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.
1.4. Current models of development of the West Iberian Margin
Exploration of the WIM has led to the definition of three major models of continental breakup (Figure 1.7) that appear to offer a solution to the extension discrepancy (Figure 1.5) and which have subsequently applied to the rifting and breakup of other magma-poor margins. These models are: crustal depth-dependent stretching (DDS), cross-cutting polyphase faulting, and migrating faulting models, which include the sequential faulting. They imply fundamentally different mechanisms for thinning the continental crust, the latter two being strongly diachronous in terms of fault activity across the WIM, although ODP Legs 103, 149 and 173 only sampled the complete synrift in a single place (Site 639). For instance, the relative age of the syn-tectonic packages in different sub-basins (e.g. between blocks 3 and 5, Figure 1.7) depends on the mode of extension: simultaneous, focusing or migrating.
Model M1, crustal depth-dependent thinning. Driscoll and Karner (1998) and Davis and Kusznir (2004, pp. 92–136) suggested that both the amount and the mechanisms of extension vary with depth across the crust, due to the different rheologies of the upper and the lower crust. In this way, the extension occurs by faulting in the brittle upper crust, whereas the lower crust deforms through ductile processes, not just stretching but also displacing or flowing, thus thinning the crust much more than indicated by the faulting (Figure 1.5). In some cases, the interpretation of eroded fault block crests and the dominance of shallow water sediments has led to the idea that much of the DDS occurred from the displacement of the lower crust late during the rifting process (Clift and Lin 2001, pp. 489–510; Davis and Kusznir 2004, pp. 92–136). In this interpretation, all the brittle deformation is imaged and occurred during a single phase, so the synrift sediments have the same age across the margin (Figure 1.7).
Model M2, cross-cutting polyphase faulting. An alternative perspective is that the crust thins by the same amount at all levels, but that not all the upper crustal extension is correctly identified. In this polyphase faulting model, extension might be underestimated where young faults cut and displace old faults (Reston 2005; Reston and McDermott 2014) as extension occurs through distinctive phases of deformations that focus with time into the rift axis. This model results in complex structural and stratigraphic geometries which means most of the earlier faults cannot be identified from seismic reflection profiles, thus explaining the extension discrepancy (Reston 2009; Figure 1.5). Seismic modeling of the resulting complex geometries confirms that earlier faulting is indeed likely to be misinterpreted, for example as eroded fault block crests (McDermott and Reston 2015). As faulting may have started at the same time across much of the margin, but focused during rifting towards the rift axis, the earliest synrift should be present across the margin, but the youngest synrift should be present only near the axis of breakup, being of the same age as local post-faulting sediment on the flanks (McDermott and Reston 2015). The initial top basement and pre-rift are dismembered and only preserved locally, as much of the top basement surface comprises early fault surfaces (Figure 1.7).
Model M3, migrating faulting. The migrating faulting models, for example rolling hinge (Buck 1988) and sequential faulting (Ranero and Pérez-Gussinyé 2010; Brune et al. 2014), are polyphase but with the locus of the extension migrating oceanwards, and the faulting migrating into the hanging wall of the previous fault, rather than focusing towards the future oceanic spreading axis. Extension occurs on a succession of individual faults active one at a time, with each fault that forms as a steep structure, flexurally rotates, and is abandoned when the angle becomes too low to allow for slip along the fault plane. Only then does a new fault form in the hanging wall of the abandoned fault, and slices ever farther into one side of the rift, with the same process continuously repeating until the crust has been thinned to zero (Ranero and Pérez-Gussinyé 2010; Brune et al. 2014). As each fault avoids cutting earlier faults, all the extension should be recognized. Extension discrepancy should only result if combined with another process (e.g. subseismic faulting, Reston and McDermott 2014), or if the migrating locus of faulting overlaps with the preceding ones, producing an element of cross-cutting and unrecognized geometries. A consequence of this model is the systematic oceanwards decreasing of the age of all of the syn-faulting sediments accumulated within each newly-formed half-graben (Figure 1.7).
The sequential faulting, as an inherently asymmetric process, might also explain the development of asymmetry between conjugate margin pairs. As proposed by Ranero and Pérez-Gussinyé (2010), sequential faulting might develop during early rifting when the continental crust is still >20 km thick. However, this early development conflicts with the apparent link between asymmetry development and crustal embrittlement, which requires asymmetry to develop only once the crust is thinner than 10 km (Reston and Pérez-Gussinyé 2007; Reston 2009). However, dynamic models propose a later onset of the development of asymmetry (Brune et al. 2014), when the crust becomes thinner than 20 km. The latter scenario is consistent with 3D seismic-derived interpretations showing that extension migrated on a 3D multi-fault variant of the rolling hinge model, only once slip at low-angle was allowed (Lymer et al. 2019). Brune et al. (2014) proposed hot ductile lower crust as a rooting zone, while Lymer et al. (2019) suggested that the recognized serpentinized mantle beneath S (Bayrakci et al. 2016) might represent the weak root zone.
The fundamental disparities between the three models – in terms of timing of faulting, number of faulting phases and rheologies – demonstrate that despite decades of exploration of the WIM, our knowledge of rifting and breakup remains fundamentally incomplete, as long as detailed timing of the geological events at rifted margins remains undefined.
1.5. Remaining questions at the West Iberian Margin and other magmapoor margins
The diversity of existing current models developed at the WIM shows that we still fundamentally do not know how the continental crust thins to zero, exposing the Earth’s mantle to the surface during rifting, leading to eventual continental breakup. The models presented in the previous section (Figure 1.7) are all potentially viable but imply fundamental differences in the mechanisms of thinning of the crust. The differences between these models can be summarized in three major, yet unanswered, questions, emerging from our current knowledge of the structures of the Galicia Margin, and more globally related to the development and evolution of magma-poor rifted margins.
Figure 1.7. Models proposed to explain fault development and the evolution of the West Iberian Margin, highlighting different predictions in terms of the age distribution of synrift sequences, the different role of detachments and different distributions of prerift/top basement
CONTINUATION OF CAPTION FOR FIGURE 1.7.– Although based on observations made at the West Iberian