|
Discussion |
1 3 Finedon Hall, Finedon NN9 5NG, UK (e-mail: davidjoanjames{at}compuserve.com)
2 Fault Analysis Group, Department of Geology, University College Dublin, Belfield, Dublin 4, Ireland (e-mail: fault{at}fag.ucd.ie)
3 Fault Analysic Group, Liverpool University Marine Laboratory, Port Erin Isle of Man, IM9 6JA, UK
4 Institute of Geological and Nuclear Sciences, PO Box 30-368, Lower Hutt, New Zealand
5 Badley Earth Sciences Ltd, North Beck House, North Beck Lane, Hundleby, Lincs PE23 5NB, UK
Scientific editing by Richard England.
D. M. D. James writes: The paper is much to be welcomed as bringing new and stimulating ideas to the ongoing debate concerning the origin of polygonal faults (
Cartwright & Lonergan 1996; Cartwright & Dewhurst 1998). However, if the proposed mechanism of blind faulting above sites of incipient diapirism is correct, then demonstration is required of significant thickness variation of the causative mobile layer and of a density inversion across its upper boundary. The following remarks aim largely to seek geophysical reassurance on these aspects, and as such are also relevant to many of the increasing number of Journal papers that utilize seismic and well data from the oil and gas industry.
The candidate mobile layer is stated to be a shale, fortuitously preserved at its depositional thickness of 35 m in the only control well apparently available. It is critical for the authors interpretation (fig. 8b) that both top and bottom of this shale are mappable seismically and the character of the sonic log (fig. 1c) gives some encouragement that this may indeed be possible; ideally via an exact tie to the well based on a well shoot, an acoustic impedance profile from well logs and knowledge of the phase and polarity of the seismic. No such tie is mentioned but assuming that the position of the shale is approximately as shown in fig. 3a, rather than its accompanying rather obvious seismic mistie in fig. 3b, the distinctive and rather tram line character at this level makes it very difficult to accept the authors requirement of a range of mobile thicknesses of 0–70 m (pp. 156, 160). Such variation should easily be resolvable (and the extreme thinning indicated by tuning effects) at the likely seismic bandwidth. Seismic lines illustrated as figs 6b & d seem to show the shale locally folded congruently with the underlying sequence rather than generating superincumbent folds by incipient diapirism; they illustrate arguably rather different geometry to the model shown in fig. 8b.
There are very many wells in the Eromanga Basin; do they show significant variation of thickness of the low-velocity layer where polygonal faults are developed and is thickness essentially constant where polygonal faulting is not developed? Does the velocity of this layer alter where polygonal faulting is not developed? As no density (FDC) log or density data from cuttings are given it is impossible to judge the claim that the low velocity shale has low density relative to its immediate overburden: it could for instance have the same density but merely be more compressible and less rigid. From its stratigraphic position this shale could well contain a flooding surface and be of different mineralogy and facies than its immediate overburden in which case it cannot be deemed undercompacted, or overpressured, relative to that overburden. Sediments do not necessarily compact down a porosity-loss curve defined by their overburden. The authors equate low velocity not only with with low density but also with elevated pore fluid pressure and low viscosity: such relationships are not uncommon but are by no means universal, hence the need for independent supporting data such as RFT data from the immediately underlying sand. The reported drilling problems need to be demonstrated as related neither to mud chemistry/shale mineralogy nor to excessive circulation (I presume casing is set at this level) before an overpressure origin is concluded. It is premature to use the higher velocity above the shale to infer a density inversion capable of inducing diapirism: what is needed are FDC data and a plausible line of geological argument that any density inversion seen today is likely to have persisted from the time of polygonal faulting. In any case, density inversion alone is no guarantee of diapirism.
Although not, to myself, adequately yet demonstrated from the Lake Hope data set, the mechanical interpretation is plausible and interesting if the likely strain rates at boundaries between different regimes of ductile extension associated with different fold curvatures are sufficiently large to induce fracture. This seems at least questionable. Moreover, by analogy with load balls and flame structures (see fig. 8a) it would seem likely that ductile deformation above growing anticlinal cusps would ensure that fold curvatures at the diapiric interface would markedly decline upwards, rather than being maintained concentrically, and that fault throw maxima would thus lie not too far above this level, rather than hundreds of metres higher. The Lake Hope faults presumably nucleated at the relatively large ductility contrast at the Coorikiana horizon (where the throws are maximal) and developed along similar growth patterns to others so well documented over the years by the Liverpool Fault Analysis Group. The seismic appears to show that fold amplitude at this level is greater than that at the top of the postulated mobile shale, rather than being equal as required by the diapiric hypothesis although this conclusion is sensitive to the time–depth relationship used. Although begging the issue of the origin of the causative stress field, could not such ‘folding’ be simply that associated with flexure around (in this case commonly conjugate) blind faults near their throw maxima, where bed length change is also greatest, rather than propagated from far below?
27 February 2000
J. J. Walsh, J. Watterson, A. Nicol, P. A. R. Nell & P. G. Bretan reply: We thank James for his comments on the ideas we put forward to explain the intriguing and, in some respects, enigmatic features associated with the Lake Hope polygonal fault system. We recognize the uncertainties concerning the mechanism that has given rise to these and to similar faults elsewhere, which have stimulated an unusually wide variety of interpretations (Henriet et al. 1991; Cartwright 1994; Cartwright & Lonergan 1996, 1997; Watterson et al. 2000). However, so far as was possible we followed our normal practice of giving greater credence to the relative certainties of the geometrical and kinematical data and conclusions, than to an inherently uncertain dynamical interpretation.
We strongly disagree with the suggestion by James that it is critical for our interpretation that both the top and bottom of the low density, and low velocity, shale interval is seismically mappable (see below). It would certainly be a good test of our model if the present boundaries of the low velocity shale interval could be mapped but we do not believe that such mapping is possible with the seismic data now available. The tramline nature of the interval is attributable to seismic effects and both modelling and analysis provides little encouragement for accurate mapping of the interval from seismic data. Seismic modelling performed by the operators (Santos) shows that the main seismic response arises from the large increase in seismic velocity at the Top of the Cadna-owie, i.e. the base of the low velocity interval. This provides a strong trough with side lobe peaks at approximately 20 ms (c. 25 m) above and below the Cadna-owie. The top of the interval has, however, a relatively low acoustic impendance contrast, producing a weak discontinuous peak that is, at best, only mappable from conventional seismics for interval thicknesses greater than c. 35 m. In these circumstances, any tuning effects are insignificant and are not seen in amplitude maps. Although accurate mapping of the thickness of the low velocity interval is not possible, local estimates of the maximum thickness of the interval can be made and are, at c. 65 m, broadly consistent with our predictions from other geometric and kinematic constraints. Given the limitations of the available data, however, we chose to examine thickness variations of intervals that are mappable from seismic data, i.e. intervals that include and are thicker than the low density shale. These thickness variations are compatible with our model (Watterson et al. 2000, fig. 7).
A critical feature of the Lake Hope polygonal fault system is the intersection of conjugate pairs of normal faults, or graben-bounding faults, at or very near to the mapped Cadna-owie horizon that lies at the base of the faulted sequence. We regret that although this geometry is very clear from three of the four sections we present, the ‘obvious seismic mistie’ that James refers to was due to a drafting error that went unnoticed even through the review process: the error resulted in a small bulk shift of all of the interpreted horizons and faults relative to the seismic background. A corrected version of fig. 3 is reproduced here (as Fig. 1) and shows that the graben-bounding faults intersect at the Cadna-owie horizon. The traces of these faults on this horizon define the ordered polygonal pattern that is a primary geometrical characteristic of the fault system. It is our belief that, where intersections of oppositely dipping graben-bounding faults coincide with a particular horizon or stratigraphic interval, the faults must have been initiated at and propagated from that horizon. In the case of the Lake Hope faults this
leads to the further conclusion that the faults propagated only upwards and sideways. Had the faults initiated elsewhere in the sequence and propagated downwards towards the Cadna-owie horizon then it would have required an extraordinary series of coincidences for each of the numerous pairs of faults to have intersected at the same horizon. This kinematic conclusion is the starting point for any subsequent consideration of the forces that could have generated the faults and leads directly to the question of what were the special properties of the sequence at or near the critical horizon? The deduction that the sequence at or adjacent to the Cadna-owie horizon is likely to have distinctive features owes nothing to interpretation of well data but derives directly from the geometry of the faults. The principal uncertainty concerns the nature of those distinctive properties. Attention is further focussed on the properties of the sequence adjacent to the critical Cadna-owie horizon by the relatively ordered polygonal fault pattern at this horizon, because the pattern is characteristic of, although not diagnostic of, structures generated by Rayleigh–Taylor instability. Only at this stage of the interpretation do questions concerning the properties of the sequence associated with the Cadna-owie horizon become relevant and the doubts expressed by James become pertinent.
|
McGill & Stromquist 1979; Cartwright et al. 2000; Moore & Schultz 1999). Here, rapid erosion of the Colorado River in Cataract Canyon has allowed a c. 460 m thick plate of clastic sedimentary rocks to move towards the canyon. The trigger for this movement was the downward erosion of the canyon to expose a free surface to the evaporite sequence underlying the clastic sequence, allowing the latter to slide or extend laterally towards the canyon. The movement of the clastic plate is accommodated by flow within the evaporites and is aided by a 4° regional dip towards the canyon. Deformation within the sliding block of clastic rock takes the form of a series of graben-bounding normal faults that strike parallel to the canyon. In this case the special properties of the sequence underlying the faulted block are known and it is the geometry of the faults which is in question. Because movement of the clastic block is accommodated by flow within the evaporite sequence, the clastic sequence has not been simply translated towards the Canyon but is also laterally extended. This extension is driven by a traction on the base of the clastic block imposed by the flowing evaporite and is accommodated by the several sets of graben-bounding faults. Given this mechanism, each pair of graben-bounding faults would be expected to converge downwards to intersect at or near the clastic–evaporite boundary, as proposed by McGill & Stromquist (1979) and McGill et al. (2000). A similar upward divergence has been shown to occur in analogue experiments (Childs et al. 1993 and references therein). Although the faults can be said to converge downwards geometrically, kinematically each pair of faults would be expected to have propagated upwards from a common point on the clastic–evaporite boundary (McGill et al. 2000). The contrary view that the graben-bounding faults propagated downwards from the surface (Cartwright et al. 2000) is based on the belief that the graben-bounding faults do not intersect at or near to the clastic–evaporite boundary, a belief strongly disputed by McGill et al. (2000). We emphasize that although the downward convergence of several pairs of graben-bounding faults to intersect at or near a particular surface is evidence that the faults propagated upwards from that surface, no inference can be made about propagation directions when graben-bounding faults do not intersect at the same horizon.
Returning to the question of the mechanics of the Lake Hope faults, the coincidence of the low velocity interval in the sonic log with the stratigraphic interval previously identified as likely to have distinctive properties, still leaves open the question of what those properties might be. Use of log data assumes that the distinctive properties associated with fault generation persist to the present day. As James correctly points out, low velocity does not necessarily mean low density. Although a rather incomplete density log for the well shown in our article suggests that the layer is also low density, without better data we concluded only that the sonic log data were consistent with our interpretation of the existence of a low density layer at the time of faulting. We did not have access to data from other wells in the Eromanga Basin at the time, but previous work by Lavering (1991) was brought to our attention following publication of our paper. Lavering (1991) reports and discusses observations relating to the geological nature of the ‘C’ horizon seismic reflection in the Eromanga Basin, to which our Cadna-owie horizon is equivalent, and is based on BMR seismic data together with extensive well data which included sonic, density, gamma ray and resistivity logs. The conclusions of Lavering (1991) are that the ‘C’ reflection or ‘low velocity layer’ represents a sequence of ‘low-density (undercompacted) shales’ that is ‘irregular in its vertical thickness and areal extent’. Lavering (1991) attributes the undercompaction to the generation of overpressures, a view that is consistent with the drilling problems associated with this interval, previously attributed to overpressuring (Alexander & Sansome 1996; Peter Boult pers. comm. 1997). We emphasize, however, that present properties and thickness of this layer may well have been modified since the time when the faults were generated. We agree with James ‘that density inversion alone is no guarantee of diapirism’, but emphasize that the model we propose is based primarily on geometrical and kinematic relationships rather than on inferred material properties. The possibility of incipient diapirism at the nodal points of the polygonal cells was referred to only as a logical inference from the analogue experiments cited and not on the basis of our own observations. However, the geometric relationships of the Lake Hope area, together with the association between diapirs and similar types of faults in the North Sea, provide supporting evidence for our proposed model. James’ considers that the localized association between some folds and related graben, on the one hand, and folding of the underlying sequence, on the other, may be indicative of an alternative mechanism. This view is not supported by the evidence. It is very clear, even from the sections we present, that the vast majority of conjugates show no association whatsoever with subdued, compaction-related, folding in the underlying sequence.
In questioning whether ‘the likely strain rates at boundaries between different regimes of ductile extension associated with different fold curvatures are sufficiently large to induce fracture’ James reflects some of our own discussions when we were formulating our model. Although we have no hesitation in using the term ‘faults’ for the structures we observed and interpreted in the seismic sections, we think it unlikely that the term ‘fracture’ would be appropriate. Acknowledging the scale-dependence of terms such as ‘ductile’ and ‘fault’, we made it clear in our article that we thought it likely that if these faults were to be examined in outcrop, or possibly core, that they would have significant components of associated ductile deformation and may possibly even be classified as ductile shear zones.
Like James, we would not have expected concentric folding in the faulted sequence. However, this geometry appears to be necessary to account for the misfit due to strain incompatibility increasing upwards, from the point of origin of the faults, leading to maximum fault offset close to the Coorikiana horizon. Misfit is cumulative so its maximum does not coincide with the maximum strain incompatibility, which is immediately above the low density layer. As James points out, faults which we have previously documented have been consistently interpreted as having propagated more or less radially from what is now the point of maximum displacement. We have proposed a different relationship for the Lake Hope faults because they show quite different features. Whereas most of the faults that we have described previously are of tectonic origin, the Lake Hope faults show many features consistent with a non-tectonic origin.
In responding to James’s comments we wish to emphasize that models proposed for natural structures invariably represent simplifications that will not match individual examples in every detail. In this case of a new model for a widespread, but only recently recognized type of fault system, the crucial question is whether or not this type of fault system is generated by a density inversion beneath the faulted sequence. Our model is similar to that of Henriet et al. (1991), who also attribute the formation of folds and faults to density inversion, but it differs significantly from the model of Cartwright & Lonergan (1996, 1997). Although we have confidence in the model proposed, we expect it to undergo progressive refinement as further data
for this and similar related systems become available. This refinement will, eventually, allow a more precise description of both the structures and the mechanism than we have been able to provide. At this stage, however, we believe the model presented is the one that best accommodates the existing evidence and data constraints.
14 June 2000
 |
References |
|---|
 TOP References |
|---|
Alexander, E.M. & Sansome, A. 1996. Lithostratigraphy and environments of deposition. In: Alexander, E.M, Hibburt, & J.E. (eds) . The petroleum geology of South Australia. South Australian Department of Mines and Energy, Eastwood, South Australia49-86.
Cartwright, J.A. 1994. Episodic basin-wide hydrofracturing of overpressured Early Cenozoic mudrock sequences in the North Sea Basin. Marine and Petroleum Geology 11, 587-607.[CrossRef][ISI][GeoRef]
Cartwright, J.A. & Dewhurst, D.N. 1998. Layer-bound compaction faults in fine-grained sediments. Geological Society of America Bulletin 110, 1242-1257.
Cartwright, J.A. & Lonergan, L. 1996. Volumetric contraction during the compaction of mudrocks: a mechanism for the development of regional-scale polygonal fault systems. Basin Research 8, 183-193.[CrossRef][ISI][GeoRef]
Cartwright, J.A. & Lonergan, L. 1997. Seismic expression of layer-bound fault systems of the Eromanga and North Sea Basins. Exploration Geophysics 28, 323-331.[GeoRef]
Cartwright, J.A., Trudgill, B. & Mansfield, C. 2000. Fault growth by segment linkage: an explanation for scatter in maximum displacement and trace length data from the Canyonlands grabens of SE Utah: Reply. Journal of Structural Geology 22, 141-143.[CrossRef][ISI]
Childs, C., Easton, S.J, Vendeville, B.C, Jackson, M.P.A, Lin, S.T, Walsh, J.J. & Watterson, J. 1993. Kinematic analysis of faults in a physical model of growth faulting above a viscous salt analogue. Tectonophysics 228, 313-329.[CrossRef][ISI][GeoRef]
Henriet, J.-P., De Batist, M. & Verschuren, M. 1991. Early fracturing of Palaeogene clays, southernmost North Sea: relevance to mechanisms of primary hydrocarbon migration. In: Spencer, A.M. (ed) . Generation, accumulation and production of Europe’s hydrocarbons. European Association of Petroleum Geoscientists, Special Publications 1, 217-227.
Lavering, I.H. 1991. Observations on the geological origin of the ‘C’ horizon seismic reflection, Eromanga Basin. BMR Journal of Australian Geology & Geophysics 12, 5-12.
McGill, G.E. & Stromquist, A.W. 1979. The grabens of Canyonlands National Park, Utah – Geometry, mechanics, and kinematics. Journal of Geophysical Research 84, 4547-4563.[GeoRef]
McGill, G.E., Schultz, R.A. & Moore, J.M. 2000. Fault growth by segment linkage: an explanation for scatter in maximum displacement and trace length data from the Canyonlands grabens of SE Utah: Discussion. Journal of Structural Geology 22, 135-140.[CrossRef][ISI][GeoRef]
Moore, J.M. & Schultz, R.A. 1999. Processes of faulting in jointed rocks of Canyonlands national Park, Utah. Geological Society of America Bulletin 111, 808-822.
Watterson, J., Walsh, J, Nicol, A, Nell, P.A.R. & Bretan, P.G. 2000. Geometry and origin of a polygonal fault system. Journal of the Geological Society, London 157, 151-162.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
