Ken Thomson writes: The suggestion that the Silverpit Crater is a pull-apart basin (Smith 2004) and not an impact structure (Stewart & Allen 2002) demonstrates that its origin requires further investigation. As noted by Underhill (2004) the coincidence of the structure with a salt withdrawal basin suggests that halokinesis may provide an alternative mechanism. Although the model provided by Smith involves salt movement it is overly complex as strike-slip detachment tectonics are not required. Instead gravitational collapse of the post-salt cover provides a simpler mechanism.
The Southern North Sea is dominated by the development of large linear salt swells (Jenyon 1988; Stewart & Coward 1995). Elevation and tilting of the swell flanks results in gravitational instability and a range of deformation styles can develop (Jenyon 1988; Stewart & Coward 1995; Thomson 2004). The instability commonly leads to listric faulting of the post-salt cover down slope, towards the adjacent salt low (Jenyon 1988; Stewart & Coward 1995; Thomson 2004). Detachment levels for the listric faults occur at a number of stratigraphic levels. The top of the Upper Permian Zechstein Group salt, which forms the swells, provides a detachment at the base of the post-salt cover whilst detachment within the post-salt cover in both the Triassic Röt Halite and within the Upper Cretaceous chalk have also been recognized (Jenyon 1988; Stewart & Coward 1995; Thomson 2004). The majority of the flank collapse faults dip in the same direction as the cover rocks they displace (i.e. towards the salt withdrawal basin and away from the swell), but occasionally listric faults can be found which dip away from the salt withdrawal basin, towards the swell (Jenyon 1988). Associated with the listric faulting rollover anticlines can develop in the hanging walls (Stewart & Coward 1995) whilst slump folds have also been recognized on swell flanks (Jenyon 1988).
Analogue modelling suggests that as salt withdrawal progresses polygonal subsiding lows surrounded by salt walls should develop (Talbot et al. 1991). Associated with the development of salt walls and isolated non-linear salt lows gravitational instability could be expected to result in concentric faulting around both diapirs and salt lows. Examples of concentric faulting as a result of salt withdrawal are rare in the literature, probably a result of the small displacements involved and the steep flanks on which they are situated making them difficult to resolve. Maione (2001) recently demonstrated that closely spaced ring faults form the periphery of two salt withdrawal basins associated with the La Rue diapir, Texas. The displacement on the faults, both normal and reverse, is small and was only clearly imaged using coherence attributes. Coherence is essentially a cube of correlation coefficients generated from the 3D seismic volume and highlights faults and other stratigraphic anomalies on seismic sections and timeslices (Baharich & Farmer 1995). Similarly, a recent study using a large 3D seismic dataset from the Southern North Sea has demonstrated that concentric faulting in the Silver Pit region is common (Thomson 2004). The Silverpit Crater is situated at the northwestern end of a synclinal salt withdrawal basin and it has been demonstrated that a similar, though less well-developed structure, can be found at the southeastern end of the syncline (Thomson 2004). Consisting of a salt low surrounded by concentric faults that displace the top chalk towards the low, the similarity with the Silverpit structure would suggest a common origin. Furthermore, concentric faults surrounding localized structural highs can also be found in the region. To the south and west of the Silverpit Crater two salt highs surrounded by concentric normal faults that dip towards the surrounding lows can be observed (Thomson 2004). Such features can be interpreted as concentric equivalents to the uppermost collapse faults seen on linear swells. However, it is the temporal relationship between the concentric faults surrounding the highs and those forming the Silverpit Crater that is crucial to understanding the origin of the crater. Where the concentric fault sets overlap the displacements on both sets are significantly reduced (Thomson 2004). This suggests that the fault sets were contemporaneous features and mutually crosscutting. As the concentric faults surrounding the high are related to swell/early diapir inflation the contemporaneous nature of the faults forming the Silverpit Crater would suggest that they were formed in response to salt withdrawal from the low (crater) to supply additional salt to the adjacent inflating high. A similar structural relationship at the southeastern end of the salt withdrawal syncline also exists (Thomson 2004).
Evidence for additional concentric structures within the Southern North Sea, the lack of evidence for shock metamorphism from wells that penetrate the Silverpit Crater (Stewart & Allen 2002) and the coincidence between the locations of concentric collapse structures and zones of evaporite withdrawal not only suggests that an impact origin for the Silverpit Crater (Stewart & Allen 2002; Stewart 2003) is unlikely but also casts doubt on the pull-apart model (Smith 2004). Instead concentric, salt withdrawal related collapse contemporaneous with adjacent inflation represents a feasible alternative. The only inconsistency between the collapse model and the observations of Stewart & Allen (2002) is the central peak observed at the Base Cretaceous level. However, evidence for the presence of the central peak is not unequivocal as all of the post-Permian stratigraphic horizons reach their greatest depths in the core of the syncline, leaving open the possibility that the peak is an artefact of the seismic processing, velocity modelling or the method of interpretation (Underhill 2004).
2 June 2004
Peter Owen writes: In a recent paper, Smith (2004) attempts to give a mundane, rather than extraterrestrial explanation, for a feature mapped on 3D seismic data in the southern North Sea. This feature, termed the Silverpit Crater, has earlier been interpreted as a meteor impact structure (Stewart & Allen 2002), without a shred of scientific justification. While I totally agree with Smith's suggestion that there is no need to call for an extra-terrestrial explanation for the Silverpit structure, Smith's alternative explanation requires some discussion/attention. In particular, Smith's proposal that ‘Cenozoic dextral movement along a right-stepping strike-slip fault creates a rhomb-graben in pre-Zechstein basement’ (fig. 7), causing Zechstein salt withdrawal and associated ring-fracturing in the Upper Cretaceous is questioned. An apparent diapir of Liassic material also forms part of the hypothesis. There are several problems introduced by such a one-stage explanation.
(1) In Smith's interpreted seismic section (fig. 3), it is not clear what has happened to the Upper Cretaceous rock apparently replaced by ‘diapiric’ Lias. Although extension of the Upper Cretaceous could be invoked to create the listric faults depicted representationally on the seismic section (fig. 3), the mechanism of introducing the ‘diapiric’ Jurassic shales in the same phase of extension is not obvious. There is also a problem explaining the absence of the lowest parts of the Cretaceous in the centre of the structure. This is highlighted by Smith's diagrammatic geological section (fig. 8), where the Upper Cretaceous cannot be restored to a balanced cross-section, as should be possible in an apparently radially symmetric structure. Continued growth of a shale diapir throughout the Upper Cretaceous, creating local thinning of the Chalk, (and hence apparently ‘missing’ Upper Cretaceous rock) is conceivable (but see next section), though not part of Smith's argument.
(2) A pre-condition for shale becoming diapiric is that it has not de-watered and has become over-pressured: conditions normally associated with rapid burial. In Smith's model, the Lias is required to have mobilized subsequent to the removal of its original cover by pre-Cretaceous (Cimmerian) erosion, and the later deposition of 500 m. of Cretaceous rock. It is barely credible that the Lias shales in the Silverpit area had preserved an underconsolidated condition at the end of the Cretaceous. Also, there is no indication of a rim syncline of early Tertiary date around the feature (fig. 3), as might be expected.
(3) The schematics of the ‘crater's’ development (figs 7 & 8) bear little relation to the seismic section (fig. 3), assuming the same orientation. ‘Wedging’ of the Lias, under the centre of feature, is not shown. A seismic section published by Underhill (2004, fig 1a) shows this thinning is due to late Jurassic movement of the salt swell to the SW of the ‘Silverpit Crater’, and erosion of the upper part of the Jurassic section.
Additionally, the V-shaped geometry of the pre-Cretaceous units in the Palaeocene (fig. 8) cannot be inferred from the seismic section. The geometry of the top Zechstein and the intra-Neogene events (fig. 3) strongly suggests that post-Cretaceous salt-withdrawal from the central part of the Silverpit ‘crater’ occurred in a single, late Neogene phase.
(4) There is little sign of significant extension, within the Triassic, on the seismic section (fig. 3). A slight time-thinning of that interval, below the centre of the feature, is probably no more than a velocity effect of the deeper burial. Additionally, the implication that Zechstein salt can transmit significant horizontal stress into the overlying rocks, is at odds with the conventional view, expressed by Smith, that the salt acts as a ‘zone of detachment between basement and cover successions’ (pp. 593–594).
(5) The Trent/ Schooner ridge is shown as a pre-Permian high (fig. 4). It is difficult to embrace the idea that a section of the ridge became a rhomb graben in the Cenozoic, occasioning some salt withdrawal, and then recovered to the slightly inverted position shown on the seismic section (fig 3), with further salt withdrawal. Such reversal of movement on the underlying strike-slip fault implies a dramatic and unlikely alteration of stress regime during the Tertiary, which, unlike the ‘opening’ phase, had no apparent effect on the overlying units.
Smith points out that the coincidence of the location of the ‘crater’ with the centre of the (later) salt-withdrawal syncline (in fact at a bend in the axis of the Upper Cretaceous syncline), and the absence of impact phenomena in a well penetrating the ‘crater’ cast doubt on the impact hypothesis. There are other coincidences with the deeper geological structure: the offset in the Trent and Schooner Ridges (fig. 4) and the erosional ramp in the Jurassic (fig. 3), mentioned above. Another coincidence is that the structure is only partially covered by the 3D seismic data, and therefore even the circular geometry is not established.
Smith's hypothesis addresses some of these coincidences, but relies on an implausible concatenation of special pleading, perhaps because of unnecessary over-elaboration. For instance, the ‘Liassic diapir’ is possibly a processing (overmigration) artefact—Underhill (2004) also hints at this—and requires no geological explanation. However, the high-amplitude reflections in the early Tertiary, above the central part of the ‘crater’ do require explanation. The seismic section presented by Underhill (2004) passing through the centre of the ‘crater’ (Underhill, pers. comm.), in conjunction with those published by Smith (2004) and Stewart & Allen (2002) shows the extent of these high-amplitude reflections is oval in plan, implying that the ‘crater’ has the same geometry. There are also signs of disturbance in the Triassic below the centre of the structure, which, if not processing artifacts, may be a relic of a vanished salt diapir.
Underhill (2004) suggested a salt-withdrawal mechanism, to explain the feature. Salt withdrawal from below the Triassic rocks, per se, is an insufficient explanation, because few such structures have associated ring fractures at the top Chalk level, even though the Silverpit ‘Crater’ is not unique (Underhill 2004) in the area. The explanation of this feature, and others like it, is likely to depend, not on a single incident, but on a combination of past events, which require elucidation.
5 August 2004
Kevin Smith replies: Accounts of British Cenozoic basin formation often seem to be an implausible concatenation of extensional, strike-slip and compressional tectonics. Interpretation of the Silverpit Crater as a meteorite impact structure in the southern North Sea has added another option to the list of Cenozoic basin-forming mechanisms (Stewart & Allen 2002). Since the impact hypothesis was proposed, the observation that the site of the Silverpit Crater coincides with a synclinal axis formed by the withdrawal of underlying Zechstein salt has prompted three different reassessments of the structure (Smith 2004; Thomson 2004; Underhill 2004). From his study of Cenozoic fault patterns in the Silverpit area, Thomson (2004, and in this discussion) concludes that the crater forms part of a spectrum of deformation related to gravitational collapse on the flanks of salt swells. However, evidence from the Cenozoic succession indicates that much of the local salt movement post-dates the formation of the crater (Stewart & Allen 2002; Smith 2004). Also, the Silverpit structure does not consist solely of concentric Cenozoic faults; the identification of a conical uplift at base Cretaceous level beneath the centre of the crater is a key element in the impact interpretation. Recognizing the difficulty of forming this uplift during salt withdrawal, Underhill (2004) doubted the validity of the seismic evidence for the structure. However, it seems unlikely that a seismic artefact should mimic the morphology of real structures in similar tectonic settings elsewhere, such as Upheaval Dome in Utah. Detailed geological mapping, augmented by geophysical investigations, has yet to resolve whether Upheaval Dome is a salt structure or a meteorite impact (Kriens et al. 1999; Stewart 2003). In one halokinetic model of the dome, the central uplift is attributed to structural collapse following the passage and subsequent dissolution of a pinched-off salt diapir (Jackson et al. 1998). In the case of the Silverpit Crater, this explanation is ruled out by the lack of evidence for Zechstein diapirism beneath the structure (Stewart & Allen 2002).
Attempting to provide a purely terrestrial explanation for the central uplift, I have suggested that it could have been caused by reactive diapirism of the Lower Jurassic succession beneath an extending Cretaceous cover. In his peer review, Daniel Schultz-Ela considered that this process was physically unlikely in dewatered shales, and Owen forcefully endorses this view. An alternative proposal, that extending Upper Cretaceous strata converge in a series of radial rotational slumps towards a central compressional uplift, was based on pre-existing models of detached cover successions in the North Sea (Stewart 1996). In addition, a description of the concentric Compton Valence structure in Dorset noted that movement of Triassic salt in a strike-slip regime was implicated in the Cenozoic uplift of the Jurassic succession. If the process of reactive diapirism is not feasible, perhaps it may still be possible to construct a non-impact model of the Silverpit uplift that incorporates some aspects of these other proposals.
Thickness variation in the Lower Jurassic was omitted for reasons of clarity from a restored section across the crater (fig. 8). Owen claims that this section fails to balance and argues that since all sections across the circular structure are the same, any recourse to out-of-the-plane extension can be ruled out. The opposite is actually the case. In a radially symmetrical extending structure, all cross-sections will exhibit the effects of extension outside the plane of the section, and mean stratigraphic thicknesses will be reduced by 1/β2 (the reciprocal of the square of the extension factor calculated from a single section). In reality, it is likely that the pre-Cretaceous succession can be effectively restored by using an extension factor comparable to that estimated from fault heaves (Stewart & Allen 2002), and assuming dips of 45° for the basement faults and 15° for intra-Zechstein detachments. In the shallower part of the interpreted seismic panel (fig. 3), fault planes are depicted diagrammatically simply because they are imaged poorly in the seismic volume and it is not known how they connect at depth. However, a few faults do intersect the base Cretaceous horizon and probably detach upon Triassic halites, as in the Flamborough Tertiary outlier (Stewart & Bailey 1996).
All alternative tectonic models will become irrelevant if firm evidence for an impact origin of the Silverpit Crater is discovered. Some impact-related phenomena associated with the crater have already been reported (D. Jutson in Self-Trail 2003), and others are in the process of analysis (D. Jutson, pers. comm. 2004). If these new findings were confirmed, it would suggest that the observed relationship between salt withdrawal and the Silverpit Crater is a consequence of the impact itself. This possibility was mentioned only briefly in my paper, but has been proposed independently by Stewart & Allen (2004) to explain why the impact is located in the centre of a salt withdrawal syncline.
The recognition of a 2 km diameter elliptical area of salt withdrawal directly beneath the crater firmly establishes a genetic relationship between crater formation and halokinesis (Smith 2004). A revised version of the impact hypothesis must explain how the top of the Zechstein salt has been deformed and displaced at depth beneath the transient crater. If the expanding front of the initial shock wave pushed apart the evaporite layer, perhaps it could be said to have formed a push-apart basin at this level. In this case, the contemporaneous formation of pull-apart basins elsewhere in the UK, such as the Bovey Basin (Smith 2004), would remain as purely coincidental. Nearer the Silverpit Crater, potential pull-apart structures include the eastern part of Flamborough Tertiary inlier (Stewart & Bailey 1996), which forms a deep Cenozoic basin with the characteristic geometry of a rhomb graben, and the Peak Trough (Milsom & Rawson 1989), a transtensional basin partially of Cenozoic age, which straddles the Yorkshire coast.
Confirmation of the impact hypothesis would mean that a terrestrial pull-apart model is not required. However, observations made in the search for an alternative explanation of the fortuitous structural and stratigraphic setting of the Silverpit Crater (Smith 2004) may contribute to a new model of impact-related deformation within evaporite basins. One possible outcome could be a resolution to the long-standing debate about the origin of Upheaval Dome (Jackson et al. 1998; Kriens et al. 1999; Kanbur et al. 2000; Stewart 2003).
Publication is by permission of the Executive Director, British Geological Survey (NERC).
27 August 2004
- © 2005 The Geological Society of London
