Seismological constraints on the crustal structures generated by continental rejuvenation in northeastern China

Crustal rejuvenation is a key process that has shaped the characteristics of current continental structures and components in tectonic active continental regions. Geological and geochemical observations have provided insights into crustal rejuvenation, although the crustal structural fabrics have not been well constrained. Here, we present a seismic image across the North China Craton (NCC) and Central Asian Orogenic Belt (CAOB) using a velocity structure imaging technique for receiver functions from a dense array. The crustal evolution of the eastern NCC was delineated during the Mesozoic by a dominant low seismic wave velocity with velocity inversion, a relatively shallow Moho discontinuity, and a Moho offset beneath the Tanlu Fault Zone. The imaged structures and geochemical evidence, including changes in the components and ages of continental crusts and significant continental crustal growth during the Mesozoic, provide insight into the rejuvenation processes of the evolving crust in the eastern NCC caused by structural, magmatic and metamorphic processes in an extensional setting. The fossil structural fabric of the convergent boundary in the eastern CAOB indicates that the back-arc action of the Paleo-Pacific Plate subduction did not reach the hinterland of Asia.


B. Results of waveform inversion
The velocity models used to calculate synthetic receiver functions were obtained by waveform inversion. The final best-fitting shear wave velocity models for all 60 stations are shown in Figure S1. The corresponding synthetic receiver functions calculated from the inverted velocity models for each station are also shown superposed upon the data in which the excellent match between the synthetic and observed waveforms was displayed for most of stations.

Figure S1
Shear-wave velocity models and comparisons between synthetics and observed receiver functions for all the 60 stations. For each station the best-fitting shear-wave velocity model is plotted in the right panel, the receiver functions are  We ranked the stacked receiver functions from data and the synthetics along the profile as shown in Figure S2 and compared their waveform variations. The waveforms of the synthetic receiver functions are in excellent agreement with the data. Major structural characteristics could be coherently detected beneath most of the stations. The velocity models of stations 17 and 19, which are located at both sides of the Tanlu Fault, are shown in Figure S3 for comparing.

C. Synthetic tests of the CCP images
The uncertainty in detecting a seismic discontinuity within the crust primarily arises from disturbances in the sedimentary cover and upper crust. We implemented a series of synthetic tests to determine whether a seismic signal in the CCP image represents an artifact or an actual velocity discontinuity. For example, when observing the effects of multiples from the sedimentary cover, we calculated the synthetic receiver functions by constructing models with a lower interface for the upper crust, deepening to 45 km beneath the inverted sedimentary cover models. As shown in Figure S4b, we can find the larger amplitudes of converted waves from the surface to a depth of more than 10 km, even though the sedimentary cover spans depths of 1-6 km. Obviously, the singles below the sedimentary cover, which are generated by the multiples, cannot be identified as velocity discontinuities. In observing the effects of the multiples from the upper crustal interface, we calculated the synthetic receiver functions by constructing models with a lower interface for the middle crust, deepening to 45 km beneath the inverted upper crustal models, and produced synthetic CCP image (Fig. S4c). Comparing the synthetic CCP image with the observation CCP image (Fig. S4a), we can clearly see that signals from multiples appeared throughout the entire crust. Hence, the signals that appeared in the observation CCP image but not in the synthetic CCP image should be considered as the crustal interfaces. The structural framework of the waveform inversion was then iteratively adjusted following the synthetic CCP test. Because the stacking profile in depth domain is constructed by time-to-depth conversion based on a defined velocity model, it is necessary to carefully assess the ability of stacking-based CCP imaging to constrain the depths of discontinuities in a region with thick sedimentary cover. Figure S5a shows the CCP depth image constructed using an average crustal model for the NCC, in which no sedimentary structure is included. The resultant CCP image (Fig. S5b) was constructed using the imaging models. As shown in Figure S5a, the depth distributions of Moho (lower boundary of the crust-mantle transition zone, blue line) beneath the Bohaiwan (Xialiaohe) basin and the Sonliao basin are apparently deeper than our seismic imaging result (green line). In Figure S5c, we compared the Moho image with the previous estimation from seismic refraction observation of the Donggou-Dongwuqi Geoscience Transect 1 (red dashed line). In Figure S5d, the geologic sections from oil fields near the NCISP-6 profile 2-4 are superposed on our velocity profile. The coincidence of sedimentary structure image and geological profile verifies the reliability of our receiver function imaging results in the basin areas.

D. Reliability analysis of the receiver function imaging
The receiver function imaging technique used in this study involves the synthetic test of CCP images and waveform inversion. The reliability analysis was carried out for the depths of major interfaces and the waveforms of receiver function. The interface depths of each station are listed in the Table S2, in which the observation depths were measured by the local maximum amplitudes of observation CCP image, and the synthetic depths were obtained from waveform inversion. The depth errors at a 90% confidence interval were less than 0.78 km along the upper boundary of the crust-mantle transition zone, and less than 1.4 km within the crust. The standard derivations of the interface depths were 0.38 km along the upper boundary of the crust-mantle transition zone, 0.48 km and 0.71 km within the crust, respectively.
In the global inversion, the best-fitting between the synthetic receiver function Y(t) and observed one O(t) is searched by minimizing the objective function 5-6 This OBJ measures the degree of fitness of waveforms and amplitudes between O(t) and Y(t). The OBJ of each station are listed in the Table S3