Hydrogeological and numerical analysis of CO2 disposal in deep aquifers in the Alberta sedimentary basin

Pergamon Energy Convers. MgmtVol. 37, l~Ios 6-8, pp. 1167-1174, 1996 Copyright © 1996 Elsevier Science Ltd 0196-8904(95)00315-0 Printed in Great Britain. All rights reserved 0196-8904/96 $15.00 + 0.00 HYDROGEOLOGICAL AND NUMERICAL ANALYSIS OF CO 2 DISPOSAL IN DEEP AQUIFERS IN THE ALBERTA SEDIMENTARY BASIN DAVID H.-S. LAW and STEFAN BACHU" Alberta Research Council, P.O.Box 8330, Edmonton, AB, Canada, T6H-5X2; Abstract--For landlocked large sources of CO,, the best approaches for reducing CO2 emissions into the atmosphere are its utilization and deep disposal into deep sedimentary aquifers or depleted oil and gas reservoirs. A number of coal-based power plants (total capacity of more than 4000 MW) are located near Lake Wabamun in central Alberta, Canada. A hydrogeological study of the sedimentary succession at the site was undertaken to identify and select aquifers which meet various requirements for CO2 disposal, particularly with respect to depth and confinement. The multi-phase, multi-component numerical model STARS was used to study the ability of the selected aquifers to accept and retain for long periods of time large quantities of CO 2 injected in a supereritieal state. The CO 2 injeetivity of the selected aquifers was examined for a whole series of parameters, including aquifer depth and thickness, rock and formation water properties, and injection characteristics. The numerical simulations indicate that even generally low-permeability aquifers can accept and retain large quantities of CO,, showing that injection of CO2 in a supereritieal state into deep aquifers in sedimentary basins is a viable option and may be the best short-to-medium term solution for reducing CO2 emissions into the atmosphere. The CO2 injeetivity is enhanced by the existence of 'sweet' zones of high permeability. CO, disposal Hydrogeological Aquifers Sedimentary basins Numerical simulations 1. INTRODUCTION Becanse of its climate-warming effect, CO 2 utilization and disposal are essential for reducing greenhouse effects in the short and medium terms until new policies and technologies for limiting CO 2 generation and for its utilization are implemented [1]. Pre- and post-combustion capture from flue gases is applied in the chemical and energy industries; however, it reduces the conversion efficiency and increases electricity costs [1,2]. Use of CO 2 by a variety of industries overall has a negligible effect [1,2], beside the fact that this use delays the CO 2 release to the atmosphere only by a short period of time. In many cases, attempts to capture and transport CO 2 ultimately lead to more CO 2 production. Thus, long-term effective CO 2 disposal is essential for reducing greenhouse effects in the near and medium future. Deep oceans may provide an almost limitless repository for CO2; however, the residence time is reduced to several hundred years [3], the dynamics of the processes involved are not well understood [1,2], and the transport and injection of CO 2 into oceans is quite expensive [2]. Disposal and immobilization of CO 2 in geological formations can be achieved by chemical reactions between CO 2 and rock minerals [4,5], but special rock mineralogy is required, and the immobilized amounts vary with rock volume, porosity and surface [5]. On land, CO2 earl be injected into depleted oil and gas reservoirs, but they have limited capacity [2,6], the CO 2 is rather quickly produced back to the surface [7], and practical problems arise [7,8,9], the most important being proximity to power plants. Carbon dioxide injection into deep aquifers in sedimentary basins represents a huge potential for CO, disposal because of the large volume of pore space available [10] and long-term (10L10 G years) hydrodynamic trapping of CO 2 [11]. The conceptual model of long-term CO 2 hydrodynamic trapping [11] is based on the fact that by injecting CO 2 into deep, * currently with the Alberta Geological Survey, 9945-108 Str., Edmonton, AB, Canada, T5K 2G6 1167 1168 LAW and BACHU: CO2 D ISPOSAL IN DEEP AQUIFERS low-permeability, regional-scale aquifers in sedimentary basins, both the dissolved and immiscible CO 2 will travel with the natural velocity of formation water (of the order of 1-10 cm/yr), once outside the radius of influence of the injection well [11]. However, for increased CO s injectivity and reduced injection pressure, CO s should be disposed of into 'sweet' zones of local (near-well) high permeability [11]. In order to test this conceptual model, a series of numerical simulations were performed for a site near Lake Wabamun in the Alberta basin, where several power plants with a total capacity of more than 4000 MW are located. The results and conclusions derived from these simulations are presented herein. 6( 59 58' 57 o 56 e. 55 o ,40 \ / . Northwest Territories \ 120- 118 ° 1160 114 ° 117 0 dLIoow ALBERTA Depth Dominant Nomenclature (m) Lithology .v ~ 844 ! - - "55° - - - - -_-_-- - -2--- - - -"-- - Lea Park Fun. Wabamun .53 ° . . . . 52 ° 1584 . . . . . . . Upper Mannville ~:~.':.'."':'-::.'.:.'.."-'.".'~:..::..;.1 Lower Mannvil le GlaucOlfiti¢ Ss' I " ' ' ' " ' " . . . . . Banf fFm. 51 ° 1763 [n n • ' " . ' . ' . ' . ' . ' ; n n u nUr tUre . . . . . . F~I ,S ] ' iaWI ;~ IL ' " • • l I " " , ' . ' . ' . . I iUn ln" ' . . . • • • • 50 ~ ~ n , ' '.' '." . ' . ' . " . ' . ' : Wabamun Grp. i 2025 i~ , ' _ i~ '_ ' _ ' _ ' . ' . ' . ' . ' . ' ,-49 ON ~"--'~' L"~on Fm. i o u o ! 14 112 110ew " carbonate sandstone shale Fig. 1. Location of Lake Wabamun in Alberta and lithological delineation in a representative well. 2. AQ~R IDENTIFICATION AND CHARACTERIZATION Major sources of CO 2 emissions in Alberta were previously identified [12] and ranked based on CO 2 quantity and quality. Among them, the power plants at Lake Wabamun (Fig. 1) represent the biggest source. Carrying elsewhere the captured CO 2 for use in enhanced oil recovery or storage in depleted gas reservoirs in the Alberta basin [12] would lead to increased energy costs due to the surface transport system [13]. Disposal of CO~ at the site by injection into deep aquifers seems to provide a more economical and long-term solution. At the site, characterized by an average ground elevation of 750 m LAW and BACHU: CO2 DISPOSAL IN DEEP AQUIFERS 1169 and a geothermal gradient of 30 °C/km [14], Cretaceous siliciclastic rocks overlie predominantly carbonate Carboniferous and Devonian rocks (Fig. 1). A series of conditions have to be met in order to dispose of CO 2 by injection into deep aquifers [6,11,13,15]. The conditions of regional confinement and depth greater than 800 m are met at the site by the Cretaceous Glauconitic Sandstone, the Devonian carbonate Nisku, and other deeper aquifers (Fig. 1). The Glauconitic Sandstone and Nisku aquifers, which are regionally confined [16] by the thick Colorado aquitard system and the thin Exshaw aquitard, respectively (Fig 1), were studied and hydrogeologieally characterized for an area of 900 km 2 around the injection site. The average characteristics of aquifer rocks and water, and of injected CO 2, used in subsequent numerical simulations, are presented in Tables 1 and 2. Table 1. Average characteristics of aquifer properties in Lake Wabamun area Depth Thickness Permeability Pressure Aquifer (m) (m) (pm ~) i Pa Glauconitic Ss 1480 13 0.006 12.4 Nisku 1860 60 0.047 16 Fracture Temperature Salinity Pressure iPa (o C) (g/L) 33.5 50 40 42.1 60 190 The average porosity for both aquifers varies in the 6-12% range, whereas the permeability anisotropy is 1%/1~--0.3. Zones of high permeability (0.1-0.4 lam 2 ) are found in places in both aquifers. A rock compressibility value of 4.5x10 a kPa" was considered for both aquifers. The fracturing pressure was estimated based on a lithostatic gradient of 22.61 MPa/km related to the weight of the overburden. Table 2. Characteristics of aquifer water and injected CO 2 Aquifer Glauconitic Ss Nisku Water Density Viscosity (kg/m ~) (mPa.s) 1030 0.617 1155 0.84 Density Viscosity (kg/m ") 696 713 (mPa.s) Carbon Dioxide Equilibrium Constant Compres- Diffusion sibility Coefficient (10 "= kPa -1) (10 "~ mZ/d) 1.22 0.307 0.89 0.371 0.068 0.060 0.081 0.076 The values for the CO~-water equilibrium constant in Table 2 correspond to injection pressures of 20 and 30 MPa for the Glauconitic Sandstone and Nisku aquifers, respectively. The water and CO 2 density and viscosity were calculated based on salinity, temperature and pressure [17,18,19,20]. 3. NUMERICAL SIMULATIONS A 2-dimensional radial grid was used to simulate CO 2 injection in a single well in either the Glaucontie Sandstone or Nisku aquifer. The grid size was kept constant in the vertical direction and increased exponentially in the horizontal direction from 1 m near the injection well to 667.3 m at the outflow boundary located at 7 km distance from the well. At this boundary, a semi-analytical model was used to calculate the flow into the surrounding 'infinite' aquifer. In some cases, the outflow boundary was increased to 17.2 km to ensure that the CO2-injection simulations are not artificially influenced by the model boundary. Accordingly, the number of gridblocks varied between 132 and 375. The multi-phase, multi-component numerical model STARS [21,22] was used to simulate the isothermal injection of CO~ into the two aquifers. The model allows for the flow of supercritical CO 2 in both immiscible and dissolved phases. The pressure- and temperature-dependent equilibrium-constant values for the CO 2 solubility in the aqueous phase are given in Table 2. For the immiscible phase, the relative permeability curves for the CO2-water system (not known) were considered similar with oil-water systems in the Alberta basin. Capillary pressure effects were neglected in simulations because the CO 2 capillary pressure is several orders of magnitude lower than the injection and aquifer pressures. In the simulations, pure CO s at aquifer temperature was injected at a constant pressure for a duration of 30 years 1170 LAW and BACHU: CO2 DISPOSAL IN DEEP AQUIFERS corresponding to the average life time of a power plant. According to injection regulations in Alberta, the injection pressure was allowed to increase to 90% of the rock fracturing pressure, corresponding to 30 and 38 MPa for the Glauconitic Sandstone and Niskn aquifers, respectively. A 3" pipe diameter for the injection well was considered in simulations. A whole series of numerical simulations were carried out by varying one parameter while the others were kept constant, to provide insight into the effects of porosity, permeability, injection-pressure and other parameters on the CO s injection-rate, amount and distribution in the two selected aquifers In order to test the hydrodynamic trapping model [11], simulations were carried out for CO s injection into homogeneous and heterogeneous aquifers, the latter being characterized by a local 'sweet' zone of high permeability surrounded by a regional-scale zone of low permeability. 4. RESULTS l-Iomggeneous Aouifers The numerical simulations have shown that, depending on aquifer temperature, a significant amount of CO 2 dissolves into the aquifer water (17-22 wt% for the Glauconitie Sandstone, and up to 25 wt% for the deeper Nisku aquifer), and travels within the hydrodynamic system in the aqueous phase. The rest of injected CO2 remains in an immiscible supercritical phase, with the tendency of gravity segregation and override at the top of the aquifer because of lower density and higher mobility of the supercritical CO 2 phase than for the aqueous phase, as shown also by [13]. The CO~ override increases with aquifer thickness. However, the supercritical CO2 density increases and the mobility ratio decreases with depth, such that associated gravity segregation, overriding, and fingering effects become less important, even negligible (see also [13]). Because of different mobility, the advancing CO s front provides a large contact zone between the aquifer fluid and injected CO s (see also [23]). Due to the generally low aquifer permeability, the CO s in either phase propagates less than 5 km away from the injection well after a peri .od of 30 years. Regarding rock properties, it was found that porosity has little effect (Fig. 2a) on the amount of injected CO s (an increase of only 7% in the amount of CO s injected after 30 years, for an increase in porosity from 6% to 12%). For less pore space, the CO2 front advances farther from the injection well. Absolute rock permeability has the most important effect on CO s injectivity. The injected CO s amount increases almost linearly with aquifer permeability, regardless of porosity (Fig. 2a). Due to higher injection rates at constant pressure, aquifers with higher permeability pressurize faster than aquifers with lower permeability. The effect of injection pressure is also quite significant. For example, for the same aquifer characteristics, approximately 50% more CO s could be injected into the thin Glauconitic Sandstone aquifer when the injection pressure is increased by 20 % from 25 to 30 MPa (Fig. 2a). The numerical results were generalized using regression analysis and a simple steady-state radial outflow model [24], leading to the following relation for CO 2 injectivity as a function of its mobility: Q/[D x (Pt" Pa)] -- 0.000538 x p x (kkr/la ) / In(re/rw) (1) where the first term represents CO 2 injectivity (mass injection rate Q in t/d per aquifer unit-thickness for a unit pressure difference between the injection and aquifer pressures Pi and Pl, in MPa, respectively), and the term (ki~/la) represents CO 2 mobility. In relation (1), D is aquifer thickness (m), k is absolute permeability (10 "1~ m2), k r is CO s relative permeability (1 for 100% CO s injection), p and la are CO~ density (kg/m s) and dynamic viscosity (mPa.s), respectively, and r w and r e are the radii of the injection well and of injection influence, respectively. For an anisotropic aquifer, k=(k h x k ) °'s. The above relation was obtained assuming well completion for the entire aquifer thickness. If constant values are assumed for CO 2 properties and the radius of injection influence, L, then a log-log plot of CO s injectivity versus mobility should have a slope of one. The results of the numerical simulations of injecting CO s into the Glauconitic Sandstone and Nisku aquifers fit such a straight line ('Fig. 2b), leading to the following relation for predicting the CO s injection rate into homogeneous aquifers: Q=0.0208 x (!% x kv)°~ x D x (Pi" P. )/It (2) In je ct ed CO 2 (t on ne s) 10 9 , , i i , ,I Sl In je ct io n R at e (t /d ay ) II- 5 00 00 l0 8 10 7 a) 1 0 6 1 5 o 0 0 ~ .- 5 oo o 11 -2 00 0 @ " 5 0O q - 2 O O I 0. 00 1 I I I t I II I i / I I ~ I l II .1 . / / N is ku / / / /~ p = 3 8 M Pa ,g' / ,~ p = 30 M Pa / /. p = 2 5 M Pa / / / // / ~ // /G la u c o n it ic g / /' S a n d st o n e /4 ¢ / / / / / / / / if' l/ C O 2 In je ct iv it y (f ff l/m /M Pa ) lO 0 b) 10 I I I il ll l I I I I I II I1 ~ I I I II II I O I I _1 I II II / O p en sy m b ol s: ~ = 0 .0 6 ~ O p en sy m b ol s: 00 = 0 .0 6 C lo se d sy m b ol s: ~ = 0 .1 2 ~ C lo se d sy m b ol s: ~ = 0 .1 2 | | il la i[ I I I Il ll l~ i ! I I t I! o l I t il ll ll [ I I il li li [ I I tl ll il i I I I li t 0. 01 0. 1 1. 0 0. 00 1 0 .0 1 0. 1 1 .0 Pe rm ea b ili ty (g m 2 ) C O 2 M ob ili ty O tm 2 /m Pa 's ) Fi g u re 2 . E ff ec ts o f p er m ea b ili ty an d p or os it y va ri at io n on C O 2 in je ct io n in to as su m ed h om og en eo u s aq u if er s in th e A lb er ta se d im en ta ry b as in : a) ra te an d cu m u la ti ve am ou n t af te r 3 0 ye ar s; b ) i n je ct iv it y as a fu n ct io n o f C O 2 m ob ili ty fo r G la n co n it ic S an d st on e (c ir cl es an d tr ia n g le s) an d N is ku (s q u ar es an d d ia m on d s) aq u if er s. 1 0 .0 g c~ o > In je ct e d C O 2 (t o n n e s) 1 0 S ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' - a ) In je ct io n R a te (t /d a y ) - 1 0 7 lO ' 0 In fi n it e -s iz e sw e e t zo n e O O O 3 o o o lo oO 5 o o - I~ R e g io n a l P e rm e a b il it y Lo ca l P e rm e a b il it y O p e n sy m b o ls : 0 .O O 6 )I ra 2 0 .1 la in 2 C lo se d sy m b o ls : 0 .0 3 0 ~ m 2 P o ro si ty : 1 2 % 0. 5 1. 0 1. 5 2. 0 2. 5 S w e e t- Z o n e R a d iu s ( km ) 2O O " ~ 4 .0 b ) 3 .5 - 3 .0 2 .5 2 .0 1 .5 1 .0 0 .5 3 .0 0 "0 0 .0 E n h a n ce m e n t F a ct o r ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' '" ' ' , N is k u S a n d st o n e N is k u ~ .- - ~ .N is ku .. .. ~ -- . -- - -" " " "" G la u co n it ic S a n d st o n e i . M R e g io n a l P e rm e a b il it y Lo ca l P e rm e a b il it y _ 0 .0 0 6 Ix m 2 O p e n sy m b o l: 0 .1 tt m 2 m _ _ __ 0 .0 3 0 lla n 2 C lo se d sy m b o l: 0 .4 IJJ n 2 I i ' ii I I i I i I , i [I i t I l, i i , I t , I i 0. 5 2. 5 3. 0 1 .0 1 .5 2 .0 R a d iu s o f s w e e t zo n e (k m ) Fi g u re 3 . E ff e ct s o f a h ig h p e rm e a b il it y (" sw e e t" ) zo n e o n C O 2 in je ct io n in to se le ct e d a q u if e rs in th e A lb e rt a se d im e n ta ry b a si n : a ) i n je ct io n ra te a n d cu m u la ti v e a m o u n t a ft e r 3 0 y e a rs o f i n je ct io n in to th e G la u co n it ic S a n d st o n e a q u if e r; a n d b ) e n h a n ce m e n t fa ct o r. t" e ~ > o . f3 "0 o > t' rl t' ~ ~ v o ~ LAW and BACHU: CO2 DISPOSAL IN DEEP AQUIFERS 1173 Relations 1 and 2 take into account relevant aquifer, CO 2 and injection characteristics, and can be applied to CO~ injection under similar conditions anywhere else in the world. Heterogeneous Aauifers To test the conceptual model [11] of enhanced CO L injectivity and low pressure build-up in a ('sweet') zone of high permeability near the injection well in an aquifer of generally low permeability, a series of numerical simulations were run for various cases of high local permeability and their size (radius). Local permeability values varying between 0.1 and 0.4 lxm ~ in areas of 0.5, 1 and 2.1 km around the injection well were considered for both the Glanconitic Sandstone and Nisku aquifers characterized by regional- scale permeability values of 0.006 and 0.03 Ixm 2. The contrast between the local and regional permeability values, and the size of the near-field area of high permeability ('sweet' zone) have both the effect of increasing significantly the CO L injection rate and, consequently, the cumulative injected amount, as illustrated for the Glauconitic Sandstone aquifer in Fig. 3a. A trend of asymptotic increase toward the CO L injectivity into an homogeneous aquifer of corresponding high permeability is evident (Fig. 3a). The results were generalized (Fig. 3b) by calculating an injectivity 'enhancement factor', defined as the CO 2 injectivity into a local zone of high permeability in rapport to the CO L injectivity into the same aquifer characterized by a low homogeneous permeability. As conceptually expected, the enhancement factor of CO L injectivity increases with the contrast between the local and regional permeability values, and with the size of the local 'sweet' zone of high permeability. The aquifer thickness seems to have little influence on the enhancement factor. In the case of the studied aquifers in the central Alberta sedimentary basin, the enhancement factor reaches values of up to 3 (Fig. 3b). Injection rates of up to 2000 t/d/well and 11,870 t/d/well can be achieved for the thin Glauconitic Sandstone and the thick Nisku aquifers, respectively, depending on the injection pressure, The injection- rate for constant injection pressure decays slowly over the 30 years period, result obtained also by [13]. If the injection-pressure is sustained over a period of 30 years, amounts of 22 and 130 million tones of CO 2 Can be realistically injected into these aquifers, respectively. For reference, the output of a 500 MW power plant is approximately 15,000 t/d (165 million tones in 30 years).These results confirm the validity of the conceptual model [11] of hydrodynamic trapping of CO 2 in deep aquifers in sedimentary basins. 5. SUMMARY AND CONCLUSIONS The majority of CO 2 point-emission sources in the Alberta sedimentary basin are located near Lake Wabamun, where power plants with approximately 4000 MW are situated. Considering various pre- requisites for CO 2 utilization and/or disposal, two aquifers were identified for hydrogeological characterization at the site and for numerical simulation of CO 2 injection under various scenarios of aquifer depth, thickness, porosity, permeability, injection pressure and existence ~d size of a 'sweet' zone of local high permeability within the regionally low permeability aquifer. The numerical simulations and subsequent generalizations have shown the following. • It is possible to inject significant amounts of CO 2 into low-permeability homogeneous deep aquifers in sedimentary basins, where the CO 2 will be hydrodynamically trapped for up to millions of years. • Depending on aquifer conditions, a significant part of the injected CO 2 will dissolve in the aqueous phase, while the rest will remain in an immiscible phase and will tend to segregate gravitationally and override to the top of the aquifer, depending on density and mobility contrasts. • The most significant factors in establishing the CO 2 injection rate at constant pressure, and consequently total injected amount, are aquifer permeability and injection pressure (see also [25]), with aquifer thickness playing a moderate role, and aquifer porosity playing a minor role. However, aquifer porosity may play an important role in aquifer capacity for CO L storage. • The existence of a 'sweet' zone of high permeability around the injection well greatly enhances the CO 2 injectivity (rate and cumulative amount) into that aquifer. The enhancement factor depends on the permeability contrast and size of the 'sweet' zone. 1174 LAW and BACHU: CO2 DISPOSAL IN DEEP AQUIFERS The presented results confh'm the concept of hydrodynamic trapping of CO s in deep aquifers in sedimentary basins [11] and, based on generalizations for aquifers in the Alberta sedimentary basin, can be applied elsewhere in the world where similar conditions exist. Acknowledgments: The authors wish to express their gratitude to Transalta Utilities Corporation, Alberta Department of Energy, CANMET and Environment Canada for funding and allowing publication of this research. REFERENCES 1. W. Turkenburg, CO 2 removal; some conclusions. Energy Convers. Mgmt. 33, 819, (1992). 2. H. J. Herzog, E. M. Drake and J. W. Tester, The capture and sequestration of power plant CO r ACS 204th Nat. Meet. ACS Div. Fuel Chem. Preprints, Washington, D.C., 23-28 August, 37, 1068 (1992). 3. C.R. Liro, E. E. Adams and H. J. Herzog, Modeling the release of CO s in the deep ocean. Energy Convers. Mgmt. 33, 667 (1992). 4. H. E. Dunsmore, A geological perspective on global warming and the possibility of carbon dioxide removal as calcium carbonate mineral. Energy Convers. Mgmt. 33, 565 (1992). 5. W. D. Gunter, E. H. Perkins and T. J. McCann, Aquifer disposal of COs-rich gases: reaction design for added capacity. Energy Convers. Mgmt. 34, 941 (1993). 6. S. Holloway and D. Savage, The potential for aquifer disposal of carbon dioxide in the UK,. Energy Convers. Mgmt. 34, 925 (1993). 7. A. C. van der Haarst and A. J. F. M . van Nieuwland, Disposal of carbon dioxide in depleted natural gas reservoirs. Climate and Energy: The Feasibility of Controlling CO s Emissions, pp. 178-188. Kluwer Academic, Dordrecht (1989). 8. K" Blok, C. A. Henddks and W.C. Turkenburg, The role of carbon dioxide removal in the reduction of the greenhouse effect. In Proceedings of an Experts' Seminar on Energy Technologies for Reducing Emissions of Greenghouse Gases, OECD/IEA, Paris, Vol. 2, pp. 135-155 (1989) 9. M. J. van der Burgt, J. Cantle and V. K. Boutkan, Carbon dioxide disposal from coal-based IGCC's in depleted gas fields, Energy Convers. Mgmt. 33, 603 (1992). 10. H. Koide, Y. Tazaki, Y. Nogichi, S. Nakayama, M. Iijima, K. Ito and Y. Shindo, Subterranean containment and long-term storage of carbon dioxide in unused aquifers and in depleted natural gas reservoirs. Energy Convers. Mgmt. 33, 619 (1992). 11. S. Bachu, W. D. Gunter and E.H. Perkins, Aquifer disposal of CO2: Hydrodynamic and mineral trapping, Energy Convers. Mgmt. 35, 269 (1994). 12. R. T. Bailey and M. M. McDonald, CO2 capture and use for EOR in western Canada 1: General overview. Energy Convers. Mgmt. 34, 1145 (1993). 13. L. G. H. van der Meer, The conditions limiting CO 2 storage in aquifers. Energy Convers. Mgmt. 34, 959 (1993). 14. S. Bachu and R. A. Burwash, Regional-scale analysis of the geothermal regime in the Western Canada Sedimentary Basin. Geothermics 20, pp. 387-407 (1991). 15. L. G. H. van der Meer, Investigations regarding the storage of carbon dioxide in aquifers in The Netherlands. Energy Convers. Mgmt. 33, 611 (1992). 16. S. Bachu, Synthesis and model of formation water flow in the Alberta basin. Am. Ass. s. Geol. Bull. 79, (1995). 17. A. M. Rowe and J. C. S. Cbou, Pressure-volume-temperature-concentration relation of aqueous NaC! solutions. J. Chem. Eng. Data 15, pp. 61-66, (1970). 18. J. Kestin, H. E. Khalifa and R. J. Correia, Tables of dynamic and kinematic viscosity of aqueous NaCI solutions in the temperature range 20-150 °C and the pressure range 0.1-35 MPa. J. Phys. & Chem. Ref. Data 10, pp. 71- 87 (1981). 19. M. A. McHugh and V. J. Krnkonis, Supercritical fluid extraction: principles and practice. Butterworths, Boston, p. 10 (1986). 20. S. P. Clark, Handbook of physical constants. The Geological Society of America, New-York, pp. 382-383 (1966). 21.B. Rubin and W. L. Buchanan, A general purpose thermal model. Soc. Petr. Eng. J. 25, pp.202-214 (1985). 22. Computer Modeling Group, STARS Technical Manual Version 4.0. Calgary (1990). 23. T. D. Krom, F. L. Jacobsen and K. H. Ipsen, Aquifer based carbon dioxide disposal in Denmark: capacities, feasibility, implications and state of readiness. Energy Convers. Mgmt. 34, 933 (1993). 24. B. C. Craft and M. F. Hawkins, Applied petroleum engineering. Prentice-Hall, pp. 283-285 (1959). 25. C.A. and K. Blok, Underground storage of carbon dioxide, Energy Convers. Mgmt. 34, 949 (1993).