In this Comment, I argue that Duarte et al.’s paper (2013) lacks relevant data, that the authors do not use critical model results that would constrain their hypothesis of subduction initiation at the southwest Iberia margin (SIM), and that they overlooked relevant literature on the same topic.
Duarte et al.’s main premise is that preexisting convergence may trigger subduction initiation elsewhere. In contradiction, another main premise of Duarte et al. is that subduction in Gibraltar is by (sic) “rapid rollback.” In the absence of opposing velocities, which seems to be the case in Gibraltar, hinge rollback should lead to extension in both plates: in the subducting plate because of slab-pull, and in the overriding plate because of divergence (e.g., the Sea of Japan), in great contradiction to the convergence taken as premise.
Duarte et al. assume that the sedimentary wedge west of Gibraltar is a compressional accretionary wedge. However, this interpretation is not unique, especially if the complete structure of the Gibraltar’s sedimentary wedge is taken into account (see Marques  for a complete discussion).
Subduction in Gibraltar is taken as certain, because Duarte et al. assume that the interpretation of tomography has a unique solution. However, Marques (2010) showed that there are sound alternatives.
Structure of the SIM
The size of the structures interpreted as signs of subduction initiation, almost exclusively the Gorringe and Horseshoe thrusts (Duarte et al., 2013, their figure 1B), have lengths of <150 km and shallow depths (their figure 2). Therefore, they seem too small for a lithospheric-scale process.
Although critical for their hypothesis, Duarte et al. do not give the lithospheric structure of continental and oceanic Iberia. Nikolaeva et al. (2010) showed that the lithospheric structure is a key to understanding subduction initiation at passive margins, but Duarte et al. do not use it to evaluate their hypothesis.
Although critical to their hypothesis, Duarte et al. do not give any data about seismicity in the SIM. The magnitude and destruction of the 1755 Lisbon earthquake cannot be explained by the faults drawn by Duarte et al., because they are too small. Furthermore, very little is known about this shock; even its location is still greatly controversial. Despite the total lack of seismic data, the 1755 shock is taken by Duarte et al. as evidence of subduction at the SIM. In contrast, a great deal is known about a large earthquake (Ms ∼7.9) that occurred in the Gorringe region in 1969, but no data are given. From aftershocks, Fukao (1973) concluded that the fault dip is ∼52° to N325, which, together with the focal mechanism, shows that thrust is toward the SIM, therefore opposite to the thrusts and the subduction polarity proposed by Duarte et al.
Despite being the work that first showed that additional forces are needed to overcome bending and frictional resistance in order to initiate subduction, the seminal work of McKenzie (1977) is not used to discuss the force balance. McKenzie was the first to suggest subduction initiation at the SIM. McKenzie even gives a minimum value that could have been used by Duarte et al. to constrain their hypothesis.
Duarte et al. do not explain how the rollback in Gibraltar (if it exists) can generate compression, nor do they give us the magnitude of the forces involved in the Nubia-Eurasia interaction (not strict convergence at this longitude) or MAR-push, so that the reader can compare with the values estimated in analytical models as the minimum to initiate subduction at a passive margin.
In this paper specifically about Iberia, subduction in northern Iberia is overlooked (e.g., Alvarez-Marron et al., 1997). By doing so, the evaluation of the force balance along the SIM is flawed. Overthrusting in northern Iberia could take up all the Nubia-Iberia convergence, as it seems to have occurred in the past, which makes it critical in the evaluation of the force balance in Iberia.
Duarte et al. overlooked Artyushkov (1987), who analyzed a situation (seemingly) very similar to the SIM. This insightful paper could have been used to help understand the geodynamic setting of the SIM.
Nikolaeva et al. (2010) is misquoted, because: (1) the driving forces in the numerical model are not taken into account by Duarte et al.; (2) the numerical models show that there are two stages in the process of subduction initiation, overthrusting and subduction proper, and the latter might never occur, which is not discussed by Duarte et al.; and (3) the parameters favoring self-sustained subduction are not used by Duarte et al. to discuss the SIM deep structure.
Nikolaeva et al. (2011), published in Geology with the title “Numerical analysis of subduction initiation risk along the Atlantic American passive margins,” is not used to help in analyzing the probability of subduction to initiate at the SIM. However, Nikolaeva et al. (2011) indicate the parameters that favor subduction initiation, which lack in the SIM: (1) the thickness of the continental crust onshore the SIM (∼30 km; González et al., 1996) is too thin to favor initiation; and (2) if there were lithospheric-scale thrusting of Iberia over the Atlantic, then the SIM would be uplifted, which is not the case. This is good evidence that the Gorringe pop-up is local, not of lithospheric scale.
To conclude, Duarte et al.’s paper is based on disputable premises, and does not show compelling evidence for subduction initiation at the SIM. The scale of the used thrusts is not compatible with the scale of subduction; therefore, thrusting could mean only local crustal shortening.