Alan M. Roberts

Lower-crustal intrusion on the North Atlantic continental margin

When continents break apart, the rifting is sometimes accompanied by the production of large volumes of molten rock. The total melt volume, however, is uncertain, because only part of it has erupted at the surface. Furthermore, the cause of the magmatism is still disputed—specifically, whether or not it is due to increased mantle temperatures. We recorded deep-penetration normal-incidence and wide-angle seismic profiles across the Faroe and Hatton Bank volcanic margins in the northeast Atlantic. Here we show that near the Faroe Islands, for every 1 km along strike, 360–400 km3 of basalt is extruded, while 540–600 km3 is intruded into the continent–ocean transition. We find that lower-crustal intrusions are focused mainly into a narrow zone ∼50 km wide on the transition, although extruded basalts flow more than 100 km from the rift. Seismic profiles show that the melt is intruded into the lower crust as sills, which cross-cut the continental fabric, rather than as an ‘underplate’ of 100 per cent melt, as has often been assumed. Evidence from the measured seismic velocities and from igneous thicknesses are consistent with the dominant control on melt production being increased mantle temperatures, with no requirement for either significant active small-scale mantle convection under the rift or the presence of fertile mantle at the time of continental break-up, as has previously been suggested for the North Atlantic Ocean.

The geometry of normal faults

This meeting took place at Burlington House on 14 and 15 June 1989. Until recently a specialist structural geology meeting on a theme such as this would have been solely the preserve of the Tectonic Studies Group. There is, however, currently much encouraging collaboration between industry and academia on research into all aspects of extensional faulting and thus this meeting was held under the joint aegis of the Petroleum Group and Tectonic Studies Group. The hope that such a meeting would draw participation from industry and academia was fulfilled, with the number of talks presented and the overall attendance dividing roughly equally between these two groups of Earth scientists. The programme included 31 papers and a number of poster displays; nearly 200 people attended. Ultimately the proceedings of the meeting will, it k hoped, appear as a Special Publication of the Geological Society. The intention of this report is therefore not to review each contribution in turn, but rather to synthesize the main themes and conclusions of the meeting, in order that they should be disseminated into the Earth sciences community as rapidly as possible and prior to any future volume. The recent involvement of industry in published basin analysis studies has resulted in an increase in the availability of seismic reflection data to all Earth scientists. A major emphasis at the meeting was therefore placed on what can be learnt about normal fault geometry from seismic reflection data. This main theme was amply supported by papers based on field

Quantitative analysis of Triassic extension in the northern Viking Graben

It has been recognized for over a decade that large-displacement, pre-Jurassic faults are present in the Northern Viking Graben, part of the North Sea rift. These faults define a series of major fault-blocks below the more obvious Jurassic rift basin. We attempt here to quantify the amount of extension associated with this older rift event, which is probably of early Triassic age. Quantitative modelling of the Triassic rift and the succeeding period of thermal subsidence has been undertaken, using a combination of flexural backstripping and flexural-cantilever forward modelling. These techniques suggest that Triassic extension across the Horda Platform (Norwegian sector) reached c. 40% (ß = 1.4). The consequences of this extension were deposition of a thick (>3 km) Triassic-Upper Jurassic syn-rift and post-rift sequence, prior to renewed, but minor, extension in the Late Jurassic-earliest Cretaceous. The thickness of the Viking Group reservoirs in the Troll area appears to have been almost entirely controlled by sediment loading during post-Triassic thermal subsidence. Jurassic extension on the Horda Platform was <5%, an order of magnitude less than the Triassic event. The Horda Platform is therefore principally an area of Triassic extension marginal to the main Jurassic rift further west. In the UK sector of the Viking Graben, Triassic structures are less obvious than those below the Horda Platform, because of Jurassic overprinting. They are, however, still present. Average Triassic extension across the East Shetland Basin was c. 15%, comparable with the magnitude of Jurassic extension in the same area. We believe that the Tern/Eider and Cormorant fault-blocks, with proven shallow basement, comprised a large eroded horst during the early Triassic, uplifted in the footwalls of major faults flanking the Magnus and Statfjord half-graben. During the Triassic, the Magnus half-graben was contiguous with the Unst Basin, now situated in the western footwall of the younger Jurassic basin. The presence of the Unst Basin suggests that Triassic extension occurred across the area that is now the northern Shetland Platform, continuing into the West Shetland area. Although the more obvious structures in the Viking Graben are Jurassic in age, the earlier Triassic event was equally as important in controlling the structural and stratigraphic history of the basin.

Mesozoic extension in the North Sea: constraints from flexural backstripping, forward modelling and fault populations

The magnitude and distribuion of late Jurassic extension in the Northern Viking Graben has been investigated by (i) syn-rift forward modelling using the flexural-cantilever model of continental rifting; (ii) post-rift flexural backstripping of a series of cross-sections; and (iii) the analysis of fault-population statistics. Application of these three techniques indicates that Jurassic extension on the tilted fault-block terrains of the East Shetland Basin, Tampen Spur and western Horda Platform is on average c. 15% ( β = 1.15). In the graben axis the regional value of β rises to c. 1.3, perhaps locally rising to 1.4. On the eastern Horda Platform Jurassic extension is low, β = c. 1.05. Flexural backstripping of the post-rift part of a cross-section through the Central Graben yields similar estimates of extension. At the flanks of the Jurassic basin β is estimated to be 1.2, rising to a likely maximum of c. 1.3 in the basin centre. These estimates of extension lie at, or towards, the low end of the previously published range of estimates. The principal reasons for this are (i) the incorporation of the thick sequence of Triassic and Lower Jurassic sediments in the backstripping; and (ii) the use of flexural isostasy (rather than Airy isostasy) in both the backstripping and forward modelling. The estimates of Jurassic extension obtained in this study do not account for the observed crustal thinning and, therefore, point to there having been a significant pre-Jurassic extensional event in the North Sea, of probable Triassic age. While a residual thermal anomaly from this event may have made a small contribution to post-Jurassic subsidence, compaction of the Triassic-Middle Jurassic sequence has been a significant contribution of the Triassic event to the thickness of the Cretaceous/Tertiary basin.

Forward and reverse modelling of rift basin formation

Abstract The McKenzie model of continental lithosphere extension describes the first-order lithosphere responses of crustal thinning and geothermal gradient increase following rifting, which lead to syn- and post-rift basin subsidence. At the sub-basin scale, seismic data show the fundamental importance of major basement faults in controlling the geometry and subsidence of rifted sedimentary basins. Reflection and earthquake seismology data show that these faults are generally planar and extend down 10–15 km to mid-crustal level. Below this depth deformation gives way to distributed ductile shear in the lower crust and mantle. Extensional faulting leads not only to hanging wall subsidence but also to footwall uplift. It can be successfully modelled using elastic dislocation, visco-elastic finite element and thin plate flexural-isostasy (flexural cantilever) theories. For extension on multiple faults, interference of footwall uplift and hanging wall collapse gives rise to the familiar block rotations of rift tectonics. Mathematical forward models of rifting, incorporating upper crustal faulting, lower crust/upper mantle plastic stretching, lithosphere thermal perturbation and re-equilibration, sediment loading and flexural isostasy have been developed and applied to both syn- and post-rift stages of basin evolution. The models allow the effects on basin geometry and subsidence of fault spacing, fault polarity and extension magnitude, as well as sediment fill and footwall erosion, to be explored. Reverse post-rift modelling from present day sections may be used to constrain β stretching estimates and to predict palaeobathymetry and topography. Both forward and reverse models have been successfully applied to many rift basins worldwide, including the Viking Graben, North Sea and the East African Rift System, both of which are documented here. The forward model is also capable of simulating the first-order geometry of metamorphic core-complexes and so-called ‘detachment faults’ documented from the Basin-and-Range Province of the western USA.

Subsidence of the Vøring Basin and the influence of the Atlantic continental margin

The post-Cretaceous subsidence history of the Vøring Basin, part of the Atlantic passive margin offshore mid-Norway, has been investigated. Extension and β -factors related to rifting and continental break-up during the Palaeocene have been quantified using both forward and reverse basin-modelling techniques. In the preferred geological model it is assumed that rifting occurred in the Vøring Basin during the Palaeocene (prior to break-up), following an earlier rift event during the Late Jurassic. During Palaeocene rifting the basin may have been dynamically uplifted by the Iceland mantle plume. In the east of the basin there was no Palaeocene extension. Subsidence analysis shows that in the centre of the basin forward and reverse models converge to predict a modest Palaeocene stretching factor (β) of c. 1.15. In the west of the basin, closest to the Atlantic margin, forward models of upper-crustal faulting also predict a β of c. 1.15, but reverse (backstripped) models of subsidence predict a β of up to 1.75. We suggest that lower-crustal and mantle-lithosphere thinning close to the margin were greater than the extension accommodated by upper-crustal faulting and that some lower-crustal/mantle-lithosphere stretching associated with continental separation was partitioned below the Vøring Basin, up to 150 km landwards of the margin.

The structural evolution of the Brent Province

Abstract The structural evolution of the Viking Graben has been the fundamental control on the deposition of the Brent Group and on the development of trapping geometries. Major crustal extension in the early Triassic caused tilting of basement fault blocks, which can still be clearly seen at the basin margins. By the mid-Triassic a post-rift thermal-subsidence basin was established. Local fault control on Brent Group deposition indicates the onset of a second period of crustal extension, although the main control on subsidence at this time was still thermal relaxation following Triassic rifting. Extension during deposition of the Brent was of only c. 1% magnitude. The second period of extension increased in magnitude following Brent deposition and peaked during the late Jurassic, creating the main structural traps for Brent Province oil and gas. Footwalls to major normal faults were uplifted and eroded. The amount of uplift on a given fault-block can be predicted using quantitative models (flexural and domino models). Subsidence in the adjacent half-graben generally outpaced sedimentation, leading to deep-water basins into which footwall material could collapse. On all but the largest fault-blocks it is likely that footwall uplift rates were low compared with erosion rates, and so footwall crests would have been degraded faster than they would have been uplifted above sea-level. Flushing by meteoric water during the late Jurassic is therefore expected only on the largest fault-blocks (e.g. Snorre, Gullfaks).

Deformation around basin-margin faults in the North Sea/mid-Norway rift

Abstract It has been demonstrated previously that the rigid-domino model of extensional faulting can account well for the uplift/subsidence patterns observed within particular areas of the North Sea rift. This model, however, provides an unsatisfactory solution to deformation occurring at the basin margins. It is suggested here that the domino model is an elegant, geometric simplification of the more complete flexural-isostatic solution to fault displacements. Application of a flexural model allows a unified treatment of the basin and its margins. The basin-margins to the North Sea/Mid-Norway rift are all shown to have responded to extensional faulting by experiencing isostatic uplift in the footwalls to the marginal faults. Uplift magnitude varies from a few hundred metres adjacent to small faults to perhaps 5 km adjacent to the largest faults. Uplift patterns can be modelled or predicted by use of the flexural-cantilever basin model. Recognition of marginal uplift throughout the rift means that geometric section balancing techniques, all of which require the footwall to be rigid during extension, are inapplicable as a method for analysis of large, basement faults within the North Sea. Marginal, fault-related uplift is considered to have been a primary source of syn-rift clastic detritus. The precise locus of deposition for material eroded from emergent basin margins will depend on local drainage patterns, but deposition in the hangingwall basin and on the footwall platform may both be anticipated.

Late Jurassic half-graben control on the siting and structure of hydrocarbon accumulations: UK/Norwegian Central Graben

Abstract The late Jurassic structure of the Central Graben is discussed, both in terms of the extensional features developed at this time and their subsequent controls on younger inversion structures. The regional kinematics of the North Sea rift suggest that during the late Jurassic the Central Graben opened approximately orthogonally to the main NNW—SSE fault-trend within the basin. Some evidence that Triassic extensional faults were reactivated at this time is seen towards the basin margins, and a significant Triassic extensional history is inferred. The late Jurassic basin comprises the axial c. 100 km of the Central Graben (s.l.). The basin axis is offset progressively to the SE by both transfer zones and discrete transfer faults. Footwall uplift of the basin margins during extension provided the source for the Upper Jurassic reservoir sands encountered by exploration drilling. These sands were probably preferentially fed into the basin at transfer zones in the margins. Salt structures are ubiquitous within the Central Graben, and their siting is everywhere controlled by the underlying fault-block topography. Salt structures form preferentially towards the crests of half-graben or buttressed against their boundary faults. Inversion structures in the Norwegian sector reactivated, with a reverse sense of displacement, late Jurassic extensional faults. A model is proposed which shows the two main inversion structures (Feda Graben, Norway; Tail End Graben, Denmark) to be detached on antithetic reverse faults riding in the salt. These faults terminate at tip-folds within major salt walls. The geometry of the inversion structures can be understood only when considered together with the geometry of the underlying late Jurassic half-graben.

Sedimentation and subsidence in the South Caspian Basin, Azerbaijan

Abstract The South Caspian Basin is believed to contain more than 20 km of Mesozoic and Tertiary sediments deposited on oceanic or thinned continental crust. Mesozoic, Palaeogene and Oligo-Miocene sediments have not been penetrated within the South Caspian Basin itself but are exposed onshore in the basin margins. The Pliocene–Recent sequence has been mapped on a regionally extensive grid of two-dimensional (2D) seismic data and penetrated by recently drilled exploration wells, and is over 7 km thick. Most of this sequence (6 km) is formed of fluvial–lacustrine deltaic sediments of the Pliocene Productive Series that are deposited unconformably above a marine Miocene shale sequence and form the principal hydrocarbon reservoirs in the basin. The Productive Series is overlain by about 1 km of Late Pliocene–Recent marine sediments The thickness of the Pliocene sedimentary sequence implies that relatively rapid, late Tertiary subsidence occurred in the South Caspian Basin; however, there is no geological evidence of a tectonic event capable of generating a major thermal subsidence event at this time. Modelling presented in this paper suggests that it is possible to account for the observed pattern of subsidence and sedimentation in the South Caspian Basin by a process of sediment loading and compaction on a thermally subsiding, late Mesozoic crust without the need for additional Tertiary subsidence mechanisms. Crucially, this model interprets the Pliocene Productive Series to have been deposited in a topographic depression, isolated from the global oceanic system, in which base level was controlled by local factors rather than by global sea level.