Distribution of the Triassic succession in the North Sea is poorly understood because of structural complexities associated with halokinesis and limited stratigraphic control. This study uses a seismic and well-based dataset to improve understanding of development of the Triassic succession in the Ula Field Area, in the Norwegian North Sea.

Core interpretation revealed a fluvial-dominated depositional environment in the Ula Field Area. Palynological studies allowed dating of cored intervals, revealing Ladinian and Carnian sections, time-equivalent to the Julius and Joanne members of the Skagerrak Formation. Well-log interpretation provided insight into the intra-Triassic stratigraphy of the Ula Field Area. A section considered to be equivalent to and extending from the Smith Bank Formation to the Jonathan Member of the Skagerrak Formation was interpreted and correlated across the area. In the proposed correlation, the Julius Member thins towards the Ula Field Area and is replaced by a time equivalent sandstone unit. The Jonathan Member displays a sandier composition compared to the equivalent section in the UK sector. Seismic facies-based interpretation of Triassic stratigraphy within salt minibasins allowed recognition and mapping of intra-Triassic units and showed that mudstone members thin towards the NE. Interpreted internal geometries within minibasins allowed determination of the timing of halokinesis. Integration of different datasets allowed palaeogeographic reconstructions for the Anisian, early Ladinian, Carnian and Norian to be constructed. To conclude, the distribution of stacked fluvial channel deposits indicates that they occur both within minibasins and across salt highs such that ongoing halokinesis had no topographic expression and that channels were free to migrate across the area.

The Triassic Skagerrak Formation in the North Sea is a continental succession deposited as a series of distributive fluvial systems (McKie and Williams 2009) developed under semi-arid climatic conditions (Goldsmith et al. 2003). It is composed of alternating sandstone and mudstone dominated members interpreted to reflect climatic changes between arid (mudstone dominated members) and humid (sandstone dominated members) conditions (Archer et al. 2010; McKie 2014; Burgess et al. 2021, 2022). The current study aims to improve understanding of the controls on the deposition and distribution of the Triassic succession and the Skagerrak Formation in particular on the Cod Terrace, focusing on the Ula Field Area in the southeastern part of Norwegian Quadrant 7 (Fig. 1). Both the Jurassic and Triassic deposits in the Ula Field are strongly influenced by salt activity. The trap configuration is a salt anticline formed during the Late Cretaceous following inversion (Zanella and Coward 2003; Mannie et al. 2014). Remaining opportunities for extending field life are primarily in the Triassic section and a key issue is predicting reservoir presence, distribution and quality in the area.

Sedimentation in the Central North Sea was strongly affected by salt movement from the Early Triassic onwards (Zanella and Coward 2003) and resulted in the formation of salt structures separated by minibasins. Early papers viewed the Skagerrak Formation as being purely controlled and confined by halokinesis with sedimentation restricted to minibasins (Hodgson et al. 1992), whereas later work proposed that halokinesis provided a preservation mechanism but that salt structures had minimal surface expression (McKie and Audretsch 2005). Correlation of Triassic strata across the Central North Sea has previously been hampered by poor seismic resolution, post-Triassic halokinesis and rifting, a lack of biostratigraphic data as well as facies changes where mudstone units that are clearly recognizable in the UK sector become increasingly sandy eastwards, into the Norwegian sector. The lack of substantial Triassic penetrations and limited core material together with poor palynological recovery, common in continental depositional settings, creates challenges in dating and both lithostratigraphic and chronostratigraphic-based correlation. In the Triassic of the Central North Sea lithostratigraphy and chronostratigraphy are effectively coincidental. The biostratigraphic resolution is sufficient to allow distinction between the different lithostratigraphic units but is not of high enough resolution to allow diachroneity between different facies belts (i.e. lithostratigraphic units) to be mapped. For these reasons, there is a lack of reliable predictive models for the lateral and vertical distribution of the intra-Skagerrak Formation members in Norway. In addition, well-log profiles do not always allow clear distinction between the different members, which have more clearly differentiated log expressions in the type area of UK Quadrant30 (Q30).

Using a 3D seismic survey supported by 19 wells including 5 with cores, we aim to reconstruct the structural and sedimentological development of the Skagerrak Formation in the Ula Field Area to help predict reservoir development and distribution. Considerable uncertainty exists regarding the timing of salt movement and the sedimentary response; therefore, with this work we aim to improve the understanding of the timing of salt movement. We integrate well-log interpretation with palynological analysis to assess the intra-Triassic stratigraphy and establish a model for minibasin evolution. We propose a lithostratigraphic correlation between the UK and part of the Norwegian sector using well log interpretation and biostratigraphic analysis that allows an increased understanding of depositional environment distribution and the influence of halokinesis during the Triassic across the Central North Sea. Based on the results of this study we conclude that the distribution of the Skagerrak Formation members extends eastwards from the UK into the Norwegian sector. We apply laboratory-derived models for the evolution of salt minibasins to the study area to place constraints on the timing of halokinesis. From the core and well-log data we propose a new conceptual model for the role of palaeotopography during Triassic sedimentation. We find no evidence to suggest salt tectonics impacted deposition patterns of the Skagerrak Formation in the area of the Ula Field. Depositional patterns were controlled by avulsion and migration of channels in the distributive fluvial systems (DFS). Finally, our work is summarized in a new well correlation panel that will help to identify lithological trends in the different members of the Skagerrak Formation.

The Central Graben represents the southern branch of the North Sea trilete rift system (Zanella and Coward 2003) and its origin is associated with the breakup of Pangaea initiated during the latest Permian to Early Triassic (Steel and Ryseth 1990; McKie 2014). The basin is a prolific hydrocarbon province that has been studied extensively with a common focus being the role of salt tectonics in controlling sedimentation, subsequent sediment preservation and the post-depositional modification of the sedimentary succession (Hodgson et al. 1992; Cameron 1993; Donovan et al. 1993; Smith et al. 1993; Goldsmith et al. 1995, 2003; Clark et al. 1998; Stewart and Clark 1999; Banham and Mountney 2013; Karlo et al. 2014; Jackson and Lewis 2016; Burgess et al. 2021, 2022; Archer et al. 2022; Gray et al. 2022). The salt is Late Permian in age and belongs to the Zechstein Supergroup and was deposited in a basin that extended across Europe from the UK to Poland, with a thickness exceeding 1 km in the basin centre (Clark and Tallbacka 1980; Glennie et al. 2003; Bachmann et al. 2010). The Zechstein Supergroup is composed of carbonates, evaporites (mainly halite and anhydrite), and some siliciclastic sedimentary rocks (Stewart and Clark 1999). The initial phase of halokinesis occurred during the Early Triassic and ended during the Middle to Late Triassic, although Stewart (2007) in some areas identified evidence for early halokinesis already during the Late Permian. Halokinesis was triggered by regional extension and differential loading of sediment on the salt (Zanella and Coward 2003).

The Zechstein Supergroup is overlain by fluvio-lacustrine and aeolian rocks of the Smith Bank Formation and fluvial rocks of the Skagerrak Formation (Deegan and Scull 1977; Wilkins et al. 2018). The Smith Bank Formation is predominantly Lower Triassic in age (Cameron 1993; Goldsmith et al. 2003) and includes the Bunter Sandstone and the Marnock Shale members in the UK sector of the North Sea. The formation is present in the lower part of minibasins with more limited development on top of salt structures. The Skagerrak Formation in the UK is Middle to Late Triassic in age and is divided into a series of sandstone and mudstone dominated members (Cameron 1993; Goldsmith et al. 1995; McKie 2014). In general, the Skagerrak Formation is considered to have comprised a dominantly braided fluvial system (Pedersen and Andersen 1980; Olsen 1988; Jarsve et al. 2014; Olivarius and Nielsen 2016). Other work suggests that there may be a transition from braided to meandering fluvial systems downstream, and interpret the Skagerrak Formation to represent part of a distributive fluvial system (DFS) (McKie 2011; Gray et al. 2022) developed under semi-arid climatic conditions (Goldsmith et al. 2003). It is considered that during increased phases of aridity in the Middle–Late Triassic, the distributive fluvial system retreated with expansion of playa-dominated areas in the downstream part of the basin (McKie 2014). However, it should be noted that the simple linkage between climate and lithology i.e. humid sandy fluvial systems and dry, muddy playa deposits has been questioned by Burgess et al. (2021) who reconstructed climatic changes which likely reflect the internal basin water balance rather than catchment wide responses.

A distributive fluvial system (DFS) is a fluvial system with an apex point located where a river exits a valley to enter a sedimentary basin, and from which a radial pattern of channels emerges (Hartley et al. 2010; Weissmann et al. 2010). Characteristically, DFSs show a predictable downstream decrease in grain size and channel dimensions and increase in the proportion of floodplain deposits (Hartley et al. 2010; Weissmann et al. 2010, 2011; Owen et al. 2017). The study area is dominated by channel belts with a general drainage direction towards the south-SW based on regional information synthesized in Gray et al. (2022). The boundary between the Smith Bank and Skagerrak formations is considered diachronous in some areas with the Skagerrak Formation being laterally equivalent to the Smith Bank Formation. In other areas the Smith Bank Formation is absent, and the Skagerrak Formation lies directly on the Zechstein Supergroup (Fig. 2) (Glennie 1998).

During the Lower to Middle Jurassic, thermal doming in the North Sea resulted in the formation of the regional Mid-Cimmerian unconformity (Underhill and Partington 1993). This widespread unconformity resulted in erosion of Jurassic and Triassic sedimentary material. Estimates from southern Norway suggest erosion of 1.3–3.5 km of Triassic–Jurassic strata (Roharman et al. 1995). During the Middle to Late Jurassic, rift extension caused a general collapse of salt highs with suprasalt minibasins forming above the salt highs (Mannie et al. 2014), filled with syn-rift sediment. Extension ended during the Early Cretaceous (Zanella and Coward 2003) with thermal subsidence prevailing and local inversion pulses during the Cretaceous, Paleogene and Neogene caused by Alpine orogenesis (Glennie 1986). During inversion, many hydrocarbon traps, including the Ula Field, formed (Mannie et al. 2014).

The understanding of the Triassic stratigraphy in the southern part of the Norwegian sector is limited. Some authors have tried to define the stratigraphy and the distribution of sedimentary deposits by studying the Triassic section through seismic facies and well-log analysis (Jarsve et al. 2014; Archer et al. 2022), but lack of appropriate age constraints hinders a full comprehension of the Triassic stratigraphy. The first Triassic lithostratigraphic scheme was established by Deegan and Scull (1977). They described the Smith Bank Formation as a monotonous sequence of red, silty claystone and the Skagerrak Formation as being sandstone-rich, but with some heterolithic sequences. Later, Cameron (1993) and Goldsmith et al. (1995) introduced members of the Skagerrak Formation, thereby improving the resolution of the stratigraphic framework in the UK sector. However, the age of these members is debated in literature. Indeed Goldsmith et al. (2003), McKie and Williams (2009), Archer et al. (2010, 2022) and Burgess et al. (2021, 2022) present slightly different age models that increase uncertainties when correlating these members. In this work, the most recent studies of Burgess et al. (2021, 2022) are used for age constraints (Fig. 2).

A 3D seismic survey (PGS16008CGR) shot in 2016 in the ED50-UTM31 coordinate system and 19 wells (Table 1) were used. The survey covers 2762 km2 across Quadrant 7 covering the 251 km2 Ula Field Area and surrounding area (Fig. 1). The bin spacing is 12.5 m and the vertical resolution decreases with depth but using a dominant frequency of 25 Hz for well 7/12-6, the vertical resolution ranges between 20 and 45 m (Table 2). The Fresnel zone calculated at the Top Triassic surface in well 7/12-6 is 538 m. The seismic data, however, is characterized by multiples in the Triassic section caused by a high-amplitude reflection interpreted as the Base Cretaceous Unconformity that hampers the interpretation. Four wells, 7/8-3, 7/12-3A, 7/12-A7 and 7/12-5, penetrate the entire Triassic section and reach the Zechstein Supergroup, the other partly penetrate the Triassic. Cores of Triassic strata totalling 301 m in thickness were described from wells 7/12-2, 7/12-6, 7/12-A7, 7/12-A15 and 7/12-A13B. The studied wells are all located within the Ula Field Area with the exception of wells 7/8-3 and 7/12-11 which are located in minibasins immediately adjacent to the Ula Field Area.

Interpretation of the 3D seismic survey was carried out in Schlumberger's Petrel software. Well tops were tied to the seismic data by synthetic seismograms for wells 7/12-3A and 7/12-6 (Fig. 2), using a zero-phase, 35 Hz, Gaussian wavelet. Five seismic facies (SF1-5; Table 3) were identified and, when penetrated, constrained utilizing mainly well 7/11-8 (Fig. 3). Two seismic facies were not penetrated by this well (SF1 and SF5), however, SF5 is penetrated by well 7/8-4. Seismic interpretation is hampered by the presence of multiples as highlighted in Figure 3 at well 7/11-8. By linking the log signature to the seismic facies, the Smith Bank Formation and the constituent members of the Skagerrak Formation could be identified. In this work, the diachronous nature of the interpreted intra-Skagerrak Formation members is not considered to be an issue due to a combination of the limited size of the study area and the high-resolution dataset that allowed a consistent interpretation through the investigated area based on seismic facies. Mapped horizons (Table 4), surfaces (Fig. 4) and isochron maps (Fig. 5) were generated. Additionally, four minibasins that surround the salt structure in the Ula Field Area and the Ula Salt High were analysed. Two minibasins (minibasins 1 and 2) are located on the west side of the Ula Field Area with minibasins 3 and 4 located to the east (Fig. 1). Of the four minibasins investigated, only minibasin 4 has been drilled, by well 7/12-11.

Well log interpretation was completed using the WellCAD software. The identification of the Smith Bank Formation and the Skagerrak Formation members were carried out using biostratigraphic data where available, and where biostratigraphic data were unavailable, utilizing the approach of Goldsmith et al. (1995) and Farris (1999) based on wells from UK Q22 and Q30. Interpretation of the depositional setting directly from the well logs is challenging due to the feldspar-rich composition of the sandstone of the Skagerrak Formation impacting log response. Therefore, definition of depositional environment was carried out by integrating the log profiles of gamma ray, neutron and density logs with the cores available for this work. To reduce uncertainty, wells were first interpreted individually, then grouped into clusters based on their locations and trajectories. Where possible, clusters were correlated using palynological data. In this work, the interpretation of the wireline logs and the identification of the members is based on Norwegian well 7/8-3 and the two UK wells 22/24b-5z and 30/07a-9. The Norwegian well 7/8-3 is considered to be representative of the Ula Field Area and provides some age constrains (Burgess et al. 2021, 2022) for the Triassic section through a publicly available biostratigraphic report, whereas the UK wells have good palynological recovery providing a reliable age model (e.g. Burgess et al. 2021, 2022).

Core sampling and sample processing for palynological analysis were carried out on the five cored wells (7/12-2, -6, -A7, -A13B, -A15) and undertaken on silt intervals using the approach suggested in Farris (1999) and developed by Burgess et al. (2021). New palynological results are presented here and integrated with a confidential biostratigraphic report completed by PetroStrat made available for this work for the Triassic cored interval in well 7/12-2, -6, -A7, -A13B and -A15.

Core descriptions provided an insight into the depositional environment and the distribution of the channel deposits in the area, wireline correlation provided constraints on stratigraphy and sediment distribution, and seismic facies analysis was used to determine the potential distribution of floodplain and fluvial channel-dominated deposits within minibasins.

Sedimentary facies

Seven sedimentary facies are recognized based on core descriptions (Table 5) and subsequently grouped into three facies associations.

Sedimentary Facies C1 (conglomerate) comprises up to 1.40 m thick erosive-based, matrix-supported conglomerates with granule to cobble size clasts that display a crude horizontal fabric at the scale of the core. The matrix is fine to medium-grained sand. Clasts are mostly grey and red mudstone or white-grey and green carbonate material. The mudclasts are identical to Si1 and the carbonate clasts to the pedogenic nodules in facies S5 (see below). The conglomerates represent tractional deposits deposited on an erosion surface and transported by unidirectional flow. Mudstone and carbonate intraclasts in the conglomerate indicate reworking of floodplain material derived from pedogenic dolocretes or calcretes. No clasts composed of evaporitic material (e.g. anhydrite or gypsum) were identified in the cores.

Sedimentary Facies S1 (massive siltstone and sandstone) comprises up to 2.25 m thick structureless siltstone to medium grained sandstone and occasionally contains mudclasts. This facies is interpreted to record dewatering following rapid deposition within channels (Beverage and Culbertson 1964; Harms and Fahnestock 1965; Harms et al. 1982), or unconfined flows (Miall 1977) if in association with facies S5 and Si1.

Sedimentary Facies S2 (laminated siltstone and sandstone) facies refers to up to 3.35 m thick very fine to medium grained sandstone and siltstone displaying centimetre scale planar stratification. When fine to medium grained and found in association with S1, the facies is interpreted to be deposited under upper flow regime conditions (Miall 1977). When it occurs as very fine to fine sandstone and associated with facies other than S1, it is considered to be deposited under lower flow regime conditions (Miall 1978).

Sedimentary Facies S3 (cross-stratified sandstone) describes sets of cross-strata up to 1.2 m thick with cosets up to 3.5 m in thickness composed of very fine to medium grained sandstone with both inclined planar stratification (>15°) and trough cross-stratification. Over 10 cm thick, fine to medium cross-stratified sandstone beds are the deposits of sinuous and straight-crested dunes that formed as bedforms within a channel (Miall 1977). Up to 10 cm thick intervals of very fine to fine grained cross-stratified sandstone represent dunes associated with single depositional events.

Sedimentary Facies S4 (rippled siltstone and sandstone) includes up to 0.75 m thick current-rippled siltstone and fine-grained sandstone in beds 5 to 40 cm thick. S4 is interpreted to record deposition under steady unidirectional subcritical flow conditions (Miall 1977).

Sedimentary Facies S5 (bioturbated siltstone and sandstone) comprises up to 4.75 m of grey to red brown siltstone and very fine to medium sandstone in beds of 20 to 30 cm thick and argillaceous siltstone affected by bioturbation and/or pedogenesis. Burrows are horizontal to vertical, sand-filled and have a cylindrical or elliptical shape. Meniscate burrows are found with simple horizontal and vertical burrows and are representative of the Scoyenia ichnofacies (Seilacher 1967; Buatois and Mangano 1998). Pedogenesis is indicated by root traces and carbonate nodules. This facies is superimposed on underlying facies and records a phase of non-deposition in channel or overbank areas (e.g. Rogers and Astin 1991; Buatois and Mangano 1998; Clemmensen et al. 1998; Zhang et al. 1998).

Sedimentary Facies Si1 (laminated mudstone and siltstone) corresponds to up to 0.55 m of millimetre scale laminated mudstone to siltstone. It overlies facies S4 or S5 either gradationally or with a sharp contact. It is interpreted to record sediment settling from suspension in a standing body of water.

Sedimentary facies associations

The sedimentary facies have been combined into three different facies associations: channel-fill facies association, splay facies association and floodplain facies association (Fig. 6l). This grouping generalizes the different depositional environments that likely existed under the dryland conditions that prevailed (e.g. McKie 2011), however the core and well log data do not allow the identification of the full range of possible environments and this threefold schematic grouping is used.

The channel-fill facies association commonly forms individual fining upwards packages up to 5 m thick which may be stacked to form packages up to 15 m thick. They comprise a basal conglomerate layer (C1) followed by cross-stratified sandstone (S3) and sometimes also by massive sandstone (S1). The upper part of the succession is characterized by planar stratification (S2), rippled sandstone (S4), bioturbated sandstone (S5) and siltstone (Si1). Each individual fining upwards package is considered to represent a single storey of a fluvial channel belt deposit. The basal part represents thalweg (C1) and bar deposits (S1 and S3) while the upper part represents bar top deposits (S4 and Si1) that was subjected to bioturbation and subaerial exposure (S5) (McKie 2011). It is rare to encounter the entire succession in one cored interval, in most examples the fining upward succession is truncated by an erosional surface indicating a multi-story channel body where the bar top/abandonment units are missing (Miall 1988; Bridge 1993).

The splay facies association is characterized by very fine to medium sandstone and forms packages up to 8.35 m thick. It includes facies S2, S4 (rippled, planar laminated siltstone and sandstone) and facies S1 and S5 (massive sandstone and bioturbated and pedogenically sandstone and siltstone). Sometimes this facies association also includes thin (0.05 m thick) conglomerate layers with mudstone intraclasts (C1) and the muddy facies Si1. This facies association is commonly characterized by fining upwards packages of bioturbated siltstone or sandstone (S5) with rippled siltstone and sandstone (S4), planar laminated siltstone and sandstone (S2) and massive sandstone (S1). The association of these facies suggests rapid sediment deposition, followed by bioturbation and subaerial exposure. This facies association is interpreted to record deposition predominantly by unconfined flows and represents splay deposits.

The floodplain facies association is up to 6.3 m thick and includes bioturbated and pedogenically modified sediment of facies S5, rippled sandstone (S4) and siltstone (Si1). These thick floodplain deposits represent a low energy floodplain or the most distal splay deposits of facies S5. Floodplain lake deposits are rare, and are associated with facies Si1, S2, S4 and S5 which together can reach 5 m in thickness. Rippled, rooted and bioturbated sediment is indicative of damp substrates (Miall 1977; Bourquin et al. 2009), likely associated with the most distal part of a splay deposit, a wet floodplain, or linked to the presence of shallow lakes/ponds formed in topographic lows. Also, conditions suitable for bioturbation can also develop along channels and channel margins as they often represent the wettest part of a dryland setting. Lakes are likely related to periods of high water table levels with increased water availability. The carbonate nodules are indicative of semi-arid conditions (Alonso-Zarza and Wright 2010). The presence of the Scoyenia ichnofacies (Seilacher 1967; Buatois and Mangano 1998) suggests a non-marine depositional setting and fluvio-lacustrine environment. Overall, pedogenic modification is limited and any soils are relatively immature indicating that subaerial exposure was periodic. The variations in pedogenesis demonstrate rapid fluctuations in water availability both laterally and vertically.

Channel deposit thickness and net channel to floodplain ratios in cored intervals and in well log profiles were calculated for the sandstone members of the Skagerrak Formation investigated. In the interval interpreted equivalent to the Judy Sandstone Member, cored channel deposits are up to 2 m thick in well 7/12-A7 whereas the net channel to floodplain ratio is 13% in core. In the interval interpreted as equivalent to the Joanne Sandstone Member, cored channel deposits are measured to be up to 12 m; cores from this interval comprise of 36% of channel deposits, 43% of splay deposits and 21% of floodplain deposits.

Well interpretation and correlation

Biostratigraphy

In general, biostratigraphic recovery was poor. However, where recovered, in situ palynomorphs were brown and appeared degraded. Those identified were in most cases long-range specimens not suitable for dating. Key findings from the confidential biostratigraphic report are the presence of specimens of Classopollis torosus in wells 7/12-2 and -6 that indicate an age not older than Carnian and therefore time equivalent to the Joanne Sandstone Member (e.g. Burgess et al. 2021). In addition, in well 7/12-A13B likely specimens of Classopollis torosus (or Classopollis spp) were identified suggesting a Carnian age. The presence of the member is confirmed by biostratigraphic analysis also in wells 7/11-6 and 7/11-8. In well 7/12-A7 a horizon containing bisaccate pollen undiff., Alisporites spp., Striatoabieites multistratus, Porcellispora longdonensis and Protohaploxyponus spp. is interpreted as Ladinian in age and time equivalent to the Julius Mudstone Member in the UK (Burgess et al. 2021). A similar interval was identified in well 7/12-A13B, allowing correlation between the two wells. A biostratigraphic report for well 7/8-3 permits identification of the Judy Sandstone Member, (e.g. Burgess et al. 2021).

Petrophysical characterization

Distinction between the Smith Bank and Skagerrak formations using wireline log data is aided by generally smoother and more monotonous petrophysical characteristics of the Smith Bank Formation log profile compared to the Skagerrak Formation. The Smith Bank Formation generally displays a variable and erratic gamma ray log response. The neutron-density log profiles display a marked and locally positive separation. The Smith Bank Formation is recognized in wells 7/8-3, 7/11-8 and 7/12-5 (Fig. 7).

In contrast to the Smith Bank Formation, the overlying Judy Sandstone Member of the Skagerrak Formation is characterized by limited neutron-density log separation and an erratic gamma ray motif (Farris 1999). This member has been identified in entirety in wells 7/8-3, 7/11-8 and 7/12-5. In 7/8-4 and 7/11-6 it is only partially penetrated and identified (Fig. 7). However, at a regional scale, the wireline log motifs of the Smith Bank Formation and the Judy Sandstone Member overlap making differentiation of these lithostratigraphic units difficult.

The Julius Mudstone Member which overlies the Judy Sandstone Member has a characteristic concave, bow-shaped gamma ray log response and the neutron-density logs are characterized by a positive separation (Farris 1999). For well log correlation the recognition of a mid-Julius Member horizon is essential. It is marked by a low gamma ray spike with a peak displayed in the density log. It is penetrated in wells 7/8-3, 7/8-4, 7/11-6, 7/11-8, 7/12-6, -11, -A2A, -A3B (Fig. 7). The Julius Mudstone Member is 210 m thick in well 7/8-3 but within the Ula Field Area the mudstone member does not exceed 10 m. The time-equivalent unit in well 7/12-A13B has well-log profile indicative of a sandstone interval (Fig. 7).

The Joanne Sandstone Member is interpreted to be present in wells 7/8-3, 7/8-4, 7/11-6, -8, 7/12-2, -6, -11, -A2A, -A3B, -A13B. In these wells the sandstone member displays an erratic gamma ray response with values ranging from low to high, and a limited separation of the neutron-density log.

The Jonathan Mudstone Member which overlies the Joanne Sandstone Member can be divided into lower, middle and upper subunits and it is interpreted to be present in wells 7/12-2, -3A, -4, -5, -8, -9, -A7D, -A15. In well 7/12-3A the Jonathan Mudstone Member directly overlies the Zechstein salt. The lower Jonathan Mudstone Member has a moderate gamma ray response and a large positive neutron-density log separation (Farris 1999; Fig. 8). The middle Jonathan Mudstone Member has a decreasing upward gamma ray profile, high density and low neutron log values that combined display a positive separation (Farris 1999; Fig. 8). The upper Jonathan Mudstone Member has a moderate gamma ray log response, a large neutron-density separation with slightly higher density log and lower neutron log response compared to the middle Jonathan Mudstone Member (Farris 1999; Fig. 8). The Jonathan Mudstone Member is >100 m in thickness although this represents a minimum as the true thickness is unknown due to post-depositional erosion. This member displays a similar signature across the area with a distinctive mudstone character, and a sandier section in the middle Jonathan Mudstone Member. The Jonathan Mudstone Member occurs mainly in the south whilst towards the north it is absent, due to post-Triassic erosion.

The Josephine Sandstone and Joshua Mudstone members which comprise the remainder of the Skagerrak Formation were not identified in the studied wells. The top of the Jonathan Mudstone Member, where present, is represented by an erosional surface overlain by Jurassic rocks. It is likely that the Josephine Sandstone and Joshua Mudstone members were eroded beneath the unconformity as occurs elsewhere in the Central North Sea, particularly in the UK sector (Burgess et al. 2021).

On the Ula Salt High, the Skagerrak Formation, from the Judy to Jonathan members, is 250 m thick in well 7/12-5, while in well 7/11-8, located in a minibasin, a section interpreted to contain the Judy, Julius and Joanne members is 500 m in thickness.

Seismic horizons

The Base Zechstein Supergroup horizon is a continuous to discontinuous, high-amplitude soft-kick reflection. This horizon is often segmented by pre-salt faults. Due to limited data, this horizon is not constrained by any wells.

The Top Zechstein Supergroup horizon is a continuous to discontinuous, high-to-low amplitude, hard kick seismic reflection. The high-amplitude reflection is mainly identified above salt structures whereas salt welds and flanks of salt structure display low amplitude reflection. The horizon is encountered in well 7/8-3, 7/12-3A and 7/12-5.

A prominent intra-Triassic horizon is observed as a continuous, high-amplitude hard-kick seismic reflection. The reflection is interpreted in all the minibasins but its presence above salt structures is unknown due to the poor seismic quality. This horizon is penetrated by well 7/11-8 in a minibasin outside the Ula Field Area.

The Smith Bank Formation corresponds to a continuous, medium-amplitude, hard-kick seismic reflection. The horizon is interpreted only in minibasins because above salt structures seismic quality is not sufficient to confidently identify the formation. Instead, in wells 7/8-3 and 7/12-5 on salt structures, the Smith Bank Formation is interpreted based on well logs. The Smith Bank Formation is present in minibasins as confirmed by well 7/11-8.

The Judy/Julius horizon is interpreted by merging the two Skagerrak Formation members due to the limited thickness of the Julius Mudstone Member above the Judy Sandstone Member. The horizon is a high-to-low-amplitude, semi-continuous, soft-kick seismic reflection characterized by a low-amplitude seismic reflection. It has a high amplitude in the minibasins in the SW of the area. The horizon is penetrated by wells 7/8-4 and 7/11-8 in minibasins and by several wells on salt structures.

For seismic interpretation, the Joanne/Jonathan horizon is merged due to the challenge of picking the relatively thin Jonathan Mudstone Member. The horizon is identified as a high-to-low-amplitude, semi-continuous, soft-kick seismic reflection. In most of the area with available seismic data, it displays a low-amplitude seismic reflection. The horizon displays high amplitude only in the minibasins in the southwestern most part of the area. This horizon is not identified in the wells in minibasins but a similar horizon is recognized in well 7/12-2 and 7/12-5 where the Jonathan and Joanne Members are interpreted to be present. These wells are located above salt structures such that the quality of the seismic data is reduced.

The Top Triassic horizon corresponds to a low-amplitude, hard-kick seismic reflection. The reflection is continuous but locally appears as discontinuous. The horizon is identified across the entire dataset, and it coincides with an unconformity. This horizon is identified in all wells available.

The Base Cretaceous Unconformity (BCU) is identified as high-amplitude, soft reflector in the seismic data. The reflection is continuous, and corresponds to a widespread erosional surface that can be picked across the entire dataset and identified within the available wells.

The limited thicknesses of the mudstone members and their amplitude variation across the dataset limit the possibility of their identification separately from the sandstone members. However, in some minibasins, especially in the SW, including minibasin 1 and 2, the Skagerrak Formation is characterized by an alteration of low and high amplitude packages. They suggest corresponding to the J-members when tied with well 7/11-8 and based on literature from the adjacent UK sector (e.g. Archer et al. 2022).

Seismic facies

Seismic interpretation of the Late Permian Zechstein Supergroup and Triassic intervals was aided through the use of seismic facies analysis. Five seismic facies (SF1 to SF5) are identified and documented in Table 3.

Zechstein Supergroup salt

Salt of the Zechstein Supergroup is identified by a hard-kick in the upper section of interpreted salt structures, whilst the flanks of salt structures were identified and picked utilizing seismic attributes and seismic facies. The Zechstein Supergroup salt is mainly characterized by SF1 and SF2, although in some areas it is possible to identify SF5 linked to internal heterogeneity. Seismic facies SF1 and SF2 are interpreted to represent evaporitic strata, whereas SF5 is interpreted to represent intra-salt non-evaporitic lithologies. The present-day thickness of the Zechstein Supergroup varies throughout the Ula Field Area as it can be absent or very thin in minibasins and very thick (3600 m) within salt structures (Fig. 4b).

Triassic

The Smith Bank Formation is dominated by SF1, SF2 and SF3 which have stratigraphic significance as the Smith Bank shale, Bunter Sandstone and Marnock Shale (Table 3). The formation may be thin or absent on salt highs but reaches a maximum thickness of 3290 m in minibasins. Seismic facies SF1 is characterized by a continuous, opaque, low amplitude and low frequency reflection package. This seismic facies is not penetrated by the available wells, therefore it is thought to be representative of floodplain deposits, similar to the interpretation of Jarsve et al. (2014), who studied equivalent strata in the Norwegian-Danish North Sea. Seismic facies SF2 comprises continuous, semi-opaque, low amplitude, low frequency reflectors and in well 7/11-8 it is associated with a medium gamma ray response and a low neutron-density separation. It is interpreted to represent fine-grained deposits typical of a distal terminal fluvial system (e.g. Kelly and Olsen 1993; Cain and Mountney 2009; Jarsve et al. 2014). The upper portion of the Smith Bank Formation displays a characteristic seismic response (SF3) identified in most of the survey area. Seismic facies SF3 comprises a continuous, medium to high amplitude and medium frequency reflection package. This interval is found between a very bright hard-kick visible in most of the minibasins at the base and may represent a local unconformity that marks the base of the Skagerrak Formation (Fig. 9). In well 7/11-8 this seismic facies is associated with high gamma ray and a neutron-density log separation, typical of fine-grained deposits such as floodplain or lake margin similar to the ephemeral-to-perennial lake facies association of Goldsmith et al. (2003) and recognized by Jarsve et al. (2014). Due to its stratigraphic position, this package is considered to be the equivalent to the Marnock Shale in the UK sector. The overlying Skagerrak Formation has a variable thickness. In minibasins it is up to 1233 m in thickness (Fig. 5d). It is characterized by alternations of SF4 and SF5. Seismic facies SF4 comprises semi-opaque discontinuous to chaotic, low amplitude, low frequency reflection; this seismic facies, supported by well logs in 7/8-4 and 7/11-8, is interpreted to represent fluvial channel facies, consistent with that described by Jarsve et al. (2014). Seismic facies SF5 is a continuous, parallel, medium to high amplitude and medium-low frequency reflection package. This facies, based on well 7/8-4, is interpreted to represent a more distal/sheet sand facies (Jarsve et al. 2014) where sheet-sand deposits are interbedded with fine-grained floodplain deposits (Goldsmith et al. 2003). SF4 corresponds to the sandstone dominated units whilst SF5 is correlated with mudstone dominated units (Fig. 8).

Minibasin description

The four studied minibasins are 3080 m deep, 7.5 to 8 km long and 2 to 7.5 km wide (Figs 1, 9 and 10). Minibasin 1 and minibasin 2 are asymmetric minibasins whose sediment packages can be divided into three intervals with different geometries, from bottom to top: bowl, wedge and layer shape (Fig. 9). In minibasin 1 the bowl shape is up to 674 m thick and it is defined by SF1, whereas the wedge shape reaches 2312 m and displays SF2 at the bottom overlain by SF3 with SF4 and SF5 at the top of the wedge shape. The overlying layer geometry is 195 m thick and contains only SF4. In minibasin 2 the bowl shape reaches 2312 m in thickness with SF1 at the base followed by SF2 with SF3 at the top. In minibasin 2 the wedge shape is about 1356 m thick and characterized by SF4 and SF5. Finally, the layer geometry on the top is about 205 m thick displaying only SF4. The seismic facies in minibasin 1 and minibasin 2 show similar characteristics and the two minibasins are considered to have a similar history. Minibasin 3, NE of Ula Field, and minibasin 4, E of Ula Field originally formed as symmetric minibasins, though at present minibasin 4 has an asymmetric shape due to post-Triassic faulting (Fig. 10). Minibasin 3 is characterized by the presence of SF2 at the bottom, overlain by SF3 and then SF4 at the top, while minibasin 4 displays SF1 in the lower part overlain by SF2 and SF4. Minibasin 3 and minibasin 4 display a similar seismic reflection character. However, minibasin 3 and minibasin 4 differ from minibasin 1 and minibasin 2 as SF5 is not present such that the alternation of SF4 and SF5 observed in minibasin 1 and minibasin 2 is not observed. The Smith Bank Formation reflection characteristics do not display any significand differences across the investigated minibasins.

Minibasin development

Seismic mapping of minibasins 1 and 2 highlighted the similarity between these two minibasins in terms of both their stratigraphy and internal geometries. The bowl-shaped geometry at the base of the minibasins beneath the Smith Bank Formation indicates symmetric minibasin subsidence with salt movement during the Early Triassic. The geometries of SF1, SF2 and SF5 which form intra-salt reflectors in the salt structures in the Ula Field Area could suggest an early stage of halokinesis occurred in the area as a relict minibasin can be identified within the salt (Fig. 12). This early phase of halokinesis may indicate that salt movement occurred during the Late Permian and follows similar observations from the Late Permian in the UK sector of the Central North Sea (Clark et al. 1998; Stewart and Clark 1999; Stewart 2007; Esestime et al. 2015; Joffe et al. 2023). The wedge geometry that overlies the bowl-shaped geometry indicates asymmetric subsidence. The change in geometry could be triggered by differential sediment loading which causes a migration of the depocenter from the centre to the margin. The horizontal stratal geometry on top is thought to form as a consequence of minibasin grounding (Fernandez et al. 2019). The main phase of halokinesis is considered to have ended at this stage, i.e. when the Jonathan Member was deposited. The formation and development of this type of minibasin was modelled by Fernandez et al. (2019) who also recognized a post-weld wedge, but this is not preserved in the Ula Field Area due to post-Triassic erosion.

The geometries observed in minibasin 3 and minibasin 4 are similar to those observed in salt minibasins offshore Angola where Jackson and Hudec (2017) and Ge et al. (2021) described minibasin formation as a consequence of differential loading. The effect of differential loading in the minibasins is the switch from a central depocenter to two lateral depocenters, where salt evacuates more rapidly at the margins than it does in the centre of the minibasins. The downlap-onlap transition reflection termination was interpreted to have been generated by the lateral expansion of the minibasins (Jackson and Hudec 2017; Ge et al. 2021). Analogous stratal geometries in minibasin 3 and minibasin 4 are considered to be the result of lateral differential loading of the underlying salt during the early stages of the deposition of the Smith Bank Formation (Fig. 10). As these geometries are present in the Smith Bank Formation, lateral expansion of the minibasins was coeval with the change in geometry of the minibasins west of the Ula Field. Minibasins 3 and 4 also display post-Triassic deformation which differs from the post-Triassic deformation observed west of the Ula Field. Post-Triassic rifting not only created a potential connection between subsalt and suprasalt intervals, but it also had a role in the preservation of the Triassic strata. The difference between the eastern and western sections of the Ula Field Area is that during Jurassic extension, uplift of the subsalt fault footwall aided preservation of the strata in the hanging wall but facilitated erosion on the footwall.

Palaeogeography

Integration of our datasets with regional studies has allowed palaeogeographic reconstructions for Anisian, early Ladinian, Carnian and Norian times (Fig. 11).

Anisian palaeogeography

By combining previously published work with data from this study, it is clear that Anisian aged strata in the investigated area were characterized by a fluvial-dominated environment with localized floodplain deposition indicated by seismic facies SF5 (Fig. 11a). The general drainage direction is assumed to be towards the SSW based on the regional work of McKie (2014) and Gray et al. (2022), with the distribution of channel belt deposits in the area based on core and well log data presented here. These data indicate that channel belts were deposited and preserved on top of salt structures across the study area. Observed channel deposit thicknesses and abundances suggest that at that time the area corresponded to the medial to distal region of a distributive fluvial system (e.g. Owen et al. 2017). This is likely to comprise isolated to partly amalgamated channel belt deposits with a series of terminal splays, damp floodplain deposits and soil development. The Anisian time period in the study area is represented by the Judy Sandstone Member (Burgess et al. 2021). The presence of channel sandstone deposits indicates that the salt structures lacked surface topographic expression. Additionally, the absence of anhydrite or gypsum clasts in the channel deposits suggests that salt structures were not exposed at the surface and reworked by fluvial channels (e.g. Banham and Mountney 2013).

Early Ladinian palaeogeography

Ladinian age rocks identified in 7/12-A7 and 7/12-A13B are interpreted as the Julius Mudstone Member-equivalent (Burgess et al. 2021). The presence and distribution of the mudstone member in the seismic data is constrained by the seismic facies SF5, similarly the presence and distribution of the sandstone members are constrained by the seismic facies SF4. As SF4 is associated with a sand-dominated interval and SF5 with fine-grained deposits, the general thinning trend may reflect the transition from mudstone-dominated floodbasin in the SW to more sand-dominated fluvial deposits in the NE. The distribution of the mudstone and sandstone during this time period reflects the retrogradation of the distributive fluvial system. The maximum lateral extent of the mudstone records the maximum retrogradation of the fluvial system. The vertical transition from the Julius Mudstone Member to the overlying Joanne Sandstone Member represents the renewed progradation of the distributive fluvial system (e.g. McKie 2014; Gray et al. 2022).

The thinning of the Julius Mudstone Member identified in the Ula Field Area differs from the interpretation of Archer et al. (2022) who in their regional study suggested that the member thickened into the Norwegian sector. The thinning identified here represents the up-dip transition into more distal to medial distributive fluvial system deposits that record the back-stepping of the distributive fluvial system facies belt. This up-dip transition from mudstone to sandstone has implications for fluid migration in the subsurface as the fluvial deposits represent potential reservoirs/stores whilst the mudstones may be seals or baffles if present.

Carnian palaeogeography

The available data for Carnian aged strata (broadly equivalent to the Joanne Sandstone Member, Burgess et al. 2021) indicate that the area investigated was again characterized by a fluvial-dominated environment with localized floodplain deposition indicated by seismic facies SF5 indicating a progradation of the distributive fluvial system in the area (Fig. 11c). The area is interpreted to correspond to a medial-to-distal region of a distributive fluvial system based on the dimension and abundance of channel deposits in core and well log profiles (e.g. Owen et al. 2017). The interpretation of this member as a more proximal area compared to the Judy Sandstone Member supports the regional work presented in McKie (2014) and Gray et al. (2022). Palaeocurrent data for this unit are inferred from regional studies (McKie 2014; Gray et al. 2022), whereas the channel abundance and distribution is based on the cores and wells available for this work. As with the Anisian reconstruction, the presence of channel deposits on salt highs and the lack of evaporite clasts within fluvial channel deposits precludes any surface topographic expression being present over salt structures during deposition.

Norian palaeogeography

Retrogradation of the distributive fluvial system occurred in the Norian, resulting in mudstone deposition across the area represented by the Jonathan Mudstone Member. In the Ula Field Area, the member displays a higher sand content than in the UK sector, reflecting a more proximal location relative to the UK sector. The SW drainage direction is assumed based on regional reconstructions (McKie 2014).

Palaeotopography

A key finding of this work is that there appears to have been no topographic relief associated with salt diapirism during deposition of any of the Skagerrak Formation members. The presence of channel deposits on the salt high indicates deposition and aggradation on top of what must have been a buried feature. This observation is in contrast to the models of Stricker et al. (2017) from the Seagull Field Area in the UK Central North Sea, and Jarsve et al. (2014) for the Danish sector of the Central North Sea in which they suggested that salt diapirs were exposed and created surface topography with no sediment deposition above their crests. Our interpretation is similar to that of McKie and Audretsch (2005), McKie et al. (2010) and McKie (2014) who emphasized the lack of evidence for fluvial containment by salt tectonics during Judy and Joanne member deposition across the same area as Stricker et al. (2017) studied. They presented a model featuring minimal salt expression at surface, with the sediment bodies geometries not being modified by subsequent erosion. A similar example with stacked fluvial deposits on an actively growing salt high has been described from the Carnian age Chinle Formation in the Paradox Basin in Utah (Hartley and Evenstar 2017). In these examples sediment aggradation was greater than vertical salt movement facilitating the accumulation of channel deposits above an active salt structure. A consequence of the flat surface topography is that sedimentation was continuous across both salt highs and minibasins with the salt movement resulting in variations in thickness of Skagerrak Formation members. This difference in thickness from minibasins to salt highs will have been accompanied by facies changes and the development of disconformities/unconformities within the Triassic succession across the salt highs, although the resolution of our datasets is not sufficient to ascertain this.

Analysis of seismic reflection data, wireline log data and cored intervals from Triassic strata in the Ula Field Area of the Norwegian North Sea record the retrogradation and progradation of distributive fluvial system across the area. During the Anisian (Judy Sandstone Member), the Ula Field Area corresponded to a distal part of a distributive fluvial system. During the early Ladinian (Julius Sandstone Member), northeastwards directed back-stepping of the distributive fluvial system resulted floodplain deposition across the area but with a transition to distal DFS deposits on the northeastern margin of the study area. During Carnian times (Joanne Sandstone member), the fluvial system prograded across the region with deposition of medial distributive fluvial system deposits in the Ula Field Area. During the Norian (Jonathan Mudstone Member) mudstone deposition dominated as the fluvial system back-stepped northeastwards again. The Jonathan Mudstone member in this area is significantly sandier than UK equivalents, suggesting increasing proximity to a source to the NE. This has implications as the mudstone members can act as seals and/or, baffles to subsurface fluid flow.

Stratal geometries and interpreted stratigraphy suggest that salt movement started primarily during the Early Triassic and terminated during the Late Triassic, during deposition of the Jonathan Mudstone Member. Although it should be noted that there is potential evidence for syn-depositional halokinesis during the late Permian. This model widens and opens new area of subsurface exploration as the portions of Triassic above salt structures are likely to include sandstone bodies. This is essential for existing oil and gas fields that are located on salt highs often with Jurassic or Cretaceous reservoir intervals and where the Triassic succession has not been investigated. The observation that syn-sedimentary halokinesis had little impact on channel belt deposition is important for unlocking potential in existing and underexplored areas of the North Sea and may also be important in other sedimentary basins affected by syn-sedimentary halokinesis.

We would like to acknowledge the support of Aker BP and the Ula Partnership for supporting this work and providing data, PGS for providing seismic data and PetroStrat for providing palynological analysis. Authors would also like to thank C.A.L. Jackson for his comments at the early stage of work. We would like to acknowledge the reviewers Tom McKie, Carita Augustsson and Erlend Morisbak Jarsve for taking the necessary time and effort to review the manuscript. We sincerely appreciate all the valuable comments and suggestions, which helped us in improving the quality of the manuscript.

LDL: data curation (lead), formal analysis (lead), methodology (lead), writing – original draft (lead); AH: data curation (supporting), formal analysis (supporting), methodology (supporting), supervision (supporting), writing – review & editing (lead); JD: data curation (supporting), methodology (supporting), supervision (supporting); ERK: data curation (supporting), methodology (supporting), supervision (supporting), writing – review & editing (supporting); JH: data curation (supporting), formal analysis (supporting), methodology (supporting), visualization (supporting), writing – review & editing (supporting); DJ: data curation (supporting), formal analysis (supporting), methodology (supporting), supervision (supporting)

This research is funded by Aker BP and its Ula partner DNO.

This research is funded by Aker BP and its Ula partner DNO.

The datasets generated during and/or analysed during the current study are not publicly available due to confidentiality but are available from the corresponding author on reasonable request.

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/)