Recently, Zheng et al. (2009) presented seismic images of the Trans-North China orogen (TNCO), a recently recognized Paleoproterozoic continent-continent collisional belt, along which the discrete Eastern and Western blocks amalgamated to form the North China craton (Zhao et al., 2001, Kröner et al., 2005; Wilde and Zhao, 2005; Faure et al., 2007). It is an important contribution because it provides the first images of the crustal structure beneath the Western block and across its boundary with the TNCO. Zheng et al. (2009) attempted to use these seismic images to determine the subduction polarity of the TNCO, which has been debated recently, with one school proposing westward subduction (Faure et al., 2007), whereas others favor east-dipping subduction (Kröner et al., 2005; Zhang et al., 2007, 2009). In the seismic images of the Western block and the TNCO, Zheng et al. recognized two low-velocity zones (marked L1 and L2 on their figures 3A and 3B), of which L1 was interpreted as the remnant of upper-middle crustal material associated with westward-dipping subduction beneath the Western block, and L2 as the remnant of upper-middle crustal material associated with westward-dipping subduction beneath the so-called “Fuping block” (Faure et al., 2007). We argue that Zheng et al.’s interpretations are not entirely convincing for the following reasons.

1. The two low-velocity zones (L1 and L2) recognized by Zheng et al. only extend through the imaged crust and do not penetrate into the mantle, as would be expected if they were true subduction zones. There are a number of examples where seismic images have successfully been applied to reconstruct Late Archean and Paleoproterozoic subduction zones, especially in Baltica (BABEL Working Group, 1990) and North America (Calvert et al., 1995). In these examples, the seismic images clearly show that the subduction zones extend as discrete linear structures into the mantle. This is not the case with the seismic image shown in figures 3A and 3B of Zheng et al. (2009), in which the L1 and L2 subduction zones are strictly limited to above the Moho. Furthermore, the L1 and L2 zones are not truncated by or in direct contact with the Moho, but separated from the latter by up to 4 km of lower crust, which leaves little possibility that the two zones had once extended into the upper mantle and were then exhumed to crustal levels due to slab break-off and tectonic uplifting, as interpreted by Zheng et al.

2. The position of the ∼550-km-long seismic profile AB is far from ideally sited in order to obtain the best picture of the structure beneath the Western block. It is mostly located along the boundary between the Ordos block and the Khondalite belt (see figure 1 of Zheng et al., 2009), and thus is likely to also record information of the earlier collision of these two terranes at ca. 1.92 Ga (Santosh et al., 2007). The west-dipping segment of L1 can be alternatively interpreted as overthrust crustal slices of the Western block during its collision with the Eastern block at ca. 1.85 Ga. In addition, Zheng et al. did not explain why they ignored a strong low-velocity zone that extends along the whole AB profile, as shown in their figure 2, but instead, only focused on the L1 and L2 low-velocity zones.

3. The seismic interpretations of Zheng et al. do not adequately reflect the surface geology. As indicated above, they interpret the L1 and L2 low-velocity zones as two west-dipping subduction zones, thus supporting the model that the North China craton was assembled as a result of two Paleoproterozoic collisional events, with the older one the result of collision between the Fuping block and the Eastern block at ca. 2.1 Ga, and the younger one formed when the combined Fuping/Eastern block collided with the Western block at 1.9–1.8 Ga (Faure et al., 2007). The proposed ca. 2.1 Ga collisional event has not been supported by metamorphic petrology or by the identification of metamorphic U-Pb zircon ages. Instead, metamorphic studies have established that the Hengshan, Wutai, and Fuping complexes experienced a single metamorphic event that is characterized by a clockwise pressure-temperature path involving isothermal decompression (Zhao et al., 2001). There is thus no support for the existence of two metamorphic events in these areas. All available metamorphic ages obtained for these three complexes are ca. 1.85 Ga, with none around 2.1 Ga (see summary in Zhao et al., 2007). Therefore, the model for the evolution of the TNCO involving two subduction zones, and thus two collisional events, is entirely speculative. As a consequence, the interpretation that the L1 and L2 low-velocity zones in the seismic images of Zheng et al. are “fossil” subduction zones is questionable and not supported by other data. The direction of subduction polarity in the TNCO has thus still not been resolved.