Impact of reservoir quality on the carbon storage potential of the Bunter Sandstone Formation, Southern North Sea Open Access

As part of the North Sea Transition Deal, the UK government has committed to funding at least two carbon capture, utilization and storage (CCUS) blocks in the East Irish Sea and Southern North Sea (UK SNS) of the UK Continental Shelf (UKCS) by the mid-2020s, with at least two further schemes to be in development by 2030 as efficiencies and knowledge increase and operations costs decrease (Department of Business, Energy & Industrial Strategy 2021, 2022). The Lower Triassic Bunter Sandstone Formation has been established as a key reservoir interval that may facilitate carbon storage in the UK SNS (Brook et al. 2003; Bentham 2006; E.ON 2011; James et al. 2016; National Grid 2016). Logistically, the offshore Bunter Sandstone Formation faces the Tees and Humberside industrial cluster and is located beneath shallow waters (20–30 m). The Formation has been penetrated by numerous exploration and production wells targeting Triassic, but mostly deeper Permian and Carboniferous gas fields. Additionally, wide geographic seismic imaging coverage of the UK SNS since the 1970s has resulted in wide, high-quality imaging of both the reservoir and extensive sealing stratigraphy above and beneath it.

As of 2023, the North Sea Transition Authority (NSTA) have issued six dedicated CCUS licenses across the UKCS, four of which are situated in the UK SNS (NSTA 2012, amended in 2020; NSTA 2021; NSTA 2022a, b) (Fig. 1). Twenty-one further license awards have been issued in September 2023, fourteen of which are located in the UK SNS (NSTA 2023). The Northern Endurance Partnership (NEP) between BP, Equinor and Total, aims to develop a Lower Triassic Bunter Sandstone Formation reservoir as a CCUS facility that will serve Net Zero Teeside (NZT) and the Humberside East Coast Cluster (ECC) industrial hubs (BP 2021). These industrial hubs collectively account for greenhouse gas emissions of 6.1 and 12.5 MtCO₂ per annum (as of 2018) respectively; around 45% of the UK's annual industrial emissions (Climate Change Committee 2020). NEP have acquired three license areas in the Silverpit Basin of the UK SNS. The first, CS001 covers the Endurance prospect (Fig. 1): a four-way structural closure estimated to have a theoretical CO₂ storage capacity of c. 2700 MtCO₂ (Gluyas and Bagudu 2020). Prior front-end engineering and development (FEED) strategies developed by National Grid (2016) had proposed a CO₂ injection plan that would utilize 2% of this capacity over an initial 20-year injection phase. NEP have also acquired licenses CS006 and CS007, partially covering UKCS quads 43/23, 43/30, 44/26, and 44/27 (NSTA 2022a, b) (Fig. 1). This license area incorporates a Bunter Sandstone Formation structural closure which received previous attention during a CCUS UK Government-led initiative to develop the UKCS as a CCUS hub, though funding for this scheme was reneged on in 2015. Bunter Closure ‘36’ (Fig. 1) was the focus of a FEED plan led by Pale Blue Dot and the Energy Technologies Institute (James et al. 2016). Harbour Energy (formerly Chrysaor) were awarded CS005, aiming to redevelop depleted Permian gas fields as part of their Viking CCS programme, which lie structurally and stratigraphically beneath the Bunter Sandstone Formation, present within the license area (Fig. 1). Viking CCS was selected as part of the track-2 CCUS phase by the UK Government in 2023 (Viking CCS 2023).

The Bunter Sandstone Formation has thus become potentially valuable resource for both the UK's Net Zero strategy, and for the rejuvenation of the mature UK SNS gas province. It forms the main post-Permian reservoir facies above mudstones of the underlying Bunter Shale Formation and evaporites and marginal carbonates and mudstones of the Late Permian Zechstein Group. The immediately overlying Rot Clay and Rot Halite Members of the Dowsing Formation, the lowest formation of the Middle Triassic Haisborough Group, have demonstrably shown to be effective stratigraphic seals in six Bunter Sandstone Formation-hosted gas fields in the UK SNS (Cooke-Yarborough 1991; Ketter 1991; Ritchie and Pratsides 1993). The Bunter Sandstone Formation, however, is not a prolific gas reservoir due to its general isolation from Carboniferous source rocks by the Zechstein Group and Lower Triassic Bunter Shale Formation (Underhill 2009). This is despite the occurrence of several periclinal structural closures throughout the UK SNS that developed during protracted phases of Zechstein Group halokinetic doming between the late Early Triassic and Cenozoic (Underhill 2003). Most of these closures exist as saline aquifers (Bentham 2006), which were penetrated by exploration wells during the 1960s and 1970s and found to be non-hydrocarbon-bearing. The Bunter Sandstone Formation has therefore not benefitted from the same focus of hydrocarbon exploration as the gas-rich Permian Rotliegend Group and Carboniferous clastics. Consequently, there is a relative lack of wireline logging and core recovery compared with sub-salt strata in the Southern North Sea. However, since the turn of the millennium and growing appetite for the implementation of CCUS in the wake of the climate crisis, it has garnered increasing academic and corporate interest as a potential geological carbon storage reservoir in the Southern North Sea.

Recent academic CCUS research on the Bunter Sandstone Formation has focused on determining theoretical CO₂ storage and pressure capacity factors of the many structural closures populating the Southern North Sea, usually at a regional scale (Heinemann et al. 2012; Noy et al. 2012; Bentham et al. 2017; de Jonge-Anderson et al. 2022; Hollinsworth et al. 2022). Of these, saline aquifers are suggested to have a collective theoretical storage capacity fivefold that of known Triassic gas fields (Bentham 2006). Various potential CO₂ containment risks such as fault occurrence and severity, chemical reactions resulting from CO₂ injection, formation water dispersal and potential historic well integrity concerns have also been explored both with regard to specific structural closures and in a general sense (Brook et al. 2003; Bentham et al. 2013; de Jonge-Anderson et al. 2022; Hollinsworth et al. 2022). Pioneering reservoir analyses have been factored into some CCUS-focused research, including reservoir heterogeneity effects on CO₂ plume dispersal, and potential effects of permeability barriers on injection pressure capacity and formation water displacement (Brook et al. 2003; Bentham et al. 2017). However, many studies average out or do not consider known geological and diagenetic heterogeneities that affect the reservoir properties of the Bunter Sandstone Formation. Such factors include heterolithic and shale horizons and significant zones of halite plugging, the latter presenting production challenges for gas fields of the Silverpit Basin (Ketter 1991; Ritchie and Pratsides 1993). Reservoir characteristic records, including sedimentology, lithofacies, and diagenesis analyses, and petrophysical analyses such as porosity and permeability experiments are limited to gas field overview publications (Ketter 1991; Ritchie and Pratsides 1993), and legacy data available from the UK National Data Repository. Furthermore, released reports from a cancelled UK government CCS competition provide a wealth of legacy reservoir data for Endurance (National Grid 2016) and saline aquifer ‘Bunter Closure 36’ (James et al. 2016), including: seismic and well log analyses, core porosity experiment results (in the case of Endurance), modelling electrofacies, and reservoir engineering modelling. Given the initial awarding of acreage for pioneering CCUS schemes, detailed reservoir characterization of the Bunter Sandstone Formation is warranted to bridge knowledge gaps between well-studied and understudied structural closures. Such research will benefit the UK Government's ambitions to develop further CCUS prospects as operation costs reduce and efficiency increases.

The purpose of this work was to develop a reservoir characterization model for the Bunter Sandstone Formation between the East Midlands Shelf in the west, and the Silverpit Basin in the east of the northern sector of the UK SNS. This area hosts thirty-one structural closures mapped as part of this study hosting the Bunter Sandstone Formation, including Endurance, Bunter Closure 36, the Esmond, Forbes, Gordon, Caister Bunter and Hunter gas fields, and several saline aquifers that face the Humber and Teeside industrial clusters, including within the CS005 license area of the Viking CCS project. Despite the discovery of gas fields made within the Silverpit Basin in the NE of the UK SNS sector, Triassic exploration has been disappointing as a result of absence of gas charge (possibly a result of isolation from Carboniferous source rocks by the Zechstein Group evaporites (Underhill 2009)). As a result, core data is currently scarce. However, many well penetrations have underwent wireline logging as part of drilling campaigns exploring sub-Zechstein prospects. Thus, a petrophysical approach is the ideal method by which to characterize reservoir quality across the UK SNS, as part of renewed academic and industry interest in the Bunter Sandstone Formation as a potential CO₂ store. Using publicly available well data and historical records from core porosity measurements, composite log and well completion reports, workflows for predicting petrophysical properties, electrofacies models, and intra-reservoir stratigraphic correlations across the northern sector of the UK SNS are developed. This aims to highlight ‘sweet spots’ where the Bunter Sandstone Formation boasts excellent porosity and is less affected by halite plugging and heterolithic or shale intervals, and equally as the latter may present challenges to CCUS. Due to the lack of abundance of well data logging the Bunter Sandstone Formation, intra-formational stratigraphic correlations are employed to assist with predicting the lateral extent of potential baffles to CO₂ vertical plume dispersal, as well as correlating high quality reservoir zones. Site-specific reservoir characterization forms the basis for theoretical storage capacity estimation of thirty-one Bunter Sandstone Formation closures. Such detailed reservoir characterization hopes to serve as a methodology for future site consideration for CCUS in the UK SNS and help develop and refine existing CO₂ injectivity models and theoretical storage capacity estimates of these sites.

Digital well data was acquired from the UK National Data Repository, hosted by the North Sea Transition Authority. Joined well-log files available exploration wells and selected appraisal and production wells were loaded into Schlumberger's TechLog wellbore software platform. Wells that did not log the Lower Triassic Bacton Group and the lowermost members of the Haisborough Group, the Rot Halite Member, and the Rot Clay Member, were discounted. Well datasets were catalogued according to available data of the Bunter Sandstone Formation to determine a series of petrophysical modelling workflows. Three well-groups were created for batch workflows; the first group was populated with wells with full coverage of gamma-ray (GR), compressional slowness (DT), bulk density (RHOB), and neutron porosity (NPHI) data; the second group consisted of wells with complete GR, DT, and RHOB logs, and a third group consisted of wells with GR and DT data.

The first well group workflow involved creating a V_shale model (vol% of shale) using well-specific NPHI and RHOB cross plots to inform effective porosity (⁠$Φ$_eff) calculations. This was achieved by tailoring NPHI and RHOB shale cutoff values within the Bunter Sandstone Formation to the average values of the stratigraphically adjacent Bunter Shale Formation. Resulting V_shale values were then input into Techlog's effective porosity model to calculate both total and effective porosity (⁠$Φ$_eff). Calculated $Φ$_eff values, where possible, were compared with core porosity data, scarce for the Bunter Sandstone Formation, to test the accuracy of this modelling. The second and third well groups used a GR-based V_shale model. In a similar fashion, matrix cutoff and shale max values were determined per each well's intrinsic log data, opposed to employing a blanket uniform cutoff across all wells. For these groups, $Φ$_eff was modelled using DT logs (sonic porosity). Matrix values between 47.5–55 µs/ft were used, whilst shale sonic cutoff values were selected from representative values of intra-formational shales within the Bunter Sandstone Formation reservoir, or alternatively from the Bunter Shale Formation.

Electrofacies modelling within the Bunter Sandstone Formation was achieved by inputting available log data into Techlog's IPSOM geological classification algorithm. This 2D indexed and probabilized self-organizing map allows for comparisons of each log parameter to be made with the capability to apply different parametric weightings, with a view to establishing a series of designated groups that reflect common petrophysical characteristics. Generally, DT was weighted at c. 43–55%, RHOB between 35–43%, GR between 2–21%, and NPHI at c. 2%. This weighting distribution was preferred due to the importance of salt-plugging and cementation in determining Bunter Sandstone Formation porosity, and the high net-to-gross of the Formation. Weightings for parameters were calibrated by running several iterations of IPSOM models and comparing the output classifications on RHOB-acoustic impedance (AI) cross-plots to validate groupings. The weighting factor of GR logs varied more significantly than other parameters to better delineate between low porosity sandstones and mudstones that may show similar DT and RHOB signatures. These were then calibrated with composite log reports with lithological descriptions of cuttings. Four electrofacies were established using the IPSOM modelling that are widely applicable for the Bunter Sandstone Formation well datasets across the northern sector of the UK SNS. Electrofacies were grouped into three reservoir classifications, represented by high-quality, medium-quality, and low-quality reservoir, and one non-reservoir classification representing intraformational shale horizons of the Bunter Sandstone Formation.

Establishing these electrofacies zones allowed for the calculation of net-to-gross (NTG) and $Φ$_eff of each group that informed point-data mapping of these important reservoir parameters that inform theoretical CO₂ storage capacity calculations. $Φ$_eff and NTG maps were produced from point data associated with corresponding wells using a standard convergent gridding method in Schlumberger's Petrel software.

To determine theoretical storage capacity of the Bunter Sandstone Formation, it was necessary to convert mapped seismic two-way-time interfaces to depth. This allowed for reservoir volumetric calculations and the determination of structural closure depth which is a crucial parameter in calculating CO₂ phase behaviour. Depth conversion was achieved by creating a layer-based velocity model for designated velocity groups using V₀ and k functions, where V₀ represents the instantaneous velocity at the datum point (sea surface), and k represents the velocity gradient of the velocity interval. We employ a similar methodology to that of Hollinsworth et al. (2022), though in this study static velocity functions representative of local layer velocity signatures at each structural closure are employed, rather than extraction of functions from a generated 2D grid. This approach was favoured due to the site-specific approach of this study, opposed to the wider regional depth-conversion employed in prior studies. Layers and velocity functions are accessible from Appendix A.

Monte Carlo simulations were run for NPV values for each individual Bunter Sandstone Formation closure, using tailored parameter ranges in each case. A ±10% normal distribution range was used for GRV, $Φ$_eff ranges for each closure were determined from on the average $Φ$_eff values calculated for wells proximal to each closure, with a ±5% normal distribution range. A ±5% normal distribution range of average NTG values was applied, capped at 100%. S_wirr value ranges were determined from a range of values linked to expected range for the average $Φ$_eff of each closure (Fig. 2). These were analysed using a triangular distribution with the most likely value corresponding with a constant of 0.3 and coefficient of $Φ$¹_eff, the minimum value corresponding with a constant of 0.2 and coefficient of $Φ$^0.8_eff, and a maximum value of 0.4 and coefficient of $Φ$^1.2_eff.

Reservoir mapping and petrophysical analysis was conducted across UKCS quadrants 42–44 and 47–49. This area incorporates the East Midlands Shelf, the Silverpit Basin, and the Indefatigable Shelf, and hosts the CCUS licenses CS001, CS005, CS006, and CS007 (Fig. 3). Thirty periclinal structural closures, including five Triassic gas fields (Esmond, Forbes, Gordon, Hunter, Caister Bunter) and twenty-five saline aquifers are identified as potential CCUS targets within the area of study (Fig. 3). Where referring to depleted gas fields or saline aquifers that may be potential geological CO₂ storage sites, we employ a naming convention using the standard UKCS nomenclature of quadrant/ block, the acronym CS (carbon-store), and sequential numbering per closure within this block, e.g. a CCUS prospect with a top-crest depth within 43/21 would be 43/21-CS1 (Fig. 3).

The total thickness distribution of the Bunter Sandstone Formation in the UK SNS is shown in Figure 4 (adapted from Underhill et al. 2023) and Figure 5. The Formation is thickest in the hangingwall of the Dowsing Fault Zone, and to the south of the Flamborough Head Fault Zone with a maximum thickness of approximately 350 m (Figs 4, 5). Thinning occurs to both the SW and north through east onto the Dogger shelf (north) Cleaverbank high (where the base-Cretaceous unconformity truncates Lower Triassic strata), and SW toward the London-Brabant Massif palaeohigh.

Petrophysical analysis of 96 wells within the area of study outlined in Figure 3 was undertaken to analysis variation in effective reservoir porosity and net-to-gross. Results are presented here from $Φ$_eff modelling, two V_shale models using either GR or RHOB-NPHI input parameters depending on log availability, analysis of acoustic impedance (AI) and RHOB cross-plots, and electrofacies classification using TechLog's IPSOM classification tool. An $Φ$_eff model was tested against core porosity data from eight wells in the Silverpit Basin area, from the Esmond, Caister Bunter and Hunter gas fields, and the Endurance, 43/12-CS1 and 43/18-CS1 saline aquifers. Core porosity data for the Bunter Sandstone Formation is scarce as it has never proven to be a prolific hydrocarbon reservoir. Using a RHOB-NPHI and V_shale model, our results show a strong correlation with empirical helium porosity measurements from various cores (Fig. 6). This model tends to underestimate porosity where V_shale calculations are high, particularly in wells 44/23a-10 and 44/23-3 (Fig. 6). Well 43/13a-C2, a production well of the Esmond gas field, shows three instances 1815, 1850, and 1900 m MD where the modelled $Φ$_eff curve overestimates core porosity. Notably wells that show greater divergence from $Φ$_eff models are gas-bearing, whereas saline aquifer wells (42/25d-3, 43/12-1, 43/18-1) show much stronger correlations. This is likely due to an artificial increase in $Φ$_eff resulting from low density gas occupying pore space, or possibly due to gas charge preserving porosity.

$Φ$_eff models show correlations with acoustic impedance (AI) and GR anomalies within the Bunter Sandstone Formation. AI logs were generated using the equation AI = V×RHOB where V is sonic velocity, calculated from the inverse of sonic transit time (DT). $Φ$_eff and AI logs exhibit a consistent inverse relationship, while high GR-spikes >75 gAPI also coincide with low $Φ$_eff values (c. 0–10%) (Fig. 6). High AI spikes >10 000 kPa s m⁻¹ also coincide with low $Φ$_eff of 0–5%, and notably have extremely low GR, and hence V_shale values. Zones of high porosity (approx. 22–35%) are associated with sandstones with low GR values (35–50 gAPI), and low AI values (6000–7000 kPa s m⁻¹ ). Horizons with distinct, but not strong AI peaks between 7000–10 000 kPa s m⁻¹ and low GR signatures often correspond with intermediate porosity values approx. 15–20%.

Reservoir quality within Bunter Sandstone Formation closures is known to be highly variable (Ketter 1991; Ritchie and Pratsides; Brook et al. 2003; Bentham 2006; Gluyas and Bagudu 2020), yet prior regional studies have tended to average values such as porosity and net-to-gross (Brook et al. 2003; Bentham 2006). High porosity sandstones (20–35%) (Brook et al. 2003) interbedded with low porosity mudstone horizons and severely halite-cemented zones that have presented historic challenges in the hydrocarbon industry to recovery and edge- and bottom-water drive (Ketter 1991; Ritchie and Pratsides 1993). Compartmentalisation and challenges to vertical permeability caused by thin low-porosity layers have also been highlighted as challenges to implementing CCUS (Sutherland et al. 2022). To better characterize reservoir heterogeneity within the Bunter Sandstone Formation at local scales, acoustic impedance – bulk density (AI-RHOB) cross-plots were generated from well-log data points to highlight potential facies zones.

AI-RHOB cross-plots for the Bunter Sandstone Formation show a range of well-log data points that represent sandstone of variable porosity (from <5% $Φ$_eff to approx. 34% $Φ$_eff), as well as those which do not fit the expected AI-RHOB relationship for sandstone (Fig. 7). When sorted by depth, the position of these data points highlights the depth-occurrence of specific horizons and anomalies that do not correspond with reservoir sandstone facies. Data points that record higher RHOB values for a given AI often correspond with shale horizons when cross-referenced with Vshale models, $Φ$_eff models, and primary composite log records. Mudstones within the Bunter Sandstone Formation have typical gAPI values between 75–100, comparable with the underlying Bunter Shale Formation values. These often correlate with the gradational, muddy base of the Bunter Sandstone Formation with the Bunter Shale Formation, and with a prominent shale horizon near the top of the Formation at the eastern margin of the Silverpit Basin (e.g. Fig. 7e and h–j). The other prominent deviation from an expected sandstone relationship is where data points tend toward much lower RHOB values for a given AI (e.g. Fig. 7d–h). Often, this corresponds with layers of salt-plugging within the uppermost 30 m of the Bunter Sandstone Formation when cross-referenced with composite log data and seismic interface interpretation. Halite-plugging is a widely recognized phenomenon at the top Bunter Sandstone Formation and is concurrent with a seismic polarity anomaly corresponding with a downward increase in AI through the Rot Clay Member (Gluyas and Bagudu 2020; Hollinsworth et al. 2022; Underhill et al. 2023). The effect of halite (density 2.17 g cm⁻³) cementation of high porosity sandstone on the AI-RHOB plot is to increase the overall density and impedance of the rock by replacing water- or methane-filled pores, but to a lesser degree than quartz (2.65 g cm⁻³) cementation. Halite cementation should therefore be distinguishable both from quartz-cemented, or indeed fine-grained or compacted sandstones, summarized in Figure 8a–d. AI-RHOB values also show deviation from the expected values for mature sandstones in gas-charged Bunter Sandstone Formation well penetrations. Well 44/23-5 (Caister Bunter) was charged to a depth of approx. 1426 m (Texas Gas Exploration Corporation 1986a), and well 43/20-1 (Gordon) to a depth of 1646 m (Hamilton Brothers Oil Company 1970), and AI-RHOB values recorded within these gas columns tend to plot at lower than would be expected for a typical water-filled porous sandstone (Fig. 9j). This is likely a result of pore-filling methane having lower density (0.657 g cm⁻³) relative to water-filled pores. (Fig. 8e).

IPSOM classification of well data from the Bunter Sandstone Formation allows for the establishment of facies-characterization within the unit. Four electrofacies classifications were established for the well dataset based on sonic velocity (DT), bulk density (RHOB) and gamma-ray (GR). Wells with neutron porosity (NPHI) included this parameter in weighting for electrofacies classification, though it was not a determining factor. Electrofacies classifications and some typical petrophysical value ranges are shown in Table 1. Figure 9 shows electrofacies classifications plotted in AI-RHOB plots to demonstrate model results against parameters with a known relationship.

Ranges in $Φ$_eff do vary between wells, as electrofacies modelling was undertaken on a per-well basis, rather than as a batched suite of wells. This was because batch-modelling of parameters from multiple well logs resulted more scattered classification within individual wells that did not reflect original petrophysical parameter values. Creating electrofacies models on an individual well-basis allows tighter petrophysical constraint on facies classification.

Overlaying electrofacies onto well logs shows the downhole facies distribution of each well (Figs 10a, 11a). Reservoir electrofacies closely track raw DT values, which themselves have an inverse relationship with modelled $Φ$_eff values, a universal relationship across the study area. High-quality reservoir facies correspond to higher DT values (Figs 10b, 11b), and hence lower AI values, and that low quality reservoir typically has DT-values (approx. <85 µs/ ft for 44/26-3, Fig. 10) that record lower $Φ$_eff values (Figs 8, 9). This is dependent on DT values and $Φ$_eff of individual wells. Non-reservoir electrofacies are easily distinguished from their GR signatures (Figs 10, 11).

The spatial distribution of $Φ$_eff and net-to-gross (NTG) of the Bunter Sandstone Formation can be shown as 2D surface maps determined from point-data associated with each well, determined through convergent interpolation of well-point data in Schlumberger's Petrel software (Fig. 12). Average Bunter Sandstone Formation $Φ$_eff was calculated for the formation, including non-reservoir facies, which were discounted from CO₂ theoretical storage capacity models by factoring in NTG values in these calculations. Whilst static modelling the distribution of these parameters is rudimentary, based on point data from a relatively small number of wells relative to the large area of interest, it can highlight reservoir ‘sweet-spots’ and can be used in conjunction with seismic and well log data to highlight zones affected by halite-plugging and increased volumes of non-reservoir facies. The recognition of substantial variation in both $Φ$_eff and NTG across the UK SNS highlights the importance of localized wireline analysis informing CO₂ storage capacity modelling, as opposed to averaging regional values.

High porosity zones (>22%) often correspond with crests of Bunter Sandstone Formation closures in the Silverpit Basin, notably including Endurance, 44/26-CS1, and 44/27-CS1, all of which lie within CCUS licensed zones (Fig. 12a). Other areas of high average $Φ$_eff include the Cavendish ridge (blocks 43/19, and 43/20) and the Gordon gas field and area to the NE, and part of the Indefatigable Shelf that hosts UKCS CCUS license CS005 (Fig. 12a). Areas of low average $Φ$_eff include a zone NE of Endurance that trends from the NW the margin of the Dogger Shelf through to 44/26-CS1 in the SE. In this area, porosity in the Bunter Sandstone Formation is severely impacted by salt and calcareous cementation in its uppermost 30 m (Zapata International Corporation 1973; Lasmo North Sea 1984; Chevron 1994; National Grid 2016), though high-quality reservoir is present below this zone (e.g. Fig. 11). The distal east and NE margin of the Silverpit Basin also records low average $Φ$_eff owing to the increased volume of mudstone, including a 10–12 m claystone interbed near the top of the Formation (Hamilton Brothers Oil Company 1970a, b; Star Energy 2008). The other major areas of low average $Φ$_eff include the NW margin of the Silverpit basin along the Dogger Shelf, the hangingwall of the Dowsing Fault Zone, the NW sector of the adjacent Sole Pit Basin and Indefatigable Shelf (Fig. 12a). Average $Φ$_eff in these areas can be as low as 11%, and typically ranges between 11–15%. In the Sole Pit Basin, reservoir quality is affected by higher abundance of clay matrix as, as well as nodular anhydrite, and occasional halite stringers (BP Petroleum Development 1965) and is very fine-fine grained in places (Amerada Hess 1990). The Formation becomes increasingly argillaceous and with greater presence of mudstone strata toward the Dowsing Fault Zone (Arpet Petroleum 1966). Abundant clay cement is also present in the low porosity (avg. 13%) 49/17-5 located on the Indefatigable Shelf (Continental Oil Company of England and National Coal Board (Exploration) Ltd. 1969b), though $Φ$_eff increases toward Bunter closures 49/17-CS1 and 49/17-CS2, in agreement with composite logs for wells 49/17-1 and 49/17-4, the latter reporting fair-good visible porosity between 15–20% (Continental Oil Company of England and National Coal Board (Exploration) Ltd. 1969b).

Comparisons of $Φ$_eff maps with seismic interpretation of the top approx. 30 m Bunter Sandstone Formation shows a relationship between acoustic properties and $Φ$_eff values of this interval. Notably, the NSTA license area CS005 partially covers an area where the Bunter Sandstone Formation is encountered as a soft-AI interface, and exhibits average $Φ$_eff values approx. 25%, far greater than the zone to the NW of the Indefatigable Shelf, and to the east in the Sole Pit Basin (average $Φ$_eff = 10–15%) (Fig. 13). Wells both NE and SW of the Endurance structure and the NW–SE ridge that hosts it show markedly lower average $Φ$_eff (≤15%) and correspond with hard-AI top Bunter Sandstone Formation interfaces. Endurance, 44/26-CS1 and 44/27-CS1 have higher average $Φ$_eff (≤25%) and are situated within soft-AI top Bunter Sandstone Formation zones (Fig. 13). There are some anomalies to this trend include 43/23-CS1 (Fig. 11), which lies within a hard-AI top interface zone but records moderate average $Φ$_eff values of 18% (Fig. 13). This is due to the presence of significant high-quality reservoir volumes beneath the uppermost salt-plugged intervals in this location. Caister Bunter gas field (44/23-CS2) also records high $Φ$_eff ranges of 19–24% within an area characterized by a hard-AI top reflector (Fig. 13). This is likely because the presence of pore-filling methane inhibited the formation of halite and anhydrite cements (Ritchie and Pratsides 1993).

Net-to-Gross maps show that the Bunter Sandstone Formation boasts high sand volumes (>80%) across most of the northern sector of the UK SNS (Fig. 11b). An average 64% of the Formation is composed of medium- to high-quality sandstone with porosity 15–30%. The distal NE and eastern zones of the Silverpit basin record a gradual increase in heterolithic and mudstone facies, in part toward the top of the Bunter Sandstone Formation, where a 10–15 m thick claystone is present (Hamilton Brothers Oil Company 1970a, b; Star Energy 2008) as far west as 44/26-CS1 (Shell U.K. Exploration and Production 1988). Lower NTG values (71–75%) also correspond with the Caister Bunter structure. The Formation develops a higher GR signature toward its top in this sector irrespective of the occurrence of this claystone horizon (e.g. Texas Gas Exploration 1983, 1986b). Two prominent high-GR anomalies are present on the Indefatigable Shelf within CCUS license block CS005. In well 49/17-4 these correspond with the presence of an abundant clay-matrix, as well as claystone interbeds comparable in composition to the overlying Rot Clay Member, though porosities up ≤20% are still recorded (Continental Oil Company of England and National Coal Board (Exploration) Ltd. 1969a, b).

Theoretical storage capacity probabilities for P10, P50, and P90 cases derived from monte carlo modelling are shown in Table 2. The largest closure by volume is 48/06-CS1. However, the proximity of the structure's crests to the surface (185 m) suggests that injected supercritical CO₂ would a gaseous phase (considered to be shallower than 800 m), thus capable of occupying a far lesser volume and putting less buoyancy pressure on the cap rock. This is also the case for the fifth largest closure by volume, 43/30-CS2. 48/14-CS1 may be a substantial closure with >1000 MCO₂t in a P90 scenario, however, with a top structure depth of 820 mTVDSS, it is possible that part of this closure may lie beneath the supercritical phase boundary for CO₂. Furthermore, the Haisborough Group directly overlying the Bunter Sandstone Formation has undergone significant domino-style normal faulting across the crest of this structure, posing a structural containment risk for CO₂ injection.

There are three large closures with theoretical storage capacities greater than 1000 MCO₂t where top structure depths lie within the supercritical field for CO₂. These are 43/21-CS1 (Endurance), 43/30-CS1, an undrilled closure forming part of the CS006 license block, and 44/26-CS1, forming part of the CS007 license block. 44/27-CS1 may also qualify as a large structure (>1000 MCO₂t), though P50 estimates equate to 980 MCO₂t. Crucially, these larger closures are proximal to moderate (100–1000 MCO₂t) and small (<100 MCO₂t) Triassic closures.

Prior studies of Endurance indicate a theoretical storage capacity of 2700 MCO₂t, with a NPV of calculation indicating an availability of 4.6 × 10⁹ m³ of pore-space, and a s_wirr value of 0.15 (National Grid 2016). Our calculations underestimate this by a factor of approx. 1.4 in our P50 case (3.3 × 10⁹ m³), and by 1.2 in our P90 case (3.8 × 10⁹ m³). It should be noted that based on our CO₂ density calculations and velocity models, our estimates of CO₂ density (495 kg m⁻³) are lower than the average of 700 kg m⁻³ suggested by National Grid (2016). It is most likely that our estimates of CO₂ density produce conservative estimates of theoretical storage capacity as this number is derived from the expected CO₂ density at the top of the structure crest.

43/30-CS1 lies within the CS006 Crown Estate license block for CCUS. At present, there are no publicly available 3D seismic data, nor any publicly available well log data, as of 2023 no wells have penetrated the structure. Mapping of the top Bunter Sandstone Formation reflector suggests it lies within a hard top-interface zone, and that reservoir conditions may reflect nearby structures such as 43/23-CS1 more so than soft-top characterized structures like 44/26-CS1 to the SE. Theoretical storage capacity estimates are comparable with previous studies (Hollinsworth et al. 2022), though acquisition of seismic and petrophysical data is key to better constraining these estimates. Given that 43/23-CS1 also lies within the CS006, it is likely that this closure could form part of a cluster of CCUS prospects in the area. With modest theoretical storage capacity estimates (P50 = 360 MCO₂t), and good reservoir conditions beneath the uppermost salt-plugged intervals (Fig. 11), it may provide a useful satellite structure to volumetrically larger prospects. Problematically for 43/23-CS1, a mid-reservoir spill point at 1725 m TVDSS limits the potential of higher quality reservoir to be utilized toward the base of the formation.

44/26-CS1 lies to the SE of 43/30-CS1 and within the boundaries of the Crown Estate CS007 license area. A previous FEED studies of the structure incorporates a development plan within its storage capacity model around a plan to inject CO₂ to occupy a portion of the available NPV at modelled at various injection rates, resulting in lower storage capacity estimates (P50 = 364 Mt) than the theoretical capacity estimates presented here (James et al. 2016). Theoretical storage capacity estimates based on petrophysical models present a more optimistic prospect than previous studies (Hollinsworth et al. 2022), owing to locally high $Φ$_eff recorded by wells in this sector of the Silverpit Basin, and the positive effect this has on S_wirr.

44/27-CS1 neighbours 44/26-CS1, situated directly adjacent to the NE. 44/27-CS1 is a dual-crested closure with a NW and SE culmination bridged by a saddle. Only the NW culmination has been penetrated by exploration wells (44/21-1) and nearby production wells for the Carboniferous–hosted Boulton gas field. Predicting reservoir conditions that are representative of the overall closure are challenged by the wrapping of a seismic-polarity reversal around the NW and NE flank of the closure (Fig. 13). The majority of the crestal section and the southern flank have reservoir conditions comparable to 44/26-CS1, but whilst petrophysical parameters from nearby wells (44/27-1) were used to guide storage capacity models, uncertainty remains as to the reservoir conditions over the saddle and southeastern culmination of the structure. The P50 theoretical storage capacity estimates here are a factor of 1.5 times less than previous estimates with modelling input parameters that are regionally based (Hollinsworth et al. 2022).

Two closures in UKCS block 49/17 also form part of the current CCUS license portfolio in the UK SNS. The Crown Estate CS005 license, which includes closures 49/17-CS1 and 49/17-CS2 within its boundary, is held by Harbour Energy through their Viking CCS venture, which seeks to store 10 MCO₂t by 2030 within the depleted, sub-salt Viking gas fields (Viking CCS 2023). 49/17-CS2, the larger of these closures (GRV 5 × 10⁹ m³) has a P50 theoretical storage capacity of 341 MCO₂t, a modest capacity compared to the larger prospects in the Silverpit Basin to the north (Table 2). A smaller closure to the NW, 49/17-CS2 has a P50 theoretical storage capacity of 93 MCO₂t. Both closures are positioned within a soft top Bunter Sandstone Formation interface, directly adjacent to a distinct seismic polarity reversal to the west and north on the Indefatigable Shelf (Fig. 13). This indicates an absence of salt-plugging within the upper stratigraphic intervals of the Formation. Despite this, petrophysical analysis indicates that the Bunter Sandstone Formation has overall lower $Φ$_eff across the Indefatigable Shelf, typically ranging from 12–21% between wells, with an overall average of 15.9%. The nearby 49/06-CS1 to the north of the Viking CCS cluster boasts a theoretical storage capacity of 589 MCO₂t. This closure lies out with the CS005 license and does not form part of any Crown Estate provincial license offerings, and petrophysical modelling indicates poor reservoir conditions (average $Φ$_eff = 12–15%).

Several closures fall within provisional CCUS license areas offered by the NSTA in summer 2022 (NSTA 2022a, b) (Table 2). The Triassic prospects in these provisional license blocks are of small <100 MCO₂t to moderate size <600 MCO₂t and are located within the Silverpit Basin. The largest of these closures is the depleted Caister Bunter gas field with an estimated theoretical storage capacity of 579 MCO₂t (P50 case). This closure was underfilled by a gas charge either due to continuing post-charge halokinesis (Ritchie and Pratsides 1993), or due to a transient connection to Carboniferous source rocks, possibly a result of dyke emplacement (c.f. Underhill 2009). Reservoir pressure conditions are presently unknown. The presence of significant halite-cemented horizons has inhibited the pressure recovery of Triassic gas fields upon hydrocarbon extraction during production, but there is evidence for longer-term pressure recovery in the Esmond field (Bentham et al. 2017). The pressure conditions within the Bunter Sandstone Formation may have severe implications for CO₂ injection rates and pressure capacity depending on local permeability conditions (Brook et al. 2003). Other provisional license areas in the Silverpit Basin cover the area north and east of Endurance. Closures including 43/16-CS1, 43/17-CS1, 43/18-CS1, and 43/19-CS1. Barring the volumetrically small 43/19-CS1, these structures each have moderate theoretical storage capacities between 160–250 MCO₂t (P50 cases). Previous studies have identified severe structural risks associated with crestal faults cross-cutting 43/17-CS1 (Bentham et al. 2013; Hollinsworth et al. 2022). Additionally, the footwall of this structure sits at 685 m TVDSS, and therefore theoretical capacity is impacted by the likelihood of CO₂ existing in a gaseous phase within it. The good reservoir conditions and proximity of other moderately-sized closures may make these attractive prospects for satellite structures upon the establishment of pioneering schemes focused on developing neighbouring larger structures.

The UK SNS hosts other moderate and minor structures (Table 2). Notable closures include the Esmond, Forbes, and Gordon cluster, which together have a combined theoretical storage capacity of approx. 760 MCO₂t. Despite the proven retention of methane, licensing of windfarm acreage to RWE Renewables as part of the Crown Estate's round four projects scheme means that it is unlikely these will form part of the UKCS CCUS portfolio (The Crown Estate 2019). Other structures include 47/03-CS1 proximal to the York and Rough fields on the East Midlands Shelf. However, structural risks have been posited due to the occurrence of domino-style normal faulting in the sealing stratigraphy of the Bunter Sandstone Formation demarcating the closures northern flank (de Jonge-Anderson et al. 2022). Closures including 43/30-CS2, 48/01-CS1, and 48/04-CS2 with moderate storage capacities are positioned either side of a major salt wall that demarcates the boundary between the Silverpit and Sole Pit basins. Reservoir quality is poor in the case of 48/08-CS1 and 48/04-CS1, which have average $Φ$_eff of 16 and 14% respectively, likely due to their location in the Dowsing Fault Zone hangingwall depocenter. Furthermore, there are structural risks associated with normal fault arrays in the overlying Haisborough Group, though these are minor faults that do not significantly offset or apparently penetrate the Bunter Sandstone Formation, instead terminating within the Rot Halite Member. 43/30-CS2 is a shallow-buried structure with a crest depth of 680 m TVDSS. CO₂ density, and the occurrence of flexural faults that pose severe risk to seal integrity make this structure an unlikely candidate for carbon storage (Hollinsworth et al. 2022). Closures on the Dogger Shelf (42/15-CS1 and 42/19-CS1) are hampered by poor reservoir conditions in this section of the UK SNS, whilst closures 47/15-CS1 and 47/19-CS1 have unremarkable theoretical storage capacity with no nearby feasible large structures that they may form a satellite prospect for.

Stratigraphic correlation within the Bunter Sandstone Formation is challenged by the absence of biostratigraphic markers, laterally variable diagenesis, and a general lack of consistent heterolithic marker horizons. Nonetheless, in prior studies, intra-formational correlations have been determined across the Endurance structure (National Grid 2016; Gluyas and Bagudu 2020), and between Caister Bunter (44/23-CS2) and 44/26-CS1 (Williams et al. 2013), and over the Hewett gas field facing the Norfolk coast (E.ON 2011). Three major stratigraphic zones, L1, L2, and L3 respectively, have previously been identified by chemostratigraphic mineral characterization, each with two subdivisions (L1a, L1b, L2a, L2b, L3a, L3b) (National Grid 2016) (Fig. 14). The top surfaces of the intervals in the Silverpit Basin here are determined to be correlatable across the extent of the basin (Fig. 13). Attempts have been made to extend these correlations westward to the East Midlands Shelf, linking structures identified in Hollinsworth et al. (2022), and de Jonge-Anderson et al. (2022), and those that form part of recently licensed blocks (CS001, CS006, and CS007). Whilst the intervals L1, L2, and L3 were originally defined by chemostratigraphic signatures (National Grid 2016), well-log correlation in this study was guided by primary gamma and acoustic impedance (derived from sonic and density) log signatures, as well as electrofacies classifications, and cross-referenced in a strike and dip section grid to ensure surfaces translated coherently across the Silverpit Basin (Figs 15, 16). The potential for seismic horizon mapping is limited as many of the subdivisions fall beneath resolution of seismic volumes covering the Silverpit Basin (approx. 35 m) (c.f. de Jonge-Anderson et al. 2022; Hollinsworth et al. 2022). All three major stratigraphic intervals are present across the Silverpit Basin, though L1 and L2 thin onto the Dogger Shelf along the northern margin of the Silverpit Basin whilst the uppermost strata of L3 is truncated by the Hardegsen disconformity over Endurance (Fig. 14). The general petrophysical and lithological properties (derived from electrofacies) of these intervals and their subdivisions are described herein.

Subdivision L1a forms the lowest stratigraphic zone of the Bunter Sandstone Formation. It records a gradual mud-to-sand facies transition across the Bunter Shale Formation into the Bunter Sandstone Formation, highlighted by the occurrence of several high GR spikes that are correlatable across the Silverpit Basin, along with lower reservoir quality in sand facies relative to stratigraphically higher zones. The NTG of L1a is the lowest of the six Bunter Sandstone Formation subdivisions, averaging 70%. Heterolithic facies are more extensive toward the distal east of the basin, increasing from approx. 20% over the East Midlands Shelf and Endurance to approx. 50% toward the basin's eastern margin (Fig. 15). L1a undergoes substantial eastward thinning from approx. 150 m over the East Midlands shelf to approx. 10 m over the Esmond and Forbes, and Caister Bunter and Hunter gas fields, separated by a 60 m thick zone over the Gordon gas field (Fig. 16). The subdivision also thins out onto the Dogger Shelf in the NW where it is not recorded in 43/11-1. Despite the prevalence of heterolithic and shale horizonsL1a records average $Φ$_eff of 17.5% (Fig. 17).

Subdivision L1b is composed of an average of 48% high-quality reservoir sandstone with 94% NTG. Reservoir quality is highest over 44/26-CS1 and 44/27-CS1. There is a general deterioration in reservoir quality toward distal east and NE areas. Several thin correlated high AI and high GR intervals, corresponding with low reservoir quality are present in wells penetrating Endurance, Caister Bunter, Hunter, and Gordon toward the basin's eastern margin. Zones of high proportions of lower quality (≤70%) reservoir occur over the Gordon Field and north of 44/26-CS1 as reservoir conditions deteriorate eastward. The subdivision thins from approx. 90 m in the south of the basin northward onto the Dogger Shelf, and east toward the Dogger Fault Zone (Fig. 16). L1b records a $Φ$_eff range between 15.8–30%, with an average of 22.2% (Fig. 17). High porosity zones are associated with the crests of Endurance, 44/26-CS1 and 44/27-CS1, whilst deterioration in porosity occurs toward the distal NE, and north onto the Dogger Shelf (Fig. 17).

L2a is dominated by high-quality reservoir sands (average 49%) with only minor heterolithic and shale components (NTG = 95%) (Fig. 15). The latter are restricted to the Dogger Shelf NW of Endurance, and toward distal eastern zones. Typical of the wider Formation, L2a is punctuated by low-porosity, high AI horizons, many of which can be correlatable across the basin. These are most prevalent in northern and eastern zones, where the reservoir quality of the Bunter Sandstone Formation severely deteriorates onto the Dogger Shelf and the distal Silverpit Basin. High quality reservoir volumes in L2a are concentrated over a NW–SE trending structural ridge that hosts 43/18-CS1 and 44/27-CS1. The subdivision boasts its highest $Φ$_eff values of approx. 28–30% above the crests of Endurance, and 44/26-CS1 and 44/27-CS1. L2a records high $Φ$_eff between 13.9–30%, with an average of 21.8% (Fig. 17). The subdivision is shallowest along the Dogger Shelf edge (≤10 m), and eastern margin of the Silverpit basin (between 10–20 m), and uniformly thickens to between 80–90 m toward the hangingwall of the Dowsing Fault Zone (Fig. 16). This indicates the development of accommodation space linked to movement in the fault zone during deposition.

L2b forms the upper zone of the L2 subdivision of the Bunter Sandstone Formation and signifies changes in sediment distribution in the Silverpit Basin. L2b has a maximum thickness of approx. 45 m in blocks 44/23 and 44/26. The subdivision thins west and NW onto the East Midlands and Dogger Shelf, and east toward the Dogger Fault Zone to ≤20 m (Fig. 16). Wells 43/23-3 and 44/26-2 record five discrete high AI, low GR spikes between 2.5–5 m thick (Fig. 14), as well as one heterolithic horizon near the base of the subdivision (Fig. 15), together accounting for 48 and 41% of the subdivision, respectively. Despite this L2b records good-excellent $Φ$_eff across the basin away from the hangingwall of the Dowsing Fault Zone (Fig. 17). As a result, high-quality reservoir accounts for 40% of L2b on average; this increases to approx. 60% over structural crests including Endurance, 44/26-CS1, and 43/18-CS1, which boast $Φ$_eff ranges between 25–30% (Fig. 16).

L3a forms a thin (10–20 m thick) horizon above L2b (Fig. 16). The subdivision thins in the structural trough between the Endurance ridge and the ridge hosting closures 44/18-CS1 and 44/27-CS1, to the NE of 44/27-CS1, and south into the Dowsing Fault Zone (Fig. 16). Two discrete shale horizons are present within L3a toward the SE, north of the Outer Silverpit Fault Zone that form up to 60% of the subdivision in the SE over 44/26-CS1 (Figs 14, 15). These shale horizons persist eastward toward the eastern margin of the basin (Fig. 15). Low-quality reservoir facies dominate the central areas of the Silverpit Basin, and the Dogger Shelf to the north. High-quality reservoir zones within L3a are present over the East Midlands Shelf and Endurance (Fig. 14), and associated with the Esmond, Forbes, Gordon gas fields, and 44/27-CS1. Porosity distribution is wide in L3a, likely because of the geographically bimodal reservoir quality (Fig. 17). $Φ$_eff ranges between 9–28%, with an average of 18.9%. High porosity zones are associated with structural crests, including Endurance, the Esmond, Forbes, and Gordon gas fields, and 44/26-CS1 (Fig. 17). Low porosity zones correspond with shale-dominated areas.

The uppermost subdivision of the Bunter Sandstone Formation is L3b. It is truncated by the Hardegsen unconformity which has resulted in thinning over the NW flank of Endurance, where it is not present within well 42/25d-3 (Fig. 14). L3b thickens to a maximum of 50 m toward the north of the Outer Silverpit Fault Zone in block 44/26 (Fig. 16). This area is characterized by a discrete shale horizon near the top of the Bunter Sandstone Formation that extends north and east over the Esmond, Forbes, Gordon gas fields (Fig. 15). Reservoir quality of L3b is intimately linked to diagenetic halite cement distribution. There is a stark contrast between zones dominated by high-quality and low-quality reservoir, correlating with areas associated with a seismic polarity reversal (e.g. James et al. 2016; Gluyas and Bagudu 2020; Hollinsworth et al. 2022). Well-log data shows low-quality reservoir zones are characterized by low GR sands with anomalouslt high AI values >13 000 kPa.s/m, above typical low quality reservoir values (Fig. 14). Furthermore, these horizons have low RHOB values relative to their AI, indicating they represent salt-plugged sandstones (i.e. Fig. 8). High-quality reservoir in L3b has similarly low GR sands that lack such AI spikes. As such, a wide range of $Φ$_eff is estimated between 2–33%, with salt-plugged zones recording average $Φ$_eff < 15%, whilst uncemented reservoir $Φ$_eff exceeds 22% (Fig. 17).

Previously, attempts to correlate intra-formational horizons within the Bunter Sandstone Formation have been focused at a site-specific scale (James et al. 2016; National Grid 2016). Results here suggest that well-log derived correlations assisted by facies-modelling can allow for effective interpretation of correlated horizons within the Formation over a wider area. In this case, mudstone horizons, low-porosity, and salt-plugged sandstone intervals that may form baffles or barriers to vertical dispersal of CO₂ plumes can be identified, and their lateral extents determined. One laterally extensive horizon of importance includes a high AI, low porosity sandstone marking the top of the L2B subdivision, extending from the East Midlands Shelf eastward over Endurance as far as the Esmond field. To the SE of the basin, the horizon grades into a mudstone which itself is laterally continuous between 44/26-CS1 and Caister Bunter. Several mudstone and low-porosity sandstone intervals within the L3a and L2b subdivisions are also traceable across the Silverpit Basin (Fig. 14). This suggests that many vertical permeability baffles may be persistent within individual closures, and injection rates and/or local pressure capacity of sandstone compartments may need to be assessed in development strategies. The cemented sandstone responsible for seismic polarity reversals and severe deterioration of reservoir quality in the uppermost 30 m of the Bunter Sandstone is not obviously linked to a stratigraphic feature as there is no obvious facies gradation within L3b (Fig. 14), which hosts either high porosity (>20%, maximum 33%) or low porosity (4–12%) sandstone.

The petrophysical and prospect-based analysis of the Bunter Sandstone Formation in this study presents the first detailed reservoir characterization of the Formation regarding CCUS potential over a regional scale, but crucially, informed by local petrophysical analysis. Previous regional studies of storage capacity potential have tended to apply universal values that homogenize porosity, NTG and S_wirr (Brook et al. 2003; Bentham 2006), or model theoretical storage capacity without consideration of geological factors such as structural or stratigraphic traps (Heinemann et al. 2012). The scope of this study was to build on regional analysis of CO₂ storage potential by incorporating reservoir heterogeneity into reservoir characterization. This is after the fashion of detailed FEED studies into Endurance (National Grid 2016) and 44/26-CS1 (James et al. 2016), but with the goal of applying the methodology to multiple CCUS prospects.

Petrophysical modelling of the Bunter Sandstone Formation shows strong correlation with empirical evidence from core porosity measurements and composite well log records (Fig. 8). However, estimating in-situ porosity and permeability values must be treated with caution, due to the possibility of dissolution of halite cement in the presence of drilling fluids. Halite from saliferous units (e.g. the Rot Halite Member) is commonly not recorded in drilling cuttings, though halite-bearing units and halite cement is preserved in sections drilled with oil-based drilling fluids (e.g. Texas Gas Exploration 1986a; Lasmo North Sea 1987; Amoco U.K. Exploration 1994; National Grid 2016). Comparisons of modelled $Φ$_eff in this study with empirical core porosity data from 44/23-5 show strong correlations (Fig. 8) suggesting that well log-based petrophysical models can capture accurate $Φ$_eff values. Overall, medium and high-quality sandstone facies (⁠$Φ$_eff approx_.15–30%) account for an average 64% of the reservoir facies across the UK SNS, highlighting broadly favourable conditions for CO₂ injection. High porosity values over structural highs in the Silverpit Basin (Fig. 12a) point to halokinetic uplift limiting the effects of burial and compaction, while conversely depocentres such as the hangingwall of the Dowsing Fault Zone (e.g. Sole Pit Basin and SW margin of the Silverpit Basin) are characterized by low reservoir porosity signalling deeper burial and compaction (Fig. 12a). Reservoir conditions increasingly worsen toward distal eastern margins of the Silverpit Basin highlighted by both increasing mudstone content, and general thinning of the Bunter Sandstone Formation.

Electrofacies modelling using common well log parameters has proven an effective comparative tool with which to assess reservoir quality and identify possible permeability baffles and barriers in the Bunter Sandstone Formation. It is possible to delineate between reservoir and non-reservoir facies, and further hierarchically group reservoir facies based on DT and RHOB values, linked closely to porosity. This analysis has provided a basis for correlating intraformational horizons in the Bunter Sandstone Formation, showing that these can be traced at the scale of sub-basins within the UK SNS. Prior correlations have been focused over smaller areas covering individual prospects (James et al. 2016; National Grid 2016), and have been restricted to major subdivision groupings. Here, this study highlights the substantial lateral continuity (10 s–100 s km) of intraformational mudstones, low permeability baffles linked to salt-plugging or low-porosity sandstone intervals, and the effects of erosion highlighted by the Hardegsen Disconformity at the base of the Mid-Triassic (Fig. 14). Such correlations are important for CCUS as they indicate the potential for compartmentalization over the scale of individual CCUS prospects, which may need to be factored into CO₂ injection strategies, addressing density and number of well penetrations and CO₂ injection rates that consider pressure capacity (c.f. Heinemann et al. 2012). Aquifer connection may be inhibited (c.f. Bentham et al. 2017) in well-cemented areas, which may have implications for CO₂ plume dispersal during injection. However, there may be beneficial aspects to this where lateral CO₂ plume migration is facilitated by vertical permeability baffles (Williams et al. 2013). Present limitations of the electrofacies methodology within this study include the inability to distinguish between generally distributed calcite, dolomite and anhydrite cements recorded in many composite log records, and halite-plugged sandstone that typically forms discrete, anomalously high AI intervals in the Bunter Sandstone Formation (e.g. Ritchie and Pratsides 1993; Fig. 12). The latter are distinguishable on AI-RHOB (Fig. 9) plots and in well logs, recording anomalously high AI values (>13 000 kPa sm⁻¹) with relatively low bulk density (43/23-3 Fig. 12) and could be better distinguished. More nuanced facies classification has been achieved through combined sedimentology, petrographic and chemostratigraphic analysis of recovered core in other studies (National Grid 2016), with electrofacies models produced from GR, DT, and resistivity logs. This methodology allowed distinction to be made between partial and fully cemented sandstones, and carbonate cemented sandstones, in addition to high quality reservoir and non-reservoir facies. Future studies focused on individual prospects may seek to access available rock core and combine a petrographic approach with petrophysical modelling to better constrain sedimentary facies and diagenetically altered zones of the Bunter Sandstone.

The combined total theoretical storage capacity of Bunter Sandstone Formation closures within licensed areas of the Silverpit Basin (CS001, CS006, and CS007) is 5700 MCO₂t (calculated from P50 cases in Table 2). In 2023, the UK's Government Department for Energy Security and Net Zero reported UK annual CO₂ emissions of 331.5 MCO₂t for the year 2022, of which transport emissions accounted for 112.5 MCO₂t (33.4% of emissions), with the business sector and industrial processes combined accounting for 71.9 MCO₂t (21.6% of emissions) (Department for Energy Security and Net Zero 2023). This highlights the potentially substantial contribution that the Bunter Sandstone Formation may offer in the UK's pursuit of attainment of Net Zero strategies linked to decarbonizing business and industry, particularly of industrial clusters facing the UK Southern North Sea (Teeside and Humberside).

Acquisition of 3D seismic over and well penetration of the volumetrically significant 43/30-CS1 will be a crucial next step in play-based exploration for CCUS in the Silverpit Basin, and well penetration of the SE crest of 44/27-CS1 will be necessary to better establish reservoir conditions within this closure. Triassic structural closures over the Harbour CCS licensed block (CS005) have the further potential to store a combined 434 MCO₂t, over-and-above prospective Rotliegened depleted gas field storage volumes (Table 2). Whilst reservoir conditions in both closures within this license are negatively affected by lower NTG and $Φ$_eff relative to those in Silverpit Basin license blocks, they present an opportunity to utilize both sub-salt and supra-salt CCUS prospects within a single license area. Minor faults possibly penetrate the Bunter Sandstone Formation through the underlying Bunter Shale Formation (Fig. 18), whilst numerous well penetrations have targeted Permian gas reservoirs, posing both geological and present engineering integrity risks (e.g. Hollinsworth et al. 2022) to CO₂ containment. Despite the presence of volumetrically significant Triassic structures in the inverted Sole Pit Basin, closures 48/06-CS1 and 48/14-CS1 pose significant risk to CO₂ storage due to shallow crest depths, and en echelon domino style faulting in the overburden of 48/14-CS1. Similar containment risks are characteristic of closures 43/11-CS1, 43/17-CS1, 43/30-CS2, 47/03-CS1, 48/01-CS1, 48/04-CS1, and 49/06-CS1. Of those closures positioned in presently offered CCUS licensed blocks, the depleted Caister Bunter Field (44/23-CS2), 43/16-CS1 and 43/18-CS1 are volumetrically significant (>100 MCO₂t), structurally low-risk, and in close proximity of presently held license areas that they may form future attractive prospects upon establishing CCUS infrastructure in the Silverpit Basin (sp.). These, and a number of volumetrically small closures (<100 MCO₂t) may become more viable with knowledge appreciation garnered from established CCUS pilot schemes.

The Bunter Sandstone Formation CCUS portfolio of the UK SNS is demonstrated here to have the collective potential to store several decades-worth of UK industrial CO₂ emissions in unfaulted, high-quality reservoir sandstones. Further work is required to better constrain the distribution and composition of pore-filling cements, particularly halite, and better understand the causes of halite-cemented stringers that substantially change the acoustic properties and reservoir quality in the upper 30 m of the Formation. Additionally, this study focuses on a static approach to characterizing reservoir quality and storage capacity of the Bunter Sandstone. At present, few studies have embraced a dynamic approach based on reservoir engineering in the Formation that consider storage efficiency beyond static capacity estimates (Williams et al. 2013). Such computational models will be essential at the scale of individual prospects in determining CO₂ dispersal rates, the role that vertical permeability barriers and open v. closed dome scenarios may have on pressure capacity (Brook et al. 2003), and trapping mechanisms. The scope of this study also does not cover structural risks including minor faults that occur within the overburden of the Chalk and Haisborough Group that may pose a containment risk for injected CO₂, or potential engineering and offshore licensing issues. An interdisciplinary approach will be crucial to assessing and understanding the full potential and risk profile of the Bunter Sandstone Formation for CCUS on the UKCS.

This study demonstrates the first regional assessment of reservoir quality in the Bunter Sandstone Formation that incorporates reservoir heterogeneity in determining theoretical CO₂ storage capacity of several structural closures in the UK SNS. Petrophysical analysis shows that reservoir quality can be ascertained through effective porosity modelling and through analysis of acoustic impedance – bulk density plots. Electrofacies modelling using common well log parameters between wells in the UK SNS has provided a basis for establishing net-to-gross, and for the first time, intra-formational stratigraphic correlation within the Bunter Sandstone Formation. Crucially, this has created the possibility for tracing laterally extensive (>10 s km) mudstone and highly-cemented stringers that would likely present as vertical permeability baffles for CCUS during injection campaigns.

Monte-carlo models of theoretical storage capacity of thirty Bunter Sandstone Formation closures were run using input parameters ranges tied to site-specific petrophysical properties, signalling the first instance of incorporating geological and petrophysical heterogeneity in assessing the storage potential of the Formation. The P50 estimates from these calculations indicates that closures located within the CS001, CS006 and CS007 have the potential to store 5700 MCO2t, the equivalent of seventy-nine years of the UK's business and industrial processes CO2 emissions for the year 2022. Triassic closures overlying the Harbor CCS prospect (license CS005) have the potential to store a further 434 MCO2t. With several new license awards covering areas with Bunter Sandstone Formation closures awarded in September 2023 (NSTA 2023), the UK SNS carbon storage portfolio is growing, and the role of the Bunter Sandstone Formation is integral to aiding the meeting of UK Net Zero targets. Acquisition of further core data and seismic coverage will be necessary to better constrain reservoir quality, reduce uncertainty, and analyse risk geological and geotechnical risks, whilst dynamic reservoir modelling will be crucial in improving the understanding of CO₂ dispersal and trapping upon injection.

Funding for this project was provided by the Net Zero Technology Centre. Interpretations were drawn from data made publicly available by the North Sea Transition Authority's National Data Repository, including PGS's 3D MegaMerge Volume. Western Geco provided access to the Cavendish 3D volume. We thank Dave Barlass, Matthew Masham, and Alex Ellwood of Western Geco for their support and Western Geco for permission to access and utilize this data. Schlumberger is thanked for the donation of Petrel and Techlog software licenses which have permitted the interpretation and analysis of seismic and wire-line log data to be undertaken.

ADH: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), visualization (lead), writing – original draft (lead), writing – review & editing (supporting); IJ-A: conceptualization (supporting), formal analysis (supporting), investigation (supporting), methodology (supporting), visualization (supporting), writing – review & editing (lead); JRU: conceptualization (supporting), funding acquisition (lead), investigation (supporting), project administration (supporting), supervision (supporting), validation (lead), visualization (supporting), writing – review & editing (supporting); RJJ: funding acquisition (supporting), project administration (lead), resources (lead), software (lead), supervision (supporting), writing – review & editing (supporting)

This work was funded by the Net Zero Technology Centre.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The datasets generated during and/or analysed during the current study are available in the UK National Data repository, https://ndr.nstauthority.co.uk/.