Constraints on mountain building in the northeastern Tibet: Detrital zircon records from synorogenic deposits in the Yumen Basin

The Cenozoic basins and ranges form the high topography of the northeastern Tibet that resulted from the India-Eurasia collision. Sedimentary rocks in the basins provide direct insight into the exhumation history of the ranges and the tectonic processes that led to the northeastward growth of the Tibetan Plateau. In this study, we analyzed and compared detrital zircon U-Pb ages from sands of modern rivers draining the Bei Shan, and North Qilian Shan and sandstones from the Yumen Basin. The zircon age distributions indicate that the strata dated to 24.2-16.7 Ma in the basin were derived from the Bei Shan, and the basin provenance changed rapidly to the North Qilian Shan terrane at ~16 Ma. These results suggest that an early stage of deformation along the Bei Shan at ~24 Ma was replaced by the growth of the North Qilian Shan at ~16 Ma. We conclude that the far-field effect associated with the Indo-Asian collision may result from Oligocene deformation in the Bei Shan, but the emergence of the North Qilian Shan at ~16 Ma could reflect the most recent outward growth of the Tibetan Plateau that may have been caused by the removal of some lithospheric mantle beneath central Tibet.

We selected one of the most complete and continuous stratigraphic sections along the northern margin of the basin, from which to collect detrital zircon samples (Fig. 1B). This section, known as the Caogou section, was dated using high-resolution magnetostratigraphic dating and was deposited between ~24.2 Ma and ~2.8 Ma 17 . Six fine-to coarse-grained sandstone samples were collected from the section and from sediments dated to 23 Fig. S1). In addition, we collected four sand samples from modern rivers (Fig. 1A) to establish distinctive U-Pb age signatures of the North Qilian Shan and the Bei Shan. Although the density of modern sand samples is not large enough to account for all of the streams draining the North Qilian Shan and the Bei Shan, the sampled rivers cross most lithologic units of both ranges that were expected to yield zircons.

Methodology
Five kilograms of sand or sandstone was collected for each sample from a single outcrop for detrital zircon analysis. Zircon crystals in these samples were isolated using a combined method with heavy liquids and a magnetic separator; the analyzed grains were hand picked at random and mounted in epoxy resin. Measurements for the U-Pb analysis of individual detrital zircon grains were carried out by LA-ICP-MS spectrometry, and the equipment is housed at the Institute of Geology and Geophysics, Chinese Academy of Sciences. Details of the analytical procedures were previously described by Xie et al. 18 .
The results and fractionation correction of the detrital zircons were calculated using GLITTER 4.0 software. We interpreted the U-Pb ages based on 206 Pb/ 238 U for grains younger than 1000 Ma, and 207 Pb/ 206 Pb for grains older than 1000 Ma. In general, 206 Pb/ 238 U ages are more precise for younger ages, whereas 206 Pb/ 207 Pb ages are more precise for older ages. All of the analytical results are listed in the Supplemental Table 1. Although all zircon U-Pb ages are presented, only the zircon ages with ≤ 25% discordance or ≤ 10% reverse discordance were used to build the age-probability diagrams. In this paper, the detrital zircon age populations for individual samples are plotted on relative age-probability diagrams derived from the probability density function 19 .
Paleocurrent indicators were also measured from the Caogou section and from other outcrops where such features were accessible. Post-depositional tilting and vertical axes rotation of these paleocurrent indicators were corrected for tilted sediment bedding and rotations as indicated by the paleomagnetic data 17 . The combination of zircon grain ages with general paleocurrent orientations in the Yumen Basin allows for identification of the predominant source of the basin sediments and the unroofing history of the adjacent ranges.

Results
Samples BS2 and BS3 were collected from two parallel rivers draining the Bei Shan (Fig. 1A). They are characterized by a U-Pb age population of 230-310 Ma, which constitutes 75% and 85% of the BS2 and the BS3 distributions, respectively. A few zircon ages at ~430 Ma were also recorded (Fig. 2). In the Bei Shan, the late Paleozoic and early Mesozoic granite plutons crop out widely, especially along its southern margins (Fig. 1A). Therefore, the 230-310 Ma population in the catchments definitely was eroded from these granite pluton source terranes. The ~430 Ma zircons may have eroded from a small area of early Paleozoic magmatic rocks present in the Bei Shan.
Samples QL1 and QL2 were collected from the SY and BY Rivers draining the North Qilian Shan (Fig. 1A). Four major age populations dominate the QL1 detrital zircon grains; there are peaks at 260-300 Ma, 415-485 Ma, 1640-2030 Ma, and 2320-2550 Ma (Fig. 2). In sample QL2, aside from the previously mentioned four age populations, 30% of the zircon grains date to 730-850 Ma, forming a fifth population (Fig. 2). Zircons yielding U-Pb ages at 1640-2030 Ma and 2320-2550 Ma are two unique zircon populations, which exhibit similar ages to the Proterozoic rocks cropping out along the modern northern Qilian Shan crest, suggesting these two groups of zircons were eroded from a Proterozoic source terrane in the North Qilian Shan 20 . These two populations in these proportions have not been identified in the Bei Shan. The other zircon populations of 260-300 Ma, 415-485 Ma, and 730-850 Ma may have been eroded from the corresponding plutons or strata in the North Qilian Shan.
Six samples from the Caogou section were dated, yielding more than 85 effective U-Pb dates for each of the samples. Zircons from 23.8 Ma (FZ1) and 20.6 Ma (FZ2) strata are dominated by age probability peaks at ~275 Ma, which account for > 83% of the total dated grains. Ages older than 310 Ma only comprise ~15% of the population for both samples ( To compare the similarity or dissimilarity of the detrital zircon age distributions, the Kolmogorov-Smirnoff (K-S) test was used. Two age distributions that returned high P values from the K-S test implies that the two age spectra are nearly identical. If the P < 0.05, then there is > 95% level confidence that their age distributions are unlikely to have been derived from the same parent population.   FZ1-FZ5, FZ7).
supported by the south-directed paleocurrent indicators shifting to north-directed ones since this time interval (Fig. 1B). The appearance of north-directed paleocurrent indicators provides independent evidence that North Qilian Shan-derived material entered the Yumen Basin during the middle Miocene.

Implications and Discussion
Our detrital zircon data, combined with paleocurrent measurements, holds important information about the tectonic evolution of the northeastern Tibetan Plateau. Most of the clasts in the 24.0-16.7 Ma strata in the Yumen Basin from the Bei Shan sources indicate significant exhumation along the Bei Shan beginning at ~24 Ma. We interpreted this exhumation as a response to the onset of the rock uplift of the Bei Shan caused by tectonic deformation in the remote region of the northeastern Tibet. Although climate change can also trigger exhumation along mountain belts 21 , the coeval subsidence of the Yumen Basin along its southern margin provides strong evidence that the rock uplift and subsequent exhumation along the Bei Shan were driven by the tectonics in the region.
Our results do not support the hypothesis that the North Qilian Shan was deformed to form relatively high topography on the northeastern Tibetan Plateau before the middle Miocene, unlike the onset of the uplift of the Bei Shan at ~24 Ma. First, the absence of any detrital zircon grains from the North Qilian Shan in the pre-16 Ma strata is the most remarkable contrast in the detrital zircon age distributions (Fig. 2). The absence of these grains requires that the North Qilian Shan had relatively lower relief during the late Oligocene to early Miocene and was presumably tectonically inactive in the southern part of the basin. Further south of the Qilian Shan, the Qaidam Basin initiated as a result of thrust faulting-related subsidence along its northwestern margin during the Paleocene or Eocene and a wedge of conglomerate filled in the accommodation created by the subsidence 9,22,23 , suggesting that early Cenozoic high topography might lie in the North Qaidam-South Qilian Shan region. Second, the remnants of mid-Tertiary sediments are widespread in the North and Central Qilian Shan area without marginal facies 24 . Assuming that these sediments were part of the Hexi Corridor Basin, the North Qilian Shan seems to be covered by the southern extended Hexi Corridor Basin, again implying no significant deformation took place in the North Qilian Shan region.
A key result of the detrital zircon records is that the provenance of the Yumen Basin sharply changed from the Bei Shan to the North Qilian Shan at ~16 Ma. This result requires the rapid growth of the North Qilian Shan but the relatively steady uplift of the Bei Shan. Following the growth of the North Qilian Shan, the clasts eroded from the newly formed high topography of the North Qilian Shan replace the clasts from the Bei Shan forming the basin deposits. Thus, our results directly constrain the timing of the emergence of the North Qilian Shan to ~16 Ma, similar to the ages obtained from recent sedimentology and low-temperature thermochronology studies 11,[25][26][27][28] .
The Cenozoic uplift histories of the Bei Shan and the North Qilian Shan challenge previous studies that infer the tectonic deformation in the Qilian Shan region was related to the systematic growth of the ATF northeastward since the late Miocene or the Pliocene 10,29 . Our results reveal poorly-known late Oligocene deformation in the outermost northeastern Tibetan Plateau, followed by the emergence of the North Qilian Shan in the mid-Miocene. During the late Oligocene, deformation was expressed as the rapid exhumation of the Bei Shan and subsidence of the Hexi Corridor Basin in the north, but the Qilian Shan, especially the North Qilian Shan was relative quiescent in the south. This observation indicates that late Oligocene deformation occurred prior to the ATF reaching this part of the plateau, which occurred relatively late in the history of the plateau. Although the geodynamics of the Oligocene deformation are not completely understood, it is consistent with the appearance of ranges and the formation of foreland basins along the northern and eastern margins of the Tibetan Plateau 30,31 , implying that stresses caused by the collision between India and Eurasia has been transferred to present-day northeastern Tibet since the late Oligocene. Thus, regardless of the exact mechanism, the tectonic processes along the Bei Shan-Hexi Corridor region may reflect far-field deformation associated with the Indo-Asian collision.   The mid-Miocene emergence of the North Qilian Shan could directly represent the most recent outward and upward growth of high topography in northeast Tibet. Yue et al. 32 and Bovet et al. 28 attribute the mid-Miocene uplift of the North Qilian Shan to the kinematic change along the ATF during the mid-Miocene. They suggest that the previous fast strike-slip movement along the ATF in northeast Tibet slowed down abruptly in the mid-Miocene, and renewed reverse faulting within the Qilian Shan region transferred the fault slip from the AFT to shortening and range growth. However, numerous studies have yielded evidence for emergence of the Qilian Shan occurring almost simultaneously along its southern to northern parts at 16-10 Ma, such as magnetostratigraphic dating of sediments to derive sediment accumulation rates or provenance changes in the Qaidam Basin 33 , Subei Basin 26,27 , Hexi Corridor Basin 17,28 ; and low-temperature thermochronology dating of detrital minerals to define rapid exhumation of the adjacent ranges 11,25 (Fig. 3). With the exception of the Qilian Shan, middle Miocene range growth and deformation appear to be widespread across the northern Tibetan Plateau. For example, thermochronologic data reveal that the onset of left-lateral strike slip motion on the Kunlun Fault initiated 20-15 Ma, the Haiyuan Fault initiated 17-12 Ma and rapid exhumation of the Dulan-Chaka highlands occurred at ~15 Ma 34 . Foraminifera assemblages contain planktonic taxa coeval with the apatite fission track, and (U-Th)/He thermal models clearly show that the Altyn Tagh has undergone significant exhumation and surface uplift from sea level to its present elevation of 1400 m since ~15 Ma 35 . Moreover, sedimentology studies constrained by paleomagnetism document changes in provenance in the Linxia Basin at ~14 Ma 36 , the Xunhua Basin at ~22 Ma 37 , sediment accumulation rate increases in the Sikouzi Basin at ~11 Ma 38 , in the Xunhua Basin at ~13 Ma 37 , the clockwise rotation of the Guide Basin at 17-11 Ma 39 and the formation of the Wushan pull-apart Basin at ~16 Ma 40 , indicating deformation occurred throughout the region during the mid-Miocene (Fig. 3). Such widespread synchronous activity supports an alternative interpretation that convective removal of some gravitationally unstable mantle lithosphere beneath north-central Tibet could trigger or accelerate deformation on the plateau margins 41 . The mantle lithosphere removal could augment the potential energy per unit area and then transfer the deformation to the plateau surroundings. In the northeastern Tibet, the Qaidam Basin is a relatively rigid block that is efficient at transferring deformation north and east to the Qilian Shan and elsewhere. During this process, the ATF may act as a transfer fault to accommodate deformation through northeastward propagation, and the termination of the ATF in the Yumen Basin may delimit the boundary of deformation caused by mid-Miocene mantle removal. This may also be the reason why the Bei Shan was inactive after the mid-Miocene.

Conclusion
The different provenance histories of the sediments in the Cenozoic Yumen Basin revealed by the detrital zircons and paleocurrent measurements provide a new perspective on the deformation along the outermost region of the northeastern Tibetan Plateau. The zircon grains from the 24.2-16.7 Ma strata in the basin are dominated by the Bei Shan plutons and indicate the late Oligocene uplift of the Bei Shan that may reflect the far-field deformation related to the Indo-Asian collision. The dramatic provenance change at ~16 Ma is highlighted by the North Qilian Shan sources, suggesting the onset of the rapid uplift of the range at this time as a result of the decreased slip rates on the ATF or in response to the removal of lithospheric mantle beneath north-central Tibet.  34 and Ritts et al. 35 . Left-lateral strike-slip fault initiation ages were derived from Yuan et al. 6 and Duvall et al. 34 . Changes in provenance /accumulation rates and basin formation/rotation were derived from Wang et al. 17 , Wang et al. 26 , Lin et al. 27 , Bovet et al. 28 , Fang et al. 33 , Garzione et al. 36 , Lease et al. 37 , Wang et al. 38 , Yan et al. 39 , Wang et al. 40 . The Fig. 3 was created by the Global Mapper 13 (http://www.bluemarblegeo.com/products/ global-mapper-download.php) and CorelDRAW X7(http://www.coreldraw.com/us/pages/free-download).