Читать книгу Continental Rifted Margins 2 - Gwenn Peron-Pinvidic - Страница 16

The general evolution of the WIM has been derived from an extensive network of 2D seismic reflection profiles, calibrated by scientific drilling acquired during ODP Legs 103 at the DGM (Boillot and Winterer 1988), and Legs 149 and 173 at the SIAP (Sawyer et al. 1994; Whitmarsh et al. 1998) (Figures 1.3 and 1.4). The recovered cores and geophysical logging data provided constraints on the lithology, biostratigraphy, palaeomagnetics, geochemistry, bulk density, thermal conductivity, and compressional seismic velocities on the rocks forming the basement and the synrift sediments. ODP results have also provided sparse information on the ages of the synrift sequences, mainly restricted to the top of some tilted blocks and basement highs.

Figure 1.5. The extension discrepancy across the Iberian margin. a) The Flemish Cap/Galicia transect. Here the amount of measurable extension b) across the Galicia Margin is too slight to explain the observed thinning. c) The same is true, but to a lesser extent, across the SIAP (composite profile LG12/TGS, interpretation modified from Mohn et al. 2015): there is still an extension discrepancy at the deep margin despite the very large amount of extension associated with different generations of detachments. d) Cross-plot of crustal thinning versus thinning from faulting for these two margins showing the relative magnitude of the extension discrepancy

The rifting west of Iberia affected Variscan crystalline continental basement that has been sampled from two tilted blocks at ODP Site 639 and at GAL11 diving site (Boillot and Winterer 1988; Boillot et al. 1988; Figure 1.3), and interpreted more widely from seismic velocities (Bayrakci et al. 2016; Davy et al. 2018). Based on a 3D fault analysis, Lymer et al. (2019) subdivided the synrift stratigraphy at the DGM in terms of age relative to the local faulting (e.g. the synrift of block 3 is divided into the packages 3A, 3B, 3C, Figure 1.3). Thus, the lowest unit A is likely prerift or early synrift, unit B is considered to be synfaulting, and unit C to be synrift but post-local faulting (see Figure 1.3 for details). Following the Variscan Orogeny, multiple phases of extension affected the WIM over 100 Myr, from the Triassic until the mid-Cretaceous, when eventual northward propagating breakup led to diachronic opening of the North Atlantic (e.g. Srivastava et al. 1990; Tucholke et al. 2007). Late Triassic to Early Jurassic early rifting affected a wide zone (Tucholke et al. 2007), producing large intracontinental basins (Figure 1.1 – LP, PB; Wilson et al. 1989, pp. 341–361; Murillas et al. 1990). Subsequent Jurassic extension thinned the crust to ≥10 km and led to the development of the GIB (e.g. Pérez-Gussinyé et al. 2003). ODP Site 639 recovered shallow water Tithonian carbonates at the western edge of the GB (Boillot and Winterer 1988) (Figures 1.1 and 1.3), that could correspond on seismic sections to relatively thin, poorly reflective packages (unit A on Figure 1.3). At the Early Cretaceous, the DGM developed due to extreme crustal thinning and mantle exhumation processes (Boillot et al. 1987), with hyper-thinning of the crust from ~10 km at the western edge of the Galicia Bank to zero in areas of exhumed serpentinized mantle (Figure 1.3). Syn-tectonic sediments (unit B) recovered at the top of a fault block at ODP Sites 638 and 639 have been dated as Valanginian–Barremian, whereas possible syn-tectonic units sampled at more oceanward Site 640 are Barremian–Aptian in age (Applegate and Bergen 1988). The westernmost synrift sample has been obtained by submersible from the top of a crustal block outcropping at the seafloor (GAL-11, Figure 1.3), consisting of sandstones whose age was not readily dateable (Reston 2005). The faulted blocks above S are consequently capped by a deep-water clastic sequence (Boillot and Winterer 1988), rather than true prerift sediments, such as the shallow water carbonates recovered at ODP Site 639. In addition, the ages of the recovered synrift sediments indicate that the top of the synrift units becomes younger towards the distal DGM (Boillot and Winterer 1988; Figure 1.3, compare ages at Sites 641 and 640). The age of the synrift sediments across the Galicia Margin is therefore diachronous (Reston 2005), which is consistent with the oceanwards migration of the rifting, as suggested from 3D observations on the kinematic development of the S detachment (Lymer et al. 2019). ODP Leg 103 clearly constrained the age of the extensional faulting at only one fault block (Sites 638, 639 and 641), where the prerift, early synrift and late synrift sequences were recovered. The age distribution of rifting across the Galicia Margin is therefore uncertain due to the punctual availability of time constraints on the age of synrift sequences.

Figure 1.6. Summary of fault heaves analysis from the Galicia 3D volume and evolutionary model for the DGM (Lymer et al. 2019)

CONTINUATION OF CAPTION FOR FIGURE 1.6.– a) Map of the top basement level (depth shown by scale bar in meters) generated within the Galicia 3D volume (location in Figure 1.1) on which the flowlines identified from corrugations on S are used to measure heaves on the block-bounding faults. b) 3D model of fault development based on the geometry of faults within set 3/4 showing that more than one fault is needed to accommodate extension at any one time as faults have limited along strike length. The flowline heaves (c, d and e) are grouped into three sets in which the heaves are complementary: as one fault dies out the displacement is taken up on another in the same set. Overall, the cumulated heaves f) show a northward decrease in extension, consistent with the propagation of rifting from the south. g) Block diagram summarizing the model of Lymer et al. (2019) for the evolution of the DGM based on the 3D fault analysis shown in a) to f). After a period of extension and thinning, the blue faults cut across the CMB and allow the ingress of water to serpentinize the mantle. Subsequent faults propagate up through the hanging wall of preceding faults as in a rolling hinge model, but one in which the root can slip at low-angle due to serpentinization and in which multiple spatially overlapping faults are active at any one time (color coded by fault set).

In the SIAP, Legs 149 and 173 recovered synrift stratigraphy through deeper settings and thicker postrift sequence than at the DGM (Figure 1.4). Although the main focus of these legs was to characterize the basement type by specifically targeting structural highs, the SIAP drilling transect provided limited constraints on the age of some individual faults. Tilted Tithonian clays and claystones mixed with clastics and sandstones predating the faulting were recovered at Sites 1065 and 901, interpreted as having been deposited in enclosed, relatively shallow and anoxic basins, close to vegetated land (Whitmarsh et al. 1998). Benthic foraminifera recovered at Site 901 indicate a neritic zone (Collins et al. 1996), that is continental shelf environment, and were interpreted as marking the onset of the rifting (Whitmarsh and Wallace 2001). A late Berriasian–early Valanginian age for the beginning of significant extension is supported by the pre- to syn-rift nature of the sediments recovered at Site 1069 (Figure 1.4), that are constituted of shallow-water limestones, interpreted as predating the main extensional phase, overlaid by conglomerates and sandstones, interpreted as marking the onset of extension at this location (Whitmarsh et al. 1998). Later synrift units have not been sampled but the timing of crustal extension is constrained by 40Ar/39Ar cooling data for samples recovered at Site 1067 and 900, indicating that these deep crustal rocks were exhumed from >18 km (0.6Gpa) through ~4 km depth by 137 Ma (Wilson et al. 2001, pp. 429–452). Thus, in the ~10 Myr between the onset of rifting in the Tithonian and the unroofing of these rocks, the crust appears to have been thinned by a factor of ~5. The final phase of crustal extension, which produced the current observed geometry of the margin probably occurred shortly after this (Tucholke et al. 2007). In the distal zone of exhumed mantle, oceanward of Site 900 (Figures 1.1 and 1.4), a series of basement highs were drilled: the sediment atop the highs represents a minimum age for the unroofing of that particular portion of peridotitic basement. Drillings have sampled mixtures of serpentine breccias and mass flow deposits of Aptian–Barremian age that may be related to the breakup process (Tucholke et al. 2007).

To summarize, in the SIAP, the Tithonian–early Berriasian sequence is tilted and part of the pre-faulting units (Figure 1.4). The late Berriasian–early Albian sequence either onlaps previous sequences/basement, or locally thickens towards the block bounding faults. The evidence thus points towards major rifting and crustal extension occurring over perhaps 10 Myr between the Tithonian and the Valanginian, followed by a phase of mantle unroofing over a further 10 Myr, before the onset of seafloor spreading during the Aptian–Barremian. As within the DGM, extension appears to have migrated oceanwards, but culminates in the unroofing of a far wider zone of mantle rocks. Although the overall time between the onset of rifting and the spreading of seafloor is similar at the two margin segments, approximately half of that time at the SIAP only involved the unroofing and subsequent extension of mantle rocks. Hence, rift duration leading to complete crustal separation may have only been about 10 Myr. Both the DGM and SIAP exhibit tilted fault blocks, detachment faulting, tectonic CMB, serpentinization beneath hyper-extended crust, and a wide zone of mantle exposed at the seafloor during rifting. At both segments of the margin, ODP results suggest that extension during the rifting may have migrated oceanwards. However, although the WIM has been extensively studied, the details of the timing of rifting remain largely unconstrained due to the fact that most scientific drillings have targeted basement highs. This strategy, while revolutionizing our understanding of the structures of rifted margins and introducing the concept of mantle unroofing, has not recovered the complete synrift sequences needed to determine the timing of motion on individual faults and the across-strike temporal distribution of fault activity during the rifting. Distinctive modes of fault development and emplacement have resulted into three main models proposed to explain the migration of the extension suggested by the available data. In the following section, we introduce these models designed from observations at the West Iberian Margin, and likely explain how magma-poor margins might have formed.