Previous basin modelling of the Faroe–Shetland Basin (FSB, offshore UK) has suggested mid-Cretaceous petroleum generation, which predates the deposition of the working Paleogene reservoirs and traps. To justify the time discrepancy between generation, reservoir, and trap formation, factors such as intermediary accumulations and overpressure have been invoked. However, across much of the FSB, the Cretaceous sequences that overly the Kimmeridgian source rock are heavily intruded by Paleogene-aged intrusions. Recent modelling has shown that the emplacement of the intrusions, coupled with lower radiogenic heat production from underlying basement, leads to estimates of petroleum generation occurring up to 40 myr more recently than suggested by previous models. In this work, we seek to better understand the role that igneous intrusions have exerted on petroleum generation and migration in the FSB. Models with varying thicknesses of Paleogene intrusions are compared with those that consider the Cretaceous sequence as purely sedimentary (i.e. similar to assumptions in previous modelling). The estimated times of petroleum generation are compared with geochronological constraints on the ages of oils (i.e. c. 90–68 Ma) along with the deposition and formation of other petroleum system elements. By considering only the effect of igneous intrusions, the expulsion onset from the source rock is retarded by up to 12 myr. In addition, our models show the impact of the intrusions on petroleum saturation and migration, suggesting that intrusions have potentially compartmentalized the basin, trapping petroleum beneath or within the sill complex. Finally, our findings suggest that basin models in regions impacted by significant magmatism need to consider the impact of intrusions to more accurately constrain both petroleum generation and migration.

Thematic collection: This article is part of the New learning from exploration and development in the UKCS Atlantic Margin collection available at: https://www.lyellcollection.org/topic/collections/new-learning-from-exploration-and-development-in-the-ukcs-atlantic-margin

Basin and petroleum system modelling in sedimentary basins is a critical part of the exploration process (e.g. Hantschel and Kauerauf 2009). The Faroe–Shetland Basin (FSB) (Figs 1 and 2), located within the West of Shetland region (UK Continental Shelf), is an active exploration province (e.g. Austin et al. 2014; Loizou 2014). Therefore, understanding the petroleum system (e.g. source rock maturation, petroleum expulsion and migration, reservoir, and trap formation) is essential for determining the prospectivity of undrilled traps and to improve understanding of fields (e.g. fluid variability and connectivity), as well as for the meaningful evaluation of migration pathways with appropriate times of charge.

Despite proven petroleum systems and the presence of numerous oil, condensate and gas fields and discoveries (e.g. Cambo, Tornado, Clair, Schiehallion, Foinaven, Rosebank, Laggan, Lancaster and Solan) (e.g. Austin et al. 2014; Fielding et al. 2014; Purvis et al. 2020), traditional petroleum system modelling studies focused on the FSB have faced issues in reconciling the predicted timing of the petroleum generation against the known occurrences and ages of oil- and gas-charged reservoirs (Doré et al. 1997; Iliffe et al. 1999; Lamers and Carmichael 1999; Carr and Scotchman 2003; Scotchman et al. 2006; Gardiner et al. 2019). A complex interplay between faults, carrier systems, overpressure and intrusions seems to have played an important role for petroleum migration to structural highs and younger stratigraphy across the Cretaceous fine-grained sedimentary sequence that otherwise forms a major regional seal (e.g. Doré et al. 1997; Iliffe et al. 1999; Scotchman et al. 2006; Gardiner et al. 2019; Schofield et al. 2020).

Several studies on isotopes and fluid inclusions have suggested age ranges of petroleum generation within the FSB of c. 113.2 ± 3.5 Ma (Ar–Ar dating of feldspars surrounding oil-filled inclusions: Mark et al. 2010), c. 93–69 Ma (U–Pb dating of calcite: Holdsworth et al. 2019), c. 83 Ma (40Ar–39Ar dating of feldspars surrounding oil-filled inclusions: Mark et al. 2005) and c. 68 ± 13 Ma (Re–Os isotopes: Finlay et al. 2011). The variability of the data available in the literature might also suggest that some local effects could have played an important role and thus much more data are needed. Importantly, as already suggested by Gardiner et al. (2019), although geochronological constraints on the ages of oils suggest that petroleum generation began between c. 90 and 68 Ma, previous basin models have predicted that petroleum generation started in the mid-Cretaceous (c. 100 Ma), significantly predating (by up to 45 myr) the deposition of Paleogene reservoirs, seals (deposition of which started at least c. 66 myr later) and the development of structural traps during the Miocene (e.g. Doré et al. 1997; Holmes et al. 1999; Iliffe et al. 1999; Scotchman et al. 2006; Tuitt et al. 2010; Gardiner et al. 2019).

The discrepancy in timing of predicted petroleum generation v. the age of generation has been explained in several ways: (1) intermediary accumulations (i.e. ‘motel model’: Doré et al. 1997; Lamers and Carmichael 1999; and ‘whoopee cushion effect’: Iliffe et al. 1999) in which petroleum that was stored in Lower–middle Cretaceous reservoir sequences remigrated after the development of the other petroleum system elements; (2) overpressure within Jurassic source rock that was assumed to retard petroleum generation (Carr 1999; Carr and Scotchman 2003; Scotchman and Carr 2005; Scotchman et al. 2006); (3) the age and composition of basement (c. 2.8–2.7 Ga: Holdsworth et al. 2018), which affected the heat flow history; and (4) the impact of Paleogene intrusions on source rock burial and thermal history (Gardiner et al. 2019).

The FSB experienced considerable intrusive magmatism during the Paleogene, resulting in a laterally extensive subsurface complex of mainly dolerite sills that intruded into Cretaceous and Paleocene sequences (Schofield et al. 2017a; Mark et al. 2019). The sill complex is variable in thickness, ranging from an approximate cumulative vertical thickness of c. 50 m in the Southern Flett Sub-basin to a cumulative thickness of more than c. 1.8 km in the Nuevo Sub-basin, within the northern part of the Judd Basin (Mark et al. 2018). Recent studies have highlighted that the presence of these intrusions can drastically affect the petroleum system in terms of: (1) the potential to overthicken sedimentary sequences (i.e. a sedimentary sequence is thickened after its deposition as result of intrusion emplacement: see Schofield et al. 2017a; Mark et al. 2019); and (2) changing the thermal properties of the sequence they intruded post-crystallization, as intrusions usually have higher thermal conductivity than the fine-grained sedimentary rocks they have intruded (Gardiner et al. 2019).

In this paper, we explore the impact of sill intrusions on petroleum generation and migration within the Nuevo Sub-basin, which is located within the southern portion of the FSB and to the south of the Cambo oil and gas discovery (Fig. 1). Our models allow us to quantify the thermal effect of the intrusions after they crystallized within the Cretaceous sedimentary host rock strata and to isolate the contribution of intrusions in retarding the onset of expulsion. Furthermore, we show and discuss the impact of intrusion emplacement on petroleum saturation. This and other similar works suggest that prospective basins containing substantial amounts of igneous rocks in the form of intrusions cannot be studied using the traditional basin modelling workflow.

The main findings of our analysis are: (1) the increased thermal conductivity of the Cretaceous sequence following the crystallization of the sills reduces the geothermal gradient in the rocks above the underlying Jurassic source rock regions, thereby affecting the timing of maturation; (2) an estimation of source rock expulsion timing that is more consistent with the timing of other proven play elements and with the geochronological dating of oils; and (3) that the key driver of petroleum saturation and migration is the complex interplay between the intrusion thickness, their capillary entry pressure (CEP), the hydraulic properties of faults (i.e. open or closed faults) and the presence of permeable carrier beds within the basin.

Although the FSB is the focus of this study, our workflow can be applied to any sedimentary basin that has been heavily intruded (e.g. Brazil: Thomaz Filho et al. 2008; South Africa: Svensen et al. 2012; Senger et al. 2017; offshore Norway: Brekke et al. 1999; Cunha et al. 2021; Gac et al. 2022; Western Australia: Holford et al. 2013; NW margin of Australia: Schenk et al. 2019; and offshore southern Australia: Reynolds et al. 2018) to evaluate the impact of intrusions on petroleum generation and migration.

The FSB is situated on thinned continental crust between the Shetland and the Faroe islands within a structurally complex, volcanic rifted margin. The FSB can be divided into a series of NE–SW-trending sub-basins, subdivided by crystalline basement highs that were formed as a result of multiple phases of extension, compression and magmatism. These basement highs are often capped by Paleozoic and Mesozoic sedimentary rocks (e.g. Doré et al. 1997; Dean et al. 1999; Carr and Scotchman 2003; Fletcher et al. 2012; Rateau et al. 2013; Rippington et al. 2015; Schofield et al. 2015; Stoker 2016) (Figs 1 and 2).

The geological architecture of the FSB has been influenced by: the Caledonian Orogeny; major rifting episodes from late Carboniferous to Cretaceous; igneous activity and thermal subsidence during the Paleogene, associated with the opening of the North Atlantic Ocean; and, finally, episodic compressional deformation during the Eocene, Oligocene, and Miocene (Scotchman et al. 1998; Dean et al. 1999; Doré et al. 2008; Rateau et al. 2013). Several rifting episodes have been identified during the Cretaceous (Valanginian–Barremian, possible Aptian–Albian, Cenomanian–Santonian and Campanian–Maastrichtian: Dean et al. 1999; Carr and Scotchman 2003), after the Cretaceous, continental break-up and formation of the proto-Atlantic Ocean began around c. 55 Ma (Ritchie et al. 2011). It is important to note that rifting is spatially variable throughout the entire FSB, with the Judd Sub-basin and the Nuevo Sub-basin showing significant evidence of tectonically driven subsidence during the Paleocence, with less evidence in the northern FSB. The rifting phases were punctuated by post-rift thermal subsidence, which generated the accommodation space for deep-water sediments, and in addition by Paleogene volcanism and late Paleocene and Miocene basin inversion events (Smallwood and Maresh 2002; Carr and Scotchman 2003; Fletcher et al. 2012). During the late Paleocene–early Eocene, the proto-Icelandic plume and the eventual continental break-up between Greenland and NW Europe generated igneous activity in the FSB, as well as along the broader NE Atlantic Margin (Ritchie et al. 2011; Ellis and Stoker 2014). This activity is expressed by thick sequences of extrusive basalts, sills and dykes mainly intruded into Cretaceous sedimentary host rocks (Ritchie et al. 2011; Schofield et al. 2015). These intrusions are collectively named the Faroe–Shetland Sill Complex (FSSC) (Ritchie et al. 2011; Schofield et al. 2015). Based on detailed seismic interpretation, Schofield et al. (2017a) suggested that the main zones of magma input into the FSB were controlled primarily by the NE–SW-striking structural fabric that compartmentalized the FSB into sub-basins. Schofield et al. (2015, 2017a) and Mark et al. (2018) analysed well data and showed that up to 88% of sills in the FSSC are less than 40 m thick, which is often below the vertical resolution of seismic reflection data at the depths where the majority of the sill complex occurs and therefore are potentially not fully resolved or detected at all in many cases (e.g. within the Flett Basin this is around 3.5 s two-way time (TWT), which is equivalent to c. 3 km below the seabed).

Petroleum habitat and petroleum discoveries within the Judd and Nuevo sub-basins and the Cambo High

The Judd and Nuevo sub-basins host several petroleum discoveries. The Cambo discovery (2002) is a four-way dip-closed structure situated on the SW culmination of the Corona High. The Cambo discovery is estimated to contain more than 800 MMbbl of c. 24° API oil and is currently being considered for development (Fielding et al. 2014; Purvis et al. 2020). The reservoir is hosted within the Eocene paralic (fluvial–deltaic–estuarine) sandstones of the Hildasay Member (T45: upper Flett Formation), which are intercalated with claystones, siltstones and coals (Fielding et al. 2014; Purvis et al. 2020) (Figs 1 and 2). The reservoir is sealed by post-Eocene marine claystones (Fielding et al. 2014; Purvis et al. 2020).

The undeveloped Tornado discovery (OMV, drilled in 2009 by well 204/13-1) (Fig. 1) is situated c. 30 km SW of the Cambo discovery in the Nuevo Sub-basin. The Tornado discovery is a three-way dip-closed structure that pinches out to the south and partly pinches out towards the east. A strong seismic amplitude anomaly has been attributed to the presence of gas- and oil-bearing sandstones of upper Paleocene age (T38: Lamba Formation) that were deposited in a marine setting as a toe of a slope fan, and the overlying seals comprise basinal mudstones and siltstones (Rodriguez et al. 2010). The field contains predominantly gas with a thin oil rim, and it has a recoverable resource of around 80 MMboe (Siccar Point Energy 2017, 2020).

The Cambo and Tornado reservoirs are thought to have been charged by the Upper Jurassic Kimmeridge Clay Formation (e.g. Holmes et al. 1999; Iliffe et al. 1999; Jowitt et al. 1999; Lamers and Carmichael 1999; Gardiner et al. 2019). Importantly, in the FSB, the sill complex is often located above the Jurassic source rock zones and below many of the reservoir zones (e.g. Paleocene) (Schofield et al. 2015; Mark et al. 2019).

To explain the time discrepancy between petroleum generation, reservoir deposition and trap formation in the FSB, several models have been invoked: these are summarized below (Table 1).

Doré et al. (1997); Lamers and Carmichael (1999): the motel model in the NE Atlantic Margin, and in the Foinaven and Flett sub-basins

Doré et al. (1997) proposed what has come to be known as ‘the motel’ (or ‘hotel’) model to explain the migration history in the FSB and other areas of the NE Atlantic Margin. In this model, expulsion from Jurassic source rock commenced during the early–mid-Cretaceous, temporarily charging Lower–middle Cretaceous reservoir sequences. The current petroleum accumulations then formed following remigration after the development of the Paleogene petroleum system elements (i.e. reservoirs, seals, and traps). Doré et al. (1997) postulated that the remigration may have been triggered by the change from NW–SE extension to NW–SE compression in the early Cenozoic. Lamers and Carmichael (1999) applied a variant of the motel model of Doré et al. (1997) to the Foinaven and the Flett sub-basins, suggesting temporary residence of petroleum in Mesozoic reservoirs. However, Lamers and Carmichael (1999) suggested that generation and expulsion occurred during the late Cretaceous. They argued that Lower Cretaceous sandstones acted as migration conduits towards temporary reservoirs that stored palaeo-accumulations which were then breached because of rapid burial sedimentation at the end of the Paleocene–early Eocene. This caused overpressure-induced fracturing, which allowed migration that was probably influenced by Cenozoic inversion structures.

Iliffe et al. (1999): the FSB whoopee cushion effect

Iliffe et al. (1999) suggested that faults allowed petroleum migration from the Jurassic source kitchen into the Lower Cretaceous (Barremian–Albian) sandstones of the Cromer Knoll Group until buoyancy drove petroleum to updip pinchouts, where it was trapped. Rapid deposition during the Paleocene led to increased pore pressure and eventually hydrofracturing (described as the ‘whoopee cushion effect’), which allowed vertical migration through Upper Cretaceous shales into the overlying Paleocene sandstones.

This model is similar to the ‘motel model’ of Doré et al. (1997) because it also invokes the migration of petroleum into temporary reservoirs. The models of Doré et al. (1997), Iliffe et al. (1999) and Lamers and Carmichael (1999) all emphasize that overpressure, necessary for remigration, was probably related to rapid deposition during the Paleocene–early Eocene. However, both Doré et al. (1997) and Lamers and Carmichael (1999) suggested that Cenozoic compression was able to create structures that were probably able to focus the remigration of petroleum, whereas Iliffe et al. (1999) suggested that remigration is related to the overpressure and development of hydrofractures.

The models of Doré et al. (1997) and Iliffe et al. (1999) have subsequently been challenged. (1) Because the extent of the Cretaceous sandstones with good porosity and permeability, and with a sufficient extent to act as temporary reservoirs, remain highly uncertain despite the availability of extensive 3D seismic data (Scotchman et al. 2006). Cretaceous sandstones do occur within wells 204/19-2 and 204/19-1 adjacent to this paper's area of interest; however, well penetrations are sparse and delineation via seismic is problematic due to the large volume of intrusions. Cretaceous sands are more frequently observed adjacent to the Rona High (e.g. 206/11-1, Clair Field), where they were eroded from the West Shetland Platform and the Rona High in the east, and transported/deposited to the west via mass flow sand systems. It is postulated, but not proven, that the Corona High is beyond the reach of these sand systems, and any Cretaceous sands were locally derived from erosion of emergent highs. (2) Scotchman et al. (2006) suggested that the inefficiency of the migration of oil and gas in two stages (as per the motel model) would lead to such a loss of volume through migration, that the known petroleum potential of the source rock is not sufficient enough to account for the volumes of hydrocarbon currently seen in fields. However, without reliable estimates on how thick, extensive and organic rich the source rock is in the basin, this is rather speculative.

Carr (1999); Carr and Scotchman (2003); Scotchman and Carr (2005); Scotchman et al. (2006): the southern FSB (Foinaven Sub-basin) overpressure

Carr (1999), Carr and Scotchman (2003), Scotchman and Carr (2005), and Scotchman et al. (2006) proposed a model to explain the delay in petroleum generation as a result of overpressure within Jurassic source rock acting to retard petroleum generation. Importantly, in their model the overpressure invoked within the Jurassic source rock means that there is no need for the ‘motel’ and/or ‘whoopie cushion’ models, because the timing of generation is delayed to within the period of trap formation during the late Paleogene. They also suggested that in the overpressure model, potentially lower volumes of oil are lost during migration than in the ‘motel’ and ‘whoopie cushion’ models. Gardiner et al. (2019) questioned the Jurassic source rock overpressure model because: (1) the impact of overpressure on source rock maturation is still not completely clear (Landais et al. 1994; Huang 1996; Nielsen et al. 2017); and (2) the amount of overpressure encountered within the Mesozoic sections across the FSB is variable, meaning that the applicability of this model at a regional scale and as a primary control may be problematic, even if it is acknowledged as a likely contributory factor in the retardation of maturation.

Gardiner et al. (2019): ‘cold’ basement and impact of intrusions in the FSB

For all previously published basin and petroleum system models developed for the FSB, the stratigraphy has only been treated as sedimentary in origin. However, work conducted by Schofield et al. (2015, 2017a) and Mark et al. (2019) demonstrated that the Cretaceous and Paleocene sequences within the FSB were heavily intruded by Paleogene-aged sills, often to a much higher degree then can be ascertained from the interpretation of seismic reflection data alone due to pervasive subseismic-scale intrusions. Mark et al. (2019) introduced the concept of overthickening, which suggests that the large amount of intrusive material, mainly within the Cretaceous sequence, might have led to the overall increase in thickness of this sequence (i.e. the present-day thickness of the sedimentary sequence is thicker than when it was deposited (even after compaction is considered)), as it is now a function of both the original sedimentary rocks and later Paleocene intrusions (Schofield et al. 2017a; Mark et al. 2019). Importantly, Mark et al. (2019) noted that the thickness of the igneous material is far from uniform across the FSB, with adjacent sub-basins having different thicknesses of intrusive material within the sedimentary pile, ranging from c. 50 m (southern Flett Sub-basin) to more than c. 1.8 km (Nuevo Sub-basin) (Mark et al. 2019). Gardiner et al. (2019) highlighted three main factors that should be considered when undertaking petroleum system modelling in basins with significant overthickening such as in the FSB: (1) igneous intrusion timing and overthickening; (2) the change (increase) in thermal conductivity of the sedimentary sequence post-crystallization of sills; and (3) the use of more accurate basement-derived heat flow. Intrusive igneous material should be added at the correct time in order to correctly model the subsidence history of the basin. In their basin modelling study, Gardiner et al. (2019) removed the overthickening (i.e. the thickness of the intrusions) from the basin succession, and then put them back at the time of emplacement. The major implication of removing igneous material is that the Jurassic source rock was significantly shallower and thus colder during the Cretaceous (i.e. prior to intrusion emplacement) than previously considered, leading to a later onset of petroleum generation than previously modelled (i.e. in the Campanian (73 Ma), c. 17 myr closer to the present-day than in previous models: see Gardiner et al. 2019). Gardiner et al. (2019) used c. 2.1–2.5 W m−1 K−1, following Hartlieb et al. (2016), as the thermal conductivity of igneous rocks and c. 0.8–1.5 W m−1 K−1, following Sharma (2002), as the thermal conductivity of sedimentary rocks. One of the largest effects that Gardiner et al. (2019) found in controlling the timing of petroleum generation was the change in thermal conductivity that occurs in a sedimentary unit when more thermally conductive igneous rocks are intruded into the less thermally conductive sedimentary rocks. A key aspect of this is the realization that the dolerite magma cools down to solid dolerite in a relatively short period of geological time (c. 102–104 years: Peace et al. 2017; Nicoli et al. 2018; Gardiner et al. 2019) compared to the FSB geological history. Since the heating related to the cooling down of the magma is local, it does not affect the large-scale heat flow of the basin, suggesting that the high-temperature anomalies generated from cooling magma can potentially be neglected. In addition, Gardiner et al. (2019) pointed out that basin models in the FSB should use a more accurate basement and lower crust composition, age and radiogenic heat production (RHP), because in the FSB the basement is older (i.e. Neoarchean, c. 2.8–2.7 Ga: Holdsworth et al. 2018) and composed of a pelitic protolith with relatively low initial RHP compared to that typically assumed in the North Sea, which was used in previous basin models. While typical North Sea values of RHP range from c. 2.5 to 3.2 mW m−3, Gardiner et al. (2019) calculated a value of 1.6 mW m−3 in the study area. This lower value results in a lower heat flow: without incorporating igneous intrusions but by only using this basement instead of the one used in previous basin models (i.e. typical North Sea), Gardiner et al. (2019) calculated a decrease of c. 10% in the mean present-day surface heat flow and geothermal gradient in comparison to a typical North Sea model. Gardiner et al. (2019) suggested that the critical moment of petroleum generation could have been delayed locally until after the Paleogene reservoir and seal units were deposited by considering the timing of emplacement and their impact on the thickness of the basin succession and by more accurate modelling of the RHP of the basement rocks. Gardiner et al. (2019) calculated that the onset of oil expulsion could have potentially occurred later than previously suggested, with values of c. 30 myr in the centre of the Judd Sub-basin and c. 22 myr in the centre of the Flett Sub-basin. They concluded that this delay could account for the migration history without needing to invoke either complex remigration or overpressure.

However, importantly, Gardiner et al. (2019) strongly stated that the FSB has such geological complexity, much of which is still unknown, that there is probably no single unifying mechanism controlling the petroleum system (including the model proposed by Gardiner et al. 2019), and that the ‘true’ understanding of the petroleum system timing probably resides in a combination of one or more of the models proposed by Gardiner et al. (2019) and other workers, (e.g. Iliffe et al. 1999; Lamers and Carmichael 1999; Carr and Scotchman 2003; Scotchman and Carr 2005; Scotchman et al. 2006).

After a source rock is deposited, it needs to be buried at an adequate subsurface temperature so that the maturation process can start, leading to petroleum generation and expulsion (first stage of migration) (Schowalter 1979; Tissot and Welte 1984; England et al. 1987; Magoon and Beaumont 1994; Hantschel and Kauerauf 2009; Allen and Allen 2013). Although many aspects of petroleum migration are still not fully understood (e.g. Hantschel and Kauerauf 2009), it is generally accepted that generated petroleum migrates via permeable pathways out of the source rock (primary migration), through carrier beds and faults (secondary migration), and that, during secondary migration, portions of petroleum will become trapped and stalled due to subsurface heterogeneities (e.g. small traps or faults) and cannot migrate any further (e.g. Schowalter 1979; Tissot and Welte 1984; England et al. 1987; Magoon and Beaumont 1994; Hantschel and Kauerauf 2009; Allen and Allen 2013). If the volume of the petroleum migrating into a trap is sufficient (i.e. charge volume is greater than leaked/spilled volume), it will form a petroleum accumulation (e.g. Smith 1966; Haney et al. 2005). If the trap is disrupted, petroleum may remigrate to a new trap or to the surface (tertiary migration: e.g. Tissot and Welte 1984; Cartwright et al. 2007; Allen and Allen 2013). Petroleum migration is dependent on the pressure regime, buoyancy and capillary entry pressure (CEP). Migration is inhibited by the CEP, which needs to be overcome during petroleum migration (e.g. Smith 1966; Berg 1975; Schowalter 1979; Tissot and Welte 1984; England et al. 1987; Hantschel and Kauerauf 2009; Allen and Allen 2013). Capillary pressure increases as pore size decreases and it is, together with interfacial tension and wettability, responsible for trapping oil and gas (e.g. Smith 1966; Berg 1975; Vavra et al. 1992; Hantschel and Kauerauf 2009). As the petroleum column increases, and therefore the buoyancy pressure exerted upwards increases, petroleum can be forced through pores with smaller and smaller pore throats (e.g. Vavra et al. 1992).

A challenge for sedimentary basins heavily intruded by sills, including the FSB, is to understand, in addition to petroleum generation, the potential influence of intrusions on petroleum migration (Rateau et al. 2013; Gardiner et al. 2019; Schofield et al. 2020). Previous authors have considered the possibility that igneous intrusions play a dual role, both as seals and also potentially allowing migration (e.g. Schutter 2003; Thomaz Filho et al. 2008; Rateau et al. 2013; Schofield et al. 2015, 2020; Senger et al. 2017; Belaidi et al. 2018; Gardiner et al. 2019; Trice et al. 2019). Multiple intrusions within the FSB have been associated with mud losses during drilling, suggesting that they may be fractured and could possibly act as migration ‘superhighways’ that allow the migration of oil and gas through thick, impermeable Cretaceous sequence (Rateau et al. 2013; Schofield et al. 2020). The intrusions within the FSB have never been exhumed or substantially uplifted since their emplacement (Schofield et al. 2020); however, intrusions naturally develop joints (fractures) during cooling. In addition, Schofield et al. (2020) suggested that the compression phase from the Miocene to the Oligocene could have potentially reactivated and expanded existing cooling fractures or potentially created new ones within intrusions.

The models presented and discussed in this paper aim to evaluate and quantify the impact of sill intrusions on petroleum generation and migration within the Nuevo Sub-basin. Our models allow a better understanding of specific factors, such as intrusion thickness and their CEP, that make basins with intrusions fundamentally different to basins without them. In building the 2D model, the decision was taken to try to minimize variables, in an attempt to explore the key sensitivities of the model to intrusion thickness and their CEP.

Basin modelling

In this study, the PGS/TGS 3D seismic reflection cube FSB 2011–12 (CS9: PP123DGFSB) and available well data from the NDR (UK National Data Repository) were used to identify the key stratigraphic horizons to use for the basin modelling after carrying out depth conversion (Table 2; Fig. 3). A 93 km-wide and 13 km-deep cross-section that intercepts wells and discoveries 204/13-1/1Z (Tornado) and 204/10-1 (Cambo) was chosen (Fig. 3). This cross-section is orientated approximately SW–NE and shows the heavily intruded Nuevo Sub-basin on the SW part, the Cambo High and the Corona Ridge on the NE part towards the Rosebank area.

The basin modelling work was completed in PetroMod 2D. A focus was placed on 2D modelling within the study to allow maximum understanding and control of fundamental model parameters, as well as assessing the impact of changing specific inputs and alternative models while keeping computation costs reasonable. Composite logs for wells 204/13-1/1Z (Tornado) and 204/10-1 (Cambo), as well as reports and literature, were used to derive the different lithologies that were averaged between the two wells and the same lithology was applied to each whole layer everywhere in order to populate the model. We considered lithology information from wells and also extrapolated them in context using available literature. For deeper layers not penetrated by these wells, lithological information from nearby offset wells was incorporated (i.e. 204/10a-4, 204/14-1, 204/14-2, 204/15-2, and 213/27-1Z), averaged and used across the whole model for each interval. We are aware that these are limitations of the model; however, because of seismic resolution and lack of additional data, we preferred to keep the model simple rather than to complicate it without any reasonable constraint. For wells 204/10-1 (Cambo) and 204/13-1 (Tornado), porosity–depth and porosity–permeability data were available for the reservoir intervals and were therefore used to calibrate the model by adjusting lithological properties around the well locations. Finally, vitrinite reflectance data and corrected bottom-hole temperatures available for two wells, 204/10-1 (Cambo) and 204/13-1 (Tornado), were used to constrain the thermal history derived from the basin model.

To model petroleum generation more accurately, original total organic carbon (TOCo), original hydrogen index (HIo), and appropriate reaction kinetic models need to be used for the Upper Jurassic source rock. The source rock was modelled to have TOCo = 4.4% and HIo = 350 mgHC g−1 TOC, which is based on Applied Petroleum Technology (2017) and Gardiner et al. (2019). It is well known that significant variability exists within the Kimmeridge Clay Formation across the FSB, which is fragmented into several sub-basins where the depositional setting and the organofacies can vary significantly (Applied Petroleum Technology 2017). However, we chose to use these values for our study to represent an average Kimmeridge Clay Formation source.

In this study, the kinetic model of the source rock was modelled using the kinetic model suggested by Pepper and Corvi (1995) for organofacies B and type II kerogen because: (1) we wanted to compare our work with Gardiner et al. (2019); and (2) the consensus in the literature (e.g. Scotchman et al. 2006) is that the environment of deposition throughout the main basins in the FSB was predominantly clastic marine, with varying degrees of marine isolation (e.g. Solan Basin) and terrestrial input. It should also be noted that no wells penetrated the Jurassic source rock(s) in the deep basins (i.e. kitchen areas), making it difficult to get additional information.

Including the intrusions in the model: palaeothickness maps

A main focus and challenge of our study was to model the intrusion emplacement and to quantitatively address the impact of the intrusions on petroleum generation and migration. To achieve this, the intrusion emplacement needed to be placed in the model at the correct geological time. This was done by removing the thickness of the intrusions from the present-day Cretaceous sequence (the unit that contains the majority of intrusions) and to then reinsert this thickness at the correct later geological time. Here, we followed the methodology documented in Gardiner et al. (2019). In order to remove the additional thickness generated by the intrusions correctly, palaeothickness maps were created. They represent the thickness of the original Cretaceous sedimentary succession before the intrusion emplacement. These maps were used up to the time of the intrusion emplacement and were then substituted by the present-day thickness (i.e. the originally deposited sedimentary Cretaceous succession together with the igneous intrusions). The palaeothickness maps were generated as follows (Fig. 4): (1) identification of the sills along the cross-section; (2) summing up the thickness of individual sills to get the cumulative thickness of all sills (i.e. the overthickening); (3) removing the intrusions from the Cretaceous sequence to obtain its thickness before the intrusion event; and (4) adding the intrusions at the right geological time (i.e. 55 Ma). Finally, the lithology within the present-day Cretaceous sequence at the sill locations was changed to dolerite using the Facies Piercing tool, which allows the facies in the chosen area (i.e. fine-grained sedimentary Cretaceous sequence) to be changed to a different one (i.e. dolerite). Within PetroMod, various tools exist, which include an Intrusion Model, which can be used instead of the Facies Piercing tool. The Intrusion Model can also be used to change the lithology to dolerite and to define several properties (i.e. intrusion age, intrusion temperature, solidus temperature, magma density, magma thermal conductivity, magma heat capacity, and crystallization heat) that are considered during simulation. However, this tool was not used because it creates a temperature increase at the time of the intrusion event and, after the magma solidifies PetroMod assumes an exponential decay of the temperature. Since the cooling process of the intrusions in sedimentary basins is generally considered as a geologically near-instantaneous event (c. 102–104 years: Peace et al. 2017; Nicoli et al. 2018; Gardiner et al. 2019) and in the absence of reliable data that describe the palaeotemperature evolution in the basin, these additional short-lived temperature peaks are hard to constrain. Hence, this represents a source of uncertainty in the model.

Thermal boundary conditions and palaeowater depth

Boundary conditions define the energetic conditions for temperature and burial history of the basin, which in turn affects the maturation of organic matter contained within the source rock (Wygrala 1989). Basin and petroleum system models require thermal boundary conditions at the top and base of the model. The top thermal boundary condition is a temperature, which varies with latitude and water depth through geological time. These temperatures can be related to the geological age and mean surface palaeotemperature based on plate tectonic reconstructions (Wygrala 1989). The reconstructed palaeolatitudes can then be used to estimate palaeosurface temperatures. Together with the palaeowater depth (PWD), the sediment–water interface temperature (SWIT) can be estimated. For this study, northern Europe and a latitude of 60° N was used.

Gardiner et al. (2019) employed a transient, fixed temperature (1330°C) full lithosphere thermal model in all their 1D models to constrain their 3D model of the FSB. They show results for the thermal calibration of well 213/27-1/1Z (Rosebank) and the resulting basal heat flow history. The heat flow history of this well could potentially be inputted directly into the model, as well 213/27-1/1Z is around 6 km away in a southeasterly direction from the study cross-section. However, by using the thermal history derived for such a well, one assumes that both the basin and the structural high experienced the same thermal history. Usually, rifting basins experience a higher basal heat flow than the structural highs, and therefore using the same basal heat flow everywhere would be unrealistic. For these reasons, we calculated the heat flow maps over the study area using the McKenzie model tool. This is based on the work by McKenzie (1978) and Jarvis and McKenzie (1980), who proposed a model for the development and evolution of sedimentary basins that consists of a rapid stretching of continental lithosphere (associated with block faulting and subsidence) followed by slow cooling. This tool allowed us to generate stretching factors and palaeoheat flow maps as output, which are applied as lower thermal boundary conditions to the model. The results from the thermal model were compared with vitrinite reflectance data and corrected bottom-hole temperatures available for wells 204/10-1 (Cambo) and 204/13-1 (Tornado).

It should be noted that all of our scenarios, including those with no intrusions, use ‘cold’ basement/crust with the radiogenic properties as calculated by Gardiner et al. (2019). Based on the work of Pollack and Chapman (1977), these workers pointed out that the RHP of continental crust that has been used in the North Sea is typically between 2.5 and 3.2 mW m−3. However, whereas the North Sea basement is Phanerozoic in age, the basement in the study area is older (i.e. Neoarchean, c. 2800–2700 Ma: Holdsworth et al. 2018). In the basement rocks drilled by well 204/10-1 (Cambo), U is 3.17 ppm, Th is 1.58 ppm and K is 20 701 ppm, and Gardiner et al. (2019) calculated a value of 1.1 mW m−3 for the radiogenic heat production. In our paper, as the cross-section intercepts well 204/10-1 (Cambo), a constant value of 1.1 mW m−3 was used for the basement/crust. It was assumed that the Lewisian basement has a constant thickness of 4 km (and homogenous composition) throughout the model, with the base of the basement having the same geometry as the top of the basement. In this way, the basement has uniform properties that have the same impact everywhere along the cross-section. This allowed a better understanding of the impact of the intrusions alone without any interference coming from using a different basement geometry, thickness, and/or composition. An alternative workflow would involve choosing several basement thicknesses and compositions (e.g. thicker basement beneath the Cambo High and thinner basement beneath the basin), and systematically exploring the sensitivity to order to quantitatively determine the impact on the thermal regime of the study area. However, the model was designed to explore the impact of some specific parameters, and exploring the impact of the basement thickness and composition was out of the scope of this work.

Together with the basal heat flow and SWIT, the PWD defines the energetic conditions for the temperature development of all layers and is important for the source rock and for the maturation of organic matter through time. It was decided to not employ any PWD values/maps and to use the default value for all scenarios (i.e. present-day depth). However, we are aware that future studies should aim to improve this aspect of our models.

Scenarios and sensitivities

The workflows described above have allowed us to successfully create a model of petroleum generation and migration within the Nuevo Sub-basin, testing the effect of the igneous intrusions on petroleum generation and migration. The different parameters used and the intrusion thicknesses are given in Table 3, and are shown in Figures 5 and 6. In total, 18 scenarios are shown and discussed in this paper (Table 3). All scenarios assumed that the intrusion emplacement and crystallization occurred at 55 Ma. This subsection describes the reasons for the variations seen in the models.

One key uncertainty within the interpretation of subsurface intrusions within sedimentary basins is the thickness of seismically unimaged material present within the subsurface. Mark et al. (2019) estimated that within the FSB the ratio of imaged v. non-imaged sills is 1:1.4–1.6. According to this proportion, this means, for example, that for every 50 m-thick intrusion there is another 70–80 m of intrusions (with individual intrusions ranging in thickness from centimetres to metres) that are not imaged.

So that the potential effect of the intrusions can be correctly taken into account and the key sensitivities assessed, three different intrusion thickness cases were set-up within the model (Figs 5 and 6) and the results analysed. These three cases are: (1) a low-case in which the intrusions are the same as they appear in the seismic line (total intrusion thickness of c. 200–300 m); (2) a medium-case in which several artificial intrusions are added (total intrusion thickness of c. 500 m); and (3) a high-case where a very large number of intrusions has been added (total intrusion thickness of c. 1000 m). It should be noted that seismic tuning, which affects the intrusions, has the potential to result in thickness overestimates where intrusions are at or close to the limit of detectability. However, it is unlikely that there would be a net overestimate of sill thickness across the entire study area, so one would still expect the seismically mapped intrusions to represent a minimum thickness. Using these three cases, multiple scenarios were run in PetroMod, and the results of these scenarios were compared against a control scenario with no intrusions, in which the Cretaceous sequence is purely sedimentary, as assumed in previous basin modelling studies.

All 18 scenarios described in this paper have the same basement/crust (e.g. basement with RHP as described in Gardiner et al. 2019), thermal boundary conditions, the same present-day bathymetry through time, lithologies and number of faults, and they differ only for intrusion thickness and intrusion CEP values and fault properties (i.e. open or closed faults). In this way a better understanding of the sole impact of the intrusions on petroleum generation and migration can be achieved.

Effects of solidified dolerite on thermal conductivity of sedimentary successions

First, we investigated the impact of the change of thermal properties of the Cretaceous sequence due to the crystallization of dolerite. Thermal conductivity expresses the ability of a rock to conduct heat, and therefore it is an important factor for heat conduction in sedimentary basins (Hantschel and Kauerauf 2009; Allen and Allen 2013). The value of the thermal conductivity is an intrinsic property of any given rock (chemical composition, mineralogy and structure). For a given heat flux, when the thermal conductivity of a rock sequence is low, the geothermal gradient is generally high. As a corollary, when a rock sequence has a high thermal conductivity, the geothermal gradient is generally low. Within the study area, the Cretaceous sequence that have been intruded by dolerite sills will experience an overall increase in thermal conductivity post-emplacement because, once solidified, dolerite has a higher thermal conductivity (i.e. 2.30 W m−1 K−1 at 20°C for dolerites) than the fine-grained (mostly shale) Cretaceous sedimentary rocks (i.e. 1.73 and 2.05 W m−1 K−1 at 20°C for the Upper and Lower Cretaceous sequences, respectively).

Finally, the transformation ratio (TR) and vitrinite reflectance (VR) of the source rock were modelled and analysed. A pseudowell at the location of well 204/13-1 (Tornado) (named the Tornado pseudowell) through the whole basin sequence was created. TR is the ratio of the petroleum (oil and gas) generated by the kerogen to the total amount of petroleum that the kerogen is capable of generating. TR is a parameter that quantifies the progress of petroleum generation, and depends on the nature of the organic material (kerogen type) and thermal history associated with the subsequent geological history of a basin (e.g. Tissot and Welte 1984). In this study we simply assumed that with a TR value of 15% for a Type II kerogen (marine clastic), petroleum expulsion from the source rock begins (Scotchman et al. 2006; Gardiner et al. 2019). Vitrinite reflectance (VR) values were modelled using the Sweeney and Burnham (1990) kinetic Easy%Ro; they are a function of time and temperature, and it is a major (measurable) indicator of thermal maturity of the source rock (e.g. Tissot and Welte 1984; Hantschel and Kauerauf 2009; Allen and Allen 2013).

Expulsion onset

The expulsion onset, which is defined as the time of the first expulsion of petroleum from the source rock, was analysed for the different modelling scenarios. Since heat flow history, lithologies and kinetics were kept the same in all modelling scenarios, the petroleum expulsion onset differences observed here simply became a function of the intrusion thickness, which is therefore the key variable in the models. Geochronological dating of oils based on Re–Os geochronological dating suggests that petroleum generation began between c. 90 and 68 Ma (Finlay et al. 2011). After running a petroleum expulsion model along the 2D section, the results are discussed in relation to the generation age derived from the geochronological dating of oils, along with the deposition and formation of other petroleum system elements.

Petroleum saturation and migration

One key objective of this study was to evaluate the impact of intrusion thickness and CEP of dolerites on hydrocarbon migration. To account for the detailed distribution of igneous intrusions within the Upper Cretaceous sequence, the Upper Cretaceous sequence requires a high vertical subdivision by the integration of sublayers. For this study, we selected the combined migration modelling method (i.e. Darcy flow is applied for low-permeability lithofacies and invasion percolation is applied for high-permeability lithofacies: for more details see Hantschel and Kauerauf 2009; Baur and Katz 2018). Permeability was used as a threshold for this domain decomposition using the default setting of 2.01 log(mD) at a porosity of 30% (Hantschel and Kauerauf 2009). By default, a facies with such properties is considered a flow-path layer (i.e. migration is solved mathematically by invasion percolation when using the ‘combined migration method’: Baur and Katz 2018). Facies with a permeability of less than 2.01 log(mD) at 30% porosity represents Darcy cells where migration is calculated according to Darcy's law. Therefore, for the low-permeability lithofacies, we selected the refined Darcy flow (high-resolution Darcy flow with medium time resolution) and a mobility factor of 2 (Baur and Katz 2018). One advantage of this migration method is the ability to model migration and accumulation at a higher resolution than what is provided by the grid cells.

The impact of the hydraulic properties of faults and intrusions on petroleum migration is explored by analysing the petroleum saturation. Faults were considered to be either open (i.e. allowing fluid flow) or closed (i.e. not allowing fluid flow) in scenarios with no intrusions and with intrusions (low-, medium-, and high-case scenarios) representing end-member scenarios. It should be noted that in the cross-section all the faults were considered to have the same hydraulic properties for the same scenario, and a mix of properties is not considered (i.e. all faults are either open or closed but not a combination of open and closed in the same scenario). This represents a limitation of this study but these end-member scenarios have been chosen to compare more easily the impact of sill intrusion thickness and their sealing efficiencies on petroleum migration.

The impact of the intrusion hydraulic properties on petroleum migration (low-, medium-, and high-case intrusion scenarios) was also explored by creating scenarios with varying CEP for the dolerites. CEP curves are one of the most important properties in characterizing the dynamic behaviour of a lithology (Smith 1966; Berg 1975; Schowalter 1979; Vavra et al. 1992). These curves are widely used and are the industry standard for well-known sandstone–shale systems. Unfortunately, because of a lack of data in the literature that precludes a better characterization, the CEP curves of igneous rocks, in general, and dolerites, in particular (which are the focus of this study) are considered by default to be impermeable perfect seals (i.e. infinite CEP values), and therefore any breakthrough and migration through them cannot occur. In addition, the default properties consider them as not compactable, with porosity of 1% and with no permeability value provided (i.e. dolerite is considered as an impermeable rock). Although different and higher values of porosity and permeability can be used, if the intrusions are still considered to be impermeable rocks (i.e. infinite CEP values) then any breakthrough and migration will still not occur.

Real-world observations of dolerites suggest that they can allow fluid flow or act as reservoirs (e.g. Schutter 2003; Bermúdez and Delpino 2008; Rateau et al. 2013; Senger et al. 2017; Schofield et al. 2020). In order to account for the possibility that sills could permit fluid flow, different CEP values are used and different sensitivities explored. The default value of porosity for the dolerite sills was kept constant (i.e. 1%) and a permeability value of 1 mD was used. These values were chosen because, although it has been shown that in certain circumstances intrusions in the FSB appear to have substantial fracture porosity and permeability (e.g. 214/28-1: Schofield et al. 2020), it is also recognized that many intrusions could potentially have mineral-filled fractures, and are therefore likely to have relatively low porosity and permeability characteristics. We particularly wanted to investigate the effect of the intrusions on petroleum migration even in low-porosity and low-permeability ‘worst-case’ scenarios.

While keeping porosity and permeability constant (1% and 1 mD, respectively) within the dolerite intrusions, the CEP values of the mercury–air system were systematically changed. These values were 25 MPa (3625.5 psi) (equivalent to a shale (typical) with c. 12% porosity), 50 MPa (7251.9 psi) (slightly higher than the highest value of shale (typical)), 100 MPa (14 503.8 psi), 250 MPa (36 259.4 psi) and 500 MPa (72 518.9 psi). A value for CEP of 1000 MPa (145 04 psi) gave a similar result to that of a perfect seal. For brevity only, the results with the default CEP curve (i.e. impermeable perfect seal) and with CEP values equal to 100 MPa (14 503.8 psi) for all scenarios and 50 MPa (7251.9 psi) and 25 MPa (3625.5 psi) for the medium-case intrusion scenario are shown and discussed. It should be noted that it was assumed that in the same scenario, all sills have the same hydraulic properties. Since no wells have drilled the sills in the immediate study area, it is hard to estimate such hydraulic properties and therefore it was preferred to keep the hydraulic properties of sills the same in the same scenario to avoid introducing artificial and unconstrained bias.

In this section we show and analyse the results of our simulations that allow us to evaluate and quantify the impact of sill intrusions on petroleum generation and migration. We focus on describing the results within the Nuevo Sub-basin (i.e. the southwestern part of the cross-section), which is where the intrusions were emplaced. Since the area between the Cambo discovery and southern Rosebank (i.e. the northeastern part of the cross-section) has the same sedimentary sequence in all scenarios (with no significant intrusion thickness) and the burial and thermal history of the source rock is the same (i.e. no overthickening), all the simulation results for this area are similar in all scenarios and are therefore not discussed.

Thermal evolution of the basin

Several simulations using the McKenzie tool were carried out using the scenario with low-case intrusion scenario, open faults, and intrusions treated as perfectly sealing intrusions, with basal heat flow and assuming, as already suggested by Gardiner et al. (2019), three separate rifting events: early Cretaceous (145–130 Ma), Albian (110–100 Ma) and Campanian (85–80 Ma). This helped us to obtain the basal heat flow history up to the present-day, which allowed matching of the previously corrected present-day bottom-hole temperatures and vitrinite reflectance values at wells 204/10-1 (Cambo) and 204/13-1 (Tornado) (Fig. 7). It should be noted that the thermal calibration was only carried out for the low-case intrusion scenario with open faults and intrusions treated as perfectly sealing intrusions because this is the result of seismic interpretation, and it can be considered to be the base case.

The temperature and the heat flow values along the 2D cross-section are shown and analysed 1000 years (i.e. 55.0010 Ma) prior to and at the time that we assumed that intrusions were emplaced and solidified (i.e. 55 Ma) (Figs 8 and 9). The isotherms before the intrusion event (Fig. 8) are approximately parallel to each other with increasing depth, suggesting that generally temperature increases with depth in a constant fashion (i.e. suggesting a normal geothermal gradient). The top and base of the Cretaceous sequence was derived by seismic interpretation and several scenarios with different intrusion thicknesses and, thus, also different sedimentary proportions of the Cretaceous sequence were created. This means that in the high-case intrusion scenario, up to the intrusion event, the Cretaceous sequence and older layers were shallower than in the medium-, low-case, and no intrusions scenarios. When the intrusions are solidified at 55 Ma (Fig. 9), the Cretaceous sequence comprises a combination of sedimentary layers with low thermal conductivity (i.e. 1.73 W m−1 K−1 at 20°C) and intrusions with higher thermal conductivity (i.e. 2.30 W m−1 K−1 at 20°C). The increase in thermal conductivity results in a lowering of the geothermal gradient (Fourier's law of thermal conduction). This phenomenon is well known in basins with salt layers and domes that have very high thermal conductivity (c. 6 W m−1 K−1 at surface conditions) (e.g. O'Brien and Lerche 1987; Mello et al. 1995; Allen and Allen 2013; Canova et al. 2018; Mangione et al. 2018). It should be noted that we considered the thermal effect of intrusions after they solidified to dolerite without considering the extra temperature increase because of magma solidification.

Therefore, the distance between the isotherms increases around the intruded Cretaceous sequence to reflect a local decrease in the geothermal gradient (e.g. beneath well 204/13-1 (Tornado)). In the scenario with no intrusions, the 120 and 140°C isotherms are at depths of c. 4.4 and 5.1 km, respectively, whereas in the high-case intrusion scenario these isotherms are deeper at c. 4.9 and 6 km, respectively. This effect is also evident with the rock column heat flow behaviour after the intrusion crystallization: in fact, for a scenario with more overthickening, the local predicted heat flow within the rock column is lower. Since this effect generated a decrease in geothermal gradient, it affected, together with the effect of the overthickening, the source rock maturation as well as all the other temperature-dependent modelling results (e.g. modelled VR and TR). For example, the heat flow within the rock column values against the scenario with no intrusions decreases at depths of c. 3961 and 5597 m, giving values of c. 1–5, 10–15 and 22–25 mW m−2 for the low-, medium-, and high-case intrusion scenarios, respectively, with open faults and perfectly sealing intrusions (Fig. 9). We also noted that the sedimentary part of the Cretaceous sequence contains fine-grained sediments (mostly shale) that have higher RHP values than dolerites. In the scenarios with overthickening, the total thickness of Cretaceous fine-grained (mostly shale) sediments is less (i.e. the amount of additional heat from the radiogenic decay of Cretaceous sediments is lower) and therefore the more overthickening (i.e. the more contribution from igneous intrusions we assume to be present between the top and the base of the Cretaceous surfaces), the less radiogenic heat input from Cretaceous fine-grained sediments. Hence, both the relatively higher thermal conductivity of the dolerites and the reduced RHP from the Cretaceous fine-grained (mostly shale) sedimentary rocks result in lower temperatures together with the effects of reduced thickness of the Cretaceous sequence before the intrusion emplacement.

After the igneous intrusions cooled down and crystallized, their high thermal conductivity with respect to the fine-grained Cretaceous reduced the geothermal gradient. To estimate the duration of the reduced geothermal gradient due to the intrusion crystallization, we analysed the duration of the perturbed geothermal gradient following sill emplacement through the Tornado pseudowell by investigating the temperature profiles and values on the top and the base of the Upper Jurassic source rock at different time steps (i.e. at 54.9 Ma, which is after the intrusion event and crystallization, and at 33.9, 5.3 and 0 Ma) for the scenarios with no intrusions, low-, medium-, and high-case with open faults and perfectly sealing intrusions (Fig. 10). By comparing the scenario with no intrusions, the temperature of the top and base of the source rock at 54.9 Ma decreases by c. 6, 18–19 and 29–31°C for the low-, medium-, and high-case scenarios, respectively. At 33.9 Ma, the decrease is reduced to c. 2, 3 and 7°C for the low-, medium-, and high-case intrusion scenarios, respectively. At 5.3 Ma, the differences are very similar to the differences at 33.9 Ma and this difference has remained approximately the same until the present-day. This suggests that the increase in thermal conductivity results in a temperature decrease that is still important at 33.9 Ma and, although it decays at 5.3 Ma, the difference still exists. It should be noted that the palaeodepth of the top and base of the source rock is not the same in the different scenarios because of different compaction behaviour between the sedimentary component of the Cretaceous succession and the intrusions.

Petroleum expulsion onset timing

The temperature decrease impacts other thermally dependent properties, such as the modelled VR and TR of the source rock. The modelled VR and TR profiles were extracted from the Tornado pseudowell for the scenarios with no intrusions, low-, medium-, and high-case, with open faults and with perfectly sealing intrusions (Fig. 11). According to Finlay et al. (2011), petroleum generation began between c. 90 and 68 Ma. In the scenario with no intrusions, at 68 Ma, the modelled VR values indicate that the source rock was already in the late oil window (i.e. VR values between 1.00 and 1.30%). However, by considering the intrusions, the onset of early oil generation (i.e. modelled VR = 0.55%) is postponed by up to 0.8, 3, and 6 myr for the low-, medium-, and high-case intrusion scenarios, respectively. The end of the early oil phase of maturity and onset of the main oil window (i.e. modelled VR = 0.70%) is postponed by up to 2, 7, and 13 myr for the low-, medium-, and high-case intrusion scenarios, respectively. Finally, the end of the main oil window and onset of the late oil window (modelled VR = 1.00%) is postponed by up to 3, 8, and 12 myr for the low-, medium-, and high-case intrusion scenarios, respectively. Hence, when intrusions are factored into the basin modelling, the source rock maturation broadly matches data from the geochronological dating of oils and the deposition and formation of other petroleum system elements.

Analysis of the predicted TR indicates that the time at which TR = 15% is attained (which we simply assume to represent the approximate onset of petroleum expulsion from the source rock) is postponed when the overthickening increases. The scenario with no intrusions and the low-case intrusion scenario have similar transformation ratio curves, with TR = 15% postponed in the latter by c. 0.6 myr. However, the medium- and high-case intrusion scenarios show that the time at which TR = 15% occurs is postponed by up to 3 and 8 myr, respectively. In addition, and probably more importantly, the medium- and high-case intrusion scenarios, and to a lesser extent the low-case intrusion scenario, show that most of the source rock was still able to expel petroleum during the c. 90–68 Ma interval and before deposition and formation of other petroleum system elements, whereas the scenario with no intrusions indicates values as high as TR = 70% at c. 80–70 Ma.

The analysis described above clarifies that the degree of overthickening impacts both the burial and thermal history of the source rock. A comparison of the timing for petroleum expulsion onset in different scenarios containing open faults and perfectly sealing intrusions helps to quantitatively address the impact of overthickening on these uncertain aspects of basin evolution.

In the scenario with no intrusions, most of the source rock in the Nuevo Sub-basin expelled petroleum between c. 95 and 85 Ma, earlier than suggested by the geochronological dating of oils (i.e. c. 90–68 Ma) along with the deposition and formation of other petroleum system elements (Fig. 13; Table 3). Hence, to explain the presence of accumulations in this region, other mechanisms such as temporary storage in intermediary reservoirs and/or overpressure need to be invoked (as discussed earlier). Proximal to the Cambo High where the source rock is shallower (i.e. present-day depth of c. 6.8–7.6 km), modelling results suggest that expulsion commenced at c. 81–68 Ma, whilst source rock in the deepest part of the basin (i.e. present-day depth of c. 8.4–8.8 km) started expelling petroleum at c. 100–95 Ma.

In the low-case intrusion scenario, the source rock in the Nuevo Sub-basin started expelling petroleum between c. 93 and 84 Ma, up to c. 2 myr later than in the scenario with no intrusions. This suggests that even a small number of intrusions can affect the modelled expulsion history (Fig. 12; Table 3). However, one can also argue that given the modelling uncertainties, this difference can be neglected. As observed for the scenario with no intrusions, the modelled section shows a variation in the onset of expulsion close to the Cambo High at c. 80–67 Ma, which again is up to c. 2 myr later than in the scenario with no intrusions. Source rock in the deepest part of the basin (i.e. present-day depth of c. 8.4–8.8 km) started expelling petroleum at c. 100–95 Ma in the low-case intrusion scenario.

In the medium-case intrusion scenario, the source rock in the Nuevo Sub-basin started expelling petroleum at c. 90–82 Ma, which is up to c. 5 myr later than in the scenario with no intrusions. As for the scenarios described above, the modelled section shows a variation for the beginning of expulsion onset in the area close to the Cambo High, with expulsion onset values occurring within the c. 78–57 Ma interval, which is up to c. 11 myr later than in the scenario with no intrusions. Similarly, source rock in the deepest part of the basin (i.e. present-day depth of c. 8.4–8.8 km) started expelling petroleum at c. 100–95 Ma (Fig. 12; Table 3).

In the high-case intrusion scenario, the source rock in the Nuevo Sub-basin started expelling petroleum at c. 87–75 Ma, which is up to c. 10 myr later than in the scenario with no intrusions (Fig. 12; Table 3). As for the scenarios described above, the modelled section shows a variation for the onset of expulsion in the area close to the Cambo High, with expulsion onset occurring during the c. 75–56 Ma interval. This is up to c. 12 myr later than in the scenario with no intrusions and, as in previous scenarios, source rock in the deepest part of the basin (i.e. present-day depth of c. 8.4–8.8 km) started expelling petroleum during c. 100–95 Ma. The consistency of results across all modelled scenarios indicates that the source rock in the deepest part of the basin (i.e. present-day depth of c. 8.4–8.8 km) generated petroleum prior to igneous activity. Importantly, the estimated onset of expulsion in the medium- and high-case scenarios is compatible with the geochronological dating of oils. For the other scenarios, only the shallower areas (i.e. areas closer to the Cambo High, present-day depths of c. 6.8–7.6 km) have an onset of expulsion that is compatible with the geochronological dating of oils.

Effects of intrusions on petroleum migration

In this subsection we analyse the petroleum saturation resulting from migration in our modelled scenarios. The key parameters we explore are fault properties representing open conduits for fluid migration or complete seals and the CEP of sills (using values of infinite CEP (i.e. intrusions treated as impermeable perfect seals), 100, 50 and 25 MPa).

Generally, we observe from our models that when faults are open they allow fluid flow to occur, enabling the transfer of deep basin fluids towards shallower areas (see also Holdsworth et al. 2019) and pressure release, so that there is, perhaps, less likelihood of overpressure build up, with implications for compaction that, in turn, reduces the porosity and permeability of the layers (e.g. the Cretaceous sequence). When intrusions are treated as impermeable seals they are able to trap petroleum both beneath and within the intrusive complex. This depends on the amount of overthickening, as discussed below (Fig. 13; Table 3). We note that with increasing total intrusion thickness, the depth at which high (defined as 50% in this study) petroleum saturation occurs increases from a present-day depth of c. 7–7.1 km in the low-case intrusion scenario to a present-day depth of c. 7.1–7.5 km in the high-case intrusion scenario (see the arrows in Fig. 13). The fault bounding the SW margin of the Cambo High (see the red box in Fig. 13) shows petroleum saturation higher than the 50% around it, suggesting that it could have acted as a focal point for migrating fluids before the migration of petroleum towards well 204/10-1 (Cambo). The petroleum saturation around this fault is very similar in all modelled scenarios, suggesting that Cambo High could have received charge from source rock in the Nuevo Sub-basin, irrespective of the intrusion thickness and their consequent impacts on migration tortuosity.

Because in all likelihood igneous intrusions are likely to have allowed some fluid flow, we also explored a scenario where their CEP was reduced to 100 MPa (mercury–air system). This allows some fluid flow through the intrusions, although some petroleum is still trapped within the intrusions and some beneath (Fig. 14; Table 3). Across all scenarios containing intrusions, the modelled volumes of petroleum trapped within the intrusive complex are higher when CEP is assumed to be 100 MPa, and similarly the depth at which high petroleum saturation is observed is shallower (a few hundreds of metres) than when intrusions are assumed to be perfect seals (see the arrows in Fig. 14).

When faults are modelled to be closed there is less chance of vertical fluid migration, which can contribute to overpressure development. Overpressure leads to undercompaction and higher porosity values compared to the scenarios with open faults. It should be noted that sealing rocks also have an important role in obtaining significant overpressure. According to pressure values, buoyancy and CEP, migration through the Cretaceous sequence depends on a complex interplay between the spatial distribution of sediments, intrusions and the CEP of intrusions (i.e. perfect seal or lower values). We noted that when increasing the total intrusion thickness through the modelled scenarios with closed faults, increasingly more petroleum is trapped within the sill complex, suggesting that migration could have been very tortuous (i.e. the more intrusions, the more tortuous the migration pathways) (Fig. 15; Table 3). When CEP is changed from the scenarios with perfectly sealing intrusions to 100 MPa, more cells are predicted to have been saturated with petroleum within the sill complex (Fig. 16; Table 3).

The scenarios described above assume that sills are either perfect seals or that they have CEP values of 100 MPa. One can speculate that 100 MPa is still probably too high because some sills could have fracture permeability (e.g. Bermúdez and Delpino 2008; Rateau et al. 2013; Senger et al. 2015; Mark et al. 2018; Schofield et al. 2020) and potentially have a lower CEP.

We explored the impact of varying the CEP to 50 and 25 MPa (roughly equivalent to a shale (typical) with c. 12% porosity) for the medium-case intrusion scenario with faults both open and closed. When faults are open, the pattern of petroleum saturation within the sill complex is very similar whether using a CEP of 50 or 25 MPa, although locally, saturation can change slightly (Fig. 17; Table 3). It should be noted that this pattern is similar but with local changes compared to the scenario with a CEP of 100 MPa. Finally, we noted that the deepest cells that have high petroleum saturation are shallower (a few hundreds of metres) in the scenarios with a CEP of 100 MPa. When faults are closed, more petroleum is trapped in the scenario with the CEP equal to 50 MPa than in the equivalent scenario with the CEP equal to 25 MPa, because the intrusions are less hydraulically conductive to flow when CEP is 50 MPa. Therefore, when CEP increases, the sill complex as a whole is able to trap more petroleum.

Several scenarios were created by changing the overthickening, intrusion CEP, and fault properties to test several potential petroleum expulsion and migration modelling scenarios for basins that contain extensive subsurface sill complexes such as the FSB (Fig. 18). The workflow in this paper, which considers the impact of intrusions on petroleum generation and migration, should be considered together with previous models applied to the FSB that invoked intermediary accumulations (i.e. ‘motel model’: Doré et al. 1997; Lamers and Carmichael 1999; and ‘whoopee cushion effect’: Iliffe et al. 1999) and overpressure within Jurassic source rock that was assumed to retard petroleum generation (Carr 1999; Carr and Scotchman 2003; Scotchman and Carr 2005; Scotchman et al. 2006). For our analysis, we focused on petroleum saturation mainly within low-permeability rocks types that have been simulated with Darcy flow. In our model set-up, for petroleum migration and accumulation, the layers with high-permeability (e.g. Flett Formation, part of the Lamba Formation and Oligocene rocks) have been modelled with the rather instantaneous invasion percolation method. The accumulation pattern within these high-permeability rock types was very similar for all the scenarios due to the predicted very low sealing efficiency of the overlying sequences. In addition, it should be noted that our focus was to evaluate the above mentioned uncertainties on potential petroleum expulsion and migration. A detailed analysis of the 3D processes affecting pressure regime, petroleum charge, and model calibration against known petroleum accumulations was out of the scope of this study.

Effect of the intrusion emplacement and crystallization on thermal conductivity of the units overlying the source rock and on the thermal evolution of the basin

After the dolerite solidified, it had higher thermal conductivity than the Cretaceous fine-grained sedimentary host rocks. Therefore, the whole layer (i.e. sedimentary Cretaceous and the igneous intrusions) has a higher thermal conductivity than before the intrusion event. Our models clearly illustrate that the increase in thermal conductivity of the Cretaceous sequence overlying the Kimmeridgian source rock after magma solidification in the Nuevo Sub-basin results in a lower geothermal gradient. The reduced geothermal gradient after the intrusions crystallized depends on the total thickness of the intrusions and is most pronounced at the time closest to the intrusion event (but still after magma solidification) and then decreases towards present-day.

The reasons for this decrease from when the intrusions solidified towards the present-day are potentially a combination of: (1) with increasing burial depth, the Cretaceous sedimentary rocks became more compacted, resulting in lower porosity and higher thermal conductivity, and less contrast in thermal conductivity with the intrusions (this effect, however, is different for the open v. closed fault scenarios because they impact porosity evolution); (2) in the case in which the porosity of the Cretaceous sequence decreases with increasing burial depth, the volume available for storage of liquid and gaseous phases decreases; as these phases have a lower thermal conductivity than the sedimentary units, the thermal conductivity of the units increases when the volume available for liquid and gaseous phases decreases; and (3) reduction of radiogenic heat production with time.

Our simulations indicate that, in general, with respect to the scenario with no intrusions, the heat flow within the rock column decreases for the low-, medium-, and high-case intrusion scenarios with open faults and perfectly sealing intrusions, respectively. This is due to a combination of overthickening and radiogenic heat production from Cretaceous shales. A decrease in temperature of the top and base of the source rock can also be observed along the Tornado pseudowell. When the intrusion thickness increases (i.e. from the low- to the high-case) the thermal conductivity of the whole Cretaceous layer increases. Although the majority of petroleum has already been generated at this point, in areas where this is not the case, this new temperature regime, together with other parameters such as the porosity of the sedimentary Cretaceous and overpressure, can influence petroleum generation.

With respect to the scenario with no intrusions, the time at which the source rock entered the early oil window (i.e. modelled VR = 0.55%) is postponed by up to 6 myr, the onset of the main oil window (i.e. modelled VR = 0.70%) is postponed by up to 13 myr and the onset of the late oil window (i.e. modelled VR = 1.00%) is postponed by up to 12 myr. The time at which the source rock reaches a value of TR = 15% in the high-case intrusion scenario is postponed by up to 8 myr with respect to the scenario with no intrusions, and most of the source rock was still able to release petroleum during the c. 90–68 Ma interval.

Although in our study area some of the timing of intrusions is well constrained (e.g. Ritchie et al. 2011; Schofield et al. 2015, 2017a), in other areas of the FSB the intrusions can be potentially older. Therefore, one can speculate on the effect on the temperature evolution if the intrusion event had occurred around the end of the Cretaceous (i.e. 66 Ma). It is possible that the reduction in geothermal gradient after the intrusion solidification would have been higher assuming an older intrusion event, as the Cretaceous sequence was shallower, more porous, and, therefore potentially holding more liquid and gaseous phases, and, thus, was therefore less thermally conductive. Porosity, and therefore the burial depth of the Cretaceous sequence at the time of the intrusion emplacement and solidification, is critical but often difficult to constrain. It should be also noted that thermal conductivity is a function of temperature (e.g. Clauser and Huenges 1995; Chen et al. 2021). In addition, when sills are considered to be perfect seals, they do not allow fluid to flow, which could potentially generate overpressure in the Cretaceous sequence. Overpressure is known to potentially result in undercompaction (e.g. Taylor et al. 2010; Stricker and Jones 2018), which can result in a reduction of thermal conductivity.

In our models we did not consider any uplift connected with the intrusion event, only subsidence of the Cretaceous sequence and the underlying layers (i.e. after the intrusion event all the layers below the intrusion are moved downwards). In this way, the burial history and the resulting thermal history of the layers below the intrusions are affected. In addition, we did not consider contact metamorphism (i.e. creation of metamorphic aureoles around the intrusions: e.g. Aarnes et al. 2011; Mori et al. 2017), which would lead to different lithological and hydraulic properties that can impact petroleum migration. It should be noted that although PetroMod does not provide an automatic way to calculate metamorphic aureoles around the intrusions, the user can change the lithologies accordingly. However, we decided to keep a simpler workflow and not consider contact metamorphism.

Impact of intrusion emplacement on petroleum expulsion onset

Our modelling helps to quantitatively address the effect of the overthickening on the petroleum expulsion history of the source rock and suggests that intrusions affected the onset of petroleum expulsion in two ways: (1) the thermal conductivity of the Cretaceous sequence increased after the intrusions solidified because of the high thermal conductivity of the dolerite, which results in a lower geothermal gradient; and (2) the recognition that the addition of igneous material during the Paleogene led to overthickening of the Cretaceous sequence impacting the burial history of the underlying source rock. It should be noted that in all the modelled scenarios, in the deepest part of the basin (i.e. present-day depth of the source rock at c. 8.4–8.8 km), including and/or varying the intrusion thickness in the modelled scenarios did not have an important impact on expulsion timing, probably because at the time of the intrusion event the source rock was already too deep.

One can speculate that petroleum generated in these areas could either have migrated into other parts of the basin or accumulated in temporary reservoirs and subsequently remigrated (e.g. Doré et al. 1997; Iliffe et al. 1999; Lamers and Carmichael 1999), or that overpressure could have retarded petroleum generation (e.g. Carr 1999; Carr and Scotchman 2003; Scotchman and Carr 2005; Scotchman et al. 2006). With the exception of these deep basinal levels, the entire source rock in the medium- and high-case intrusion scenarios, and to a lesser extent the low-case intrusion scenario, has a modelled expulsion onset time that is compatible with the geochronological dating of oils along with the deposition and formation of other petroleum system elements, considering the given uncertainties (e.g. the source rock facies can vary in different parts of the basin). The expulsion onset is retarded by up to 12 myr in comparison to the scenario with no intrusions. Hence, our modelling shows the important impact of the emplacement of the intrusive complex in the FSB. Additional mechanisms such as overpressure and secondary migration could have important roles, and should also be considered.

Petroleum migration: is a large fraction of petroleum trapped beneath (or within) the sill complex?

The presence of an extensive intrusive complex in strata overlying Kimmeridgian Jurassic source rock in the FSB raises the question as to what effect (if any) the intrusions might have had on vertical petroleum migration. Igneous intrusions in general within sedimentary basins are regarded as having little primary porosity and permeability (e.g. Bermúdez and Delpino 2008; Rateau et al. 2013), and have been shown to act as subsurface fluid barriers by Grove et al. (2017). However, secondary porosity and permeability can be created by fracture systems within the intrusions themselves (either due to cooling or subsequent tectonic deformation: Bermúdez and Delpino 2008; Rateau et al. 2013; Schofield et al. 2020). Therefore, it is possible that in a scenario where intrusions have fracture permeability, the hydraulic conductivity of the fracture system will be the primary control on migration within the intrusions and in the host rocks surrounding the intrusions (e.g. Senger et al. 2015).

The intrusions in the FSB are primarily dolerites (Schofield et al. 2017a; Mark et al. 2019) and, by default, PetroMod treats them as perfect seals. However, previous work has shown that factors such as the burial depth, intensity of faulting, diagenesis, and the thickness of the intrusions control the hydraulic properties of sills (Rateau et al. 2013). This means that in the same basin some sills allow fluid flow, while others can act as barriers (Rateau et al. 2013; Mark et al. 2018; Schofield et al. 2020). Sills buried at depths greater than 5 km can have open fractures that may potentially allow transmission of overpressure from deeper parts of the basin (Schofield et al. 2020). Since the intrusions within the FSB have not been substantially uplifted, the fractures are probably the result of cooling and contractional processes that occurred at the time of emplacement (Schofield et al. 2020). The presence of fractures is supported by substantial mud losses when drilling through the intrusions (e.g. of the 29 wells that have penetrated intrusions in the FSB, potentially more than 80% have shown mud losses when drilling through the intrusions) (Rateau et al. 2013; Mark et al. 2018). In addition, as already pointed out by Schofield et al. (2020), Rateau et al. (2013) and Schofield et al. (2015, 2017b), open fractures within the intrusions could have potentially controlled the location of oil and gas accumulation because the intrusions were more permeable (potentially acting as migration ‘superhighways’) than the encasing, low-permeability Cretaceous sedimentary sequence. Hence, our study builds on previous work in the FSB and highlights the importance of an accurate characterization of the sill complex both statically (i.e. their geometry and thickness) and dynamically (i.e. their fluid flow properties), as well as a representation of the hydraulic properties of the faults.

Despite the FSB hosting a good quality source rock with TOC values ranging from 4.4% (in the study area) to 6% and HI = 350 mgHC g−1 TOC (Iliffe et al. 1999; Gardiner et al. 2019) and excellent quality reservoirs, which over the Cambo High are characterized by up to 30% porosity and multi-Darcy permeability (Fielding et al. 2014; Purvis et al. 2020), the number of discoveries within the FSB has never matched that of the adjacent North Sea petroleum province. The differences in the apparent prospectivity of the North Sea v. the FSB can be attributed to a variety of reasons (including source rock presence and preservation). A key difference is that within many areas of the FSB, an extensive sill complex exists between the source kitchen and the reservoir rocks (Ritchie et al. 2011; Schofield et al. 2015; Mark et al. 2018). Our modelling results indicate that intrusions can be potentially responsible for trapping a large fraction of petroleum generated beneath or within the sill complex. This may be one factor contributing to the discrepancy in petroleum volumes discovered in the FSB with respect to the North Sea (aside from the likes of the Clair Field). One can hypothesize that the petroleum generated has been trapped below or within the sill complex, reducing the volume of the total charge to viable reservoirs.

Although our modelling is focused on one 2D cross-section, the modelled geology is highly representative of much of the FSB. This paper highlights, even on a simplistic basis, the role that intrusions can play in a petroleum system, namely the ability to change generation timing, and acting as barriers and baffles to subsequent fluid migration (Schutter 2003; Thomaz Filho et al. 2008; Rateau et al. 2013; Schofield et al. 2015, 2020; Senger et al. 2017; Belaidi et al. 2018; Gardiner et al. 2019; Trice et al. 2019). Given the number of basins globally containing igneous intrusions, (e.g. Brazil, offshore Norway, South Africa, Western Australia, offshore southern Australia, NW margin of Australia, West of Shetland, Norwegian Atlantic Margin, Greenland basins and offshore West Africa: Thomaz Filho et al. 2008; Ritchie et al. 2011; Svensen et al. 2012; Holford et al. 2013; Schofield et al. 2015; Magee et al. 2016; Mark et al. 2018; Reynolds et al. 2018; Schenk et al. 2019), the relationships illustrated in this paper have substantial consequences for the modelling of petroleum expulsion and migration in such regions. Unfortunately, existing basin modelling software might not be sufficient when we want to understand the impact of igneous intrusions in basins, as new workflows are needed to be able to more accurately represent basins that contain igneous intrusions.

Finally, the model is carried out in 2D, and therefore migration in the third dimension is not captured. But this work represents a first step towards better modelling of petroleum generation and migration in heavily intruded basins and aims to understand the basic challenges before moving towards more complex 2D or 3D models. Furthermore, making a 3D model with the same resolution used here would be computationally very expensive.

Factoring intrusion emplacement effects into basin modelling: do we really need it?

Our paper shows the impact of igneous intrusions on petroleum generation and migration, highlighting the need to include them during basin and petroleum system modelling. We also demonstrate that when the intrusion thickness is around 10% of the sedimentary layer (i.e. low-case intrusion scenario in which the intrusion thickness is c. 200–300 m and the Upper Cretaceous sequence is c. 2 km), the impact of the intrusions is still visible; however, we expect that for an intrusion proportion of less than 5–10%, the effect is potentially negligible (see also Schenk et al. 2019). Therefore, in areas where the intrusion thickness is low (e.g. in the southern Flett Sub-basin where the cumulative intrusion thickness is c. 50 m) the other models suggested in the literature must be invoked to explain the timing discrepancy between source rock maturation and the deposition of the working Paleogene reservoirs, as also acknowledged by Gardiner et al. (2019). It should be noted that if the source rock is already buried sufficiently to generate and expel oil and gas before the intrusion emplacement, the effect of intrusions on petroleum generation is potentially negligible. Thus, while we have highlighted the role of igneous intrusions within the basin modelling context, it is important to note that intrusions, where present, form a part of the rock record and need to be treated with comparable detail as the sedimentary rocks so that the model can better represent geological reality. It is also important to note that it is fundamental to consider the variation in the thermal conductivity of basin sequences both pre- and post-intrusion emplacement.

The Nuevo Sub-basin in the FSB has been heavily affected by igneous intrusions. The intrusion emplacement generated a post-depositional overthickening of the Cretaceous sequence that needs to be considered to correctly model the burial and thermal evolution of the source rock and the basin that contains it. Petroleum migration pathways through sills also need to be evaluated. We developed and applied a workflow to correctly take into account the effect of igneous intrusions. The results of our work show: (1) if the additional thickness of the intrusions is correctly placed into the Cretaceous sequence, the source rock started generating petroleum later than previously suggested by basin models because the source rock is shallower up to the time of the intrusion event; (2) the thermal conductivity in the Cretaceous sequence following intrusion emplacement and solidification increases, due to the large amounts of dolerite; (3) migration pathways were tortuous because of the presence of sills; and (4) petroleum saturation is the result of the combination of hydraulic properties of faults and intrusions because they affected the compaction history of the dominantly sedimentary Cretaceous sequence.

Our work has sought to build a more realistic basin model by including the intrusions and respecting the geological sequence of events, but it does not claim that the other mechanisms suggested in the literature to explain the complex generation and migration of petroleum are not valid. After honouring the geology within the limitations of any given model, other mechanisms can still be invoked to provide additional answers. It should also be highlighted that rarely one single mechanism is able to explain the complexity of natural systems, particularly in a tectonically and sedimentologically complex and heterogeneous geological province such as the FSB.

As a general conclusion of this work, basin models in heavily intruded areas such as the Nuevo Sub-basin, in particular, and the FSB, in general, need to consider the impact of the intrusions at a local scale. One should be very careful with generalizations at the basin scale because each Sub-basin needs to be investigated in detail because of the role of variable thicknesses of any intrusions. In this way reliable predictions can be made.

Finally, although with some limitations, this work is, to the best of our knowledge, the first to use seismic interpretation together with classical basin modelling data (e.g. bottom-hole temperature and VR) in the evaluation of 2D petroleum generation and migration in complex basins that have been heavily intruded by igneous intrusions.

The authors thank Siccar Point Energy E&P Ltd (now part of Ithaca Energy (UK) Ltd) and Shell UK who provided data for this project under the P2403 licence. PGS/TGS are thanked for allowing the author access to the 3D seismic cube FSB 2011–2012 (CS9: PP123DGFSB). The authors are grateful to Schlumberger for granting access to Petrel and PetroMod.

AM: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), visualization (lead), writing – original draft (lead), writing – review & editing (lead); NS: conceptualization (equal), data curation (equal), investigation (supporting), supervision (lead), visualization (equal), writing – original draft (supporting), writing – review & editing (supporting); SH: writing – original draft (supporting), writing – review & editing (supporting); CG: investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting); CE: investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting); CF: writing – original draft (supporting), writing – review & editing (supporting); OS: investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting); DG: writing – original draft (supporting), writing – review & editing (supporting); BH: investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting); LB: investigation (supporting), writing – original draft (supporting), writing – review & editing (supporting); JRU: writing – original draft (supporting), writing – review & editing (supporting).

Siccar Point Energy E&P Ltd (now part of Ithaca Energy (UK) Ltd) and Shell UK provided funding for this study.

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 data that support the findings of this study are available but restrictions apply to the availability of these data, which were used under licence for the current study and so are not publicly available.

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License (http://creativecommons.org/licenses/by/4.0/)