The reactivation of craton is a crucial issue for the understanding of continent lithosphere. Although the partial loss of the thick lithospheric root of the North China Craton (NCC) was documented, its mechanism remains a contentious topic in the study of cratonic evolution. In this paper, we imaged the crust-mantle structures by using the data from 36 permanent stations with fine spatial coverage in the Yanshan tectonic belt (YSTB) and the Taihangshan region (TSR) in the northern NCC. Two tectonic units with distinct crust-mantle structures were recognized and their junction zone was identified. The results reveal distinct features of the crust-mantle boundaries of the YSTB and the TSR which are characterized by a sharp interface and a thick transition zone, respectively. Based on the analysis of the seismic structures of the crust-mantle boundary and the upper mantle, we propose that the NCC reactivation was possibly derived by a slow secular ascension of hydrated asthenospheric material atop the stagnant Pacific slab, which influenced the formation of crust-mantle transition zone in the eastern NCC. On the other hand, regional delamination of the deep crust after orogenic thickened and/or magmatic underplating beneath the YSTB caused the spatial variation of the crust-mantle boundary. Both of the mantle processes possibly changed the nature of the lithospheric mantle.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Our general understanding of the Archean cratons is that they are cold, buoyant and stable; however, the North China Craton (NCC) is a significant exception. This craton is one of the oldest Archean cratons in the world, preserving continental rocks as old as 3.8 Ga [Liu et al., 1992]. Contrasting compositions of mantle xenoliths from Palaeozoic kimberlites and Cenozoic basalts suggests the removal of the Archean lithospheric mantle during the Phanerozoic [Menzies et al., 1993; Griffin et al., 1998]. The timing of the lithospheric removal possibly occurred during the Jurassic-Cretaceous [Wu et al., 2003; Menzies et al., 2007, and references therein] when the eastern NCC experienced significant reactivation, accompanied by extensive magmatism and large-scale basin formation. Accordingly, studies of lithospheric removal are of profound implications for our understanding of continental modification. However, the geodynamic mechanism of the removal of the NCC lithosphere is strongly debated.
 The delamination of the deep crust and the lithospheric mantle could result in the change of the crust-mantle boundary zone due to that the lithospheric mantle was partly or entirely added from the ambient convecting upper-mantle. Otherwise the magmatic underplating from the ascending mantle could result in the dynamic thickening of the crust-mantle transition zone. The large-scale coverage and high spatial resolution of the seismic observation are necessary to study the crust-mantle interaction of the craton beyond the limitation of the unevenly distributed geochemical samples.
 In the previous seismic imagines from two linearly arranged seismic observations in the NCC, the imaged crust is characterized by a sharp crust-mantle boundary beneath the Yanshan [Zheng et al., 2007], and a thick crust-mantle transition zone of about 5–10 km beneath the Taihangshan [Zheng et al., 2006]. The varying structure is cryptic and significant for understanding the NCC reactivation. Unfortunately, the spatial extent of the contrast tectonic units and the governing mechanism remain ambiguous due to the limit of linearly arranged observations. The dense permanent stations of Capital Seismic Network in China (CSNC) with an average spacing about 30 km and fine spatial coverage, however, make it possible to determine the spatially range of the tectonic units with distinct crust-mantle boundary structure.
 In this paper we studied the crust-mantle structures in the Yansan tectonic belt (YSTB) and in the Taihangshan region (TSR) by using the seismic data from 36 permanent CSNC stations. And then, the regional seismic images of the crust-mantle boundary zones in the YSTB and the TSR were used to investigate the inherence of the contrast structures and place further constraints on the mechanism of the NCC reactivation.
2. Geologic Setting
 The NCC, which is traditionally named as the Chinese part of Sino-Korean Craton, consists of two major blocks, the Eastern and the Western blocks. The amalgamation of the two blocks with age of ∼1.85 Ga formed a N–S trending collisional orogen called the Trans-North China Orogen [Zhao et al., 2001] (Figure 1). The NNE-trending TSR separates the Trans-North China Orogen to the west and the Bohai Bay basin to the east (Figure 1). This Precambrian terrane was uplifted in the Mesozoic and Cenozoic as indicated by the dome-shaped tectonics, the emplacement of voluminous Late Mesozoic granites scattered in the north of the TSR, which is often considered to be tectonically coupled with the sinking of the Bohai Bay basin, and are both thought to be associated with the NW-SE tectonic extension in the Late Mesozoic to Cenozoic [Xu et al., 2001]. The E-ENE trending YSTB (Figure 1), on the other hand, was inferred to have experienced multiple phases of — mostly N–S directed — contractional deformation during the Jurassic and Cretaceous [Zheng et al., 2000]. Tectonic extension accompanying voluminous volcanism, however, dominated the YSTB soon after the contraction events, suggesting that the YSTB is likely to be a tectonic part of the Mesozoic-Cenozoic extension system. Generally, it was proposed that the present tectonics of the YSTB and the TSR (Figure 1) are the direct consequence of the cratonic reactivation [Davis et al., 2001; Chen et al., 2003; Gao et al., 2004; Zheng et al., 2006, 2007; Jia and Zhang, 2005].
3. Seismic Imaging Results
 The data recorded from the 36 stations during January 2002 to July 2003 were used. Applying the same method by Zheng et al. , the receiver functions were produced from teleseismic records for each station. For most of the stations, more than 50 receiver functions with high signal-to-noise ratio were used. The resulting dataset contains 3130 receiver functions for further analysis (see auxiliary materials).
 We ranked the receiver functions stacked within the 0–360° azimuth from east to west, and from north to south respectively as shown in Figure 1. The stations distributed (see auxiliary materials) in the E–W and the N–S trending converge at a junction area, and we divided them into the E–W and the N–S profiles according to their waveform properties. The dominant phases near 4–5 s are clear in all the stations. In the E–W profile another strong phase can be continuously traced near 14–15 s, in contrast to that the same counterparts do not appear in the N–S profile, except a few stations.
 An integrated receiver function imaging technique [Zheng et al., 2006] was used, which involves iteratively implementing waveform inversion and the Common Converse Point (CCP) stacking of receiver functions. This technique permits the exploration of detailed crustal features. In the CCP imaging, two study profiles with approximate E–W and N–S trending were selected. In the waveform inversion the receiver functions stacked in an azimuth range ∼110°–160° were used as the observation data (marked by ‘OBS’ in Figure 1) (see auxiliary materials for imaging method).
 The final velocity images for the E–W (YSTB) and N–S (TSR) profiles are shown in Figures 2a and 2b, which are compiled from the best-fitting shear wave velocity models for the individual stations (see auxiliary materials for imaging results). The four stations (CHC, SHC, ZJK, and ZHB) in the junction area were compiled in the E–W imaging profile as well as the N–S imaging profile for identifying the division of tectonic units. The most striking feature is that the imaged crust-mantle boundary zones beneath the YSTB appear sharp, while, in contrast, thicker ones characterize the TSR. This result demonstrates that the contrast structure features are inherent in the whole study region, consistent with those from the linear temporal station profiles [Zheng et al., 2006, 2007].
 By comparing the synthetic receiver functions calculated from the inverted velocity models with the data, we identified that the P-to-S convert wave energy near 4–5 s and the following PpPs phases near 14–15 s are generated by the crust-mantle boundary. The other strong phases are generated by the intra-crust interfaces. It is confirmed that the distinct characteristics shown in the latter phases of the receiver function profiles (Figure 1) are mainly generated by the distinct structures of the crust-mantle boundary. As examples we display the velocity models and the waveforms of receiver functions for stations FEN and MIY at the YSTB, and ZHB and CHC at the TSR in Figures 2c and 2d. It is clear that the thick crust-mantle transition zone results in diffused and weakened PpPs phases at the stations ZHB and CHC in the TSR, while the sharp crust-mantle boundary yields strong PpPs phases at stations FEN and MIY in the YSTB. These features are also consistent with the difference of average amplitude ratio of the PpPs versus the Ps phases: that is about 0.73 for the YSTB, in contrast to 0.45 for the TSR.
 The remarkable differences of receiver function waveforms and the imaged structures are two lines of reliable evidence to divide the tectonic units. According to the waveform features (Figure 1) and the imaged structure of the crust-mantle boundary zone (Figure 2), it is inferred that the stations MDY, MIY, TST and FEN belong to the eastern tectonic unit, while SHC, ZJK, CHC and ZHB belong to the western one. The junction zone is located nearly at the northern section of Taihangshan Piedmont Fault [Xu et al., 2001], which segment the region with NNE-trending structure to the west and with EW-trending structure to the east (Figure 1). The observed receiver function waveforms for these stations imply a rapid spatial variation of the properties of crust-mantle boundary beneath this region.
 If the Moho slopes, the stacked receiver functions would show a broadening Moho conversion phase. To judge whether the distinct waveforms were derived from the dipping Moho or not, receiver functions were stacked azimuth-dependently. As examples the receiver functions stacked azimuth-dependently are shown in Figure 1 for eight stations (see auxiliary materials for all the stations.) The receiver functions from different azimuth ranges show the same features of the contrasting waveforms as those stacked from all the receiver functions. The tangential and radial receiver functions for the representative stations are displayed with back-azimuth intervals of about 10–20° in Figure S2. Lacking mirror image diagnostics of the receiver functions on back-azimuth indicates that the influence from a dipping angle of the Moho can be ignored.
4. Discussion and Conclusions
 In this study the seismic imaging documents that the contrast structural features of the crust-mantle boundary zones are inherent between the YSTB and the TSR, and the junction zone of the two tectonic units can be identified. The crust-mantle boundary pattern in the studied region is spatially correlated with that of the major volcanisms, which is characterized by the Cenozoic alkali basalts in the TSR and the Mesozoic mafic igneous province in the YSTB [Menzies et al., 2007, and references therein]. People would wonder what kind of governing factors had caused the spatial variation of the crust-mantle boundaries accompanying with the cratonic reactivation in the NCC. A number of plausible causes could explain the broadening and thinning of crust-mantle boundary. However, according to previous studies, two end-member models were most possible associated with the NCC reactivation, which include: (1) a “top-down”, rapid delamination of the mantle lithosphere induced by gravitational instability; or (2) a “bottom-up”, slow, thermally-induced asthenospheric erosion [Wu et al., 2003; Menzies et al., 2007, and references therein]. Unfortunately, it seems that neither of the two end-members could explain the contrast crust-mantle structures of the NCC alone.
 The seismic image of the YSTB (Figure 2a) reveals a sharp crust-mantle boundary, which could be attributed to the direct contact of the intruding mantle materials with the evolved higher-level crust. The geodynamic process of this phenomenon, associated with the NCC reactivation, is more likely due to the rapid foundering of the lower crust and the underlying mantle, uprising of asthenospheric mantle and replacing the volume formerly occupied by the sunken lower crust and lithospheric mantle. If this explanation is correct, it is predicted that underplating should have occurred before the foundering. On the other hand, the thicker crust-mantle transition zone revealed by the seismic image of the TSR (Figure 2b) prefers a secular interaction between the crust and the underlying mantle, and implies underplating at the base of the lower crust has occurred.
 To understand the geodynamic processes of the NCC, it is necessary to consider our findings under the background of the recent tomography, mineral physics and numerical experiments. Recent high-pressure experiments, thermodynamic and numerical calculations indicated that dehydration, partial melting, and mantle circulation might take place atop the subduction slabs [Gerya et al., 2004; Sakamaki et al., 2006; Zhao et al., 2007, and references therein]. These studies exhibit a large-scale mechanism demonstrating that the origin of upper mantle heterogeneity has both chemical and thermal contributions and is possibly associated with the deeply rooted tectonic processes.
 The eastern NCC existed as a continental margin during the extension of the Paleo-Pacific and Pacific plate since Permian. The seismological studies suggested that the subducted oceanic slab pools at the mantle transition zone [Chen et al., 2006; Huang and Zhao, 2006]. Although other trigger factors are not precluded, it was possible that the dehydration melting or/and decompressional melting can trigger a regional mantle convection beneath the eastern NCC [e.g., Niu, 2005; Zhao et al., 2007] under the tectonic framework of the eastern margin of the Eurasia continent. The tomography results also displayed low-velocity anomalies in the asthenospheric mantle above the mantle transition zone, beneath the eastern NCC [Huang and Zhao, 2006; Li et al., 2006], which could be caused by the ascending flow in mantle or by volatiles resulting from the deep dehydration reaction. Together considering the above evidence, we speculate that the secular slow convective circulation of the mantle caused by deep dehydration reaction have occurred atop the stagnant Pacific slab. As shown in Figure 3, the mafic melt and heat provided by ascending mantle could result in the thermal/chemical erosion and change the nature of the lithospheric mantle. This process might influence the formation of the crust-mantle transition zone with a thick mafic basal layer beneath the eastern NCC as displayed in the seismic imaging of the TSR, and revealed by the Mesozoic mafic xenoliths of the YSTB and the TSR [Shao et al., 2000; Chen et al., 2001]. However, in local region such as the YSTB, the foundering of orogenic thickened and/or magmatic underplating crust might disturb/overthrow the stable mantle convection, and form a sharp crust-mantle boundary. That is, the small-scale disturb within a large-scale mantle process can reconcile both the magma properties and the seismic structures in the YSTB and the TSR, which is probably compared with the geodynamic process occurred beneath the southern Sierra Nevada mountains in California [Zandt et al., 2004].
 In summary, the model in Figure 3 suggests that the change of the lithospheric mantle can be attributed to the aggregative processes including the protracted thermal/chemical erosion and underplating, and the rapid delamination. The bottom-up hydrous mantle flow originating from 410 km depth was possibly the major cause of a series of hot processes and the cratonic reactivation in the NCC. This explanation provides insight into the geodynamics of the NCC reactivation and evolution of the continental crust. However, it is worthy noted that further evidence from deep structure explorations with wider coverage in the NCC is necessary for understanding the dynamic process more comprehensively.
 We thank F. Y. Wu, S. Gao, and H. F. Zhang for helpful discussion and comments and N. P. Agostinetti and an anonymous reviewer for valuable comments and constructive criticism. We thank A. Muxworthy for his proofreading. We gratefully acknowledge the CEA for providing the CSNC data. This work was supported by the NSFC (grant 40434012) and CAS.