Characterising the internal structural complexity of the Southern North Sea Zechstein Supergroup Evaporites

The internal structure and architecture of evaporite sequences is often overlooked, with attention frequently concentrating on the external geometries that salt bodies form. The availability of extensive 3D seismic data affords the opportunity to interpret the internal structures within these evaporite sequences and comprehensively characterise the different structural facies over large areas. This paper concentrates on the Zechstein Supergroup evaporite deposits within the Southern North Sea of the United Kingdom's Continental Shelf. This analysis of the internal structural complexity and stratigraphic heterogeneity utilises 26,000 km2 of 3D seismic data together with over 96 wells from the Southern North Sea. Characterisation of the different structural facies present was undertaken alongside mapping their spatial distribution to understand the relationship they have with one another and the structural evolution that may have been taken. This work has (1) characterised and mapped six different internal structural facies present within the Zechstein with increasing levels of deformation; (2) shown the internal lithological heterogeneity is indicative of variations in the vertical strength profile of layered evaporite sequences; (3) discontinuous high‐amplitude reflections within the Zechstein are as a result of the geometries being too steeply dipping for the seismic data to image; and (4) the ability to possibly predict the internal heterogeneity of areas of poorly imaged salt, such as within large diapiric salt structures, from surrounding structural facies. These findings suggest that there is significant internal complexity even within areas of the basin with minor mobilisation to the external salt geometry.


| INTRODUCTION
Laterally extensive salt deposits are never solely comprised of halite but rather form interbedded sequences of halite and other evaporite and none evaporite lithologies. These variable evaporitic formations are referred to as layered evaporite sequences (Rowan et al., 2019). Evaporite sequences are common features of many sedimentary basins around the world (Butler et al., 2015); notable examples include the Zechstein Supergroup of Northern Europe, the Messinian of the Mediterranean and the Iren of the Precaspian Basin (Jackson & Hudec, 2017).
Some evaporite lithologies, most notably halite, are mobile in the subsurface, typically reacting to external changes in stress and leading to flow (Jackson & Stewart, 2017) or, alternately, responding to internal stresses arising from variations in density differences. Three types of loading can lead to force-driven salt flow: gravitational, displacement and, to a much lesser extent, thermal (Hudec & Jackson, 2007). This ability of evaporite formations to flow causes the formation of geological structures such as salt diapirs, salt-cored anticlines and salt welds. Internal structures which do not modify the external geometries of the evaporite formation can also form, the most common structures formed being folds (Rowan et al., 2019).
The mechanical behaviour of evaporite sequences is partly controlled by the proportion and layering order of the constituent lithologies present (Adamuszek et al., 2021). Interbeds, such as carbonates and anhydrites, are much stronger in both compression and extension than halite (Table 1), which undergoes flow at stresses of <5 Mpa (Zulauf et al., 2011). The mode of deformation for these interbeds also differs from halite as they can undergo both a combination of brittle (breaking apart) and ductile deformation (competent deforming) to accommodate strain (Strozyk et al., 2014). In contrast, halite rarely undergoes brittle deformation because it is ductile under low strain rates (Jackson & Hudec, 2017). These mechanically strong lithologies influence the strain patterns within an evaporite sequence (Strozyk et al., 2014). Differing quantities and distributions of mechanically strong interbeds influence the structural style and how internal strain is partitioned in mobile evaporites (Rowan et al., 2019;Stewart et al., 1996;Stewart & Harvey, 1998). Interbeds modify not only the behaviour of evaporite sequences, but also aid in defining the strain and kinematic record of the evaporite sequence in both outcrops and on seismic data by acting as marker beds, which aid in helping to quantify levels of strain which have occurred (Zulauf & Zulauf, 2005).
Stringers are blocks of non-halite material with different flow properties commonly consisting of the original interbedded lithologies (Jackson & Hudec, 2017). Stringers have been observed in many different evaporite sequences (Al-Siyabi, 2005;Edgell, 1996;Giles & Rowan, 2012;Jenyon, 1989) and previously recognised and described in the Zechstein of North West Europe (Jenyon, 1989;Van et al., 2011). Boudinage is a common mode of failure of competent layers, such as anhydrites and carbonates, that are enclosed within a weak, ductile matrix, such as halite, and which undergo layer-parrel extension (Rowan et al., 2019), as generally competent rocks are more likely to undergo brittle failure in extension than compression.
While previous work has investigated intra-salt deformation (Burliga, 1996;Butler et al., 2015;Cartwright et al., 2012;Davison et al., 1996;Jackson et al., 2014;Jackson et al., 2015;Strozyk et al., 2012), there has been a limited emphasis on distinguishing the different styles and distributions of deformation, especially on the scale of a basin. In some cases, the ability to distinguish the internal deformation of an evaporite sequence is limited by the quality of the available seismic data, with halite-dominated evaporitic sequences often described as having a chaotic or transparent appearance (e.g. Hanafi et al., 2022;Jones & Davison, 2014). However, improvements in the seismic processing of evaporite sequences are leading to coherent reflectivity within these formations, potentially leading to an evolution in the interpretation of evaporite sequences' internal structural heterogeneity. Much of the current understanding of evaporite sequences has come from their association and the subsequent data acquisition with prolific hydrocarbon provinces, such as the Precaspian Basin (Rowan et al., 2019), the South Atlantic conjugate margins (Blaich et al., 2008;Wen et al., 2019), the Zagros basin (Amthor et al., 2005;Bordenave & Hegre, 2010) and the North Sea (Peryt et al., 2010), where sequences are important for both trapping geometries and seals (Archer et al., 2012;Sarg, 2001). Similarly, evaporite formations will likely be important components in many subsurface developments required for the energy transition, in particular for geological storage sites (Duffy et al., 2023).
This study uses observations and interpretations from 3D seismic and well data from the Southern North Sea to examine the style and distribution of the internal deformation within the layered evaporite sequences of the Zechstein. The findings here have implications for understanding the deformation within layered evaporites.

| GEOLOGICAL SETTING
The Zechstein Supergroup of Northern Europe is an expansive group of layered evaporites that were deposited during the Lopingian of the Late Permian. The Zechstein has been studied extensively throughout the North Sea due to its importance in petroleum systems (Glennie, 1998). On the United Kingdom Continental Shelf (UKCS), the Zechstein is most commonly either the sealing sequence for the underlying Rotliegend natural gas reservoirs of the Southern North Sea (Bailey et al., 1993) or has led to the formation of structural trapping geometries for many intervals within the North Sea, such as for the Palaeocene turbidites, Cretaceous chalk, Triassic and Jurassic reservoirs (Baniak et al., 2020;Evans et al., 2004;Fraser et al., 2002;Jackson & Stewart, 2017).

| Tectonic evolution
Two major east-west trending rift basins are present on the eastern side of the UKCS: the South Permian Basin and the North Permian Basin (Figure 1). The two basins are separated by the Mid-North Sea High and the Ringkøbing-Fyn High (Clark et al., 1998) (Figure 1). These two larger basins host numerous smaller sub-basins, such as the Forth Approaches Basin (Cartwright et al., 2001) and the Silverpit Basin (Bailey et al., 1993), located in the North and South Permian basins, respectively.
The South Permian Basin of the UKCS has developed over a period of 380 million years, consisting of a complex geological history with several rift and post-rift phases ( Figure 2). Extension initiated with the collapse of the Late Devonian Variscan orogenic belt (Pharaoh et al., 2010;Schulmann et al., 2014) and subsequent Late Carboniferous inversion (Hodgson et al., 1992;Ziegler, 1990). Early-Permian rifting saw the two Permian basins of the North Sea open up and begin to develop (Glennie et al., 2003;Hodgson et al., 1992). This rifting phase continued until the Middle Triassic, only briefly interrupted in the Late Permian by a phase of post-rift thermal subsidence (Geluk, 2007;Hodgson et al., 1992;Pharaoh et al., 2010). Tensional stresses that developed during the period of Triassic rifting reactivated pre-existing faults and formed new fault systems, including those defining the East Irish Sea (Glennie & Underhill, 1998;Zanella & Coward, 2003). Another rifting phase occurred from the Jurassic up into the Early Cretaceous (Erratt et al., 1999), this phase of rifting thinned the lithosphere, allowing for a thermal anomaly which caused thermal doming and uplift (Pharaoh et al., 2010;Zanella & Coward, 2003;Ziegler, 1992). However, this uplift event was short-lived, lasting until the Late Jurassic as rifting continued (Thomas & Coward, 1996). When rifting ceased in the Middle Cretaceous, the basin experienced a shortterm period of thermal subsidence, which lasted until the Late Cretaceous (Pharaoh et al., 2010). This was followed by a period of tectonic inversion, which was initiated in the Campanian and continued to the Early Cenozoic (Erratt et al., 1999). As inversion ended, it was replaced by thermal subsidence once again; as the Cenozoic sediments began to be deposited into the North Sea thermal Sag basin, leading to up to 3 km of Cenozoic sedimentation in areas (Wong et al., 2007;Ziegler, 1990).

| Zechstein Supergroup gross depositional environment
Prior to the deposition of the Zechstein Supergroup Evaporites, the Permian Rotliegend aeolian sediments were deposited during the early phases of thermal subsidence (Maynard & Gibson, 2001). During this phase of aeolian sedimentation, the basin was entirely landlocked, isolated from all surrounding ocean bodies (Peryt et al., 2010). Subsidence of the basin was greater than that of the sedimentation rate, leading to an underfilled basin (Glennie et al., 2003). The basin remained underfilled into the Late Permian due to the continued Permo-Triassic subsidence; this led to the centre of the basin being as much as 300 m below average sea level by the Late Permian, which was the onset of the Zechstein Supergroup deposition (Glennie, 1998).
The Zechstein ocean formed due to an influx of marine water into the underfilled Permian basins of Europe (Smith, 1979) after a significant transgression occurred from the Barents Sea to the north (Strozyk et al., 2017). Both basins had restricted exchange of waters from the Northern Boreal Ocean and Southern Tethys Ocean (Pancost et al., 2002), leading to little influx of marine water. Despite the initial influx of large amounts of oceanic water, the sedimentation rate during this period remained low, which, combined with the high temperature, arid environment and limited water supplies, led to the Zechstein ocean becoming a giant evaporite production area in both of Europe's Permian basins (Glennie, 1998). This period of evaporite deposition occurred in the Late Permian from 258 to 251 Ma, according to chemical analysis of fluid inclusions found within Zechstein halite (Lippolt et al., 1993;Menning, 1995). The original depositional extent of the evaporite basin is represented by the modern-day distribution of the Zechstein (Jackson & Stewart, 2017).
The Zechstein Supergroup's lithostratigraphy is separated into Zechstein cycles, commonly referred to as F I G U R E 1 (a) A WSW-ENE trending regional cross-section of the South Permian Basin from just offshore the UK to the edge of the UK continental shelf. The main structural elements and stratigraphy (Quaternary-Pre-Permian) within the South Permian Basin are visible. Redrawn from Pharaoh et al. (2010). (b) Palaeo-geography map of the Zechstein's deposition, redrawn from Słowakiewicz et al. (2018). Important geological areas are named, and cross-section A-A′ is marked. Z cycles, with either 5 or 7 in total, depending on the nomenclature used (Bailey et al., 1993;Geluk, 2007) ( Figure 2). The cycles characterise cyclic evaporation, and hence depositional sequences relating to periods of transgression and regression within the restricted basin (Pharaoh et al., 2010). Each depositional Zechstein cycle begins at maximum sea level with lithologies associated with depositional environments in settings of low salinity, such as carbonates. As the progressive evaporation of the marine waters occurs, the salinity of the brine in the basin increases, and so do the subsequently deposited lithologies (Peryt et al., 2010). At the end of each Zechstein depositional cycle, total evaporation of the brine will likely have occurred, leading to the ground surface being completely dry until the next subsequent influx of water (Glennie, 1998). F I G U R E 2 Chronostratigraphic chart of the Southern Permian Basin stratigraphy running N-S through the basin centre modified from Patruno et al. (2022). All major stratigraphic successions are shown, as well as unconformities. Regional geological events for the North Sea Permian Basins are also present. The Lithostratigraphy of the Zechstein Supergroup, alongside the Zechstein cycles, are included. The nomenclature for the Zechstein is from Johnson et al. (1993). The Zechstein cycles are subdivided into import cycle components (±) used in this study. AU, atlantean unconformity; BCU, base cretaceous unconformity; BPU, base permian unconformity; MMU, mid miocene unconformity; MNSH, mid north sea high; Rtlgd, rotliegend; SP Fm, silverpit formation.

| DATA
The study area ( Figure 3) is located within the South Permian Basin of the UKCS, a mature gas basin that has been explored and exploited for the last 50 years (Rouillard et al., 2020). The study area has the same extent as the seismic data used, covering a total area of 26,386 km 2 . The western limit is ca. 20 km offshore the east coast of England and extends east to the easternmost edge of the UK sector of the Southern North Sea. The most northern extent of the study area is defined by the Mid-North Sea high and stretches south from this margin for 280 km. The seismic and well data are available from the North Sea Transition Authority National Data Repository (https://ndr.nstau thori ty.co.uk/) under an Open Government Licence.

| Seismic data
The seismic data used in this study were the Southern North Sea Mega Survey Revision 2 (SNSMSR2) (Figure 3), which was merged and processed by Petroleum Geo-Services (PGS) in 2015. This 3D merged seismic dataset is comprised of 86 individual surveys from the UK sector of the Southern North Sea. The original datasets used in the merge were generally zero-phased 3D time-migrated seismic surveys, each constituent survey was resampled to 4 ms, and inline/crossline grids interpolated to 12.5 m if they were not already. Before the final merging process, cleaning was accomplished using a time-variant filter to avoid miss-stacking to reduce noise in deeper sections. Phase matching for the surveys within the merge was not undertaken; however, polarity and amplitude matching were in order to decrease edge effects. The survey was generally zero-phased and displayed using the European polarity standard so that a negative amplitude represents an increase in acoustic impedance, and a positive amplitude represents a decrease in impedance. The vertical resolution of the SNSMSR2 varies throughout the dataset due to the constituent surveys having different original acquisition parameters; however, it varies between ca. 12 m in the Cenozoic and ca. 52 m in the Zechstein ( Table 2). The F I G U R E 3 Location and data map for the study area. Summary overview of all data used within this study and its geographical location. The top left map shows the study location with respect to the UK and Europe. The Southern North Sea Mega Survey Revision 2 seismic survey outline (blue) is present alongside the locations of all well data used within this paper. Well 44/27-2's location ( Figure 4) is marked on. NSTA, North Sea Transition Authority; SPB, South Permian Basin. full processing report can be found at (https://ndr.ogaut hority.co.uk/).

| Well data
In total, 96 wells were used during this study ( Figure 3). The wells used within this study were selected based on whether the well penetrated the Zechstein Supergroup: the availability of petrophysical logs, specifically density, sonic and gamma ray and the availability of checkshot data. Wells without these complete petrophysical logs were occasionally used in areas with poor data availability. A full list of the wells used in this study can be found in Table S1.

| Well interpretation
Lithologies and stratigraphic boundaries were interpreted for the Zechstein Supergroup ( Figure 4) using a combination of petrophysical properties, lithology composite logs and cuttings descriptions. Lithologies for the Zechstein Supergroup were defined for the minimum thickness interval possible using the petrophysical data, which, while dependent on the logging tool used, was between 0.3 and 2.5 m (Bourke et al., 1989). Lithologies for younger stratigraphy were also interpreted to enable a consistent seismic stratigraphic framework to be constructed across the study.
Synthetic-seismic well ties were generated for 60 wells to correlate lithological and seismic-stratigraphic boundaries interpreted from wells to the seismic data. A full list of wells for which synthetic-seismic ties were generated is available in Table S1. These well ties were generated using sonic and density logs, together with checkshot data. Synthetic traces were generated using an analytical 30 Hz ricker wavelet and extracted wavelets, which were compared with the original seismic data to determine which wavelet was the best fit. In some cases, syntheticseismic traces required time-shifting or stretching to account for mismatch in datums, largely a result of using a merged seismic volume. The synthetic is compared to multiple traces from the surveys as many of the wells throughout the study area are deviated (e.g. Figure 4). The most prominent responses within the generated synthetics are from halite-carbonate and halite-anhydrite interfaces (Figure 4; e.g. 2307-2356 ms), for example the +Z3 to −Z3 and +Z2 to −Z2. The large reflection coefficient resulting from the boundary between these two lithologies/cycles gives a strong seismic response and as such the strong resultant reflections can be utilised as marker beds in the SNSMR2. In Figure 4, only a few beds have both top and base fully resolvable within the generated synthetic, notably the 77 m thick k-salt bed within the +Z3 (3037-3114 m) and the 100 m thick −Z3 anhydrite bed (3400-3500 m).

| Seismic interpretation
Major stratigraphic and lithological boundaries identified in the well data were subsequently interpreted in the 3D seismic, based on the seismic-well ties ( Figure 4).

T A B L E 2
Seismic resolution examples for slice Inline 31,000-35,000, crossline 32,000-35,000 from the SNSMSR2. Reflections were initially mapped using an inline and crossline spacing of 1250 m and then auto-tracked and quality controlled, areas with low auto-tracking confidence were mapped at a reduced inline/crossline spacing, typically between 125 and 250 m to improve the confidence in 3D auto-tracking. Surfaces were generated from the auto-tracked horizons using a convergent gridding algorithm with a grid spacing of 50 × 50 m to create seamless surfaces (Figure 5a). A common problem with the SNSMSR2 survey is the lack of continuity of the water bed reflection. This issue, also observed by Grant et al. (2019), was likely due to survey vintage and age, these issues were particularly problematic in areas of shallow water depth.

Vertical seismic resolution (m) λ = ((Hz / m/s)/4) Notes
To counter this problem, instead of using the water bottom reflection, bathymetry data from EMODnet were used and converted to two-way travel time, assuming a water velocity of 1494 m/s, a similar approach to previous studies (e.g. Grant et al., 2019). The Zechstein cycles can be interpreted on seismic data, where the cycle thickness is above the tuning thickness (Strozyk et al., 2012). Thick halite units act as the final lithology deposited for each of the Z2, Z3 and Z4 cycles ( Figure 2); the relative acoustic properties of the halite juxtaposed with anhydrite and dolomites lead to high acoustic impedance contrasts and hence these cycles are clearly differentiated on seismic data (Grant et al., 2019). Critical cycles within this analysis are the Z2 and Z3 cycles, notably the +Z2 and +Z3 cycles (Figure 2), which are the thickest halite units present within the Zechstein of the South Permian Basin and the −Z3 anhydrite/carbonate cycle, which, due to its distinct strong seismic response, acted as the primary marker reflection.
Dip attributes were generated for surfaces of interest, notably the top of the Zechstein Supergroup and the −Z3 surface. This attribute measures the dip in degrees of the surface at each point and applies a colour relating to the dip level; darker areas relate to higher levels of dip (Figures 6-9). This attribute aids in highlighting structural features and changes.

| Depth conversion and velocity modelling
As the SNSMSR2 was in the time domain, a velocity model was created to convert the generated seismic surfaces to the depth domain. A layered velocity model was generated, constrained by the six key stratigraphic surfaces mapped from the seismic data and using the bathymetry for the seabed surface. The model uses time/depth relationship from 75 wells across the survey area. Residuals between the generated velocity model and well tops were used to quantify the accuracy and precision of the model.
The velocity model was built using an iterative process; each time aiming to reduce the residuals to a maximum of 5% for the top and base of the Zechstein. This process involved quality control checking the time-depth data in the wells, removing spurious velocity data, and changing the weighting of different wells' influence spatially on surrounding surfaces. Further information on the velocity modelling and building process can be found in the data repository.

| Structural facies interpretation
To interpret the internal deformation of the Zechstein evaporites, the internal reflection geometries were classified, principally constrained by the geometry of the prominent seismic reflections of the −Z3 Plattendolomit and the −Z3 Hauptanhydrit (Figure 4). Based on these interpretations of the 3D seismic data, the structural facies were characterised based on the following characteristics; observable internal folds and associated geometries such as amplitude and wavelength, reflection terminations, faults present and where they were located and where in the Zechstein cycles have these features observable. A nomenclature was derived to classify the internal Zechstein geometries based on these observed features, which are named as follows: planar and continuous, gently folded and continuous, open and continuous, closely folded and poorly imaged, areas of withdrawal and the chaotic reflections facies. The facies' interpretation included both crosssection views of the seismic data and from the geometries observed on interpreted surfaces in map view.

RESULTS
The base Zechstein is a prominent positive zero phase reflection across the area (Figure 4). Frequently small-scale extensional faults can be interpreted to displace the reflection F I G U R E 4 Petrophysical logs, interpreted lithology log, Zechstein cycle, synthetic-seismic well ties, formation age and calculated velocity model extract of Well 44/27-2 in the Southern north sea. 44/27-2 is located within the depocenter of the South Permian basin (Figure 3). Petrophysical logs shown are Sonic (DT) and Density (RHOB). An interpreted lithology log is present, as well as interpreted Zechstein cycles and stratigraphic ages. Lithological interpenetration for the Zechstein was undertaken at the same resolution as the petrophysical logs allowed. The interval velocities for the well used within the velocity model are present; note the change in velocity at major lithological boundaries. Key seismic reflections are also shown. (Figure 6a,b); these faults do not propagate to the top of the Zechstein being isolated to the Z1 cycle and the base of Z2. The top Zechstein reflection is a negative reflection; however, since there is no significant velocity contrast (Figure 4) between the overlying Triassic interval and the Zechstein, the event is generally of low amplitude (Figure 4). The steep dip at the flanks of both internal and external salt structures makes it challenging for seismic data to correctly image these features' geometries, which exacerbates the difficulty in interpretation. Within the Zechstein interval, the internal geometries vary significantly (Figures 6-9). In areas that are absent of major salt structures, the thickness of the Zechstein Supergroup ranges from <47 m, that is below seismic resolution to >1.5 km (true vertical thickness) (Figure 5a,b). In the core of major salt structures characterised by salt diapirs and anticlines, the thickness of the Zechstein can be >3 km thick (true vertical thickness) (Figure 5b). This current-day thickness varies from the original depositional gross thickness of the Zechstein Supergroup due to halokensis salt mobilisation that has occurred.
The net to gross of halite to non-halite lithologies varies spatially throughout the basin (Figure 5c) In the more proximal areas of the basin, higher quantities of anhydrite and carbonates are present, with halite usually consisting of only <25% of the Zechstein's lithology, halite being 12.6% in well 48/30-6, for example. However, this changes the further towards the basin depocenter with the highest halite % in none structured areas typically >75% halite, 81.3% in well 44/22-5, for example.

| Structural facies characterisation
The following subsections describe the six defined structural facies' characteristics. These sections are ordered by the magnitude of kinematic deformation that has been observed. The terminology we use to define the structural facies is derived from the coherent reflectivity observed from the −Z3 formation, which is the key marker bed used within the Zechstein.

| Planar and continuous
The top Zechstein reflection normally has a planar geometry with little to no structure. Where strata is dipping it follows the base Zechstein reflecting the regional geology; however, rare anticlines are present throughout these areas, and these have large-scale wavelengths of 2.5-10 km (Figure 5a), with dip angles of limbs being <2°. No deformation is observed in the overlying Triassic, Cretaceous or Cenozoic intervals above this structural facies. Reflections within the Zechstein are typically laterally continuous and planar with a dip parallel to the underlying Z1 unit and the uppermost Rotliegend group (Figure 6a). All reflections show a continuous reflection character throughout the area (Figure 6a). The seismic reflection amplitude of the −Z2 is lower than those of the Z3 and Z4 (Figure 6a). A thickness decrease of 60 ms can be identified within the Z2 towards the west of the cross-section (Figure 5a), where the seismic reflections terminate. The overlying +Z3 also exhibits minor thickness changes, with a decrease in thickness towards the west, correlating with those observed in the +Z2 (Figure 6a). The Z4/5 has subtle thickness changes between 12.5 and 20 ms; however, these are distinct from thickness changes or features of the Zechstein below. In these areas, the reflections' dip and dip direction change relate to the dip of the underlying Z1 Zechstein cycle sediments rather than due to the influence or formation of salt-related structures. The average dip of the −Z3 in this area is 2°, with a maximum dip of 4°. There are no observed faults of folds within the salt in these areas.

| Gently folded and continuous
The top salt within this area is commonly folded. These folds have large wavelengths of 2.5-15 km and a trend of NW-SE (Figure 5a). A large open syncline is shown in Figure 6. In areas where top salt forms anticlines or synclines, the −Z3 follows this geometry, forming folds of the same wavelength independent of the smaller folds within the −Z3. While internally, the Zechstein shows evidence of significant deformation, the top of the interval remains relatively undeformed. As the deformation is entirely within the Zechstein Supergroup, there is no observed deformation in the overlying Triassic, Cretaceous or Cenozoic.
The internal reflections within the Zechstein are parallel to those of the base Zechstein; however, there are distinct areas where the reflections display deformation in the form of folds (Figure 6d). These folds are asymmetric and are observed to affect the +Z2 as well as the reflections of the −Z3 (Figure 6d). The folds have a wavelength of 1-3 km, an amplitude of 50-100 ms and limbs which dip ranging from 5 to 14°. The axial planes of the folds are typically gently inclined. The traces of the fold axis are 1-5 km within this area and have no preferred orientation (Figure 6e). Folds are not observed in the +Z3 and Z4/5 Zechstein units (Figure 6d,e). Line length comparison of the top Zechstein and −Z3 was measured and compared, the −Z3 was 2.3% longer than the top Zechstein (12.38 vs. 12.67 km).
The reflections of the +Z3 and Z4/5 are continuous and can be traced extensively across the basin. The +Z3 and Z4/5 Zechstein cycles remain comparatively undisturbed by the deformation occurring in the +Z2 and −Z3. Despite minor localised deformation within the +Z3 above the folds in the −Z3, the seismic reflections are parallel with one another, and the top and base Zechstein reflections. The interval from the top −Z3 reflection to the top Zechstein thins across anticline crests and thickens into synclines, with areas above synclines being up to 200 m thicker than above the anticlines (Figure 6e).
The base Zechstein Z1 has occasional extensional faults (Figure 6d). The top termination for the faults occurs at the top Z1 reflection, unaffecting the Z2, and the base terminate into the sub-Zechstein strata below, with the maximum level of displacement occurring at the Z1 level. The faults have a throw ranging between 35 and 65 m, and a heave between 40 and 200 m.
Rare extensional faults are observed (Figure 6d) within the +Z2 and −Z3, displacing the top of the +Z2 and the base of the −Z3. The faults termination points are within the +Z2 and −Z3. These faults are 380-450 m and a displacement of up to 95 m. Alternatively, these features could be interpreted as monoclines with steeply dipping limbs, as the surrounding folds being compressional, it would be unlikely for extensional faults to be present.

| Open folded and continuous
The top Zechstein is commonly undeformed, remaining parallel with the base Zechstein. However, some areas of the top Zechstein form rare anticlinal and synclinal structures, which trend NW-SE (Figure 5a), similar to observations seen in Section 5.1.2. In Figure 7a, the top Zechstein defines a gentle syncline with a wavelength of 7.2 km. The deformed −Z3 follows the top Zechstein's geometry to form gentle anticlines and synclines of the same wavelength, independent of the asymmetric folds present in the −Z3.
The intra-Zechstein units exhibit prominent levels of deformation, specifically within +Z2 and −Z3 units (Figure 7a). The seismic reflections of the lower +Z2 are increasingly discontinuous, with asymmetrical folds visible, which are also present in the −Z3. By comparison, the top +Z3 and Z4/5 are relatively undeformed (Figure 7a).
The folds in +Z2 and −Z3 are asymmetrical, have inclined axial planes and are similar to those seen in 4.1.2 ( Figure 7a). The asymmetrical folds have wavelengths of 0.4-1 km, amplitudes from 100 to 300 ms and dip on the fold limbs typically ranging from 10 to 35°, with a maximum dip of 38°. The fold traces of the structures range from 1.5 to 9 km long ( Figure 7c). This area shows the fold axis trends in an NE-SW orientation; however, the orientation has been observed to be different in other parts of the basin, with orientations trending closer to N-S. Line length of the top Zechstein and −Z3 was measured and compared, the −Z3 was 8.6% longer than the top Zechstein (9.78 vs. 10.62 km).

F I G U R E 8 Seismic section E-E′ of the
Units +Z3 and Z4/5 is 350-420 m thick within the synclines and only 270-300 m thick above the flanking anticlines of the −Z3 (Figure 7a). Within the Z4 Zechstein unit, there are isolated reflections which terminate onto the stratigraphically lower sections of Z4 ( Figure 7a); there are two possible interpretations to these features: (1) strain-induced thickness changes from salt mobility or (2) these terminations represent intra-Zechstein onlap features, which would suggest a small level of salt mobility syn-depositionally.
Very rare faults are present within the Zechstein, displacing the −Z3 and the top of the +Z2 within the Zechstein in this area (Figure 7a). Reverse faults (Figure 7a) are observed, which terminate downwards in the basal part of the +Z2 Zechstein and upwards in the base of the +Z3. These faults have throws of ca. 120 m, a heave of ca. 240 m, a length of ca. 400 m and a displacement of ca. 230 ms.
The base Zechstein Z1 is deformed NW-SE trending extensional planar faults; however, they are more common and are larger than those observed in the gently folded and continuous facies (Figure 7a). The faults continue to terminate downwards at the top Z1 reflection, with the top of the Z1 having been juxtaposed against the Z2. The basal termination for these faults is still pre-Zechstein strata; however, the length of these faults is greater than those present within in the gently folded and continuous facies, being <1 km. The faults now have a throw ranging between 60 and 80 m and a heave between 50 and 150 m. The faults do not have any spatial relationship with the deformation occurring above within the Zechstein, as faults within the Z1 occur below both anticlines, synclines and limbs of the folding in the above Z2 and −Z3.

| Closely folded and poorly imaged
The top Zechstein is commonly deformed, forming anticlines and synclines with wavelengths of 3-15 km and fold trend orientations of NW-SE. Some anticlines have ruptured, developing into areas in which the Z2/Z3 salt units are seen to pierce the overburden forming diapiric salt structures. The mobile salt that has intruded upwards into the overburden has thinned. The sections where the overburden has been pierced occur at the hinge of anticlines in the top Zechstein or in areas where faults within the overburden come into contact with the top Zechstein.
The Z1 is heavily deformed by NW-SE striking extensional faults (Figure 8a,c). The faults are apparent every ca. 7.5 m to ca. 3.5 km; they have a greater displacement than previous Z1 extensional faults described in the gentle/open folded continuous facies, with displacements of ca. 100-150 ms. The Z1 reflections are displaced significantly, juxtaposed against the above +Z2 (Figure 8a).
The +Z2 is dominated by noisy, discontinuous seismic reflections ( Figure 8a). However, individual reflections within the +Z2 can be differentiated. The reflections have varying high dips up to the ability of the seismic to correctly image them at ca. 66°. Asymmetric folds are visible within the +Z2; however, it is challenging to differentiate any continuous structures due to the data (Figure 8b).
The −Z3 seismic reflections that were continuous in the other previously identified structural facies no longer have a continuous appearance, resulting in the +Z2 and +Z3 not being differentiable in places (Figure 8a). The isolated imaged sections of −Z3 have varying geometries. Sections with asymmetric folds are present within the −Z3, forming open and tight folds. These folds have amplitudes of 65-195 ms, wavelengths of 1.1-1.6 km present and fold limb dips up to ca. 66°. Isolated high-amplitude reflections can be observed when the trajectory of the visible to none visible dipping limb is followed (Figure 8a,b). Some none continuous sections of −Z3 may have no folds present; however, they have planar geometries that have been rotated to a high level of dip, the same dip levels as those observed fold limbs. Isolated areas of the −Z3 within this facies typically have areas of 0.02 to ca. 10 km 2 (Figure 8c). No preferred orientation within the hinge line of the folds, with both symmetrical and asymmetrical folds, can be observed, although this is likely due to the data quality. The top Zechstein in this section was measured to be 24.5 km while −Z3 was measured to be 31.6 km long, meaning a shortening of 29.1% has occurred.
The Z4/5 does not have the same level of intra-cycle deformation as the lower Zechstein units. The seismic reflections remain parallel, with a constant thickness; however, the geometry of the Z4/5 unit has been deformed as it now follows the geometry of the top Zechstein (Figure 8a,b). No faulting is observed within the Z4/5 units. Both the − Z2 and the +Z3 show significant inflation around the fold hinges, with the +Z3 increasing in thickness at its maximum point by 400 ms and the −Z2 by 300 ms. 5.1.5 | Area of withdrawal These areas are typically found adjacent to major salt structures, such as salt diapirs and anticlines. These areas form predominantly elongate oval-shaped areas that align with the salt structures they flank, for example, the middle basin salt walls are completely encased by this facies, which also occurs in WNW-ESE trending belts (Figure 5b).
The Z2 is thinner than elsewhere in the basin, being only 42-47 m thick. In some sections, the lack of −Z2 is such that the Z5 and Z1 could be interpreted as being in contact with one another (Figure 9a). The visible patches of Z2 salt have planar seismic reflections present within the Zechstein. Small areas of high-amplitude reflections are present in the remnants of the Z2 salt ( Figure 9a). The +Z3 unit is not visible, unlike the other structural facies observed within the basin (Figure 9a), thinning until it is no longer seismically resolvable. The isolated high-amplitude patches in these areas may be remnants of the −Z3 having undergone brittle deformation.
Post-Zechstein strata are thickened above these deformation areas, especially within the Cenozoic sedimentary deposits (Figure 2). The strata above the fringes of the area of withdrawal facies always show dips pointing towards the lowest point of top salt, forming large-scale synclines ranging from 0.1 to 10 km scale.

| Chaotic reflections
Seismic reflections are not coherent for any section of the Zechstein, this incoherency could be any number of features, such as the internal stratigraphy still being competent and isoclinically folded, the internal stratigraphy no longer being competent and consisting of unimaged stringers, breccia or the likely highly complex internal structures are not resolvable by the conventional processing used for the SNSMSR2 (Figure 9e). These structural facies typically coincide with salt structures, such as NW trending salt diapirs or large anticlines (Figure 5b). The flanks of these areas are typically surrounded by the withdrawal deformation facies (Figure 9e).
In rare areas, small packages of more continuous, high-amplitude reflections are visible ( Figure 9e); these small packages are only visible on seismic for <1 km and have low dips. Wells that have drilled through these facies and used within this study show more heterogeneity than in the seismic data. Interbeds of greater than the seismic resolution are present but have not been imaged correctly.

| Distribution of internal Zechstein structural facies
The spatial distribution of the different internal structural facies of Zechstein deformation is shown in Figure 10. The distinct structural facies commonly coincide with facies of the adjacent deformation levels; however, the chaotic reflection facies appear enclosed within all other structural facies identified, this is due to salt-cored structures being present within all identified domains.
The structural facies identified provide evidence of spatially varying and a continuum in strain style and magnitude within the Zechstein Supergroup (Figures 6-9). The distribution of the six identified structural facies can be broadly split into three distinct structural domains ( Figure 10). Domain A is located in the proximal area of the basin, it is characterised by the planar continuous and the gently folded deformation facies.
Small areas of chaotic reflection facies are present sporadically throughout this domain. Domain B, within the central region of the study area, is dominated by the withdrawal and undifferentiable facies, trending NW-SE (Figures 5a,b and 10). This domain is characterised by large salt structures, including areas of salt expulsion which have fed these structures. Finally, domain C is located in the most distal part of the basin and is dominated by the closely folded and poorly imaged structural facies. Domain C is the most diverse of the observed domains, with every characterised structural facies present within it.
The spatial distribution of structural facies and hence domains is complex. A very generalised trend could be inferred with increasing levels of internal deformation from proximal areas of the basin towards the distal regions of the basin, with several salt structures in the middle transitional areas of the basin.
F I G U R E 1 0 Internal structural facies map of the Zechstein in the South Permian Basin. The six different structural facies characterised in this study are geospatially mapped. Transects for cross sections (B-G) are located on the map. The western section of the map is dominated by the planar and gently folded and continuous facies, whereas further eastwards, towards the basin depocenter, the closely folded and poorly imaged facies dominates. Observed structural domains have been marked on the map as A, B and, C. 6 | DISCUSSION

| Variation in structural styles
This study has observed a generalised trend for the intensity of internal deformation of the Zechstein to increase basinward (Figure 10), alongside the orientation of −Z3 folds being perpendicular to the basin depocentre in none structured salt areas. We interpret facies in sections from Sections 5.1.1-5.1.4 as a continuum of deformation, with the strain level increasing from one facies to the next, as evidenced by the fold amplitude increasing, wavelength decreasing and a higher level of shortening in the −Z3 per facies (Figures 6-8). However, it is important to note that layer thickness can control fold geometries. While the % of carbonates and anhydrite decreases basinward (Figure 5c), the thickness of the −Z3 does not vary enough to quantify the change in levels of folding, with the thickness to fold amplitude remaining constant (Appendix S1). So layer thickness is unlikely to be responsible for the change in fold parameters observed.
Given that gravity gliding and spreading are interpreted as an important driving mechanisms within the South Permian Basin (Stewart & Coward, 1995), it is suggested that the primary driver for the internal structures observed is salt mobilisation (flow) towards the basin depocenter. This model would be consistent with the northwest to southeast trending fold axis observed internally within the Zechstein and with the increase in fold magnitude towards the basin depocenter. We cannot rule out that the observed shortening may also be related to strain partitioning during the early cretaceous inversion phase within the basin. Density inversion is also an alternate explanation to the formation of such internal Zechstein fold features; however, the orientation of the internal folds indicates an association with a regional stress field, an observation that density inversion does not explain (Figures 6 and 7). Figures 2 and 4 show that the Zechstein supergroup is highly heterogeneous, comprised of four primary different lithologies repeating in the Zechstein 1-5 cycles ( Figure 2). These cycles and changes in lithology lead to a rheological stratification which is often seen in evaporite sequences and is present within the Zechstein (Figure 11e,f) (Rowan et al., 2019). This rheological stratification and hence mechanical stratigraphy affects how intra-salt units accommodate stresses applied to them (Evans & Jackson, 2021). These parameters control the structural styles that develop within an evaporite sequence.

| Vertical strain distribution and flow regimes
Two observations consistent throughout the basin for the characterised internal structural facies of the Zechstein are as follows: (1) the observed vertically varying levels of deformation. Each Z cycle, for example, +Z2, −Z3, Z4/5, has undergone a different magnitude of strain. The vertical stratification and partitioning of deformation is likely a result of two separate factors, lithology and hence rheology, as different lithologies have different rheology (Burliga, 1996), and types of flow that have occurred internally within the Zechstein (Davison et al., 1996); and (2) the relatively undeformed overburden in areas of intense internal deformation within the Zechstein (Figures 6 and  7). One interpretation for this formation is the mechanical decoupling of the Zechstein from the overburden, with the Zechstein acting as an internally heterogeneous mega detachment. An alternative interpretation for this observation is that the internal deformation formed prior to the deposition of the overburden. Deformation occurring before overburden deposition would allow for the difference in the deformation magnitude between cycles and also explain the observed possible internal Zechstein onlap observed in Section 5.1.3. However, further investigation into syn-depositional Zechstein mobilisation would be needed and is outside the remit of this study.
The observed internal deformation of the Zechstein (Figures 6-9) is likely indicative of the dominant flow regime within the salt (Figure 11a-d). Nearly all flows occurring within salt units are hybrids, however, comparing them to ideal flows provides a good base point for identifying the type of flow occurring (Jackson & Hudec, 2017). The observations made in the study area show that the that the stratigraphic middle of the Zechstein, where the −Z3 is located has undergone significant shortening compared with the top Zechstein (up-to 29.1% within the closely folded and poorly imaged facies), and hence has experienced the highest magnitudes of longitudinal strain (Figure 11a,c,e,f), with lower magnitudes occurring stratigraphically below and above. The structures that are resolved in the −Z3 and the adjacent +Z2 indicate a higher magnitude of strain at the base of the Zechstein compared with the Z3+, Z4/5 (Figures 6-9), which is unusual for typical flow deformation styles as typically it is higher towards the top (Figure 11a-c). The +Z2 commonly has asymmetrical folds (Figures 6-8) present and deformed seismic reflections directly below the Z3+, Z4/5 reflections which are undeformed. The observed geometries of the Zechstein, could be explained by a number of possible different flow profiles, are (1) Simple Couette flow (Simple Shear) (Figure 11b,e) of the +Z2 and −Z3 with the Z3+/4/5 mechanically acting as the overburden and the Z1 and −Z2 and acting as the underburden. For this explanation, a shear zone will have developed in the upper salt units to decouple the lower −Z3 from the overburden, the competency difference between the −Z3 and +Z3 will have aided in the development of the shear zone. The observations supporting this possible flow profile include the increasing strain from the +Z2 into the −Z3, the lack of deformation with seismic reflections remaining parallel to one another in the Z3+/4/5, and no deformation in the −Z2 and Z1. This flow profile would also explain the geometry that Z4/5 take in Figure 7, where it has been uniformly deformed and behaves like the overburden. The lithological and hence rheological differences of the Zechstein can further help to explain this possible flow profiles; the −Z2 and Z1 act as the underburden as they are comprised of competent carbonates and anhydrites, the above −Z2, consisting of pure halite, as the halite is acting as a décollement between the two surfaces; (2) Asymmetric Poiseuille flow (Figure 11c,f) could also be an alternate dominant flow profile. Observations in support of this interpretation are maximum strain being constrained to the middle −Z3 unit, with it allowing for asymmetric levels of deformation stratigraphically, with greater levels of deformation occurring in the +Z2 over the Z3+/4. In this flow profile, there is still a décollement between the −Z2 and +Z2, and the upper Z4/5 is behaving as part of the overburden. The gently/open folded continuous and closely folded and poorly imaged facies have experienced compressional stresses, as seen by the folded features observed in each facies (Figures 6-9). These internal compressional folds often occur; without affecting the overburden, with a lack of structure to the top salt, without compressional faults in the overburden and no syn-kinematic thinning strata above these areas. The lack of such apparent compressional features below or above the Zechstein suggests that much of this compression may have been restricted to solely within the Zechstein. The trend of the structural facies and the distribution of these suggest a lateral distribution of the compressional stress increasing with distance down dip.

| Alternate presence of stringers
The closely folded and poorly imaged facies is the end member of the continuous compressional deformation observed throughout the research area ( Figure 10). The maximum dip that was interpretable on any feature within the seismic data was >ca. 66°. Within the closely folded and poorly imaged facies, we observe folds with dips up to 66° after this, the reflections cannot be mapped as they have a discontinuous appearance. The ability of the seismic data to resolve high dip features, therefore, leads to questions about the observed lack of continuity within the −Z3 in this structural facies, in that, whether the −Z3 is present and not just imaged correctly or if the −Z3 has undergone brittle deformation and formed stringers, and, if they do, how did they develop. We believe that the −Z3 is continuous and has not undergone brittle deformation and suggest two alternate scenarios for why these features appear on seismic data as they do; (1) The −Z3 is continuous and has not undergone brittle deformation. The observed lack of continuity between the −Z3 is due to the fold limbs being steeply dipping so that the seismic data cannot image it (Figures 8a and 9d). The folds now have tight/fan geometries with limbs dipping >66°. Further evidence is opposing fold limbs dipping towards high-amplitude reflection packages while no longer being visible on seismic data as the dip of the −Z3, likely as the dip becomes too great to image (Figure 8a). This explanation fits well with our proposed nomenclature for the observed structural facies, in that the magnitude of folding is increasing basinward's F I G U R E 1 2 Proposed model for interpreting the internal heterogeneity of salt structures where poorly imaged by seismic data. The initial stages of the evaporite formation are bedded and undeformed. When salt mobilisation occurs, the competent layers undergo either brittle or ductile deformation. Which type of deformation it undergoes determines the internal heterogeneity of the surrounding salt structure as the competent layers will be mobilised into the structure. Cross-section of the Benthe salt dome, a cross-section of an unnamed salt diapir redrawn after Pichat (2022). and just not well imaged. This is further supported by the facies' line length changes, which show an increasing % in shortening within the −Z3 in our measured examples of the facies in sections 5.1.1-5.1.4, going from 2.3%, 8.6%, 12.6% and finally 29.1%. Mined Zechstein structures that can be used as analogues have been observed to have features dipping as steeply as 90°, suggesting that folds with dips greater than the seismic can image are possible within evaporite formations ( Figure 12). Alternatively, (2) the −Z3 is continuous and has not undergone brittle deformation. As the fold magnitude has increased from the higher magnitude of strain basinward, the fold limbs of the −Z3 have thinned, and the fold hinges thickened. The fold limbs being thinned to below the thickness below that of the seismic resolution combined with how steeply dipping they are would explain why they are not observed on the seismic data. The thickened fold hinges would also explain why we see isolated high-amplitude packages of −Z3 where we would expect the fold hinges to be (Figure 8a,b). This possible explanation is also consistent with our observation of increased shortening towards the basin depocenter.
However, these alternate explanations of the appearance of the −Z3 reflection on the seismic data do not preclude the ability of this formation from being overpressured, a common observation for stringers in evaporite sequences. Laterally extensive overpressure within evaporite lithologies is possible (Dale et al., 2021), so it should not be seen as a defining diagnostic feature suggesting stringers are present. Extra to this, as suggested in explanation 2, if the fold limbs have thinned, they may have undergone diagenetic changes making them impermeable and allowing further mechanisms for overpressure to occur in fold hinges.
We are, however, more confident with the interpretation of stringers being present within the area of withdrawal facies, as the lack of salt in these areas due to expulsion means that the −Z3 stringers are unable to have dips >66° so lack of appearance on seismic is unlikely due to poor imaging. Furthermore, areas of withdrawal have experienced extensional forces from salt flow as salt was expulsed to feed surrounding salt structures, causing boudinage and brittle deformation to occur with the −Z3.

| Model-driven internal deformation of salt structures
Conceptual models help determine the different internal structures that can form within chaotic reflection areas. The best analogues for the internal structural heterogeneity of diapirs originate from mined evaporite structures that have been mapped, with those such as the Hanigsen-Wathlingen salt dome, which shows boudinaged and incoherent Z cycles within the salt, or the Bartensleben Diapir, where the internal Zechstein units are deformed but still competent and distinguishable ( Figure 12). We suggest a model-driven process for interpreting the internal heterogeneity when seismic data do not distinguish coherent reflections. The model is driven by the closest facies to the area of chaotic reflections that has a resolvable internal heterogeneity, often the area of withdrawal facies. Assumptions are made that the competent layers, such as anhydrites and carbonates, within an evaporite sequence are either brittlely deformed or competently deformed as salt flows to feed the surrounding salt structure. Thus, structures surrounded by the withdrawal facies with high-amplitude packages (Figures 9a  and 12), interpreted as brittlely deformed competent layers, have these brittlely deformed stringers present within. Alternately, for structures next to areas where the competent layer has undergone ductile deformation, the layer remains competent as it is mobilised to feed the growing salt structures (Figure 12). To a lesser extent model could also be used to identify what residual components that may remain within the specific areas of withdrawal facies by observing the surrounding structural facies.

| CONCLUSIONS
The Zechstein of the South Permian Basin can be characterised by six distinct internal structural facies. Each facies represents a type of internal deformation or unique salt deformation. Four of these structural facies represent a continuum in the increase in the magnitude of strain increasing basinward, due to the structures observed. We propose this is due to salt flow towards the basin depocenter. These internal structural facies of the Zechstein can be further grouped into three distinct broader domains throughout the basin. Observations indicate that a combination of the type of internal salt flow and the variation in vertical mechanical stratigraphy control the structural facies and the internal geometries that form within the Zechstein.
We suggest an alternative interpretation for the lack of continuity in the seismic imaging of the Zechstein's Hauptanhydrit and Plattendolomit (−Z3), which are often interpreted as stringers formed via brittle deformation. Our observations suggest that the lack of continuity with the seismic reflections are from the limbs of fold structures with dips steeper than can be imaged or thinned beyond seismic resolution, having undergone ductile deformation from salt flow, rather than the Zechstein no longer being competent, having undergone brittle deformation in these areas.
A model-driven approach for deriving the possible structures and heterogeneities has been conceived for the internals of salt structures or areas where the seismic imagining of salt bodies is too poor to be interpretable. This model relies on observations of the internals of the surrounding salt, which has fed these structures. If heterogeneities such as stringers are present, we suggest they will have been expulsed into the salt structures; if not we suggest that the layers within the evaporite sequence will remain competent.