DePaolo et al. 2019; Nghiem et al. 2004; Pan mental equations used in CO storage modeling are basically et al. 2018a). These equations could be coupled with geo- the same as equations that describe the flow of oil, gas and chemical reactions, geomechanical modules and other rel- water in porous media. These equations are the conserva- evant physical phenomena. The solution of these equations tion of mass, momentum and energy. Constitutive relations requires either a sequential, simultaneous or fully coupled are used to formulate solutions for these equations. Other approach. physics which could be coupled with the basic equations are Over the years, researchers have made numerous attempts equations that predict geomechanical effects and geochemi- to describe underground CO migration and trapping mecha- cal reactions among others (Temitope and Gupta 2019). nisms using numerical analysis. Weir et al. (1996) developed The conservation of mass equation for components can be a two-dimensional model to evaluate C O quantities that written as the summation of the advection, diffusive terms migrated beyond a cap rock after C O injection for 10 years and the time rate of change of mass which equal a source into a 3-km-deep aquifer at a mass transfer rate of 100 kg/s. or sink term. They varied the confining layer’s permeability in order to determine the amount of CO that could pass through the ( s X ) + ∇ ⋅ ( q X )− ∇ ⋅ (  D ∇X )= S i layer. They concluded that a low-permeability seal should i i i overlay any target formation as this would mean that higher (10) capillary pressures would be required for the C O to pen- Darcy’s law for a single-phase flow can be written as etrate the seal. Another C O storage study conducted by q kk researchers at the Alberta Research Council (Gunter et al. v = =− (∇p +  g∇z) (11) 1993; Law and Bachu 1996) for the Upper Manville Group where the modeled formation was a Cretaceous glauconitic where t represents the time,  represents the porosity,  is sandstone aquifer 1.46 km in depth. The formation top of the density, q is the Darcy flux, k is the permeability tensor, the aquifer was overlain by several regional-scale aquitards k is the relative permeability, D is the diffusivity, X is the (low-permeability shale layers) that inhibited upward migra- mole fraction, s is the saturation term,  is the tortuosity, tion of the injected CO . The unevenness of the formation S denotes the source/sink term, v is the velocity vector,  is permeability was modeled based on drill-stem tests per- the dynamic viscosity, p is the pressure, g is the acceleration formed during exploration. The study showed no C O leak- due to gravity, and z represents the depth. Subscripts  and age during the modeled time scale. i are the phase and index, respectively. Nghiem et  al. (2004) developed a fully coupled EOS The permeability tensor can be written fully as compositional simulator for modeling C O storage in aqui- fers. The module consisted of geochemical reactions such as 1 3 Petroleum Science (2019) 16:1028–1063 1043 gas dissolution in the aqueous phase, chemical equilibrium General Purpose Reservoir Simulator (AD_GPRS) by Stan- reactions, mineral dissolution, and precipitation. The highly ford University (Benson et al. 2013; Fan 2006; Iskhakov coupled sets of nonlinear equations were solved simulta- 2013), MUFTE-UG (Multiphase Flow Transport and Energy neously using the Newton approach. The geochemistry Model on Unstructured Grids) developed by a joint effort module of the simulator was validated with the Geochemist of the University of Stuttgart and the University of Heidel- Workbench (GWB) developed at the University of Illinois berg (Ebigbo et al. 2006), IPARS-CO2 (Integrated Parallel with high accuracies. The resulting codes were applied on Accurate Reservoir Simulator) developed by the University two numerical grids: a 2D reservoir used to analyze the of Texas at Austin (Kong 2014; Wheeler et al. 2008); also impact of mineral trapping and a 3D grid used to study the existing are several simulators by the National Laboratories evolution of the CO plume. Rutqvist et al. (2010) coupled in the USA including TOUGH and TOUGH2 usually used a geomechanical simulator (FLAC3D) and a multi-phase in collaboration with ECON2 (Hovorka et al. 2006; Pruess flow simulator (TOUGH2) to study the ground deforma- et al. 2002), STOMP Subsurface Transport over Multiphase tions which would occur at the In Salah storage site in Alge- Processes (Bonneville et al. 2013) [see Table 5 for full list]. ria. Surface deformation results derived from monitoring The difference between most of these simulators lies in the using interferometry synthetic aperture radar (InSAR) were numerical methods and discretization technique used, the employed in this study to validate the numerical models and inclusion or non-inclusion of certain physics and the cou- displayed good agreements with obtained results. A sum- pling methods of the physics. mary of the workflow for most of the reservoir simulators Numerical simulations have been applied to assess the for CO storage issue is provided in Fig. 6. feasibility of commercial storage in aquifers. In a recent Many researchers exploring CO storage issues have study, Temitope et al. (2016) employed the Computer Mod- focused more on simulations for large-scale analysis with elling Group (CMG) simulator with an advanced geochemi- most experiments carried out aimed at better understanding cal modeling module to evaluate the possibility of commer- the physics of the processes that occur during the injection cial injection in the Shuaiba aquifer of the Falaha syncline and post-injection phases. Thus, due to the complex nature in the United Arab Emirates (UAE). Simulation results were of storage of CO and the time period taken for carbon able to provide the possible migration path of injected C O 2 2 dioxide to be stored underground, the only effective way to into the aquifer. In modeling the impact of thermal fac- understand the storage capacity of an aquifer before injec- tors on the injection of CO into the FutureGen 2.0 Site in tion commences is through modeling and simulations. This Illinois in the USA, Nguyen et al. (2016) made use of the explains why there exists a myriad of simulators which have simulators STOMP-CO coupled with the ABAQUS finite the capacity to model CO storage in aquifers; among them element simulator. Results suggested that in the range of includes CMG (Computer Modelling Group) GEM-GHG temperatures in which injection would take place, fracturing Module (Nghiem et al. 2004, 2009), ECLIPSE 100 and 300 would be unlikely to happen due to thermal factors. Basirat (Schlumberger), CO2STORE Module (Pickup et al. 2011, et al. (2016) employed the TOUGH2 simulation codes to 2012; Sifuentes et  al. 2009), Automatic Differentiation model the injection of CO into an experimental site in Basic equations Darcy law Conservation of mass Conservation of energy Numerical techniques used to couple equations simultaneous/sequential coupling Modelling of geochemical Modelling of geochemical Complex processes reactions effects Dual porosity/dual Adsorption Cap rock integrity permeability Mineral dissolution/precipitation Stress relation Residual gas trapping CO dissolution in brine EOS modelling Fig. 6 Workflow for CO storage modeling 1 3 1044 Petroleum Science (2019) 16:1028–1063 1 3 Table 5 List of simulators and codes for CO storage Simulators Full names Description Developers Relevant literature ABAQUS-FEA ABAQUS-FEA Geomechanical, single- and two- SIMULIA Gemmer et al. (2011) and Le Gallo phase flow et al. (2006) AD_GPRS Automatic Differentiation—General Generalized multiple phase composi- Stanford University Huo and Gong (2010) and Li et al. Purpose Reservoir Simulator tional/thermal model for unstruc- 2012) tured grids CO —PENS CO —Predicting Engineered Natural System-level modeling of the long- Los Alamos National Laboratory Pawar et al. (2006) and Stauffer et al. 2 2 Systems term fate of C O in sequestration (LANL) (2006) sites CO Toolkit Permedia ™ Permedia High-resolution petroleum migra- Landmark Cavanagh and Ringrose (2010, 2011) tion simulator for multi-phase flow behavior in porous, faulted and fractured media Advanced Resources International Mingjun et al. (2010) and Schepers COMET3 COMET# Black oil production, hydrocarbon et al. (2009) dioxide recovery from desorption- controlled reservoirs COMSOL Multiphysics COMSOL General partial differential equation COMSOL Farajzadeh et al. (2009) and Houdu solver with finite element solver et al. (2008) COORES CO Reservoir Environmental Multi-component, three-phase and Insititut français du pétrole Estublier and Lackner (2009) and Le Simulator 3D fluid flow in heterogeneous Gallo et al. (2006) porous media CrunchFlow Crunch Flow 3D, multi-phase transport with equi- Lawrence Livermore National Labo- Siirila et al. (2012) and Steefel and librium and kinetic mineral–gas– ratory | Lasaga (1994) water reactions RetrasoCodeBright Retraso (REactive TRAnsport of A solution of the flow, heat and Technical University of Catalonia Kvamme and Liu (2009) and Olivella SOlutes) CodeBright (COupled geomechanical model equations (UPC), Barcelona, Spain et al. (1996) DEformation of BRIne Gas and Heat Transport) DuMux DUNE for Multi-(Phase, Component, Multi-scale, multi-physics toolbox for University of Stuttgart Class et al. (2009) and Flemisch et al. Scale, Physics) the simulation of flow and transport (2007) processes in porous media ECLIPSE ECLIPSE Non-isothermal multi-phase flow in Schlumberger Juanes et al. (2006), Martens et al. porous media (2012) and Sifuentes et al. (2009) ELSA Estimating Leakage Semi-Analyti- Provides quantitative estimates of Princeton University Nordbotten et al. (2005b, 2009) cally fluid distribution and leakage rates in systems involving a sedimentary succession of multiple aquifers and aquitards FEFLOW FEFLOW Solving the groundwater flow equa- DHI-WASY Melikadze et al. (2013) tion with mass and heat transfer, including multi-component chemi- cal kinetics Petroleum Science (2019) 16:1028–1063 1045 1 3 Table 5 (continued) Simulators Full names Description Developers Relevant literature FEHM Finite Element Heat and Mass Trans- Non-isothermal, multi-phase flow Los Alamos National Laboratory Pawar et al. (2005) and Robinson et al. fer Simulator (including phase-change) in unfrac- (2000) tured and fractured media with reactive geochemistry & geome- chanical coupling GASMOD/GCOMP GASMOD/GCOMP Multi-phase reservoir simulator PHH Engineering Software Limited Palmer and Mansoori (1996) GEM-GHG Generalized Equation of State Non-isothermal multi-phase flow in Computer Modelling Group (Canada) Kumar et al. (2005) and Nghiem et al. Model—Greenhouse Gases porous media (2009) GMI-SFIB Geomechanics International—Stress Three-dimensional stress mod- Geomechanics International Fang (2011) and Fang and Khaksar and Failure of Inclined Boreholes eling for compressional (wellbore (2012) breakout) and tensional (tensile wall fractures) stress failure, fracture modeling GWB The Geochemist’s Workbench Chemical reactions, pathways, kinet- University of Illinois Bethke and Yeakel (2009) and Lu et al. ics (2011) IPARS-CO2 Integrated Parallel Accurate Reser- Non-isothermal compositional EOS University of Texas at Austin Delshad et al. (2011) and Kong et al. voir Simulator coupled with geochemical reactions (2015) MASTER Miscible Applied Simulation Tech- Black oil simulator, compositional National Energy Technology Labora- Ammer and Brummert (1991) niques for Energy Recovery multi-phase flow tory METSIM 2 METSIM 2 A non-isothermal multi-component Imperial College Durucan et al. (2004) and Law et al. coalbed gas simulator (2004) MODFLOW MODFLOW Solving groundwater flow equation to US Geological Survey (USGS) Nicot et al. (2009) simulate the flow through aquifers MoRes Modular Reservoir Simulator A modular object-oriented design for Shell Class et al. (2009) and Wei and Saaf black-oil, equation-of-state (EOS) (2009) and K-value compositional simula- tions ACCRETE Athena Carbon Dioxide Capture and Thermal multi-phase 3D reactive University of Bergen Hellevang and Kvamme (2006, 2007) Storage Geochemistry Module transport numerical code MUFTE-UG Multiphase Flow Transport and Isothermal and non-isothermal multi- University of Stuttgart Assteerawatt et al. (2005) and Ebigbo Energy Model on Unstructured phase–multi-component flow and et al. (2006) Grids transport processes in porous and fractured porous media NFFLOW-FRACGEN Fracture Network Generator Flow Two-phase, multi-component flow in National Energy Technology Labora- Myshakin et al. (2015) and Schwartz Simulator fractured media tory (2006) NUFT Non-isothermal Unsaturated–satu- Non-isothermal multi-phase flow and Lawrence Livermore National Labo- Hao et al. (2012) rated Flow and Transport model chemical reactions in porous media ratory | OGS: [Couples GEM, BRNS, OpenGeoSys Porous and fractured media THMC Helmholtz Centre for Environmental Graupner et al. (2011) and Li et al. PHREEQC, ChemApp, simulation Research (UFZ) (2014) Rockflow] 1046 Petroleum Science (2019) 16:1028–1063 1 3 Table 5 (continued) Simulators Full names Description Developers Relevant literature PFLOTRAN Parallel Reactive Flow and Transport Non-isothermal multi-phase, multi- Los Alamos National Laboratory Lu and Lichtner (2005, 2007) component, chemically reactive flows in porous media PHAST PHAST Multi-component, 3-D transport with US Geological Survey (USGS) Parkhurst et al. (2004) equilibrium and kinetic mineral– gas–water reactions PHREEQC PHREEQC The low-temperature aqueous geo- US Geological Survey (USGS) Parkhurst and Appelo (2013) and van chemical simulator Pham et al. (2012) PSU- COALCOMP Pennsylvania State University- Three-dimensional, two-phase, dual Pennsylvania State University/ Bromhal et al. (2005) and Manik et al. COALCOMP porosity, sorption, fully implicit, National Energy Technology (2000) compositional coalbed methane Laboratory reservoir simulator ROCKFLOW Rock Flow Multi-phase flow and solute transport Bundesanstalt für Geowissenschaften Kolditz et al. (2003) processes in porous and fractured und Rohstoffe and University of media, as well as thermal–hydrauli- Hannover cally–mechanical (THM) coupled processes SOLVEQ/CHILLER CHILLER Multi-component multi-phase Department of Geological Sciences, Palandri and Kharaka (2005) and Reed equilibrium geochemical calcula- University of Oregon and Spycher (2006) tion software based on minimum free-energy RTAFF2 Reactive Transport and Fluid Flow Non-isothermal multi-phase and a French Geological Survey (BRGM) multi-component flow simulator SIMED II SIMED II Two-phase three-dimensional multi- The Netherlands/Commonwealth Stevenson and Pinczewski (1995) and component coalbed gas simulator Scientific and Industrial Research van Bergen et al. (2002) Organization (CSIRO), Australia STOMP Subsurface Transport over Mul- Non-isothermal multi-phase flow in Pacific North-West National Labora- Bonneville et al. (2013) and White tiphase Processes porous media, coupled with reactive tory et al. (2012) transport. TOUGH/TOUGH2 Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Pruess et al. (2002) and Pruess and ter and Heat unfractured and fractured media tory Spycher (2007) TOUGH-FLAC Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Rutqvist (2012) and Rutqvist and ter and Heat unfractured and fractured media tory Tsang (2003) with geomechanical coupling TOUGHREACT Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Xu et al. (2006a) ter and Heat Reactive Transport unfractured and fractured media tory with reactive geochemistry VESA Vertical Equilibrium with Subsurface Vertically averaged numerical model Princeton University Gasda et al. (2009) Analytical for large-scale flow coupled with an embedded analytical model for wellbore flow Petroleum Science (2019) 16:1028–1063 1047 Maguelone, France. Geophysical monitoring tools were analysis expected from commercial injection into the forma- used in their field experiments to gain useful information tion. Different methods exist for the calculation of storage about the site and also to monitor the movement of the gas. volumes and can be broadly classified into static and dynamic They highlighted the importance of accounting for geologi- estimation methods. As the names would suggest, static esti- cal heterogeneity in modeling procedures. In addition, the mation methods do not change with time and only require study was able to provide information on the usefulness of basic rock and fluid properties. They are typically determined geophysical monitoring tools in analyzing plume migration using volumetric and compressibility parameters. Con- in storage sites. versely, dynamic estimation methods vary with time and are Benchmark studies have thus been performed to under- determined using reservoir simulations and some analytical stand the capabilities of different softwares used for carbon methods which incorporate time-dependent variables in their dioxide storage. Pruess et al. (2002) performed a critical derivations. Estimation of CO storage capacity in geological comparison on the performance of different commercial media is at best an approximation due to the many uncertain- reservoir simulator codes for accurate prediction of CO ties present both in the formation (heterogeneity) and in the storage processes (that is TOUGH2, Geoquest’s ECLIPSE, physics of the processes. The level of uncertainty also varies CMG’s GEM, etc.). They concluded that all softwares could with the method being used to determine the storage capac- be used to simulate the essential flow and transport pro - ity and the amount of available data. The methodology to be cesses that would accompany geologic storage. However, used for the determination of the capacity is dependent on the hydromechanical process would only be solved by one the formation type, that is coal seams, depleted oil and gas code TOUGH-FLAC. Law et al. (2004) analyzed the results reservoirs or saline aquifers. In addition, the extent of the of five simulators to a benchmark problem for CO storage storage medium may determine the approach to be used in issues in coalbed formations. Class et al. (2009) also per- storage capacity determination. Open boundaries where the formed a benchmark study with the use of different simula- extent of the media is assumed to be infinite, closed where tors to address the problems related to C O storage in geo- the extent of the media is assumed to have a finite end and logic formations. The outcome of such benchmark studies semi-closed are all different forms available in the literature illustrates that the results of the simulation of any storage for storage capacity determination. problem would depend on the simulator used and are highly Because candidate storage sites are usually not fully dependent on the numerical methods used and the physics of characterized before estimates are made, they are usually processes implemented. It is suggested that the choice of the reported as a low- and high-capacity estimate of storage simulator to be used would depend on the physical processes (DOE 2007) with Monte Carlo simulations employed to being focused on for best results. account for uncertainties. Two primary methodologies are Simulation of CO storage is generally a little more dif- being used; they include the methodology by the Depart- ficult than conventional simulations due to the interplay ment of Energy (DOE) of the USA (DOE 2007) and the between phase change, composition and reservoir hetero- Carbon Dioxide Sequestration Leadership Forum (CSLF) geneity which require highly efficient computational algo- (Bachu et al. 2007b) and the formulas used by the two bod- rithms (Jiang 2011). The striking difference between CO ies for storage determination are summarized in the next storage issues and conventional porous media modeling is subsections. the large temporal and spatial scale differences. A multi- scale methodology which incorporates advanced numerical 5.1 Coal seams schemes may be the best way to approach such scale differ - ences in such a way as to capture the complex multi-phase, The formulas for calculating the storage capacity of coal multi-component species, and physics in heterogeneous sys- seams by the DOE and CSLF methods are as follows:DOE: tems and also save computational cost. Such multi-scale, M = Ah CE multi-physics approach has been implemented in the devel- g (14) opment of certain simulators (Flemisch et al. 2007). CSLF: M = Ah(1 − f − f ) n G CO a m CO c c (15) 2 2 5 Capacity estimation for  CO storage G = V ∗ projects cs L (16) P + P An initial estimate of the storage capacity of a formation where A represents the area, h is the thickness, h is the gross is required for successful implementation of CCS projects. thickness, C is the concentration of C O standard volume Such estimates assist in project planning and in potential risk per unit of coal volume, f and f are the ash and moisture a m 1 3 1048 Petroleum Science (2019) 16:1028–1063 weight fraction of coal, M is the mass storage, E is the C O CSLF: storage efficiency factor that reflects a fraction of the total (P Z T ) s r r coal bulk volume that is contacted by CO ,  is the density, Gas fields ∶ M =  R (1 − F )OGIP (19) 2 CO CO f IG 2 2 (P Z T ) r s s n is the bulk coal density, G is the gas coal content, G is c c cs the gas content at saturation, V and P are the Langmuir L L R OOIP volume and pressure, respectively, and P represents the Oil fields ∶ M =  − V + V (20) CO CO iw pw 2 2 pressure. The Langmuir volume is the maximum adsorp- f tion capacity of the gas for a particular coal at a defined where A represents the area, h is the net thickness,  is the n e temperature and infinite pressure. Its unit is usually given in effective porosity, M is the mass storage, E is the CO stor- scf/ton (volume of gas per mass of unit coal). The Langmuir age efficiency factor that reflects a fraction of the total pore pressure (also known as the critical desorption pressure) is volume from which oil and/or gas has been produced and the pressure at which one half of the Langmuir volume can that can be filled by CO , ρ is the density, B is the formation 2 f be adsorbed/stored. volume factor, S is the average water saturation, P repre- In the CSLF method, the storage capacity available in coal sents the pressure, Z and T are the compressibility factors, seams for C O is determined in a manner akin to the deter- respectively, R is the recovery factor, OOIP and OGIP stand mination of initial gas in place in coalbed methane reservoirs for the original oil and gas in place, respectively, F is the IG as shown in Eq. 15. The ability of the coal gas to adsorb the fraction of injected gas, and V and V are the volumes of iw pw injected CO is dependent on pressure, temperature and coal injected and produced water, respectively. characteristics of the formation. The gas content at saturation is determined by Eq. 16. The two equations assume that the 5.3 Saline aquifers CO contacts all the available coal and that the coal adsorbs CO to full capacity. In reality, however, this may not be prac- Bachu et al. (2007a) as part of research conducted by the ticable; hence, a correction factor is introduced to account for Carbon Sequestration Leadership Forum (CSLF) expressed the non-ideality as given in Eq. 17: the effective storage capacity available in structural traps M = M ∗ C ∗ R in terms of volume and mass of CO as in Eqs. 21 and 22, e CO f (17) 2 respectively. The boundaries of the aquifer are considered where M is the effective storage capacity, C is the comple- to be open. tion factor, and R is the recovery factor. The product of completion and recovery factor is together known as the gas V = Ah(1 − S )C CO w c (21) 2 irr deliverability. The completion factor C is an estimate of that M = Ah(1 − S ) C part of the net cumulative coal thickness within the drilled CO w CO c (22) 2 irr 2 coal zone that will contribute to gas production or storage; it where the spatial variation of the formation is known; the is dependent on the individual thickness of the separate coal volumes can be expressed as seams and on the distance between them and is lower for thin coal seams than for thick ones (Bachu et al. 2007a). Monte V = (1 − S )dxdydz ∗ C (23) CO w c 2 irr Carlo uncertainty analysis can be employed to account for uncertainties in the determination of unknown parameters. where A is the area, h is the thickness, S is the irreducible irr water saturation,  is the density of C O , and C is the CO 2 c 5.2 Oil and gas reservoirs capacity coefficient which is dependent on the trap hetero- geneity, buoyancy and sweep efficiency. Estimation of available storage capacity in depleted oil and gas The capacity coec ffi ient is usually site-specic fi and is best reservoirs is not as complicated as with coal seams and saline determined through numerical simulations or detailed field aquifers as these reservoirs have been adequately character- work. It incorporates effects such as the heterogeneity of the ized during the production stages of the reservoir. The basic aquifer, buoyancy effect and sweep efficiency. The Interna- assumption in the formulation of storage capacities is the avail- tional Energy Agency Greenhouse Gas R&D Programme ability of all the pore spaces vacated by hydrocarbon fluids. In (IEAGHG 2009) in their study evaluated the capacity coeffi- other words, it is assumed that the formation fluids have not cient as a function of lithology based on extensive numerical been replaced by water from any supporting aquifer around studies. The values derived for carbonate formations based the region of the field. The storage capacity by the CSLF and on the 10th, 50th and 90th percentiles were 1.41%, 2.04% DOE methods are as stated below. and 3.27%, respectively. The formula for capacity estimates DOE: derived by the US Department of Energy (DOE-NETL 2015) is similar to that of the CSLF. The only difference lies M = Ah  (1 − S )B E (18) n e w f 1 3 Petroleum Science (2019) 16:1028–1063 1049 in the capacity coefficient given for the carbonate formations is the time, and ΔP is the maximum allowable pressure max with the DOE estimating the 10th, 50th and 90th percentiles increase. as 0.51%, 2.0% and 5.5%, respectively. Dynamic simulations still represent the best method for The storage volume available by residual trapping can be the determination of storage capacities of geological forma- determined using the correlation below: tions selected for storage as they contain detailed informa- tion regarding the petrophysical properties of the formation. V =ΔV S CO t trap CO t (24) 2 2 Coupled with this, numerical simulators nowadays have where S (saturation of CO ) is dependent on the hyster- CO t 2 embedded in their simulators the ability to calculate the stor- esis effects of the relative permeabilities and the CO satura- 2 age capacity provided by the different storage mechanisms tions during reversal flow. over an extended period. Analytical determination methods As highlighted earlier, the dissolution of C O in brine 2 such as fractional flow theory (Moghanloo et al. 2015) and is a continuous and slow process that is dependent on the relative permeability curve analysis method (Zhu et al. 2017) convection, diffusion and dispersion. The storage capacity for the determination of storage volumes can also be found on a basin and regional scale, as determined by Bachu et al. in the literature. (2007a) for solubility trapping, is given below The aforementioned described techniques have been employed mainly in the determination of storage capacities CO CO 2 2 M = ( X −  X )dxdydz (25) CO t s s o o across the world. Lindeberg et al. (2009) used both analytical and reservoir simulations to estimate the available storage where  is the porosity,  is the density, X stands for mass capacity in the Utsira Formation of Norway. Their reser- fraction, M denotes the mass, and subscripts s and o denote voir simulations were done in such a way to model elevated the carbon dioxide content at the saturation and initial pressures in the aquifer. In addition, a CO breakthrough stages, respectively. The time frame required for mineral from production wells was also monitored in estimate deter- trapping to occur makes it difficult to provide correlations mination. In China, Liu et al. (2005) estimated the storage for the determination of the mineral trapping capacity. capacities in gas fields and coalbeds present in the country. Zhou et al. (2008) devised a simple method for determin- Similarly, Suekane et al. (2008) determined the residual ing the storage capacity in closed and semi-closed aquifers. and solubility capacities available in Japanese aquifers. By The main idea lies in the premise that injected C O will lead improving on the flaws of the conventional analytical tech- to a pressure increase in the formation. This will, in turn, niques for storage estimation, Ding et al. (2018) proposed lead to a displacement of native brine which can either be new analytical methodologies for the determination of solu- stored in the expanded pore space due to compression of the bility and mineral trapping in aquifers and depleted oil reser- rocks (closed systems) or the pore space in the seals overly- voirs. Their model was applied to the HB oil field in China, ing the formation (semi-closed systems). and estimates were compared to a similar methodology by Zhou et al. (2008) showed the derivations for closed sys- Xu et al. (2004) with slight discrepancies observed. They, tems by using the given in Eqs. 26 and 27 below. however, argued that their model would be superior as, in addition to the model’s ability to determine storage capacity V =  +  V ΔP (26) by solubility trapping, the model could also determine the CO p w pore max annual storage capacities by mineral trapping. M =  +  V ΔP (27) CO p w pore max CO 2 2 6 Measurement, monitoring and verification For semi-closed systems the following equation is techniques during  CO storage suggested: Monitoring the movement of the plume for leakages is criti- V (t )=  +  ΔP (t )V CO 1 p w max max pore cal in the post-injection phase of storage. Containment of + 0.5  +  ΔP (t )V ps w max max s (28)the CO is achieved if proper monitoring is performed as max 2Ak ΔP (t) s max leakages could be detected early, thus ensuring that the + dt 0 environment and groundwater are not at risk from released w s gases. Furthermore, monitoring could be employed in the where  is the compressibility, A is the area, k is the perme- validation of simulation predictions by tracking the pres- ability, subscripts s, p, w refer to the seal, pore and water, sure buildup in the formation (Bourne et al. 2014). Mass respectively, β refers to the compressibility of the rock ps balance verifications are also an important reason for carry - from pore to seals, V is the volume, µ is the water viscos- ing out monitoring studies. Injected C O volumes must be ity, B stands for thickness of the top and bottom seals, t tracked to ensure they are stored in identified zones and in 1 3 1050 Petroleum Science (2019) 16:1028–1063 line with emission quotas specified before the commence- they can easily be detected even at low concentrations, are ment of such projects. Successful verification of simulations highly soluble in C O , are non-toxic and are non-radioactive. via monitoring would provide researchers with greater con- A notable CO injection project which has made use of the fidence in the use of simulation tools. Consequently, a lot of tracer technique for monitoring is the Frio Project (Nance effort is continuously made to develop accurate monitoring et al. 2005). Their monitoring design made use of PFCs as tools. As with the modeling approach, monitoring of CO the chemical tracer to monitor leakages. Fibrous elements can either be classified on a spatial or temporal basis. On a such as capillary absorbent tubes (CATs) were placed on spatial basis, it is monitored based on the area which the CO surface installations in order to adsorb the PFCs. The CATs affects. On this basis, it can be classified into atmospheric were removed on a periodic basis to ascertain the amount monitoring, near-surface monitoring and subsurface moni- of PFCs which had sorbed on the surface of the CATs using toring (which involves the faults, wells, reservoir and seals) thermal desorption and gas chromatograph techniques. Laser (Brown et al. 2009). On a temporal basis, monitoring can systems are remote sensing technologies that make use of be grouped as during the injection phase and post-injection either optical absorption, breakdown spectroscopy or non- phase. For further discussion, we limit ourselves to discuss- linear optics to monitor gas leakages. A laser application ing monitoring on a spatial basis.for CO detection, however, only makes use of the optical absorption technique. In this technique, the laser beams a 6.1 Atmospheric monitoring tools light which has been tuned to the wavelength of the CO on the gas. The scattered light which emanates from the gas As the name implies, these tools ensure that the CO injected after absorption is examined. An issue with this technique into the formations does not leak into the atmosphere above is the accurate determination of the wavelength of C O as it. This monitoring strategy is important due to the concerns the absorption wavelengths of CO must be carefully deter- about leaked CO . Atmospheric monitoring tools are typi- mined without infringing on the absorption wavelengths of cally required to be very sensitive as leakage of CO from water vapor. the formation could be quickly dispersed in the atmosphere, thus making it difficult for other forms of monitoring tools 6.2 Near‑surface monitoring tools to recognize the gas immediately. Atmospheric monitoring tools are placed at the potential leakage sources so as to Usually, the flow of CO at the near-surface consists of increase their detection capability and are especially required bubbles which emanate from faults or near an abandoned to provide confidence in carbon dioxide storage and for car - wellbore. Monitoring of C O at the near-surface is impor- bon accounting verification. The tools used to detect CO tant as it serves as a link between the subsurface and the leakage in the atmosphere are optical sensors, atmospheric atmosphere. Therefore, it can provide information on leaks tracers and eddy covariance (Brown et al. 2009). Other sys- in the subsurface while preventing leaks to the atmosphere tems which can be used in monitoring the atmospheric levels if detected in time, monitoring in this area has been proven of CO include CO detectors, advanced leak detection sys- 2 2 to be less expensive than atmospheric and subsurface moni- tem, laser systems and LIDAR. As the quantity of safe CO toring. Some techniques which can be used for near-surface required to exist in the atmosphere must not exceed certain monitoring could also be used for subsurface monitoring. limits, CO detectors can be applied to sense the existence of Such techniques which could be used for this monitoring excess CO in the atmosphere. Application of CO detectors 2 2 have been summarized in the next subsection. Such tech- might, however, prove to be impractical due to the enormous niques include interferometric synthetic aperture radar number of detectors that would be required to effectively (InSAR), tiltmeters, time-lapse seismic among others. detect the gas. Eddy covariance also known as eddy flux is an important atmospheric monitoring tool used to quan- tify the fluxes of gases between the surface of the earth and 6.3 Subsurface monitoring tools the atmosphere. It has the advantage of being able to cover kilometers of space, thereby providing quick monitoring and The objectives of subsurface monitoring are to track the having a low to moderate cost. Atmospheric tracers are arti- movement of an injected CO plume in a deep geologic for- ficial substances injected into the formation along with the mation; to define the lateral extent and boundaries of the CO in order to observe the leakage of C O early on. They plume; to track associated pressure changes in the reservoir; 2 2 are also used to monitor the flow direction of the CO in the and to demonstrate long-term stability of the CO plume 2 2 formation. Conventional tracers which have been employed (Brown et al. 2009). Numerous monitoring techniques can for monitoring studies are the perfluorocarbons (PFCs) and be employed for the monitoring of CO plume in the sub- sulfur hexafluoride (SF6). Perfluorocarbons (PFCs) are, surface. The choice of monitoring techniques to be used however, preferred to sulfur hexafluoride (SF6) because for subsurface monitoring is dependent on the information 1 3 Petroleum Science (2019) 16:1028–1063 1051 required, costs of monitoring technique and time frame to to observe the extent of geomechanical deformation in the achieve information. subsurface. They are particularly useful in the monitoring of Seismic methods have been employed to evaluate the dis- cap rock deformations. InSAR has been applied for the mon- tribution of faults and the subsurface structures using 3D itoring of surface deformations. It achieves its objectives techniques. In a 4D mode that includes time-lapse data, seis- by making use of two synthetic aperture radars to generate mic methods can also be used to track the movement of the maps. This technique is sensitive to changes in deformations injected plume and gas leakages. Multi-component 3D sur- and has been used to measure millimeter changes in sur- face seismic provides better information when the geology face deformation. Different forms of the InSAR techniques of the formation is non-uniform. Together with time-lapse, include corner reflector Interferometric synthetic aperture multi-component seismic profiling provides valuable infor - radar (CR-InSAR), permanent scatterer interferometric syn- mation on the migration of the injected gas. If cost consid- thetic aperture radar (PS-InSAR) and differential interfero- erations are taken into account, 2D time-lapse seismic moni- metric synthetic aperture radar (D-InSAR). The technique toring could be used to provide data on the injected plume. has been applied for the monitoring of natural occurrences The downside of the 2D methods is in their inability to track such as volcanoes and earthquakes. The ability of the InSAR plume movement in formations with complex geometries. technique to monitor surface deformations has been applied 2D seismic techniques would be more useful where observa- in storage sites for tracking fluid pressure alterations, thus tion wells are available and cross-well seismic technology determining leakages. Recently, it was pioneered as a moni- could be employed. Vertical seismic profile (VSP) has been toring tool at the In Salah storage site in Algeria. employed to provide information on the leakages and the The choice of monitoring tool to be employed on any migration path of CO (El-Kaseeh et al. 2017). Most of the specific storage site is dependent on the nature of the site. conventional seismic methods have been used to determine For example, geophysical monitoring from the surface is leakages and migration path of the C O . In order to quantify dependent on the extent of overburden on the aquifer. There- the injected gas, seismic methods have been employed by fore, in geologically complex scenarios, monitoring of the combining the measurement of the velocity with Gassmann injected plume via this technique would be more cumber- modeling. This method requires that the density of CO at some. In the same vein, information available on a particu- reservoir conditions is known. Determination of this density lar storage site could influence the monitoring technique is not an easy process, and therefore, seismic monitoring chosen. Depleted oil and gas reservoirs which have been tools have been combined with gravimetry. Gravimetry basi- adequately characterized and have been proven to have cally involves using gravity to monitor the in situ changes assured seal integrity would make for easier monitoring of in the density of the injected gas. Results from gravimetric the injected CO plume. monitoring could provide reliable inputs for flow simula- Established commercially known CCS projects have tions. Gravimetric methods, however, possess low sensitiv- employed different monitoring tools. Torp and Gale (2004) ity and require a sizeable amount of C O injected into the provided useful information on the monitoring tools used formation before responses can be picked up. at the Sleipner project in Norway. Repeated seismic data Electromagnetic and electric methods have found impor- were among the many tools used for monitoring (Fig. 7). tant use as monitoring tools. They make use of electrical and The monitoring procedures confirmed some of the estimates electromagnetic responses from the subsurface to determine from reservoir simulation. The injected CO moved upward the changes in saturation. These techniques involve meas- due to buoyancy after the injection and accumulated under uring important electric parameters such as conductivity, the cap rock overlying the formation. Also, it was observed resistivity and employing correlations such as the Archie that solubility trapping would occur faster than mineral expression to relate these parameters to saturations. Differ - trapping. The simulation model for the Sleipner project was ent methods that use these concepts are the magnetotelluric then history-matched with the seismic data results to pro- sounding, electromagnetic resistivity, electrical resistivity vide accurate predictions for the future. However, seismic tomography (ERT), electromagnetic induction tomography monitoring is costly and other monitoring tools such as pres- (EMIT) among others. sure monitoring and observation wells could provide viable Geophysical logs have also been employed for the moni- alternatives. toring of subsurface-injected plumes. They provide useful Ringrose et al. (2013) analyzed the lessons learned from information on well properties and reservoir fluids. Exam- the In Salah Project in Algeria. Among these were the need ples of geophysical logging tools which could be employed for characterization of the overburden and the reservoir include sonic logs, neutron logs and density logs. Coupled prior to injection, constant risk assessments of the identi- with their ability to map saturation, geophysical logging fied storage sites and the significance of flexibility in the tools could also provide information on the onset of cor- design of capture, compression and injection systems. The rosion in the casings of wellbores. Tiltmeters can be used interferometric synthetic aperture radar (InSAR) method for 1 3 1052 Petroleum Science (2019) 16:1028–1063 Fig. 7 Seismic survey results for the Sleipner project (Torp and Gale 2004) storage monitoring was pioneered in this project. InSAR was such as failure modes (risk evaluation), likelihood and con- able to provide information on millimeter changes in ground sequences of failure must be answered when performing risk surface elevation; it was also able to give insights into the assessments for projects. Risks and challenges involved in geomechanical response to CO injection. Arts et al. (2004) CO storage are highlighted below. 2 2 made use of time-lapse seismic studies to monitor plume movement in the Utsira Formation. Notably, they were able 7.1 Leakage to demonstrate that the impact of the movement of CO on seismic measurements was considerable and thus seismic The primary and most important risk factor is leakage. Most could be used as a suitable monitoring tool during the life- modeling and monitoring studies conducted in the devel- cycle of a storage project. A summary of monitoring tools opment, implementation and monitoring phases of carbon used at select CCS projects is provided in Table 6. dioxide storage are done primarily to avoid leakage of the On a broader scale, monitoring is usually quantified as gas into the atmosphere, groundwater aquifers, shallow soil monitoring, verification and accounting (MVA) to include zones and overlying resource bearing strata and to ensure mass balance verifications and accounting for operators. secure containment of gas. The leakage of carbon dioxide Interested readers are referred to Plasynski et al. (2011) for could be as a result of the following: details on MVA strategies for different projects. 1. Aquifer over-pressurization: Aquifer over-pressurization could lead to cracks in the cap rock overlying it and in the reactivation of faults and thus should be avoided. 7 Risks and challenges in  CO storage The risk of aquifer over-pressurization is much less in depleted hydrocarbon dioxide reservoirs due to reduced The high dependency of world energy on coal-fired power pressure before the injection started. Saline aquifers plants makes carbon capture and storage a very important pose more risk from aquifer over-pressurization because technology for the mitigation of global warming. Therefore, the pressure of the aquifer begins from the initial pres- it represents the only viable option in the short term to limit sure, thereby leading to quick buildup of pressure when global warming effects and must be pursued vigorously. injection commences. Vilarrasa et al. (2010) performed However, just as with most technologies, carbon dioxide numerical simulations to ascertain the risk of over- storage comes with its own risks and challenges which must pressure during injection. The authors employed an be properly catered for before venturing into it. Questions axisymmetric horizontal aquifer–cap rock system cou- 1 3 Petroleum Science (2019) 16:1028–1063 1053 Table 6 Monitoring techniques used in field-scale projects Field project Category Monitoring techniques Select literature Sleipner Saline aquifer Time-lapse gravity; micro-seismic; time-lapse Arts et al. (2004) and Cavanagh (2013) seismic Ketzin Saline aquifer Pulsed-neutron gamma logs; 4D seismic; 3D repeat Ivanova et al. (2012) and Kiessling et al. (2010) seismic survey; electrical resistivity tomography (ERT); cross-well seismic; geophysical monitor- ing Weyburn EOR Passive seismic; 4D time-lapse; vertical seismic Bellefleur et al. (2003), Preston et al. (2005) and profile (VSP); tracer injection; geochemical White (2011) sampling analysis; production data analysis; cross- well seismic Otway Depleted gas field Hydrodynamic sampling; high-resolution travel Boreham et al. (2011), Etheridge et al. (2011) and time, flask sampling; 3D surface seismic; head- Urosevic et al. (2011) space gas sampling; logging pressure/temperature; flux tower; surface soil gas; downhole fluid sam- pling; CO sniffers; VSP; micro-seismic; borehole seismic; groundwater chemistry Cranfield Saline aquifer Cross-well seismic tomography; 4D seismic, elec- Ajo-Franklin et al. (2013), Alfi et al. (2015), Carrigan trical resistance tomographic monitoring; vertical et al. (2013) and Kim and Hosseini (2014) seismic profile; above zone monitoring interval; tracers In Salah Depleted gas field Well log data; 3D seismic baseline survey; 4D Mathieson et al. (2010), Onuma and Ohkawa (2009) seismic monitoring; groundwater monitoring and Ringrose et al. (2009) wells; micro-seismic monitoring; satellite InSAR monitoring; tracers in CO injection wells; core analysis (storage unit); soil and surface gas sam- pling; core analysis (cap rock unit) Rumaitha EOR Cross-well seismic; DTS (distributed temperature Al-Hajeri et al. (2010) and Figuera et al. (2014, 2016) Zone-B, Abu sensing in observation well, 41 meters from injec- Dhabi tor; permanent multi-phase flow meter (MPFM), logging tools, sponge coring near injector well; production data analysis pled with hydromechanics. Their results showed that of the aquifer encounters pressures capable of breaking the highest risk of over-pressures and fault reactivation the seal. were at the beginning of injection where fluid pressures 2. Abandoned wells: Another significant leakage pathway rise. Lindeberg et al. (2009) also noted the importance is abandoned wells; this leakage pathway is more plau- of the consideration of injection pressures in the pre- sible in a depleted hydrocarbon reservoir which has been vention of leakages through the cap rock. An engineer- used previously for the commercial production of hydro- ing strategy has been proposed by Eke et al. (2011) to carbon dioxides than in saline aquifers. This is because minimize the leakage of CO . In their paper, they argued depleted hydrocarbon dioxide reservoirs possess wells that surface mixing of C O with brine prior to injec- whose structural integrity might have degraded over tion could enhance the dissolution trapping mechanism. time. Degradation of wells could be as a result of cas- Subsequently, this would lead to a denser CO which is ing corrosion and reactions of the minerals with plug-in saturated with brine being injected into the reservoir. By materials or reservoir fluids which compromise integrity. implication, the strong buoyancy drive, typically expe- Human errors in the design of wells such as loose plugs rienced in aquifers, is minimized and the risk of C O could also create pathways for leakage of gases. Sev- leaking via prolonged contact of the CO with the seal eral studies have been conducted to assess the impact of is curtailed. It is therefore important before commenc- leakages through wells (Carey 2018; Kopp et al. 2010). ing any storage activity to perform a geomechanical 3. Faults and fractures: It is essential while performing site analysis in order to understand the fracturing pressure selection and characterization to ensure that there are no of the cap rock and thus avoid over-pressurization of the transmissive faults and fractures in the identified forma- aquifer. Large areal extents of a proposed aquifer could tion. Additionally, during the injection of CO , care must also mean that pressure propagates much faster, ensuring be taken to ensure that inactive faults are not activated that it takes a significant amount of time before the seal due to the high aquifer pressures. Fractures could also 1 3 1054 Petroleum Science (2019) 16:1028–1063 be developed in the cap rock if the temperature of the primary and secondary entrapments, (2) seismic activi- injected CO is much lower than the in situ temperature ties which could lead to the natural release of CO into the 2 2 in the aquifer. In the Abu Dhabi Rumaitha Zone-B pro- atmosphere, (3) fractures and faults which could lead to the ject, the CO is heated at the surface prior to injection, rapid release of C O , (4) abandoned and structurally weak 2 2 to ensure thermal induced fractures are not created in the wells which possess the ability to release large amounts of reservoir. CO back to the atmosphere and (5) release of C O that 2 2 rarely occurs through eruptive processes. 7.2 Induced seismicity Another postulated risk associated with CO storage is that 8 Conclusions of induced seismicity. The risk, however, has been proven to be negligible in field-scale projects that have been car - The risk of global warming is no longer hearsay. Several ried out due to the relatively small size of the projects and countries have accepted that our world is facing the risk of low injection rates. Nicol et al. (2011) noted that induced an endangered atmosphere and this must be addressed. The seismicity could lead to earthquakes that exceed magnitudes problem is not just a scientific one but also affects other of M6 and have the potential to impact on the containment, spheres of human endeavor. In this review, we provide the infrastructure and public perceptions of safety at C O stor- reader with the state of the art on carbon dioxide storage age sites. The possibility of the occurrence of a seismic science and technology. From a scientific viewpoint, the event would be higher if faults are present. This reiterates understanding of the processes involved in the process has the need for proper site characterization and identification of been greatly enhanced over the years with concrete informa- faults and fractures to avoid their reactivation and the pos- tion available on the fate of the injected C O before, during sible consequences of this reactivation (Oldenburg 2014). and after the injection phases. However, there are certain issues which we believe still need to be addressed before 7.3 Economic considerations the science can be considered full-fledged. The modeling procedures involved in carbon dioxide storage is multi-scale Carbon dioxide capture and carbon dioxide storage are two in both the temporal and spatial scales; we believe that for technologies that go hand-in-hand, hence the popular acro- the physics of the different level scales to be effectively nym CCS. The success of one process is dependent to a understood, the problem needs to be approached using multi- large extent on the success of the other. As such, it is nec- scale formulations. This would require the development of essary to state that the deployment of carbon dioxide stor- advanced numerical algorithms which are very robust and age projects would be greatly enhanced if carbon dioxide computationally efficient for best results. Improvements capture processes are also successful. The key economic in monitoring tools used at commercial CCS sites would issue associated with carbon dioxide capture processes is also go a long way toward validating scientific models and the high cost of the capture of CO from stationary power simulation predictions. An improvement in the capability of plants. In fact, most successful commercial deployment of monitoring and modeling tools implies that the risk of the carbon dioxide storage projects has pursued the option of leakage of C O is greatly reduced. It is obvious that these separating CO from produced gas rather than capturing could not be accomplished if the number of commercial CO from coal plants. This represents a cheaper option for CCS sites does not greatly increase. Governments would the companies involved. As with most burgeoning technol- need to establish and enforce policies such as carbon dioxide ogy, there is always a higher cost for companies which make pricing and taxation which would compel companies that the first step toward developing the technology before the would otherwise have considered the cheaper option of the technology improves and costs are reduced. For this reason, emission of C O directly into the atmosphere into consider- there is a reticence among companies to avoid making the ing CCS. first move. This disposition can be quelled by government In summary, a successful carbon dioxide storage project action in subsidizing the costs involved for the early movers, would involve accurate site selection, characterization (stor- thereby encouraging more participation. age capacity estimation, plume modeling) and monitoring Lewicki et al. (2007) made use of leakages of CO from to avoid the risks of leakages through seals, faults and aban- natural and industrial formations to analyze the features, doned wells. The site characterization would be successful events and processes (FEPs) of the leakages from both through the use of modeling and simulation tools whose natural and man-made sources. A total of 12 natural and 4 accuracy would be greatly enhanced through measurement, industrial analogues were looked into in their comparisons. monitoring and verification during the post-injection phase. They concluded at 5 FEPs which could lead to the release Carbon dioxide storage is a technology that has come to stay of stored C O in aquifers: (1) accumulation of C O beneath with the advantage of allowing the continued use of fossil 2 2 1 3 Petroleum Science (2019) 16:1028–1063 1055 integrating constant-concentration laboratory assay data with fuels while still saving our environment from the risks of variable-concentration field exposure. Environ Model Assess. global warming and therefore must be embraced by all. 1997;2(4):333–43. https ://doi.org/10.1023/a:10190 29931 755. Aydin G, Karakurt I, Aydiner K. Evaluation of geologic storage options Acknowledgements The authors gratefully acknowledge the research of CO : applicability, cost, storage capacity and safety. Energy support provided by the Department of Petroleum Engineering, Khalifa Policy. 2010;38(9):5072–80. https ://doi.or g/10.1016/j.en pol University of Science and Technology, Sas Al Nakhl Campus, Abu .2010.04.035. Dhabi, UAE. The corresponding author (AB) is thankful to the Drill- Bachu S, Adams JJ. Sequestration of CO in geological media in ing, Cementing, and Stimulation Research Center, School of Petroleum response to climate change: capacity of deep saline aqui- Technology, Pandit Deendayal Petroleum University, Raisan, Gandhi- fers to sequester C O in solution. Energy Convers Manag. nagar, Gujarat-382007, India, for supporting his research. Thanks are 2003;44(20):3151–75. https ://doi.or g/10.1016/S0196 also extended to other individuals who were, directly and indirectly, -8904(03)00101 -8. related to this work. Bachu S, Bonijoly D, Bradshaw J, Burruss R, Holloway S, Christensen NP, et al. CO storage capacity estimation: methodology and Open Access This article is distributed under the terms of the Crea- gaps. Int J Greenh Gas Control. 2007a;1(4):430–43. https ://doi. tive Commons Attribution 4.0 International License (http://creat iveco org/10.1016/S1750 -5836(07)00086 -2. mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Bachu S, Bonijoly D, Bradshaw J, Buruss R Christensen N, Holloway tion, and reproduction in any medium, provided you give appropriate S et al. Phase II, final report from the task force for review and credit to the original author(s) and the source, provide a link to the identification of standards for CO storage capacity estimation. Creative Commons license, and indicate if changes were made. In: Carbon sequestration leadership forum, Washington, United States; 2007b. p. 43. Basirat F, Fagerlund F, Denchik N, Pezard PA, Niemi A. Numeri- cal modelling of C O injection at small-scale field experi- References mental site in Maguelone, France. Int J Greenh Gas Control. 2016;54:200–10. https ://doi.org/10.1016/j.ijggc .2016.09.006. Adams EE, Caulfield JA, Herzog HJ, Auerbach DI. Impacts of reduced Bellefleur G, Adam L, White D, Mattocks B, Davis T. Seismic imaging pH from ocean C O disposal: sensitivity of zooplankton mortal- and anisotropy analysis of 9C 3D-VSP data at Weyburn Field, ity to model parameters. Waste Manag. 1998a;17(5–6):375–80. Saskatchewan, Canada. In: SEG technical program expanded https ://doi.org/10.1016/s0956 -053x(97)10038 -1. abstracts, Society of Exploration Geophysicists; 2003, p. Adams EE, Caulfield JA, Herzog HJ, Auerbach DI. Impacts of reduced 1326–29. pH from ocean C O disposal: sensitivity of zooplankton mortal- Belhaj H, Bera A. A brief review of mechanisms for carbon dioxide ity to model parameters. Waste Manag. 1998b;17(5–6):375–80. sequestration into aquifer reservoirs. Int J Pet Eng. 2017;3(1):49– https ://doi.org/10.1016/s0956 -053x(97)10038 -1. 66. https ://doi.org/10.1504/IJPE.2017.08899 6. Ajo-Franklin JB, Peterson J, Doetsch J, Daley TM. High-resolu- Bennion DB, Bachu S. The impact of interfacial tension and pore size tion characterization of a CO plume using crosswell seismic distribution/capillary pressure character on C O relative perme- 2 2 tomography: Cranfield MS, USA. Int J Greenh Gas Control. ability at reservoir conditions in C O -brine systems. In: SPE/ 2013;18:497–509. https ://doi.org/10.1016/j.ijggc .2012.12.018. DOE symposium on improved oil recovery, Society of Petroleum Al-Hajeri SK, Negahban S, Bin-dhaaer GS, Al-Basry AH. Design and Engineers; 2006. https ://doi.org/10.2118/99325 -MS. implementation of the first CO -EOR pilot in Abu Dhabi, UAE. Benson SM, Cole DR. CO sequestration in deep sedimentary forma- 2 2 In: SPE EOR conference at oil and gas, West Asia; 2010. https tions. Elements. 2008;4(5):325–31. https://doi.or g/10.2113/gsele ://doi.org/10.2118/12960 9-MS.ments .4.5.325. Alfi M, Hosseini SA, Alfi M, Shakiba M. Effectiveness of 4D seismic Benson SM, Hingerl F, Li B, Pini R, Tchelepi H, Zuo L. Investigations data to monitor CO plume in Cranfield CO -EOR project. In: in geologic carbon sequestration: multiphase flow of CO and 2 2 2 Carbon management technology conference; 2015. https ://doi. water in reservoir rocks; 2013. org/10.7122/43955 9-MS. Bestehorn M, Firoozabadi A. Effect of fluctuations on the onset Ambrose WA, Lakshminarasimhan S, Holtz MH, Núñez-López V, of density-driven convection in porous media. Phys Fluids. Hovorka SD, Duncan I. Geologic factors controlling C O stor- 2012;24(11):114102. https ://doi.org/10.1063/1.47674 67. age capacity and permanence: case studies based on experience Bethke C, Yeakel S. Geochemist’s workbench: release 8.0 reference with heterogeneity in oil and gas reservoirs applied to CO stor- manual; 2009. age. Environ Geol. 2017;54(8):1619–33. https://doi.or g/10.1007/ Bonneville A, Gilmore T, Sullivan C, Vermeul V, Kelley M, White S, s0025 4-007-0940-2. et al. Evaluating the suitability for CO storage at the Future- Aminu MD, Nabavi SA, Rochelle CA, Manovic V. A review of develop- Gen 2.0 Site, Morgan County, Illinois, USA. Energy Procedia. ments in carbon dioxide storage. Appl Energy. 2017;208:1389– 2013;37:6125–32. https://doi.or g/10.1016/j.egypro.2013.06.541 . 419. https ://doi.org/10.1016/j.apene rgy.2017.09.015. Boreham C, Underschultz J, Stalker L, Kirste D, Freifeld B, Jenkins Ammer J, Brummert A. Miscible applied simulation techniques for C, et al. Monitoring of CO storage in a depleted natural gas energy recovery-version 2.0; 1991. reservoir: gas geochemistry from the C O CRC Otway Project, Arts R, Eiken O, Chadwick A, Zweigel P, van der Meer L, Zinszner B. Australia. Int J Greenh Gas Control. 2011;5(4):1039–54. https:// Monitoring of C O injected at Sleipner using time-lapse seismic doi.org/10.1016/j.ijggc .2011.03.011. data. Energy. 2004;29(9–10):1383–92. https://doi.or g/10.1016/j. Bourne S, Crouch S, Smith M. A risk-based framework for measure- energ y.2004.03.072. ment, monitoring and verification of the Quest CCS Project, Assteerawatt A, Bastian P, Bielinski A, Breiting T, Class H, Ebigbo Alberta, Canada. Int J Greenh Gas Control. 2014;26:109–26. A, et  al. MUFTE-UG: structure, applications and numerical https ://doi.org/10.1016/j.ijggc .2014.04.026. methods. Newsl Int Groundw Model Centre Colo Sch Mines. Braathen A, Bælum K, Christiansen HH, Dahl T, Eiken O, Elvebakk H, 2005;23(2):10. et al. The Longyearbyen CO Lab of Svalbard, Norway—initial Auerbach DI, Caulfield JA, Adams EE, Herzog HJ. Impacts of ocean assessment of the geological conditions for C O sequestration. CO disposal on marine life. I. A toxicological assessment Nor J Geol. 2012;92(4):353–76. 1 3 1056 Petroleum Science (2019) 16:1028–1063 Bromhal GS, Neal Sams W, Jikich S, Ertekin T, Smith DH. Simu- Cole S, Itani S. The Alberta carbon trunk line and the benefits of CO . lation of CO sequestration in coal beds: the effects of sorp- Energy Procedia. 2013;37:6133–9. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . tion isotherms. Chem Geol. 2005;217(3–4):201–11. https ://doi.egypr o.2013.06.542. org/10.1016/j.chemg eo.2004.12.021. Coninck HD, Loos M, Metz B, Davidson O, Meyer L. IPCC special Broseta D, Tonnet N, Shah V. Are rocks still water-wet in the presence report on carbon dioxide capture and storage. Intergovernmental of dense C O or H S? Geofluids. 2012;12(4):280–94. https://doi. Panel on Climate Change; 2005. 2 2 org/10.1111/j.1468-8123.2012.00369 .x. Delshad M, Kong X, Wheeler MF. On interplay of capillary, gravity, Brown B, Carr T, Vikara D. Monitoring, verification, and accounting and viscous forces on brine/CO relative permeability in a com- of CO stored in deep geologic formations. US Department of positional and parallel simulation framework. In: SPE reservoir Energy National Energy Technology Laboratory; 2009. simulation symposium, Society of Petroleum Engineers; 2011. Brydie J, Jones D, Jones JP, Perkins E, Rock L, Taylor E. Assessment https ://doi.org/10.2118/14214 6-MS. of baseline groundwater physical and geochemical properties for DePaolo DJ, Cole DR, Navrotsky A, Bourg IC. Geochemistry of geo- the Quest carbon capture and storage project, Alberta, Canada. logic CO sequestration, 77. Walter de Gruyter GmbH & Co Energy Procedia. 2014;63:4010–8. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . KG; 2019. egypr o.2014.11.431. Ding S, Xi Y, Jiang H, Liu G. CO storage capacity estimation Burnside NM, Naylor M. Review and implications of relative perme- in oil reservoirs by solubility and mineral trapping. Appl ability of C O /brine systems and residual trapping of C O . Int Geochem. 2018;89:121–8. https ://doi.or g/10.1016/j.apg eo 2 2 J Greenh Gas Control. 2014;23:1–11. https ://doi.org/10.1016/j. chem.2017.12.002. ijggc .2014.01.013. DOE-NETL. Carbon storage Atlas Fifth Edition. U.S. Department of Busch A, Gensterblum Y. CBM and CO -ECBM related sorption pro- Energy Office of Fossil Energy, National Energy Technology cesses in coal: a review. Int J Coal Geol. 2011;7(2):49–71. https Laboratory; 2015. ://doi.org/10.1016/j.coal.2011.04.011. DOE U. Carbon sequestration ATLAS of the United States and Canada. Carey JW. Probability distributions for effective permeability of Office of Fossil Energy, National Energy Technology Laboratory, potentially leaking wells at C O sequestration sites. Los Ala- Morgantown, WV, 90p; 2007. mos National Lab (LANL), Los Alamos, NM (United States); Durucan S, Shi JQ, Korre A. Numerical modeling and prediction of 2018. abandoned mine methane recovery: field application at the Saar Carrigan CR, Yang X, LaBrecque DJ, Larsen D, Freeman D, Ram- Coalfield, Germany. Geol Belg. 2004;7(3–4):207–13. irez AL, et al. Electrical resistance tomographic monitoring Ebigbo A, Bielinski A, Kopp A, Class H, Helmig R. Numerical mod- of CO movement in deep geologic reservoirs. Int J Greenh eling of CO sequestration with MUFTE-UG. CMWR XVI, 2 2 Gas Control. 2013;18:401–8. https ://doi.org/10.1016/j.ijggc Copenhagen; 2006. .2013.04.016. Eke PE, Naylor M, Curtis A, Haszeldine S. CO leakage prevention Cavanagh A. Benchmark calibration and prediction of the Sleipner C O technologies. In: Offshore Europe, Society of Petroleum Engi- plume from 2006 to 2012. Energy Procedia. 2013;37:3529–45. neers; 2011. https ://doi.org/10.2118/14526 3-MS. https ://doi.org/10.1016/j.egypr o.2013.06.246. El-Kaseeh G, Will R, Balch R, Grigg R. Multi-scale seismic measure- Cavanagh A, Ringrose P. In Salah high-resolution heterogeneous simu- ments for C O Monitoring in an EOR/CCUS project. Energy lations of CO storage. Geoscience. 2010;5:53–63. Procedia. 2017;114:3656–70. https ://doi.or g/10.1016/j.egypr Cavanagh A, Ringrose P. Simulation of CO distribution at the In o.2017.03.1497. Salah storage site using high-resolution field-scale models. Ennis-King J, Paterson L. Role of convective mixing in the long-term Energy Procedia. 2011;4:3730–7. https://doi.or g/10.1016/j.egypr storage of carbon dioxide in deep saline formations. In: SPE o.2011.02.306. annual technical conference and exhibition, Society of Petroleum Cavanagh AJ, Haszeldine RS. The Sleipner storage site: capillary Engineers; 2003. https ://doi.org/10.2118/84344 -MS. flow modeling of a layered CO plume requires fractured shale Estublier A, Lackner AS. Long-term simulation of the Snøhvit barriers within the Utsira Formation. Int J Greenh Gas Control. CO storage. Energy Procedia. 2009;1(1):3221–8. https ://doi. 2014;21:101–12. https ://doi.org/10.1016/j.ijggc .2013.11.017.org/10.1016/j.egypr o.2009.02.106. Celia MA, Bachu S, Nordbotten JM, Bandilla KW. Status of CO Etheridge D, Luhar A, Loh Z, Leuning R, Spencer D, Steele P, et al. storage in deep saline aquifers with emphasis on modeling Atmospheric monitoring of the CO2CRC Otway Project and approaches and practical simulations. Water Resour Res. lessons for large scale C O storage projects. Energy Procedia. 2015;51(9):6846–92. https ://doi.org/10.1002/2015w r0176 09. 2011;4:3666–75. https ://doi.org/10.1016/j.egypr o.2011.02.298. Chadwick RA, Zweigel P, Gregersen U, Kirby GA, Holloway S, Fan Y. Development of CO sequestration modeling capabilities in Johannessen PN. Geological reservoir characterization of a CO Stanford general purpose research simulator. MS report, Depart- storage site: The Utsira Sand, Sleipner, Northern North Sea. ment of Petroleum Engineering, Stanford University, California; Energy. 2004;29(9–10):1371–81. https: //doi.org/10.1016/j.energ 2006. y.2004.03.071. Fang Z. Mitigate drilling risks in highly depleted reservoir fields by Chiquet P, Broseta D, Thibeau S. Wettability alteration of caprock intelligently planning well trajectory for a C O sequestration pro- minerals by carbon dioxide. Geofluids. 2007;7(2):112–22. https ject. In: Brasil offshore, Society of Petroleum Engineers; 2011. ://doi.org/10.1111/j.1468-8123.2007.00168 .x.https ://doi.org/10.2118/14312 9-MS. Christiansen AC. Climate policy and dynamic efficiency gains A case Fang Z, Khaksar A. Geomechanical issues and solutions for CO study on Norwegian C O -taxes and technological innovation in sequestration in depleted hydrocarbon sandstone reservoirs. the petroleum sector. Clim Policy. 2001;1(4):499–515. https :// In: 46th US rock mechanics/geomechanics symposium, doi.org/10.3763/cpol.2001.0150. American Rock Mechanics Association; 2012. https ://doi. Class H, Ebigbo A, Helmig R, Dahle HK, Nordbotten JM, Celia MA, org/10.2118/13399 4-PA. et al. A benchmark study on problems related to CO storage in Farajzadeh R, Zitha PLJ, Bruining J. Enhanced mass transfer of geologic formations. Comput Geosci. 2009;13(4):409–34. https CO into water: experiment and modeling. Ind Eng Chem Res. ://doi.org/10.1007/s1059 6-009-9146-x. 2009;48(13):6423–31. https ://doi.org/10.1021/ie801 521u. Figuera LA, Al-basry AH, Al-Hammadi KE, Al-Yafei A, Sakaria D, Tanakov MY. Complex phased development for C O EOR in oil 1 3 Petroleum Science (2019) 16:1028–1063 1057 carbonate reservoir, Abu Dhabi Onshore. In: Abu Dhabi interna- Grobe M, Pashin JC, Dodge RL. Carbon dioxide sequestration in geo- tional petroleum exhibition and conference, Society of Petroleum logical media: state of the science. In: AAPG studies in geol- Engineers; 2014. https ://doi.org/10.2118/17196 7-MS. ogy. American Association of Petroleum Geologists, Tulsa, OK; Figuera L, Al Hammadi K, Tanakov M. Case study of C O injection to 2009. p. xi, 715. enhance oil recovery into the transition zone of a tight carbonate Gunter W, Bachu S, Perkins E. Aquifer disposal of C O -rich gases in reservoir. In: Abu Dhabi international petroleum exhibition and the vicinity of the Sundance and Genesee power plants, phase I: conference, SPE Society of Petroleum Engineers; 2016. https :// injectivity, chemical reactions and proof of concept. Edmonton doi.org/10.2118/18320 3-MS. (AB): Alberta Research Council [ARC]: 1994-16; 1993. Finley RJ. An overview of the Illinois Basin—Decatur Project. Greenh Gunter WD, Gentzis T, Rottenfusser BA, Richardson RJH. Deep coal- Gases Sci Technol. 2014;4(5):571–9. https ://doi.org/10.1002/ bed methane in Alberta, Canada: a fuel resource with the poten- ghg.1433. tial of zero greenhouse gas emissions. Energy Convers Manag. Flemisch B, Fritz J, Helmig R, Niessner J, Wohlmuth B. DUMUX: 1997;38:S217–22. https ://doi.or g/10.1016/S0196-8904(96)00272 a multi-scale multi-physics toolbox for flow and transport pro- -5. cesses in porous media, ECCOMAS thematic conference on Gunter W, Perkins E, Hutcheon I. Aquifer disposal of acid gases: mod- multi-scale computational methods for solids and fluids, Cachan, elling of water–rock reactions for trapping of acid wastes. Appl France; 2007. p. 82–7. Geochem. 2000;15(8):1085–95. https ://doi.org/10.1016/s0883 Flett M, Brantjes J, Gurton R, McKenna J, Tankersley T, Trupp M. -2927(99)00111 -0. Subsurface development of C O disposal for the Gorgon Project. Gunter WD, Mavor MJ, Robinson JR. CO storage and enhanced meth- 2 2 Energy Procedia. 2009;1(1):3031–8. https ://doi.org/10.1016/j. ane production: field testing at Fenn-Big Valley, Alberta, Canada, egypr o.2009.02.081. with application. Greenh Gas Control Technol. 2005;7:413–21. Friedmann SJ, Stamp VW. Teapot dome: characterization of a Haghighat SA, Mohaghegh SD, Gholami V, Shahkarami A, Moreno CO -enhanced oil recovery and storage site in Eastern Wyoming. DA. Using big data and smart field technology for detecting leak - Environ Geosci. 2006;13(3):181–99. https ://doi.org/10.1306/ age in a CO storage project. In: SPE annual technical conference eg.01200 60501 7. and exhibition, Society of Petroleum Engineers; 2013. https ://doi. Gasda SE. Numerical models for evaluating CO storage in deep saline org/10.2118/16613 7-ms. aquifers: leaky wells and large-scale geological features. Prince- Han WS. Evaluation of C O trapping mechanisms at the SACROC ton University, 2010. northern platform: site of 35 years of CO injection, Citeseer; Gasda SE, Nordbotten JM, Celia MA. Vertical equilibrium with sub- 2008. scale analytical methods for geological CO sequestration. Com- Hansen O, Gilding D, Nazarian B, Osdal B, Ringrose P, Kristoffersen put Geosci. 2009;13(4):469–81. https ://doi.org/10.1007/s1059 J-B, et al. Snøhvit: the history of injecting and storing 1 Mt 6-009-9138-x. CO in the fluvial Tubåen Fm. Energy Procedia. 2013;37:3565– Gasda SE, Nordbotten JM, Celia MA. Vertically averaged approaches 73. https ://doi.org/10.1016/j.egypr o.2013.06.249. for CO migration with solubility trapping. Water Resour Res. Hao Y, Sun Y, Nitao J. Overview of NUFT: A versatile numeri- 2011. https ://doi.org/10.1029/2010w r0090 75. cal model for simulating flow and reactive transport in porous Gemmer L, Hansen O, Iding M, Leary S, Ringrose P. Geomechani- media; 2012. p. 212–39. cal response to C O injection at Krechba, InSalah, Algeria. Hassanzadeh H, Pooladi-Darvish M, Keith DW. Stability of a fluid First Break. 2011;30(2):79–84. https ://doi.org/10.3997/2214- in a horizontal saturated porous layer: effect of non-linear 4609.20144 113. concentration profile, initial, and boundary conditions. Transp Gershenzon NI, Soltanian M, Ritzi RW, Dominic DF. Influence of Porous Med. 2006;65(2):193–211. h t t p s : / / do i . o r g /1 0 . 1 0 0 7/ small scale heterogeneity on CO trapping processes in deep s1124 2-005-6088-1. saline aquifers. Energy Procedia. 2014;2014(59):166–73. https Hassanzadeh H, Pooladi-Darvish M, Keith DW. Scaling behavior of ://doi.org/10.1016/j.egypr o.2014.10.363. convective mixing, with application to geological storage of Ghomian Y, Pope GA, Sepehrnoori K. Hysteresis and field-scale opti- CO . AIChE J. 2007;53(5):1121–31. https ://doi.org/10.1002/ mization of WAG injection for coupled C O -EOR and sequestra- aic.11157 . tion. In: SPE symposium on improved oil recovery, Society of Haszeldine RS. Carbon capture and storage: How green can black Petroleum Engineers; 2008. https://doi.or g/10.2118/110639-ms . be? Science. 2009;325(5948):1647–52. Gislason SR, Oelkers EH. Geochemistry: carbon storage in basalt. Hellevang H, Kvamme B. ACCRETE-Geochemistry solver for Science. 2014;344(6182):373–4. https ://doi.org/10.1126/scien CO -water-rock interactions. In: Proceedings GHGT 8 con- ce.12508 28. ference; 2006. Global CCS. Institute, The global status of CCS; 2017. Hellevang H, Kvamme B. An explicit and efficient algorithm to solve Goodarzi S, Settari A, Zoback M, Keith DW. Thermal effects on shear kinetically constrained C O –water–rock interactions. WSEAS fracturing and injectivity during C O storage. In: ISRM interna- Trans Math. 2007;6(5):681. tional conference for effective and sustainable hydraulic fractur - Herzog H, Drake E, Adams E. CO capture, reuse, and storage tech- ing, International Society for Rock Mechanics; 2013. https: //doi. nologies for mitigating global climate change. A white paper; org/10.5772/56311 . 1997. Gorecki CD, Hamling JA, Ensrud J, Steadman EN, Harju JA. Integrat- Herzog H, Eliasson B, Kaarstad O. Capturing greenhouse gases. Sci ing CO EOR and C O storage in the Bell Creek oil field. In: Am. 2000;282(2):72–9. 2 2 Carbon management technology conference; 2012. https ://doi. Hesse M, Tchelepi HA, Orr FM. Scaling analysis of the migration of org/10.7122/15147 6-ms. CO in saline aquifers. In: SPE annual technical conference and Gozalpour F, Ren SR, Tohidi B. C O EOR and storage in oil reservoirs. exhibition, Society of Petroleum Engineers; 2006. https ://doi. Oil Gas Sci Technol Revue D IFP Energies N. 2005;60(3):537–org/10.2118/10279 6-ms. 46. https ://doi.org/10.2516/ogst:20050 36. Hosseini SA, Lashgari H, Choi JW, Nicot J-P, Lu J, Hovorka SD. Static Graupner BJ, Li D, Bauer S. The coupled simulator ECLIPSE– and dynamic reservoir modeling for geological CO sequestra- OpenGeoSys for the simulation of CO storage in saline for- tion at Cranfield, Mississippi, USA. Int J Greenh Gas Control. mations. Energy Procedia. 2011;4:3794–800. h t t p s : / / d o i . 2013;18:449–62. https ://doi.org/10.1016/j.ijggc .2012.11.009. org/10.1016/j.egypr o.2011.02.314. 1 3 1058 Petroleum Science (2019) 16:1028–1063 Houdu E, Poupard O, Meyer V. Supercritical C O leakage modelling Kim S, Hosseini SA. Above-zone pressure monitoring and geomechan- for well integrity in geological storage project. In: Proceedings ical analyses for a field-scale CO injection project in Cranfield, of the COMSOL conference; 2008. MS. Greenh Gases Sci Technol. 2014;4(1):81–98. https ://doi. Hovorka SD, Benson SM, Doughty C, Freifeld BM, Sakurai S, Daley org/10.1002/ghg.1388. TM, et al. Measuring permanence of C O storage in saline for- Koide H, Tazaki Y, Noguchi Y, Nakayama S, Iijima M, Ito K, et al. mations: the Frio experiment. Environ Geosci. 2006;13(2):105– Subterranean containment and long-term storage of carbon- 21. https ://doi.org/10.1306/eg.11210 50501 1. dioxide in unused aquifers and in depleted natural-gas reser- Hovorka SD, Meckel TA, Treviño RH. Monitoring a large-volume voirs. Energy Convers Manag. 1992;33(5–8):619–26. https :// injection at Cranfield, Mississippi—project design and recom-doi.org/10.1016/0196-8904(92)90064 -4. mendations. Int J Greenh Gas Control. 2013;18:345–60. https :// Kolditz O, Kaiser R, Habbar D, Rother T, Thorenz C. ROCKFLOW- doi.org/10.1016/j.ijggc .2013.03.021. theory and users manual, release 3.9. Groundwater Group, Center Huo D, Gong B. Discrete modeling and simulation on potential leakage for Applied Geosciences, University of Tübingen, and Institute through fractures in C O sequestration. In: SPE annual techni- of Fluid Mechanics, University of Hannover; 2003. cal conference and exhibition, Society of Petroleum Engineers; Kong X. Petrophysical modeling and simulation study of geological 2010. https ://doi.org/10.2118/13550 7-ms. CO sequestration.  University of Texas at Austin, 2014. IEAGHG. Development of storage coefficients for CO storage in Deep Kong X, Delshad M, Wheeler M. An integrated capillary, buoyancy, Saline Formations. In: International energy agency greenhouse and viscous-driven model for brine/CO relative permeability in gas R&D programme; 2009. a compositional and parallel reservoir simulator, modelling and IPCC A. Intergovernmental panel on climate change. IPCC Secretariat simulation in fluid dynamics in porous media, Springer, 2013. p. Geneva; 2007. 125–42. https ://doi.org/10.1007/978-1-4614-5055-9_8. Iskhakov R. High-resolution numerical simulation of C O sequestration Kong XH, Delshad M, Wheeler MF. History matching heterogeneous in saline aquifers. Stanford University; 2013. coreflood of CO /brine by use of compositional reservoir simula- Ivanova A, Kashubin A, Juhojuntti N, Kummerow J, Henninges tor and geostatistical approach. SPE J. 2015;20(2):267–76. https J, Juhlin C, et  al. Monitoring and volumetric estimation of ://doi.org/10.2118/16362 5-PA. injected CO using 4D seismic, petrophysical data, core meas- Kongsjorden H, Kårstad O, Torp TA. Saline aquifer storage of carbon urements and well logging: a case study at Ketzin, Germany. dioxide in the Sleipner project. Waste Manag. 1998;17(5):303–8. Geophys Prospect. 2012;60(5):957–73. https ://doi.org/10.111https ://doi.org/10.1016/s0956 -053x(97)10037 -x. 1/j.1365-2478.2012.01045 .x. Kopp A, Binning PJ, Johannsen K, Helmig R, Class H. A contribution Jalil M, Masoudi R, Darman NB, Othman M. Study of the CO injec- to risk analysis for leakage through abandoned wells in geologi- tion, storage, and sequestration in depleted M4 carbonate gas cal CO storage. Adv Water Resour. 2010;33(8):867–79. https:// condensate reservoir, Malaysia. In: Carbon management technol-doi.org/10.1016/j.advwa tres.2010.05.001. ogy conference; 2012. https ://doi.org/10.7122/15005 0-ms. Korbøl R, Kaddour A. Sleipner vest C O disposal-injection of James T. Catch me if you can. Eng Technol. 2013;8(2):56–8.removed CO into the Utsira formation. Energy Convers Jiang X. A review of physical modelling and numerical simula- Manag. 1995;36(6–9):509–12. https ://doi.or g/10.1016/0196- tion of long-term geological storage of C O . Appl Energy. 8904(95)00055 -I. 2 0 1 1 ; 8 8 ( 1 1 ) : 35 5 7 – 6 6 . h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . a p e n e Kumar A, Noh MH, Ozah RC, Pope GA, Bryant SL, Sepehrnoori K, rgy.2011.05.004. et al. Reservoir simulation of C O storage in aquifers. SPE J. Juanes R, Spiteri EJ, Orr FM, Blunt MJ. Impact of relative permeability 2005;10(03):336–48. https ://doi.org/10.2118/89343 -pa. hysteresis on geological CO storage. Water Resour Res. 2006. Kvamme B, Liu S. A new reactive transport reservoir simulator for https ://doi.org/10.1029/2005W R0048 06. aquifer storage of CO —with implicit geomechanical analysis. Jung JW, Wan JM. Supercritical CO and ionic strength effects on Carbon dioxide capture for storage in Deep Geologic Formations; wettability of silica surfaces: equilibrium contact angle meas- 2009. p. 349–76. urements. Energy Fuels. 2012;26(9):6053–9. https ://doi. Lal R. Carbon sequestration. Philos Trans R Soc Lond B Biol Sci. org/10.1021/ef300 913t. 2008;363(1492):815–30. https://doi.or g/10.1098/rstb.2007.2185. Kaldi JG, Gibson-Poole CM, Payenberg TH. Geological input to selec- Lamy CMM, Iglauer S, Pentland CH, Blunt MJ, Maitland GC. Cap- tion and evaluation of CO geosequestration sites; 2009. illary trapping in carbonate rocks. In: SPE EUROPEC/EAGE Kampman N, Bickle M, Wigley M, Dubacq B. Fluid flow and CO – annual conference and exhibition, Society of Petroleum Engi- fluid–mineral interactions during CO -storage in sedimentary neers; 2010. https ://doi.org/10.2523/13072 0-ms. basins. Chem Geol. 2014;369:22–50. https ://doi.org/10.1016/j. Lasseter T, Waggoner J, Lake L. Reservoir heterogeneities and their chemg eo.2013.11.012. influence on ultimate recovery. Reserv Charact. 1986;545:1. https Kane RL, Klein DE. Carbon sequestration: an option for miti-://doi.org/10.1016/b978-0-12-43406 5-7.50022 -8. gating global climate change. In: Environmental chal- Law DH-S, Bachu S. Hydrogeological and numerical analysis of lenges and greenhouse gas control for fossil fuel utilization CO disposal in deep aquifers in the Alberta sedimentary basin. in the 21st century, Springer; 2002. p. 75–88. https ://doi. Energy Convers Manag. 1996;37(6):1167–74. https ://doi. org/10.1007/978-1-4615-0773-4_6.org/10.1016/0196-8904(95)00315 -0. Kempka T, Kuhn M. Numerical simulations of CO arrival times and Law DH-S, van der Meer L, Gunter W. Comparison of numerical simu- reservoir pressure coincide with observations from the Ketzin lators for greenhouse gas storage in coalbeds, Part IV; History pilot site, Germany. Environ Earth Sci. 2013;70(8):3675–85. match of field micro-pilot test data. In: The 7th international https ://doi.org/10.1007/s1266 5-013-2614-6. conference on greenhouse gas control technologies, Vancouver Kiessling D, Schmidt-Hattenberger C, Schuett H, Schilling F, Krue- BC, Canada, September 5–9; 2004. p. 2239–42. ger K, Schoebel B, et al. Geoelectrical methods for monitoring Le Gallo Y, Trenty L, Michel A, Vidal-Gilbert S, Parra T, Jeannin L. geological CO storage: first results from cross-hole and surface– Long-term flow simulations of CO storage in saline aquifer. In: 2 2 downhole measurements from the CO SINK test site at Ketzin The GHGT8 conference-trondheim (Norway); 2006. p. 18–22. (Germany). Int J Greenh Gas Control. 2010;4(5):816–26. https Lee K, Park M.-H, Kim Y, Browne G, Kaldi J. CO sequestration in ://doi.org/10.1016/j.ijggc .2010.05.001. deep sedimentary formations of the southwestern margin of the Ulleung Basin, offshore, east sea, Korea, OCEANS, 2012-Yeosu. 1 3 Petroleum Science (2019) 16:1028–1063 1059 IEEE; 2012. p. 1–4. https ://doi.or g/10.1109/ocean s-y eosu mineral trapping in C O geological storage. Environ Earth Sci. .2012.62634 21. 2011;62(3):507–17. https: //doi.org/10.1007/s12665 -010-0543-1. Lewicki JL, Birkholzer J, Tsang CF. Natural and industrial analogues Lu J, Kharaka YK, Thordsen JJ, Horita J, Karamalidis A, Griffith C, for leakage of CO from storage reservoirs: identification of fea-et al. CO –rock–brine interactions in Lower Tuscaloosa For- 2 2 tures, events, and processes and lessons learned. Environ Geol. mation at Cranfield CO sequestration site, Mississippi, USA. 2007;52(3):457–67. https: //doi.org/10.1007/s00254 -006-0479-7. Chem Geol. 2012;291:269–77. https ://doi.org/10.1016/j.chemg Li-ping H, Ping-ping S, Xin-wei L, Qi-Chao G, Cheng-sheng W, eo.2011.10.020. Fangfang L. Study of CO EOR and its geological seques- Malik QM, Islam M. C O injection in the Weyburn field of Canada: 2 2 tration potential in oil field around Yulin city. J Pet Sci Eng. optimization of enhanced oil recovery and greenhouse gas stor- 2015;134:199–204. https://doi.or g/10.1016/j.petrol.2015.06.002 . age with horizontal wells. In: SPE/DOE improved oil recovery Li B, Benson SM. Small-scale heterogeneities and buoyancy- symposium, Society of Petroleum Engineers; 2000. https ://doi. driven CO migration in geological storage. Energy Procedia. org/10.2118/59327 -ms. 2014;63:3608–15. https://doi.or g/10.1016/j.egypro.2014.11.391 . Manik J, Ertekin T, Kohler T. Development and validation of a com- Li B, Benson SM, Tchelepi HA. Modeling fine-scale capillary het- positional coalbed simulator. In: Canadian international petro- erogeneity in multiphase flow of CO and brine in sedimentary leum conference, Petroleum Society of Canada; 2000. https://doi. rocks. In: The XIX international conference on water resources, org/10.2118/2000-044. University of Illinois at Urbana-Champaign, IL, USA; 2012. p. Marckmann H, Jaeger P, Eggers R. Interfacial phenomena in counter- 17–22. current processes using supercritical fluids. In: Proceedings of Li D, Bauer S, Benisch K, Graupner B, Beyer C. OpenGeoSys-Che- the international conference on supercritical fluids, Versailles, mApp: a coupled simulator for reactive transport in multiphase France; 2003. p. 26–30. https ://doi.or g/10.1016/b978-04445 systems: code development and application at a representative 1574-2/50015 -8. CO storage formation in Northern Germany. Acta Geotech. Martens S, Kempka T, Liebscher A, Lüth S, Möller F, Myrttinen A, 2014;9(1):67–79. https ://doi.org/10.1007/s1144 0-013-0234-7. et al. Europe’s longest-operating on-shore CO storage site at Li C, Zhang KN, Wang YS, Guo CB, Maggi F. Experimental and Ketzin, Germany: a progress report after three years of injection. numerical analysis of reservoir performance for geological C O Environ Earth Sci. 2012;67(2):323–34. https ://doi.org/10.1007/ storage in the Ordos Basin in China. Int J Greenh Gas Control. s1266 5-012-1672-5. 2016;45:216–32. https ://doi.org/10.1016/j.ijggc .2015.11.011. Martens S, Liebscher A, Möller F, Henninges J, Kempka T, Lüth S, Liang Z, Shu W, Li Z, Shaoran R, Qing G. Assessment of CO EOR et al. CO storage at the Ketzin pilot site, Germany: fourth year 2 2 and its geo-storage potential in mature oil reservoirs, Shengli of injection, monitoring, modelling and verification. Energy Oilfield, China. Pet Explor Dev. 2009;36(6):737–42. https://doi. Procedia. 2013;37:6434–43. https ://doi.or g/10.1016/j.egypr org/10.7122/15003 1-ms. o.2013.06.573. Lindeberg E, Vuillaume J-F, Ghaderi A. Determination of the Mathias SA, Hardisty PE, Trudell MR, Zimmerman RW. Approxi- CO storage capacity of the Utsira Formation. Energy Pro- mate solutions for pressure buildup during CO injection in 2 2 cedia. 2009;1(1):2777–84. https ://doi.or g/10.1016/j.egypr brine aquifers. Transp Porous Med. 2009a;79(2):265–84. https o.2009.02.049.://doi.org/10.1007/s1124 2-008-9316-7. Litynski JT, Klara SM, McIlvried HG, Srivastava RD. An overview Mathias SA, Hardisty PE, Trudell MR, Zimmerman RW. Screening of terrestrial sequestration of carbon dioxide: The United States and selection of sites for C O sequestration based on pressure Department of Energy’s fossil energy R&D program. Clim buildup. Int J Greenh Gas Control. 2009b;3(5):577–85. Change. 2006;74(1–3):81–95. https ://doi.or g/10.1007/s1058 Mathieson A, Midgley J, Dodds K, Wright I, Ringrose P, Saoul N. 4-005-6960-6. CO sequestration monitoring and verification technologies Liu Y-F, Li X-C, Bai B. Preliminary estimation of CO storage capacity applied at Krechba, Algeria. Lead Edge. 2010;29(2):216–22. of coalbeds in China. J Rock Mech Eng. 2005;24(16):2947–52.https ://doi.org/10.1190/1.33048 27. Liu HW, Tellez BG, Atallah T, Barghouty M. The role of CO cap- Melikadze G, Körting O, Kapanadze N, Müller B, Todadze M, Jim- ture and storage in Saudi Arabia’s energy future. Int J Greenh sheladze T, et al. Investigation of carbon dioxide fluxes and Gas Control. 2012;11:163–71. https ://doi.org/10.1016/j.i jggc possibility its storage in Georgia. J Georgian Geophys Soc .2012.08.008. Issue A Phys Solid Earth. 2013;16a:32–6. Liu F, Ellett K, Xiao Y, Rupp JA. Assessing the feasibility of C O Metz, B, editor. Carbon dioxide capture and storage: IPCC spe- storage in the New Albany Shale (Devonian–Mississippian) with cial report. Summary for policymakers: a report of working potential enhanced gas recovery using reservoir simulation. Int J group III of the IPCC. Technical summary: a report accepted Greenh Gas Control. 2013;17:111–26. https://doi.or g/10.1016/j. by working group III of the IPCC but not approved in detail. ijggc .2013.04.018. Intergovernmental panel on climate change and storage; 2005. Liu L-C, Li Q, Zhang J-T, Cao D. Toward a framework of environmen- Metz, B, Davidson, O, de Coninck, H, Loos, M, Meyer, L. Carbon tal risk management for CO geological storage in China: gaps dioxide capture and storage, intergovernmental panel on cli- and suggestions for future regulations. Mitig Adaptat Strateg mate change, Geneva (Switzerland). Working group III; 2005. Glob Change. 2016;21(2):191–207. Mingjun Z, Chongtao W, Haiyang P, Jia C. Productivity of coalbed Lu C, Lichtner PC. PFLOTRAN: massively parallel 3-D simulator methane wells in southern of Qinshui Basin. Min Sci Technol for CO sequestration in geologic media. In: DOE-NETL fourth (China). 2010;20(5):765–77. https ://doi.or g/10.1016/s1674 annual conference on carbon capture and sequestration, Citeseer; -5264(09)60278 -7. 2005. Moghanloo RG, Dadmohammadi Y, Bin Y, Salahshoor S. Apply- Lu C, Lichtner PC. High resolution numerical investigation on the ing fractional flow theory to evaluate CO storage capacity effect of convective instability on long term CO storage in of an aquifer. J Pet Sci Eng. 2015;125:154–61. https ://doi. saline aquifers. In: Journal of physics: conference series, IOP org/10.1016/j.petro l.2014.11.013. Publishing; 2007. p. 012042. https ://doi.or g/10.1088/1742- Moritis G. EOR weathers low oil prices. Oil Gas J. 6596/78/1/01204 2. 2000;98(12):39–61. Lu JM, Wilkinson M, Haszeldine RS, Boyce AJ. Carbonate cements Mukherjee M, Misra S. A review of experimental research on enhanced in Miller field of the UK North Sea: a natural analog for coal bed methane (ECBM) recovery via C O sequestration. Earth 1 3 1060 Petroleum Science (2019) 16:1028–1063 Sci Rev. 2018;179:392–410. https ://doi.or g/10.1016/j.earsc Palandri JL, Kharaka YK. Ferric iron-bearing sediments as a mineral irev.2018.02.018. trap for C O sequestration: iron reduction using sulfur-bear- Myshakin E, Siriwardane H, Hulcher C, Lindner E, Sams N, King ing waste gas. Chem Geol. 2005;217(3):351–64. https ://doi. S, et al. Numerical simulations of vertical growth of hydraulic org/10.1016/j.chemg eo.2004.12.018. fractures and brine migration in geological formations above the Palmer I, Mansoori J. How permeability depends on stress and pore Marcellus shale. J Nat Gas Sci Eng. 2015;27:531–44. https://doi. pressure in coalbeds: a new model. In: SPE annual technical con- org/10.1016/j.jngse .2015.08.030. ference and exhibition, Society of Petroleum Engineers; 1996. Nance H, Rauch H, Strazisar B, Bromhal G, Wells A, Diehl R et al. https ://doi.org/10.2523/36737 -ms. Surface environmental monitoring at the Frio C O sequestra- Pan Y, Hui D, Luo PY, Zhang Y, Sun L, Wang K. Experimental inves- tion test site, Texas. In: National energy technology laboratory tigation of the geochemical interactions between supercritical fourth annual conference on carbon capture and sequestration, CO and shale: implications for CO storage in gas-bearing shale 2 2 Alexandria, VA, May; 2005. p. 2–5. formations. Energy Fuels. 2018a;32(2):1963–78. https ://doi. Narinesingh J, Boodlal DV, Alexander D. Feasibility study on the org/10.1021/acs.energ yfuel s.7b030 74. implementation of C O -EOR coupled with sequestration in Pan ZJ, Ye JP, Zhou FB, Tan YL, Connell LD, Fan JJ. CO storage 2 2 Trinidad via reservoir simulation. In: SPE energy resources in coal to enhance coalbed methane recovery: a review of field conference, Society of Petroleum Engineers; 2014. https ://doi. experiments in China. Int Geol Rev. 2018b;60(5–6):754–76. org/10.2118/16993 5-ms.https ://doi.org/10.1080/00206 814.2017.13736 07. Nghiem L, Sammon P, Grabenstetter J, Ohkuma H. Modeling CO Pang ZH, Li YM, Yang FT, Duan ZF. Geochemistry of a conti- storage in aquifers with a fully-coupled geochemical EOS com- nental saline aquifer for C O sequestration: The Guantao positional simulator. In: SPE/DOE symposium on improved oil Formation in the Bohai Bay Basin, North China. Appl Geo- recovery, Society of Petroleum Engineers; 2004. https ://doi. chem. 2012;27(9):1821–8. h tt ps : // do i.o r g /1 0. 101 6/ j. ap g e o org/10.2118/89474 -ms. chem.2012.02.017. Nghiem L, Shrivastava V, Tran D, Kohse B, Hassam M, Yang C. Simu- Park SS, Park J, Kim TH, Lee KS. Influence of heterogeneous capillary lation of C O storage in saline aquifers. In: SPE/EAGE reservoir pressure on CO sequestration in different wetting conditions. 2 2 characterization and simulation conference; 2009. https ://doi. In: The twenty-fifth international offshore and polar engineering org/10.2118/0806-0068-jpt. conference, International Society of Offshore and Polar Engi- Nguyen BN, Hou Z, Stewart ML, Murray CJ, Bonneville A. Ther- neers; 2015. mal impact of CO injection on geomechanical response at the Parkhurst DL, Appelo C. Description of input and examples for FutureGen 2.0 Site: a three-dimensional thermo-geomechanical PHREEQC version 3: a computer program for speciation, batch- approach. Int J Greenh Gas Control. 2016;54:29–49. https://doi. reaction, one-dimensional transport, and inverse geochemical org/10.1016/j.ijggc .2016.08.026. calculations 2328-7055, US Geological Survey; 2013. https :// Nicol A, Carne R, Gerstenberger M, Christophersen A. Induced seis-doi.org/10.3133/tm6a4 3. micity and its implications for CO storage risk. Energy Procedia. Parkhurst DL, Kipp KL, Engesgaard P, Charlton SR. PHAST—a pro- 2011;4:3699–706. https://doi.or g/10.1016/j.egypro.2011.02.302 . gram for simulating ground-water flow, solute transport, and Nicot J-P, Hovorka SD, Choi J-W. Investigation of water displacement multicomponent geochemical reactions. US Department of the following large C O sequestration operations. Energy Procedia. Interior, US Geological Survey; 2004. https: //doi.org/10.3133/ 2009;1(1):4411–8. https://doi.or g/10.1016/j.egypro.2009.02.256 . tm6a8 . Nordbotten JM, Celia MA, Bachu S. Injection and storage of C O in Parry W, Forster CB, Evans JP, Bowen BB, Chan MA. Geochemistry deep saline aquifers: analytical solution for CO plume evolution of CO sequestration in the Jurassic Navajo Sandstone, Colo- 2 2 during injection. Transp Porous Med. 2005a;58(3):339–60. https rado Plateau, Utah. Environ Geosci. 2007;14(2):91–109. https ://doi.org/10.1007/s1124 2-004-0670-9.://doi.org/10.1306/eg.07120 60600 4. Nordbotten JM, Celia MA, Bachu S, Dahle HK. Semianalytical solu- Parson EA, Keith DW. Fossil fuels without CO emissions. Sci- tion for C O leakage through an abandoned well. Environ Sci ence. 1998;282(5391):1053–4. https ://doi.or g/10.1126/scien Technol. 2005b;39(2):602–11. https ://doi.or g/10.1021/es035 ce.282.5391.1053. 338i. Pashin JC, Dodge RL. Carbon dioxide sequestration in geological Nordbotten J, Kavetski D, Celia M, Bachu S. A semi-analytical model media—state of the science. AAPG Stud Geol. 2010;59:59. estimating leakage associated with CO storage in large-scale https ://doi.org/10.1016/s0921 -3198(06)80027 -x. multi-layered geological systems with multiple leaky wells. Envi- Pawar RJ, Zyvoloski GA, Tenma N, Sakamoto Y, Komai T. Numeri- ron Sci Technol. 2009;43(3):743–9. cal simulation of laboratory experiments on methane hydrate Obi EOI, Blunt MJ. Streamline-based simulation of carbon diox- dissociation. In: The fifteenth international offshore and polar ide storage in a North Sea aquifer. Water Resour Res. engineering conference, International Society of Offshore and 2006;42(3):W03414. https ://doi.org/10.1029/2004w r0033 47. Polar Engineers; 2005. Oelkers EH, Gislason SR, Matter J. Mineral carbonation of C O . Ele- Pawar R, Carey J, Chipera S, Fessenden J, Kaszuba J, Keating G ments. 2008;4(5):333–7. https ://doi.or g/10.2113/gsele ments et al. Development of a framework for long-term performance .4.5.333. assessment of geologic C O sequestration sites. In: Eighth Oldenburg CM. The risk of induced seismicity: Is cap-rock integrity on international conference on greenhouse gas control technolo- shaky ground? Greenh Gases Sci Technol. 2014;2:217–8. https gies (GHGT-8); 2006. p. 19–22. ://doi.org/10.1002/ghg.1299. Perkins, E, Czernichowski-Lauriol, I, Azaroual, M, Durst, P. Long Olivella S, Gens A, Carrera J, Alonso EE. Numerical formula- term predictions of C O storage by mineral and solubility trap- tion for a simulator (CODE_BRIGHT) for the coupled analy- ping in the Weyburn Midale Reservoir. In: Proceedings of the sis of saline media. Eng Comput. 1996;13(7):87. https ://doi. 7th international conference on greenhouse gas control tech- org/10.1108/02644 40961 01515 75. nologies (GHGT-7); 2004, p. 2093–96. https://doi.or g/10.1016/ Onuma T, Ohkawa S. Detection of surface deformation related with b978-00804 4704-9/50279 -2. CO injection by DInSAR at In Salah, Algeria. Energy Procedia. Pickup GE, Jin M, Olden P, Mackay EJ, Sohrabi M. A sensitivity 2009;1(1):2177–84. https ://doi.org/10.2523/12294 -ms. study on CO storage in saline aquifers. In: SPE EUROPEC/ 1 3 Petroleum Science (2019) 16:1028–1063 1061 EAGE annual conference and exhibition, Society of Petroleum based upon simultaneous solution of strongly coupled subsets Engineers; 2011. https ://doi.org/10.2118/14305 4-ms. of chemical components. Adv Water Resour. 2000;23(4):307–24. Pickup G, Jin M, Mackay E. Simulation of near-well pressure build-https ://doi.org/10.1016/S0309 -1708(99)00034 -2. up in models of CO injection. In: ECMOR XIII-13th Euro- Rutqvist J. The geomechanics of CO storage in Deep Sedimentary 2 2 pean conference on the mathematics of oil recovery; 2012. Formations. Geotech Geol Eng. 2012;30(3):525–51. https://doi. https ://doi.org/10.3997/2214-4609.20143 238.org/10.1007/s1070 6-011-9491-0. Plasynski SI, Litynski JT, McIlvried HG, Vikara DM, Srivastava RD. Rutqvist J, Tsang C-F. TOUGH-FLAC: a numerical simulator for The critical role of monitoring, verification, and accounting analysis of coupled thermal-hydrologic-mechanical processes in for geologic carbon dioxide storage projects. Environ Geosci. fractured and porous geological media under multi-phase flow 2011;18(1):19–34. https ://doi.org/10.1306/eg.06231 01000 8. conditions. In: The TOUGH symposium; 2003. p. 12–4. Preston C, Monea M, Jazrawi W, Brown K, Whittaker S, White D, Rutqvist J, Vasco DW, Myer L. Coupled reservoir-geomechanical et al. IEA GHG Weyburn CO monitoring and storage project. analysis of CO injection and ground deformations at In Salah, 2 2 Fuel Process Technol. 2005;86(14):1547–68. Algeria. Int J Greenh Gas Control. 2010;4(2):225–30. https://doi. Pruess K. Numerical modeling studies of the dissolution-diffusion-org/10.1016/j.ijggc .2009.10.017. convection process during C O storage in saline aquifers. Saadatpoor E, Bryant SL, Sepehrnoori K. New trapping mechanism Lawrence Berkeley National Laboratory; 2008. https ://doi. in carbon sequestration. Transp Porous Med. 2010;82(1):3–17. org/10.2172/94412 4.https ://doi.org/10.1007/s1124 2-009-9446-6. Pruess K, Spycher N. EC O N—a fluid property module for the Saadawi OH, Pickup GE, Jin M, Mackay EJ. Streamline simulation TOUGH2 code for studies of CO storage in saline aquifers. of CO Storage in saline aquifers. In: SPE middle east oil and 2 2 Energy Convers Manag. 2007;48(6):1761–7. https ://doi. gas show and conference, Society of Petroleum Engineers; 2011. org/10.1016/j.encon man.2007.01.016.https ://doi.org/10.2118/14253 5-ms. Pruess K, Garcia J, Kovscek T, Oldenburg C, Rutqvist J, Steefel Saghafi A, Faiz M, Roberts D. CO storage and gas diffusivity prop- C et al. Intercomparison of numerical simulation codes for erties of coals from Sydney Basin, Australia. Int J Coal Geol. geologic disposal of CO . Lawrence Berkeley National Labora- 2007;70(1):240–54. https ://doi.org/10.1016/j.coal.2006.03.006. tory; 2002. https ://doi.org/10.2172/81356 6. Saini D. Simultaneous C O -EOR and storage projects, engineering Qi R, LaForce TC, Blunt MJ. Design of carbon dioxide storage in aspects of geologic C O storage. Springer; 2017. p. 11–19. https aquifers. Int J Greenh Gas Control. 2009;3(2):195–205. https ://doi.org/10.1007/978-3-319-56074 -8_2. ://doi.org/10.1016/j.ijggc .2008.08.004. Schepers KC, Nuttall BC, Oudinot AY, Gonzalez RJ. Reservoir Qiao XJ, Li GM, Li M, Wang ZM. C O storage capacity assessment modeling and simulation of the Devonian gas shale of east- of deep saline aquifers in the Subei Basin, East China. Int J ern Kentucky for enhanced gas recovery and C O storage. In: Greenh Gas Control. 2012;11:52–63. https://doi.or g/10.1016/j. SPE international conference on CO capture, storage, and ijggc .2012.07.020. utilization, Society of Petroleum Engineers; 2009. https ://doi. Rasmusson M, Fagerlund F, Tsang Y, Rasmusson K, Niemi A. Pre-org/10.2118/12662 0-ms. requisites for density-driven instabilities and convective mix- Scherer GW, Celia MA, Prevost J-H, Bachu S, Bruant R, Duguid A, ing under broad geological CO storage conditions. Adv Water et al. Leakage of C O through abandoned wells: role of corrosion 2 2 Resour. 2015;84:136–51. https ://doi.or g/10.1016/j.advw a of cement. Carbon Dioxide Capture Storage Deep Geol Form. tres.2015.08.009. 2015;1:827–48. Reed M, Spycher N. User’s guide for CHILLER: a program for com- Schilling F, Borm G, Würdemann H, Möller F, Kühn M, Group CS. puting water–rock reactions, boiling, mixing and other reaction Status report on the first European on-shore CO storage site at processes in aqueous–mineral–gas systems and minplot guide. Ketzin (Germany). Energy Procedia. 2009;1(1):2029–35. https Oregon: University of Oregon; 2006.://doi.org/10.1016/j.egypr o.2009.01.264. Reeves SR. Geological sequestration of CO in deep, unmineable Schwartz BC. Fracture pattern characterization of the Tensleep For- coalbeds: an integrated research and commercial-scale field mation, Teapot Dome. Wyoming: West Virginia University; demonstration project. In: SPE annual technical conference and 2006. exhibition, Society of Petroleum Engineers; 2001. https ://doi. Seibel BA, Walsh PJ. Carbon cycle potential impacts of C O injection org/10.2118/71749 -ms. on deep-sea biota. Science. 2001;294(5541):319–20. https: //doi. Riaz A, Cinar Y. Carbon dioxide sequestration in saline forma-org/10.1126/scien ce.10653 01. tions: part I—review of the modeling of solubility trapping. J Seibel BA, Walsh PJ. Biological impacts of deep-sea carbon dioxide Pet Sci Eng. 2014;124:367–80. https ://doi.org/10.1016/j.petro injection inferred from indices of physiological performance. l.2014.07.024. J Exp Biol. 2003;206(Pt4):641–50. https ://doi.or g/10.1242/ Riaz A, Hesse M, Tchelepi HA, Orr FM. Onset of convection in a grav- jeb.00141 . itationally unstable diffusive boundary layer in porous media. Senel O, Will R, Butsch RJ. Integrated reservoir modeling at the J Fluid Mech. 2006;548:87–111. https ://doi.org/10.1017/S0022 Illinois Basin–Decatur Project. Greenh Gases Sci Technol. 11200 50074 94. 2014;4(5):662–84. https ://doi.org/10.1002/ghg.1451. Riding J, Rochelle C. The IEA Weyburn CO monitoring and storage Shi JQ, Durucan S. CO storage in deep unminable coal seams. Oil Gas 2 2 project: final report of the European research team. In: British Sci Technol Revue D IFP Energ N. 2005;60(3):547–58. https :// Geological Survey; 2005.doi.org/10.2516/ogst:20050 37. Ringrose P, Atbi M, Mason D, Espinassous M, Myhrer Ø, Iding M, Shi JQ, Durucan S, Fujioka M. A reservoir simulation study of C O et al. Plume development around well KB-502 at the In Salah injection and N flooding at the Ishikari coalfield CO storage 2 2 CO storage site. First Break. 2009;27(1):81–5. pilot project, Japan. Int J Greenh Gas Control. 2008;2(1):47–57. Ringrose P, Mathieson A, Wright I, Selama F, Hansen O, Bissell R, https ://doi.org/10.1016/s1750 -5836(07)00112 -0. et  al. The In Salah C O storage project: lessons learned and Shipton ZK, Evans JP, Kirschner D, Kolesar PT, Williams AP, Heath knowledge transfer. Energy Procedia. 2013;37:6226–36. https:// J. Analysis of C O leakage through ‘low-permeability’ faults doi.org/10.1016/j.egypr o.2013.06.551. from natural reservoirs in the Colorado Plateau, east-central Robinson BA, Viswanathan HS, Valocchi AJ. Efficient numerical tech - Utah. Geol Soc Lond Spec Publ. 2004;233(1):43–58. https :// niques for modeling multicomponent ground-water transport doi.org/10.1144/gsl.sp.2004.233.01.05. 1 3 1062 Petroleum Science (2019) 16:1028–1063 Shipton ZK, Evans JP, Dockrill B, Heath J, Williams A, Kirchner D, Torp TA, Gale J. Demonstrating storage of C O in geologi- et al. Natural leaking CO -charged systems as analogs for failed cal reservoirs: the Sleipner and SACS projects. Energy. geologic storage reservoirs. Carbon Dioxide Capture Storage 2004;29(9):1361–9. https ://doi.or g/10.1016/s0360 Deep Geol Form Results C O Capture Proj. 2006;2:699–712. -5442(04)00153 -7. https ://doi.org/10.1016/b978-00804 4570-0/50130 -6. Underschultz J, Boreham C, Dance T, Stalker L, Freifeld B, Kirste Siemons N, Bruining H, Castelijns H, Wolf K-H. Pressure depend- D, et al. C O storage in a depleted gas field: an overview of ence of the contact angle in a CO –H O–coal system. J Colloid the CO  CRC Otway Project and initial results. Int J Greenh 2 2 2 Interface Sci. 2006;297(2):755–61. https ://doi.org/10.1016/j. Gas Control. 2011;5(4):922–32. https://doi.or g/10.1016/j.ijggc jcis.2005.11.047. .2011.02.009. Sifuentes WF, Giddins MA, Blunt MJ. Modeling CO storage in aqui- Urosevic M, Pevzner R, Shulakova V, Kepic A, Caspari E, Sharma fers: assessing the key contributors to uncertainty. In: Offshore S. Seismic monitoring of C O injection into a depleted gas res- Europe, Society of Petroleum Engineers; 2009. https ://doi. ervoir–Otway Basin Pilot Project, Australia. Energy Procedia. org/10.2118/12358 2-ms. 2011;4:3550–7. https://doi.or g/10.1016/j.egypro.2011.02.283 . Siirila ER, Navarre-Sitchler AK, Maxwell RM, McCray JE. A quantita- US Environmental Protection Agency. Understanding Global Warm- tive methodology to assess the risks to human health from C O ing Potentials; 2014. leakage into groundwater. Adv Water Resour. 2012;36:146–64. van Bergen F, Pagnier H, van der Meer L, van den Belt F, Win- https ://doi.org/10.1016/j.advwa tres.2010.11.005. thaegen P, Westerhoff R. Development of a field experiment Smith SA, Sorensen JA, Steadman EN, Harju JA. Acid gas injection of CO storage in coal seams in the Upper Silesian Basin of and monitoring at the Zama oil field in Alberta, Canada: a case Poland (RECOPOL). In: Gale J, Kaya Y, editors. The 6th inter- study in demonstration-scale carbon dioxide sequestration. national conference on greenhouse gas control technologies Energy Procedia. 2009;1(1):1981–8. https ://doi.org/10.1016/j. (GHGT-6); 2002, p. 1–4. https ://doi.org/10.1016/b978-00804 egypr o.2009.01.258.4276-1/50091 -x. Spiteri E, Juanes R, Blunt MJ, Orr FM. Relative-permeability hystere- van Bergen F, Pagnier H, van der Meer L, van den Belt F, Winthae- sis: trapping models and application to geological C O sequestra- gen P, Westerhoff R. Development of a field experiment of CO 2 2 tion. In: SPE annual technical conference and exhibition, Society storage in coal seams in the Upper Silesian Basin of Poland of Petroleum Engineers; 2005. https: //doi.org/10.2523/96448- ms. (Recopol). In: 6th International conference on greenhouse gas Stauffer PH, Viswanathan H, Pawar RJ, Klasky ML, Guthrie GD. control technologies, Elsevier; 2003, p. 569–74. https ://doi. CO -PENS: a CO sequestration systems model supporting org/10.1016/b978-00804 4276-1/50091 -x. 2 2 risk-based decisions. In: The 16th international conference on van der Meer L, Kreft E, Geel C, D’Hoore D, Hartman J. C O storage computational methods in water resources; 2006. p. 19–22. and testing enhanced gas recovery in the K12-B reservoir 2. In: Steefel CI, Lasaga AC. A coupled model for transport of multiple 23rd world gas conference, Amsterdam; 2006. chemical-species and kinetic precipitation dissolution reactions van Pham TH, Aagaard P, Hellevang H. On the potential for C O with application to reactive flow in single-phase hydrothermal mineral storage in continental flood basalts-PHREEQC batch- systems. Am J Sci. 1994;294(5):529–92. https:/ /doi.org/10.2475/ and 1D diffusion–reaction simulations. Geochem Trans. ajs.294.5.529. 2012;13(5):2–12. https ://doi.org/10.1186/1467-4866-13-5. Stéphenne K. Start-up of world’s first commercial post-combustion coal Verdon JP. Significance for secure C O storage of earthquakes fired CCS project: contribution of Shell Cansolv to SaskPower induced by fluid injection. Environ Res Lett. 2014;9(6):064022. Boundary Dam ICCS Project. Energy Procedia. 2014;63:6106–https ://doi.org/10.1088/1748-9326/9/6/06402 2. 10. https ://doi.org/10.1016/j.egypr o.2014.11.642. Vilarrasa V, Bolster D, Olivella S, Carrera J. Coupled hydrome- Stevenson M, Pinczewski V. SIMED II—multi-component coalbed chanical modeling of CO sequestration in deep saline aqui- gas simulator. User’s manual version 1.21. Centre for Petro- fers. Int J Greenh Gas Control. 2010;4(6):910–9. https ://doi. leum Engineering, University of New South Wales; 1995.org/10.1016/j.ijggc .2010.06.006. Suekane T, Nobuso T, Hirai S, Kiyota M. Geological storage of Wei LL, Saaf F. Estimate C O storage capacity of the Johansen carbon dioxide by residual gas and solubility trapping. Int J Formation: numerical investigations beyond the benchmark- Greenh Gas Control. 2008;2(1):58–64. https:/ /doi.org/10.1016/ ing exercise. Comput Geosci. 2009;13(4):451–67. https ://doi. s1750 -5836(07)00096 -5.org/10.1007/s1059 6-008-9122-x. Temitope A, Gupta I. A review of reactive transport modeling in Weir GJ, White SP, Kissling WM. Reservoir storage and containment wellbore integrity problems. J Pet Sci Eng. 2019;1:1. https :// of greenhouse gases. Transp Porous Med. 1996;23(1):37–60. doi.org/10.1016/j.petro l.2018.12.079.https ://doi.org/10.1007/bf001 45265 . Temitope A, Gomes JS, Al Kobaisi M, Hu J. Characterization and Wheeler M, Delshad M, Thomas S. Modeling CO sequestration quantification of the CO sequestration potential of a carbonate using a sequentially coupled ‘Iterative-IMPEC-time-split- aquifer in Falaha Syncline, Onshore Abu Dhabi. In: Abu Dhabi thermal’ compositional simulator. In: 11th European confer- international petroleum exhibition and conference, Society of ence on the mathematics of oil recovery; 2008. https ://doi. Petroleum Engineers; 2016. https ://doi.or g/10.2118/18344 org/10.3997/2214-4609.20146 390. 2-ms. White D. Monitoring CO storage during EOR at the Weyburn- Thakur IS, Kumar M, Varjani SJ, Wu Y, Gnansounou E, Ravin- Midale Field. Lead Edge. 2009;28(7):838–42. htt ps :// doi. dran S. Sequestration and utilization of carbon dioxide by org/10.1190/1.31677 86. chemical and biological methods for biofuels and biomateri- White D. Geophysical monitoring of the Weyburn C O flood: results als by chemoautotrophs: opportunities and challenges. Biore- during 10 years of injection. Energy Procedia. 2011;4:3628– sour Tech. 2018;256:478–90. https ://doi.or g/10.1016/j.bior t 35. https ://doi.org/10.1016/j.egypr o.2011.02.293. ech.2018.02.039. White MD, Bacon DH, McGrail BP, Watson DJ, White SK, Zhang Z. Thomson AM, Izaurralde RC, Smith SJ, Clarke LE. Integrated esti- STOMP subsurface transport over multiple phases: STOMP- mates of global terrestrial carbon sequestration. Glob Environ CO and STOMP-CO e guide: version 1.0, Pacific Northwest 2 2 Change Hum Policy Dimens. 2008;18(1):192–203. https://doi. National Laboratory (PNNL), Richland, WA (US); 2012. https org/10.1016/j.gloen vcha.2007.10.002.://doi.org/10.2172/10590 44. 1 3 Petroleum Science (2019) 16:1028–1063 1063 Wong S, Law D, Deng X, Robinson J, Kadatz B, Gunter WD, et al. Yang DY, Tontiwachwuthikul P, Gu YG. Interfacial interactions Enhanced coalbed methane and CO storage in anthracitic between reservoir brine and C O at high pressures and elevated 2 2 coals—micro-pilot test at South Qinshui, Shanxi, China. temperatures. Energy Fuels. 2005;19(1):216–23. https ://doi. Int J Greenh Gas Control. 2007;1(2):215–22. https ://doi.org/10.1021/ef049 792z. org/10.1016/s1750 -5836(06)00005 -3. Zakrisson J, Edman I, Cinar Y. Multiwell injectivity for CO stor- Xiuzhang W. Shenhua Group’s carbon capture and storage (CCS) age. In: SPE Asia Pacific oil and gas conference and exhi- demonstration. Min Rep. 2014;150(1–2):81–4. https ://doi. bition, Society of Petroleum Engineers; 2008. https ://doi. org/10.1002/mire.20140 0006.org/10.2118/11635 5-MS. Xu T, Apps JA, Pruess K. Analysis of mineral trapping for CO Zhou Q, Birkholzer JT, Tsang C-F, Rutqvist J. A method for quick disposal in deep aquifers. Lawrence Berkeley National Labora- assessment of C O storage capacity in closed and semi-closed tory; 2001. https ://doi.org/10.2172/78913 3. saline formations. Int J Greenh Gas Control. 2008;2(4):626– Xu TF, Apps JA, Pruess K. Reactive geochemical transport simula- 39. https ://doi.org/10.1016/j.ijggc .2008.02.004. tion to study mineral trapping for C O disposal in deep arena- Zhou Q, Birkholzer JT, Mehnert E, Lin YF, Zhang K. Modeling ceous formations. J Geophys Res Solid Earth. 2003. https :// basin-and plume-scale processes of CO storage for full-scale doi.org/10.1029/2002j b0019 79. deployment. Groundwater. 2010;48(4):494–514. https://doi.or g Xu T, Apps JA, Pruess K. Numerical simulation of CO dis-/10.1111/j.1745-6584.2009.00657 .x. posal by mineral trapping in deep aquifers. Appl Geo- Zhu LT, Liao XW, Chen ZM, Mu LY, Chen XY. Analytical model chem. 2004;19(6):917–36. https ://doi.or g/10.1016/j.apg eo for quick assessment of capillary storage capacity in saline chem.2003.11.003. aquifers. Int J Greenh Gas Control. 2017;65:160–9. https: //doi. Xu T, Sonnenthal E, Spycher N, Pruess K. TOUGHREACT—a org/10.1016/j.ijggc .2017.09.004. simulation program for non-isothermal multiphase reactive Zwingmann N, Mito S, Sorai M, Ohsumi T. Preinjection characterisa- geochemical transport in variably saturated geologic media: tion and evaluation of C O sequestration potential in the Hai- applications to geothermal injectivity and C O geological zume Formation, Niigata Basin, Japan—Geochemical modelling sequestration. Comput Geosci. 2006a;32(2):145–65. https :// of water–minerals–CO interaction. Oil Gas Sci Technol Revue doi.org/10.1016/j.cageo .2005.06.014. D IFP Energies N. 2005;60(2):249–58. https ://doi.org/10.2516/ Xu XF, Chen SY, Zhang DX. Convective stability analysis of the ogst:20050 15. long-term storage of carbon dioxide in deep saline aqui- fers. Adv Water Resour. 2006b;29(3):397–407. https ://doi. org/10.1016/j.advwa tres.2005.05.008. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Petroleum Science Springer Journals

A review of CO2 storage in geological formations emphasizing modeling, monitoring and capacity estimation approaches

Petroleum Science , Volume 16 (5) – Jul 8, 2019

Loading next page...
 
/lp/springer-journals/a-review-of-co2-storage-in-geological-formations-emphasizing-modeling-R0rmwi3rGV

References (300)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s)
Subject
Earth Sciences; Mineral Resources; Industrial Chemistry/Chemical Engineering; Industrial and Production Engineering; Energy Policy, Economics and Management
ISSN
1672-5107
eISSN
1995-8226
DOI
10.1007/s12182-019-0340-8
Publisher site
See Article on Publisher Site

Abstract

The merits of C O capture and storage to the environmental stability of our world should not be underestimated as emissions of greenhouse gases cause serious problems. It represents the only technology that might rid our atmosphere of the main anthropogenic gas while allowing for the continuous use of the fossil fuels which still power today’s world. Underground storage of C O involves the injection of C O into suitable geological formations and the monitoring of the injected plume 2 2 over time, to ensure containment. Over the last two or three decades, attention has been paid to technology developments of carbon capture and sequestration. Therefore, it is high time to look at the research done so far. In this regard, a high-level review article is required to provide an overview of the status of carbon capture and sequestration research. This article presents a review of C O storage technologies which includes a background of essential concepts in storage, the physical processes involved, modeling procedures and simulators used, capacity estimation, measuring monitoring and verification techniques, risks and challenges involved and field-/pilot-scale projects. It is expected that the present review paper will help the researchers to gain a quick knowledge of CO sequestration for future research in this field. Keywords CO storage · Geological formation · Modeling for CO storage · Mechanism of CO storage · CO storage 2 2 2 2 projects 1 Introduction are greener technologies such as nuclear energy and wind energy which reduce the combustion of fossil fuels associ- The global warming scourge is threatening to ravage human- ated with emission sources and energy efficiency. The con- ity. Rising sea levels, increases in average global air and sea tinued need for fossil fuels across the world and the rela- surface temperatures, widespread snow and ice melting are tively slow pace of renewable energy development suggests notable effects of global warming (IPCC 2007). The implica- that the amount of undesired different gases being emitted tion of these indicators in the long run on health, nutrition into the atmosphere will remain on the increase. It is impera- and the economy can be ill-afforded and therefore has been tive, therefore, the ways should be developed in which these the subject of a great deal of research to date. Numerous harmful gases can be expunged from the atmosphere. strategies have been employed or are under intense scru- Greenhouse gases, a term for the climate-unfriendly gases tiny as a means of tackling climate change, some of which emitted into the atmosphere, provide a threat to our ecosys- tem with CO accounting for 82% of greenhouse gases in the atmosphere. Though the global warming potential (GWP) of Edited by Yan-Hua Sun CO is less than other greenhouse gases (US Environmental * Achinta Bera Protection Agency 2014), the sheer amount of CO being achintachm@gmail.com emitted into the atmosphere makes it the most significant of all greenhouse gases for efficient climate control. Department of Petroleum Engineering, Khalifa University The advent, development and implementation of carbon of Science and Technology, Sas Al Nakhl Campus, P.O. Box 2533, Abu Dhabi, UAE dioxide capture, utilization and storage (CCUS) technology promises to reduce the amount of greenhouse gases enter- Drilling, Cementing, and Stimulation Research Center, School of Petroleum Technology, Pandit Deendayal ing the atmosphere. CCUS encompasses the capture of car- Petroleum University, Raisan, Gandhinagar, Gujarat 382007, bon dioxide and its associated compounds from producing India Vol:.(1234567890) 1 3 Petroleum Science (2019) 16:1028–1063 1029 sources, compression, transportation and the utilization of The technology is currently at the research stage without the captured C O for processes such as injection into deep any existing pilot tests. underground geological formations for permanent storage C. Geological sequestration is the most widely used seques- and injection into existing oil fields for additional recovery tration technology. In this process, CO is stored in geo- of hydrocarbons. logical underground structures such as saline aquifers, Some previous review articles summarized the different depleted oil and gas reservoirs and unmineable coal beds physicochemical methods responsible for suitable CO stor- (IPCC 2007; Kaldi et al. 2009; Metz 2005; Pashin and age and the difficulties in different aspects (Riaz and Cinar Dodge 2010). A short description of all storage sites is 2014; Belhaj and Bera 2017; Aminu et al. 2017; Thakur given below: et al. 2018). The main motivation of this review paper is to 1. Saline aquifer formations: Saline aquifer forma- present all aspects of CCUS projects worldwide along with tions represent the best salted sink for storage of the technologies, modeling issues and physicochemical pro- CO among all geological options due to their enor- cesses occurred during the C O sequestration within geolog- mous storage capacity (Grobe et al. 2009). Recently, ical formation. This review will serve as a single handbook estimates of the order of 103 Gt C O have been for understanding CCUS and to provide researchers the facts made for the Alberta deep saline basin by account- about CCUS in the oil industry. C O flooding for enhanced ing for the solubility trapping mechanism (Bachu oil recovery is one of the effective methods in additional oil and Adams 2003). Another example is the injection recovery. The injected carbon dioxide can be stored in the of the produced CO into the Utsira aquifer in the formation of the reservoir. Therefore, it is important to know North Sea (Korbøl and Kaddour 1995; Torp and the rock capacity and power to store the carbon dioxide for Gale 2004). It is required that the aquifer be saline a long time. because this already makes it unsuitable for indus- Storage of CO has been employed in different parts of trial, agricultural and human purposes (Aydin et al. the world. The modes of storage can be broadly classified 2010; Metz et al. 2005). into natural and man-made modes of storage. Natural modes Other storage modes which have been employed include terrestrial sequestration, while man-made storage for the storage of CO include basalts (Gislason includes storage in geologic formations. Several modes for and Oelkers 2014) and mineral carbonation (Oelk- utilizing and storing CO have been explored as follows: ers et al. 2008). Among all geologic sequestration mechanisms, deep saline aquifers represent the A. Terrestrial sequestration is the capture of C O from ones exhibiting highest sequestering capability, as the atmosphere and storing it into soils and vegetation. against those provided by depleted oil and gas res- Removal of CO from the atmosphere through photo- ervoirs and unmineable coal beds (IPCC 2007; Torp synthesis and prevention of the emission of C O from and Gale 2004; Kaldi et al. 2009; Parson and Keith terrestrial sources are the mechanisms for terrestrial 1998). storage. It has been postulated to provide an important 2. Depleted oil and gas reservoirs: Previously produc- mechanism for the storage of carbon dioxide (Litynski ing oil and gas fields which have been considered et al. 2006; Thomson et al. 2008). uneconomical for further production of hydrocar- B. Ocean sequestration qualifies as the largest possible sink bons are suitable candidates for geological seques- for carbon dioxide storage with an estimated potential tration. Characteristics required for a storage site are storage of 40,000 gigatonnes (Gt) of CO (Herzog et al. present in such formations and have been employed 1997, 2000; Lal 2008) and the possibility of storing over for geologic sequestration. An important advan- 90% of current CO emissions. It involves the injection tage is that they have been adequately character- and deposition of CO into the water body at depths ized previously. Additionally, the safe and secure below 1 km either from moving ships, fixed pipelines nature of these formations which have been able to or offshore platforms. At this depth, water has a lower store oil and gas over a long period of time makes density than the injected CO and the latter is expected them prime candidates. Existing numerical com- to dissolve and disperse into the water body (Metz et al. puter models of such formations which have been 2005). However, there are huge concerns over the envi- history-matched provide improved confidence in ronmental impact of CO on marine life from the acidity the formations. Infrastructures and wells used in the of seawater near the injection point (Seibel and Walsh development of these fields are also available for 2001). The scalability of experiments involved in ocean CO injection. Storage capacity available in depleted sequestration is also very difficult, thus requiring expen- reservoirs is significantly lower due to the need to sive field experiments (Adams et al. 1998a, b; Auerbach avoid exceeding pressures that can damage the cap et al. 1997; Herzog et al. 1997; Seibel and Walsh 2003). 1 3 1030 Petroleum Science (2019) 16:1028–1063 rock and the significant leakage threat posed by the China (Liang et al. 2009) and different parts of the abandoned wells. (A potential for leaks exists behind world for simultaneous EOR and storage processes well casings.) (Ghomian et al. 2008; Gozalpour et al. 2005; Liu 3. Deep unmineable coal beds: CO has been employed et al. 2013; Moritis 2000; Narinesingh et al. 2014). for the recovery of methane from coal seams during the enhanced coal bed methane (ECBM) recovery This integrated review will discuss storage of CO in various process (Busch and Gensterblum 2011; Mukherjee geological formations with a focus on saline aquifers. Sec- and Misra 2018; Pan et al. 2018b). Produced meth- tion 1 contains the introductory part of the review. Section 2 ane from this source can be utilized as an energy discusses the properties of the gas which favors storage as source. Coal beds have very large fracture networks well as trapping mechanisms and the physical processes through which gas molecules can diffuse into the involved in the storage process. Section 3 gives a summary matrix and desorb tightly adsorbed methane. C O of the pilot- and commercial-scale projects which are in the has been proven to raise methane recovery to about planning phase, in operation or have been abandoned. In 90% from 50% when conventional methods are Sect. 4, we discuss the modeling strategies for CO which applied. Injected C O is stored in the formations have been applied in the literature. Section 5 covers the esti- after methane has been recovered. Storage in coal mation methods for storage capacities. In Sect. 6, an over- beds can take place at shallower depths than other view of the measuring, monitoring and verification tools and formation types and as such relies on CO adsorp- challenges is provided. Section 7 reports the risks and chal- tion on the coal surface. However, the technical fea- lenges that may be present before commercial application of sibility of this storage process strongly depends on field-scale projects. Finally, conclusions and recommenda- the coal’s permeability as a result of its depth vari- tions are provided in Sect. 8. It is expected that the entire ation with the influence of effective stress on coal manuscript will provide an overview of CCUS issues of past, fractures (Metz et al. 2005). present and future challenges for newcomers in this field. The laboratory and field testing feasibility of commercial CO injection into coal beds and seams has been reported in the San Juan Basin, which 2 CO storage in saline aquifers is the world’s first ECBM project (Reeves 2001). Other enhanced coal bed methane recovery projects 2.1 Conditions required for storage sites reported in the world for laboratory and field testing include the Sydney Basin in Australia (Saghafi et al. The selection of a geological site for storage must be done 2007) and deep coalbed methane in Alberta Canada to meet three main conditions: capacity, injectivity and (Gunter et al. 1997). containment. The requirement of the capacity of a storage 4. CO storage during enhanced oil recovery: CO is site ensures that the selected site possesses adequate pore 2 2 used for enhanced oil recovery (EOR) from mature volumes to store large amounts of C O . Typical conditions fields. CO for EOR operations has been employed would mean that the site should contain significant porosity in the miscible and immiscible states. When injected and/or occupy a very large area. Injectivity of C O is assured into oil, CO has the capability to swell the oil, if the candidate formation possesses high permeability reduce its viscosity and reduce interfacial tension ensuring that lower wellhead pressures can be used to main- and in some cases become miscible with the oil tain desired injection rates. Competent cap rocks and sealing allowing for single-phase flow. Of the two miscible faults (if present) are necessary to ensure that the injected states for EOR via CO injection, miscibility of CO CO does not escape to the surface or leak into groundwa- 2 2 2 in oil usually provides higher recoveries. The abil- ter due to the lower density of the C O gas compared with ity of CO to become miscible in oil is determined resident brine. For successful storage of carbon dioxide, it by the minimum miscibility pressure (MMP). At is required that C O be stored in a supercritical phase, the and above this pressure, CO is miscible in oil and state in which C O exists when it is compressed to higher 2 2 below, it is immiscible. Though C O injection in pressures and temperatures (about 89 °F and 7.4 MPa). In this process is done primarily for EOR, it comes this phase, C O possesses properties of a liquid but flows as with the added benefit of storage of CO contribut- a gas. Essentially, CO is required to be stored at this state 2 2 ing to minimizing the global warming scourge. Over due to its higher density, reducing the buoyancy differential the last decade, CO has been used in over 70 EOR between CO and in situ fluids (Grobe et al. 2009; Kane 2 2 operations around the world with over 40 reported and Klein 2002; Koide et al. 1992). Though the density of in West Texas (Moritis 2000), Weyburn Field in CO is higher when injected underground, it remains signifi- Canada (Malik and Islam 2000), Shengli Oilfield in cantly lower than the density of in situ brine which lies in 1 3 Petroleum Science (2019) 16:1028–1063 1031 the region of 1200–2000 kg/m depending on the salinity of underground in the long term is essential. During the injec- the brine. The implication of this density differential is the tion process in the targeted formation, viscous forces are the buoyant movement of CO when injected underground and dominant forces for the migration of C O. CO is then stored 2 2 2 thus demanding the presence of low-permeability cap rocks in either the supercritical or the gas phase as a function of which overlay the aquifer. depth at the associated pressure and temperature. Once the injection stops, the supercritical CO tends to migrate 2.2 Trapping mechanisms upward through the porous and permeable rock as a result of the buoyancy effect created by its density difference com- The storage capacity, containment and injectivity of CO pared to other reservoir fluids and laterally via preferential are dependent on the geological and petrophysical proper- pathways until a cap rock, fault or other sealed discontinuity ties of the target formation. The injected supercritical CO is reached (Han 2008). This will prevent further migration is securely trapped underground via two major trapping of the CO as shown in Fig. 2. In depleted oil and gas fields, mechanisms (physical trapping and geochemical trapping) the movement of the CO can also be halted by abandoned (Fig. 1). The effectiveness of the storage process is governed wells sealed with solid cement plugs. The risk associated by a combination of both trapping mechanisms to ensure with such trapping is leakages behind casing or through long-term storage (Coninck et al. 2005). the mentioned plugs. Thus, many studies have been con- ducted on the leakage of CO through geological structures 2.2.1 Physical trapping and existing wells (Ambrose et  al. 2017; Eke et  al. 2011; Lewicki et al. 2007; Scherer et al. 2015; Shipton et al. 2004, Physical trapping is the process where C O maintains its 2006; Temitope and Gupta 2019; Zakrisson et al. 2008). physical nature after injection into an aquifer. It can be sub- divided into structural (hydrostratigraphic) and residual 2.2.1.2 Residual/capillary trapping As supercritical C O (capillary) trapping. Generally, the time period for physical percolates through storage formations, reservoir fluids are trapping is believed to be less than a century (Juanes et al. displaced. The movement of the C O occurs in two direc- 2006). tions: upward as a result of density differences and later - ally due to viscous forces. Reservoir fluid fills the spots left. 2.2.1.1 Structural trapping Structural trapping is usually However, some of the CO is left behind as disconnected/ the first form of trapping encountered during geological residual droplets in the pore spaces as displayed in Fig. 3. sequestration, and a similar mechanism has kept oil and Surface tension between CO and brine acts to halt the gas securely stored underground for millennia. Geologi- CO movement, thereby causing higher capillary entry pres- cal structures such as anticlines covered with cap rocks sure than the average rock pressure as suggested by Saadat- (an ultra-low-permeability layer), stratigraphic traps with/ poor et al. (2010). At this point, C O becomes immobilized without sealed faults are employed for the storage of C O in the pores at residual gas saturation. It is usually observed as a mobile phase or supercritical fluid. Maximization of in rocks with small-scale capillary heterogeneities. Recent this storage mechanism to ensure that C O injected remains studies have revealed that capillary trapping appears to be Structural (hydro stratigraphic) trapping Physical trapping Residual (capillary) trapping < 100 years Sorption trapping CO trapping mechanisms Slow diffusion in aqueous phase Solubility trapping Convective processes Geochemical trapping Mineral trapping Reaction with minerals Fig. 1 Different CO trapping mechanisms during the geological storage process 1 3 1032 Petroleum Science (2019) 16:1028–1063 Buoyant CO plume trapped by the seal (cap rock) Injection well Cap rock Porous media (aquifer) CO makes its way to the top of the aquifer Buoyant CO plume trapped by sealing fault Sealing fault Fig. 2 Physical trapping of injected C O as a result of the formation structure Cap rock2.2.2 Geochemical trapping CO gets trapped in the pore throats of the Geochemical trapping occurs when C O changes its physical porous media, as it makes and chemical nature by undergoing series of geochemical its way to the cap rock reactions with the formation brine and the rock and ceases to remain in the mobile or immobile phase. This interaction ensures the disappearance of CO as a separate phase and further increases storage capacity, making this an appropri- Grain ate feature of long-term storage. 2.2.2.1 Solubility trapping In a similar manner by which Porosity filled with water sugar dissolves in tea, C O dissolves in other fluids in either the supercritical or gaseous phase. Solubility trapping occurs as a result of the dissolution of the CO in the brine, Fig. 3 Residual trapping of injected C O as a result of the formation leading to dense CO -saturated brine. At this point, it ceases pore structure. Arrows in the diagram indicate the movement of the to remain a separate phase which eliminates any buoyancy CO plume effect. Over time, CO -saturated brine becomes denser than the surrounding reservoir fluids and falls to the bottom of a more efficient mechanism to trap CO underground in the the formation over time, culminating in more secure C O 2 2 short term compared to other short-term trapping mecha- trapping (Fig. 4). nisms (Burnside and Naylor 2014; Lamy et al. 2010). Its The dissolution of CO in the aqueous phase leads to the efficiency is due to exhibition of higher capillary forces to formation of weak carbonic acid which decomposes over + − buoyant forces, causing CO to appear as pore-scale bubbles time into H and HCO ions (Eq. 1). It can also react with 2 3 rather than being retained by a somewhat compromised cap other cations in the formation brines to form insoluble ionic rock. Furthermore, it provides an advantage of no risk of species as highlighted in Eqs. 1–4. CO solubility in forma- major failure associated with structural traps over a short tion water decreases as temperature and salinity increase. time scale (Jalil et al. 2012). + − CO + H O ↔ H + HCO 2 aq 2 (1) ( ) 1 3 Petroleum Science (2019) 16:1028–1063 1033 Cap rock Brine saturated with CO Convection Convection CO drops into the aquifer Convection Porous media (aquifer) Fig. 4 Pictorial representation of solubility trapping via convective mixing, one of the mechanisms for the dissolution of CO into aquifers modeling of these reactions is critical to the success of 2+ + CO sequestration predictions. This trapping mechanism Ca + CO + H O ↔ H + CaHCO (2) 2(aq) 2 3(aq) 2 is dependent on the rock minerals, the pressure of the gas, + + temperature and porosity and has been found to produce Na + CO + H O ↔ H + NaHCO (3) 2(aq) 2 3(aq) significant changes in the rock permeability and porosity (Benson and Cole 2008; Kampman et al. 2014). Perkins 2+ + Mg + 2CO + 2H O ↔ 2H + Mg(HCO ) . (4) 2(aq) 2 3 2(aq) et al. (2004) predicted from a simulation study that all the CO injected into the Weyburn Oil Field will be converted 2.2.2.2 Mineral trapping Mineral trapping occurs as a to carbon dioxide minerals after 5000 years. They reported greater mineralization capacity for the cap rock and overly- result of the conversion of CO into calcite due to reactions with solid minerals. This trapping is believed to be relatively ing formation rock, which is quite significant for leakage risk assessment. The capacity is estimated based on the amount slow since it occurs during/after solubility trapping and con- sidered as the most permanent form of storage. C O in the of minerals available for carbon dioxide precipitation and the quantity of C O used in the reaction processes. The most aqueous phase forms a weak acid which reacts with rock minerals to form bicarbonate ions with different cations striking advantage of mineral trapping mechanism over the other mechanisms is that it prevents CO from existing as a depending on the mineralogy of the formation. An example of such reaction with potassium basic silicate (Eq.  5) and separate phase, thus ensuring that its upward movement is halted and also enhances the formation of stable precipitates calcium (Eq. 6) is shown below: (Xu et al. 2001, 2003, 2004). 3K-feldspar + 2CO + 2H O ↔ Muscovite There are multiple mechanisms responsible for the stor- 2(aq) 2 (5) + − age operating simultaneously and on different time scales + 6Quartz + 2K + 2HCO which influence the storage capacity estimate. The interac- tion between various mechanisms is quite complex, evolves 2+ + Ca + CO + H O ↔ Calcite + 2H 2(aq) 2 (6) with time and depends highly on local conditions. An exam- ple of time scale evolution of different mechanisms at play Precipitation of carbon dioxide minerals is invariably in a deep saline formation is as shown in Fig. 5. induced by reactions with the rock formations depending on the mineralogy of these formations. Hence, geochemical 1 3 1034 Petroleum Science (2019) 16:1028–1063 The drainage and imbibition-like processes during the 2.3 Physical processes during  CO storage injection and post-injection stages of CO storage lead to hysteresis, a process where the capillary pressure and rela- A number of physical processes are involved in the injection and post-injection phases of carbon dioxide. CO trapping in tive permeability curves change pathways. This phenomenon has been described as being very critical to the successful aquifers is aided by three physical processes buoyancy (grav- ity), viscous forces and capillary forces (Kong et al. 2013). modeling of C O trapping processes (Ghomian et al. 2008; Juanes et al. 2006; Spiteri et al. 2005). This is because as the During the injection period of C O into aquifers, viscous forces are the dominant forces for the vertical and lateral CO migrates upward after the injection phase, the remain- ing CO plume gets disconnected due to water displacing migration of C O due to pressure gradients created by the 2 2 injection processes. The injected fluid (CO ) displaces the CO at the trailing edge and becomes a series of blobs. C O 2 2 is trapped in these blobs, and the mechanism is termed resid- formation fluid (brine) in a drainage-like process. In the post-injection phase, a combination of buoyancy ual or capillary trapping mechanism, which over time results in the dissolution of the CO in the formation brine. and capillary forces are responsible for the trapping of CO . 2 2 Buoyancy forces are usually greater than capillary forces and Heterogeneity and wettability of the aquifer are also key considerations in this mechanism. Heterogeneity has been viscous forces after injection in deep saline aquifers, leading to upward migration of C O . Buoyancy results from density subdivided into the small and large scales (Gershenzon et al. 2014; Lasseter et al. 1986; Li and Benson 2014). Viscous and differences between the injected CO and the aquifer brine causing the CO to migrate upward after injection displacing capillary forces dominate the flow, while gravity forces are generally regarded as unimportant when small-scale hetero- water in an imbibition-like process. The upward migration leads to gravity segregation, and geneities are considered. When large-scale heterogeneity is considered, the formation possesses variable pore throat sizes, further migration to the surface is prevented by the ultra- low permeable seal at the formation top. Once reaching which are likened to different capillary tubes sizes. As a result, a variable amount of entry capillary pressure is required to the top of the formation, the vertical migration is halted, while the lateral migration continues until a sealing fault displace the formation fluid. This leads to more CO being trapped as the entry pressure is overcome. Wettability and or formation boundary is reached. Thorough geomechani- cal analysis has to be made to ensure that leakage of C O interfacial tension changes have been proven to alter the capil- lary pressures in a porous medium (Bennion and Bachu, 2006; does not occur when the buoyant CO reaches the seal. One means of leakage is when the pressure of the C O is high Chiquet et al. 2007; Jung and Wan 2012; Park et al. 2015; Yang et al. 2005). The basic definition of capillary pressure enough to overcome the entry pressure of the seal (Hesse et al. 2006). Others could be due to the cap rock fractures, (Eqs. 7 and 8) and Young–Laplace equation (9) can be shown as follows in terms of mathematical forms: thermal stresses in the caprock as a result of temperature variation between the injected CO and aquifer and the pres- P = P − P (7) c nw w ence of open faults, fractures and abandoned wells (Chiquet et al. 2007; Goodarzi et al. 2013). Geomechanical considera- P = = (8) tions involving cap rock integrity are one of the factors that d d affect the sequestering capacity of the overlying seal. 2 cos w,CO P = P − P = (9) c CO w where d is diameter; R is the pore throat radius; P is defined Injection as the capillary pressure; P and P are the pressures of nw w Trap filling the non-wetting and wetting phases, respectively; P is CO Physical trapping the pressure of C O ;  is the water surface tension;  is the 2 w interfacial tension;  is the interfacial tension between Dissolution w,CO water and CO , and θ is the contact angle between the wet- Residual CO trapping ting medium and the rock surface. Mineralization In a typical CO –water system, CO is usually described 2 2 Adsorption as the non-wetting phase, while water is the wetting phase; 1 2 3 4 5 6 however, it has been proven that during the C O upward 110 10 10 10 10 10 migration, this wetting state can be changed (Broseta et al. Time, years 2012; Chiquet et al. 2007; Marckmann et al. 2003; Siemons et al. 2006; Yang et al. 2005). Equation 9 shows that the cap- Fig. 5 Time frame for trapping mechanisms in deep saline formations during and after injection (IPCC 2007; Metz et al. 2005) illary pressure is dependent on the pore throat radius, R, the 1 3 Petroleum Science (2019) 16:1028–1063 1035 interfacial tensions (  ) and the contact angles (θ) between storage and on the effective monitoring tools which could be the wetting medium and the rock surface. Therefore, the used for large-scale injections. These projects can be broadly interfacial tensions and wettability have a significant impact classified according to the storage location of the different on the sequestration capabilities of aquifer rocks. projects (saline, EOR, depleted gas reservoirs, ECBM), During the residence time of trapped CO in the blobs based on the mode of capture of the carbon dioxide (power and ganglia, C O dissolves into brine and this dissolution plants CCS projects and non-power plant CCS projects) and has been proven to occur by three principal mechanisms. based on the current status of the projects (planned, ongoing They are (a) diffusion of CO within the aqueous phase, and completed CCS projects). (b) reactions with the host minerals (classified as mineral The Sleipner project in Norway is the first case of large- trapping) and (c) convective mixing driven by slight density scale commercial C O storage in the world (Torp and Gale differences between the water saturated with CO and the 2004). The project began in 1996 and injected about a mil- unsaturated water (Ennis-King and Paterson 2003; Hassan- lion tons of CO into the sands of the Utsira Formation zadeh et al. 2007). Ennis-King and Paterson (2003) stated which is about 900 m below the bottom of the North Sea. that the dominant mechanism for long-term dissolution of The major incentive behind the commencement of the Sleip- CO in the formation brine is convective mixing rather than ner project was the need for minimization of taxes placed on pure diffusion as it is in orders of magnitude faster than dif- the direct emission of C O into the atmosphere (Christian- fusion and chemical reaction with the host mineral. sen 2001; Global CCS 2017; Kongsjorden et al. 1998). The The disproportionate dissolution of CO in brine leads to companies involved were faced with the options of paying gravitational instabilities which could further aid in solubil- heavy taxes for atmospheric emissions or injecting the CO ity trapping. Several researchers have worked on trying to into saline aquifers. Injection of C O into saline aquifers determine the onset time of convective mixing and the influ- provided a beneficial means for cost reduction by the par - encing factors (Bestehorn and Firoozabadi 2012; Ennis-King ties involved. Policies such as carbon dioxide pricing which and Paterson 2003; Hassanzadeh et al. 2007; Rasmusson would coerce companies with high CO emissions into con- et al. 2015; Riaz et al. 2006; Xu et al. 2006b). Ennis-King sidering the need for C O storage are major ways to ensure and Paterson (2003) used a linear stability analysis technique emissions into the atmosphere are significantly reduced. to provide an estimate of the time required for convective Another incentive for CO storage is the low cost of captur- instability to begin. They predicted the time to be typically ing; this has especially been noticed in the current field-scale up to tens of years, and this method has been used by sev- projects where CO injected was obtained from the separa- eral other researchers (Hassanzadeh et al. 2006; Hesse et al. tion of CO from produced gases, thus reducing the need 2006; Riaz et al. 2006). Riaz et al. (2006) determined the for capturing from coal plants which have not undergone critical time and wavelength or the onset of convective mix- separation and would cost more to capture from such plants. ing using the method of linear stability. It was determined The high cost of capturing CO from combustion processes that the critical time varies between 2000 years and 10 days has triggered the idea of carbon dioxide capture utilization and the critical wavelength varies between 200 and 0.3 m and storage (CCUS) where the CO could also be used for for a permeability variation of 1–3000 mD. Rasmusson et al. enhanced oil recovery and revenue derived from the pro- (2015) applied the Rayleigh number (Ra) in determining duced oil could be used to offset the cost of capturing and the onset of gravity-driven instabilities. They predicted that injecting into formations. The success of the Sleipner project a prerequisite for Ra, which must be greater than a critical elicited the increased field deployments on CO storage. Ra, is required for the onset of density-driven instabilities. Several pilot-scale projects have also been implemented Finally, as C O remains dissolved in the brine, it forms weak across the world. These projects typically inject small acids which react with the host minerals to form precipitates amounts of CO into identified formations for a small period (Gunter et al. 2000; Kumar et al. 2005; Xu et al. 2001). of time. These projects provide answers to questions of inter- est to the investigators. The first pilot-scale project in the USA was the Frio Project where about 1600 tons of CO was 3 Field‑scale projects on  CO storage injected at a depth of about 1500 m below the surface for a period of 10 days (Hovorka et al. 2006). The Frio Project CO sequestration projects are currently ongoing or in the provided information about the movement of CO plume and 2 2 planning stage across the world. Notable among these are the was able to validate numerical models developed to analyze Sleipner project in Norway, the Weyburn Project in Canada subsurface CO migration. Other notable pilot-scale projects and the In Salah Project in Algeria. Tables 1, 2, 3 and 4 pre- are the Cranfield Project (Hosseini et al. 2013; Hovorka et al. sent the lists of most of the projects. These field-scale injec - 2013), Decatur Project (Finley 2014; Senel et al. 2014), Ket- tions of CO into candidate formations have provided more zin site in Germany (Kempka and Kuhn 2013; Martens et al. insight into the physics of the processes involved in geologic 1 3 1036 Petroleum Science (2019) 16:1028–1063 1 3 Table 1 Storage projects across the world: saline aquifer projects Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Captain UK Captain Clean Energy 3.8 Mt/year Power Pipeline to offshore deep Planning James (2013) Limited (CCEL) owned by saline formations Summit Power and CO Deep Store Don Valley UK 2Co Energy Ltd, Samsung 4.9 Mt/year Power Pipeline for sequestration Planning Construction & Trading, in offshore deep saline BOC formations Killing Holme UK C.GEN NV and National 2.5 Mt/year Power Offshore pipeline for storage Planning Grid in deep saline aquifers Korea CCS Korea Korea Carbon Capture and 1 Mt/year Power Ulleung deep saline forma- Planning Lee et al. 2012 Sequestration R&D Center tion or the Gorae gas (KCRC) reservoir Lianyungang China Summit Power, National 1 Mt/year Power Binhai for saline aquifers or Planning Pang et al. (2012) and Qiao Grid, CO Deep Store North Jiangsu oilfields for et al. (2012) EOR Longyearbyen NorwayUNIS CO Lab-AS Not available Power Onshore storage in a saline Planning Braathen et al. (2012) aquifer White Rose UK Capture Power Limited, the 2 Mt/year Power Pipeline to offshore storage Planning Verdon (2014) consortium of Alstom UK in a saline aquifer Limited, Drax Power Lim- ited and National Grid plc Cranfield USA SECARB 1–1.5 Mt/year Non-power Saline reservoir, Tuscaloosa In operation since 2009 Lu et al. (2012) Sandstone Formation, down dip of the mature Cranfield Oil Field Citronelle USA SECARB, Denbury, South- 0.25 Mt/year Non-power The southern flank of the In operation since 2011 Haghighat et al. (2013) ern Energy Citronelle dome Decatur USA Archer Daniels Midland, 1 Mt/year Non-power Sequestration in Mount In operation since 2011 Zhou et al. (2010) MGSC (Led by Illinois Simon sandstone State Geological Survey), Schlumberger Carbon Services and Richland Community College Kevin Dome  USA Big Sky Partnership, 0.125 Mt/year Non-power The Duperow Formation Planning Riding and Rochelle (2005) Schlumberger Carbon (3900 ft) Services, Vecta Oil & Gas Ltd, Lawrence Berkeley National Lab, Los Alamos National Lab Petroleum Science (2019) 16:1028–1063 1037 1 3 Table 1 (continued) Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Wasatch Plateau USA  South West Partnership 1 Mt/year Non-power The Jurassic Entrada Forma- Planning Parry et al. (2007) (SWP), New Mexico tion and Navajo sandstone Institute of Mining and Technology, University of Utah, Schlumberger and Los Alamos National Laboratory Fort Nelson CanadaPlains CO Reduction Part- 2.2 Mt/year Non-power Middle Devonian carbonate Planning nership (PCOR), Spectra rock Energy, British Columbia Ministry of Energy, Mines and Petroleum Resources Quest Canada Athabasca Oil Sands Project: 1.1 Mt/year Non-power Injection into the Cambrian Launched in November Brydie et al. (2014) Shell Canada, Chevron Basal Sands 2015 Canada, and Marathon Oil Sands Sleipner Norway StatOil 0.9 Mt/year Non-power Utsira Formation In operation since 1996 Chadwick et al. (2004) Ketzin Germany GFZ German Research 0.06 Mt/year Non-power Stuttgart sandstone reservoir In operation since 2008 Schilling et al. (2009) Centre for Geosciences and Ketzin partners Snohvit Norway Statoil ASA, Petoro AS 0.7 Mt/year Non-power Saline Tubaen sandstone In operation since 2007 Hansen et al. (2013) (Norwegian state direct formation reservoirs interest), Total E&P Norge AS, GDF Suez E&P Norge AS, Norsk Hydro, Hess Norge ULCOS Florange France ArcelorMittal and ULCOS 0.7–1.2 Mt/year Non-power Onshore deep saline forma- On hold Global CCS (2017) (Ultra-Low-CO2-Steel) tions Ordos China Shenhua Group 1 Mt/year Non-power EOR/saline aquifer In operation since 2010 Li et al. (2016) Gorgon Australia Gorgon Joint Venture (Chev- 3.4–4.0 Mt/year Non-power Dupuy Formation 2.5 km Under construction Flett et al. (2009) ron Australia, ExxonMobil, below Barrow Island Shell, Tokyo Gas, Osaka Gas and Chubu Electric) Yulin China Shenhua Group, Dow 2–3 Mt/year Non-power Onshore deep saline aquifers Planning Li-ping et al. (2015) Chemicals Minami-Nagaka Japan EnCana, IEA 0.015  Mt/year Non-power Haizume Formation In operation since 2002 Zwingmann et al. (2005) Frio USA Bureau of Economic Geol- 177 t/day Non-power Frio Formation In operation since 2004 Hovorka et al. (2006) ogy of the University of Texas Teapot Dome USA Rocky Mountain Oilfield 170 t/day Non-power Tensleep and Red Peak In operation since 2006 Friedmann and Stamp (2006) Testing Center (RMOTC) Formation 1038 Petroleum Science (2019) 16:1028–1063 1 3 Table 2 Storage projects across the world: C O EOR/storage projects Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Santos Basin Brazil Petrobras, BG E&P 1 Mt/year Non-power Lula and Sapinhoá oil In operation since 2013 Saini (2017) Brasil Ltda, Petrogal fields Brasil Boundary Dam Canada SaskPower 1 Mt/year Power EOR in Weyburn field, October 2014-date Stéphenne (2014) excess to be used at the Aquistore project Bow City Canada Bow City Power, Can- 1 Mt/year Power Pipeline for EOR Planning Global CCS (2017) solv (Subsidiary of Shell), Luscar, Fluor Pembina Canada Penn West 50 t/day Non-power Cardium Formation In operation since 2005 1 Mt/year Non-power EOR in 2 carbonate In operation since 2000 White (2009) Weyburn-Midale Canada Cenovus Energy, fields: Weyburn field Apache Canada, PTRC (Petroleum (6500 t/day) and Technology Research Midale field (1200 t/ Center) day) Zama Canada PCOR, Apache Canada 0.067 Mt/year Non-power EOR in Zama Keg In operation since 2006 Smith et al. (2009) Ltd River oil field Alberta Carbon Trunk Canada Enhance Energy Inc. 14.7 Mt/year Non-power Injection into the Clive Planning Cole and Itani (2013) Line oil reservoir Daqing China Alstom and China 1 Mt/year Power EOR in nearby fields Planning Xiuzhang (2014) Datang Corporation Dongguan China Dongguan Taiyangzhou 1 Mt/year Power EOR in Shangdong Planning Liu et al. (2016) Power Corporation, Province Xinxing Group, Nan- jing Harbin Turbine Co, KBR, Southern Company, GreenGen China GreenGen 2 Mt/year Power Onshore EOR Planning Haszeldine (2009) Shengli China Sinopec 1 Mt/year Power EOR in Shangdong Planning Liang et al. (2009) Province Uthmaniyah Saudi Arabia Saudi Aramco 0.8 Mt/year. Power Pipeline for onshore In operation Liu et al. (2012) EOR Uthmaniyah Saudi Arabia Saudi Aramco 0.8 Mt/year. Non-power Pipeline for onshore In operation since 2015 Liu et al. (2012) EOR ESI CCS Project United Arab Emirates Abu Dhabi Future 0.8 Mt/year Non-power EOR, Rumaitha Zone- Started in 2017 Temitope et al. (2016) (UAE) Energy Company B, and Bab Zone-B (Masdar) and Abu Dhabi National Oil Company (ADNOC) Petroleum Science (2019) 16:1028–1063 1039 1 3 Table 2 (continued) Project name Country Company operators Total planned storage Capture mode CO fate Status of project References Taweelah United Arab Emirates Abu Dhabi Future 2 Mt/year Power Injection for EOR Planning (UAE) Energy Company (Masdar) and Taweelah Asia Power Company (TAPCO) and Emirates Alu- minium (EMAL) Bell Creek USA PCOR, Denbury 1 Mt/year Non-power EOR at Bell Creek oil Planning Gorecki et al. (2012) field, Montana Hydrogen Energy USA SCS Energy  3 Mt of CO captured Power Pipeline to onshore Planning Global CCS (2017) California Project annually EOR in Occidental’s (HECA) Elk Hills oil field Kemper County USA Mississippi Power, 3.5 Mt of CO annually Power Pipeline for onshore In construction Global CCS (2017) Southern Energy, EOR KBR 1 Mt/year Non-power EOR in West Hasting’s In operation since 2013 Global CCS (2017) Port Arthur USA Air Products and Chemicals, Den- and Oyster Bayou oil bury Onshore LLC, fields, Texas University of Texas Bureau of Economic Geology and Valero Energy Corporation Texas Clean Energy USA Summit Power Group 2–3 Mt/year captured Power EOR in the Permian Planning Global CCS (2017) Project (TCEP) Inc, Siemens, Fluor, Basin Linde, R.W. Beck, Blue Source and Texas Bureau of Eco- nomic Geology WA arish Petra Nova USA Petra Nova Holdings: 1.4 Mt of CO captured Power Pipeline for onshore In construction Global CCS (2017) a 50/50 partnership annually EOR in the West between NRG Energy Ranch Oil Field in and JX Nippon Oil Jackson County, & Gas Exploration Texas Corp. 1040 Petroleum Science (2019) 16:1028–1063 Table 3 Storage projects across the world: depleted reservoir projects Project name Country Company opera- Total planned Capture mode CO fate Status of project References tors storage In Salah Algeria BP, Sontrach, 1.2 Mt/year Non-power The Krechba 2004–2011 Ringrose et al. and Statoil Formation suspended (2009) Otway Australia CO CRC 0.065 Mt/year Non-power The Waarre In operation Underschultz et al. (Coopera- Formation since 2008 (2011) tive Research Center for Greenhouse Gas Technolo- gies) *ROAD (Rot- Netherlands E.ON Benelux, 1.1 Mt/year Power Pipeline for stor- Planning Global CCS terdam Opslag Electrabel, age in depleted (2017) en Afvang GDF Suez and reservoirs Demonstrate Alstom project) K12B Netherlands Gaz de France 0.365 Mt/year Non-power Rotleigendes In operation van der Meer et al. since 2004 (2006) Peterhead UK Scottish and 1 Mt/year Power Pipeline to off- Planning Southern shore depleted Energy (SSE) Goldeneye gas and Shell reservoir Northern Reef USA Midwest 0.365 Mt/year Non-power Depleted oil field In operation Trend Regional Car- in the Northern since 2013 bon Sequestra- Reef Trend tion Partner- (carbonate ship (MRCSP). reservoir) DTE Energy, Core Energy and Batelle Table 4 Storage projects across the world: C O ECBM projects Project name Country Company opera- Total planned Capture mode CO fate Status of project References tors storage Fenn Big Valley Canada Alberta Research 50 t/day Non-power Mannville group In operation since Gunter et al. (2005) Council 1998 CSEMP Canada Suncor Energy 50 t/day Non-power Ardley Formation In operation since Shi and Durucan 2005 (2005) Qinshui Basin China Alberta Research 30 t/day Non-power Shanxi Formation In operation since Wong et al. (2007) Council 2003 Yubari Japan Japanese Ministry 0.004 Mt/year Non-power Yubari Formation In operation since Shi et al. (2008) of Economy, 2004 Trade and Industry Recopol Poland TNO-NITG 1 t/day Non-power Silesian Basin In operation since van Bergen et al. (Netherlands) 2003 (2003) 2013) and the Otway Project in Australia (Etheridge et al. (2015) noted that the embryonic stage of technology on 2011; Underschultz et al. 2011). CO capture would mean high costs of capture from power Even though carbon dioxide capture is outside the scope plants for early movers. Early movers need to be encour- of this review, it is obvious that the deployment of many aged by governments through subsidies. Successful cases carbon dioxide storage projects would be dependent on the of subsidies by the government can be seen in the Bound- cost and success of carbon capture processes. Celia et al. ary Dam Project by SaskPower and the Quest Project by Shell both in Canada. 1 3 Petroleum Science (2019) 16:1028–1063 1041 solved the equations using the streamlined methodology. 4 Modeling strategies employed for  CO Though their model was able to solve the pressure-driven storage flow in complex flow fields, it was limited by the assump- tions of a simple geochemical model and incompressible Numerical modeling is typically carried out before the flow. Qi et al. (2009) used the model developed by Obi and commencement of injection projects. They are used for Blunt (2006) to postulate a design strategy for injection predictions and optimizations. The flow path of the injected of CO which would render a large percentage of the CO CO needs to be predicted prior to injection. Furthermore, 2 2 immobile on the pore scale. As their work was focused on the optimization of well location needs to be properly maximizing the gas trapped via the residual gas trapping assessed during the planning phase. Several authors have mechanism, they modified the existing model by assigning attempted to model the plume movement of injected C O in relative permeabilities on a block by block basis. All in all, saline formations. Modeling of CO storage in saline aqui- these papers have been able to demonstrate the feasibility fers is usually performed using either analytical or numeri- of modeling storage of CO in saline aquifers by employing cal models. The choice of modeling technique employed is the streamlined methodology. Streamline simulations are, dependent on the aims of the researchers, the nature of the however, best suited to processes where limited pressure problem and the data available. Analytical models have the changes are expected to occur. Given that only injection is advantage of providing a quick insight into the suitability usually modeled in CO storage in saline aquifers thus lead- of a formation for storage. Zhou et al. (2008) employed an ing to significant pressure changes, streamline simulations analytical model to determine the storage capacity in saline have found limited applications in C O storage modeling. aquifers and expected pressure buildup during storage Vertical equilibrium models work by discretizing the simu- operations. Mathias et al. (2009a) developed approximate lation domain only in the horizontal direction leaving one solutions for pressure buildup in aquifers assuming vertical layer in the vertical direction. Two forms of the vertical pressure equilibrium and accounting for the Forchheimer equilibrium model exist: vertically integrated numerical flow of CO and brine. Solutions from the study were sub- models which include capillary forces and analytical mod- sequently applied in the screening of potential C O storage els including a sharp interface where the capillary pres- sites (Mathias et al. 2009b). Analytical models have been sure zone is thin with homogeneous formation parameters. used for plume migration studies. Nordbotten et al. (2005a) The technique capitalizes on the strong density differential also developed approximate solutions for the prediction of between supercritical C O and the in situ brine which leads the plume migration path in a CO storage site. The model to a marked upward increase in the CO . Particularly, on was validated with the commercial simulator ECLIPSE short time scales, the density differential could lead to a with very good accuracy. The underlying assumptions of strong buoyancy segregation of the two fluids. The idea analytical models are, however, too simplistic and cannot behind this technique is to derive a better understanding account for reservoir property and model geometry het- of the lateral plume spread and the segregation between erogeneities. More so, the complex geochemical reactions the different fluid phases. Its limitation is in its inability expected in CO storage cannot be reliably captured by to model heterogeneity in the vertical direction. The tech- analytical models. Streamline simulations, vertical equilib- nique has, however, been applied (Gasda et al. 2009, 2011) rium models and regular, conventional grid-based numeri- in modeling of CO storage. Another modeling technique cal models are forms of numerical modeling techniques which has been applied to the simulation of C O in aquifers which have been applied for the modeling of C O storage is the inversion percolation technique. In this approach, (Cavanagh and Haszeldine 2014; Gasda 2010; Jiang 2011; viscous forces are ignored; therefore, the only forces that Li et al. 2012; Obi and Blunt 2006; Pruess 2008; Saadawi dominate the flow are the capillary and gravity forces. et al. 2011; Wheeler et al. 2008). Streamline simulations Consequently, this technique is most suitable in systems work by splitting the simulation domain into small grid with low fluxes. Inversion percolation is employed when sizes and determining the pressure in each grid block using the capillary number (ratio of viscous forces to capillary a finite difference technique. The resulting pressure field force) is less than 0.0001. High-resolution inversion perco- is applied in tracing the streamlines which determine the lation models are noted for their simplicity and the speed expected flow fields. As opposed to other forms of numeri- of their numerical solutions. Limitations of this approach cal modeling, streamline simulations are faster and com- are, however, found when flow rates are high and capillary putationally efficient as flow equations are reduced to one- heterogeneity is not pronounced. Notably, this approach has dimensional equations along the streamlines. Obi and Blunt been employed in the modeling of the In Salah Field Pro- (2006) and Qi et al. (2009) have applied streamline simula- ject and the Sleipner storage with a high degree of accuracy tions in the modeling of CO storage. In their model, Obi (Cavanagh and Ringrose 2011; Cavanagh and Haszeldine and Blunt (2006) coupled transport and flow equations and 2014). Conventional 3D simulations making use of highly 1 3 1042 Petroleum Science (2019) 16:1028–1063 developed numerical discretization techniques have been k k k ⎛ ⎞ xx xy xz used to overcome the shortcomings of the other techniques ⎜ ⎟ k = k k k (12) yx yy yz ⎜ ⎟ by incorporating all relevant physics such as expected pres- k k k ⎝ ⎠ zx zy zz sure increases and heterogeneities in both the vertical and horizontal directions. Typically, they employ finite differ - Conservation of energy can also be solved for by equat- ence/element/volume techniques to solve transport and flow ing the summation of the time rate of change of the energy equations. In addition, they are able to couple other related term, advection and conduction terms to the source term as physical phenomena such as geochemistry, geomechanics shown below: and thermal changes. As a result of the detailed modeling of inherent physics, the regular 3D grid-based numerical modeling techniques are more computationally costly than  ( s U )+(1 − ) C T s s the other techniques. Most commercial simulators which (13) have been employed for modeling of CO storage issues + ∇ ⋅ ( q H ) −∇ ⋅ (∇T)= S have full modeling capabilities (Class et al. 2009; Nghiem et al. 2009). where U represents the specific internal energy, H is the Modeling of C O storage is a multi-component, multi- specific enthalpy, T is the temperature, C is the specific heat phase process with the two fluid phases as the brine and capacity, and all other symbols have definitions as described a CO -rich phase and the components like C O, H O, dis- 2 2 2 earlier. Subscript s represents the solid phase. solved salts in the brine and rock minerals. It should be noted These equations (Eqs. 10, 11 and 13) represent the fun- that the number of components modeled can be different damental equations for the modeling of storage of CO in depending on the problem to which it is applied. The funda- porous media (DePaolo et al. 2019; Nghiem et al. 2004; Pan mental equations used in CO storage modeling are basically et al. 2018a). These equations could be coupled with geo- the same as equations that describe the flow of oil, gas and chemical reactions, geomechanical modules and other rel- water in porous media. These equations are the conserva- evant physical phenomena. The solution of these equations tion of mass, momentum and energy. Constitutive relations requires either a sequential, simultaneous or fully coupled are used to formulate solutions for these equations. Other approach. physics which could be coupled with the basic equations are Over the years, researchers have made numerous attempts equations that predict geomechanical effects and geochemi- to describe underground CO migration and trapping mecha- cal reactions among others (Temitope and Gupta 2019). nisms using numerical analysis. Weir et al. (1996) developed The conservation of mass equation for components can be a two-dimensional model to evaluate C O quantities that written as the summation of the advection, diffusive terms migrated beyond a cap rock after C O injection for 10 years and the time rate of change of mass which equal a source into a 3-km-deep aquifer at a mass transfer rate of 100 kg/s. or sink term. They varied the confining layer’s permeability in order to determine the amount of CO that could pass through the ( s X ) + ∇ ⋅ ( q X )− ∇ ⋅ (  D ∇X )= S i layer. They concluded that a low-permeability seal should i i i overlay any target formation as this would mean that higher (10) capillary pressures would be required for the C O to pen- Darcy’s law for a single-phase flow can be written as etrate the seal. Another C O storage study conducted by q kk researchers at the Alberta Research Council (Gunter et al. v = =− (∇p +  g∇z) (11) 1993; Law and Bachu 1996) for the Upper Manville Group where the modeled formation was a Cretaceous glauconitic where t represents the time,  represents the porosity,  is sandstone aquifer 1.46 km in depth. The formation top of the density, q is the Darcy flux, k is the permeability tensor, the aquifer was overlain by several regional-scale aquitards k is the relative permeability, D is the diffusivity, X is the (low-permeability shale layers) that inhibited upward migra- mole fraction, s is the saturation term,  is the tortuosity, tion of the injected CO . The unevenness of the formation S denotes the source/sink term, v is the velocity vector,  is permeability was modeled based on drill-stem tests per- the dynamic viscosity, p is the pressure, g is the acceleration formed during exploration. The study showed no C O leak- due to gravity, and z represents the depth. Subscripts  and age during the modeled time scale. i are the phase and index, respectively. Nghiem et  al. (2004) developed a fully coupled EOS The permeability tensor can be written fully as compositional simulator for modeling C O storage in aqui- fers. The module consisted of geochemical reactions such as 1 3 Petroleum Science (2019) 16:1028–1063 1043 gas dissolution in the aqueous phase, chemical equilibrium General Purpose Reservoir Simulator (AD_GPRS) by Stan- reactions, mineral dissolution, and precipitation. The highly ford University (Benson et al. 2013; Fan 2006; Iskhakov coupled sets of nonlinear equations were solved simulta- 2013), MUFTE-UG (Multiphase Flow Transport and Energy neously using the Newton approach. The geochemistry Model on Unstructured Grids) developed by a joint effort module of the simulator was validated with the Geochemist of the University of Stuttgart and the University of Heidel- Workbench (GWB) developed at the University of Illinois berg (Ebigbo et al. 2006), IPARS-CO2 (Integrated Parallel with high accuracies. The resulting codes were applied on Accurate Reservoir Simulator) developed by the University two numerical grids: a 2D reservoir used to analyze the of Texas at Austin (Kong 2014; Wheeler et al. 2008); also impact of mineral trapping and a 3D grid used to study the existing are several simulators by the National Laboratories evolution of the CO plume. Rutqvist et al. (2010) coupled in the USA including TOUGH and TOUGH2 usually used a geomechanical simulator (FLAC3D) and a multi-phase in collaboration with ECON2 (Hovorka et al. 2006; Pruess flow simulator (TOUGH2) to study the ground deforma- et al. 2002), STOMP Subsurface Transport over Multiphase tions which would occur at the In Salah storage site in Alge- Processes (Bonneville et al. 2013) [see Table 5 for full list]. ria. Surface deformation results derived from monitoring The difference between most of these simulators lies in the using interferometry synthetic aperture radar (InSAR) were numerical methods and discretization technique used, the employed in this study to validate the numerical models and inclusion or non-inclusion of certain physics and the cou- displayed good agreements with obtained results. A sum- pling methods of the physics. mary of the workflow for most of the reservoir simulators Numerical simulations have been applied to assess the for CO storage issue is provided in Fig. 6. feasibility of commercial storage in aquifers. In a recent Many researchers exploring CO storage issues have study, Temitope et al. (2016) employed the Computer Mod- focused more on simulations for large-scale analysis with elling Group (CMG) simulator with an advanced geochemi- most experiments carried out aimed at better understanding cal modeling module to evaluate the possibility of commer- the physics of the processes that occur during the injection cial injection in the Shuaiba aquifer of the Falaha syncline and post-injection phases. Thus, due to the complex nature in the United Arab Emirates (UAE). Simulation results were of storage of CO and the time period taken for carbon able to provide the possible migration path of injected C O 2 2 dioxide to be stored underground, the only effective way to into the aquifer. In modeling the impact of thermal fac- understand the storage capacity of an aquifer before injec- tors on the injection of CO into the FutureGen 2.0 Site in tion commences is through modeling and simulations. This Illinois in the USA, Nguyen et al. (2016) made use of the explains why there exists a myriad of simulators which have simulators STOMP-CO coupled with the ABAQUS finite the capacity to model CO storage in aquifers; among them element simulator. Results suggested that in the range of includes CMG (Computer Modelling Group) GEM-GHG temperatures in which injection would take place, fracturing Module (Nghiem et al. 2004, 2009), ECLIPSE 100 and 300 would be unlikely to happen due to thermal factors. Basirat (Schlumberger), CO2STORE Module (Pickup et al. 2011, et al. (2016) employed the TOUGH2 simulation codes to 2012; Sifuentes et  al. 2009), Automatic Differentiation model the injection of CO into an experimental site in Basic equations Darcy law Conservation of mass Conservation of energy Numerical techniques used to couple equations simultaneous/sequential coupling Modelling of geochemical Modelling of geochemical Complex processes reactions effects Dual porosity/dual Adsorption Cap rock integrity permeability Mineral dissolution/precipitation Stress relation Residual gas trapping CO dissolution in brine EOS modelling Fig. 6 Workflow for CO storage modeling 1 3 1044 Petroleum Science (2019) 16:1028–1063 1 3 Table 5 List of simulators and codes for CO storage Simulators Full names Description Developers Relevant literature ABAQUS-FEA ABAQUS-FEA Geomechanical, single- and two- SIMULIA Gemmer et al. (2011) and Le Gallo phase flow et al. (2006) AD_GPRS Automatic Differentiation—General Generalized multiple phase composi- Stanford University Huo and Gong (2010) and Li et al. Purpose Reservoir Simulator tional/thermal model for unstruc- 2012) tured grids CO —PENS CO —Predicting Engineered Natural System-level modeling of the long- Los Alamos National Laboratory Pawar et al. (2006) and Stauffer et al. 2 2 Systems term fate of C O in sequestration (LANL) (2006) sites CO Toolkit Permedia ™ Permedia High-resolution petroleum migra- Landmark Cavanagh and Ringrose (2010, 2011) tion simulator for multi-phase flow behavior in porous, faulted and fractured media Advanced Resources International Mingjun et al. (2010) and Schepers COMET3 COMET# Black oil production, hydrocarbon et al. (2009) dioxide recovery from desorption- controlled reservoirs COMSOL Multiphysics COMSOL General partial differential equation COMSOL Farajzadeh et al. (2009) and Houdu solver with finite element solver et al. (2008) COORES CO Reservoir Environmental Multi-component, three-phase and Insititut français du pétrole Estublier and Lackner (2009) and Le Simulator 3D fluid flow in heterogeneous Gallo et al. (2006) porous media CrunchFlow Crunch Flow 3D, multi-phase transport with equi- Lawrence Livermore National Labo- Siirila et al. (2012) and Steefel and librium and kinetic mineral–gas– ratory | Lasaga (1994) water reactions RetrasoCodeBright Retraso (REactive TRAnsport of A solution of the flow, heat and Technical University of Catalonia Kvamme and Liu (2009) and Olivella SOlutes) CodeBright (COupled geomechanical model equations (UPC), Barcelona, Spain et al. (1996) DEformation of BRIne Gas and Heat Transport) DuMux DUNE for Multi-(Phase, Component, Multi-scale, multi-physics toolbox for University of Stuttgart Class et al. (2009) and Flemisch et al. Scale, Physics) the simulation of flow and transport (2007) processes in porous media ECLIPSE ECLIPSE Non-isothermal multi-phase flow in Schlumberger Juanes et al. (2006), Martens et al. porous media (2012) and Sifuentes et al. (2009) ELSA Estimating Leakage Semi-Analyti- Provides quantitative estimates of Princeton University Nordbotten et al. (2005b, 2009) cally fluid distribution and leakage rates in systems involving a sedimentary succession of multiple aquifers and aquitards FEFLOW FEFLOW Solving the groundwater flow equa- DHI-WASY Melikadze et al. (2013) tion with mass and heat transfer, including multi-component chemi- cal kinetics Petroleum Science (2019) 16:1028–1063 1045 1 3 Table 5 (continued) Simulators Full names Description Developers Relevant literature FEHM Finite Element Heat and Mass Trans- Non-isothermal, multi-phase flow Los Alamos National Laboratory Pawar et al. (2005) and Robinson et al. fer Simulator (including phase-change) in unfrac- (2000) tured and fractured media with reactive geochemistry & geome- chanical coupling GASMOD/GCOMP GASMOD/GCOMP Multi-phase reservoir simulator PHH Engineering Software Limited Palmer and Mansoori (1996) GEM-GHG Generalized Equation of State Non-isothermal multi-phase flow in Computer Modelling Group (Canada) Kumar et al. (2005) and Nghiem et al. Model—Greenhouse Gases porous media (2009) GMI-SFIB Geomechanics International—Stress Three-dimensional stress mod- Geomechanics International Fang (2011) and Fang and Khaksar and Failure of Inclined Boreholes eling for compressional (wellbore (2012) breakout) and tensional (tensile wall fractures) stress failure, fracture modeling GWB The Geochemist’s Workbench Chemical reactions, pathways, kinet- University of Illinois Bethke and Yeakel (2009) and Lu et al. ics (2011) IPARS-CO2 Integrated Parallel Accurate Reser- Non-isothermal compositional EOS University of Texas at Austin Delshad et al. (2011) and Kong et al. voir Simulator coupled with geochemical reactions (2015) MASTER Miscible Applied Simulation Tech- Black oil simulator, compositional National Energy Technology Labora- Ammer and Brummert (1991) niques for Energy Recovery multi-phase flow tory METSIM 2 METSIM 2 A non-isothermal multi-component Imperial College Durucan et al. (2004) and Law et al. coalbed gas simulator (2004) MODFLOW MODFLOW Solving groundwater flow equation to US Geological Survey (USGS) Nicot et al. (2009) simulate the flow through aquifers MoRes Modular Reservoir Simulator A modular object-oriented design for Shell Class et al. (2009) and Wei and Saaf black-oil, equation-of-state (EOS) (2009) and K-value compositional simula- tions ACCRETE Athena Carbon Dioxide Capture and Thermal multi-phase 3D reactive University of Bergen Hellevang and Kvamme (2006, 2007) Storage Geochemistry Module transport numerical code MUFTE-UG Multiphase Flow Transport and Isothermal and non-isothermal multi- University of Stuttgart Assteerawatt et al. (2005) and Ebigbo Energy Model on Unstructured phase–multi-component flow and et al. (2006) Grids transport processes in porous and fractured porous media NFFLOW-FRACGEN Fracture Network Generator Flow Two-phase, multi-component flow in National Energy Technology Labora- Myshakin et al. (2015) and Schwartz Simulator fractured media tory (2006) NUFT Non-isothermal Unsaturated–satu- Non-isothermal multi-phase flow and Lawrence Livermore National Labo- Hao et al. (2012) rated Flow and Transport model chemical reactions in porous media ratory | OGS: [Couples GEM, BRNS, OpenGeoSys Porous and fractured media THMC Helmholtz Centre for Environmental Graupner et al. (2011) and Li et al. PHREEQC, ChemApp, simulation Research (UFZ) (2014) Rockflow] 1046 Petroleum Science (2019) 16:1028–1063 1 3 Table 5 (continued) Simulators Full names Description Developers Relevant literature PFLOTRAN Parallel Reactive Flow and Transport Non-isothermal multi-phase, multi- Los Alamos National Laboratory Lu and Lichtner (2005, 2007) component, chemically reactive flows in porous media PHAST PHAST Multi-component, 3-D transport with US Geological Survey (USGS) Parkhurst et al. (2004) equilibrium and kinetic mineral– gas–water reactions PHREEQC PHREEQC The low-temperature aqueous geo- US Geological Survey (USGS) Parkhurst and Appelo (2013) and van chemical simulator Pham et al. (2012) PSU- COALCOMP Pennsylvania State University- Three-dimensional, two-phase, dual Pennsylvania State University/ Bromhal et al. (2005) and Manik et al. COALCOMP porosity, sorption, fully implicit, National Energy Technology (2000) compositional coalbed methane Laboratory reservoir simulator ROCKFLOW Rock Flow Multi-phase flow and solute transport Bundesanstalt für Geowissenschaften Kolditz et al. (2003) processes in porous and fractured und Rohstoffe and University of media, as well as thermal–hydrauli- Hannover cally–mechanical (THM) coupled processes SOLVEQ/CHILLER CHILLER Multi-component multi-phase Department of Geological Sciences, Palandri and Kharaka (2005) and Reed equilibrium geochemical calcula- University of Oregon and Spycher (2006) tion software based on minimum free-energy RTAFF2 Reactive Transport and Fluid Flow Non-isothermal multi-phase and a French Geological Survey (BRGM) multi-component flow simulator SIMED II SIMED II Two-phase three-dimensional multi- The Netherlands/Commonwealth Stevenson and Pinczewski (1995) and component coalbed gas simulator Scientific and Industrial Research van Bergen et al. (2002) Organization (CSIRO), Australia STOMP Subsurface Transport over Mul- Non-isothermal multi-phase flow in Pacific North-West National Labora- Bonneville et al. (2013) and White tiphase Processes porous media, coupled with reactive tory et al. (2012) transport. TOUGH/TOUGH2 Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Pruess et al. (2002) and Pruess and ter and Heat unfractured and fractured media tory Spycher (2007) TOUGH-FLAC Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Rutqvist (2012) and Rutqvist and ter and Heat unfractured and fractured media tory Tsang (2003) with geomechanical coupling TOUGHREACT Transport of Unsaturated Groundwa- Non-isothermal multi-phase flow in Lawrence Berkeley National Labora- Xu et al. (2006a) ter and Heat Reactive Transport unfractured and fractured media tory with reactive geochemistry VESA Vertical Equilibrium with Subsurface Vertically averaged numerical model Princeton University Gasda et al. (2009) Analytical for large-scale flow coupled with an embedded analytical model for wellbore flow Petroleum Science (2019) 16:1028–1063 1047 Maguelone, France. Geophysical monitoring tools were analysis expected from commercial injection into the forma- used in their field experiments to gain useful information tion. Different methods exist for the calculation of storage about the site and also to monitor the movement of the gas. volumes and can be broadly classified into static and dynamic They highlighted the importance of accounting for geologi- estimation methods. As the names would suggest, static esti- cal heterogeneity in modeling procedures. In addition, the mation methods do not change with time and only require study was able to provide information on the usefulness of basic rock and fluid properties. They are typically determined geophysical monitoring tools in analyzing plume migration using volumetric and compressibility parameters. Con- in storage sites. versely, dynamic estimation methods vary with time and are Benchmark studies have thus been performed to under- determined using reservoir simulations and some analytical stand the capabilities of different softwares used for carbon methods which incorporate time-dependent variables in their dioxide storage. Pruess et al. (2002) performed a critical derivations. Estimation of CO storage capacity in geological comparison on the performance of different commercial media is at best an approximation due to the many uncertain- reservoir simulator codes for accurate prediction of CO ties present both in the formation (heterogeneity) and in the storage processes (that is TOUGH2, Geoquest’s ECLIPSE, physics of the processes. The level of uncertainty also varies CMG’s GEM, etc.). They concluded that all softwares could with the method being used to determine the storage capac- be used to simulate the essential flow and transport pro - ity and the amount of available data. The methodology to be cesses that would accompany geologic storage. However, used for the determination of the capacity is dependent on the hydromechanical process would only be solved by one the formation type, that is coal seams, depleted oil and gas code TOUGH-FLAC. Law et al. (2004) analyzed the results reservoirs or saline aquifers. In addition, the extent of the of five simulators to a benchmark problem for CO storage storage medium may determine the approach to be used in issues in coalbed formations. Class et al. (2009) also per- storage capacity determination. Open boundaries where the formed a benchmark study with the use of different simula- extent of the media is assumed to be infinite, closed where tors to address the problems related to C O storage in geo- the extent of the media is assumed to have a finite end and logic formations. The outcome of such benchmark studies semi-closed are all different forms available in the literature illustrates that the results of the simulation of any storage for storage capacity determination. problem would depend on the simulator used and are highly Because candidate storage sites are usually not fully dependent on the numerical methods used and the physics of characterized before estimates are made, they are usually processes implemented. It is suggested that the choice of the reported as a low- and high-capacity estimate of storage simulator to be used would depend on the physical processes (DOE 2007) with Monte Carlo simulations employed to being focused on for best results. account for uncertainties. Two primary methodologies are Simulation of CO storage is generally a little more dif- being used; they include the methodology by the Depart- ficult than conventional simulations due to the interplay ment of Energy (DOE) of the USA (DOE 2007) and the between phase change, composition and reservoir hetero- Carbon Dioxide Sequestration Leadership Forum (CSLF) geneity which require highly efficient computational algo- (Bachu et al. 2007b) and the formulas used by the two bod- rithms (Jiang 2011). The striking difference between CO ies for storage determination are summarized in the next storage issues and conventional porous media modeling is subsections. the large temporal and spatial scale differences. A multi- scale methodology which incorporates advanced numerical 5.1 Coal seams schemes may be the best way to approach such scale differ - ences in such a way as to capture the complex multi-phase, The formulas for calculating the storage capacity of coal multi-component species, and physics in heterogeneous sys- seams by the DOE and CSLF methods are as follows:DOE: tems and also save computational cost. Such multi-scale, M = Ah CE multi-physics approach has been implemented in the devel- g (14) opment of certain simulators (Flemisch et al. 2007). CSLF: M = Ah(1 − f − f ) n G CO a m CO c c (15) 2 2 5 Capacity estimation for  CO storage G = V ∗ projects cs L (16) P + P An initial estimate of the storage capacity of a formation where A represents the area, h is the thickness, h is the gross is required for successful implementation of CCS projects. thickness, C is the concentration of C O standard volume Such estimates assist in project planning and in potential risk per unit of coal volume, f and f are the ash and moisture a m 1 3 1048 Petroleum Science (2019) 16:1028–1063 weight fraction of coal, M is the mass storage, E is the C O CSLF: storage efficiency factor that reflects a fraction of the total (P Z T ) s r r coal bulk volume that is contacted by CO ,  is the density, Gas fields ∶ M =  R (1 − F )OGIP (19) 2 CO CO f IG 2 2 (P Z T ) r s s n is the bulk coal density, G is the gas coal content, G is c c cs the gas content at saturation, V and P are the Langmuir L L R OOIP volume and pressure, respectively, and P represents the Oil fields ∶ M =  − V + V (20) CO CO iw pw 2 2 pressure. The Langmuir volume is the maximum adsorp- f tion capacity of the gas for a particular coal at a defined where A represents the area, h is the net thickness,  is the n e temperature and infinite pressure. Its unit is usually given in effective porosity, M is the mass storage, E is the CO stor- scf/ton (volume of gas per mass of unit coal). The Langmuir age efficiency factor that reflects a fraction of the total pore pressure (also known as the critical desorption pressure) is volume from which oil and/or gas has been produced and the pressure at which one half of the Langmuir volume can that can be filled by CO , ρ is the density, B is the formation 2 f be adsorbed/stored. volume factor, S is the average water saturation, P repre- In the CSLF method, the storage capacity available in coal sents the pressure, Z and T are the compressibility factors, seams for C O is determined in a manner akin to the deter- respectively, R is the recovery factor, OOIP and OGIP stand mination of initial gas in place in coalbed methane reservoirs for the original oil and gas in place, respectively, F is the IG as shown in Eq. 15. The ability of the coal gas to adsorb the fraction of injected gas, and V and V are the volumes of iw pw injected CO is dependent on pressure, temperature and coal injected and produced water, respectively. characteristics of the formation. The gas content at saturation is determined by Eq. 16. The two equations assume that the 5.3 Saline aquifers CO contacts all the available coal and that the coal adsorbs CO to full capacity. In reality, however, this may not be prac- Bachu et al. (2007a) as part of research conducted by the ticable; hence, a correction factor is introduced to account for Carbon Sequestration Leadership Forum (CSLF) expressed the non-ideality as given in Eq. 17: the effective storage capacity available in structural traps M = M ∗ C ∗ R in terms of volume and mass of CO as in Eqs. 21 and 22, e CO f (17) 2 respectively. The boundaries of the aquifer are considered where M is the effective storage capacity, C is the comple- to be open. tion factor, and R is the recovery factor. The product of completion and recovery factor is together known as the gas V = Ah(1 − S )C CO w c (21) 2 irr deliverability. The completion factor C is an estimate of that M = Ah(1 − S ) C part of the net cumulative coal thickness within the drilled CO w CO c (22) 2 irr 2 coal zone that will contribute to gas production or storage; it where the spatial variation of the formation is known; the is dependent on the individual thickness of the separate coal volumes can be expressed as seams and on the distance between them and is lower for thin coal seams than for thick ones (Bachu et al. 2007a). Monte V = (1 − S )dxdydz ∗ C (23) CO w c 2 irr Carlo uncertainty analysis can be employed to account for uncertainties in the determination of unknown parameters. where A is the area, h is the thickness, S is the irreducible irr water saturation,  is the density of C O , and C is the CO 2 c 5.2 Oil and gas reservoirs capacity coefficient which is dependent on the trap hetero- geneity, buoyancy and sweep efficiency. Estimation of available storage capacity in depleted oil and gas The capacity coec ffi ient is usually site-specic fi and is best reservoirs is not as complicated as with coal seams and saline determined through numerical simulations or detailed field aquifers as these reservoirs have been adequately character- work. It incorporates effects such as the heterogeneity of the ized during the production stages of the reservoir. The basic aquifer, buoyancy effect and sweep efficiency. The Interna- assumption in the formulation of storage capacities is the avail- tional Energy Agency Greenhouse Gas R&D Programme ability of all the pore spaces vacated by hydrocarbon fluids. In (IEAGHG 2009) in their study evaluated the capacity coeffi- other words, it is assumed that the formation fluids have not cient as a function of lithology based on extensive numerical been replaced by water from any supporting aquifer around studies. The values derived for carbonate formations based the region of the field. The storage capacity by the CSLF and on the 10th, 50th and 90th percentiles were 1.41%, 2.04% DOE methods are as stated below. and 3.27%, respectively. The formula for capacity estimates DOE: derived by the US Department of Energy (DOE-NETL 2015) is similar to that of the CSLF. The only difference lies M = Ah  (1 − S )B E (18) n e w f 1 3 Petroleum Science (2019) 16:1028–1063 1049 in the capacity coefficient given for the carbonate formations is the time, and ΔP is the maximum allowable pressure max with the DOE estimating the 10th, 50th and 90th percentiles increase. as 0.51%, 2.0% and 5.5%, respectively. Dynamic simulations still represent the best method for The storage volume available by residual trapping can be the determination of storage capacities of geological forma- determined using the correlation below: tions selected for storage as they contain detailed informa- tion regarding the petrophysical properties of the formation. V =ΔV S CO t trap CO t (24) 2 2 Coupled with this, numerical simulators nowadays have where S (saturation of CO ) is dependent on the hyster- CO t 2 embedded in their simulators the ability to calculate the stor- esis effects of the relative permeabilities and the CO satura- 2 age capacity provided by the different storage mechanisms tions during reversal flow. over an extended period. Analytical determination methods As highlighted earlier, the dissolution of C O in brine 2 such as fractional flow theory (Moghanloo et al. 2015) and is a continuous and slow process that is dependent on the relative permeability curve analysis method (Zhu et al. 2017) convection, diffusion and dispersion. The storage capacity for the determination of storage volumes can also be found on a basin and regional scale, as determined by Bachu et al. in the literature. (2007a) for solubility trapping, is given below The aforementioned described techniques have been employed mainly in the determination of storage capacities CO CO 2 2 M = ( X −  X )dxdydz (25) CO t s s o o across the world. Lindeberg et al. (2009) used both analytical and reservoir simulations to estimate the available storage where  is the porosity,  is the density, X stands for mass capacity in the Utsira Formation of Norway. Their reser- fraction, M denotes the mass, and subscripts s and o denote voir simulations were done in such a way to model elevated the carbon dioxide content at the saturation and initial pressures in the aquifer. In addition, a CO breakthrough stages, respectively. The time frame required for mineral from production wells was also monitored in estimate deter- trapping to occur makes it difficult to provide correlations mination. In China, Liu et al. (2005) estimated the storage for the determination of the mineral trapping capacity. capacities in gas fields and coalbeds present in the country. Zhou et al. (2008) devised a simple method for determin- Similarly, Suekane et al. (2008) determined the residual ing the storage capacity in closed and semi-closed aquifers. and solubility capacities available in Japanese aquifers. By The main idea lies in the premise that injected C O will lead improving on the flaws of the conventional analytical tech- to a pressure increase in the formation. This will, in turn, niques for storage estimation, Ding et al. (2018) proposed lead to a displacement of native brine which can either be new analytical methodologies for the determination of solu- stored in the expanded pore space due to compression of the bility and mineral trapping in aquifers and depleted oil reser- rocks (closed systems) or the pore space in the seals overly- voirs. Their model was applied to the HB oil field in China, ing the formation (semi-closed systems). and estimates were compared to a similar methodology by Zhou et al. (2008) showed the derivations for closed sys- Xu et al. (2004) with slight discrepancies observed. They, tems by using the given in Eqs. 26 and 27 below. however, argued that their model would be superior as, in addition to the model’s ability to determine storage capacity V =  +  V ΔP (26) by solubility trapping, the model could also determine the CO p w pore max annual storage capacities by mineral trapping. M =  +  V ΔP (27) CO p w pore max CO 2 2 6 Measurement, monitoring and verification For semi-closed systems the following equation is techniques during  CO storage suggested: Monitoring the movement of the plume for leakages is criti- V (t )=  +  ΔP (t )V CO 1 p w max max pore cal in the post-injection phase of storage. Containment of + 0.5  +  ΔP (t )V ps w max max s (28)the CO is achieved if proper monitoring is performed as max 2Ak ΔP (t) s max leakages could be detected early, thus ensuring that the + dt 0 environment and groundwater are not at risk from released w s gases. Furthermore, monitoring could be employed in the where  is the compressibility, A is the area, k is the perme- validation of simulation predictions by tracking the pres- ability, subscripts s, p, w refer to the seal, pore and water, sure buildup in the formation (Bourne et al. 2014). Mass respectively, β refers to the compressibility of the rock ps balance verifications are also an important reason for carry - from pore to seals, V is the volume, µ is the water viscos- ing out monitoring studies. Injected C O volumes must be ity, B stands for thickness of the top and bottom seals, t tracked to ensure they are stored in identified zones and in 1 3 1050 Petroleum Science (2019) 16:1028–1063 line with emission quotas specified before the commence- they can easily be detected even at low concentrations, are ment of such projects. Successful verification of simulations highly soluble in C O , are non-toxic and are non-radioactive. via monitoring would provide researchers with greater con- A notable CO injection project which has made use of the fidence in the use of simulation tools. Consequently, a lot of tracer technique for monitoring is the Frio Project (Nance effort is continuously made to develop accurate monitoring et al. 2005). Their monitoring design made use of PFCs as tools. As with the modeling approach, monitoring of CO the chemical tracer to monitor leakages. Fibrous elements can either be classified on a spatial or temporal basis. On a such as capillary absorbent tubes (CATs) were placed on spatial basis, it is monitored based on the area which the CO surface installations in order to adsorb the PFCs. The CATs affects. On this basis, it can be classified into atmospheric were removed on a periodic basis to ascertain the amount monitoring, near-surface monitoring and subsurface moni- of PFCs which had sorbed on the surface of the CATs using toring (which involves the faults, wells, reservoir and seals) thermal desorption and gas chromatograph techniques. Laser (Brown et al. 2009). On a temporal basis, monitoring can systems are remote sensing technologies that make use of be grouped as during the injection phase and post-injection either optical absorption, breakdown spectroscopy or non- phase. For further discussion, we limit ourselves to discuss- linear optics to monitor gas leakages. A laser application ing monitoring on a spatial basis.for CO detection, however, only makes use of the optical absorption technique. In this technique, the laser beams a 6.1 Atmospheric monitoring tools light which has been tuned to the wavelength of the CO on the gas. The scattered light which emanates from the gas As the name implies, these tools ensure that the CO injected after absorption is examined. An issue with this technique into the formations does not leak into the atmosphere above is the accurate determination of the wavelength of C O as it. This monitoring strategy is important due to the concerns the absorption wavelengths of CO must be carefully deter- about leaked CO . Atmospheric monitoring tools are typi- mined without infringing on the absorption wavelengths of cally required to be very sensitive as leakage of CO from water vapor. the formation could be quickly dispersed in the atmosphere, thus making it difficult for other forms of monitoring tools 6.2 Near‑surface monitoring tools to recognize the gas immediately. Atmospheric monitoring tools are placed at the potential leakage sources so as to Usually, the flow of CO at the near-surface consists of increase their detection capability and are especially required bubbles which emanate from faults or near an abandoned to provide confidence in carbon dioxide storage and for car - wellbore. Monitoring of C O at the near-surface is impor- bon accounting verification. The tools used to detect CO tant as it serves as a link between the subsurface and the leakage in the atmosphere are optical sensors, atmospheric atmosphere. Therefore, it can provide information on leaks tracers and eddy covariance (Brown et al. 2009). Other sys- in the subsurface while preventing leaks to the atmosphere tems which can be used in monitoring the atmospheric levels if detected in time, monitoring in this area has been proven of CO include CO detectors, advanced leak detection sys- 2 2 to be less expensive than atmospheric and subsurface moni- tem, laser systems and LIDAR. As the quantity of safe CO toring. Some techniques which can be used for near-surface required to exist in the atmosphere must not exceed certain monitoring could also be used for subsurface monitoring. limits, CO detectors can be applied to sense the existence of Such techniques which could be used for this monitoring excess CO in the atmosphere. Application of CO detectors 2 2 have been summarized in the next subsection. Such tech- might, however, prove to be impractical due to the enormous niques include interferometric synthetic aperture radar number of detectors that would be required to effectively (InSAR), tiltmeters, time-lapse seismic among others. detect the gas. Eddy covariance also known as eddy flux is an important atmospheric monitoring tool used to quan- tify the fluxes of gases between the surface of the earth and 6.3 Subsurface monitoring tools the atmosphere. It has the advantage of being able to cover kilometers of space, thereby providing quick monitoring and The objectives of subsurface monitoring are to track the having a low to moderate cost. Atmospheric tracers are arti- movement of an injected CO plume in a deep geologic for- ficial substances injected into the formation along with the mation; to define the lateral extent and boundaries of the CO in order to observe the leakage of C O early on. They plume; to track associated pressure changes in the reservoir; 2 2 are also used to monitor the flow direction of the CO in the and to demonstrate long-term stability of the CO plume 2 2 formation. Conventional tracers which have been employed (Brown et al. 2009). Numerous monitoring techniques can for monitoring studies are the perfluorocarbons (PFCs) and be employed for the monitoring of CO plume in the sub- sulfur hexafluoride (SF6). Perfluorocarbons (PFCs) are, surface. The choice of monitoring techniques to be used however, preferred to sulfur hexafluoride (SF6) because for subsurface monitoring is dependent on the information 1 3 Petroleum Science (2019) 16:1028–1063 1051 required, costs of monitoring technique and time frame to to observe the extent of geomechanical deformation in the achieve information. subsurface. They are particularly useful in the monitoring of Seismic methods have been employed to evaluate the dis- cap rock deformations. InSAR has been applied for the mon- tribution of faults and the subsurface structures using 3D itoring of surface deformations. It achieves its objectives techniques. In a 4D mode that includes time-lapse data, seis- by making use of two synthetic aperture radars to generate mic methods can also be used to track the movement of the maps. This technique is sensitive to changes in deformations injected plume and gas leakages. Multi-component 3D sur- and has been used to measure millimeter changes in sur- face seismic provides better information when the geology face deformation. Different forms of the InSAR techniques of the formation is non-uniform. Together with time-lapse, include corner reflector Interferometric synthetic aperture multi-component seismic profiling provides valuable infor - radar (CR-InSAR), permanent scatterer interferometric syn- mation on the migration of the injected gas. If cost consid- thetic aperture radar (PS-InSAR) and differential interfero- erations are taken into account, 2D time-lapse seismic moni- metric synthetic aperture radar (D-InSAR). The technique toring could be used to provide data on the injected plume. has been applied for the monitoring of natural occurrences The downside of the 2D methods is in their inability to track such as volcanoes and earthquakes. The ability of the InSAR plume movement in formations with complex geometries. technique to monitor surface deformations has been applied 2D seismic techniques would be more useful where observa- in storage sites for tracking fluid pressure alterations, thus tion wells are available and cross-well seismic technology determining leakages. Recently, it was pioneered as a moni- could be employed. Vertical seismic profile (VSP) has been toring tool at the In Salah storage site in Algeria. employed to provide information on the leakages and the The choice of monitoring tool to be employed on any migration path of CO (El-Kaseeh et al. 2017). Most of the specific storage site is dependent on the nature of the site. conventional seismic methods have been used to determine For example, geophysical monitoring from the surface is leakages and migration path of the C O . In order to quantify dependent on the extent of overburden on the aquifer. There- the injected gas, seismic methods have been employed by fore, in geologically complex scenarios, monitoring of the combining the measurement of the velocity with Gassmann injected plume via this technique would be more cumber- modeling. This method requires that the density of CO at some. In the same vein, information available on a particu- reservoir conditions is known. Determination of this density lar storage site could influence the monitoring technique is not an easy process, and therefore, seismic monitoring chosen. Depleted oil and gas reservoirs which have been tools have been combined with gravimetry. Gravimetry basi- adequately characterized and have been proven to have cally involves using gravity to monitor the in situ changes assured seal integrity would make for easier monitoring of in the density of the injected gas. Results from gravimetric the injected CO plume. monitoring could provide reliable inputs for flow simula- Established commercially known CCS projects have tions. Gravimetric methods, however, possess low sensitiv- employed different monitoring tools. Torp and Gale (2004) ity and require a sizeable amount of C O injected into the provided useful information on the monitoring tools used formation before responses can be picked up. at the Sleipner project in Norway. Repeated seismic data Electromagnetic and electric methods have found impor- were among the many tools used for monitoring (Fig. 7). tant use as monitoring tools. They make use of electrical and The monitoring procedures confirmed some of the estimates electromagnetic responses from the subsurface to determine from reservoir simulation. The injected CO moved upward the changes in saturation. These techniques involve meas- due to buoyancy after the injection and accumulated under uring important electric parameters such as conductivity, the cap rock overlying the formation. Also, it was observed resistivity and employing correlations such as the Archie that solubility trapping would occur faster than mineral expression to relate these parameters to saturations. Differ - trapping. The simulation model for the Sleipner project was ent methods that use these concepts are the magnetotelluric then history-matched with the seismic data results to pro- sounding, electromagnetic resistivity, electrical resistivity vide accurate predictions for the future. However, seismic tomography (ERT), electromagnetic induction tomography monitoring is costly and other monitoring tools such as pres- (EMIT) among others. sure monitoring and observation wells could provide viable Geophysical logs have also been employed for the moni- alternatives. toring of subsurface-injected plumes. They provide useful Ringrose et al. (2013) analyzed the lessons learned from information on well properties and reservoir fluids. Exam- the In Salah Project in Algeria. Among these were the need ples of geophysical logging tools which could be employed for characterization of the overburden and the reservoir include sonic logs, neutron logs and density logs. Coupled prior to injection, constant risk assessments of the identi- with their ability to map saturation, geophysical logging fied storage sites and the significance of flexibility in the tools could also provide information on the onset of cor- design of capture, compression and injection systems. The rosion in the casings of wellbores. Tiltmeters can be used interferometric synthetic aperture radar (InSAR) method for 1 3 1052 Petroleum Science (2019) 16:1028–1063 Fig. 7 Seismic survey results for the Sleipner project (Torp and Gale 2004) storage monitoring was pioneered in this project. InSAR was such as failure modes (risk evaluation), likelihood and con- able to provide information on millimeter changes in ground sequences of failure must be answered when performing risk surface elevation; it was also able to give insights into the assessments for projects. Risks and challenges involved in geomechanical response to CO injection. Arts et al. (2004) CO storage are highlighted below. 2 2 made use of time-lapse seismic studies to monitor plume movement in the Utsira Formation. Notably, they were able 7.1 Leakage to demonstrate that the impact of the movement of CO on seismic measurements was considerable and thus seismic The primary and most important risk factor is leakage. Most could be used as a suitable monitoring tool during the life- modeling and monitoring studies conducted in the devel- cycle of a storage project. A summary of monitoring tools opment, implementation and monitoring phases of carbon used at select CCS projects is provided in Table 6. dioxide storage are done primarily to avoid leakage of the On a broader scale, monitoring is usually quantified as gas into the atmosphere, groundwater aquifers, shallow soil monitoring, verification and accounting (MVA) to include zones and overlying resource bearing strata and to ensure mass balance verifications and accounting for operators. secure containment of gas. The leakage of carbon dioxide Interested readers are referred to Plasynski et al. (2011) for could be as a result of the following: details on MVA strategies for different projects. 1. Aquifer over-pressurization: Aquifer over-pressurization could lead to cracks in the cap rock overlying it and in the reactivation of faults and thus should be avoided. 7 Risks and challenges in  CO storage The risk of aquifer over-pressurization is much less in depleted hydrocarbon dioxide reservoirs due to reduced The high dependency of world energy on coal-fired power pressure before the injection started. Saline aquifers plants makes carbon capture and storage a very important pose more risk from aquifer over-pressurization because technology for the mitigation of global warming. Therefore, the pressure of the aquifer begins from the initial pres- it represents the only viable option in the short term to limit sure, thereby leading to quick buildup of pressure when global warming effects and must be pursued vigorously. injection commences. Vilarrasa et al. (2010) performed However, just as with most technologies, carbon dioxide numerical simulations to ascertain the risk of over- storage comes with its own risks and challenges which must pressure during injection. The authors employed an be properly catered for before venturing into it. Questions axisymmetric horizontal aquifer–cap rock system cou- 1 3 Petroleum Science (2019) 16:1028–1063 1053 Table 6 Monitoring techniques used in field-scale projects Field project Category Monitoring techniques Select literature Sleipner Saline aquifer Time-lapse gravity; micro-seismic; time-lapse Arts et al. (2004) and Cavanagh (2013) seismic Ketzin Saline aquifer Pulsed-neutron gamma logs; 4D seismic; 3D repeat Ivanova et al. (2012) and Kiessling et al. (2010) seismic survey; electrical resistivity tomography (ERT); cross-well seismic; geophysical monitor- ing Weyburn EOR Passive seismic; 4D time-lapse; vertical seismic Bellefleur et al. (2003), Preston et al. (2005) and profile (VSP); tracer injection; geochemical White (2011) sampling analysis; production data analysis; cross- well seismic Otway Depleted gas field Hydrodynamic sampling; high-resolution travel Boreham et al. (2011), Etheridge et al. (2011) and time, flask sampling; 3D surface seismic; head- Urosevic et al. (2011) space gas sampling; logging pressure/temperature; flux tower; surface soil gas; downhole fluid sam- pling; CO sniffers; VSP; micro-seismic; borehole seismic; groundwater chemistry Cranfield Saline aquifer Cross-well seismic tomography; 4D seismic, elec- Ajo-Franklin et al. (2013), Alfi et al. (2015), Carrigan trical resistance tomographic monitoring; vertical et al. (2013) and Kim and Hosseini (2014) seismic profile; above zone monitoring interval; tracers In Salah Depleted gas field Well log data; 3D seismic baseline survey; 4D Mathieson et al. (2010), Onuma and Ohkawa (2009) seismic monitoring; groundwater monitoring and Ringrose et al. (2009) wells; micro-seismic monitoring; satellite InSAR monitoring; tracers in CO injection wells; core analysis (storage unit); soil and surface gas sam- pling; core analysis (cap rock unit) Rumaitha EOR Cross-well seismic; DTS (distributed temperature Al-Hajeri et al. (2010) and Figuera et al. (2014, 2016) Zone-B, Abu sensing in observation well, 41 meters from injec- Dhabi tor; permanent multi-phase flow meter (MPFM), logging tools, sponge coring near injector well; production data analysis pled with hydromechanics. Their results showed that of the aquifer encounters pressures capable of breaking the highest risk of over-pressures and fault reactivation the seal. were at the beginning of injection where fluid pressures 2. Abandoned wells: Another significant leakage pathway rise. Lindeberg et al. (2009) also noted the importance is abandoned wells; this leakage pathway is more plau- of the consideration of injection pressures in the pre- sible in a depleted hydrocarbon reservoir which has been vention of leakages through the cap rock. An engineer- used previously for the commercial production of hydro- ing strategy has been proposed by Eke et al. (2011) to carbon dioxides than in saline aquifers. This is because minimize the leakage of CO . In their paper, they argued depleted hydrocarbon dioxide reservoirs possess wells that surface mixing of C O with brine prior to injec- whose structural integrity might have degraded over tion could enhance the dissolution trapping mechanism. time. Degradation of wells could be as a result of cas- Subsequently, this would lead to a denser CO which is ing corrosion and reactions of the minerals with plug-in saturated with brine being injected into the reservoir. By materials or reservoir fluids which compromise integrity. implication, the strong buoyancy drive, typically expe- Human errors in the design of wells such as loose plugs rienced in aquifers, is minimized and the risk of C O could also create pathways for leakage of gases. Sev- leaking via prolonged contact of the CO with the seal eral studies have been conducted to assess the impact of is curtailed. It is therefore important before commenc- leakages through wells (Carey 2018; Kopp et al. 2010). ing any storage activity to perform a geomechanical 3. Faults and fractures: It is essential while performing site analysis in order to understand the fracturing pressure selection and characterization to ensure that there are no of the cap rock and thus avoid over-pressurization of the transmissive faults and fractures in the identified forma- aquifer. Large areal extents of a proposed aquifer could tion. Additionally, during the injection of CO , care must also mean that pressure propagates much faster, ensuring be taken to ensure that inactive faults are not activated that it takes a significant amount of time before the seal due to the high aquifer pressures. Fractures could also 1 3 1054 Petroleum Science (2019) 16:1028–1063 be developed in the cap rock if the temperature of the primary and secondary entrapments, (2) seismic activi- injected CO is much lower than the in situ temperature ties which could lead to the natural release of CO into the 2 2 in the aquifer. In the Abu Dhabi Rumaitha Zone-B pro- atmosphere, (3) fractures and faults which could lead to the ject, the CO is heated at the surface prior to injection, rapid release of C O , (4) abandoned and structurally weak 2 2 to ensure thermal induced fractures are not created in the wells which possess the ability to release large amounts of reservoir. CO back to the atmosphere and (5) release of C O that 2 2 rarely occurs through eruptive processes. 7.2 Induced seismicity Another postulated risk associated with CO storage is that 8 Conclusions of induced seismicity. The risk, however, has been proven to be negligible in field-scale projects that have been car - The risk of global warming is no longer hearsay. Several ried out due to the relatively small size of the projects and countries have accepted that our world is facing the risk of low injection rates. Nicol et al. (2011) noted that induced an endangered atmosphere and this must be addressed. The seismicity could lead to earthquakes that exceed magnitudes problem is not just a scientific one but also affects other of M6 and have the potential to impact on the containment, spheres of human endeavor. In this review, we provide the infrastructure and public perceptions of safety at C O stor- reader with the state of the art on carbon dioxide storage age sites. The possibility of the occurrence of a seismic science and technology. From a scientific viewpoint, the event would be higher if faults are present. This reiterates understanding of the processes involved in the process has the need for proper site characterization and identification of been greatly enhanced over the years with concrete informa- faults and fractures to avoid their reactivation and the pos- tion available on the fate of the injected C O before, during sible consequences of this reactivation (Oldenburg 2014). and after the injection phases. However, there are certain issues which we believe still need to be addressed before 7.3 Economic considerations the science can be considered full-fledged. The modeling procedures involved in carbon dioxide storage is multi-scale Carbon dioxide capture and carbon dioxide storage are two in both the temporal and spatial scales; we believe that for technologies that go hand-in-hand, hence the popular acro- the physics of the different level scales to be effectively nym CCS. The success of one process is dependent to a understood, the problem needs to be approached using multi- large extent on the success of the other. As such, it is nec- scale formulations. This would require the development of essary to state that the deployment of carbon dioxide stor- advanced numerical algorithms which are very robust and age projects would be greatly enhanced if carbon dioxide computationally efficient for best results. Improvements capture processes are also successful. The key economic in monitoring tools used at commercial CCS sites would issue associated with carbon dioxide capture processes is also go a long way toward validating scientific models and the high cost of the capture of CO from stationary power simulation predictions. An improvement in the capability of plants. In fact, most successful commercial deployment of monitoring and modeling tools implies that the risk of the carbon dioxide storage projects has pursued the option of leakage of C O is greatly reduced. It is obvious that these separating CO from produced gas rather than capturing could not be accomplished if the number of commercial CO from coal plants. This represents a cheaper option for CCS sites does not greatly increase. Governments would the companies involved. As with most burgeoning technol- need to establish and enforce policies such as carbon dioxide ogy, there is always a higher cost for companies which make pricing and taxation which would compel companies that the first step toward developing the technology before the would otherwise have considered the cheaper option of the technology improves and costs are reduced. For this reason, emission of C O directly into the atmosphere into consider- there is a reticence among companies to avoid making the ing CCS. first move. This disposition can be quelled by government In summary, a successful carbon dioxide storage project action in subsidizing the costs involved for the early movers, would involve accurate site selection, characterization (stor- thereby encouraging more participation. age capacity estimation, plume modeling) and monitoring Lewicki et al. (2007) made use of leakages of CO from to avoid the risks of leakages through seals, faults and aban- natural and industrial formations to analyze the features, doned wells. The site characterization would be successful events and processes (FEPs) of the leakages from both through the use of modeling and simulation tools whose natural and man-made sources. A total of 12 natural and 4 accuracy would be greatly enhanced through measurement, industrial analogues were looked into in their comparisons. monitoring and verification during the post-injection phase. They concluded at 5 FEPs which could lead to the release Carbon dioxide storage is a technology that has come to stay of stored C O in aquifers: (1) accumulation of C O beneath with the advantage of allowing the continued use of fossil 2 2 1 3 Petroleum Science (2019) 16:1028–1063 1055 integrating constant-concentration laboratory assay data with fuels while still saving our environment from the risks of variable-concentration field exposure. Environ Model Assess. global warming and therefore must be embraced by all. 1997;2(4):333–43. https ://doi.org/10.1023/a:10190 29931 755. Aydin G, Karakurt I, Aydiner K. Evaluation of geologic storage options Acknowledgements The authors gratefully acknowledge the research of CO : applicability, cost, storage capacity and safety. Energy support provided by the Department of Petroleum Engineering, Khalifa Policy. 2010;38(9):5072–80. https ://doi.or g/10.1016/j.en pol University of Science and Technology, Sas Al Nakhl Campus, Abu .2010.04.035. Dhabi, UAE. The corresponding author (AB) is thankful to the Drill- Bachu S, Adams JJ. Sequestration of CO in geological media in ing, Cementing, and Stimulation Research Center, School of Petroleum response to climate change: capacity of deep saline aqui- Technology, Pandit Deendayal Petroleum University, Raisan, Gandhi- fers to sequester C O in solution. Energy Convers Manag. nagar, Gujarat-382007, India, for supporting his research. Thanks are 2003;44(20):3151–75. https ://doi.or g/10.1016/S0196 also extended to other individuals who were, directly and indirectly, -8904(03)00101 -8. related to this work. Bachu S, Bonijoly D, Bradshaw J, Burruss R, Holloway S, Christensen NP, et al. CO storage capacity estimation: methodology and Open Access This article is distributed under the terms of the Crea- gaps. Int J Greenh Gas Control. 2007a;1(4):430–43. https ://doi. tive Commons Attribution 4.0 International License (http://creat iveco org/10.1016/S1750 -5836(07)00086 -2. mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Bachu S, Bonijoly D, Bradshaw J, Buruss R Christensen N, Holloway tion, and reproduction in any medium, provided you give appropriate S et al. Phase II, final report from the task force for review and credit to the original author(s) and the source, provide a link to the identification of standards for CO storage capacity estimation. Creative Commons license, and indicate if changes were made. In: Carbon sequestration leadership forum, Washington, United States; 2007b. p. 43. Basirat F, Fagerlund F, Denchik N, Pezard PA, Niemi A. Numeri- cal modelling of C O injection at small-scale field experi- References mental site in Maguelone, France. Int J Greenh Gas Control. 2016;54:200–10. https ://doi.org/10.1016/j.ijggc .2016.09.006. Adams EE, Caulfield JA, Herzog HJ, Auerbach DI. Impacts of reduced Bellefleur G, Adam L, White D, Mattocks B, Davis T. Seismic imaging pH from ocean C O disposal: sensitivity of zooplankton mortal- and anisotropy analysis of 9C 3D-VSP data at Weyburn Field, ity to model parameters. Waste Manag. 1998a;17(5–6):375–80. Saskatchewan, Canada. In: SEG technical program expanded https ://doi.org/10.1016/s0956 -053x(97)10038 -1. abstracts, Society of Exploration Geophysicists; 2003, p. Adams EE, Caulfield JA, Herzog HJ, Auerbach DI. Impacts of reduced 1326–29. pH from ocean C O disposal: sensitivity of zooplankton mortal- Belhaj H, Bera A. A brief review of mechanisms for carbon dioxide ity to model parameters. Waste Manag. 1998b;17(5–6):375–80. sequestration into aquifer reservoirs. Int J Pet Eng. 2017;3(1):49– https ://doi.org/10.1016/s0956 -053x(97)10038 -1. 66. https ://doi.org/10.1504/IJPE.2017.08899 6. Ajo-Franklin JB, Peterson J, Doetsch J, Daley TM. High-resolu- Bennion DB, Bachu S. The impact of interfacial tension and pore size tion characterization of a CO plume using crosswell seismic distribution/capillary pressure character on C O relative perme- 2 2 tomography: Cranfield MS, USA. Int J Greenh Gas Control. ability at reservoir conditions in C O -brine systems. In: SPE/ 2013;18:497–509. https ://doi.org/10.1016/j.ijggc .2012.12.018. DOE symposium on improved oil recovery, Society of Petroleum Al-Hajeri SK, Negahban S, Bin-dhaaer GS, Al-Basry AH. Design and Engineers; 2006. https ://doi.org/10.2118/99325 -MS. implementation of the first CO -EOR pilot in Abu Dhabi, UAE. Benson SM, Cole DR. CO sequestration in deep sedimentary forma- 2 2 In: SPE EOR conference at oil and gas, West Asia; 2010. https tions. Elements. 2008;4(5):325–31. https://doi.or g/10.2113/gsele ://doi.org/10.2118/12960 9-MS.ments .4.5.325. Alfi M, Hosseini SA, Alfi M, Shakiba M. Effectiveness of 4D seismic Benson SM, Hingerl F, Li B, Pini R, Tchelepi H, Zuo L. Investigations data to monitor CO plume in Cranfield CO -EOR project. In: in geologic carbon sequestration: multiphase flow of CO and 2 2 2 Carbon management technology conference; 2015. https ://doi. water in reservoir rocks; 2013. org/10.7122/43955 9-MS. Bestehorn M, Firoozabadi A. Effect of fluctuations on the onset Ambrose WA, Lakshminarasimhan S, Holtz MH, Núñez-López V, of density-driven convection in porous media. Phys Fluids. Hovorka SD, Duncan I. Geologic factors controlling C O stor- 2012;24(11):114102. https ://doi.org/10.1063/1.47674 67. age capacity and permanence: case studies based on experience Bethke C, Yeakel S. Geochemist’s workbench: release 8.0 reference with heterogeneity in oil and gas reservoirs applied to CO stor- manual; 2009. age. Environ Geol. 2017;54(8):1619–33. https://doi.or g/10.1007/ Bonneville A, Gilmore T, Sullivan C, Vermeul V, Kelley M, White S, s0025 4-007-0940-2. et al. Evaluating the suitability for CO storage at the Future- Aminu MD, Nabavi SA, Rochelle CA, Manovic V. A review of develop- Gen 2.0 Site, Morgan County, Illinois, USA. Energy Procedia. ments in carbon dioxide storage. Appl Energy. 2017;208:1389– 2013;37:6125–32. https://doi.or g/10.1016/j.egypro.2013.06.541 . 419. https ://doi.org/10.1016/j.apene rgy.2017.09.015. Boreham C, Underschultz J, Stalker L, Kirste D, Freifeld B, Jenkins Ammer J, Brummert A. Miscible applied simulation techniques for C, et al. Monitoring of CO storage in a depleted natural gas energy recovery-version 2.0; 1991. reservoir: gas geochemistry from the C O CRC Otway Project, Arts R, Eiken O, Chadwick A, Zweigel P, van der Meer L, Zinszner B. Australia. Int J Greenh Gas Control. 2011;5(4):1039–54. https:// Monitoring of C O injected at Sleipner using time-lapse seismic doi.org/10.1016/j.ijggc .2011.03.011. data. Energy. 2004;29(9–10):1383–92. https://doi.or g/10.1016/j. Bourne S, Crouch S, Smith M. A risk-based framework for measure- energ y.2004.03.072. ment, monitoring and verification of the Quest CCS Project, Assteerawatt A, Bastian P, Bielinski A, Breiting T, Class H, Ebigbo Alberta, Canada. Int J Greenh Gas Control. 2014;26:109–26. A, et  al. MUFTE-UG: structure, applications and numerical https ://doi.org/10.1016/j.ijggc .2014.04.026. methods. Newsl Int Groundw Model Centre Colo Sch Mines. Braathen A, Bælum K, Christiansen HH, Dahl T, Eiken O, Elvebakk H, 2005;23(2):10. et al. The Longyearbyen CO Lab of Svalbard, Norway—initial Auerbach DI, Caulfield JA, Adams EE, Herzog HJ. Impacts of ocean assessment of the geological conditions for C O sequestration. CO disposal on marine life. I. A toxicological assessment Nor J Geol. 2012;92(4):353–76. 1 3 1056 Petroleum Science (2019) 16:1028–1063 Bromhal GS, Neal Sams W, Jikich S, Ertekin T, Smith DH. Simu- Cole S, Itani S. The Alberta carbon trunk line and the benefits of CO . lation of CO sequestration in coal beds: the effects of sorp- Energy Procedia. 2013;37:6133–9. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . tion isotherms. Chem Geol. 2005;217(3–4):201–11. https ://doi.egypr o.2013.06.542. org/10.1016/j.chemg eo.2004.12.021. Coninck HD, Loos M, Metz B, Davidson O, Meyer L. IPCC special Broseta D, Tonnet N, Shah V. Are rocks still water-wet in the presence report on carbon dioxide capture and storage. Intergovernmental of dense C O or H S? Geofluids. 2012;12(4):280–94. https://doi. Panel on Climate Change; 2005. 2 2 org/10.1111/j.1468-8123.2012.00369 .x. Delshad M, Kong X, Wheeler MF. On interplay of capillary, gravity, Brown B, Carr T, Vikara D. Monitoring, verification, and accounting and viscous forces on brine/CO relative permeability in a com- of CO stored in deep geologic formations. US Department of positional and parallel simulation framework. In: SPE reservoir Energy National Energy Technology Laboratory; 2009. simulation symposium, Society of Petroleum Engineers; 2011. Brydie J, Jones D, Jones JP, Perkins E, Rock L, Taylor E. Assessment https ://doi.org/10.2118/14214 6-MS. of baseline groundwater physical and geochemical properties for DePaolo DJ, Cole DR, Navrotsky A, Bourg IC. Geochemistry of geo- the Quest carbon capture and storage project, Alberta, Canada. logic CO sequestration, 77. Walter de Gruyter GmbH & Co Energy Procedia. 2014;63:4010–8. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . KG; 2019. egypr o.2014.11.431. Ding S, Xi Y, Jiang H, Liu G. CO storage capacity estimation Burnside NM, Naylor M. Review and implications of relative perme- in oil reservoirs by solubility and mineral trapping. Appl ability of C O /brine systems and residual trapping of C O . Int Geochem. 2018;89:121–8. https ://doi.or g/10.1016/j.apg eo 2 2 J Greenh Gas Control. 2014;23:1–11. https ://doi.org/10.1016/j. chem.2017.12.002. ijggc .2014.01.013. DOE-NETL. Carbon storage Atlas Fifth Edition. U.S. Department of Busch A, Gensterblum Y. CBM and CO -ECBM related sorption pro- Energy Office of Fossil Energy, National Energy Technology cesses in coal: a review. Int J Coal Geol. 2011;7(2):49–71. https Laboratory; 2015. ://doi.org/10.1016/j.coal.2011.04.011. DOE U. Carbon sequestration ATLAS of the United States and Canada. Carey JW. Probability distributions for effective permeability of Office of Fossil Energy, National Energy Technology Laboratory, potentially leaking wells at C O sequestration sites. Los Ala- Morgantown, WV, 90p; 2007. mos National Lab (LANL), Los Alamos, NM (United States); Durucan S, Shi JQ, Korre A. Numerical modeling and prediction of 2018. abandoned mine methane recovery: field application at the Saar Carrigan CR, Yang X, LaBrecque DJ, Larsen D, Freeman D, Ram- Coalfield, Germany. Geol Belg. 2004;7(3–4):207–13. irez AL, et al. Electrical resistance tomographic monitoring Ebigbo A, Bielinski A, Kopp A, Class H, Helmig R. Numerical mod- of CO movement in deep geologic reservoirs. Int J Greenh eling of CO sequestration with MUFTE-UG. CMWR XVI, 2 2 Gas Control. 2013;18:401–8. https ://doi.org/10.1016/j.ijggc Copenhagen; 2006. .2013.04.016. Eke PE, Naylor M, Curtis A, Haszeldine S. CO leakage prevention Cavanagh A. Benchmark calibration and prediction of the Sleipner C O technologies. In: Offshore Europe, Society of Petroleum Engi- plume from 2006 to 2012. Energy Procedia. 2013;37:3529–45. neers; 2011. https ://doi.org/10.2118/14526 3-MS. https ://doi.org/10.1016/j.egypr o.2013.06.246. El-Kaseeh G, Will R, Balch R, Grigg R. Multi-scale seismic measure- Cavanagh A, Ringrose P. In Salah high-resolution heterogeneous simu- ments for C O Monitoring in an EOR/CCUS project. Energy lations of CO storage. Geoscience. 2010;5:53–63. Procedia. 2017;114:3656–70. https ://doi.or g/10.1016/j.egypr Cavanagh A, Ringrose P. Simulation of CO distribution at the In o.2017.03.1497. Salah storage site using high-resolution field-scale models. Ennis-King J, Paterson L. Role of convective mixing in the long-term Energy Procedia. 2011;4:3730–7. https://doi.or g/10.1016/j.egypr storage of carbon dioxide in deep saline formations. In: SPE o.2011.02.306. annual technical conference and exhibition, Society of Petroleum Cavanagh AJ, Haszeldine RS. The Sleipner storage site: capillary Engineers; 2003. https ://doi.org/10.2118/84344 -MS. flow modeling of a layered CO plume requires fractured shale Estublier A, Lackner AS. Long-term simulation of the Snøhvit barriers within the Utsira Formation. Int J Greenh Gas Control. CO storage. Energy Procedia. 2009;1(1):3221–8. https ://doi. 2014;21:101–12. https ://doi.org/10.1016/j.ijggc .2013.11.017.org/10.1016/j.egypr o.2009.02.106. Celia MA, Bachu S, Nordbotten JM, Bandilla KW. Status of CO Etheridge D, Luhar A, Loh Z, Leuning R, Spencer D, Steele P, et al. storage in deep saline aquifers with emphasis on modeling Atmospheric monitoring of the CO2CRC Otway Project and approaches and practical simulations. Water Resour Res. lessons for large scale C O storage projects. Energy Procedia. 2015;51(9):6846–92. https ://doi.org/10.1002/2015w r0176 09. 2011;4:3666–75. https ://doi.org/10.1016/j.egypr o.2011.02.298. Chadwick RA, Zweigel P, Gregersen U, Kirby GA, Holloway S, Fan Y. Development of CO sequestration modeling capabilities in Johannessen PN. Geological reservoir characterization of a CO Stanford general purpose research simulator. MS report, Depart- storage site: The Utsira Sand, Sleipner, Northern North Sea. ment of Petroleum Engineering, Stanford University, California; Energy. 2004;29(9–10):1371–81. https: //doi.org/10.1016/j.energ 2006. y.2004.03.071. Fang Z. Mitigate drilling risks in highly depleted reservoir fields by Chiquet P, Broseta D, Thibeau S. Wettability alteration of caprock intelligently planning well trajectory for a C O sequestration pro- minerals by carbon dioxide. Geofluids. 2007;7(2):112–22. https ject. In: Brasil offshore, Society of Petroleum Engineers; 2011. ://doi.org/10.1111/j.1468-8123.2007.00168 .x.https ://doi.org/10.2118/14312 9-MS. Christiansen AC. Climate policy and dynamic efficiency gains A case Fang Z, Khaksar A. Geomechanical issues and solutions for CO study on Norwegian C O -taxes and technological innovation in sequestration in depleted hydrocarbon sandstone reservoirs. the petroleum sector. Clim Policy. 2001;1(4):499–515. https :// In: 46th US rock mechanics/geomechanics symposium, doi.org/10.3763/cpol.2001.0150. American Rock Mechanics Association; 2012. https ://doi. Class H, Ebigbo A, Helmig R, Dahle HK, Nordbotten JM, Celia MA, org/10.2118/13399 4-PA. et al. A benchmark study on problems related to CO storage in Farajzadeh R, Zitha PLJ, Bruining J. Enhanced mass transfer of geologic formations. Comput Geosci. 2009;13(4):409–34. https CO into water: experiment and modeling. Ind Eng Chem Res. ://doi.org/10.1007/s1059 6-009-9146-x. 2009;48(13):6423–31. https ://doi.org/10.1021/ie801 521u. Figuera LA, Al-basry AH, Al-Hammadi KE, Al-Yafei A, Sakaria D, Tanakov MY. Complex phased development for C O EOR in oil 1 3 Petroleum Science (2019) 16:1028–1063 1057 carbonate reservoir, Abu Dhabi Onshore. In: Abu Dhabi interna- Grobe M, Pashin JC, Dodge RL. Carbon dioxide sequestration in geo- tional petroleum exhibition and conference, Society of Petroleum logical media: state of the science. In: AAPG studies in geol- Engineers; 2014. https ://doi.org/10.2118/17196 7-MS. ogy. American Association of Petroleum Geologists, Tulsa, OK; Figuera L, Al Hammadi K, Tanakov M. Case study of C O injection to 2009. p