Tertiary evolution of the western Tarim basin, northwest China: A tectono-sedimentary response to northward indentation of the Pamir salient

Authors

  • Hong-Hong Wei,

    Corresponding author
    • Key Laboratory of Continental Collision and Plateau Uplift, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Qing-Ren Meng,

    1. State Key Laboratory of Lithospheric Evolution, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Lin Ding,

    1. Key Laboratory of Continental Collision and Plateau Uplift, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Zhen-Yu Li

    1. Key Laboratory of Continental Collision and Plateau Uplift, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author

Corresponding author: H.-H. Wei, Key Laboratory of Continental Collision and Plateau Uplift, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing 100101, China. (weihh@itpcas.ac.cn)

ABSTRACT

[1] This paper deals with the Tertiary tectonic and sedimentary evolution of the western Tarim basin based on an integrated stratigraphic, sedimentary, structural, and tectonic analyses. Basin evolution is divided into three stages: Paleogene, Miocene, and Pliocene. The western Tarim basin was the easternmost part of the Tethyan realm from Late Cretaceous to Paleogene, and marine sedimentation continued into the Early Miocene. Miocene development of the western Tarim basin was chiefly governed by West Kunlun right-slip faulting and the simultaneous northward thrusting of the Pamir salient and Tianshuihai terrane. Yecheng subbasin developed as a pull-apart basin owing to synchronous activity of the West Kunlun and the Shache-Yangdaman right-slip faults. Hotan foreland basin formed in response to northward displacement of the Tianshuihai terrane, and another might have developed in front of the advancing Pamir salient in the Miocene. Basinward thrusting became predominant in the orogenic belts adjacent to the western Tarim basin in the Pliocene. North-directed displacement and uplift of the Tiklik thrust terrane fragmented the preexisting Hotan foreland basin, and collision of the Pamir with the southern Tian Shan deformational fronts caused complete destruction of the Miocene Pamir foreland basin. Eastward displacement of the Qimugen fold-thrust system led to flexural subsidence of the Yecheng subbasin in the Pliocene. Kashi subbasin developed as part of the southern Tian Shan foreland basin, and was controlled by the eastern Pamir as well. A tectonic scenario is proposed to illustrate complicated interplay of the western Tarim basin with its peripheral orogens in the Tertiary.

1 Introduction

[2] The present-day topography of northwest China is largely shaped by Cenozoic continental shortening related to the Eurasia-India collision [Molnar and Tapponnier, 1975]. The initial collision is considered to be Eocene [Najman et al., 2010], but collision-induced deformation within the Tibetan plateau are controversial [Searle et al., 2011]. The Pamir salient developed as a far-field consequence of the Eurasia-India collision [e.g., Tapponnier et al., 2001], and continues as an area of active deformation [Burtman and Molnar, 1993; Fan et al., 1994; Strecker et al., 1995]. Multiple studies have focused on late Cenozoic tectonics of the Pamir [Amidon and Hynek, 2010; Schmidt et al., 2011; Lukens et al., 2012; Robinson et al., 2007, 2012; Sobel et al., 2013] and its surrounding regions, including the southern Tian Shan [Scharer et al., 2004; Heermance et al., 2008], the West Kunlun Shan [Cowgill, 2010], and the Hindu Kush [Pavlis and Das, 2000]. The northward propagation of the Pamir salient is estimated to be at least 300 km [Burtman and Molnar, 1993]. It was accommodated on the east by the West Kunlun Shan [Cowgill, 2010] and on the west by the Hindu Kush fault systems [Pavlis and Das, 2000] (Figure 1).

Figure 1.

Tectonic map showing relationship of the western Tarim basin and adjacent regions. Note that the Pamir salient separates the Tarim and the Tajik basins that both contain Cretaceous-Paleogene marine deposits.

[3] The Tarim basin lies to the east of the Pamir salient, is flanked on the north by the Tian Shan, on the west by the west Kunlun Shan, and on the south by the Altun Shan (Figure 1). All of these structural belts were reactivated at 20–25 Ma [Sobel and Dumitru, 1997; Yin et al., 2002; Heermance et al., 2008]. The western Tarim basin experienced marked subsidence, with a cumulative thickness of Cenozoic strata up to 10 km [Jia et al., 1997]. It is generally interpreted as a foreland basin in response to eastward convergence of the West Kunlun Shan [Tian et al., 1989; Ding et al., 1997; Yang and Liu, 2002]. However, there is no consensus about the age of initial flexural subsidence induced by tectonic loads [Wang and Fu, 1996; He et al., 2005; Cui et al., 2006]. The detailed Cenozoic space-time development of the western Tarim basin is also poorly understood.

[4] Some efforts have been made to restore structural, stratigraphic, and sedimentary evolution of the western Tarim basin [Wang and Fu, 1996; Meng, 1997; Jin et al., 2003; Qu et al., 2005; Heermance et al., 2007], but few investigations have dealt with tectono-sedimentary processes in the basin interior and their bearing on basin-margin tectonism. This paper attempts to reconstruct the sedimentary and subsidence history of the western Tarim basin in order to reveal its tectonic response to the northward propagation of the Pamir salient. It is shown that the western Tarim basin evolved from an inland sea into a segmented foreland basin in the Cenozoic, controlled by various basin-margin tectonism. Both dextral strike-slip faulting and thrusting might have played a crucial role in basin evolution. This paper first exhibits present-day configuration of the western Tarim basin by geological cross sections in conjunction with isopach maps of Cenozoic units. Then an analysis is carried out of regional stratigraphy and sedimentation, followed by an attempt to restore tectonic and sedimentary history of the basin according to temporal and spatial relationship between basin-edge tectonism and subsidence/sedimentation of basin interior.

2 Regional Geology

[5] The Tarim basin has a Precambrian crystalline basement and a Sinian to Paleozoic marine sedimentary cover [Tian et al., 1989; Jia et al., 2004]. Sutures on the northern, southern, and western edges of the Tarim basin record the amalgamation of the Tarim block with the Kazakhstan plate in the Late Carboniferous [Han et al., 2011] and with the Qiangtang terrane in the Late Triassic [Matte et al., 1996], respectively. The India-Eurasia collision exerted profound influence on the Tian Shan, West Kunlun Shan, and Altun Shan [e.g., Yin and Harrison, 2000], and the resultant large-scale thrusting and strike-slip faulting in these structural belts in turn controlled Cenozoic subsidence and deposition of the Tarim basin [e.g., Yang and Liu, 2002].

[6] Cenozoic structures in the western margin of the Tarim basin include both contractional and strike-slip faults. The right-slip Karakoram fault zone commenced around 23 Ma [Lacassin et al., 2004; Valli et al., 2008] or 15 Ma [Phillips et al., 2004; Searle et al., 2011], and now has a cumulative dextral offset of 120 to 150 km [Searle et al., 1998]. Another major right-slip fault zone, the so-called Kashgar-Yecheng transfer system (KYTS), contains four strike-slip faults, and has a total slip of ~280 km [Cowgill, 2010]. The KYTS was activated in the latest Oligocene [Sobel and Dumitru, 1997; Cowgill, 2010], but its slip rate might have been significantly reduced at 3–5 Ma owing to the substantial increase of northward motion of the Tarim basin [Sobel et al., 2011]. The KYTS may not be a suitable name for this West Kunlun Shan strike-slip fault system in the West Kunlun Shan because of the following: (1) Kashgar and Yecheng are two cities located far from the slip faults concerned; and (2) the Kashgar area was not affected by the West Kunlun right-slip faults, but mainly by the frontal thrust faults of the southern Tian Shan [e.g., Chen et al., 2007]. Therefore, more appropriate geographic names—Wupaer fold-thrust system, West Kunlun dextral slip fault system, Qimugen fold-thrust system, and Hotan fold-thrust system—are used in this study to designate deformation zones along the western margin of the Tarim basin (Figure 2). The east-striking Hotan fold-thrust system involves Tertiary strata [Cui et al., 2006; Li et al., 2010] and is responsible for up to 100 km of Cenozoic shortening [Cowgill et al., 2003]. The Tiklik fault has transported the Tiklik thrust terrane northward over the Hotan fold-thrust system (Figure 2).

Figure 2.

Diagram showing tectonic framework of the western Tarim basin. The western and southern basin margins are structured by the Wupor, Qimugen, and Hotan fold-thrust systems, and a buried Shache-Yangdaman fault exists in basin interior. The basin is divided into a proximal basin zone, the Markit slope, and the Bachu uplift, with the proximal zone subdivided into the Kashi, Yecheng, and Hotan subbasins.

[7] Figure 3 shows the general stratigraphic sequence of the western Tarim basin and its possible correlations with adjacent basins in Central Asia. Late Cretaceous to Paleogene marine deposit is one of the prominent features of the western Tarim basin [Hao et al., 1982; Tang et al., 1992]. The western Tarim basin was once covered by the easternmost portion of the Tethyan sea [Yong and Shan, 1986; Tang et al., 1992] and connected with the Tajik basin because they share similar stratigraphic and sedimentary successions [Burtman, 2000]. Several transgression-regression cycles are recognized, but controversy remains on the maximum extent of marine transgression into the Tarim [Hao and Zeng, 1984; Yong and Shan, 1986; Tang et al., 1992] and the timing and origin of final sea retreat [Meng, 1997; Guo et al., 2002; Bosboom et al., 2011]. Continental sediments comprise the bulk of the basin fills and their deposition, since the latest Oligocene is attributed to the adjacent foreland uplift [Sobel and Dumitru, 1997; Dong and Xiao, 1998; Yin et al., 2002]. However, there is few detailed study to reveal how the western Tarim basin evolved in response to structural development of adjacent orogens [Fu et al., 2010], especially the northward impingement of the Pamir salient.

Figure 3.

A tentative stratigraphic and lithological correlation of the western Tarim basin and the Fergana, Alai, Tajik, and Amu Darya basins. Compiled mainly from Tang et al. [1989, 1992], Burtman [2000], Brookfield and Hashmat [2001], Hao et al. [2002], Guo et al. [2002], Ulmishek [2004], Zhu et al. [2005], and Huan [2009].

3 Methodology

[8] With respect to the Tertiary stratigraphy of the western Tarim basin, we have basically followed its well-established lithostratigraphic sequence, but have refined the age of some units on the basis of recent magnetostratigraphic studies. Stratigraphic correlation of the western Tarim basin with adjacent basins, e.g., Fergana, Tajik, and Amu-Darya, is based on published data. Detailed correlation of stratigraphic units between the individual basins, however, cannot be made owing to a paucity of biostratigraphic database. The results of our tentative correlations are shown in Figure 3. Structural analysis is based in part on review of field studies concerning basin-edge fold-fault systems and, in part, on geological interpretation of published seismic profiles, as shown in Figures 4, 6, and 7. We have conducted sedimentary studies by observing outcrops, examining cores from boreholes, analyzing seismic facies, and investigating basin-wide spatial-temporal variations of facies associations and sequences. Sedimentary facies and their lateral variations are displayed by both field photos, facies sequences, and thickness changes of different units, as shown by Figure 5. In general, the integration and synthesis of relevant data from surface and subsurface geology form a foundation of this study on the tectonic history of the western Tarim basin.

Figure 4.

Three representative geological cross sections based on interpretations of seismic profiles. Locations of the sections are indicated in Figure 2. Compiled mainly from Dong and Xiao [1998], Xiang [2006], and Dai et al. [2009].

Figure 5.

Isopach maps of different units of the Tertiary strata. Note that contour lines of Paleogene units are truncated by the West Kunlun Shan and the southern Tian Shan, whereas the lines of Miocene units are in general parallel to the West Kunlun Shan except for those in the region west of Kashgar City. Contour lines in (a) and (b) are from Shao et al. [2006] and contour lines of (c), (d), and (e) are from Hao et al. [2002].

4 Basin Architecture

[9] Figure 4 exhibits three representative cross sections based on the interpretation of seismic profiles, showing the internal structure of the western Tarim basin. The Cenozoic section is over 10 km thick in front of the West Kunlun Shan [Dong and Xiao, 1998] and tapers to the east. Isopach maps of different stratigraphic units exhibit variations of basin configuration during the Tertiary (Figure 5). The western Tarim basin is typically divided into three subbasins [Dong and Xiao, 1998], the Kashi, Yecheng, and Hotan subbasins, which are flanked on the east by the distal Markit slope and the Bachu uplift (Figure 2).

[10] Three fold-thrust systems are distinguished along the basin's western edge, from north to south, the Wupaer, Qimugen, and Hotan systems (Figure 2). Both surface and subsurface structures of the fold-thrust systems have been investigated [Yin et al., 2002; Qu et al., 2005; Cui et al., 2006; Chen et al., 2010; Cowgill, 2010; Li et al., 2010]. The Wupaer system is characterized by arcuate northeast-convex thrusts between the Main Pamir thrust and the Wupaer thrust [Wu et al., 2004; Chen et al., 2010; Li et al., 2012]. Thrusting was initiated in the latest Pliocene, because the Pliocene Artux Formation lies in the footwall of the Wupaer fault, and a hanging wall piggyback basin is filled with Quaternary sediments (Figure 6).

Figure 6.

Interpreted seismic profile across the Wupaer fold-thrust system, showing a piggyback basin resting unconformably over the Wupaer fold-thrust system and filled with Quaternary strata. The arrowed reflections are possible thrust faults. Refer to Figure 2 for location of the profile. Synthesized from Wu et al. [2004], Liu et al. [2005], and Chen et al. [2010].

[11] The Qimugen fold-thrust system involves pre-Tertiary strata [Qu et al., 2005; Zhang et al., 2007]. Tectonic triangle zones are identified at depth and are characteristic of the leading edge of the Qimugen fold-thrust system [Qu et al., 2004; Hu et al., 2008; Cheng et al., 2011]. The basal thrust might have taken advantage of lower Paleogene gypsum beds typical of the Aertashi and Qimugen Formations (Figure 3) A passive-roof duplex results in folding of overlying Paleogene and younger Tertiary strata into a broad monocline (Figure 7a). Such antiformal stacking is characteristic of passive-roof duplexes [Qu et al., 2004, 2005].

Figure 7.

Line drawing of seismic profiles showing (a) a triangle zone in deep part of the Qimugen fold-thrust system that is characterized by a basal thrust, a roof backthrust and a passive-roof duplex in between and (b) structures of the westernmost part of the Hotan fold-thrust system. Note the growth strata that began developing when the upper Artux Formation was deposited. Refer to Figure 2 for location of the profiles. Seismic sections are from Xiang [2006], Wu et al. [2004], and Hu et al. [2008].

[12] The Hotan fold-thrust system consists of inner and outer zones separated by the Tiklik fault (Figure 2). The inner (southern) zone comprises Proterozoic metamorphic and igneous complexes of various ages, and is referred to as the Tiklik thrust terrane in this study (Figure 8). The outer zone is made up of folds and thrusts involving Paleozoic-Tertiary strata. Some anticlines, like the Duwa and Piyaman anticlines, can be readily recognized in the field (Figure 2); others exist at depth (Figure 7b). Such folds are partially the result of updip propagation of blind thrusts [Luo and He, 1999], or are due to the growth of duplexes at depth (Figure 9). A regional basal detachment exists within either the Aertashi gypsum or Permian shale beneath the outer fold-thrust zone [Luo and He, 1999]. It ramps upsection northward to higher stratigraphic levels about 15–20 km north of the Tiklik fault but rarely reaches the surface [Luo and He, 1999]. Displacement on this detachment is estimated at 40 to 100 km [Wang et al., 2001; Cowgill et al., 2003].

Figure 8.

Simplified geological map showing structures of the eastern Hotan fold-thrust system. Note the occurrence of Miocene strata east of the Tiklik thrust terrane and refer to text for detailed descriptions (from XBGMR, [1993]; Li et al., [2010]).

Figure 9.

Interpreted seismic profile across the Shache–Yangdaman fault. Note thickness difference of the Wuqia Group across the fault. Late reactivation can be perceived by the negative flower structure affecting the Xiyu Formation. N2a, Artux Formation; N1w, Wuqia Group; E–K2, Paleogene and Upper Cretaceous; from Meng [1997].

[13] The interior of the western Tarim basin is less influenced by late Cenozoic deformations than its southern margin, except for some slip and thrust faults (Figure 2), such as the NW-striking right-slip Shache-Yangdaman fault [Meng, 1997; Hu et al., 1997]. This strike-slip fault possibly extends further to the northwest [Hu et al., 1997], but must tip out to the southeast because no faults are recognized in the region west of Pishan town (Figure 2). The stratigraphic units beside the fault are assigned according to basin-wide correlation of key seismic reflections that define the boundaries of units [Dong and Xiao, 1998]. The Shache-Yangdaman fault must be active during the Miocene on account of its control on deposition of the Wuqia Group (Figure 9). The Artux Formation, however, does not display obvious variation of thickness (Figure 9). Late-stage reactivation of the fault is also evident on account of its negative flower structure expressed by Quaternary strata (Figure 9). The Bachu uplift is another area affected by slip and thrust faults (Figure 2), which is attributed to Miocene reactivation of late Paleozoic faults [Hu et al., 1997; Meng et al., 2008]. In addition, Late Pliocene folding is particularly common in the proximity of the western Tarim basin, such as the Yengisar, Atushi, and Kashi anticlines [Qu et al., 2005; Chen et al., 2007]. The Kepingtage thrust zone is also demonstrated to have been active in late Cenozoic [Allen et al., 1999], and also exerted an influence on the faulting of the Bachu uplift [Xiao et al., 2005].

5 Stratigraphy and Sedimentation

5.1 Overview of Tertiary Sequence

[14] Tertiary sequence of the western Tarim basin has been extensively investigated [Hao and Zeng, 1980, 1984; Hao et al., 1982; Compilation of Regional Stratigraphy of the Xinjiang Uygur Autonomous Region (CRSXYAR), 1981; Tang et al., 1989; XBGMR, 1993; Zhou and Chen, 1990; Guo, 1994], and is divided into the Kashi Group, the Wuqia Group, and the Artux Formation in ascending order (Figure 3). The age assignments of individual units in the Paleogene Kashi Group are based primarily on benthic foraminifera and other marine fossils [Hao and Zeng, 1980; Hao et al., 1982; Tang et al., 1989; Pan et al., 1991; Wu, 1991; Guo, 1994; Bosboom et al., 2011]. The Wuqia Group comprises the Keziluoyi, Anjuan, and Pakabulake Formations, and is composed largely of nonmarine deposits, the ages of which are assigned on the basis of foraminifera, ostracods, and charophytes [Hao et al., 1982, 2002]. The onset of the Keziluoyi Formation remains uncertain, because its contact with the underlying Oligocene Bashibulake Formation appears to be a hiatus at the northern margin of the western Tarim basin [Hao et al., 2002]. Magnetostratigraphic study shows that the Anjuan Formation south of Yecheng town formed during a ~21–17 Ma period [Wang et al., 2006]. Accordingly, the Keziluoyi and Anjuan Formations developed together in Early Miocene time and prior to Middle-Late Miocene deposition of the Pakabulake Formation [Hao et al., 2002].

[15] The Artux Formation is of the Pliocene age according to ostracod assemblages and its magnetostratigraphic ages at ~4.6–3.5 Ma [Zheng et al., 2000]. The formation can be divided in the basin proximal areas into a lower part dominated by sandstone and an upper characterized by its conglomerate [Zhou and Chen, 1990; Meng, 1997]. The upper subunit is considered to be the lower portion of the Xiyu Formation in some studies [Zheng et al., 2000; Heermance et al., 2007], given that the widely distributed Xiyu Formation is composed almost entirely of conglomerate. The onset of the Xiyu conglomeratic deposition was diachronous along the edges of the western Tarim basin, ranging from ca. 2.5–3.5 Ma in the eastern flank of the western Kunlun Shan [Teng et al., 1996; Zheng et al., 2000] and ca. 1.9 Ma in front of the southern Tian Shan [Chen et al., 2001] or even ca. 16 Ma [Heermance et al., 2007].

[16] Paleogene stratigraphic sequences of the western Tarim basin can be correlated with their counterparts in the Fergana, Alai, Tajik, and Amu-Darya basins to the west [Leith, 1982; Burtman, 2000; Brookfield and Hashmat, 2001; Ulmishek, 2004], but few detailed correlations of individual units have been conducted [Wu, 1991]. The difficulties in correlating the stratigraphic sequences are in large part due to the lack of detailed Tertiary biostratigraphic study in those central Asia basins. Figure 3 provides a summary of Jurassic to Cenozoic stratigraphic relations between the western Tarim basin and adjacent basins relying mainly on Tang et al. [1989, 1992], Burtman [2000], Brookfield and Hashmat [2001]; Hao et al. [2002], Guo et al. [2002], Ulmishek [2004], Zhu et al. [2005], and Huan [2009].

5.2 Sedimentary Processes

[17] Marine sedimentation in the western Tarim basin commenced in the Late Cretaceous and continued into the Late Oligocene [Hao and Zeng, 1984]. Non-marine deposition then became prevalent from the Miocene. Shallow-marine depositional environments are readily inferred from abundant marine biota [Tang et al., 1992] and extensive literature dealing with marine sedimentation in the western Tarim basin [Hao et al., 1982; Yong and Shan, 1986; Tang et al., 1989, 1992]. Interpretation of continental environments rests mainly on absence of marine fossils and lithofacies associations [Hao et al., 2002].

5.2.1 Kashi Group

[18] Five transgression-regression cycles have been recognized, with the first two occurring in the Late Cretaceous [Tang et al., 1989; Xue et al., 1998]. Marine transgression is recorded by quartz arenite, limestone, and fine-grained clastics, whereas regression is marked by occurrence of evaporate and reddish mudstone as well as shrinkage of depositional area [Tang et al., 1992]. The Aertashi Formation represents the first Tertiary transgression, and is characterized by limestone, dolomite, and thick-bedded whitish gypsum. The carbonate and evaporate are interpreted as shallow-marine and lagoonal deposits during marine transgression [Hao et al., 1982; Tang et al., 1992]. The lower part of the Qimugen Formation is composed primarily of grayish-greenish mudstone and thin- to medium-bedded limestone (Figure 10a), containing plentiful planktonic and benthic foraminifers [Hao and Zeng, 1984] as well as marine bivalves and ostracods (Figure 10b). The upper part is dominated by reddish-brown mudstone and massive gypsum beds. The lower Qimugen deposits are regarded as the consequence of continued marine transgression, whereas the upper resulted from subsequent regression presumably due to relative sea-level fall.

Figure 10.

Field photos of typical sedimentary facies of Tertiary successions. (a) Marine mudstone with thin-bedded sandstone facies, Qimugen Formation. Man for scale; (b) Shell bed, Qimugen Formation; (c) Cross-bedded conglomerate and gravelly sandstone facies interbedded with sandstone, Anjuan Formation. Camera cap 8 cm wide; (d) Core showing bioturbated siltstone and mudstone facies, Pakabulake Formation; (e) Thick- and medium-bedded sandstone interlayered with mudstone/siltstone, Pakabulake Formation. Width of view ~30 m; (f) Thin- to medium-bedded sandstone interlayered with mudstone/siltstone, Artux Formation. Circled car for scale.

[19] The overlying Kalatar-Wulagen succession, characterized by shelly limestone, fossiliferous mudstone, and cross-bedded glauconitic sandstone, represents the recurrence of shallow-marine sedimentation in Middle Eocene time, and passes upward into reddish mudstone, sandstone and thin-bedded gypsum of the lower Bashibulake Formation. The Kalatar-lower Bashibulake sequence represents the fourth marine transgression-regression cycle roughly in Late Eocene time [Tang et al., 1992]. The final marine transgression is inferred from greenish mudstone and limestone in the middle Bashibulake Formation, which contains marine foraminifers, bivalves, and ostracods [Hao and Zeng, 1984]. Marine environment came to an end in the Late Oligocene, as implied by reddish mudrock, sandstone, conglomerate, and gypsum in the upper Bashibulake Formation [Yong and Shan, 1986; Tang et al., 1992]. Figure 11 shows spatial and temporal distribution of marine deposits in the western Tarim basin from the Late Cretaceous to Early Miocene.

Figure 11.

Diagram showing distribution of marine deposits during the (a) Late Cretaceous, (b) Eocene, and (c) Early Miocene. Absence of the Cenozoic in adjacent orogens is considered as the result of erosion caused by late Cenozoic indentation of the Pamir. Contour lines mainly from Yong and Shan [1986] and Guo et al. [2002].

5.2.2 Wuqia Group

5.2.2.1 Kashi Subbasin

[20] The Wuqia sequence is dominated by finer facies like mudstone, siltstone, and fine-grained sandstone (Figure 12, sections 1 and 2). The Keziluoyi Formation is up to 1000 m thick. It is characterized by cross-bedded sandstone with conglomeratic lenses frequently containing broken Paleogene (?) shell fossils [Meng, 1997]. The Anjuan Formation is dominated by laminated mudstone and siltstone, with conglomerate and coarse-grained sandstone present as lenticular bodies of variable scale. Transgressive unconformities separate the Anjuan sandstone from underlying Paleogene and Mesozoic units at basin margins [Zhou et al., 2002]. The Pakabulake Formation, with a thickness of up to 2000 m, consists chiefly of sandstone with mudstone/siltstone interlayers (Figure 10c). Fine-grained facies are dominant in the upper intensely bioturbated sequence (Figure 10d). Cross-bedding is common in sandstone, and soft-sediment deformation, e.g., slumps, occurs in some sequences.

Figure 12.

Representative facies sequences of the Wuqia Group based on measured sections and boreholes. Note that conglomerate-dominated facies are relatively few, present only in a few localities or occurring in the middle of sequences in proximal region. Refer to text for detailed description and discussion. Modified from Meng [1997].

[21] The Keziluoyi Formation (Figure 3) is interpreted as largely fluvial based on occurrence of cross- and parallel-bedding in sandstone and conglomerate, despite its association with red mudstone and gypsum. Its sandstone and conglomerate facies associations represent channel fills, whereas the mudstone and siltstone facies associations result from floodplain deposition. The overlying Anjuan Formation is considered largely lacustrine by virtue of the dominance within it of greenish and dark laminated mudstone. The unconformity between the Anjuan Formation and underlying strata is assumed to have resulted from lake expansion and transgressive deposition at the basin's edges. The still higher Pakabulake succession records lacustrine deltaic sedimentation. Massive conglomerate, sandstone, and soft-sediment deformation collectively indicate rapid deposition of delta fronts, with laminated fine-grained facies being suspension deposits.

5.2.2.2 Yecheng Subbasin

[22] Although sharing similar facies sequences with the Kashi subbasin (Figure 2), the Wuqia Group in the Yecheng subbasin is typified by the frequent occurrence of conglomeratic beds in the Anjuan and Pakabulake Formations in its proximal area (Figure 12, sections 3, 8, 9, and 12), with the accumulative thickness up to 1500 m. Conglomerate is both clast- and matrix-supported, and often associated with stratified sandstone (Figure 10e). Conglomeratic facies change both laterally and vertically into thick-bedded sandstone. In contrast, the Wuqia Group consists dominantly of fine-grained facies in the basin interior (Figure 12, sections 10, 13, and 14). Of particular interest is the presence of turbiditic sandstone in the Anjuan-lower Pakabulake successions in close association with laminated siltstone and mudstone [Zhou et al., 1984; Qiu, 1987; Li et al., 2003]. Turbiditic sandstone exists in a number of drilled cores, and displays different types of Bouma sequence, e.g., Ta-d, Ta-c and Ta-b [Qiu, 1987]. Soft-sediment deformation and conglomeratic debrite are also common in the turbidite-bearing successions [Zhou et al., 1984].

[23] The Keziluoyi Formation is interpreted as the deposits of fluvial systems, and the Anjuan-Pakabulake massive conglomerate and sandstone units as debris-flow and sheetflood deposits. The Anjuan-Pakabulake clastics together represent alluvial-fan systems at basin edges. The finer facies adjacent to the alluvial fans result from alternation of distributary streams and ephemeral lakes in distal area. A dark-colored siltstone/mudstones facies in the Anjuan and Pakabulake Formations is thought to represent deep-lake deposits with turbidity-flow deposition. Attendant soft-sediment deformation originates from synsedimentary slumping. Accordingly, the Pakabulake Formation is better interpreted as the consequence of fan-deltaic sedimentation as a whole.

5.2.2.3 Hotan Subbasin

[24] Lithostratigraphic subdivision of the Wuqia Group is indistinguishable in the Hotan subbasin because it is made up entirely of reddish fine-grained sandstone, siltstone, and mudstone (Figure 12, section 17). Mudrock is mostly unstratified, and desiccation cracks are common. Coarse-grained sandstone occurs as interlayers of various thicknesses in the successions, and exhibits cross-bedding and channelized bases. Massive conglomerate is only locally present. The Wuqia Group in the Hotan subbasin is interpreted as both fluvial and lacustrine. Lenticular sandstone characterizes channel deposits and reddish desiccated mudrock corresponds to floodplain deposits. Some greenish laminated mudrock might have formed in ephemeral lakes. Vertical alternations of fluvial and lacustrine facies are characteristic of the Wuqia Group sequence, which are considered as the result of frequent propagation of fluvial systems into lakes. Also noticeable is the thick accumulation of the fine-grained facies, which indicates consistent aggradations of fluvial and ephemeral-lake sediments. This fact implies a sufficient sediment supply that kept pace with continuous subsidence so that similar depositional environments were maintained.

5.2.3 Artux Formation

[25] The Artux Formation consists generally of conglomerate and sandstone in proximal regions and fine-grained sandstone and siltstone/mudrock in the basin interior (Figure 13). Matrix-supported conglomerate is internally disorganized, whereas clast-supported conglomerate displays parallel stratification and pebble imbrication. Sandstone is either structureless or stratified. Coarse-grained sandstone is commonly interbedded with matrix-supported conglomerate, whereas cross-bedded sandstone is associated closely with clast-supported conglomerate. Brownish mudstone and siltstone associations, together with medium- and thin-bedded sandstone, comprise successions up to hundreds of meters thick (Figure 13). Matrix-supported conglomerate and associated sandstone are interpreted as debris-flow and braided-stream deposits of proximal alluvial fans (Figure 14). The medium- or thin-bedded sandstone and mudstone/siltstone facies represent either alluvial plain deposits if adjacent to alluvial fans, or ephemeral lake deposits if they occur distant from them (Figure 14).

Figure 13.

Representative facies sequences of the Artux Group based on measured sections and boreholes. Note the eastward change from conglomerate-dominated to finer facies in the Kashi subbasin. Refer to text for detailed description and discussion. Modified from Meng [1997].

Figure 14.

Diagram showing spatial distribution of Pliocene depositional systems. Some modern rivers might have existed since the Pliocene in consideration of their contribution to the proximal alluvial fans. Also note the absence of alluvial fans in front of the Tiklik fault.

[26] Marked spatial variations of the Artux facies associations are noticeable [Meng, 1997]. The Artux successions in the Kashi subbasin are dominated by debrites in the westernmost part, conglomerate and sandstone eastward, and mudrock near Kashgar (Figure 13, sections 1–4). This eastward change of facies sequences suggest that the Kashi subbasin might have been filled by a large alluvial-fan system draining from the west.

[27] Sandstone and siltstone facies dominate Artux successions of the Yecheng subbasin (Figure 13, sections 7, 8, 12, 16, and 19), with conglomerate and coarse-grained sandstone present only in proximal areas or in the upper sequences (Figure 13, sections 5, 10, 11, and 14). The sandstone facies is interpreted as deposits of distributary streams from adjacent alluvial fans, whereas the extensive finer facies presumably result from floodplain deposition. Ephemeral lakes might be responsible in part for deposition of mudstone and siltstone. Conglomeratic bodies in the upper Artux successions are likely the result of basinward propagation of alluvial fans.

6 Basin Development

6.1 Paleogene

[28] Together with the Tajik basin, the western Tarim basin was located in the extensional southern margin of the Cretaceous Eurasian continent [Leith, 1985; Otto, 1997; Dong and Xiao, 1998]. The Tethyan sea entered the western Tarim basin in Late Cretaceous time [Hao and Zeng, 1984] (Figure 15a), leading to inundation of the western Tarim basin in the Early Eocene (Figure 15b). Similar lithofacies and correlatable fossil assemblages both attest to the Paleogene connectivity of the western Tarim basin with the Tajik and Amu Darya basins [Hao and Zeng, 1984; Tang et al., 1989; Burtman, 2000; Brookfield and Hashmat, 2001]. An early Tertiary connection between western Tarim and Tajik basins to the west is also implied by isopach maps of the Kashi Group. The contour lines terminate conspicuously against the Pamir-West Kunlun Shan (Figure 5), indicating that Paleogene strata should have extended continuously to the west. A marine to continental transition occurred in the southern margin of the basin around the latest Oligocene due to the northward encroaching of the Pamir salient (Figure 15c), resulting in a westward retreat of the inland sea. Marine incursions might also have resumed in the Early Miocene (Figure 15d), and led to marine sedimentation along the northern periphery of the western Tarim basin [Hu, 1982; Guo et al., 2002].

Figure 15.

Diagrams showing the tectono-sedimentary evolution of the western Tarim basin and its possible coupling with northward indentation of the Pamir salient. (a) Marine transgression occurred in Late Cretaceous; (b) Marine inundation reached its climax in the Early Eocene; (c) Both the Pamir-West Kunlun Shan and the southern Tian Shan began uplifting in the latest Oligocene, leading to the sea withdrawal; (d) The West Kunlun right-slip fault system commenced, and its coupling with the coeval right-stepped Shache-Yangdaman fault led to the formation of a pull-apart basin (Yecheng subbasin). The Hotan, Pamir and southern Tian Shan foreland basins developed simultaneously, and marine incursions also occurred; (d) Large-scale basinward thrusting took place in the peripheral orogens during the Pliocene, and the western Tarim basin manifests itself as a walled basin.

6.2 Miocene

[29] Continued India-Eurasia convergence resulted in persistent northward impingement of the Pamir [e.g., Burtman and Molnar, 1993] and crustal thickening there [Ducea et al., 2003; Mechie et al., 2012; Schmidt et al., 2011]. Both the eastern Karakoram and West Kunlun right-slip faults were initiated in the Early Miocene [Searle et al., 1998; Cowgill, 2010]. A notable depocenter was localized in the Yecheng subbasin, with the Wuqia succession accumulating to thicknesses of up to 6 km (Figure 5). Given the synchronicity of the West Kunlun and Shache-Yangdaman right-slip faults, it is likely that the Yecheng subbasin originated from extension at the right stepover between the two faults and evolved as a pull-apart basin in the Miocene (Figure 15d). This localized subsidence can well account for the persistence of the depocenter through the Miocene (Figure 5). Alluvial-fan and braided-river deposits are dominant along the margin of the Yecheng subbasin, and change into deeper-lake fine facies and turbiditic sandstones in its interior. These facies associations and spatial variations indicate fan-delta deposition [McPherson et al., 1987], which is very common in fault-bounded rapidly subsiding pull-apart basins [Nilsen and Sylvester, 1995].

[30] Contour lines of the Keziluoyi and Anjuan Formation isopach maps are generally parallel to the trend of the West Kunlun Shan, but are truncated by the eastern Pamir (Figure 5). These situations suggest that the West Kunlun Shan must have been uplifted by Early Miocene time, whereas the eastern Pamir did not yet exist. Implicitly, the western Tarim basin was still connected with the Tajik basin to the west and inundated by marine incursions in front of the southern Tian Shan (Figure 15d).

[31] The Hotan fold-thrust system did not form until the Late Pliocene, because the Wuqia Group is clearly involved in folding and the upper Artux Formation manifests itself as growth strata (Figure 7b). Fine-grained facies are characteristic of the Wuqia succession in both the Hotan subbasin [Meng, 1997] and the Buya area (Figure 8), thus indicating that the Tiklik thrust terrane did not come into being in the Miocene. The Tianshuihai terrane south of the Hotan subbasin began to be uplifted at ca. 25–23 Ma [Sobel and Dumitru, 1997; Wang et al., 2003; Li et al., 2007], and displaced to the north [Zhou et al., 2004; Cheng et al., 2008]. Thus, it is conjectured that a foreland basin was developed in response to thrust loading of the Tianshuihai terrane (Figure 15d). The Hotan subbasin represents only the distal portion of the previous Hotan foreland basin, whose proximal sediments were mostly eroded away due to late-stage exhumation of the Tiklik thrust terrane. Some conglomeratic beds did survive at the southernmost edge of the basin, for example, in the Pulu area [CRSXYAR, 1981].

[32] The Bachu uplift is often regarded as the forebulge of the western Tarim foreland basin [He et al., 2005; Ding et al., 2008; Zhang et al., 2011]. Given that right-slip faulting dominated the western margin of the Tarim basin during the Miocene [e.g., Cowgill, 2010], the Bachu uplift cannot be attributed to flexural uplift related to basin-margin thrust loading. The eastward tapering of the Wuqia Group can simply be due to the inherited high topographic relief of the Bachu uplift that experienced multi-phase uplift from the Paleozoic to Mesozoic [Ding et al., 2008; He et al., 2008], especially in the latest Permian [Dumitru et al., 2001; Lin et al., 2012]. The Bachu uplift itself was also characterized by transpressional faulting in the Miocene [Yang et al., 2007; Meng et al., 2008].

6.3 Pliocene

[33] Strong basinward thrusting of the orogenic belts adjacent to the western Tarim basin has occurred since the Pliocene [e.g., Qu et al., 2004]. The Tiklik thrust terrane became active and disrupted the Miocene Hotan foreland basin. The Hotan subbasin evolved into a foreland basin owing to its loading by the Tiklik thrust terrane (Figure 15e), with the Artux Formation thinning northwards toward the Bachu uplift (Figures 4a–4d). Intense basinward thrusting occurred simultaneously along the basin's western margin [e.g., Qu et al., 2005], coeval with a significant decrease in slip rate of the West Kunlun fault [Sobel et al., 2011]. Displacement on the Shache-Yangdaman right-slip fault must also have stopped at this time, because the Artux Formation shows no thickness change on either side of the fault (Figure 9). High-rate subsidence persisted in the Yecheng subbasin, with the Artux succession thickening up to 3500 m. The continued and pronounced subsidence of the Yecheng subbasin is interpreted as the result of the combined tectonic load induced by the Tiklik thrust terrane from the south and the Qimugen thrust system from the west (Figure 15d). The Bachu uplift rose and migrated to the west [He et al., 2008], as evidenced by both eastward onlap of the Artux Formation over the Wuqia Group above a variable parallel to angular unconformity and westward downlap of seismic reflections in the uppermost Artux Formation [Meng et al., 2008]. Implicitly, the Yecheng subbasin resulted from flexural subsidence in the Pliocene. The Kashi subbasin is practically a portion of the southern Tian Shan foreland basin [Heermance et al., 2007], and also bounded on the southwest by the eastern Pamir. The combined influence of the southern Tian Shan and the eastern Pamir led to marked subsidence of the Kashi subbasin where, as mentioned above, the Artux succession is over 3000 m thick [Meng, 1997].

[34] Pliocene development of the western Tarim basin was apparently controlled by three independent fold-thrust systems at the basin margins, the Tiklik, Qimugen, and southern Tian Shan. Consequently, the Pliocene western Tarim basin cannot be classified as a typical foreland basin because its subsidence originated from the combined effect of several thrust belts. The descriptive term “walled basin”, as coined by Carroll et al. [2010], may be suitable for the Pliocene western Tarim basin to account for its unique tectonic and physiographic features among Cenozoic basins in northwestern China.

7 Discussion

7.1 Age and Cause of Final Sea Retreat

[35] Early Tertiary marine sedimentation is a prominent feature of the western Tarim basin, and is considered to have resulted from transgression from the west [Tang et al., 1992]. However, both the age and cause of the final sea retreat remain debated. It is generally held that marine deposition came to an end in Late Oligocene time on account of the first appearance of continental sediments in front of the rising West Kunlun Shan [Tang et al., 1992]. Bosboom et al. [2011], in contrast, argue that the last major regression in the Tarim basin took place as early as Late Eocene, and attribute the sea retreat to global sea-level fall, rather than to uplift of the West Kunlun Shan. However, marine foraminifera have been identified in the Lower Miocene south of the southern Tian Shan [Hu, 1982; Guo et al., 2002], including Cibicidoides, Cibicides, and Pararotalia in the Keziluoyi Formation and Ammonia Honyaensis-A. hatatatensis assemblage in the Anjuan unit. These foraminifera are present in grayish-greenish mudstone [Hu, 1982], indicating they are in situ rather than foraminiferal aeolian deposits. In addition, planktonic foraminifera including Globigerina quinqueloba, Globigerina venezuelana, and Globorotalia mayeri have also been discovered in the southeastern margin of the Tarim basin, thereby attesting to Early Miocene marine deposition there [Ritts et al., 2008]. Although Early Miocene marine deposits are evident, it remains uncertain whether the marine beds arose from marine incursions or from persisting marine sedimentation. The first situation appears most plausible because of the lack of evidence for a long-term continued thick accumulation of typical marine facies.

[36] The uplift of the West Kunlun Shan around 24 Ma [Sobel and Dumitru, 1997] might be responsible for the initiation of latest Oligocene continental deposition in the southwestern Tarim basin. The marine to nonmarine transition occurred in the Tajik basin in the Late Oligocene [Burtman, 2000; Brookfield and Hashmat, 2001]. It is worth noting that early continental sedimentation was restricted to the front of the West Kunlun Shan and the Hindu Kush, whereas marine deposition continued in the northwestern Tarim basin during the Early Miocene (Figure 11c). This fact implies that a lowland zone must have lain south of the Tian Shan and extended westward to the northern Tajik basin. The Tertiary inland sea persisted until at least the Middle Miocene in the Amu-Darya basin [Ulmishek, 2004], and, thus, was likely to have entered the Tarim basin through a threshold lowland zone when sea level rose. This lowland zone can be perceived from Early Miocene isopach maps that imply westward continuation of the Keziluoyi-Anjuan Formations (Figures 5c and 5d).

[37] The aforementioned lowland zone coincides spatially with the foredeep of the southern Tian Shan foreland basin, which commenced in the Early Miocene [Heermance et al., 2008] or since the latest Oligocene [Sobel and Dumitru, 1997]. Late Oligocene-Early Miocene northward overthrusting of the Pamir salient might have amplified the subsidence of the lowland zone, and led to deep-water sedimentation in the Kashi subbasin, as recorded by the Anjuan dark-colored fine facies [Meng, 1997]. Continued convergence of the Pamir toward the southern Tian Shan resulted in complete separation of the Tarim and Tajik basins in the Middle Miocene [Coutand et al., 2002]. It is, therefore, argued that the northward approaching of the Pamir salient should have played an important role in gradual sea retreat out of the Tarim basin, although the possibility of global sea-level fluctuations should not be dismissed.

7.2 Dispersal of Sediment From Source to Basin

[38] Upper Oligocene and Miocene successions consist largely of floodplain facies (Figure 12). Alluvial conglomerate is only present in the southernmost edges of the basin [CRSXYAR, 1981] and in the middle part of the Wuqia Group in the Yecheng subbasin (Figure 12, section 8). Alluvial conglomerate-dominated facies of the Artux Formation occur in front of the West Kunlun Shan (Figure 13), particularly at the outlets of modern rivers, like the Tiznap, Yargant, Kizil, and Keriya Rivers (Figure 14). Few alluvial fans developed in the Hotan subbasin in the Pliocene, although several modern streams like the Karakax and Yurungkax rivers now flow across the Tiklik thrust terrane into the basin (Figure 14).

[39] Late Oligocene and Miocene sediments are thought to have come from the West Kunlun Shan and the southern Tian Shan [Sobel and Dumitru, 1997; Jin et al., 2003; Zheng et al., 2006; Heermance et al., 2007; Bershaw et al., 2012], but few studies have deciphered how sediments were supplied to the basin. Based on our reconnaissance investigation of the time-space distribution of different western Tarim depositional systems, we argue that sediment-routing systems from source to basin must have changed over time due to the northward encroaching of the Pamir salient and the varying deformational processes at the basin's boundaries. The early-stage northward thrusting of the Pamir-West Kunlun Shan belt caused uplift and erosion of pre-Miocene stratigraphic units, leading to alluvial-fan deposition in front of the advancing fold-thrust belts. Overthrusting of the Pamir took place in the Early Miocene [Coutand et al., 2002], and was enhanced by dextral displacements of both the Karakorum and West Kunlun faults [Searle et al., 1998; Cowgill, 2010]. Pre-Miocene deposits in front of the Pamir were either buried or removed by the advancing Pamir salient. Proximal alluvial fans survived only along the basin's edge where slip faulting was dominant or basinward thrusting was insignificant. We postulate that a longitudinal trunk river might have existed in front of the advancing Pamir, flowing to the east and transporting detritus into a western Tarim basin. This likelihood is supported by the overall east-directed paleocurrents recorded from the Wuqia successions west of the Yecheng subbasins [Meng, 1997; Zheng et al., 2006; Chen et al., 2011]. The eastward flow of this longitudinal river might have been partially guided by rapid subsidence of the Yecheng subbasin. The southern Tian Shan foreland basin developed independently with respect to the northward advance of the Pamir in the Early Miocene, and was filled dominantly with fluvial-lacustrine facies (Figure 12, sections 1 and 2).

[40] The Pliocene Artux alluvial conglomerate occurs mostly at the outlets of modern rivers along the edge of the western Tarim basin. This spatial coincidence suggests that some modern rivers had existed since the Pliocene and were responsible for the Artux alluvial-fan formation. The Artux sequences in the Yecheng subbasin are characterized by beds rich in unstable heavy minerals, such as pyroxene, amphibole, and epidote [Zhou, 1985]. The Kudi ophiolites, made up basically of ultramafic, gabbroic, and basaltic rocks [e.g., Matte et al., 1996], could be the source for the heavy minerals. If so, the Tiznap River might have delivered the heavy minerals into the Yecheng subbasin because the ophiolites currently crop out only in its headwaters (Figure 14). In addition, the marked eastward change from conglomerate- to sandstone-dominated faces (Figure 13) suggests that sediments debouch mostly from the west, consistent with the restored eastward paleocurrents [Meng, 1997]. The Kizil River might be primarily responsible for transporting these sediments (Figure 14). Alluvial fans in the Pulu area at the southeastern edge of the Hotan subbasin were presumably fed by the Keriya River (Figure 14). By virtue of these collective observations, it is plausible that some of the west Tarim's modern trunk streams, such as the Kizil, Yarkant, and Tiznap rivers, must have existed since the Pliocene. The same might be true of the Surkhob river system in the northeastern Tajik basin.

8 Conclusion

[41] Late Cenozoic deformation of the Pamir-West Kunlun Shan belt has been investigated by different approaches, but detailed deformational processes remain unclear. It is postulated in this study that West Kunlun dextral slip faulting might have been coupled with that of the Shache-Yangdaman fault through a right stepover during the Miocene. The consequence would be the localized extensional subsidence of the Yecheng subbasin. Early Pliocene decrease in the slip rate of the West Kunlun dextral slip faults was coincident with both initiation of the Qimugen fold-thrust system west of the Yecheng subbasin and the northward displacement of the Tiklik thrust terrane. The transition from strike-slip to thrust deformations might be the result of the head-on interaction of the deformation fronts of the northern Pamir and the southern Tian Shan, giving rise to dramatic crustal thickening and radial thrusting. The Pliocene eastward migration of the western Tarim basin depocenter can be attributed to basinward displacement along both the Qimugen and Tiklik fold-thrust systems. It is proved that the western Tarim basin underwent a complex evolution controlled by various basin-margin tectonism, and it was obviously not a simple foreland basin system as previously regarded. Gradual transition of basin types was clearly related to progressive northward impingement of the Pamir salient.

[42] In summary, the western Tarim basin experienced a three-stage evolution during the Tertiary. During the Paleogene, the Tarim basin was one part of the southern margin of the Eurasia continent, and manifested itself as its easternmost inland sea. The Miocene witnessed a period of marked northward impingement on the basin by the Pamir salient, which caused its eastern portion, now the western Tarim, to evolve gradually into an independent basin. Three subbasins developed. The Hotan and Pamir foreland basins formed in response to the advancing Pamir and Tianshuihai domains, whereas the Yecheng subbasin originated from right pull-apart extension between the West Kunlun and the Shache-Yangdaman dextral faults. Pliocene basinward thrusting of peripheral orogens collectively contributed to the subsidence of the western Tarim basin, and led to a significant modification of previous basin architecture.

Acknowledgments

[43] This research is supported by grants from the National Science Foundation of China (41272238, 40721003 and 40972151). We thank E. Sobel, C. Johnson, L. Schoenbohm (Associate Editor), and an anonymous reviewer for offering critical reviews that have led to significant improvement of the paper. Thanks also go to G.A. Davis for helping polish the English.

Ancillary