models have been proposed as explanations for commonly-observed characteristics of magma-poor rifted margins worldwide, including the extension discrepancy. For each model, a simplified interpretation of part of a Galicia 3D is presented, flowline illustrating the different possible resulting stratigraphic geometries of deposition of the syn-tectonic sequences. The same dogleg line through the Galicia 3D volume is shown in each case. Model M1: all faults active over the same time period, before the crust was thinned by late displacement of the lower crust beneath a detachment. M1 predicts widespread prerift sediments (light blue), the same timing of fault onset and abandonment across the margin, and see rapid deepening in the postrift succession. Model M2: Polyphase cross-cutting faults, with extension focusing towards the line of eventual breakup. We would expect limited occurrence of prerift sediments (most top basement being the footwall of earlier faults, showing considerable exhumation), similar ages for the onset of faulting across the margin but very different timing of the cessation of faulting, and complex distribution of synfaulting sequences with a larger age range than predicted by sequential faulting. Model M3: sequential faulting predicts widespread prerift sediments, a pronounced age progression in both the onset and ending of active faulting across the margin (Unit B), and thus also in the oldest post-faulting sequences (Unit C). Block numbering from Ranero and Pérez-Gussinyé (2010); fault and synrift unit numbering from Lymer et al. (2019).
Question 1: how is the crust thinned from 20 km to zero km?
Beneath rift basins such as the North Sea, the crust thins from ~30 km to ~20 km (e.g. Wood and Barton 1983). Together with the presence of normal faults and the subsidence history, the thinning indicates that the basin formed by lithospheric extension (McKenzie 1978) with normal faults accommodating extension of the upper crust, ductile flow extension in the lower crust and lower lithosphere, and possible shear zones (Reston 1993) accommodated the extension of the uppermost mantle. However, early studies suggested that the amount of fault-controlled extension was far less than that required to thin the crust (Ziegler 1983): there is an extension discrepancy (Figure 1.5). Improvement in imaging has reduced the discrepancy under such basins to amounts expected in the presence of subseismic deformation (10%–50% of the fault-controlled extension, Marrett and Allmendinger 1992), that is the deformation occurring between faults and accommodating the distortion of faulted blocks, but as extension increases, especially at rifted margins, this discrepancy increases. This is the case at the WIM, where the amount of extension above S that can be deduced from the geometry of the faulted blocks (Figure 1.5) is considerably less than that implied by the subsidence rate (e.g. Sibuet 1992) and by measured crustal thinning (e.g. Reston 2005). Typically, the faults observed at the WIM indicate about 50%–60% extension (stretching factors of 1.5–1.6, that is thinning factors of 0.4 or lower, Figure 1.5), which is too low for the hyper-thinned crust observed at the WIM. The amount of thinning either requires more extension, or the crust to have been thinned by some other method. This is the extension discrepancy which provides fundamental constraints on the process of crustal thinning towards continental breakup.
Model M1 implies that all the brittle faulting is imaged (Figure 1.7); thus, top basement should, except where subject to erosion, be capped by prerift and early synrift rocks.
Model M2 infers that top basement is locally the exhumed slip surface of an earlier fault (Figure 1.7), cut across and displaced by the more recent faults.
In Model M3, the timing of the fault movement should systematically get younger oceanwards (Figure 1.7). Only minor extension discrepancy is excepted towards the deep margin (Ranero and Pérez-Gussinyé 2010), where the crust cut by later faults has locally been prethinned by movement on preceding faults.
Question 2: what is the role of detachments in rifting to breakup and how do they develop?
Large-offset, apparently low-angle normal faults, commonly referred to as detachment faults, accommodate crustal thinning in some areas of high factors of extension, such as at the DGM (S, Figure 1.3) and at the SIAP (H, Figure 1.4). The end-member models M1, M2 and M3 (Figure 1.7) imply different mechanisms of development of detachments.
In Model M1, the detachment zone develops late in the rifting history and corresponds to the interface at the top of the mantle and the base of the upper crust, along which the lower crust has been laterally displaced by flow during rifting.
In Model M2, the detachment consists of an early fault that cut through the mantle, allowing for serpentinization, which makes it possible for the detachment fault to remain active during all of the latest stages of rifting due to weakening of the underlying rocks. The top of the mantle thus acts as a detachment as extension focuses through subsequent faulting.
In Model M3, the detachment is a composite structure that comprises the root segments of successive steep faults which rotated to low angle before being abandoned and a new fault formed in the hanging wall of the former fault, propagating the detachment oceanwards and resulting in an apparently continuous sub-horizontal surface.
Detachment faults have been imaged at a variety of margins and hyperextended basins (e.g. de Charpal et al. 1978; Reston et al. 2001), but are best known from the GM (e.g. Reston et al. 2007). There, 3D analysis (Lymer et al. 2019) suggests that S formed from a combination form of models M2 and M3, with an early fault that cut through the CMB and led to mantle serpentinization (M2), but evolved into a 3D rolling hinge onto which multiple faults rooted at once, once the crust had thinned to <10 km (variant of M3). Thus, sequential faulting may develop as one of the later phases of extension in a polyphase mode.
Question 3: what is the evolution of the lithospheric rheology, strain distribution, melt production and serpentinization during rifting?
In contrast to magma-rich margins, magma-poor margins such as the WIM display few and small magmatic bodies atop the well-defined faulted blocks. Several hypotheses have been advanced to explain the lack of magma at such margins: very slow rifting (Pérez-Gussinyé et al. 2006), pre-depletion of the asthenosphere following earlier melting events (Muentener and Manatschal 2006), lithospheric (as opposed to crustal) DDS in which crustal breakup occurred before lithospheric separation (Huismans and Beaumont 2011) and cool sub-lithospheric mantle (Reston and Phipps-Morgan 2004). A key to understanding the lack of melt is to constrain the timing and rate, both absolute and relative, of crustal rifting and of mantle thinning (Brune et al. 2014, 2017; Ros et al. 2017).
To summarize the above section, the remaining key unknowns at the WIM are as follows:
– How are the lithosphere and the crust thinned in space and time from 20 km to zero? This can be tested by addressing the extension discrepancy (Figure 1.5) and determining the detailed rift history of the WIM by sampling the age of the synrift in each half-graben to calibrate the existing excellent seismic database at this margin.
– How and when do detachments develop and control the structural evolution of distal margins? Determining the detailed rift history will place the role of detachments into a well-constrained framework to test the current models of development of the WIM, and to determine the timing of any switch to migrating, detachment-controlled extension.
– How does evolving lithospheric rheology control strain localization,