As mentioned in the Comment from Zhao et al. (2010), our original paper (Zheng et al., 2009) provides the first images of the crustal structure beneath the Western block and across its boundary with the Trans-North China orogen (TNCO). Our interpretation on the crustal structure is not only based on the evidence from seismic observations, but also combines geological observations. We respond to the comments of Zhao et al. in the order laid out by the authors.

The first point raised by Zhao et al. is that the two low-velocity zones (L1 and L2) recognized in our paper do not penetrate into the mantle, which differs from some published results. If, however only common conversion point (CCP) stacking is used, a continuous trace of positive phases can be identified in the uppermost mantle (located at ∼107.7–109.3°E and at a depth of ∼45–55 km; see Fig. DR1B, Zheng et al., 2009, Data Repository item 2009100). To understand the nature of this trace, we can use forward modeling and waveform inversion to analyze the observed data. This shows that the image within the uppermost mantle, identified by the CCP stacking, is an artifact of PpPs (P to S conversion after reflection from the surface) multiples from a shallow crustal structure, and not from dipping discontinuities that extend into the mantle. It is therefore necessary, for obtaining a reliable image of a Precambrian subduction fossil, to adopt an integrated receiver function imaging technique (Zheng et al., 2009, Data Repository item 2009100). Our results suggest that caution should be taken when using seismological techniques to investigate Precambrian subduction remnants within the uppermost mantle. Caution should be exercised because a section of Precambrian continental crust subducted into the topmost upper mantle may be exhumed or modified by dehydration/melting, and then may fully disappear or be reworked. Despite this, our findings clearly indicate that Precambrian subduction remnant can be preserved in the continental crust. This has allowed the imaged crustal structure to be used to identify the direction of subduction in the TNCO as westward.

The second point raised by Zhao et al. deals with the positioning of the seismic profile AB. First, we reiterate the fact that the Ordos Basin is covered by thick Quaternary sediments, which makes this area unsuitable for acquiring high-quality seismic data. Thus the profile AB, along the boundary between the Ordos block and the Khondalite belt, provides the best location for a seismic profile. Zhao et al. propose that the low-velocity zone, L1, can be alternatively interpreted as overthrust crustal slices of the Western block emplaced during its collision with the Eastern block. As we mentioned in our paper, this is unlikely because “the rocks in the hanging wall should have an associated high seismic velocity rather than the low-velocity zone of L1” (Zheng et al., 2009, p. 397).

The final statement of Zhao et al. regarding the receiver function section in our figure 2 is a misunderstanding. First, the receiver function is obtained by deconvolving the vertical from the radial seismogram component, and represents the structure response beneath the seismic station. Second, the waveform of the receiver function is a composite of conversions and reverberations derived from the reflections, refractions, and conversions of seismic waves traversing velocity discontinuities. Therefore, the receiver function phases may either be Ps conversions from a real discontinuity or PpPs/PsPs+PpSs multiples from a shallower discontinuity. Using forward modeling and waveform inversion we have identified that the strong negative phases that Zhao et al. highlighted are a result of multiples, and that the weaker negative phases are the Ps phases converted from the upper interfaces of the low-velocity zones L1 and L2.

Our interpretation supports a dynamic model in which the amalgamation of the North China craton took place through two stages of continental collisions, occurring at ca. 2.1 Ga and 1.9 Ga (Faure et al., 2007). The third point of Zhao et al. argued against this interpretation because the ca. 2.1 Ga collisional event has not been supported by metamorphic evidence or by the identification of metamorphic U-Pb zircon ages. Only ca. 1.85 Ga metamorphism has been established from the TNCO. Nevertheless, continent–continent collisions and arc–continent collisions are not necessarily identified by high-pressure metamorphism; in most cases, it depends upon crustal depth. The ca. 2.1 Ga event has been recognized as an extensive crust melting event along the TNCO. The SHRIMP zircon U-Pb ages obtained from Wanzi supercrustal rocks and Nanying granitic gneisses suggest a sequential volcanic-sedimentary process and crust melting events at 2.1–2.0 Ga (Zhao et al., 2002). In addition, the Paleoproterozoic (2.12–2.04 Ga) crust melting events have been recognized in many other terranes along the TNCO, such as the Wutai, Hengshan, Luliang, and Huai'an terranes (Kröner et al., 2005; Zhao et al., 2008a, 2008b). These granitoid gneisses can be ascribed to the response of the westward subduction and post-collisional crust melting of the Taihang Ocean (Faure et al., 2007). The low-velocity zone L2 in our seismic imaging mostly likely corresponds to a Precambrian subduction continent segment exhumed from the topmost upper mantle during the last stage of subduction. Our findings are therefore in good agreement with the dynamic model described by Faure et al. (2007).