The strata of the Cenozoic Tarim Basin were mainly exposed around the basin, and most of the basin was covered by desert. According to the sedimentary characteristics of Cenozoic strata in the basin, the Tarim Basin could be divided into three regions: Southwest Tarim, North Tarim, and Southeast Tarim (Table 1). Although there were different division schemes for Cenozoic stratigraphic subdivision in this area, in this study, the evolution stages in the Cenozoic were based on the stratigraphic subdivision unified as follows (Table 1).

In Southwest Tarim, the Paleogene strata included the Bashibulake Formation, Zhuoyoulegansu Formation, Wulagen Formation, Kalatar Formation, Gaijitage Formation, Qimugen Formation, and Aertashi Formation in the Kashi Group; the Neogene strata included the Pakabulake Formation, Anjuan Formation, Keziluoyi Formation in the Wuqia Group and the Atushi Formation (Table 1). In North Tarim, the Paleogene strata included the Awate Formation, Xiaokuzibai Formation, and Talak Formation in the Kulugeliemu Group; the Neogene strata included the Kuche Formation, Kangcun Formation, Jidike Formation, and Suweiyi Formation (Table 1). In Southeast Tarim, the Paleogene strata included the Kulugeliemu Group; the Neogene strata included the Kuche Formation, Kangcun Formation, Jidike Formation, and Suweiyi Formation (Table 1).

Since the Paleogene, the tectonic compression caused the uplift of the surrounding mountains around the Tarim Basin, while the basin itself was relatively stable (Wang et al., 2014; Laborde et al., 2019; Li Y. P. et al., 2020). During the Neogene, with the uplift of the Qinghai - Tibet Plateau, there are long-term growth and erosion in the mountain ranges and caused progressive thrusting over the Tarim basin (Jolivet et al., 2010; Wang et al., 2014; Laborde et al., 2019). Since the Quaternary, the Tarim Basin was surrounded by high mountains, progressively soil drying and forming China’s largest inland water system and the largest desert.

According to the long-term eustatic trend of the Tarim Basin, there were at least five episodes of marine incursion in the Tarim Basin from the Cretaceous to the Paleogene (Figure 5), which are supported by Cretaceous-Paleogene sedimentary records (Bosboom et al., 2011; Bosboom, 2013; Bosboom et al., 2014; Li et al., 2017; Zhang et al., 2018). There are reliable pieces of evidence to restrict the time of the fourth and fifth incursions and the sea retreat to the Cenozoic (Bosboom et al., 2011; Bosboom et al., 2014; Zhang et al., 2018) (Figure 5).

The palaeomagnetic chronology and paleoenvironmental records of Cenozoic sections indicate that the fourth sea retreat from Aertashi occurred at ∼41 Ma (Bosboom et al., 2014). Research on marine–terrestrial sediments at Oytag in the Early Cenozoic also suggests that the sea retreat from the Tarim Basin may have been diachronous (Sun and Jiang, 2013). The sea retreat from Oytag occurred in the Early Eocene at ∼47 Ma and from Keliyang at ∼45 Ma (Sun and Jiang, 2013; Sun et al., 2016a); the ∼45 Ma peak in detrital zircon U/Pb was considered as an indication of a new sedimentary source (Wang et al., 2021).

Pieces of evidence, including the palaeomagnetic chronology and paleoenvironmental records, also restricted the fifth sea retreat occurring from Kuqa, Kezi, and Bashibulake to ∼37 Ma and ∼33–25 Ma (Bosboom et al., 2014) (Figure 5). In the Pamir-Tianshan convergence zone, the successions changed from marine deposits in the lower, continental clay, fine sand in the middle, and molasse in the upper part, recording the climate changes and evolution of the sea retreat (Wang et al., 2014). Small-scale sea incursions still occurred when the Neo-Tethys Ocean retreat occupied the Tarim Basin, which is supported by the sulfur isotope of gypsums or anhydrites in Kuqa (δ³⁴S from 11.6‰ to 15.9‰ in the Middle Paleocene, from 12.7‰ to 15.9‰ in the Late Eocene) (Zhang et al., 2013). Particle size analysis of the Suweiyi Formation sediments of the Southwest Tarim Depression indicated the sand was similar to near-shore sedimentary environments including lagoons and shallow lake facies (Tang et al., 2014). The lakes in the Jidike Formation were generally broad and shallow, and the lake shoreline was in response to both climate change and tectonic activities (Ma CM. et al., 2020).

Marine and lagoon deposits with fossils were recognized from southwest to the north, which was more than a thousand of kilometers by the outcrop and wells of the Miocene formations of the northern Tarim Basin (Guo et al., 2002). There were thin gypsum interbeds of marine strata in the Late Oliogene Suweiyi Formation in the Kuqa and Late Oliogene Keziluoyi Formation reserved in Southwest Tarim (Guo et al., 2002; Ma T. et al., 2020) (Figure 5). These fossils, including Foraminifer, Dinoflagellates, Ostracods, and Calcareous nannofossils, would seem to suggest that the range of seawater influence in the Early Miocene included two main areas, Kuqa and Southwest Tarim (Figure 5B). The high saline concentration of saltwater deposits with Foraminifer and Ostracods identified in Kuqa and Southwest Tarim showed that two lagoon bays in the western Tarim Basin were linked to the residual sea in the Late Oliogene or even until the early Miocene (Guo et al., 2002). However, from the Late Miocene to the Pliocene, the marine or lagoon environments vanished completely, and lake and freshwater deposits with Stonewort were common in the Tarim Basin (Guo et al., 2002) (Figure 5B).

The final time of the retreat of the Neo-Tethys Ocean from the Tarim Basin remains controversial. Some believe it was in the Early Oligocene (Bosboom et al., 2011; Sun and Jiang, 2013; Sun et al., 2016b), while others believe it occurred in the Late Miocene (Guo et al., 2002). Re-evaluating the fossil evidence and sedimentary facies (Figure 5), it was concluded that the final sea retreat occurred at 37–25 Ma, and the lagoon environments in the southwest and northwest of the Tarim Basin linked to the Neo-Tethys Ocean during the Late Oligocene. Taken together, these studies supported the notion that, as suggested by the presence of marine fossils and gypsum deposits, stepwise and diachronous sea retreat from the southwestern and northwestern margins of the Tarim Basin occurred in the late Oligocene—early Miocene.

The tectonic configuration of the Tarim Basin and its adjacent plates played an important role in the reconstruction of the proto-type basin because the compressive component of Cenozoic deformation controlled the basin (Wei et al., 2000; Li et al., 2007; Li Y. P. et al., 2020), with up to ∼66.7% of compressive deformation accommodated in the Cenozoic orogeny (Laborde et al., 2019). Previous studies have explored the quantified compressive component (Li et al., 2007; Laborde et al., 2019). A regional-scale balanced cross-section constructed on Southwest Tarim—Kuqa including foreland fold-thrust showed a 64% shortening rate for Southwest Tarim and 30.6% for Kuqa (Li et al., 2007). Detailed analysis and examination found that most of the Cenozoic compressive deformation (from ∼94% to 100%) was concentrated in the mountains along the basin margins, which meant it was negligible (up to ∼6%) within the Tarim Basin (Laborde et al., 2019). Therefore, it was of great importance to restore the original thickness isopaches and determine the pre-shortening tectonic configuration of the Tarim Basin.

We calculated the shortening amount of every cross-section and quantified them into four representative profiles: AA′, BB′, CC′, and DD′ (Table 2) (Figure 1). According to the quantified data, we enlarged the range of the pre-shortening original thickness map and considered the zero-thickness line as the boundary of the pre-shortening proto-type basin.

The landform pattern of the Tarim Basin in the Paleocene was mainly characterized by uplift in the center with depressions in the north and south (Shao et al., 2006). A short-term invasion of seawater occurred in the Kuqa Depression and Southwest Tarim Depression from the Paleocene to the Late Eocene (Guo et al., 2002; Li et al., 2017; Zhang et al., 2018; Lin et al., 2019) (Figure 6), where there were marine sediments from tens to hundreds of meters thick deposited in fluvial-deltaic and shallow-water lacustrine settings (Shao et al., 2006).

The degree of seawater influence was decreasing in the Late Eocene due to the small-scale, stepwise Neo-Tethys Ocean retreat (Shao et al., 2006). When seawater came from the west, sabkha or the open platform type of depositional environment was the most important sedimentary facies (Yue et al., 2017). Turning to the period of sea retreat, the basin fill was mainly composed of several typical facies, representing: bay lagoon facies and tidal facies dominated by carbonates and gypsum in the southwestern part; shore-shallow lacustrine facies dominated by sandstone, mudstone, carbonates, and gypsum in the northern part; and fluvial fan facies and fan delta facies dominated by conglomerate, sandstone, and siltstone along the basin margin (Shao et al., 2006; Li et al., 2013) (Figure 6).

The deformation and northward indentation of the Pamir Salient could have played an important role in controlling the lithofacies and sedimentary facies in the Tarim Basin. The provenance of the Cretaceous-Oligocene strata of the Southwest Tarim Depression was mainly from the Central Pamir (Sun et al., 2016a; Wang et al., 2021) (Figure 6), and the deposition center was near the West Kunlun Mountains (Shao et al., 2006). Regarding the Kuqa Depression, the provenance was mainly from the South Tianshan Mountains (He et al., 2004; Shao et al., 2006) (Figure 6), with the deposition center moving southward gradually (He et al., 2004).

The India-Asian collision had a remote effect on structural deformation and tectonic activity (Li et al., 2007; Jia, 2009; Laborde et al., 2019). For example, the West Kunlun reactivation was initiated in the Eocene with subsequent Oligocene to Miocene accelerations (Figures 2A, B), the Southwestern Tian Shan thrust wedge was formed above a decollement level (Figures 2A–C), and the Altyn Tagh strike-slip wedge was formed above a deep decollement level (Figures 2C, D).

In order to reconstruct the proto-type basin map of the Tarim Basin and its adjacent plates, we considered the structures activated during the Cenozoic and quantified the compressive component of this deformation. To the west, about 12.5 km of crustal shortening were accommodated in the Kuqa Depression, while to the east, about 20 km of crustal shortening were accommodated in the Southwest Tarim Depression (Table 2). Then we enlarged the range of the pre-shortening proto-type basin and combined the overall tectonic background. Separating the South Tianshan Mountains and the Pamir Salient was the intervening Alai Valley. The history of the Pamir Salient has been argued but data from new studies in the geology of Northern Pamir suggest that subsequent Cenozoic northward translation of the Pamir was interpreted to be only ∼50 km and no more than ∼80–100 km (Li Y. P.et al., 2020), establishing a greater degree of accuracy on the compressive component we quantified.

During the Late Eocene to Oligocene, the western part of the Tarim Basin contained an open marine, lagoon environment, with a continental environment in the eastern part (Zhang et al., 2022) (Figure 7). Eight sedimentary facies were divided in the study area, including lagoon facies, lacustrine facies, plain braided channel facies, braided delta facies, braided delta plain facies, alluvial plain facies, and alluvial fan facies, showing that the characteristics of spatial distribution in this area were from marine to continental (Figure 7). The Southwest Tarim Depression and Kuqa Depression could be identified as major deposition centers, where the maximum sediment deposition were both over eight-hundred-meters thick (He et al., 2004; Shao et al., 2006; Zhang et al., 2022). In Kuqa, the Suweiyi Formation was mainly composed of mudstones, siltstones, sandstones, and gypsum interbeds with nannofossils (Figure 5). In Southwest Tarim, mudstones with interbedded siltstones and laminated gypsum with nannofossils occurred in the Bashibulake Formation (Figure 5), which is evidence of the depositional environment of a lagoon.

However, whether the lagoon facies developed its range in the Late Oligocene is controversial. Based on the latest lithofacies-paleographic map of the Tarim Basin in the Suweiyi Formation provided by Tarim Oilfield, and combined with the analysis of the Neo-Tethys Ocean retreat (Figure 5), the outside seawater was linked to the western Tarim Basin via the Alai Valley, and the range of seawater influence in the Late Oligocene included two main areas, north to Kuqa and south to Southwest Tarim (Figure 7). In Kuqa, gypsum interbeds of marine strata and marine fossils developed in the Suweiyi Formation and Jidike Formation, which provides some support that lagoon facies developed in the Late Oligocene or even early Miocene. It could be concluded that the range of lagoon facies in the Late Oligocene was larger than the Late Eocene (Figures 6, 7). In the eastern part of the Tarim Basin, a continental sedimentary system existed from the basin margins to the center, from alluvial fan to braided river delta plain and braided river delta front (Shao et al., 2006; Zhang et al., 2022) (Figure 7).

During the Neogene, the lasting tectonic compression caused the mountains surrounding the Tarim Basin to uplift and the basin closed, and the compressive component of the Neogene deformation could also be quantified. To the west, about 11 km of crustal shortening were accommodated in the Kuqa Depression, while to the east, about 10 km of crustal shortening were accommodated in the Southwest Tarim Depression (Table 2). Then we restored the range of the pre-shortening proto-type basin and combined the overall tectonic background. Compared with the Late Paleogene, the mountains, such as the Tianshan Mountains and West Kunlun Mountains, were reactivated by the ongoing structure deformation and tectonic activity (Figures 7, 8). A northward indentation of the Pamir Salient in the Cenozoic, causing a lower amount of shortening documented along the Alai Valley, and the basins, such as the Tarim Basin and the Qaidam Basin, were considered to have undergone continuous aridification.

Compared with the proto-type basin map of the Tarim Basin and its adjacent plates in the Late Eocene and Late Oligocene, the depositional environment of the Tarim Basin significantly changed in the Oligocene—Miocene: marine facies disappeared, and continental sedimentation patterns formed (Figures 6–8). The sea retreated from the whole Tarim Basin in the Late Neocene, and the Tarim Basin gradually graded to a closed inland environment (Figure 8). The Tarim Basin contained terrestrial depositional environments, such as alluvial fan groups, deltas, and flood plain facies with salt lakes (Figure 8). There were large thick molasse units near the deposition center in the Southwest Tarim Depression and Kuqa Depression, while toward the Bachu Uplift the thickness decreased gradually (Wu et al., 2020).

In the Neogene, it was still active on the south edge of the Tianshan mountain, the deformation was influencing the northern edge of the Tarim Basin (Jolivet et al., 2010). The deposition center of the Kuqa Depression gradually moved north to south and the thickness of the sediment increased (He et al., 2004). Several tectonic stages occurred in the Pamir-Tianshan convergence zone in the Cenozoic, and alluvial deposits dominated the sequences from the early Miocene (Wang et al., 2014) (Figure 8). The reactivation event of the Bachu Uplift could be linked to the India-Asian collision, where there were thrust fault systems may related to a deep decollement within the basement below the Tarim Block (Laborde et al., 2019) (Figures 2A, B).

Based on the lithofacies-paleographic map of the proto-type basin and our calculations of the quantified shortening (Table 2), and combined with the regional tectonic configuration, the Tarim proto-type basin and its periphery in the Late Eocene (Figure 6), Late Oligocene (Figure 7), and Late Neogene (Figure 8) was obtained. When comparing the proto-type basin and its periphery in the Late Eocene (Figure 6) and in Late Oligocene (Figure 7), the biggest difference that could be seen was that the range of seawater influence became larger and the range of lagoon facies increased from the Southwest Tarim Depression to the Kuqa Depression (Figure 7). This variation indicated that there was still an intermittent connection between the Tarim Basin and the Neo-Tethys Ocean that occurred separately during the third, fourth, and fifth transgression-regression events in the Paleogene (Figure 5), as well as the final sea retreat which occurred at 25–30 Ma in the Late Oligocene (Figure 5) or even the early Miocene.

In the long-term paleo-geographical evolution in the Cenozoic, as the global sea levels fell, the Neo-Tethys Ocean retreat gradually occurred from the Junggar Basin and the Tarim Basin, and the Fergana Basin and the Alai Valley in southern Kyrgyzstan, and the Afghan–Tajik Basin (Bosboom et al., 2011; Carrapa et al., 2015; Li et al., 2017; Li Q. et al., 2020). In the Paleogene, the Neo-Tethys Ocean entered the western Tarim Basin via the present-day Alai Valley (Bosboom et al., 2011; Zhang et al., 2013; Bosboom et al., 2014) (Figure 9). The sea retreated westward in at least five transgressions in the west of the Tarim Basin during the Cretaceous-Cenozoic (Bosboom et al., 2011; Bosboom, 2013; Zhang et al., 2013; Bosboom et al., 2014), where tidal flat and lagoon facies associations were extended and carbonate reservoirs and mudstone covers were formed (Shao et al., 2006; Yue et al., 2017). The sedimentary lithofacies were reflected by the Neo-Tethys Ocean seawater, and developed lagoon facies in the Alai Valley and shore-shallow lacustrine facies in the east of the Tarim Basin (Guan and Guan, 2002; Guo et al., 2002; Ma CM. et al., 2020) (Figure 5).

For many decades, the time of the India-Asia continental collision has been debated. But it is accepted that the final collision was completed at ∼45–40 Ma, followed by continued post-collisional convergence to the present (Xia et al., 2009; Bouilhol et al., 2013; Todrani et al., 2022). The Tibetan Plateau is still experiencing deformation (Liu et al., 2009), distributing a lot of Cenozoic basin groups (Song et al., 2022) (Figure 9). The northeastern margin of the Tibetan Plateau, which includes the North Qiangtang terrane and the South Qiangtang terrane as well as the Kunlun Mountains and the Qaidam Basin, continues to deform in response to the ongoing India-Asia collision (Liu et al., 2009; Fang et al., 2020). As a huge intermontane basin, the Qaidam Basin was dominated by lacustrine facies during the Oligocene (Wu et al., 2020) (Figure 9).

The tectonic evolution of the Pamir Salient in the Cenozoic has played an important role in controlling the shift from marine to continental sedimentation in the Tarim and Tajik Basins (Bosboom et al., 2011; Sun and Jiang, 2013; Carrapa et al., 2015; Sun et al., 2016b). There were also basin groups distributed in Central Asia that experienced Paleozoic—Cenozoic sedimentary evolution including, for example, the Fergana Basin, Afghanistan-Tajik Basin, and Turgay Basin in the west, and the Junggar Basin and Santanghu Basin in the east (Figure 9) (Wang et al., 2019; Ma T. et al., 2020). Several intermontane basins, such as the Yili basin and the Turpan basin, were preserved within the interior of the TianShan Mountains (Figure 9) (Jolivet et al., 2010). In the west of the tectono-paleogeography map around the Tarim Basin in the Late Paleogene, a remnant Neo-Tethys Ocean branched via Alai Valley into a remnant marine embayment, causing the lagoon facies to develop in the Southwest Tarim Depression and Kuqa Depression in the northern Tarim Basin (Figure 9).

The continental collision between the Arabian and Eurasian plates following the subduction of the Neo-Tethys Ocean beneath Eurasia resulted in a complex deformation across the Zagros Mountains (Figure 10). There was permanent closure of the northwestern segment of the Tethyan Seaway by 12.8 Ma in the Early-Middle Miocene (Homke et al., 2004; Pirouz et al., 2015; Sun et al., 2021). Both the complex mountain building deformation across the Zagros Mountains and the further northward indentation of the Pamir Salient resulted in the closure of the Alai Valley and a complete sea retreat from the Tarim Basin (Figure 10). The sea retreat was the one of the most significant events during the Late Oligocene—Early Miocene, causing the biggest differences in the southwest of the Tarim Basin (Figures 9, 10).

As the India-Asia continental collision is still ongoing, the Pamir Salient is still experiencing deformation. Intraplate orogeny occurred in the Kunlun Mountains and Tianshan Mountains due to the strong remote effect in the Miocene. In the mountains surrounding the Tarim Basin, mainly at the front of the Western Kunlun Mountains and Southwestern Tianshan Mountains, a significant fraction of the deformation was accommodated by new-formed structures during the Cenozoic (Li et al., 2007; Laborde et al., 2019). Accordingly, the margins of the Tarim Basin, mainly in the Southwest Tarim Depression and the Kuqa Depression, corresponded to the basin-mountain system and became rejuvenated foreland basins due to the intense regional compression (Wei et al., 2000; Wang et al., 2020) (Figures 10, 11).

The present-day topography of the Tarim Basin and its surrounding mountains and adjacent blocks is considered to be related to the ongoing India-Asia collision (Figures 9–11). In the Cenozoic, the Western Kunlun Mountains and the South Tianshan Mountains were rejuvenated, and the Southwest Tarim Depression and Kuqa Depression formed rejuvenated foreland basins (Figure 11). Associated with these compressive structures, the Bachu Uplift also corresponded to the rejuvenated Western Kunlun Mountains, forming steep bounding thrusts, in particular deep fault systems (Zheng et al., 2014; Laborde et al., 2019) (Figures 2A, B).