Differentially dolomitized carbonate strata in the Western Canada Sedimentary Basin (WCSB) are increasingly targeted for carbon capture, utilization, and storage (CCUS), yet few studies have evaluated the petrophysical characteristics of these conventional hydrocarbon reservoirs for this purpose. To address this, this study uses drill-core analysis (sedimentology, diagenesis, pore morphology, and distribution), together with core-plug and production data, to evaluate the properties of five depleted oil and gas fields in the Middle to Upper Devonian Swan Hills Formation, Leduc Formation and Wabamun Group.

The Swan Hills and Leduc formations are composed of reef, shoal, and lagoon deposits that are predominantly fossil-rich (e.g., stromatoporoid-dominated rudstones and boundstones). In contrast, the carbonate-ramp deposits of the Wabamun Group are fossil-poor, consisting instead of variably bioturbated carbonate mudstones, wackestones, and packstones. Replacement dolomitization is variable throughout each stratigraphic unit, but generally occurs within fossil-rich and/or heavily bioturbated intervals. Fracture densities are broadly comparable in limestone and dolostone. Porosity in the Swan Hills and Leduc formations is predominantly moldic and vuggy, occurring where fossils (e.g., stromatoporoids) are partially or fully dissolved. Pore space in the Wabamun Group is mostly restricted to intercrystalline porosity in burrows. In general, burial cements (e.g., calcite and dolomite) are volumetrically insignificant and only partially fill pores. Exceptions to this include porosity-occluding cements associated with fractures and breccias in the vicinity of faults.

Dolomitization and depositional facies are found to exert a strong control on pore morphology, distribution, and interconnectivity. Porosity is principally controlled by the relative abundance of skeletal grains and by the presence of burrows. These highly porous facies acted as fluid pathways during burial diagenesis, resulting in their preferential dolomitization, solution enhancement of pre-existing pores, and creation of volume reduction-related porosity. The high CO2 storage capacity and low unplanned plume migration risk (due to depositional and/or diagenetic baffles) of dolomitized reefal reservoirs (e.g., Swan Hills and Leduc formations) make them more attractive targets for CCUS than those with limited capacity and/or potential migration pathways (e.g., fault-related fractures and breccias in the Wabamun Group). These results demonstrate that drill-core analysis, in combination with legacy data, can provide valuable insights into the factors that control reservoir CO2 injectivity, plume migration, and storage capacity.

The Western Canada Sedimentary Basin (WCSB) is a mature petroleum province. The majority of conventional hydrocarbon reserves are hosted in the differentially dolomitized carbonate strata of the Middle to Upper Devonian (Hay et al. 1994). Although these fields are now mostly depleted, there is increasing interest in extending field production by carbon capture, utilization, and storage (CCUS), which includes CO2-driven enhanced oil recovery (EOR) (Bohm et al. 2009). This method would produce remaining residual oil, trap CO2 over geological timescales, and contribute to reducing global greenhouse gas emissions necessary for mitigating the impact of anthropogenic climate change. Bachu et al. (2000) reviewed the potential for CCUS in the WCSB and concluded that oil/gas reservoirs (including Middle to Upper Devonian carbonate strata) and deep saline aquifers in the southwest, central, and northwest regions of Alberta were among the most promising targets (Fig. 1).

In mid-2022 a post-combustion CCUS project was brought online at the Glacier Gas Plant northwest of Grande Prairie, Alberta. This plant has been in operation for several years for pre-combustion CCUS and has generated 500,000 tons of CO2 offsets (Belenkie et al. 2022). Another large-scale CCUS project in the WCSB is the Alberta Carbon Trunk Line (ACTL), which was completed in Spring 2020. The ACTL is a 240-km-long pipeline that will transport CO2 from industrial sources to oil fields across central Alberta, where it will be used for EOR projects (Cole and Itani 2013). One such project involves injecting CO2 into the Leduc Formation (Upper Devonian) in the Clive Field, which has produced 70 million barrels of oil (50% of initial oil in place) since it was discovered in the 1950s.

Although EOR of the Middle to Upper Devonian dolostone reservoirs of the WCSB will undoubtedly have significant economic and environmental benefits, to date very little has been published on the petrophysical characteristics (e.g., porosity, permeability, pore morphology) of these reservoirs in the context of CCUS. These reservoirs formed through replacement dolomitization during shallow burial and were variably cemented by saddle dolomite and calcite during deeper burial (Stacey et al. 2021a). Burial dolomitization was driven by the regional flux of Mg-rich fluids through faults and aquifers, and as such, dolostone reservoirs are typically found in proximity to faults (e.g., Kaufman et al. 1991; Mountjoy et al. 1999; Duggan et al. 2001; Green and Mountjoy 2005; Stacey et al. 2021a).

Previous studies of Devonian carbonate strata in the WCSB have established that depositional facies with high initial porosity and permeability were important in controlling fluid flow and hence are preferentially dolomitized, ultimately forming hydrocarbon reservoirs (e.g., Kaufman et al. 1991; Saller et al. 2001). Although previous research into fault-controlled dolostone bodies (Davies and Smith 2006) had suggested that the best-quality reservoir can be found proximal to faults, recent outcrop studies (Koeshidayatullah et al. 2020a; McCormick et al. 2021; Stacey et al. 2021b) have shown that dolostone bodies in the vicinity of faults are generally tight due to the large volume of interconnected saddle dolomite cement in fault-adjacent breccias.

This study evaluates the petrophysical characteristics of conventional oil and gas fields in the partially dolomitized Swan Hills Formation, Leduc Formation, and Wabamun Group, all of which satisfy the initial requirements for long-term CO2 storage. In particular, they are at sufficient depths to facilitate the storage of CO2 in a supercritical state, as reservoir temperatures and pressures are greater than 31.1°C and 7.38 MPa, respectively (Bachu et al. 2000). A key question when assessing reservoir suitability, however, is whether injected CO2 will migrate from the initial injection site over time, and if it will eventually escape into overlying strata (such as shallow aquifers) or the atmosphere. In order to reduce this risk, it is necessary to identify potential trapping mechanisms (e.g., seals and aquitards), transport properties of deep aquifers (e.g., porosity/permeability and formation-water flow directions), and fluid pathways (e.g., faults and fracture zones) (White et al. 2004). Additionally, a thorough understanding of the temperature and formation-fluid composition (salinity) of reservoirs is required, as these factors control wettability (and therefore CO2 capillary trapping) and carbonate dissolution kinetics (Iglauer et al. 2011; Arif et al. 2017).

The formation waters of the Swan Hills Formation, the Leduc Formation and the Wabamun Group are currently flowing updip to the northeast and are assumed to be driven by past tectonic compression (Bachu 1999). These aquifers are separated from recharge areas by strong aquitards and/or aquicludes (the mudstones and shales of the Waterways Formation, Ireton Formation, and Exshaw Formation, respectively) and are expected to have a very long retention time due to their confinement (Bachu et al. 2000). Additionally, although deep-seated fault systems crosscut Devonian strata in certain areas of the WCSB (Schultz et al. 2016, 2017), none extend beyond the Upper Devonian Nisku and Ireton formations (Corlett et al. 2018), indicating that the risk of CO2 escaping into post-Devonian strata is low.

In order to further assess risk to future CCUS and EOR projects in the WCSB, this study has the following objectives:

  1. Document and describe the sedimentology (depositional facies) and diagenesis (dolomitization, cementation, fracturing) of selected Devonian carbonate reservoirs.

  2. Evaluate the control of these aspects on reservoir porosity and permeability through the integration of petrophysical data (core-plug measurements, pore morphology).

  3. Determine the suitability of the studied reservoirs for CO2 injection and storage by assessing their storage capacity and potential CO2 plume migration pathways.

The WCSB is a large sedimentary basin that underlies most of Alberta and extends into northeast British Columbia, the southern Northwest Territories and southwest Saskatchewan. The WCSB is composed of a north-easterly tapering wedge of sedimentary rocks that are more than 6 km thick, extending south from the Canadian Shield into the Cordilleran foreland thrust belt, thickening from east to west (Porter et al. 1982). During the late Proterozoic to Jurassic, a succession dominated by carbonate rocks was deposited on a platform that extended along the western ancient passive margin of North America. The later development of a foreland basin dictated the depositional style from the Late Jurassic to Paleocene, with deposition of clastic sediments sourced from the Canadian Cordillera (Mossop and Shetsen 1994) during the Columbian (Middle Jurassic to Early Cretaceous) and Laramide orogenies (Late Cretaceous to Paleocene; Panǎ et al. 2001). These effectively overprinted the earlier tectonic events associated with the Antler Orogeny (Late Devonian to Mississippian; Nelson et al. 2002; Hauck et al. 2017).

Depositional Facies and Reservoir Characteristics

The Middle to Upper Devonian Swan Hills Formation is composed of platform-top and reefal carbonates consisting of lagoonal facies (commonly represented by Amphipora stromatoporoid floatstone), reef-flat facies (stromatoporoid rubble-dominated rudstone), and reef-margin facies (encrusting stromatoporoid boundstone) (Kaufman and Meyers 1988; Wendte and Uyeno 2005). The Swan Hills Field was discovered in 1957 with 2.45 billion barrels (bbls) of oil initially in place and recoverable reserves of 904 million bbls (Viau 1988). The field is situated on the Swan Hills Platform in the West Shale Basin (Fig. 1) at a present burial depth of 2310–2590 m, covering an area of ca. 3,321 km2. Oil is mainly hosted in reef margin facies, stratigraphically trapped by the overlying shales of the Waterways Formation (Viau 1987). Legacy production data (provided by the Alberta Geological Survey) from the Swan Hills Field well considered by this study (04-13-039-09W5) indicates mean oil flow rates of 22.1 bbls/d. The Caroline Field was discovered in 1986, with 2.1 trillion cubic feet (tcf) of sour gas initially in place (Harvey et al. 1993). The field is located on the southern Swan Hills Platform in the East Shale Basin (Fig. 1) at a present burial depth of ca. 3,500 m, covering an area of ca. 295 km2. Gas is hosted in pervasively dolomitized reefal facies and is stratigraphically trapped updip by the shales of the Waterways Formation (e.g., Oldale et al. 1994). To date, 1.9 tcf of gas has been recovered from the Swan Hills Formation at Caroline, at a maximum rate of 433 mmcf/d (million cubic feet per day) (January 1999) from 12 wells.

The Upper Devonian Leduc Formation is composed of reefs that developed on older carbonate platforms. In the East Shale Basin, isolated reef complexes developed on the Cooking Lake Platform (Wendte 1994), and in the West Shale Basin they form the Rimbey–Meadowbrook reef trend, a feature that is ca. 560 km in length (Fig. 1). Leduc Formation reefs are composed of stromatoporoid-dominated reef flank and fore-reef facies (Drivet and Mountjoy 1997). The Acheson Field was discovered in 1950, with 186 million bbls of oil initially in place and recoverable reserves of 131 million bbls (Switzer et al. 1994). The field is part of the Rimbey–Meadowbrook reef trend and is at a present burial depth of ca. 1500 m, covering an area of ca. 88 km2. Oil is hosted in pervasively dolomitized reefal facies and is stratigraphically trapped updip by the shales of the Ireton Formation (Switzer et al. 1994; Drivet and Mountjoy 1997). The maximum rate of recovery from the Leduc A Pool at Acheson was in August 1988, when 21 wells recovered a total of 19 mbbls/d (thousand barrels per day).

The Upper Devonian Wabamun Group is composed of alternating carbonates and evaporites that were deposited on an epeiric carbonate ramp (Halbertsma 1994) (Fig. 1). Carbonate facies are dominated by nodular limestones, mudstones, skeletal and/or peloidal wackestones, and minor packstones (Halim-Dihardja and Mountjoy 1988). The Pine Creek Field was discovered in 1955, with 421 million cubic feet of gas initially in place (Halbertsma 1994). This field is in the West Shale Basin (Fig. 1) at a present burial depth of ca. 3000 m, covering an area of ca. 205 km2. Gas is hosted in subtidal to intertidal peloidal wackestones that are pervasively dolomitized in intervals < 20 m thick, stratigraphically trapped updip by the shales of the Exshaw Formation (Green and Mountjoy 2005). Production peaked in 1977 to 1978, during which pool production reached 30 mmcf/d from only two wells. The Eaglesham North Field was discovered in 1987, with 36 million bbls of oil initially in place and recoverable reserves of eight million bbls (Halbertsma 1994). This field is located in the Peace River Arch area (Fig. 1) at a present depth of ca. 2100 m, covering an area of ca. 495 km2. Oil is hosted in dolomitized mudstones and peloidal wackestones and/or packstones stratigraphically trapped updip by the shales of the Exshaw Formation and the dolomitic marls of the Banff Formation (Saller and Yaremko 1994). In the 12-month period from January 1984 to January 1985, 59.7 bbls/d oil and 27.9 mcf/d gas (mean values) was recovered from the Eaglesham North well considered by this study (13-13-078-26W5) (Alberta Geological Survey data).

Drill cores from six wells that penetrate the Swan Hills Formation, the Leduc Formation, and the Wabamun Group (Table 1) were selected based on their locations in the regions identified by Bachu et al. (2000) as being prospective for CCUS.

Over 220 meters of drill core was described over a two-week period at the Alberta Energy Regulator Core Research Centre in Calgary, Canada. Depositional textures were classified based on the Dunham Classification (1962) with the Embry and Klovan (1971) modification for intervals interpreted to be bound at deposition. Sedimentary textures and fossil assemblages were assigned to depositional facies using classifications previously proposed for the Swan Hills Formation (Kaufman and Meyers 1988; Saller et al. 2001; Wendte and Uyeno 2005), the Leduc Formation (Wendte 1994; Drivet and Mountjoy 1997), and the Wabamun Group (Halim-Dihardja and Mountjoy 1988; Saller and Yaremko 1994). In addition to this, core description focused on cataloguing the occurrence and distribution of diagenetic features (e.g., dolostone–limestone contacts, macroporosity types, fractures, and cements).

Forty thin sections were made from samples that best represented the depositional fabrics, diagenesis (e.g., dolomitization, dissolution, and cementation) and pore types identified in core. These samples were stained with alizarin Red-S and potassium ferricyanide (Dickson 1966) and impregnated with blue-dye resin to identify porosity. Thin sections were examined under plane-polarized light (PPL) and cross-polarized light (XPL) using a Nikon Eclipse LV100N POL microscope. Calcite and dolomite crystal textures and porosity types were described based on the classifications of Flügel and Munnecke (2010), Sibley and Gregg (1987), and Choquette and Pray (1970).

Legacy core-plug data (346 measurements, including porosity and maximum permeability, KMAX) of the fields in question was provided by the Alberta Geological Survey and correlated by depth with lithostratigraphic logs (completed at the Core Research Centre, Calgary) drawn for each well. Following this, the data were statistically analyzed to determine the variability of porosity and permeability in each depositional facies as well as the overall impact of dolomitization and fracturing on reservoir quality. The Alberta Geological Survey also provided legacy production data from the wells considered in this study, which was used to evaluate reservoir fluid-flow properties.

Swan Hills Formation

Depositional Features

The depositional facies of the Swan Hills Formation (Swan Hills and Caroline fields) were identified using the facies classifications (discussed below) of Kaufman and Meyers (1988), Saller et al. (2001), and Wendte and Uyeno (2005). Lagoonal, shoal, back-reef, reef, and fore-reef facies are present in the Swan Hills Field and are interpreted to represent shallowing-upward sequences (Fig. 2). The same sequences are also present in the Caroline Field and are composed of lagoon, shoal, and back-reef facies (Fig. 3).

In the Swan Hills Field, lagoonal facies are characterized by the occurrence of stromatoporoid (Amphipora) rudstones and packstones, tabular stromatoporoid framestones, and crinoidal carbonate wackestones (Fig. 2A, G). High-energy shoal facies are composed of rudstones that contain a heterogeneous assemblage of Amphipora, Stachyodes, and tabular and bulbous stromatoporoids (Fig. 2C). Back-reef facies exhibit repeating alternations of brachiopod-dominated packstones overlain by tabular/bulbous stromatoporoid framestones that contain abundant brachiopods and rare corals. Each cycle top (ca. 2 m thick, Fig. 2) is marked by the occurrence of framestones that contain bulbous stromatoporoids (Fig. 2E). Reef-crest facies consist of grainstones and rudstones that are dominated by tabular stromatoporoids with uncommon crinoids and rare bulbous stromatoporoids. The fore-reef is characterized by tabular and bulbous stromatoporoid rudstone intervals with detrital reef material (Fig. 2).

In undolomitized intervals of the Caroline Field, shoal facies consist of tabular stromatoporoid framestone (Figs. 3, 4A). Dolomitized lagoonal facies consist of packstone (with subordinate grainstone) containing stromatoporoids (commonly Amphipora but also Stachyodes) and uncommon gastropods (Figs. 3, 4C). Dolomitized high-energy shoal facies are characterized by the occurrence of stromatoporoid (predominantly Stachyodes) framestone and grainstone (Figs. 3, 4E).

Diagenetic Features

The Swan Hills Field is not dolomitized, but saddle-dolomite-cemented fractures and associated minor brecciation occur in shoal, back-reef, reef, and fore-reef facies (e.g., Fig. 2C). Conversely, the lagoon, back-reef, and shoal facies of the Caroline Field are pervasively dolomitized. A dolostone–limestone contact is observed at ca. 3713 m (Fig. 3). This contact is bedding-parallel, sharp (non-gradational), and occurs within the same facies. In the Caroline Field, dolomite, calcite, and rare, double-terminated quartz crystals are observed cementing molds and vugs in shoal and back-reef facies, with dolomite cement also filling fractures and adjacent vugs (Fig. 4E). The most heavily fractured intervals in the Swan Hills Formation are in reefal facies. In the Swan Hills Field, back-reef, reef, and fore-reef facies contain significantly more fractures (14, 8, and 18 per m, respectively) than lagoon and shoal facies (5 and 4 per m, respectively). Similarly, the dolomitized back-reef facies of the Caroline Field are more heavily fractured (14 per m) than dolomitized lagoon and shoal facies (4 and 2 per m, respectively). Additionally, limestone shoal facies have a similar fracture density (3 per m) to dolostone shoal facies.

Leduc Formation

Depositional Features

Because the Leduc Formation is pervasively and destructively dolomitized with most allochems removed by extensive dissolution, it is difficult to identity depositional facies. Despite this, the relative size of molds and vugs are used to infer the original allochems present in the Acheson Field (Fig. 5). Larger pores (> 6 cm) were inferred to originally have been tabular stromatoporoids (Fig. 5A), medium-sized pores (1 to 4 cm) were identified as Stachyodes (Fig. 5C), whereas smaller pores (0.2 to 2 cm) were inferred to have originally been Amphipora (Fig. 5E). Based on this, the depositional facies of the Acheson Field were identified (Fig. 5) following the facies classifications of Wendte (1994) and Drivet and Mountjoy (1997). Shoal facies are dominated by Stachyodes and Amphipora rudstones with rare brachiopods and tabular stromatoporoids. Lagoonal facies are characterized by Amphipora-dominated floatstones and packstones with uncommon Stachyodes (Fig. 5).

Diagenetic Features

The lagoon and shoal facies of the Acheson Field are pervasively dolomitized, and dolomite is strongly fabric destructive. Dolomite cement is observed partially filling molds and vugs in both lagoon and shoal facies (Fig. 5C, E). Fracture densities in the Acheson field are similar in both lagoon and shoal facies (8 and 5 per m, respectively) and are slightly higher than those of dolomitized lagoon and shoal facies in the Swan Hills Formation (Caroline Field).

Wabamun Group

Depositional Features

The depositional facies of the Pine Creek (Fig. 6) and Eaglesham North (Figs. 7, 8) fields were identified using the facies classifications of Halim-Dihardja and Mountjoy (1988) and Saller and Yaremko (1994). In the Pine Creek Field, mid-ramp facies are intensely bioturbated and contain rare brachiopods and gastropods (Fig. 6). Undolomitized intervals with little bioturbation consist of laminated and massive wackestones (Fig. 6A) whereas partially dolomitized intervals consist of bioturbated packstone (Fig. 6B). Dolomitized intervals are characterized by the occurrence of fenestrae cemented by anhydrite (Fig. 6E) and highly porous burrows (Fig. 6G).

The mid-ramp facies of the Eaglesham North Field (Figs. 7, 8) are broadly similar to those of the Pine Creek Field, consisting of both limestone and dolomitized bioturbated crinoidal wackestones (with subordinate packstone) (Fig. 8A). Breccias occur in the dolomitized intervals of the Eaglesham North Field (Fig. 7) and are characterized by varying degrees of brecciation, ranging from minor fracturing and cementation (Fig. 8C) to mosaic and chaotic breccias fully supported by calcite and dolomite cement (Fig. 8E).

Diagenetic Features

The mid-ramp facies of the Pine Creek Field are pervasively dolomitized in an interval ca. 13 m thick (Fig. 6). A dolostone–limestone contact is observed at ca. 3057.09 m (Fig. 6). This contact is bedding-parallel, sharp (non-gradational), and occurs within the same facies. Anhydrite cement is observed filling fenestrae in dolomitized intervals (Fig. 6E). Dolomite cements partially fill moldic porosity related to burrows in both limestone and dolostone intervals (Fig. 6G). Similarly, the dolomitized mid-ramp facies of the Eaglesham North Field also tend to be cemented by calcite and dolomite, although the cements are typically confined to fractures (Fig. 8C). In brecciated intervals in mid-ramp dolomite, breccia clasts are fully supported by calcite cement, with minor saddle dolomite cement lining the margins of breccia clasts (Fig. 8E). In the Pine Creek Field, limestone is significantly more fractured than dolostone (12 and 7 per m, respectively). Conversely, in the Eaglesham North Field (Wabamun Group), dolostone has a greater fracture density than mid-ramp limestone (14 and 8 per m, respectively). Additionally, brecciated intervals in mid-ramp dolostone exhibit even higher fracture densities (24 per m).

Dominant Pore Types

Dominant pore types in each depositional facies are summarized in Table 2. Pore diameters were measured in hand specimen. Representative pore types are illustrated for the Swan Hills Formation (Figs. 2, 4), the Leduc Formation (Fig. 5), and the Wabamun Group (Figs. 6, 8).

Swan Hills Formation

Porosity in the Swan Hills Field is generally moldic (Table 2) and fabric selective, commonly occurring as a result of partial or complete dissolution of bulbous stromatoporoids (Fig. 2E). In thin section, moldic pores exhibit varying degrees of interconnectivity and bitumen abundance (Fig. 2B, F), and are occasionally cemented by saddle dolomite (Fig. 2D). Porosity in fine-grained intervals (e.g., fossiliferous wackestone) is restricted to fractures and intraparticle pores in bioclasts (Fig. 2H). In comparison, low porosity intervals in the Caroline Field occur in limestone that is partially replaced by coarsely crystalline nonplanar dolomite (Fig. 4B). Porous intervals occur between dolomite rhombs that have a planar-s texture within molds of Stachyodes and Amphipora. In thin section, intercrystalline pores are occasionally partially cemented by calcite (Fig. 4D). In hand specimen, vugs in the Caroline Field are up to 80 mm in diameter (Fig. 4E). They occur where Amphipora and Stachyodes are interpreted to have been dissolved and solution enhanced.

Leduc Formation

Porosity in the Acheson Field is generally vuggy (Table 2) and is related to the solution enhancement of molds created by the dissolution of large allochems such as stromatoporoids up to 90 mm in diameter (Fig. 5A). In thin section, porosity is either intercrystalline (Fig. 5B) or vuggy (Fig. 5D). Dolomite cements commonly line pore margins but never fully occlude pores (Fig. 5F).

Wabamun Group

Porosity in the undolomitized intervals of the Pine Creek Field is generally restricted to partially calcite cemented interparticle pores (Fig. 6B). Lower porosities in partially dolomitized intervals are due to the presence of coarsely crystalline anhydrite nodules (Fig. 6D, F). In contrast, porosity in dolomitized intervals occurs in sucrosic dolomite that formed in burrows. In thin section, this intercrystalline porosity is occasionally partially cemented by calcite (Fig. 6H). In hand specimen, rare vugs up to 32 mm in diameter also occur in dolomitized burrows (Fig. 6G). In comparison, porosity in the Eaglesham North Field is predominantly restricted to fractures in fossiliferous limestone (Fig. 8B) and coarsely crystalline nonplanar replacement dolomite (Fig. 8D). Isolated vugs (up to 70 mm in diameter) occur in dolomitized breccias and are commonly lined by coarsely crystalline saddle dolomite cement (Fig. 8F).

Porosity and Permeability

The petrophysical characteristics of the facies of each field in this study were analyzed with legacy core plug porosity and permeability data (n = 346). Table 2 summarizes the core-plug porosity and permeability measurements of the individual depositional facies that together constitute the depositional environments discussed below.

The lagoon and shoal facies of the Swan Hills Field (Swan Hills Formation) have relatively low porosity (mean 3.6% and 3.4%, respectively) and permeability (mean 0.57 mD and 2.69 mD, respectively). Back-reef facies have the highest porosity (mean 5.3%) and permeability (mean 14.09 mD) of any facies in the Swan Hills Field. Reef and fore-reef facies have lower porosity than back-reef facies (mean 2.3% and 3.3%, respectively), but significantly higher permeability (mean 7.37 mD and 8.10 mD, respectively) compared to lagoon and shoal facies (Table 2, Fig. 9A).

Dolomitized lagoonal facies of the Caroline Field (Swan Hills Formation) have significantly higher porosity and permeability (mean 8.9% and 95.02 mD) than the limestone equivalent in the Swan Hills Field. Similarly, dolomitized shoal facies have higher porosity and permeability (mean 10.6% and 109.89 mD) than undolomitized shoal facies (mean 1.4% and 13.02 mD). Dolomitized back-reef facies have the highest porosity (mean 10.6%) and permeability (mean 151.08 mD) of any facies in the Caroline Field (Table 2, Fig. 9B).

Although the dolomitized lagoonal and shoal facies of the Acheson Field (Leduc Formation) have lower porosity values (mean 5.8% and 6.8%, respectively) than the equivalent dolomitized facies in the Caroline Field, the permeability is significantly higher (mean 176.91 mD and 1134.10 mD, respectively) (Table 2, Fig. 9C).

Dolomitized mid-ramp facies of the Pine Creek Field (Wabamun Group) have slightly higher porosity and significantly higher permeability (mean 5.4% and 32.37 mD) than mid-ramp limestone (mean 2.1% and 3.60 mD) (Table 2, Fig. 9D). The dolomitized ramp facies and brecciated intervals of the Eaglesham North Field (Wabamun Group) have slightly higher porosities than that of limestone (mean 1.5%, 1.2%, and 0.7%, respectively) but lower permeabilities (mean 1.41, 14.78, and 57.17 mD, respectively) (Table 2, Fig. 9E).

Dolostone–Limestone Contacts

In the Caroline Field, porosity increases from 1 m below the contact (dolostone = 0.5%) to the contact itself (1.2%) and increases 1 m above it (limestone = 1.4%). Permeability decreases from 1 m below the contact (dolostone = 0.50 mD) to the contact itself (0.02 mD) and increases 1 m above it (limestone = 0.07 mD) (Fig. 3).

In the Pine Creek Field, two dolostone–limestone contacts were observed above and below a dolomitized interval. Porosity increases from 1 m below the upper contact (dolostone = 1.4%) to the contact itself (1.8%) and decreases 1 m above it (limestone = 1.2%). Permeability increases from 1 m below the contact (dolostone = 0.90 mD) to the contact itself (2.50 mD) and is similar 1 m above it (2.50 mD). Porosity increases from 1 m below the lower contact (limestone = 1.1%) to the contact itself (3.5%) and also increases 1 m above it (dolostone = 8.4%). Permeability decreases from 1 m below the contact (limestone = 5.60 mD) to the contact itself (3.20 mD) but increases 1 m above it (dolostone = 56.00 mD) (Fig. 6).

Factors Controlling Porosity and Permeability

The prediction of reservoir quality (e.g., hydrocarbon storage capacity and deliverability) in carbonate strata is challenging due to the inherent complexity of their pore systems, which are controlled by depositional facies and diagenetic overprinting. Despite this, it is still possible to evaluate the relationship between the observed textures and measured porosity and permeability to determine the possible factors that control injectivity, storability, and storage security of CO2 on a well scale, and by extension, how these factors may generally affect other Devonian carbonate reservoirs across the WCSB.

Facies

In the Swan Hills and Leduc formations, the highest mean porosity occurs in facies with high concentrations of skeletal fossil fragments such as stromatoporoids (e.g., back reef and shoal, respectively, Fig. 9). This indicates that total porosity and pore size is principally controlled by the relative abundance of skeletal grains, regardless of whether strata are dolomitized or not (Fig. 10). The porous structure of many skeletal grains, particularly stromatoporoids (e.g., Amphipora, Idiostroma, and Stachyodes), may have facilitated the flow of diagenetic fluids, and their inherent mineralogical instability (e.g., high-Mg calcite) almost certainly rendered them susceptible to dissolution (e.g., Amthor et al. 1993), and resulted in the generation of porosity where pore size and morphology is dictated by the size of the original allochem (Fig. 10). Moldic porosity is most commonly associated with Amphipora, which is most abundant in lagoonal facies (as previously concluded by Saller et al. 2001). Stachyodes is common in shoal facies (as previously concluded by Wendte 1994) and is associated with both moldic and vuggy porosity. Bulbous and tabular stromatoporoids are most abundant in reefal facies (as previously concluded by Kaufman and Meyers 1988; Drivet and Mountjoy 1997). Generally, the highest porosity and permeability is associated with the dissolution of bulbous and tabular stromatoporoids in reefal facies (Table 2, Fig. 9) and are associated principally with vuggy pores in pervasively dolomitized intervals and with moldic porosity in limestone. Moldic pores, although large, may not contribute to effective porosity in these formations since they appear to be isolated, although some connectivity may be facilitated via open fracture networks, microporosity in the fine-grained mud matrix, or intercrystalline porosity networks present in neomorphosed (replaced and/or recrystallized) matrix (e.g., Fig. 4D, 5B).

In the Wabamun Group, porosity is principally controlled by the presence of burrows (e.g., Saller and Yaremko 1994). Intercrystalline porosity is restricted to burrows in pervasively dolomitized intervals, and is significantly higher in heavily bioturbated intervals than massive, less bioturbated, wackestone and packstone (Fig. 10). The high permeability of pervasively bioturbated, dolomitized intervals (e.g., Fig. 6G, H) likely indicates the presence of effective pore networks. One possible explanation for this is that pore networks have some degree of connectivity through intersecting branching burrows, and that interconnectivity is proportional to the degree of bioturbation (as previously concluded by Baniak et al. 2013, 2022).

Dolomitization

Replacement dolomitization of the Middle to Upper Devonian carbonate strata of the WCSB is interpreted to have taken place from the circulation of connate seawater through aquifers and faults during shallow burial (Stacey et al. 2021a, and references therein). Pervasively dolomitized intervals generally have significantly higher porosity and permeability than undolomitized intervals (Figs. 9, 10), in part because dissolution of skeletal grains (e.g., stromatoporoids) is interpreted to have taken place during burial dolomitization (which postdates the formation of early low-amplitude stylolites and took place at temperatures between 32 and 86°C; Stacey et al. 2021a, and references therein). In the Swan Hills and Leduc formations, solution enhancement of pre-existing porosity during dolomitization is also interpreted to have occurred (e.g., Fig. 4E, 5C) along with the creation of intercrystalline porosity (Figs. 4D, 5B, 10) (including by the ca. 13% reduction in rock volume caused by the smaller size of dolomite crystals compared to those of calcite, Machel 2004). In the Wabamun Group, replacement dolomitization, and subsequent recrystallization during deeper burial (Stacey et al. 2021a), enhanced intercrystalline porosity in dolomitized burrows (Figs. 4G, H, 10). The wide range of porosity and permeability measurements in dolostone (Table 2, Fig. 9) appears to be mainly facies controlled and is related to the relative abundance and size of skeletal grains. For example, in the Acheson Field (Leduc Formation), shoal facies have higher porosity and permeability (mean 6.8% and 1134.10 mD, respectively) than lagoonal facies (mean 5.8% and 176.91 mD, respectively), which corresponds to larger pore diameters in the former (up to 9 cm, Fig. 5C) than the latter (up to 3 cm, Fig. 5E).

Porosity and permeability also vary across dolostone–limestone contacts in the Caroline Field (Swan Hills Formation) and the Pine Creek Field (Wabamun Group). Although this is negligible in the Caroline Field (Fig. 3), limestone porosity and permeability values increase towards dolostone contacts in the Pine Creek Field (Fig. 6, 9D). One possible explanation for this is through the leaching of limestone by acidic fluids at the terminal edge of dolomitization fronts where the Mg/Ca ratio of the fluid has dropped low enough for dolomitization to wane, but dissolution of calcite persists. Such a process has been suggested to occur during the development of dolomitization fronts (e.g., Koeshidayatullah et al. 2020b).

Cementation

Although replacement dolomitization resulted in an overall porosity increase (Table 2, Fig. 9), later phases of dolomite and calcite cementation variably reduced it (Fig. 10). In the Swan Hills and Leduc formations, dolomite cement occurs as overgrowths on replacement dolomite, but as this cement mainly lines vugs, porosity is not significantly reduced (e.g., Fig. 5D). Similarly, as blocky equant calcite cements in the Swan Hills Formation and Wabamun Group are volumetrically insignificant and only partially fill pores (Figs. 4D, 6H), porosity is not significantly reduced. It is possible that porosity was preserved because cementation was ended by the emplacement of hydrocarbons during the Jurassic to Cretaceous Laramide Orogeny (Creaney et al. 1994). This is consistent with previous studies (e.g., Stacey et al. 2021a and references therein) which conclude that replacement-dolomite phases in the Devonian strata of the WCSB formed far earlier (Devonian to Mississippian) than later deep-burial cementation (Cretaceous).

The cements that have the most detrimental impact on storage capacity and effective fluid flow during CO2 injection are within fractures. In the Swan Hills Field, saddle dolomite cement occludes fractures (Fig. 2C) and, locally, moldic porosity in bulbous stromatoporoids (Fig. 2D). In the Eaglesham North Field, porosity is extremely low (< 1.4%) in intervals with saddle dolomite and calcite cemented fault-controlled breccias (Fig. 7, 2120.6–2126.6 m depth). Similar breccias crop out in the Middle Cambrian Cathedral Formation (Canadian Rocky Mountains) and are observed to be fault-controlled (e.g., Jeary 2002; Vandeginste et al. 2007; Johnston et al. 2009; Stacey et al. 2021b; McCormick et al. 2023), but the full lateral extent of fault-related dolostone bodies and associated cemented breccias is difficult to determine in the subsurface because they are subseismic in scale (e.g., Sharp et al. 2010). In outcrop, replacement dolostone bodies extend up to 6.5 km in length, and breccias in replacement dolostone bodies exclusively occur within 50 m of normal faults, with breccia clasts fully supported by saddle dolomite cement (Stacey et al. 2021b). Observations of saddle dolomite cementation in the vicinity of faults were also made in the Simonette Field (Swan Hills Formation) (Duggan et al. 2001). The similarity of these fault-related breccias to those in the Wabamun Group suggest a common formation mechanism, by convection of dolomitizing fluids along fault planes (Stacey et al. 2021b) to form bodies of limited lateral extent (Fig. 10).

Fractures

Fracture networks are more likely to occur in dolostone as opposed to limestone, because dolostone has a greater tensile strength (i.e., less ductile) which allows fractures to remain open and therefore have higher effective permeability (Schmoker et al. 1985). However, there are a range of other controls on fracture density (e.g., distance from fault, mechanical stratigraphy, curvature, and depositional facies) that can also control fracture distribution (e.g., Dati 2013; Korneva et al. 2018, and references therein).

The data collected from the various reservoirs in this study show that fractures occur in limestone and dolostone intervals and are broadly equally abundant in the Swan Hills Formation (e.g., shoal limestone = 3 per m, shoal dolostone = 2 per m), but are slightly more abundant in limestone than dolostone (excluding brecciated intervals and associated fracture networks) in the Wabamun Group (ramp limestone = 12 per m, ramp dolostone = 7 per m). This most likely reflects a strong control by depositional facies on the occurrence and density of fractures regardless of dolomitization, because dolomitized and undolomitized reefal facies in the Swan Hills Formation are equally fractured (13 and 14 per m, respectively). This agrees with Dati (2013) and Korneva et al. (2017), who found that rock texture (crystal/grain size and depositional facies) is more important that mineralogy (e.g., calcite vs. dolomite) in regulating the development and intensity of fracture networks. This analysis excludes fracture networks in brecciated intervals (e.g., in the Eaglesham North Field) where fracture orientations and density are random because they formed through hydrobrecciation, dissolution, and partial collapse adjacent to faults (e.g., Kaufman et al. 1991; Saller and Yaremko 1994; Mountjoy et al. 1999; Duggan et al. 2001; Green and Mountjoy 2005; Stacey et al. 2021a).

Legacy well production data indicates that the Swan Hills Field and the Eaglesham North Field have good flow rates (22.1 and 59.7 bbls/day for the studied wells) despite their low mean porosity, which is consistent with fracture-controlled flow. It is more difficult to assess the role of fracture-controlled flow in reservoirs with higher matrix porosity and permeability (e.g., Caroline, Acheson, and Pine Creek fields) as the matrix likely contributes significantly to productivity. The Pine Creek Field (Wabamun Group) has a much lower mean porosity than the Caroline Field (Swan Hills Formation) (3.5% and 7.8%, respectively), but it still produced approximately half the volume of gas as the Caroline Field over a 12-month period (mean 13615 and 23720 mcf/d, respectively), suggesting that fracture-controlled flow played an important role in gas deliverability in the Pine Creek Field. Similarly, it is likely that fractures contribute to flow in the Acheson Field (Leduc Formation), as fluid flow rates are higher in the Acheson Field than the Caroline field (69.8 (oil) and 46.9 (water) bbls/d, respectively), despite both fields having similar mean porosities (6.1% and 7.8%, respectively).

Reservoir Suitability for CO2 Storage

Storage of CO2 in the retired Devonian hydrocarbon fields of the WCSB may provide a valuable early step in the development of subsurface storage solutions in Canada. These fields have a proven fluid trap and seal that has maintained hydrocarbons in place since the Cretaceous (Creaney et al. 1994). Depletion of hydrocarbons over decades of production has reduced pore pressures, providing an opportunity for gas injection which will increase pore pressure with a relatively low risk of seal failure. Furthermore, the geology and production history of the fields are well known, and the sites have well and surface infrastructure that can be repurposed for CO2. Many fields are also suitable for enhanced oil recovery, which may significantly increase the economics of any given project. While retired reservoirs present a very low risk of geological storage failure, the presence of existing wells that intersect the storage reservoir is a potential path to a loss of containment. It is beyond the scope of this study to assess this risk, but fields with a lower number of well penetrations (which could be continuously monitored) would be most ideally suited to CCUS.

High-quality reservoirs for CO2 storage typically have high porosity and high permeability to facilitate CO2 injection (Cook 2006). Successful retention of CO2 principally controlled by the presence of effective caprocks and selection of reservoirs that are at a depth in which the injectant can remain in a supercritical state (> 31.1°C; > 7.38MPa), typically greater than 1,000 m (Bachu et al. 2000). Due to the compressibility of supercritical CO2 with increased pressure and depth, deeper storage sites may prove more suitable as they are able to store larger volumes of CO2. Long-term trapping mechanisms for the secure storage of CO2 include residual, capillary, and mineral trapping (Iglauer et al. 2011; Zhao et al. 2014), and therefore together with the porosity and permeability, it is necessary to understand the size, shape, and genesis of porosity.

CO2 is normally injected into reservoirs in a dry (anhydrous) supercritical state, but it usually becomes more hydrated as it migrates away from the wellbore (Wang et al. 2013). The solubility of CO2 increases with temperature and pressure but decreases with an increase in fluid salinity (Li et al. 2018), adding complexity to the prediction of the migration pathways and distribution of CO2 in the reservoir, as well as understanding the wettability of pores (Wang et al. 2013). Experimental work (Arif et al. 2017) has shown that carbonate reservoirs with temperatures > 34°C and low salinities (0–5 wt.% NaCl) have good CO2 storage potential because they are more likely to be water-wet, which makes them less susceptible to vertical plume migration.

In light of this, all of the reservoirs considered by this study are at sufficient depths to maintain CO2 in a supercritical state (Swan Hills Field, 2310–2590 m; Caroline Field, ca. 3500 m; Acheson Field, ca. 1500 m; Pine Creek Field, ca. 3000 m; Eaglesham North Field, ca. 2100 m), and have sufficient sealing capacity to ensure capillary trapping of CO2 (Bachu et al. 2000; Miocic et al. 2016). Apart from the Acheson Field (37.5°C), each of the reservoirs considered by this study have relatively high temperatures (Swan Hills Field, 57.5°C; Caroline Field, 87.5°C; Pine Creek Field, 75°C; Eaglesham North Field, 52.5°C). Central and southern portions of the WCSB exhibit a relatively low temperature gradient (Weides and Majorowicz 2014), which means that CO2 will be denser than at equivalent depths under a normal geothermal gradient, further inhibiting leakage by buoyancy forces. These temperatures are conducive to water wettability, although the actual wettability is not known. In reality, wettability is likely to be mixed, because this is the case for most carbonate reservoirs (e.g., Chilingar and Yen 1983) and because the formation fluids of Devonian strata in the WCSB are highly saline (e.g., Bachu 1995; Hutcheon et al. 1995). CO2 itself can also act as a wetting fluid, meaning that wettability can vary in three-phase CO2 water hydrocarbon systems depending on the fluid that mineral surfaces are exposed to (Wang et al. 2013). This is important to the mechanism of CO2 trapping, since mineral reactivity and residual trapping are partly dependent on wettability (e.g., Iglauer et al. 2011).

The results from this study show that the principal control on porosity and permeability in the studied fields in the WCSB is dolomitization, and then facies, with the highest porosity and permeability occurring in dolomitized reefal facies (Fig. 10). On this basis, the fields with the most favorable reservoir properties for CO2 storage are those with dolomitized reefal strata, such as the Caroline Field (Swan Hills Formation) and the Acheson Field (Leduc Formation). The Caroline Field had a high original in place volume of gas (2.1 tcf; Harvey et al. 1993) and has the highest mean porosity (7.8%) of all the reservoirs considered by this study. It has a mixture of intercrystalline, moldic, and vuggy pores that are not significantly cemented by calcite or dolomite (Fig. 4), and in the studied wells, there are no significant baffles to flow because dolomitization is pervasive and has homogenized porosity and permeability variations between facies (Table 2, Fig. 9B). Although there is some potential for CO2 to migrate into laterally adjacent porous limestone, the risk of CO2 escaping into overlying aquifers is low, because the field is deeply buried (ca. 3500 m) and is effectively sealed by the overlying Waterways Formation. Similarly, the Acheson Field originally contained 186 million bbls of oil (Switzer et al. 1994) and also has a high mean porosity (6.6%) (Table 2, Fig. 9C). As in the case of the Caroline Field, there are no significant baffles to flow, and calcite/dolomite cements are volumetrically insignificant (e.g. Fig. 5D).

The Pine Creek Field (Wabamun Group) is only partially dolomitized and therefore has a significantly lower mean porosity (3.6%) than that of the Caroline and Acheson fields (Table 2, Fig. 9D). This is also expressed by a lower original in-place volume of gas (421 million cubic feet; Halbertsma 1994). Despite this, the reservoir is relatively homogeneous (i.e., little variation in depositional facies and degree of dolomitization), apart from minor potential flow baffles close to dolostone–limestone contacts where burrows are cemented (e.g., Fig. 6). As such, the Pine Creek Field has some potential for CO2 storage but is limited by its low porosity.

The lowest mean porosity (1.1%) encountered by this study was in the Eaglesham North Field (Wabamun Group) (Table 2, Fig. 9E), which only had 36 million bbls of oil initially in place (Halbertsma 1994). Although the pores appear to be poorly connected, compartmentalized by calcite and saddle dolomite cemented breccias, the high permeability of the field (mean 455 mD) indicates a high volume of fractures that may control fluid flow. As such, the Eaglesham North Field is the least viable prospect for CO2 storage, since it has low matrix porosity but an abundance of fractures that could lead to irregular plume development and risk lateral containment failure. Similarly, the mean porosity of the Swan Hills Field (Swan Hills Formation) is also low (3.4%) (Table 2, Fig. 9A). This, coupled with a low mean permeability (2.6 mD), could potentially inhibit injectivity, rendering the Swan Hills Field an unattractive prospect for CO2 storage.

Injectivity of CO2 into the Devonian Carbonates of the WCSB

The potential for alteration of carbonate pore systems by dissolution and/or cementation effects of carbonic-acid formation has been discussed by a number of authors (e.g., Zekri et al. 2009; Khather et al. 2017). During CO2 injection, reactions between carbonic acid and carbonate minerals (e.g., calcite and dolomite) can lead to the alteration of the host-rock petrophysical properties through dissolution, precipitation, and mechanical compaction (Zekri et al. 2009). Recent work (Khather et al. 2017) has shown that porosity and permeability can both increase and decrease as a result of flooding of carbonated brine (i.e., CO2-saturated fluid) under reservoir conditions. Many of these studies are based solely on laboratory experiments, however, and generally fail to describe a mechanism for the flushing of acidic brines through the reservoir over an extended period of time. Snippe et al. (2023) show that the potential for dissolution may be low and that, under realistic subsurface conditions, water mobility rates near the CO2 plume can result in extremely limited carbonate dissolution. This is consistent with geochemical models which suggest that CO2-saturated fluids can attain equilibrium with carbonate minerals (e.g., calcite) within a few days of initial injection (e.g., Delerce et al. 2023). Furthermore, experiments suggest that dolomite does not react with anhydrous CO2 near the injection site, because once CO2 has dissolved in the formation water away from the borehole, the pH of the fluid may be buffered (Wang et al. 2013).

The findings of this study suggest that the best potential for CO2 storage is in the dolomitized reservoirs of the Caroline, Acheson, and Pine Creek fields. As previously discussed, these fields have moderate to good mean permeability (Table 2, Fig. 9), related to the presence of intercrystalline, moldic, and vuggy porosity and open fractures. This would facilitate both injectivity and plume migration in these reservoirs, the latter of which may be affected by heterogeneities that include interbedded undolomitized strata and cementation (e.g., anhydrite cementation in the Pine Creek Field). Plume migration is most likely to be focused initially along the best-connected pore network, for example, where an intercrystalline pore network is developed a relatively even sweep of CO2 might be anticipated. In contrast, fractures and vugs could channelize the migration of CO2 and potentially lead to bypassing of smaller pores that would be suitable for capillary or residual trapping of CO2. In this case, solution enhancement of fractures and vugs may take place, further enhancing the permeability of the largest pores and increasing pore-scale heterogeneity. Eventually, however, the system will reach equilibrium and carbonate minerals will precipitate, trapping CO2 by in situ mineralization (Wang et al. 2013; Khather et al. 2017). Overall, the high average permeability and abundance of intercrystalline porosity in the Acheson Field means that it has excellent potential for CO2 injection and plume migration. The Caroline Field also has very good potential, although the range in pore sizes and shapes (large vugs, fractures, and intercrystalline pores) will create complexity in wettability, water–rock reaction, and trapping, potentially resulting in heterogeneous plume migration and storage efficiency. In the Pine Creek Field, the complexity in the pore network, dictated by a framework of dolomitized burrows, will impact the migration of the plume and CO2 trapping. Further complexity—and potential mineral trapping—of CO2 might possibly occur as the result of anhydrite dissolution and subsequent calcite precipitation (e.g., Yu et al. 2019).

The risk of irregular and unpredictable lateral plume development due to complex regional permeability systems was recognized by Bachu et al. (2011), with conventional hydrocarbon production data providing evidence for widespread pressure communication (e.g., on the dolomitized Bashaw Platform of the Leduc Formation between the Clive and Bashaw fields; Tsang and Springer 1983). Such permeability pathways are difficult to identify from core or well logs, are unlikely to be large enough to show up on seismic, and can be inferred only from fluid and pressure interaction. This presents a risk to lateral containment, but does not imply a risk to vertical containment. Nevertheless, the strong facies control on pore type and volume identified in this study indicates that facies-controlled heterogeneity on porosity and permeability could be predicted in and between fields using seismic stratigraphy, detailed facies mapping, and paleoenvironmental interpretation and stratigraphic forward modeling. Such an approach, coupled with existing knowledge of facies distribution from the extensive prior study of the WCSB, should improve confidence in identifying the likely presence of moldic and vuggy porosity in reefal facies. This would also improve knowledge of the presence, distribution, and length scales of low-permeability layers that could be beneficial in preventing the buoyant CO2 from forming an expanded plume in the topmost layers of reservoirs (Al-Khdheeawi et al. 2017).

Other Risks and Uncertainties Associated with CO2 Injection and Storage in the Devonian Carbonates of the WCSB

Although the presence of calcite- and dolomite-cemented breccias in certain reservoirs (e.g., Eaglesham North Field) is not considered to significantly impact porosity and permeability at the field scale, their presence is important because they may indicate the presence of sub-seismic-scale faults and fracture systems. Experience from actively monitored EOR development of the Mississippian Midale Beds at Weyburn, Saskatchewan, has demonstrated that a systematic and/or large-scale fracture system will preferentially mobilize injected CO2 along the direction of open fracture sets. This is particularly problematic in thin-bedded carbonates such as the Midale target beds of the field; however, with sufficient characterization of the fracture network, the operators have reduced the impact with suitable well design (e.g., Whittaker et al. 2011; Cavanagh and Rostron 2013; Elkhoury et al. 2013; Uddin et al. 2013).

In addition to fracture network characterization, it is also important to consider the impact of injecting high-pressure fluids in the vicinity of faults, as recent work (Schultz et al. 2016; Zhang et al. 2016; Corlett et al. 2018) has shown that wastewater injection and hydraulic fracturing in Devonian strata in west-central Alberta (Kaybob Field) has caused the reactivation of Precambrian basement-rooted faults, leading to numerous instances of induced seismicity. Additionally, later work has shown that fault reactivation can also be associated with secondary recovery in conventional Leduc Formation reservoirs (Schultz and Wang 2020). Although it is highly unlikely that CO2 could escape to post-Devonian aquifers along faults because fault tips are effectively sealed by Upper Devonian and Mississippian strata (Corlett et al. 2018), it is still necessary to mitigate the risk of fault reactivation. Pawley et al. (2018) and Schultz and Pawley (2019) suggested that the following factors are the strongest predictors of geological susceptibility to induced seismicity in the Devonian strata of the WCSB: the proximity of the injection zone to the Precambrian basement, in-situ stress, formation overpressure, proximity to fossil reef margins, formation-water lithium concentration (higher in the vicinity of basement-rooted faults) and the rate of natural seismicity. Additionally, recent work (Stacey et al. 2021a, and references therein) has shown that basement-rooted faults can be identified not only by elevated formation water lithium concentrations but also by the occurrence of radiogenic strontium, the presence of saddle dolomite cemented breccias, and the co-occurrence of authigenic quartz and albite with dolomite. These findings may prove useful for the placement of wells in future CCUS projects.

Integration of qualitative observations (sedimentology, diagenesis, and pore morphology and distribution) with quantitative data (legacy core-plug porosity/permeability measurements and production data) reveals that the potential CO2 storage capacity and plume migration in the Devonian carbonate strata of the WCSB are controlled by the following factors (Fig. 10):

  • Dolomitization is the principal control on the suitability of reservoirs for CO2 storage in the WCSB. Dolomitization created well connected intercrystalline, moldic, and vuggy pore networks that would optimize injection, plume migration, and storage of CO2. Facies with high initial porosity acted as fluid pathways during burial diagenesis, resulting in their preferential dolomitization. This process caused solution enhancement of pre-existing pores and created volume-reduction-related porosity, forming dolostone that is generally far more porous and permeable than facies-equivalent limestone.

  • Depositional facies also exert a strong control on pore morphology, distribution, and interconnectivity. Porosity is principally controlled by the relative abundance of skeletal grains (e.g., particularly stromatoporoids in reefal facies of the Swan Hills and Leduc formations) and by the presence of burrows (e.g., heavily bioturbated intervals of the Wabamun Group).

  • Burial cements (e.g., calcite and dolomite) are volumetrically insignificant and only partially fill pores. Exceptions to this include porosity-occluding cements associated with fractures and breccias in the vicinity of faults.

  • Fracture densities are broadly comparable in limestone and dolostone, suggesting that this is mainly controlled by depositional facies rather than mineralogy. Fracture-permeability is most important in fine-grained facies, as fractures facilitate flow between otherwise poorly connected pore networks.

Taken as a whole, these findings indicate that the risk of unplanned vertical and lateral CO2 migration is low in reservoirs hosted within reefal facies (e.g., the Caroline and Acheson Fields of the Swan Hills and Leduc formations), because these contain both depositional (e.g., transitions to less porous lagoonal facies) and diagenetic (e.g., dolostone–limestone contacts) baffles to flow. In contrast, the risk of plume migration is high in heavily fractured and brecciated reservoirs, such as those that are directly crosscut by deep-seated faults (e.g., the Eaglesham North Field of the Wabamun Group). Despite this, the risk of CO2 escaping these reservoirs is low, because the Devonian strata of the WCSB are effectively sealed by overlying shale.

The results of this study demonstrate that drill-core analysis, together with integrated legacy core-plug and production data, can provide valuable insights into the principal petrophysical factors (e.g., porosity and permeability) that control storage capacity and plume migration in reservoirs that are prospective for CCUS. Wider application of this approach may prove useful for evaluating carbonate reservoirs for CCUS in other mature petroleum provinces.

Supplemental material is available from the SEPM Data Archive: https://www.sepm.org/supplemental-materials.

The work contained in this study forms a section of a Ph.D. project undertaken as part of the Natural Environment Research Council Centre for Doctoral Training (CDT) in Oil and Gas (grant number NE/M00578X/1) under its Extending the Life of Mature Basins research theme. It is fully funded by NERC, with additional funds from the American Association of Petroleum Geologists Grants-in-Aid General Fund. We are grateful to the Alberta Energy Regulator for providing access to the drill cores, digital well data, and maps used in this study, and to Lukus Wagstaff and Susan Co for their assistance and guidance during core sampling. We thank Stephen Kaczmarek and John Southard for editorial handling and associate editor John Rivers, Maxwell Pommer, and three anonymous reviewers for comments and critiques that significantly improved the manuscript.

Open Access CC-BY 4.0

Supplementary data