Linking source and sink: The timing of deposition of Paleogene syntectonic strata in Central Asia

Because correlation of paleomagnetic data from continental basins to the Geomagnetic Polarity Time Scale (Ogg, 2020) is commonly non-unique, magnetostratigraphy alone can lead to dramatically different age models for continental siliciclastic sequences in the absence of fossils or radiometrically datable volcanic ash layers (Lowe, 2011). This inevitably leads to contrasting models for the timing, rates, and duration of tectonic and paleoclimatic processes. This fundamental challenge is well exemplified in Cenozoic terrestrial strata in Central Asia (Figs. 1A–1E), where competing age models have strongly diverging implications for the growth of the Tibetan Plateau (Ji et al., 2017; Staisch et al., 2020; Wang et al., 2022) and its association with Asia paleo-environments including the retreat of the Proto-Paratethys Sea, a major water body that covered large surfaces of Eurasia during the Paleogene (Bosboom et al., 2017; Sun and Liu, 2006; Zheng et al., 2015).

The two largest terrestrial basins in the Cenozoic Tibetan orogen are the Tarim and Qaidam Basins, which contain critical archives of mountain building and paleoclimate (Fig. 1B). The growth strata and thick-bedded conglomerates in the Lulehe Formation, the basal stratigraphic unit of Cenozoic strata in the Qaidam Basin (Fig. 1D), are interpreted as synorogenic sediments that record the initiation of mountain building in Northern Tibet in response to the ca. 60 Ma India-Asia collision (Ding et al., 2022; Yin et al., 2008). However, due to a lack of directly datable geologic markers and the scarcity of vertebrate fossils, two strongly contrasting age models, with a basal age of either ca. 50 Ma (Ji et al., 2017) or ca. 30 Ma (Wang et al., 2022), have been proposed for the depositional age of the Lulehe Formation, resulting in competing models for the lateral growth history of the Tibetan Plateau (Staisch et al., 2020; Wang et al., 2022; Yin et al., 2008). A similar debate centers around the depositional age of Cenozoic strata in the Tarim Basin, where some have proposed a Pliocene age for the Artux Formation (Sun and Liu, 2006) but others have assigned ages of ca. 27–15 Ma for the same unit (Zheng et al., 2015) (Fig. 1E). This differing age assignment on the eolian- and gypsum-bearing Artux Formation has led to a fundamental debate about the timing of aridification in Central Asia (Licht et al., 2016; Liu et al., 2014; Sun and Liu, 2006). In addition, knowledge about the exact timing of deposition of Lulehe Formation affects interpretations on how the Proto-Paratethys Sea retreated permanently from Central Asia (Bosboom et al., 2017; Ma et al., 2022) (Fig. 1B) and in turn impacted the regional climate and biodiversity (Barbolini et al., 2020; Meijer et al., 2019). The latter regression was attributed to the combined effect of eustatic fluctuations and far-field tectonics in response to the India-Asia collision (Bosboom et al., 2017; Burtman and Molnar, 1993; Dupont-Nivet et al., 2007; Kaya et al., 2019). However, given strikingly different interpretations depending on the age model for the deposition of syntectonic strata in the Qaidam Basin, it remains elusive whether Paleogene tectonism along Northern Tibet impacted regression in the areal extent of the Proto-Paratethys Sea.

Here we present a simple yet novel approach to assessing age models in tephra- and fossil-poor strata by linking source to sink of such sediments and examining the temporal relationship between rapid basement exhumation and syntectonic sedimentation. Applying this approach to the Tarim and Qaidam Basins, we constrain the depositional age of Paleogene syntectonic strata in both basins and explore the correlation between Paleogene tectonism in Northern Tibet with the regression of the Proto-Paratethys Sea.

Magnitudes and rates of exhumation determined by low-temperature thermochronology (LTT) and growth-strata deposition provide constraints on the timing of range exhumation in and around syntectonic basins that can be interpreted to reflect major phases of fault activity. The onset timing of growth strata in the immediate vicinity of a basin-bounding fault should largely coincide with the timing of intensified exhumation linked to fault activity (Figs. 2A and 2B). Competing age models for the associated stratigraphic units can be evaluated by comparing the onset of rapid exhumation and correlative faulting revealed by LTT with the proposed age of associated growth strata defined by magnetostratigraphic correlation (Figs. 2A–2D). To validate this approach, we investigate fault activity in the Zagros fold-and-thrust belt (hereafter Zagros Mountains) in Iran and the Ruby Mountains metamorphic core complex (hereafter Ruby Mountains) in western North America (Figs. S1 and S2 in the Supplemental Material¹), where the depositional ages of the syntectonic strata are well established by radiometric ages (Figs. 2E–2J). We then apply this approach to the Tarim and Qaidam Basins to evaluate the debated magnetostratigraphic ages of Paleogene syntectonic strata. By integrating these newly constrained Paleogene tectonism data from the Qaidam Basin with published LTT records (He et al., 2018), we explore the role of intra-plate deformation in driving the Proto-Paratethys Sea incursions. Geological background and statistical analyses of the Paleogene tectonism in Northern Tibet are given in Texts S1 and S2 in the Supplemental Material.

In the Zagros Mountains (Figs. 2E–2G), apatite (U-Th)/He (AHe) data from the hanging wall of the Kirkuk fault record rapid reverse-faulting exhumation at ca. 8–7 Ma (Koshnaw et al., 2020b), which is consistent with the initiation of growth strata in the footwall of the Kirkuk fault at ca. 8.0 Ma (Koshnaw et al., 2017, 2020a). In the Ruby Mountains (Figs. 2H–2J), apatite fission-track (AFT) and AHe ages from the footwall of the Ruby detachment show evidence of rapid normal-faulting exhumation at 17–15 Ma (Colgan et al., 2010), coinciding with the initiation of growth strata in the hanging wall of the detachment at ca. 16 Ma (Lund Snee et al., 2016; Satarugsa and Johnson, 2000). These consistencies between the rapid exhumation and basement cooling in the source area and the initiation of growth strata in the associated sedimentary sink lend strong support to the proposed age model of the late Cenozoic strata in the Zagros and Ruby Mountains, allowing us to apply this approach to debated stratigraphic age models in Central Asia.

In the northwestern Qaidam Basin, AFT data from the basement rocks in the hanging wall of fault BF1 show rapid exhumation at 50–30 Ma and 30–10 Ma (Fig. 3A). These time intervals are widely interpreted as evidence for a two-stage rock uplift of the Altyn Tagh Range (Jolivet et al., 2001; Zhang et al., 2012). As shown on section QB1 (Fig. 3A), two sequences of growth structures occur in the footwall of fault BF1 that are consistent with pulsed exhumation of the basement (Cheng et al., 2021). Following age model Q1 (Ji et al., 2017), the growth strata indicate rock uplift and basement exhumation during the Paleocene–Eocene and Oligocene–Miocene, respectively. This is consistent with the exhumation history of the Altyn Tagh basement revealed by LTT. However, following age model Q2 (Wang et al., 2022), the growth strata indicate pulsed rock uplift at >25.5–23.5 Ma and 16.5 to <6.3 Ma, separated by tectonic quiescence from 23.5 Ma to 16.5 Ma. This second scenario contradicts the Miocene exhumation of the Altyn Tagh basement indicated by LTT data (Jolivet et al., 2001; Zhang et al., 2012).

In the southern Qaidam Basin, AFT and AHe data from the basement on the hanging wall of fault BF2 (Fig. 3B) show rapid exhumation at ca. 35–25 Ma, indicating rapid exhumation of the Eastern Kunlun Range from the latest Eocene to Oligocene (Clark et al., 2010; Li et al., 2021). Growth strata (section QB2, Fig. 3B) are developed in the footwall of fault BF2 (Cheng et al., 2021), suggesting a corresponding rock uplift and exhumation of the Eastern Kunlun Range. Following age model Q1, the occurrence of the growth strata indicates rapid rock uplift and exhumation of the Eastern Kunlun basement from 35.5 to <8.1 Ma, consistent with the exhumation history of the Eastern Kunlun basement revealed by LTT. However, following age model Q2, the resulting age of the growth strata requires tectonic quiescence from >25.5 Ma to 16.5 Ma with subsequent rapid rock uplift from 16.5 to <6.3 Ma. This second scenario contradicts the recognized latest Eocene–Oligocene rapid exhumation of the Eastern Kunlun basement.

In the southwestern Tarim Basin (section TB1, Fig. 3C), AFT data from the hanging wall of fault BF3 reveal rapid exhumation at 24–12 Ma and 12–6 Ma, which are interpreted as evidence of two-stage rock uplift of the Western Kunlun Range (Li et al., 2019). Corresponding growth structures are observed on the footwall of fault BF3 (Wang and Wang, 2016) (Fig. 3C). Following age model T1 (Sun and Liu, 2006), the growth strata indicate a prolonged tectonic quiescence period from 65.5 Ma to 5.2 Ma, with rapid rock uplift of the basement occurring during the Pliocene (>5.2 Ma to <2.6 Ma). This contradicts the proposed late Oligocene to Miocene rapid exhumation of the Western Kunlun Range based on LTT data. However, following age model T2 (Zheng et al., 2015), the growth strata suggest rapid faulting initiation along fault BF3 at ca. 22.6 Ma, consistent with the LTT record. Moreover, in section TB2 (Fig. 3D), AFT data from the hanging wall of fault BF4 show rapid exhumation and rock uplift of the Western Kunlun Range at 15–5 Ma (Cao et al., 2015). Growth strata are well preserved in the footwall of fault BF4. Following age model T1 (Sun and Liu, 2006), the pre-growth strata indicate prolonged tectonic quiescence from 65.5 Ma to 2.6 Ma while the growth strata suggest the onset of rapid rock uplift of the Western Kunlun Range at ca. 2.6 Ma (Fig. 3D). This scenario contradicts the Miocene to Pliocene rapid exhumation of the Western Kunlun Range derived from LTT data (Cao et al., 2015). However, following age model T2, the growth strata indicate a ca. 15 Ma initiation of rapid faulting, consistent with the exhumation history of the Western Kunlun basement based on LTT.

The inconsistency between the LTT data in exhuming source regions and the age of growth-strata relationships in the adjacent sedimentary basins reveals that both “young” age models (Q2 and T1) for Cenozoic strata in the Qaidam and Tarim Basins, while magnetostratigraphically reasonable, are in conflict with the exhumation history of surrounding basement units. Our analysis hence indicates that the timing of syntectonic sedimentation is in very good agreement with age models Q1 and T2. Sedimentation initiated during the Paleocene–Eocene and was followed by intensified Oligocene–Miocene mountain building along the northern Tibetan Plateau margin. This episodic mountain building along the northern margin of the Tibetan Plateau highlights key features of out-of-sequence intra-plate deformation promoted by the post-collisional convergence.

Because Paleogene marine incursions in the Tarim Basin do not simply align with eustatic sea-level changes (Figs. 4A–4C), recent studies have suggested a dominant role of tectonic loading and basin filling associated with the growth of the Pamir salient in the Proto-Paratethys Sea evolution (Kaya et al., 2019). However, Paleogene marine records were recently discovered farther east in the Qaidam Basin (Ma et al., 2022) indicating that the Proto-Paratethys Sea extended into Northern Tibet, which would have been located closer to the Tarim Basin, given the 300–500 km left-lateral offset along the Altyn Tagh fault since the Eocene (Cheng et al., 2016) (Texts S3 and S4; Fig. S4). As a result, Paleogene intracontinental deformation and associated modification in surface elevations along the northern Tibetan Plateau margin could have played a crucial role in driving the Proto-Paratethys Sea retreat in addition to deformation in the Pamir (Kaya et al., 2019). However, this hypothesis remains ambiguous due to the limited LTT data in Northern Tibet (Jepson et al., 2021) and the competing age models of syntectonic strata in the Qaidam Basin.

Here we combine the newly constrained depositional age data from the Qaidam Basin with published LTT data sets, which together reflect Paleogene tectonic activity in Northern Tibet. We observe a consistent temporal correlation between tectonic activity and Proto-Paratethys Sea incursions (Figs. 4A–4D). Specifically, periods of tectonic quiescence in Northern Tibet at 57–56 Ma, 48–41 Ma, and 39–36 Ma correspond to the timing of the first, second, and third incursions of the Proto-Paratethys Sea, while peaks in tectonic activity at 56–48 Ma and 41–39 Ma coincide with the first, second, and third regressions (Fig. 4D). We propose that renewed acceleration of deformation in Northern Tibet and associated surface-elevation change promoted the intermittent retreat of the Proto-Paratethys Sea, while intervening deceleration of tectonic deformation facilitated Proto-Paratethys Sea incursions.

The temporal coincidence between tectonism in Northern Tibet and Proto-Paratethys Sea regression highlights the previously underestimated role of tectonics in Northern Tibet in the retreat of the vast marine domain through uplift and basin infilling, which together with the northward indentation of the Pamir salient as well as the global sea-level fall during the Eocene-Oligocene transition (Kaya et al., 2019), led to the demise of the Proto-Paratethys Sea in Central Asia.

This work was supported by the National Natural Science Foundation of China (U22B6002, 41888101, 41930213) and the Alexander von Humboldt Foundation. We thank editor Robert Holdsworth, Devon Orme, Ryan Leary, and an anonymous reviewer for constructive feedback that improved the manuscript.