Aeolian sandstones of the Permian Rotliegend Group comprise the principal gas reservoirs of the Southern North Sea Basin. The upper portion of the reservoir interval comprises a unit informally called the ‘Weissliegend’. This unit comprises mass-flow and in-situ soft-sediment deformed deposits formed when the Zechstein Sea catastrophically flooded the Rotliegend sand sea. The Weissliegend typically exhibits poorer reservoir properties than the underlying Rotliegend reservoirs and its thickness and distribution are notoriously difficult to map.
Hanging vertical sections on a regionally extensive, intra-Rotliegend super-surface has led to the recognition of a preserved topography at the top of the Rotliegend. The thickness differences within the unit between the super-surface and the top Rotliegend (termed here the Upper Aeolian Unit, UAU) represent the original dune topography modified by flood-related processes, and provide an insight into the nature and scale of the aeolian bedforms that existed within the basin prior flooding. The relationship between known Weissliegend distribution and this topography provides a predictive tool for understanding Weissliegend distribution.
Mapping the preserved topography reveals preserved bedforms up to 85 m high, with a bedform spacing of 8 to 10 km. The dunes deposited ‘transverse type’ strata but the scale and spacing of the bedforms are too big compared with modern transverse dunse. Instead a more complex model is proposed. The Weissliegend has highly variable thickness (0–26 m) and is commonly thicker within the interdune hollows.
The model for a preserved topography at the top of the Rotliegend is further supported by thickness trends in the overlying Zechstein units which increase into the proposed interdune lows.
The Permian Rotliegend Group is the main reservoir rock in the Southern North Sea (Fig. 1) with gas derived primarily from underlying Westphalian coals (Glennie & Provan 1990). The Leman Sandstone Formation, the reservoir interval, is predominantly aeolian with minor fluvial and sabkha deposits present (Glennie 1990). The proportion of sabkha deposits increases towards the north where the basin centre was occupied by a large lake (Glennie 1990). The proportion of fluvial sediments increases towards the south and west where basin-margin alluvial fan systems introduced sediment to the basin. The hydrocarbon seal is the Upper Permian Zechstein, comprising carbonates and evaporites (Fig. 1) that also greatly reduce seismic energy. This study concentrates on the southern part of the basin where the main erg was situated during Rotliegend times. In the study area the upper part of the Leman Sandstone is represented by the Weissliegend facies. This was proposed by Glennie & Buller (1983) to have been produced by soft-sediment deformation and reworking of underlying Rotliegend dunes during the transgression of the Zechstein Sea. The Weissliegend has generally poor reservoir properties, is variable in thickness and is notoriously difficult to map and predict, partly because its base is difficult to pick from wireline logs.
In this study, rather than hanging sections from the top of the Rotliegend, wells are hung from an intra-Leman surface interpreted to have been approximately horizontal at the time of the Zechstein transgression (base of unit 11 of Howell & Mountney (1997) and base of Unit 5b of George & Berry (1993)). This surface can be mapped across the study area with a variety of expressions and is interpreted as a super-surface (Fig. 2) (sensuTalbot 1985). Below the super-surface, a series of stacked 10 to 15 m thick cross-bed sets are separated by bounding surfaces, interpreted as the lower portions of a succession of climbing aeolian bedforms. Above this surface lies a single, cross-bedded unit of aeolian sandstones (up to 85 m thick) lacking major bounding surfaces, and is termed here the ‘Upper Aeolian Unit’ (UAU). The UAU represents the preservation of all, or virtually all, of the dune system that was present at the time of the transgression, as it was partially modified by flooding-related processes that locally produce the Weissliegend facies. Variations in thickness of the UAU represent the dune topography that existed after the Zechstein transgression (Howell 1992; Howell & Mountney 1997). The Weissliegend is predicted to be thicker in the palaeo-interdunes and thinner on the dune crests.
Aeolian topography preserved above depositional super-surfaces (sensuKocurek & Havholm 1993) has been described from outcrops. The Cretaceous Etjo Sandstone of Namibia was flooded by the Etendeka flood basalts and has a preserved topography of up to 90 m (Mountney et al. 1999; Jerram et al. 2000). The Entrada Formation of southwestern USA (Eschner & Kocurek 1986; Benan & Kocurek 2000), the Minnelusa Formation, Wyoming, USA (Vincelette & Chittum 1981) and the Rotliegend equivalent ‘Yellow Sands’ of Durham, UK (Glennie & Buller 1983) are other examples of preserved topography in flooded dune systems.
The dataset and methodology
The dataset for this study includes 139 wells from an area covering 16 Licence Blocks (4056 km2) in the southern part of the Southern North Sea basin (Fig. 3). Sonic, density–neutron and gamma-ray logs (scale 1:200) were available from all wells and core or core photographs were used for 39 wells. The quality and coverage of the available data vary and were better in the eastern parts of the study area.
The super-surface was initially identified in core, calibrated with the wireline logs (139 wells) and then traced across the study area. The character of the surface and the facies associated with it were then mapped. Once the super-surface was identified, the thickness of the Upper Aeolian Unit was measured. Also the thicknesses of related stratigraphic units were measured from each well and include: UAU, the super-surface to the base of the Kupferschiefer; the Kupferschiefer; the Zechstein Kalk; the Werra Anhydrite; the total thickness of the Rotliegend.
The thickness of the Weissliegend was also noted, initially in the cored wells and then, with less confidence, from well logs. From the thickness data a series of maps was produced using a digital mapping package.
Basin Evolution and Regional Depositional Setting
The study interval was deposited within the Southern Permian Basin. Collision between Gondwana and Laurussia caused north–south compression in late Westphalian time and the Variscan Highlands south of the North Sea basins were created. These highlands started to collapse in the late stage of the orogeny (Early Permian time) and transtensional and transpressive movements developed southeast–northwest and southwest–northeast grabens and half-grabens across the foreland. By the Late Permian these fault systems became largely inactive and regional thermal subsidence led to the development of the Northern and Southern Permian Basins (Glennie 1990) (Fig. 1). The southern basin stretched from Poland to the east coast of England. The basin was constrained by the London Brabant Massif in the southwest, the East Midland Shelf to the west and the Mid North Sea–Ringkøbing–Fyn High in the north (Fig. 1). Smith & Taylor (1989) suggest that the southern basin was in connection with basins in the north through the north–south striking Central and Horn Graben, at least during Zechstein times.
During the Permian, the Southern North Sea was at latitudes comparable to the modern day North African–Arabian Deserts (10–15° north of the equator: Glennie 1983a). Before the onset of Rotliegend deposition extensive erosion and denudation generated the Saalian Unconformity that separates the faulted Carboniferous basement from overlying Rotliegend sediments (Glennie 1990). Glennie (1990) reports four facies within the Rotliegend Group: fluvial, aeolian, sabkha and lacustrine (Fig. 1). In the central and northern parts of the basin claystones, siltstones and evaporites were deposited within a semi-permanent desert lake, fringed by a sabkha plain. The deposits of the lake and sabkha comprise the Silverpit Formation (Fig. 1). The main erg, including the location of this study, lay to the south of the sabkha. Fluvial deposits of basin margin alluvial fans (George & Berry 1993, 1997) fringe the southern and western margins of the erg. The fluvial and aeolian facies make up the gas-bearing Leman Sandstone Formation.
Recent studies of Rotliegend palaeogeography (George & Berry 1993, 1997; Howell & Mountney 1997) show a large erg influenced by a trade wind belt with winds from the east-northeast (Glennie 1983b). From dipmeter data the dunes are interpreted as predominantly transverse-type bedforms, migrating towards the west (Glennie 1983b). George & Berry (1993) also proposed that compound star dunes were present in the centre of the erg. On the margins of the erg, smaller barchanoid dunes and sand-sheet deposits are interbedded with sabkha and desert lake facies.
Permian subsidence outpaced sedimentation and, by the time the Zechstein transgression catastrophically terminated continental sedimentation, the basin has been proposed to have been up to approximately 250 m below sea-level (Glennie 1990). The rapid flooding caused soft-sediment deformation and reworking of the upper parts of the active Rotliegend dunes forming the Weissliegend facies (the focus of this study). Relatively deep marine conditions ensued and the black bituminous mudstone, the Kupferschiefer, was deposited. This, in turn, was followed by the carbonates and evaporites of the Zechstein Group (Glennie 1990).
Facies and Facies Association in Core and Wireline Log
The facies identified in core and wireline log are summarized in a table based upon this study, Howell (1992) and George & Berry (1993) (Figs 4 & 5). Facies have been grouped into facies associations based upon interpretation of the depositional environment. In the study area the majority of the Rotliegend is composed of aeolian dune deposits interbedded with the volumetrically less significant interdune facies. A regionally extensive aeolian sand-sheet overlies the mapped super-surface, and locally fluvial, sabkha and lake facies were also observed.
Aeolian dune facies association
Well cross-stratified sandstones are the most common facies in core. Cross-stratification is defined by grain-size and typically dips at 15 to 30° (Figs 4 a, b & c). These sandstones are fine to medium grained, well sorted, well rounded and generally red in colour (white and grey are also common). Cross beds are typically a few centimetres thick and are separated by thin (millimetre-scale) lightly coloured laminae. The thin laminae are finer grained and well cemented. The thickest laminae occur at the top of units and at the base they wedge out, become thinner and more cemented. Co-set thickness is up to tens of metres with foreset dips also decreasing towards the base. At the base, the foresets interfinger with wind-ripple lamination.
Wireline log character
In wireline logs the dune deposits show a low and blocky gamma-ray pattern. Resistivity logs indicate high porosities and permeabilities. The base of the co-sets, the bounding surface, has a higher sonic velocity and a small density–neutron log separation (Fig. 5).
The aeolian cross-bedded sandstone was deposited by grain flow and grain fall processes on the lee-side of dunes as described by, for example, Hunter (1977). At the base of the bed a bounding surface forms and is created when a dune migrates on top of another (Hunter 1977). These surfaces are well cemented. The predominance of aeolian strata and the thickness of preserved cross-bed sets are taken as evidence that the study area was within the central portion of a large erg. This corresponds with other palaeogeographical reconstructions of the area (Glennie 1990; George & Berry 1993, 1997; Howell & Mountney 1997). The interpreted topography at the top of the UAU has important implications for interpreting the style of aeolian bedform. This will be discussed in more detail below.
Aeolian interdune facies association
The interdune facies association includes planar laminated, moderately sorted and rounded sandstone; wavy laminated heteroliths and, thin, dark red mudstones. The bed thickness ranges from 0.1 to 2 m and the grain size from silt to coarse sand. The base of these deposits is erosional and there is an abundance of lensoid and wind-ripple laminae. Occasional concentrations of coarse and fine granules are seen. Commonly, the interdune association has a dark red colour and is more cemented than underlying dunes. When strongly cemented they may also be grey-green in colour. Finer-grained intervals are typically associated with poor core recovery.
Wireline log character
Because of the more cemented and less sorted nature of the interdune association it produces a faster sonic velocity and closer density and neutron logs than the aeolian dune sandstone (Fig. 5). Wavy-laminated, heterolithic examples exhibit higher gamma-ray, faster sonic velocity and higher density values than the sand-dominated examples and the density and neutron logs show a negative separation pattern.
Wind-rippled sandstones, wavy-laminated heteroliths and thin dark red mudstones are interpreted as interdune deposits. Saltation processes form wind-rippled sandstones when sand is blown over a dry surface. Wind-rippled sandstones dominate this facies association within the study area. The wavy laminae are adhesion ripples produced by sand adhering to a damp surface (Glennie 1970). Mudstones indicate deposition from suspension during periodic flooding of the interdune area. The sharp base to the aeolian unit was generated by erosion related to bedform migration. The rare presence of damp interdune deposits indicates that the system was partly a damp aeolian system (sensuKocurek & Havholm 1993) and water table rise was an important factor in its preservation (Stokes 1968). Damp interdunes are typically thin and laterally discontinuous (Talbot 1985; Sweet 1999). The grey-green colour seen in some of the cores is sediment that was never oxidized and iron that kept its 2+ form.
Sand-sheet facies association
In core the sand-sheet facies association comprises horizontally bedded, bimodally sorted wind-rippled deposits that are well cemented. The sand-sheet association is superficially similar to the dry interdune deposits described above, and the deposits in this association are differentiated by their greater thickness, typically 2 to 3 m (Figs 4h & i). The base of these deposits is erosive and the colour darker red than surrounding dune strata.
Wireline log character
Sand-sheet deposits typically show slightly higher gamma-ray response than aeolian dune sandstones. They are differentiated from fluvial deposits by high and ‘spiky’ sonic and density log responses (Fig. 5).
Tractional and saltation processes form the wind-ripples dominating the sand-sheet facies. Thick accumulations of horizontally bedded wind-rippled sandstone are interpreted as aeolian sand-sheets. The formation of sand-sheets rather than dunes depends upon certain conditions (high water table, early cementation by salts etc.) typically associated with high water table and comparably wet climates (Kocurek & Nielson 1986). The distinction between this association and the dry interdune deposits is important. Interdune surfaces climb as the bedforms they lie between migrate. Consequently they are diachronous. Interdune deposits described from other studies (e.g. Kocurek 1988) are typically thin (less than 1 m) and not laterally extensive, as the conditions required for stable, wet aeolian systems require the water table to rise at the same rate as the bedforms climb. Howell & Mountney (1997) proposed that laterally extensive aeolian sand-sheets were associated with super-surface formation in the Rotliegend. A super-surface forms when an erg becomes sediment starved for a long period of time (mainly due to wetter climate) (reviewed by Fryberger 1993) and the area is covered by an extensive sand-sheet with few (if any) undeveloped dunes. A super-surface represents a cessation in erg deposition, is laterally extensive and cross-cuts all bounding surfaces related to bedform climb (Kocurek & Havholm 1993). Deflation super-surfaces (sensuKocurek & Havholm 1993) are typically very planar and dip at less than 1° .
Fluvial facies association
Fluvial deposits are rare in the cores that were studied. Where observed the association is generally composed of medium to coarse-grained sandstone, is moderately sorted, and beds show a fining-upward trend. Deposits of this association typically have a greater abundance of exotic mineral grains than the aeolian sandstones. The fluvial units are typically massive or contain water-escape structures and have a white to white-grey colour. Finer-grained examples also contain current ripples. Rarely within the study area, the facies association may contain rip-up clast conglomerates (Fig. 5) (Glennie 1990) and examples are locally overlain by thin beds of mudstone (Fig. 4).
Wireline log character
Because of the higher proportions of mud and clay, fluvial sandstone shows higher gamma-ray values than the aeolian deposits. Because of the abundance of cements, sonic and density values are higher than the aeolian sandstone. Fluvial units often show a blocky and ‘bell-shaped’ fining-upward log pattern (George & Berry 1993) (Fig. 5).
Because of their predominantly massive nature, fluvial deposits are interpreted to have been deposited by flash floods. Examples that are well sorted probably originated within the erg, whilst examples containing extra-basinal material were introduced from outside. Mudstone beds capping fluvial deposits are interpreted as abandonment lakes deposited as the channel is cut off from the main flow and finer-grained material settles (Sneh 1983; Glennie 1990; George & Berry 1993). The fluvial deposits are the result of flash flood events. Their occurrence in the centre of the erg is taken to indicate wetter climatic periods.
Sabkha/lacustrine facies association
This facies association was not recognized in the studied cores but was interpreted in well logs from the north of the study area. In published descriptions sabkha deposits contain poorly bedded claystones, siltstones and sandstones with abundant mud cracks, adhesion ripples and anhydrite nodules. The lacustrine deposits are typically finer grained, tending to mudstones with minor siltstones. These are red-brown in colour and may locally contain several halite horizons (Glennie 1990).
Wireline log character
The finer lacustrine member can be distinguished from the broad log anomaly of the sabkha deposits as it has a thin and very prominent kick on the gamma-ray log due to higher content of clay and evaporitic minerals. Separating the interdune and sabkha/lacustrine deposits can be hard but the interdune association has a generally less pronounced wireline kick. Examples have high gamma-ray values and negative density–neutron separations.
The sabkha deposits were deposited as damp plains where wind-blown mud and sand adhered to the damp surface. Continued evaporation and ground-water recharge resulted in precipitation and dissolution of salts (Glennie 1970). Adjacent to the sabkhas were standing bodies of water where the lacustrine facies were deposited. The lacustrine deposits were laid down by fine-grained, wind-blown material settling out of suspension, together with evaporites. To the north of the study area, these facies form the Silverpit Formation, deposited in and around, a large basin lake. High frequency fluctuations in lake level occurred during climatic changes and wetter periods pushed the lake margin south and into the study area (Glennie 1990; George & Berry 1997).
Soft-sediment deformed and reworked deposits: the Weissliegend
In the study area and across most of the southern Permian Basin, the top of the Rotliegend is marked by a variably thick, reworked and soft-sediment deformed unit called the Weissliegend. The term Weissliegend is an old German Kupferschiefer miners’ term for the ‘underlying white beds’ of sandstone, which occur at the top of the Rotliegend. In the North Sea it has since become synonymous with the complex reworked and soft-sediment deformed deposits that resulted from the Zechstein transgression of the Rotliegend erg and is not based on its stratigraphic position or its mode of formation (Glennie & Buller 1983). The Weissliegend is up to 50 m thick in parts of the Southern North Sea (Glennie & Buller 1983).
For the purpose of this study the Weissliegend has been divided into two facies associations: in-situ reworking or reworking by resedimentation.
In-situ Weissliegend facies association
The deposits of this association are up to 6 m thick within the study area. These units are predominantly grey, although thicker examples are rarely red and light brown towards their base. These sandstones are predominantly fine to medium grained and well sorted. The deposits are both completely massive or contain rare contorted laminations highlighted by thin, black, silt-rich laminae. Contorted laminae are irregular, wavy and over steepened (Figs 4d & f). Most significantly the base of this facies is diffuse and gradational into the underlying dune sands. The solely massive deposits are commonly overlain by the resedimented Weissliegend described below. As with the resedimented Weissliegend, the top of the interval is locally bioturbated and the contact with the overlying Kupferschiefer is always sharp. The association is often more cemented than the underlying aeolian dunes.
Wireline log character
The wireline log character for both the Weissliegend associations varies greatly across the study area and core-to-log correlation is essential. In the wireline logs the Weissliegend is typically less ‘clean’ (higher gamma-ray, less density–neutron separation), especially towards the top. The increased cementation produces higher sonic-velocities and density values (Fig. 5). In the case of the in-situ Weissliegend, the base is typically very difficult to pick on well logs alone.
The massive to soft-sediment deformed sandstones are interpreted to have been generated by in-situ homogenization and deformation of the primary sedimentary fabric. The lack of tractional structures, the soft-sediment deformation structures and the lack of a sharp base favour a mode of formation associated with pore pressure increase and rapid water or gas escape. Glennie & Buller (1983) proposed four different mechanisms that can trigger soft-sediment deformation of dunes: (i) air-entrapment and expulsion; (ii) cyclic loading by earthquakes and (iii) storm waves; and (iv) changes in water-table level. Glennie & Buller (1983) favoured the air-entrapment mechanism to cause the large-scale deformation of the Rotliegend dunes. Fluctuations in the water table are believed to be confined solely to the lowermost parts and marine reworking to the uppermost decimetres of the dunes. The Weissliegend is predominantly grey because it was flooded before it could be buried below the Rotliegend water table and the reddening process (oxidation of irons entrapped in pellicle clay; Turner 1980) could occur. Reducing fluids from the Kupferschiefer may also have caused the lack of reddening some time after the flooding. Rare red and brown examples indicate parts of dunes, which were buried down to the water table prior to deformation.
Resedimented Weissliegend facies association
Resedimented Weissliegend deposits have similar textural parameters as the previous facies. They are typically massive to planar bedded and are mostly interbedded with thin, dark mudstone intervals (Howell 1992). Beds are sharply to erosively based. This facies is typically thicker than the in-situ deposits, reaching a maximum thickness of up to 26 m in the study area. Towards the top, evidence of marine reworking is rare, confined to the top 10 cm where present (Glennie & Buller 1983) and is more typically absent.
Wireline log character
Redeposited Weissliegend is identified by its sharp base and the greater proportion of thin claystone interbeds giving gamma-ray anomalies. As with the in-situ homogenized Weissliegend, the units are typically more cemented and consequently have higher sonic velocities than the underlying dunes (Fig. 5). The base of this Weissliegend association is generally easier to pick in the wireline logs as it is more cemented than underlying and overlying strata.
The recognition of a sharp erosive base and tractional sedimentary structures, however rare, is significant. This implies that at least some of the Weissliegend (originally deposited as dune sandstones) were redeposited by mass-flow processes, as sandy gravity flows. Thin interbedded siltstone intervals similar in character to the Kupferschiefer are additional evidence for this process. They indicate that the deposition of the Kupferschiefer was interrupted by mass-flow events. In many cases, because of the well sorted nature of the original aeolian deposits, it may be very difficult to distinguish massive sandstones deposited by mass flows, from those generated by in-situ homogenization. In such cases the vertical context is significant in making an interpretation. It is also significant to note that the Weissliegend formed through in-situ homogenization is generally considerably thinner than that formed through resedimentation.
Heward (1991) studied comparable Rotliegend deposits in the Northern Permian Basin and proposed that the considerable thickness (>50 m) of some Weissliegend facies, with relics of former dune bedding, is too thick to be produced by marine reworking. Heward (1991) proposed the Weissliegend facies to be a non-marine mass-flow deposit that formed during intense rainstorms prior to the Zechstein transgression. The undisturbed dune bedding within a deformed sequence of Weissliegend facies is instead interpreted as a temporary return of aeolian sedimentation between Rotliegend rainstorms. However, there is very little evidence in the study area to suggest that the climate was wetter towards the end of Rotliegend times. In the area around the north of Quads 48 and 49, north of the study area, a unit informally called the Upper Leman Sandstone overlies the playa lake deposits of the Silverpit Formation and records a contraction of the basin lake (Howell 1992). Consequently it would appear that far from getting wetter, the climate towards the end of the Rotliegend was, in fact, becoming more arid within the study area.
In summary, the Weissliegend is interpreted to have been formed as a result of the Zechstein transgression and rapid flooding of the basin as proposed by Glennie & Buller (1983). Flooding occurred as a result of either post- or inter-glacial rise in sea-level or a phase of Permian rifting in the proto-central graben. The rapid flooding of the basin and drowning of dunes resulted in in-situ soft-sediment deformation through air and water escape (as proposed by Glennie & Buller 1983) and mass flow and resedimentation of dunes that were saturated and unstable.
The Super-Surface and the Upper Aeolian Unit
In all of the wells studied there is a notable change from cross-bedded aeolian sets, less than 10 m thick, with thin interdunes, in the middle of the Rotliegend to a thick cross-bedded unit at the top, lacking significant bounding surfaces (Fig. 2). This upper unit is termed here the ‘Upper Aeolian Unit’ (UAU) and is overlying a regionally extensive sand-sheet facies (base upper aeolian unit, see Fig. 2). The UAU comprises cross-bedded sets up to 85 m thick, although the thickness is highly variable. The top of the UAU includes the Weissliegend facies, which may exhibit either a sharp or diffuse base.
A regionally extensive sand-sheet up to 3 m thick typically marks the base of the UAU and separates the upper unit from the underlying stacked sets of aeolian cross-beds. Although in the southern part of the region fluvial deposits are present and in the north the sand-sheet passes laterally into sabkha deposits (Fig. 6). Well-to-well correlations indicate that the surface at the base of the sand-sheet truncates draa migration surfaces and consequently it can be interpreted as a super-surface (Talbot 1985; Kocurek & Havholm 1993). Super-surfaces form when erg deposition is terminated either by non-deposition or erosion driven by external factors such as climate change, sediment supply and/or erg dynamics or tectonics. This surface was originally identified by Howell (1992) and Howell & Mountney (1997) and can be correlated across the entire basin as the base of their unit 11, marking a temporary change to a wetter climate, coupled with lake expansion to the north and increased fluvial activity at the margins of the basin.
An investigation was undertaken into thickness variations in the UAU and their links to changes in the thickness of the Weissliegend, overlying formations in the Zechstein and the thickness of the Rotliegend. These thickness changes are central to understanding the topography and morphology of the erg at the time of transgression. Much of what follows is based upon the assumption that the super-surface was close to planar and horizontal prior to the deposition of the UAU. This assumption is based upon (i) correlation of markers in the middle portion of the Rotliegend below the super-surface, that are sub-parallel and indicate that topography existed at both the base and top of the Rotliegend and (ii) observations of super-surfaces from other systems which are typically flat (see Kocurek 1988). The identification of this surface as an approximately horizontal and synchronous surface is fundamental to the recognition of topography in the overlying UAU.
Stratigraphic Thickness and Implications for Preserved Topography at the Top of the Rotliegend
Upper Aeolian Unit
The thickness of the interval above the super-surface (traced in wireline logs) and below the Kupferschiefer varies across the study area. An isopach map based upon 139 wells shows that the UAU is only a few metres thick in the northeastern parts of the study area (Fig. 7a) and thickens to 85 m in the southwest of the study area. This height is similar to the 88 m recorded in the Sole Pit erg by George & Berry (1997). Towards the centre of the study area the pattern becomes more undulose and the consistency of isopachs illustrates thicker units aligned with a northeasterly trend. These have a spacing of approximately 10 km.
The northeasterly trend of the isopach was chosen after trying different gridding techniques including techniques such as inverse distance to a power, nearest neighbour, triangulation and many more. Each of the techniques show a similar northeast trend and, when gridding with a technique that highlight bulls-eyes in the data (a technique called inverse distance to a power), there is a clear elongation of isopach bulls-eyes towards the northeast. This trend was considered when highlighting the remnant topography in the central and eastern parts of the study area, which has most data points. No great confidence can be placed in the isopachs in the southern and northern parts of the study area as these do not contain many data points, but the northeast trend is, to some extent, transferred into these areas. Excluding 10%, 20% and 30% of the data points in a random fashion tested the model and the trend of the preserved topography was consistent.
Figure 7b shows an isopach map for the Weissliegend facies identified in available core photographs and in core-correlated composite logs (scale 1:500, 131 wells). The Weissliegend facies is very thin in the northeastern parts of the study area with an increasing thickness towards the south and west. The thickness varies greatly across the whole central and western areas to a maximum of 26 m (Fig. 7b). In the central parts a vague northeastern isopach trend is seen.
Figure 7c shows an isopach map of the UAU without the Weissliegend. The thicker ridges are still small in the northeastern corner and in the centre they reach heights of 75 m. This isopach map of the remnant aeolian topography shows similar elongations to the UAU, i.e. towards the northeast, and the trend is slightly clearer. This isopach map was tested in a similar fashion to the UAU map and the elongation of the preserved dune ridges could be mapped out more accurately. The highest topography was interpolated between the areas in a northeast trend (as indicated by the bulls-eyes) and highlighted. The highlighted topography of this map has later been used as a comparison with the surrounding stratigraphy map in the following sections.
The thickness of the Kupferschiefer was collected from interpreted wireline logs (scale 1:500, 133 wells). The Kupferschiefer is generally no more than 3.5 m thick within the study area (Fig. 8b), though it is thicker in the southwestern corner where it reaches a thickness up to 8 m. In the northeastern parts it is almost missing. North of the very thick Kupferschiefer in the south, the isopachs are elongated towards the northeast. This great thickness in the southwest may include overlying Zechstein Kalk shelf facies – muddy carbonaceous shales that could have been wrongly interpreted as Kupferschiefer facies.
The Zechstein Kalk thickness was collected from interpreted wireline logs similar to the Kupferschiefer (scale 1:500, 133 wells). The Zechstein Kalk ranges in thickness from less than a metre up to 40 m within the study area (Fig. 8c). The unit is very thick in the southwest and in the east and is thin in the central and northern parts. A weak lineation of the isopachs towards the northeast is also seen.
The very thick units in the southwest might be due to the nature of the unit as it forms as a carbonate shelf facies. The very thick isopachs towards the east might be a reef development on the flanks of the Inde High.
The thickness of the total Rotliegend was collected from interpreted wireline logs and includes the stratigraphy from the top of the Carboniferous to the base of the Kupferschiefer. In the produced isopach map, the total Rotliegend is relatively thin (approximately 30 m) in the northeast but reaches up to 280 m thickness in the southwest (Fig. 8a). The isopachs show an elongation towards the northwest. Thin Rotliegend can be seen in a small area in the southwestern parts of the study area as well as in the southeast.
The very thin Rotliegend towards the northeast is probably the influence of the Silverpit Lake as well as the Inde High. A smaller high might also result in the thin area in the southwest. In the southeastern part of the study area fluvial deposits represent the super-surface. If these fluvial sandstones are part of a long-lasting fluvial system the Rotliegend did not reach great thickness because of erosion by flash floods.
The thickness variations observed in the UAU have important implications for understanding the Rotliegend aeolian system and the distribution of the Weissliegend. The thickness variations within the UAU may be due to a number of different mechanisms: differential subsidence; inherited topography at the base of the interval; preserved aeolian topography, modified by the Zechstein transgression.
The map for the overall thickness of the total Rotliegend interval (Fig. 8a) shows a series of isopachs that run broadly NW–SE, parallel to the main structural trend of the Sole Pit. This implies that either deposition of the Rotliegend was in an active fault-controlled basin (cf. George & Berry 1997) or a significant topographic low that followed the structural grain existed prior to the deposition of the Rotliegend (Howell & Mountney 1997). In either case, the overall thickness trend cannot account for the thickness changes observed in the UAU.
The base of the Rotliegend shows a clear, irregular, erosional topography, which is mainly infilled by fluvial deposits (George & Berry 1993; Howell & Mountney 1997). Similar thickness changes were observed in the Etjo sandstones of Namibia (Mountney et al. 1999). However, the base of the UAU is a regional deflation surface (super-surface), which is parallel to a series of surfaces within the middle of the Rotliegend and the differential fill of an underlying topography is not considered a valid mechanism for creating the thickness trends in the UAU. Consequently it is concluded that they result from a preserved topography at the top of the Rotliegend. Such a conclusion is supported by Figures 7b and 8b, c, which show the thickness of the stratigraphic intervals that overlie the Rotliegend. The trends of the thickest parts of the UAU (without the Weissliegend; Fig. 7c) are also shown. From these maps it is evident that the overlying units thicken were the UAU is thinnest, i.e. into the palaeo-topographic lows.
The thickness trends observed in the UAU are interpreted as a record of dune topography preserved by the flooding of the Zechstein Sea. This topography was partially modified by the reworking related to the formation of the Weissliegend. This topography is an important insight into the bedforms present at the time of flooding. Furthermore, the interaction between the topography and the processes forming the Weissliegend has implications for predicting Weissliegend distribution.
Since the thickness trends in the UAU are interpreted as preserved dune topography it is important to consider them further. From the UAU thickness map (Fig. 7a), the south-central part of the area is interpreted as a part of the main Rotliegend erg, which passed into an erg margin setting towards the north. The thicker packages are interpreted as the dunes and the crests of these trended broadly NNE–SSW. The presence of the UAU unit everywhere indicates that the Rotliegend erg prior to transgression was a positive topographic feature that covered the whole of the study area; i.e. that is, the bedforms were not isolated features migrating across the super-surface. The rare occurrence of damp interdune deposits indicates it was a dry aeolian system (sensuKocurek & Havholm 1993). Studies of dip-meter data (the longest arms on the rose diagrams indicate the maximum amount of readings) from the UAU (i.e. including the Weissliegend; 22 wells) (Fig. 9) show that the majority of avalanche faces dipped towards the west, perpendicular to the mapped crest of the bedforms. Wells rarely show scattered dip directions.
The preserved dune topography on the isopach maps of the study area and the dip-meter data initially indicate transverse aeolian bedforms, similar to those proposed by Glennie (1983b) and supported by Howell (1992), migrating towards the west. However, transverse bedforms in modern deserts are generally only tens of metres in height with typical wavelengths of several hundred metres. Consequently, the height and spacing of the bedforms, coupled with the complexity of the isopach pattern is evidence of a more complicated history. These two issues: (i) the scale of the bedforms and (ii) the orientation and migration direction, must be considered in the light of modern aeolian systems.
Scale of the bedforms
Whilst the scale of the observed bedforms is greater than examples of modern transverse dunes, modern linear dunes reach great heights (more than 200 m; Holm 1960) and have wavelengths similar to those observed in the UAU. Complex linear dunes carrying star dunes on their top can reach wavelengths up to 7 km in Grand Erg Oriental, Algeria (Breed & Grow 1979; Fryberger 1979). Near Hassi Messaoud, Grand Erg Oriental (Fryberger 1979), chains of star dunes form on top of linear sand features (close to the area on Fig. 11). Here the effective wind regime is either bimodal or complex but seems to have a resultant steady drift direction. Breed & Grow (1979) also noted another analogue in the region of Erg Irarrarene near Brodj Omar Driss (Fort Flatters), Algeria (Breed et al. 1979). Here star dunes developed on top of barchanoid dunes. This would have created a dominant palaeocurrent pattern similar to the Rotliegend dunes. However, some modern dunes developed during long periods and in varying wind regimes through time and their scale and spacing may not be indicative of their original depositional environment. Varying wind regimes before the Zechstein Sea transgression may also have altered the Rotliegend dunes and created a more complex sedimentary pattern.
Alternatively, the extremely wide dune spacing (up to 10 km) may also have been produced if water table rise preceding the Zechstein Sea transgression produced a rise in ground water level, stabilized the substrate and reduced the sediment input. In this case the active dunes would have started to bypass sediment, widening the space between them, but would have also implicated a slow transgression (which is not the case). Evidence for a wetter substrate is seen in the onshore Yellow Sands (Durham, UK) where northeasterly elongated linear dunes are interpreted to result from a southerly directed wind regime (Steele 1983; Chrintz & Clemmensen 1993). The dunes in Durham are overlain by sand-sheet facies and may have reached equilibrium to become wholly sand-passing (Steele 1983), probably due to the early phase of marine transgression that resulted in a decrease in sand supply (Chrintz & Clemmensen 1993). However, such an interpretation is not favoured for the Southern North Sea as the thickness of the UAU is never zero between the bedforms, indicating active and continued dune migration up to the drowning event.
Slip-face and dune crest orientation
There is no direct evidence that proves that the dunes in the study area are transverse. It is known from previous studies that unimodal dip directions (as found in transverse dunes) may also be produced by linear dunes such as in the Namib Desert (e.g. Glennie 1970; Besler 1975; Rubin & Hunter 1985; Bristow et al. 2000). In a linear dune from the Namib Desert, Bristow et al. (2000) found that the initial phase of sedimentation in linear dunes and the lack of sinuous elements on top of these dunes will give a preferred foreset orientation, i.e. a transverse sedimentary pattern. Additionally, Rubin & Hunter (1985) proposed that linear dunes accumulated by lateral migration. Similarly, Hesp et al. (1989) found evidence of laterally migrating longitudinal dunes in northwest China.
Instead of interpreting the dunes as simple transverse or linear forms, a more complex story is probably appropriate. The South Hewett fault scarp and the London Brabant platform might have altered the local wind pattern to become more diverse, but still with dominant bedform migration towards the west. Furthermore, the climate may have changed prior to the Zechstein transgression and compound dunes generated a more scattered dip pattern noted in the south and west. Glennie (1983a) proposed that the Rotliegend bedforms may be complex, linear dunes with superimposed transverse bedforms and George & Berry (1993) postulated the presence of ‘sand storing’ star dunes towards the centre of the erg. The results of this work favour a complicated history for the UAU.
Because of the scale and spacing of the reconstructed bedforms, much of the observed topography is a product of linear dunes, orientated NNE (Fig. 10). These dunes carry, superimposed upon them, smaller transverse forms, migrating toward the west. Linear dunes may be orientated up to 20° to 30° off the predominant wind direction (Lancaster 1982). This implies that the resultant dip direction may not strongly influence the dune crest elongation, and would imply winds from either the NE or ESE. However, the dunes may have been reworked and changed depositional pattern due to climate change (interglacials/glacials). An ENE palaeowind direction is consistent with reconstructions of the trade wind patterns (Glennie 1983a). The presence of more ‘bulls-eyes’ towards the southern parts of the study area is interpreted as star dunes (Fig. 10). Their occurrence may either result from the interacting topography outside the study area (in the Hewett region) or it may reflect complexity in the centre of the main Rotliegend erg. As noted earlier in this paper, compound dunes (with star dunes on top) of Erg Irarrarene and near Hassi Messaoud in the Grand Erg Oriental (Algeria) can form on a scale similar to what we see in the Southern North Sea (Fig. 11). Finally, it is important to consider that the apparent spacing of the preserved dune crests on the isopach maps might also be the outcome of the small number of data points. When contouring, artefacts produced by the gridding algorithm (in this case kriging) will have an affect on the final isopach map. This might partly have caused the vast distance between the interpreted dune highs (8 to 10 km), but also the width of the dune crests (3 to 4 km). Modern dune ridges rarely have a width greater than 2 km.
Distribution of the Weissliegend
The presence of a preserved topography in the UAU (at the top of the Rotliegend), has important implications for the origin and distribution of the Weissliegend. Understanding the mechanisms and controls by which the Weissliegend forms, has implications for predicting thickness variations in areas with no data. When the Weissliegend isopach map was compared with the map of the UAU (Fig. 7b), the Weissliegend is generally thicker in the areas interpreted as topographic lows in the centre of the study area. In the northeastern parts of the study area, where the preserved topography is only a few metres high, the Weissliegend facies is generally thin. Furthermore, the thickest Weissliegend facies does not always occur in the middle of the interpreted interdunes but slightly off the dune crests. The sand could not always travel all the way down and halted due to its well-sorted nature. These factors suggest that the Weissliegend formed mainly by mass flows that transported material from dune crests into interdune lows (Fig. 12). Additional support comes from the sharp base seen in cores (this study & from Howell 1992), the observed vague, flat and fine-grained lamination as well as the absence of marine reworking.
Glennie & Buller (1983) proposed air-entrapment to have been the dominant deformation mechanism. In this study, the occurrence of minor in-situ soft-sediment deformed structures, formed by the same type of process, is seen in Weissliegend facies with diffuse base and occasionally observed escape features. In the cores studied, a diffuse contact to the underlying aeolian unit was mainly associated with a homogeneous Weissliegend facies that subsequently became vaguely flat laminated because it was transported.
Deformation may also have been triggered by other mechanisms (e.g. Doe & Dott 1980; Horowitz 1982; Van Loon & Brodzikowski 1987). Van Loon & Brodzikowski (1987) summarized processes that cause deformation: bioturbation, water and gas migration, temperature changes, glacial activity, atmospheric activity, sudden loading, compaction and mass movement. Also, forces such as faulting, cyclic loading from storm waves and earthquakes and changes in water-table level may cause liquefaction.
First we can conclude that the super-surface is mappable across the study area. Isopach mapping of the interval between this surface and the top of the Rotliegend reveals dune topography preserved after the Zechstein transgression. The revealed topography, the Upper Aeolian Unit UAU), is up to 85 m thick and comprises both the Weissliegend facies and the uppermost aeolian unit (Fig. 7a). On the map, northeast-trending crests with a spacing of 10 km are identified. The preserved draas/dunes are small in the northeastern corner but towards the centre and SE they reach great heights (Figs 7a, c).
The differential isopach units seen in UAU are not related to inherited topography or subsidence, as seen on the map of the total Rotliegend thickness. This imitates the structural trend in a northwestern direction (Figs 3, 8a). The interpretation of the UAU topography is supported by the Kupferschiefer and the Lower Zechstein, which reflect the topography by thickening into the interpreted topographic lows.
The dune crest trends revealed from the isopach of the UAU are perpendicular to the dominating west-northwestern dip direction. The initial interpretation of these data was transverse dunes as noted by previous workers (e.g. Glennie 1983a; Howell 1992); however the scale of the bedforms (i.e. the height together with the dune spacing) in this study is too large to be of that type if related to modern examples. Consequently a more complicated interpretation is favoured. Bedforms were initially linear and produced the large topographic features observed on the isopach maps. These dunes were subsequently modified and overlain by smaller transverse forms. In the centre of the main erg, in the southern parts of the study area, compound star forms also existed.
The Weissliegend facies can be predicted to be thicker within topographic lows of the Upper Aeolian Unit that controlled its distribution. The facies reaches a thickness of 26 m. Sharp erosive contacts with the underlying aeolian strata as well as the massive to flat-laminated nature of the facies indicate the Weissliegend to have formed by mass flows. In-situ homogenization processes developed local diffuse contacts and massive beds that are thinner than the mass-flow deposits. Similar beds observed in core grade into flat-laminated facies higher up in the succession. Together, these processes dominate the formation of the Weissliegend facies and both are triggered by liquefaction.
The authors would like to thank BP Amoco for funding this research and supplying data, as well as their permission to publish this work. K. W. Glennie and Jim Marshall are thanked for their constructive comments on this paper as a part of the thesis ‘A sub-surface evaluation of the Weissligend facies in the UK, Southern North Sea’ (Anna Strömbäck, University of Liverpool, 2001). Gerhard Bachmann and Reinhard Gaupp are also thanked for their useful comments when reviewing the paper.