Tectonic evolution of the Eocene–Oligocene Lushi Basin in the eastern Qinling belt, Central China: Insights from paleomagnetic constraints

2.1 Eastern Qinling orogen

The Qinling orogen is located between the north China block and the south China block, extending over 2,000 km from east to west in Central China (Figure 1a). The Qinling orogen is traditionally divided, from west to east, into two parts along its strike, which are the western Qinling orogen and the eastern Qinling orogen [56]. The eastern Qinling orogen is delimited by the Huicheng Basin in the west and the Dabieshan orogen in the east (Figure 1a). Many NW–SE-striking rift basins (e.g., Sanmenxia Basin, Lushi Basin, Luoning Basin, and Luonan Basin) are developed in the orogen and bounded by a series of NW–SE- and WNW–ESE-striking transtensional and strike-slip faults (Figure 1a and b). The eastern Qinling orogen contains both Precambrian crystalline basement rocks and sedimentary covers of low- to medium-grade metamorphism. Based on the model raised by Dong et al. [36], the tectonics of the Qinling orogen is divided into four segments (the southern north China block, north Qinling belt, south Qinling belt, and northern south China block) by the Luonan–Luanchuan fault, Shangdan suture, and Mianlüe suture from north to south. The north and south Qinling belts accreted to the north China block during Neoproterozoic and Early Paleozoic time, respectively [36,41,56]. Subduction of the slab of the northern south China block and the amalgamated north China block finished at Triassic [36,41,56]. Then, the Qinling orogen evolved into the intracontinental evolutionary process [37]. The intermountain basins of the eastern Qinling orogen were mainly formed along the major faults (e.g., Luonan–Luanchuan fault and its branch faults) accompanied by orogen collapse and depression during the Late Cretaceous–Cenozoic [53].

Figure 1

Location and topographic and tectonic schemes of the study area and adjacent regions. (a) Location of the study area in eastern Asia (the inset map) and the topographic and tectonic schemes of the northern periphery of the eastern Qinling orogen. (b) Simplified regional geological map of the Lushi Basin. Abbreviations: DO, Dabieshan Orogen; DT, Datong Basin; HC, Huicheng Basin; HS, Hua Shan; LF, Linfen Basin; LS, Lushi Basin; LSRZ, Lingshi relay zone; QO, Qinling Orogen; SLGRZ: Shilingguan relay zone; SMX, Sanmenxia Basin; TY, Taiyuan Basin; WTS, Wutai Shan; XD, Xinding Basin; XS, Xiao Shan; YC, Yuncheng Basin; YS, Yan Shan; ZTS, Zhongtiao Shan.

2.2 Lushi Basin

As a half-graben basin, the NE–SW-striking Lushi Basin is situated on the northern margin of the eastern Qinling orogen (Figure 1a and b), with a length of ∼30 km and a width of 7–15 km. The Lushi Basin is bounded by NE–SW-striking normal faults, which are the branch faults of the Luonan–Luanchuan fault zone (Figures 1b and 2a). The bounding faults are exposed southeast of the basin, with the footwall composed of the Mesoproterozoic Xiong’er Group and the hanging wall consisting mainly of Eocene–Oligocene strata and Pleistocene loess (Figure 1b). The Lushi Basin is filled by ∼1,900 m-thick terrestrial deposits consisting of fluviolacustrine mudstones, siltstones, limestones, marlstones, dolomite, sandstones, and alluvial conglomerates, which span an age interval from the Middle Eocene to the Oligocene [53,54]. In addition, the thickness of these sediments gradually increases from north to south, and the depocenter is in the south of Lushi country [54]. The strata of the Lushi Basin are divided into three parts, from bottom to top: the Middle Eocene Zhangjiacun Formation, the Late Eocene Lushi Formation, and the Oligocene Dayu Formation, and all the formations occurred successively [52,53,54].

The Lushi Basin preserves integrated and continuous sedimentary records, which lay a foundation based upon which to decipher the tectono-sedimentary evolution of this basin [53,54]. The 80–400 m-thick Zhangjiacun Formation is an angular unconformable unit overlying the Xiong’er Group and is mainly distributed in the valleys of the Wenyu River and Suoyu River (Figure 2a). This formation consists of floodplain deposits of purple-red, brick-red mudstone, sandstone, marlstone, dolomite, and conglomerate [46,54]. The 200–400 m-thick Lushi Formation is mainly distributed in the southern basin, with the maximum thickness in the valley of the Wenyu River (Figure 2a). This formation is featured by fluviolacustrine sedimentary rocks of interbedded muddy limestone, dolomite, and calcareous mudstone in the lower part, and interbedded argillaceous limestone and calcareous sandstone in the upper part [46,54]. Similar to the distributions of the Zhangjiacun Formation and Lushi Formation, the 800–1,000 m-thick Dayu Formation is a fluvial sequence containing brown conglomerate and sandy conglomerate, with local inclusions of gray-green mudstone and sandstone [46,54]. Covered by Quaternary loess, the Dayu Formation is divided into two members and has a total thickness of ∼1,000 m in the southeastern basin.

Figure 2

(a) Detailed geological map of the investigated section and adjacent area. (b) Cross section of the Dayu section, with light gray and dark gray numbers denoting the locations of sampling sites and field pictures, respectively. (c) Field picture of the Zhangjiacun Formation in the Dayu section; the yellow line is the stratigraphic boundary between the Zhangjiacun and Lushi Formations. (d) Field image of the Lushi Formation in the Dayu section. (e) Field image of the Dayu Formation in the Dayu section.

2.3 Dayu stratigraphic section

Our field observations and sampling were primarily conducted along the ∼720 m-thick Dayu section, which is located ∼10 km southeast of Lushi country (Figure 2a). According to our field investigation, the Dayu section is made up of the monoclinal strata and outcrops of a continuous succession that contains three lithological units (Figure 2a and b). This section shows angular unconformable contact with underlying Mesoproterozoic metamorphic rocks, and the top of the section is truncated by a huge gully (Figure 2a and b). The lower lithological unit of the Dayu section is the Middle Eocene Zhangjiacun Formation, with a thickness of ∼220 m. It is dominated by thick purple-red, brick-red siltstone, pebbly coarse sandstone, conglomerate containing calcareous nodules in the lower part, and brick-red siltstone interbedded with gray-white muddy dolomite in the upper part (Figures 2b, c, and 3). Both the coarse-grained sandstone and conglomerate are poorly sorted. The gravel in the conglomerate mainly consists of quartzite, limestone, dolomite, and volcanic rock. Above the Zhangjiacun Formation, the Lushi Formation is mainly composed of gray–white dolomite interbedded with gray–green mudstone in the lower part, and interbedded with gray limestone, brick-red siltstone, and gray–green mudstone in the upper part (Figures 2b, d, and 3). The top lithological unit is the first member of the Dayu Formation, which is marked by the thick gray sandy conglomerate interbedded with brownish-red siltstone, mudstone, and sandstone (Figures 2b, e, and 3). As shown in this section, the strata from the Zhangjiacun Formation to the Dayu Formation dip southeast with approximate angles varying from ∼17° to ∼26° (Figure 2b).

Figure 3

Lithology, sampling sites of AMS, SARs, and magnetostratigraphic results for the Dayu section and correlations with the geomagnetic polarity timescale (GPTS2020) [106]. The gray circles with numbers mark the horizons of sampling sites of AMS.

2.4 Biostratigraphy

In recent decades, abundant fossils have been found in the Lushi Basin [51,52,54,57,58]. In addition, mammalian, Ostracoda, and gastropod fossils have been found in areas near the Dayu section. The formations and corresponding fossil assemblages are shown in Table 1. As early as the 1970s, Zhou et al. [57] discovered the Mengjiapo fauna, which includes Lushilagus lohoensis, Eudinoceras sp., Anthracotherjidae, Caenolapyus, Forstercooperia, Hyaenodon, Honanodon, Microlitan, and Rhinotitan in Eocene strata, and named the strata the Lushi Formation. Similar to other fauna in Late Eocene strata in Shaanxi, Yunnan, Hubei, Xinjiang, and Guangxi provinces, these mammalian fossils thus were considered Late Eocene assemblages [52,54]. In addition, the Ostracoda and gastropod fossils found in this formation are also typical assemblages in the Late Eocene (Table 1). Tong and Wang [51] analyzed mammalian fossils in the Zhangjiacun Formation and provided their representative ages as Middle Eocene (Table 1). In these mammalians, Uintatherium is a characteristic fossil widely found in Middle Eocene strata (e.g., in the Guanzhuang Formation in North China and Eocene strata in the USA) [59,60]. In the Zhangjiacun Formation, the identified Sinodarwinula guanzhangensis, Eucypris sp., Cyprinolus cf. speciosus Mandelstam, Darwinula sp. fossils are Middle Eocene Ostracoda fossil assemblages [52,54]. No mammalian, Ostracoda, and gastropod fossils were found in the Dayu Formation, whereas the pollen fossils were discovered in the Dayu Formation and Zhangjiacun Formation along the Dayu section (Table 1). Based on these identified pollen fossils and stratigraphic sequences, the Dayu Formation was thought to be deposited in the Oligocene [54].

3.1 Sampling

A total of 162 samples for AMS analysis were also collected at nine sampling sites throughout the Dayu section (from 34.068961° N, 111.170850° E to 34.054806° N, 111179342° E): three sites (DY1–DY3) from the Zhangjiacun Formation, five sites (DY4–DY8) from the Lushi Formation, and one site (DY9) from the Dayu Formation (Figures 2b and 3). We collected three or four samples from each stratum within the 5-meter thickness range to ensure that more than 15 samples were taken from each sampling site. Specifically, the samples in each stratum were distributed apart from each other within 50 cm. All samples were collected using a gasoline-powered drill and oriented with a magnetic compass in the field. Before measurements in the laboratory, all samples were cut into standard cylindrical specimens (diameter, 2.5 cm; height, 2.2 cm) with nonmagnetic tools. The samples collected from DY1–DY3 sites and DY8 are characterized by purple-red and brick-red siltstone, while those from DY4, DY5, DY6, and DY7 sites consist of gray–white dolomite and gray limestone, respectively. In addition, the samples taken from the DY9 site are marked by brownish-red medium sandstone.

3.2 Methods

4.1 Rock magnetic properties

Figure 4 shows the χ−T curves, stepwise thermal demagnetization curves of the three-axis IRM, and hysteresis loops of representative samples. DY6 and DY8 samples are selected from Lushi Formation, while DY1 and DY9 samples were collected from Zhangjiacun Formation and Dayu Formation, respectively. In the heating sections of the χ−T curves (e.g., DY1, DY6, and DY8), the most pronounced characteristic is one dramatic drop in magnetic susceptibility near 585°C, which denotes the existence of magnetite in the samples (Figure 4a). The heating curve of the DY1 sample exhibits an obvious decrease in susceptibility at 600–680°C intervals, indicating the presence of hematite (Figure 4a). In addition, the heating curve of the DY9 sample is almost flat to 700°C, which is probably attributable to the high content of paramagnetic minerals. The cooling curves of all samples show much higher susceptibility values than the heating curves (Figure 4a). Sharp increases in susceptibility values are observed for all samples when cooled below 585°C (Figure 4a), indicating the formation of ferromagnetic minerals such as magnetite during heating. In the thermal demagnetization curves of three-axis IRM, the low-coercivity fractions of the DY1 sample and all the fractions of DY6 and DY8 samples exhibit unblocking temperatures of 560–600°C, confirming the presence of magnetite, whereas the medium- and high-coercivity fractions of the DY1 sample and all the fractions of the DY9 sample mainly represent hematite with an unblocking temperature of approximately 680°C (Figure 4b). In addition, magnetite particles are also identified in the DY9 sample according to the inflection between 530 and 600°C on the curve of low-coercivity fraction (Figure 4b).

Figure 4

Rock magnetic results of representative samples from DY1 (at the depth of 623.98 m), DY6 (at the depth of 297.38 m), DY8 (at the depth of 167.01 m), and DY9 (at the depth of 109.54 m) sites, respectively. (a) Temperature dependence of the magnetic susceptibility of samples during heating (red) and subsequent cooling (blue). (b) Thermal demagnetization curves of a three-component IRM. (c) Hysteresis loops for samples before (up) and after slope correction for paramagnetic contribution (down). (d) First-order reversal curve diagrams for samples with SF values of 10, 16, 18, and 8, respectively. B _c is the coercive force, and B _u corresponds to the distribution of interaction fields.

The hysteresis loops of the four samples before and after paramagnetic correction are exhibited in Figure 4c, and the calculated χ _h/χ _l ratio is also shown. All samples showed slight hysteresis characteristics (straight line) before the paramagnetic correction and thus denoted a remarkable paramagnetic contribution. The hysteresis of all samples except the DY8 sample was apparent after the paramagnetic mineral correction, suggesting obvious contributions from ferromagnetic minerals in our samples (Figure 4c). Different from other samples, the hysteresis loops of the DY8 sample exhibit significant contributions from paramagnetic minerals. In addition, the hysteresis loops of the DY9 sample did not close, even at 500 mT (Figure 4c), which may reflect a strong contribution from high-coercivity components such as hematite and goethite [64]. Wasp-waisted behavior occurred in some samples (e.g., DY1 and DY9 samples) (Figure 4c), suggesting the presence of an admixture of low-coercivity magnetic minerals such as magnetite and antiferromagnetic minerals such as hematite [64]. First-order reversal curve (FORC) diagrams of DY1, DY6, and DY9 samples yield a central ridge almost on the B _c axis and the peak coercivity at about B _c = 15, 1, and 20 mT, respectively (Figure 4d). The vertical contours are characterized by large openings along the B _u axis, which are characteristics of multi-domain and/or pseudo-single-domain magnetite [65,66]. In consideration of wasp-waisted behaviors in DY1 and DY9 samples and the elongated contours along the B _c axis, the hematite particles are also present in these two samples. Weak signals of magnetic minerals could be recognized in the FORC diagram of the DY8 sample (Figure 4d), which may be due to the scarcity of magnetic minerals.

4.2 AMS results

The mean susceptibility (K _m) values of the Dayu section samples range from 0 × 10⁻⁶ SI to 250 × 10⁻⁶ SI, with 72% of samples less than 100 × 10⁻⁶ SI (Figure 5a and d). These results show that the Dayu section has K _m values, indicating that paramagnetic minerals dominate the magnetic susceptibility of samples. The P _J values of the Dayu section are commonly less than 1.10, while the majority of BDRS samples are less than 1.06 (Figure 5b and d). T values of the Dayu section range from −1 to 1, with a larger variation of distribution (Figure 5c and e), indicating that the AMS ellipsoids vary from moderately prolate to oblate. The K _m−P _J diagrams show that the P _J values are uncorrelated with the K _m values for the Dayu section (Figure 5d), suggesting that P _J is independent of the magnetic carriers in the sediments. The P _J–T plots of the Dayu section also show that there is completely no correlation between P _J and T values (Figure 5e). This is different from a common situation in that higher P _J values usually correspond to oblate magnetic ellipsoids [23,67]. In the F–L diagram, where the isolines of T are also plotted, it can be seen that the L values range from 1.00 to 1.03 and the F values are between 1.00 and 1.08 (Figure 6).

Figure 5

Parameters of the AMS in the Dayu section. (a) Frequency distribution of magnetic susceptibility (K _m) of samples in the Dayu section. (b) Frequency distribution of corrected degrees of anisotropy (P _J) of samples in the Dayu section. (c) Frequency distribution of shape parameter (T) of samples in the Dayu section. (d) K _m vs P _J diagram of all the samples. (e) P _J vs T diagram of all the samples.

Figure 6

Flinn diagram of all the samples.

The equal-area projections of the AMS ellipsoids of samples are shown in Figures 7a, c, e, 8a–d, and 9a. The overwhelming majority of P _J values are lower than 1.04, whereas T values are primarily larger than 0, except those of DY1 and DY2 sites (Figures 7b, d, f, 8f, and 9b), implying the magnetic ellipsoids are mainly oblate shapes. For the Zhangjiacun Formation (DY1–DY3 sites), the maximum principal anisotropy axes (K ₁) concentrate in the NW–SE direction in the geographic coordinate, and the K ₁ is also NW–SE oriented, lying at a smaller angle with the bedding plane in the paleogeographic coordinate (the mean K ₁ directions are D/I = 323.1° ± 9.5°/1.3° ± 6.5°, 141.9° ± 25.4°/4.0° ± 6.2°, and 154.9° ± 50.8°/28.8° ± 16.5° for DY1–DY3 sites, respectively) (Figure 7a, c, and e). The minimum principal anisotropy axes (K ₃) of these samples are relatively scared and oblique to the bedding plane in both the geographic and paleogeographic coordinates (Figure 7a, c, and e). For Lushi Formation (DY4–DY8 sites), the K ₁ of samples are NE–SW oriented in both the geographic and paleogeographic coordinates (the mean K ₁ directions are D/I = 237.5° ± 33.7°/30.6° ± 20.6°, 226.4° ± 39.3°/12.8° ± 18.4°, 214.0° ± 12.7°/41.8° ± 10.9°, 243.5° ± 30.3°/25.4° ± 8.1°, and 73.7° ± 30.0°/26.7° ± 23.3° for DY4–DY8 sites, respectively), whereas the K ₃ of samples are dispersive and also at middle angles with the bedding plane in both coordinates (Figure 8a–e). These distributions suggest that the magnetic fabrics of the Zhangjiacun Formation and Lushi Formation are formed by neither sedimentary processes nor tectonic deformation. For the Dayu Formation, the K ₁ is generally subhorizontal to the bedding plane and well defined in an NW–SE direction (the mean K ₁ direction is D/I = 337.5° ± 36.6°/4.6° ± 5.7°) in the paleogeographic coordinates, while the K ₃ is approximately orthogonal to the bedding plane (the mean K ₃ direction is D/I = 150.8° ± 8.6°/85.3° ± 4.8°) (Figure 9a).

Figure 7

Equal-area projections of AMS for the samples of DY1 (a), DY2 (c), and DY3 (e) sites, and P _J vs T plots (right) for the samples of DY1 (b), DY2 (d), and DY3 (f) sites from the Zhangjiacun Formation.

Figure 8

Equal-area projections of AMS for the samples of DY4 (a), DY5 (b), DY6 (c), DY7 (d), and DY8 (e) sites, and P _J vs T plots for the samples of DY4–DY8 sites (f) from the Lushi Formation.

Figure 9

Equal-area projections of AMS for the samples of DY9 site (a), and P _J vs T plots for the samples of DY9 site (b) from the Dayu Formation.

5.1 Source of the AMS

The χ−T curves, the three-axis IRM measurements, and the hysteresis loops reflect that magnetite and hematite are the main magnetic carriers of remanence in the Zhangjiacun Formation and Dayu Formation, whereas magnetite is the dominant magnetic mineral in the Lushi Formation (Figure 4). However, the significant paramagnetic content of samples can be inferred from the hysteresis loops (Figure 4c), while the gradual decrease or slight changes below ∼500 °C shown in χ−T heating curves of four samples (Figure 4a), obviously indicate the remarkable contribution of paramagnetic minerals as well. In addition, the susceptibility of sediments in the Lushi Basin has a range of <250 × 10⁻⁶ SI (Figure 5a and d), suggesting that the bulk susceptibility is primarily carried by the paramagnetic fraction [63]. The AMS parameters (e.g., P _J) are independent of ferrimagnetic content (Figure 4b), further indicating that the AMS of the Dayu section is dominated by the anisotropy of paramagnetic minerals with some local contribution of hematite and magnetite. The AMS is thought to be derived from magnetocrystalline anisotropy and shape anisotropy [23,63,67]. Paramagnetic minerals and hematite have intrinsic magnetocrystalline anisotropy [18,63,68], while the anisotropy of magnetite originates from both magnetocrystalline and shape anisotropy [69,70] and is primarily influenced by the shape and grain size (magnetic domain) [18,26]. Therefore, the AMS of the Dayu section reflects the preferred crystallographic orientation.

5.2 Cause of the anomalous AMS

According to the ternary diagram of inclinations of the principal anisotropy directions [27], both the normal and anomalous magnetic fabrics are found in the Dayu section (Figure 10). The anomalous magnetic fabrics may be due to the mixing of normal, inverse, and intermediate fabrics or strong strain intensity [17,27]. The second case can be ruled out by the low P _J values and lack of strain markers at the outcrop scale. As noticed by some studies, the intermediate fabrics can result from a few iron-bearing silicate minerals (e.g., orthopyroxene, riebeckite, staurolite) [71,72], but most commonly, mixing of normal and inverse fabrics is the main controlling factor [26,27,30]. According to our field investigation, there is little orthopyroxene, riebeckite, and staurolite in the strata. The larger variation of T values of DY1–DY8 sites (Figures 7b, d, f, and 8f) could result from the mixture of normal and inverse fabrics [30]. Hence, the anomalous magnetic fabrics of the Zhangjiacun and Lushi Formations may arise from the mixing of normal and inverse fabrics. Many magnetic minerals are reported to be responsible for the inverse magnetic fabrics, such as iron-bearing carbonates, tourmaline, cordierite, goethite, and single domain magnetite [72,73,74,75,76,77,78]. Our rock magnetic data only show the presence of hematite and magnetite in samples, and the dominant remanence carriers of the Lushi Formation are multidomain and/or pseudo single domain magnetite (Figure 4). Moreover, the Zhangjiacun and Lushi Formations consist of massive dolomite, limestone, and calcareous nodules (Figures 2b–e and 3), implying the potential effect of iron-bearing carbonates. Thus, the anomalous fabrics of the Zhangjiacun and Lushi Formations may arise from the iron-bearing carbonates.

Figure 10

Ternary diagram relating the inclination angles of the principal anisotropy directions (K ₁, K ₂, and K ₃) in the paleogeographic coordinate system. Samples of (a) Zhangjiacun Formation, (b) Lushi Formation and (c) Dayu Formation with normal and inverse magnetic fabrics fall in the areas of the lower left and right triangles, respectively, whereas the intermediate fabrics fall in the upper triangle area. Samples plotted in the central hexagonal area could be regarded as anomalous [27]. Different types of magnetic fabrics are defined by Rochette et al. [72], and Ferré [26]. Square, triangle, and circle symbols represent K ₁, K _2, and K ₃, respectively.

5.3 Tectonic implications

The AMS residing in paramagnetic minerals and hematite has been reported to be controlled by the sedimentation process and/or deformation (e.g., tectonic extension/compression) [1,3,4,6,8,10,17,18,79]. If the rocks merely undergo sedimentation, the initial shapes of AMS ellipsoids are controlled by the flow regime and gravitational forces [80]. The K ₃ direction is almost perpendicular to the bedding plane in a quiet water environment, while it should be slightly off-vertical and streaked in a moderate current environment and high flow environment, respectively [81]. In addition, the K ₁ orientation is approximately parallel or vertical to the flow direction depending on the flow velocity [81]. Tectonic deformation processes will progressively overprint the primary sedimentary magnetic fabric according to the degree of deformation and form a tectonic magnetic fabric [22,23,24,25]. In the compressive tectonic settings, the development of AMS has been divided into three stages (initial deformation magnetic fabric, pencil structure magnetic fabric, and cleavage magnetic fabric) [23] or six stages (primary sedimentary fabric, initial deformation magnetic fabric, pencil structure magnetic fabric, weak cleavage magnetic fabric, strong cleavage magnetic fabric, and stretching lineation magnetic fabric) [22]. Nevertheless, in the extensional settings, the K ₁ direction is commonly parallel to the maximum extension direction with the K ₃ orientation vertical to the bedding plane [1,3,15,18].

For the DY9 sampling site, the pronounced K ₁ direction and the general perpendicularity of the K ₃ direction to the bedding plane suggest that the AMS is related to either the sedimentation process or the initial deformation. If the AMS is controlled by the deposition process, the K ₃ distribution should be streaked and slightly off-vertical to the bedding plane in high and moderate paleocurrent environments, respectively [81]. These cases could be ruled out based on the distribution of K ₃ in the DY9 sampling site (the K ₃ direction is D/I = 150.8° ± 8.6°/85.3° ± 4.8°) (Figure 9a). Given the low PJ values in sediments, the AMS of the DY9 site in the Lushi Basin is not the sedimentary fabric but an incipient deformation magnetic fabric. The tilt-corrected K ₁ is subparallel to the dip of the NE–SW-striking bounding fault (the attitude is D/I = 335°/42°) (Figure 9a), which conforms to the situation that the Lushi Basin is developed in a local extensional setting. The tilt-corrected K ₁ of the DY9 site is well grouped in the NW–SE direction, denoting the NW–SE extension direction during the Late Eocene–Earliest Oligocene. This NW–SE extension of the Lushi Basin generally corresponds to the NW–SE extension in eastern Qinling orogen suggested by fault kinematic analysis during Eocene–Early Miocene [46,82].

We displayed SARs based on our magnetostratigraphic study of Lushi Basin sediments (K. Jiang, unpublished data) and identified five stages with different SARs from the upper part of the Dayu section (Figure 3). Figure 11 shows three periods of low SARs (69.7 m Ma⁻¹ at 0–167 m, 49.3 m Ma⁻¹ at 230–268 m, and 29.6 m Ma⁻¹ at 308–332 m) and two periods of high SARs (258.2 m Ma⁻¹ at 167–230 m and 180.2 m Ma⁻¹ at 268–308 m) in the Lushi Basin. In the first four stages, the strata are dominated by limestone, mudstone, and siltstone, which represent shore–shallow lake facies. Nevertheless, the last stage displays slow sedimentation of the basin with massive fluvial coarse-grained sandstone and conglomerate developed. All of these sedimentation processes in the Lushi Basin suggest the shrinkage of a lake. In general, the SARs of terrestrial basins are dominated by climatic and tectonic variations that affect the supply of sources and the erosion and transportation of original materials from the sources [83,84,85]. Benthic δ¹⁸O results have been extensively used as a reliable proxy with which to record global paleoclimate variations [86,87]. A warmer and more humid climate could lead to a high-energy current environment and an increase in the sedimentation rate [88]. The δ¹⁸O data indicated that the global climate is generally stable before 34 Ma (Figure 11), which was inconsistent with the variations in the SAR. The SAR of the Dayu section more likely arose from tectonic movements on account of the inconsistent relationship between sedimentation and global climate variation.

Figure 11

SAR changes referred from magnetostratigraphic results in the upper Dayu section. The marine oxygen isotope results are from Zachos et al. [86].

Two rapid subsidence processes (∼36.6 to ∼35.3 Ma) of the Lushi Basin were observed in the upper part of the Lushi Formation (Figures 3 and 11). These two accelerated pulses of sediment input correlated with the rapid exhumation of the northern Qinling orogen (e.g., rapid uplift of Hua Mountain in 35–25 Ma) [39,89]. The accelerated subsidence at ∼35 Ma has also been determined in the Sanmenxia Basin of eastern Qinling orogen [90]. In addition, the northern and eastern margins of the Tibetan Plateau were rapidly uplifted accompanied by the subsidence of basins in the northeastern Tibetan Plateau at Late Eocene–Early Oligocene [91,92,93,94,95], which were coeval with the high SARs of Lushi Basin. The bounding faults of the Lushi Basin, which were inherited from the main terrain boundary structures of the Qinling orogen, were activated to be normal with a sinistral strike-slip component contemporaneous with the uplift of the northern Tibetan Plateau [40,46]. Thus, we raise the possibility that the tectonic evolution of the Lushi Basin in the Late Eocene–Earliest Oligocene was influenced by the uplift in the northern Tibetan Plateau, which may be the result of the India–Eurasia collision (Figure 12).

Figure 12

Schematic model showing the relationship between western Pacific plate subduction, India–Eurasia collision, and the tectonic evolution of the Lushi Basin during the Late Eocene–Earliest Oligocene. The purple dotted line in the upper picture denotes the boundary of east China’s extensional domain, which is caused by the subduction of the Pacific plate. The crustal deformation of northeastern Tibet is linked with the west Pacific plate subduction by a sinistral strike-slip fault system along the eastern Qinling orogen.

Nevertheless, NW–SE extension affected a large part of eastern China (e.g., Bohai Bay Basin) during the Eocene–Oligocene [46]. The tectonic evolution of the Lushi Basin under this extension was probably attributed to the northwestward subduction of the western Pacific Plate and the retreat of the trench during the Eocene–Oligocene [40,46,47], during which decreased Pacific–Eurasia convergence associated with the rollback of the subducting slab resulted in back-arc rifting in eastern China (e.g., the Songliao Basin and Bohai Bay Basin) [96,97,98]. The seismic tomographic model also indicated that the western edge of the subducting Pacific plate was roughly coincident with the surface topographic boundary (Taihang Shan) in east China [99,100]. The spatial coincidence between the deep-seated subducting slab and the Lushi Basin during the Eocene–Oligocene suggests that the basin formed under back-arc extension caused by asthenosphere upwelling (Figure 12). Deep geophysical observations reveal that the eastward extrusion of the Tibetan asthenosphere along the southern margin of the Ordos block and eastern Qinling orogen during the India–Eurasia collision is feasible [101,102,103]. In addition, the surface pattern of crustal deformation is well coupled with geophysical observations, suggesting that the Late Paleogene sinistral fault system along the eastern Qinling orogen could be a lithospheric corridor linking the eastward asthenospheric flow of Tibetan with back-arc extension in east China [47]. Given the potential interaction of Pacific subduction and India–Eurasia collision [47,104,105], we suggest that the tectonic evolution of the Lushi Basin was complicated and may be dominated by the combined effects of western Pacific subduction and India–Eurasia collision during the Late Eocene–Earliest Oligocene. According to the magnetostratigraphic investigation of the Sanmenxia Basin, which lies in the eastern Qinling near the Lushi Basin, the influence of western Pacific subduction and India–Eurasia collision were considered to dominate the tectonic evolution of the basin during the Eocene (∼54 to ∼34 Ma) and during the Oligocene–Neogene (∼34 to ∼3 Ma), respectively [90]. Therefore, the western Pacific subduction may be the dominant factor in the tectonic evolution of the Lushi Basin during the Late Eocene–Earliest Oligocene.