Late Oligocene–early Miocene birth of the Taklimakan Desert
- aSchool of Geography Science, Nanjing Normal University, Nanjing 210023, China;
- bCenter for Excellence in Tibetan Plateau Earth Sciences, Chinese Academy of Sciences, Beijing, 100101, China;
- cSchool of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China;
- dDepartment of Earth and Planetary Science, University of Tokyo, Tokyo 113-0033, Japan;
- eDepartment of Geology and Geophysics, Louisiana State University, Baton Rouge, LA 70803;
- fTourism and Environment College, Shaanxi Normal University, Xi’an 710119, China;
- gWestern Australian Argon Isotope Facility, Department of Applied Geology and John de Laeter Centre, Curtin University, Perth, WA 6845, Australia
See allHide authors and affiliations
Edited by Zhonghe Zhou, Chinese Academy of Sciences, Beijing, China, and approved May 11, 2015 (received for review December 22, 2014)

Significance
The formation of the Taklimakan Desert marked a major geological event in central Asia during the Cenozoic, with far-reaching impacts. Deposition of both eolian sand dunes in the basin center and the genetically equivalent loessite along the basin margins provide evidence for the birth of the Taklimakan Desert. This paper resolves a long-standing debate concerning the age of the Taklimakan Desert, specifically whether it dates to ∼3.4–7 Ma, currently the dominant view. Our result shows that the desert came into existence during late Oligocene–early Miocene, between ∼26.7 Ma and 22.6 Ma, as a result of widespread regional aridification and increased erosion in the surrounding mountain fronts, both of which are closely linked to the tectonic uplift of the Tibetan–Pamir Plateau and Tian Shan.
Abstract
As the world’s second largest sand sea and one of the most important dust sources to the global aerosol system, the formation of the Taklimakan Desert marks a major environmental event in central Asia during the Cenozoic. Determining when and how the desert formed holds the key to better understanding the tectonic–climatic linkage in this critical region. However, the age of the Taklimakan remains controversial, with the dominant view being from ∼3.4 Ma to ∼7 Ma based on magnetostratigraphy of sedimentary sequences within and along the margins of the desert. In this study, we applied radioisotopic methods to precisely date a volcanic tuff preserved in the stratigraphy. We constrained the initial desertification to be late Oligocene to early Miocene, between ∼26.7 Ma and 22.6 Ma. We suggest that the Taklimakan Desert was formed as a response to a combination of widespread regional aridification and increased erosion in the surrounding mountain fronts, both of which are closely linked to the tectonic uplift of the Tibetan–Pamir Plateau and Tian Shan, which had reached a climatically sensitive threshold at this time.
Surrounded by the Tian Shan to the north, the Pamir to the west, and the West Kunlun to the south, the Taklimakan Desert is the world’s second largest sand sea (Fig. 1). Located far from any major source of moisture, and shadowed by Tibetan and central Asian mountain ranges, the Taklimakan is deprived of rainfall, with mean annual precipitation not exceeding 50 mm (1). Provenance studies suggest that mineral dust from the Taklimakan Desert contributes substantially to the global aerosol system, allowing it to play a significant role in modulating global climate on various time scales (2, 3). The formation of the Taklimakan Desert therefore marked a major environmental event in central Asia during the Cenozoic, with far-reaching impacts. Furthermore, determining when and how the desert formed holds the key to better understanding the nature of tectonic–climatic linkage in this critical region. However, the time at which the Taklimakan Desert came into existence has been strongly debated, with estimates ranging from only a few hundreds of thousands to a few million years ago (1, 4⇓⇓⇓–8).
Location map. (A) Topographic map showing the western portion of the Taklimakan Desert (Tarim Basin) and the surrounding mountain ranges with major active faults. The location of the studied Cenozoic sedimentary sections at Aertashi, Kekeya, and Mazatagh are shown by stars. (B) Map showing the location of the Tarim Basin (TB), Junggar Basin (JB), and Chinese Loess Plateau (CLP). Qin’an loess section (QA) is marked by a red star.
In the context of this study, we define desertification to represent not only the formation of a significant sand sea but also the generation of a dynamic eolian system that supplied voluminous mineral dust on regional and even global scales. Evidence of desertification in the geological past is preserved in sedimentary sequences within, and along the margins of, the present-day Taklimakan Desert that lies within the Tarim Basin (Fig. 1). Recent geochronological work, mostly based on the magnetostratigraphy of these sedimentary sequences, proposed ∼3.4 Ma to 7 Ma for the initiation of the Taklimakan (4⇓⇓⇓–8). Precise dating of these terrestrial rocks has largely been hampered by lack of dateable material. In this study, we report a newly identified volcanic tuff from two sedimentary sections along the southwestern margin of the Tarim Basin. Radioisotopic dating of the volcanic minerals provides a robust age for the sections, and therefore we are able to determine the age of the Taklimakan Desert more precisely.
Geological Setting and Lithostratigraphy
From latest Cretaceous to early Paleogene, much of the western Tarim Basin was influenced episodically by a shallow sea, which was connected to the Paratethys, an epicontinental marine seaway covering large parts of Europe and southern central Asia (9). Shallow marine strata of this age are observed extensively along the western and southwestern margins of the Tarim Basin, extending to Hotan to the east and Ulugqat to the north (Fig. 1). Five major marine incursions have been recognized in the broad region, with the fourth being the final one at Aertashi, where it is represented by the Wulagen Formation (Fig. 2 A and B and SI Text). Recent biostratigraphic work at Aertashi has suggested that the Paratethys Sea finally retreated from this locality at ∼41 Ma (9, 10), a process that has been attributed to global eustasy coupled with tectonism associated with the Indo–Asia collision. Together, these have amplified the aridification of the Asian interior (9, 11).
Stratigraphy and magnetostratigraphy of Cenozoic sedimentary sequences within and along the southwestern margin of the Tarim Basin (Taklimakan Desert). The magnetostratigraphy of all sections was correlated to GTS2012 (19). (A) Stratigraphy and magnetostratigraphy of the Aertashi section based on age control from volcanic ash and paleontology (9). (B) Shallow marine deposits and continental red beds at Aertashi. (C) Volcanic ash and loessite intercalated in Xiyu conglomerate at Kekeya section. (D) Stratigraphy and revised magnetostratigraphy for the Kekeya section based on age control of the volcanic ash combined with data from ref. 12. The solid line and dotted line indicate two alternative magnetostratigraphic correlations for this section (see SI Text for more information). (E) Exposed Mazatagh section. (F) Yellowish desert sandstone at Mazatagh section. (G) Stratigraphy and revised magnetostratigraphy based on ref. 6 in the context of regional stratigraphic correlation. (H) Red eolian sandstone within the red beds at Mazatagh. See Fig. 1 for locations.
Syntectonic foreland basin deposition followed the marine regression immediately and led to accumulation of thick continental red beds (Fig. 2 A and B). The Bushiblake Formation is characterized by red, fine-gained mudstone and fine sandstone, with common evaporative gypsum, typical of low-energy, meandering river, lacustrine, and playa deposits (SI Text). The overlying Wuqia Group (SI Text) consists of thickly bedded sandstones with minor mudstone. The overall content of sand increases upward, as does the grain size.
Lithofacies change to distal alluvial deposits, consisting of conglomerate, sandstone, and siltstone crossing the boundary from the Wuqia Group to the Artux Formation (SI Text). The conglomerate layers in the Artux Formation are thin-bedded debris flow deposits, containing medium-sized, angular to subrounded pebbles, of which more than half are sedimentary rocks. The overlying Xiyu Formation (SI Text) is up to 3 km thick and consists of massive boulder to cobble-grade conglomerate with increasing volumes of igneous and high-grade metamorphic clasts. The Xiyu Formation is typical of proximal diluvial fan deposits derived from unroofed mountain belts. Red beds passing upward into upward-coarsening conglomerate and debris flow deposits recorded the change in paleoslope and sediment supply related to uplift of the northern margin of the Tibetan Plateau and Pamir.
Massive siltstone lenses intercalated in the Artux and Xiyu Formations at Kekeya and many other localities are particularly noteworthy (Fig. 2 C and D). Previous sedimentological studies, including facies investigations and grain size (Fig. S1 A and B) and geochemical analyses, suggest that these siltstone lenses are eolian loessite, having been sourced from the desert, deposited, and preserved on an intermittent diluvial fan system (4), a process resembling that in the Taklimakan Desert today.
Grain size distributions (A and C) and accumulations (B and D) of eolian samples and their comparisons with fluvial sediments. (A and B) Samples of the siltstone bands from Artux (YC06-32A) and Xiyu (YC06-68) Formations at Kekeya section, surficial loess at Kekeya (YC06-34), and loess from Chinese Loess Plateau (LC S0-L1-2.1). These samples show strong similarity in terms of grain size distribution. (C and D) Samples of the modern desert sands (TRM106-009), eolian sandstone from Wuqia Group at Mazatagh section (MZTG12-25), and fluvial sandstone from Aertashi section (AT11-022, AT11-022s, At11-025). Eolian origin sands are better sorted than fluvial sands, with minor silt and coarse sand components.
The modern Taklimakan Desert is surrounded by a series of giant diluvial fans that link the uplifting mountain chains with the Tarim Basin. Sediments shed off the mountain fronts have been eroded, mostly but not exclusively, by landsliding and glaciation. They are then delivered down the slope to the basin through the fan systems, and sorted by eolian and fluvial processes into silt and sand fractions, which become the constituents of loess and desert, respectively. These processes of dust and sand production must have been in operation since the deposition of the Artux and Xiyu formations. We therefore conclude that widespread accumulation of loessite in the Artux and Xiyu formations along the margins of the Taklimakan Desert strongly suggests that the source area had become fully arid and desertified, and was supplying dust to the proximal region at that time.
Direct evidence of desertification is found in the exposed sedimentary sequence at the Mazatagh Range, in the central Taklimakan Desert (Figs. 1 and 2E). The lithostratigraphy at this location has been well established and can be generally correlated to the basin margin sequences (Fig. 2G) (5, 6), except that the lithologies are generally finer and sedimentation rates are lower. Red cross-bedded eolian sandstone (Fig. 2H and Figs. S1 C and D and S2) is present in the red bed unit as isolated sandstone lenses, likely indicating an initial stage of a desertified environment that was composed of low-energy fluvial, playa lakes and associated lunette dunes. The most prominent sedimentary unit in the Mazatagh sequence is the well-developed, cross-bedded, yellowish sandstone (Fig. 2F). It is about 200 m thick and stratigraphically continuous and is interpreted as typical desert sand dune deposits.
Scanning Electron Microscope (SEM) images of quartz grains from studied samples. (A and B) SEM images of quartz grains from modern desert sands (TRM106-009). (C and D) SEM images of quartz from the eolian sandstone at the Mazatagh section indicate they are predominantly subrounded to rounded grains, similar to modern desert sands. (E and F) Quartz grains from fluvial sandstone at the Aertashi section are predominantly angular to subangular grains.
Deposition of both eolian sand dunes in the basin center and the genetically equivalent loessite along the basin margins provides two lines of evidence to suggest that the Taklimakan Desert came into existence around that time.
We have identified volcanic ash in the Aertashi and Kekeya sections (Fig. 2C, Fig. S3, and SI Text), where it is intercalated in the Xiyu Formation, the basal age of which has previously been determined by magnetostratigraphy to be of Plio-Pleistocene age (12). In this study, 40Ar/39Ar and U–Pb dating of the volcanic ash has provided, for the first time, to our knowledge, an anchor point to better constrain the strata, and therefore the age of desertification. Petrologic investigation showed that the composition of the ash includes mainly vitroclastic, sanidine, aegirine–augite, and biotite, typical of alkaline magmatic rocks (Fig. S3). These volcanic minerals are ideal for radioisotopic dating. In addition, we also carried out field investigations together with petrologic studies and found that the volcanic activity associated with the Tashkorgan alkaline complex (13) is most likely responsible for providing the Xiyu Formation ash.
Volcanic ash from Kekeya section. (A) Photo showing two layers of the volcanic ash. (B) Microscopic view of the lower (white) ash layer, showing vitroclastic, sanidine, and minor aegirine–augite and biotite. (C) Microscopic view of the upper (dark) ash layer, showing compositions of vitroclastic, aegirine–augite, biotite, sanidine, and quartz.
40Ar/39Ar Dating of Biotite from Volcanic Ash
We selected three samples (YC10-19, AT10-34, and AT10-35) from two locations (Kekeya section and Aertashi section) for 40Ar/39Ar dating and separated unaltered, 200- to 800-µm-size biotite. These minerals were separated using a Frantz magnetic separator, and then carefully handpicked under a binocular microscope. The selected biotite crystals of sample AT10-35 were classified according to their color, with green, red, and brown populations.
The 40Ar/39Ar analyses were performed at the Western Australian Argon Isotope Facility at Curtin University. Due to the possibility of crystal xenocrysts in the volcanic tuff layers, each biotite crystal was fused in one step using a 110-W Spectron Laser System, with a continuous laser rastered over the sample during 1 min to ensure complete melting (see SI Text for more detail). Ar isotopic data corrected for blank, mass discrimination and radioactive decay are given in Datasets S1–S3. Individual errors in Datasets S1–S3 are given at the 1σ level. Since the crystals in a single tuff can yield vastly different ages and we did not step heat the samples, we did not apply the standard plateau age criteria. Rather, we looked for age convergence of the youngest population with a minimum of four grains yielding indistinguishable age. This procedure is similar to the approach used for individual U–Pb zircon age to calculate a magmatic age. All sources of uncertainties are included in the final age proposed for a given sample.
Eighteen crystals were analyzed for sample YC10-19. Fig. S4A shows a continuum of age from ∼11 Ma to 13 Ma. The six youngest ages yielded a concordant age population with a weighted average age of 11.17 ± 0.15 Ma (mean square weighted deviation (MSWD) = 1.2, possibility (P) = 0.33), interpreted as the age of the youngest eruption bearing the crystals. Those results also indicate the occurrence of older eruptions, possibly of up to 13 Ma, whose crystals have been entrained during the youngest eruption at ∼11.2 Ma; however, since this could be due to minute excess 40Ar* incorporated in the crystal, we refrain from making assumptions about the magmatic activity for the three samples. For sample AT10-34 (from the upper layer), we analyzed 14 biotite crystals. This sample includes two distinct age groups, with the oldest group (n = 7) having ages ranging from 120 Ma to 180 Ma and whose significance is beyond the scope of this paper. The youngest group (n = 7) shows a range of age (Fig. S4B), with the four youngest grains giving a concordant age population with a weighted mean age of 10.94 ± 0.20 Ma (MSWD = 0.03, P = 0.99) (Fig. S4C). This age is interpreted as the eruption age of the ash layer from which AT10-34 is derived. The last sample, AT10-35, consists of 21 crystals, and 40Ar/39Ar results show that this sample also contains crystals from much older events (Fig. S4D). Apparent ages of the older group (all red or brown crystals) range from 114 Ma to 214 Ma. Five crystals from this group yield a homogenous age population with a weighted mean age of 189.9 ± 1.4 Ma (MSWD = 1.4, P = 0.24) likely to date a Jurassic magmatic event in the region. The youngest population (n = 13) yields a range of ages from 17 Ma to 10 Ma, suggesting episodic activity during this period. The 10 youngest grains belong mostly to the green group and yield a concordant age population with a weighted mean age of 11.49 ± 0.34 Ma (MSWD = 0.70, P = 0.71) (Fig. S4E). However, it should be noted that most of these apparent ages have large age uncertainties due to the fact that the young population consists mostly of small grains that yielded a small argon beam signal. The age of 11.49 Ma is interpreted as the age of the tuff eruption. All three tuff ages are within 0.5 Ma of each other, suggesting that explosive volcanism occurred in this region over a wide area at ∼11 Ma. Since the three ages are not concordant (MSWD = 4.2), this suggests that the volcanic activity lasted for a duration of at least several hundred thousands years.
Total fusion 40Ar/39Ar age plot. Weighted mean ages were calculated using the youngest concordant crystal populations. For samples YC10-19 and AT10-34, ages selected in the calculation are indicated in red, and excluded ages are indicated in gray. For AT10-35, color code indicates the approximate color of the biotite crystal (red, brown, and green). (A) Eighteen crystals were analyzed for sample YC-19 from Kekeya section. The six youngest ages yielded a weighted mean age of 11.17 ± 0.15 Ma (MSWD = 1.2, P = 0.33). (B) The ages of all 14 biotite crystals of sample AT10-34 from Aertashi section. (C) Four youngest grains of sample AT10-34, giving a weighted mean age of 10.94 ± 0.20 Ma (MSWD = 0.03, P = 0.99). (D) Twenty-one biotite crystals of sample AT10-35 from Aertashi section. (E) Ten youngest grains of AT10-35 yield a weighted mean age of 11.49 ± 0.34 Ma (MSWD = 0.70, P = 0.71).
U–Pb Dating of Zircon from Volcanic Ash
Sample YC12-13-5 was taken from Kekeya section. It was the only sample in this study that contained suitable zircon grains for U–Pb dating. Zircon grains were separated using traditional flotation method, and then carefully handpicked under a binocular microscope to obtain optically clear, colorless grains. Together with zircon standard Plésovice (337 Ma) (14), the zircon grains from sample YC12-13-5 were mounted onto a 2.4-cm-diameter epoxy disk, ground, and then polished for analysis. To reveal the internal structure, all zircons were imaged using transmitted, reflected light, as well as cathodeluminescence. The mount was vacuum-coated with high-purity gold before analysis with secondary ion mass spectrometry (SIMS). U, Th, and Pb were measured using a Cameca IMS-1280 SIMS at the Institute of Geology and Geophysics, Chinese Academy of Sciences, in Beijing. Methods used to obtain the U–Pb analytical data were similar to those developed by Li et al. (15). The standard zircon Plésovice (14) was used to determine U–Pb ratios and absolute abundances. The sample was analyzed in a continuous analytical session with standards interspersed with every four to five unknowns. Corrections for common Pb in this paper were made using measured 204Pb and an age-appropriate Pb isotopic composition of Stacey and Kramers (16). Data processing was carried out using the Isoplot 4.15 program (17).
Twenty-five spots of 25 zircon grains from sample YC12-13-5 were measured during a single analytical session, and the data are reported in Dataset S4, with individual errors given at 1σ level. The analysis results showed relatively high Th/U ratios (0.391–1.903), with U contents ranging from 248 ppm to 2619 ppm. Common Pb was low for the majority of analyses, with f206 values (the proportion of common 206Pb in total measured 206Pb) lower than 4.23%, apart from spot 15, which yielded a value of 28.89%. Since the ages of zircons are quite young, we looked for 206Pb/238U age convergence of the youngest population as the tuff eruption age. Twenty out of 25 zircon grains yielded a homogenous age population with a weighted mean age of 11.18 ± 0.11 Ma (95% confidence interval, MSWD = 1.18, P = 0.27; Fig. S5A), while the other five, which could be xenocrystic zircons, produced 12.4 ± 0.2 Ma, 16.3 ± 0.6 Ma, 28.7 ± 0.6 Ma, 207.0 ± 3.1 Ma, and 707.6 Ma ± 10.3 Ma (spots 20, 17, 08, 10, and 07, respectively; Fig. S6 and Dataset S4). Alternatively, to avoid the influence of common Pb, 14 ages (spots 01, 02, 03, 05, 06, 09, 11, 12, 13, 16, 18, 23, 24, and 25; Fig. S6 and Dataset S4) with measured 206Pb/204Pb > 1,000 were calculated, giving almost the same weighted mean age of 11.18 ± 0.13 Ma (95% confidence interval, MSWD = 1.2, P = 0.26; Fig. S5B). The age of 11.18 Ma is interpreted as the eruption age of the tuff layer from which YC12-13-5 is derived.
206Pb/238U zircon age plot of sample YC12-13–5 from Kekeya section. (A) Weighted mean ages were calculated using the 20 youngest concordant ages. (B) Weighted mean ages were calculated using the 14 youngest concordant crystal populations with measured 206Pb/204Pb > 1,000.
Cathodoluminescence images of measured zircons and the relative 206Pb/238U ages. The red circles show the positions of SIMS dating. Number in the blue box is the spot number.
The 206Pb/238U age of 11.18 Ma of sample YC12-13-5 is very close to the 40Ar/39Ar age of biotite (11.17 Ma) of sample YC10-19. The age difference could have resulted from different eruptions. However, we believe that this minor discrepancy is more likely to have been caused by the difference of closure temperatures of the two types of minerals or age uncertainties. We therefore conclude that the volcanic ash was deposited soon after the eruption that occurred at ∼11 Ma in this region.
Birth of the Taklimakan Desert: When and How?
Constrained by the volcanic ash and regional lithostratigraphic correlations, we are able to reconstruct a magnetostratigraphy for the Cenozoic sequence based on previously published data from Kekeya (12) and Mazatagh (6), as well as new measurements from Aertashi (Fig. 2 A, D, and G). All paleomagnetic samples were progressively demagnetized from room temperature up to 600–695 °C in 5–100 °C steps. Remanent magnetizations were measured using a 2G 755-R Superconducting Rock Magnetometer. The characteristic remnant magnetization (ChRM) directions were assessed on an orthogonal demagnetization diagram and calculated by application of principal component analysis to determine the direction of the best least-squares line fit (18) (Fig. S7 and SI Text).
Normalized decay plots of intensity by thermal demagnetization (Top Row) and orthogonal projections (Bottom Row) of typical thermal demagnetization behaviors of the studied samples. The squares refer to the vertical plane, and circles refer to the horizontal plane. (A) Red sandstone at 664.0 m from the Aertashi section. (B) Reddish mudstone sample at 2633.3 m from the Aertashi section. (C) Yellowish fluvial siltstone sample at 2754.1 m from the Aertashi section. (D) Yellowish loess sample from the Xiyu formation at Kekeya section.
The Aertashi section comprises a relatively complete succession of Cenozoic sediments. More importantly, the section is well constrained by the volcanic ash in the upper part and biostratigraphy at the base (10), and therefore can be used as a reference for regional correlations. The pattern of magnetic polarity zones of Aertashi can be correlated to the Geologic Time Scale 2012 (GTS2012) (19) (Fig. S8, SI Text, and Dataset S5). Two depositional hiatuses occurred when lithofacies changed to massive sandstone and conglomerate, respectively. Geomagnetic correlations to GTS2012 of the Kekeya (Fig. S9, SI Text, and Dataset S6) and Mazatagh (Fig. S10 and SI Text) sections, based on stratigraphic correlations to the Aertashi section, yielded age ranges from 37.5 Ma to 10 Ma and 34 Ma to 17.5 Ma, respectively.
Paleomagnetic results and correlations with the results of Bosboom et al. (10) and GTS2012 (19) for the Aertashi section. These indeterminable polarities are indicated with half-strips for our results. Indeterminable polarities of the results of Bosboom et al. were shaded in gray. The preferred correlation is brown shaded, whereas the rejected correlation is light shaded. See Fig. 2 for explanatory legend of lithostratigraphic patterns and symbols used.
Of greatest significance is the timing of the transition from fluvial deposit to diluvial fan debris flow with eolian silts at Kekeya, and red eolian dunes at Mazatagh, which is, on our updated magnetostratigraphy, dated to be ∼26.7–22.6 Ma (Figs. S9 and S10 and SI Text). We therefore argue that, by late Oligocene to early Miocene time, the Tarim Basin, surrounded by a rising Tibetan–Pamir Plateau and Tian Shan, had become fully arid and desertified, supplying dust to the mountain fronts, where it accumulated as loess. Desertification of the Asian interior has had far-reaching impacts on regional and even global scales. Provenance studies of modern mineral dust suggest that the Taklimakan Desert is one of the major dust source areas to the global aerosol system (2, 3), and might have been so since the birth of the desert. In this regard, it is worth noting that the onset of loess deposition at Qin’an in the Chinese Loess Plateau (20, 21) (see Fig. 1B for location), and change in the dust to the distal north Pacific Ocean (22⇓–24) as shown by a significant increase in the detrital contribution and quartz content, occurred at about the same time (Fig. 3).
Detrital contribution (blue dots) of core from Site 1215 and aeolian quartz percentage (green dots) of core LL44-GPC3 from the North Pacific since the Eocene (22, 24). The arrows on the right side indicate the onset of eolian loess at the Chinese Loess Plateau (20, 21) and desertification of Asian interior obtained by this study.
Formation of a desert over tectonic time scales requires two integral prerequisite conditions: an arid climate and a sufficient supply of fine-grained sediments. During much of the Paleogene, a roughly zonal region of China between 30°N and 50°N was under an arid climate, primarily controlled by the northern westerlies (25). The Paleogene aridity of the Tarim Basin, which was part of this arid region, is registered by widespread accumulation of red beds and intercalated evaporite. However, the climatic configuration changed dramatically during the period from late Oligocene to early Miocene, after which the Asian monsoon set in, prevailing in southeastern China, and the arid region with greatest severity retreated to the northwest.
Indo–Asia collision likely starting at ∼60 Ma resulted in the progressive uplift of the Tibetan Plateau and other mountain ranges in Central Asia (26⇓⇓–29). These may have grown to heights of more than 3 km by the late Oligocene (30, 31), even if the extent of uplift has since continued to grow. Generation of such topography forced major climatic changes during the Cenozoic (32⇓⇓–35). Numerous studies suggested that significant uplift of the Tibetan–Pamir Plateau and the Tian Shan occurred around the late Oligocene–early Miocene (36⇓⇓–39). The uplifted plateau together with the possible retreat of the Paratethys (9, 11) forced by uplifting topography could have led to a transition from planetary climate system to monsoonal climate system (35, 40, 41). At the same time, the arid zone, with much severity, retreated to the Asian interior (25). Equally, if not more, important, is the uplift-induced erosion and weathering of bedrock, which supplied great volumes of sediments to the mountain fronts in the form of alluvial and diluvial fans. These sediments would then have been sorted by fluvial and eolian processes into sand and silt fractions, with the former becoming the constituents of a dynamic desert system and the latter being deflated and transported as eolian dust (4). We suggest that this mechanism has been in operation since the late Oligocene–early Miocene time, and that the resultant formation of the Taklimakan Desert was a direct response to a combination of widespread regional aridification and increased unroofing and erosion in the surrounding mountains, both of which are closely linked to the uplift of the Tibetan–Pamir Plateau and the Tian Shan (36⇓⇓–39).
SI Text
Regional Geology and Stratigraphy of Cenozoic Sedimentary Sequences
With an average elevation in excess of 4 km, the Pamir are composed of tectonic terranes that are the along-strike equivalents of those known from the Himalaya and Tibetan Plateau that accreted onto Eurasia during the Paleozoic and Mesozoic (42, 43). Significant tectonic contraction during the Cenozoic has resulted in thrust and strike–slip faulting within the Pamir that generally follows the arcuate trend of the Pamir salient. The Pamir is one of the most tectonically active regions in central Asia, having experienced ∼300 km of northward indentation relative to stable Eurasia (44, 45) and ∼600–900 km of crustal shortening (46) during the Cenozoic. Thrusting of the Pamir over the Tian Shan to the north along the Main Pamir Thrust accommodated 10% of the present convergence between India and Eurasia (47).
The basement of the Tarim Basin consists of Archean and Proterozoic metamorphic rocks (48). The sedimentary cover, including Proterozoic and Phanerozoic sequences, ranges from 17 km thick within major depositional centers to 5 km thick over the central uplift. During the Cenozoic, the Tarim Basin developed into a complex system of foreland basins largely in response to the India–Eurasia collision. Basinward thrusting of the Tian Shan and western Kunlun Shan caused subsidence of the foreland basins where Cenozoic sedimentary rocks, mainly of terrestrial facies, are up to 10 km thick (48).
The lithostratigraphy of the Cenozoic sequences in the southwestern Tarim Basin along the West Kunlun and Pamir was provided by the Geological Bureau of China (48), with additional information from various authors (49, 50). Basin-wide correlation of the stratigraphy has been well established, and a brief description of each lithological unit from the sections is summarized as follows.
Wulagen Formation.
This formation represents the final episode of a series of marine incursions in the western and southwestern Tarim Basin since latest Cretaceous. The shallow marine to marginal marine deposits are characterized by green gray color and oolitic grainstone, fossiliferous packstone, laminated carbonate mudstone, dolomitic marl, calcarenite, and rippled fine-grained sandstone. Typical fossils include bivalves Exogyra and Pecten, supporting a shallow marine environment. Trace fossils include Skolithos, which is typical of marginal and shallow marine environments.
Bashiblake Formation.
Conformably overlying the Wulangen Formation, this unit contains red colored, interbedded mudstone, shale, and thin sandstone. Thin-bedded gypsum is abundant toward the upper part, together with isolated eolian dune deposits. The depositional environments included fan delta, basin plain, playa, and meandering rivers.
Wuqia Group.
Conformably overlying the Bashiblake Formation, this formation is composed of large, massive sandstone beds interbedded with red mudstone. The proportion and thickness of sandstone beds varies through the section, but mudstone is always subordinate. The thickness of the Wuqia Group ranges from 800 m to 1,500 m. The depositional environment was a meandering river system. It is worth noting that eolian sandstone is found at the upper part of the Wuqia Group at the Mazatagh Range, in the central Taklimakan Desert (Figs. S1 and S2).
Artux Formation.
At Kekeya, the Artux Formation conformably overlies the Wuqia Group, whose base is defined as the first appearance of pebble conglomerate (49). The Artux Formation is composed of orange to yellowish gray, fine-grained sandstone, together with orange, massive, very fine sandstone to siltstone, and granule to pebble conglomerates with minor thin-bedded red mudstone. The siltstone contains predominately eolian silt, which is interpreted as loess (Fig. S1) (49). The thickness within the Kekeya section is 810 m. At Aertashi, the Artux Formation differs slightly in containing more fluvial sandstone and no eolian loess. This is probably because the Aertashi section is close to the paleo-Yarkand River. The depositional environments of the Artux Formation were fluvial, distal alluvial. At Mazatagh, conglomerate is rarely observed, but the well-developed, cross-bedded, yellowish sandstone in lower Atux Formation is noteworthy. It is about 200 m thick and stratigraphically continuous and is interpreted as typical desert sand dune deposits.
Xiyu Formation.
At Kekeya and the surrounding area, the Xiyu Formation is up to 3 km in thickness, consisting of polymictic pebble to boulder conglomerate intercalated with siltstone lenses. Sorting decreases up-section, and the maximum clast size increases (up to 2.5 m) in the upper part. Most thick beds are massive, although many of the thinner beds show normal or reverse grading. Lenses of pale yellow siltstone occur between the conglomerate beds. These siltstone beds are massive and vary in thickness from 30 cm to 10 m. The lenses are commonly completely truncated by the overlying conglomerate. Deposition was episodic, as debris and sheet floods tend to be infrequent but highly energetic, depositing a large amount of sediment in a short period. The lack of significant stream flow on the diluvial fan allowed silt-sized sediments (in the form of silt bands) to accumulate on the uneven fan floor during times of low sediment input. The silt is interpreted as loess (Fig. S1) and is similar to what is currently being deposited on top of the Gobi gravels that rim the Tarim Basin. The Xiyu Formation at the Aertashi section is predominantly polymictic pebble to boulder conglomerate, with minor sandstone lenses and no eolian silts. The depositional environments of the Xiyu Formation at the Aertashi section are interpreted as alluvial fan, which might connect with the paleo-Yarkand River system.
Volcanic Ash at Kekeya and Aertashi
Volcanic ash was discovered in the upper Xiyu Formation at Kekeya. The ash is composed of two layers, ∼1 m apart (Fig. S3A). The upper layer is about 5 cm thick, generally dark colored, and contains 65% volcanic ash and glass. Volcanic minerals include aegirine–augite, biotite, sanidine, and quartz (Fig. S3B). A biotite sample (YC10-19) was retrieved from this layer for 40Ar/39Ar dating. The lower layer is about 30 cm thick and yellowish whitish colored, and is typical of alkaline tuff, composed mainly of vitroclastic and crystal fragments with a dominant grain size of 200–300 μm. Crystal fragments are dominated by sanidine, with minor aegirine–augite and biotite (Fig. S3C). A zircon sample (YC12-13-5) was retrieved for U–Pb dating.
Volcanic ash was also discovered at the Aertashi section from about the same stratigraphic location as that seen at Kekeya. The ash layer is dark colored, and contains aegirine–augite, biotite, sanidine, and quartz. Two biotite samples, AT10-34 and AT10-35, were obtained from two closely separated levels for 40Ar/39Ar dating.
40Ar/39Ar Dating of Biotite from Kekeya and Aertashi
Samples were loaded into three large wells of a 1.9-cm-diameter and 0.3-cm-depth aluminum disk. These wells were bracketed by small wells that included Fish Canyon sanidine used as a neutron fluence monitor for which an age of 28.294 ± 0.036 Ma (1σ) was adopted (51). The discs were Cd-shielded (to minimize undesirable nuclear interference reactions) and irradiated for 2 h in the US Geological Survey nuclear reactor in central position. The mean J values computed from standard grains within the small pits range from 0.0006961 (±0.43%) to 0.0006990 (±0.25%) determined as the average and SD of J values of the small wells for each irradiation disk. Mass discrimination was monitored using an automated air pipette and provided a range of mean values of 1.00085 (±0.30%) to 1.006049 (±0.37%) per dalton (atomic mass unit) relative to an air ratio of 298.56 ± 0.31 (52). The correction factors for interfering isotopes were (39Ar/37Ar)Ca = 7.30 × 10−4 (±11%), (36Ar/37Ar)Ca = 2.82 × 10−4 (±1%) and (40Ar/39Ar)K = 6.76 × 10−4 (±32%).
The 40Ar/39Ar analyses were performed at the Western Australian Argon Isotope Facility at Curtin University. Due to the possibility of crystal xenocrysts in the volcanic tuff layers, each biotite crystal was fused in one step using a 110-W Spectron Laser System, with a continuous Nd-YAG (IR, 1,064 nm) laser rastered over the sample during 1 min to ensure complete melting. The gas was purified in a stainless steel extraction line using two SAES AP10 getters and one GP50 getter. Ar isotopes were measured in static mode using a MAP 215-50 mass spectrometer (resolution of ∼450, sensitivity of 4 × 10−14 mol/V) with a Balzers SEV 217 electron multiplier using 9–10 cycles of peak-hopping. The data acquisition was performed with the Argus program written by M. O. McWilliams and ran under a LabView environment. The raw data were processed using the ArArCALC software 2 (53), and the ages have been calculated using the decay constants recommended by Renne et al. (51). Blanks were monitored every three to four steps, and typical 40Ar blanks range from 1 × 10−16 to 2 × 10−16 mol. Ar isotopic data corrected for blank, mass discrimination, and radioactive decay are given in Datasets S1–S3. Individual errors in Datasets S1–S3 are given at the 1σ level. Since the crystals in a single tuff can yield vastly different ages and we did not step heat the samples, we did not apply the standard plateau age criteria. Rather, we looked for age convergence of the youngest population with a minimum of four grains yielding indistinguishable age. This procedure is similar to the approach used for individual isotope dilution thermal ionization mass spectrometry U–Pb zircon age to calculate a magmatic age. All sources of uncertainties are included in the final age proposed for a given sample.
Paleomagnetic Analysis
A hand-held, petrol-driven rock drill was used to collect paleomagnetic samples in the field. Samples were orientated with a magnetic compass and orientation sleeve.
We collected ∼2,500 samples through the Kekeya section at a resolution of about 1.4 m in the Wuqia Group and Artux Formation. However, sampling was only possible from the siltstone bands with a resolution ranging from 0.2 m to 82.0 m for the Xiyu Formation.
We collected ∼1,123 paleomagnetic samples through the Aertashi section with a sampling interval of about 3.8 m in the Wuqia Group and Artux Formation. For the Xiyu Formation, consisting of conglomerate and minor coarse sandstone lenses, no paleomagnetic sample was collected. Samples A1101–A1110 were drilled at the end of the program to sample the 30-m gap between A493 and A494. A1111–A1123 overlap with A494–A510 ∼150 m along strike.
Thermal treatments and paleomagnetic measurements of all samples from Kekeya section and samples A300–A800 and A1101–A1123 from the Aertashi section were conducted in a low-field cage controlled by three sets of 3.8 m × 3.8 m Helmholtz coils, at the University of Western Australia. Samples were heated to 600–690 °C in 10–100 °C steps, using a Magnetic Measurements Thermal Demagnetizer (MMTD1-18 or MMTD80), and their magnetizations were measured using a 2G 755-R Superconducting Rock Magnetometer (cryogenic magnetometer). For these samples, we select one per three for paleomagnetic measurements; only when polarity was undetermined did we measure the rest of the samples.
Both thermal treatments and paleomagnetic measurements of samples A001–A299 and A801–A1023 from Aertashi section were conducted in the magnetic shielded space (<150 nT) at the Paleomagnetic Laboratory of the Institute of Earth Environment, Chinese Academy of Sciences. Samples were subjected to stepwise thermal demagnetization using a TD-48 thermal demagnetizer. They were stepwise heated to 680–695 °C, using 10–22 temperature steps with temperature increments of 10–75 °C. Remanent magnetizations were measured using a 2G 755-R Superconducting Rock Magnetometer (cryogenic magnetometer).
After removal of a low-temperature overprint component at 100–300 °C, most red colored samples yielded a stable ChRM component directions decay linearly reaching the origin after stepwise demagnetization up to 670–695 °C, indicating hematite and magnetite (or maghemite) as the dominant magnetic carriers of ChRM (Fig. S7 A and B). The ChRM directions of siltstone samples from the Atux and Xiyu Formation at the Kekeya section or brown siltstone samples from the Atux Formation at the Aertashi section decay in the direction of the origin between 250 °C and 600 °C, pointing to magnetite or maghemite carriers (Fig. S7 C and D).
The ChRM directions were assessed on an orthogonal demagnetization diagram (Zijderveld diagram) and calculated by application of principal component analysis (13) to determine the direction of the best least squares line fit. Only the data with maximum angular deviation below 15° and with consistent declination and inclinations were used for the magnetostratigraphy of this paper.
Correlations with GTS2012
Aertashi Section.
The Aertashi section comprises a complete succession of Cenozoic sediments. With the constraints of the volcanic ash of this study and the recent biostratigraphic work (19), we are able to reconstruct a magnetostratigraphy for the continental sequence of this locality. The complete pattern of magnetic polarity zones recorded at the Aetashi section is provided by Fig. S8. Our magnetic polarity zones are consistent with the results of Bosboom et al. (10). The pattern of magnetic polarity zones recognized at Aertashi was correlated with the GTS2012 (14). The correlation of the lower part of the polarity zones is constrained by the independent marine fossil assemblages (10). The basal part of the polarity zones R1 to N7 can be clearly correlated to C18r to C16n. Also, the magnetic polarity zones from R8 to N14 are easily correlated with chrons C12r through C8n. Given the age of the volcanic ash (∼11 Ma), we are able to correlate the upper part of the polarity zones to GTS2012 (14). The two long reversed intervals R29 and R32 are firstly correlated to C5Cr and C5Br. This shows that the upper part of the polarity zones N17 to N32 are in good agreement with C6Bn to C5Bn (Fig. S8).
However, we can’t find the polarity zones correlating to chrons C19 to C16n1n and C6Br to C7, which suggests there likely are depositional hiatus around N7 to R8 and N14 to N17 in the Aertashi section (Fig. S8). The depositional hiatus occurred when lithofacies changed to massive sandstone and then to conglomerate. We argue that the tectonic uplift or decrease of erosion base level not only caused the lithological change from fine to coarse but also led to erosion of this area, which can explain the absence of these intervals.
Kekeya Section.
The marine sequence of the Kekeya section is not exposed, although it was proved to exist, by petroleum drilling. Although the top of the polartity zones is constrained by the volcanic ash, it is still not easy to correlate the polarity zones N17′ to R28′ of the Xiyu Formation to GTS2012. We proposed two alternative correlations to the GTS2012 for this section. From the standpoint of paleomagnetism, we correlate N3′, R5′, R6′, N9′, N16′ through N19′, N21′, N26′, and N29′ to C15n, C8nr, C6Cr, C6Bn.2n through C6Bn.1n, C6n through C5En, C5Cn.3n through C5Cn.1n, C5ADn through C5ACn, and C5n.2n, respectively (Fig. S9, gray shaded). Given the very large sampling intervals, this correlation seems reasonable.
According to field observations and stratigraphic correlations, we provide another correlation. We suggest there is a hiatus at the boundary between N3 and R3, where lithofacies change to massive sandstone and boulder clay was found (Fig. S9). Then the polarity zones from N1′ to N16′ could be correlated to C17n.1n to C6n (Fig. S9, brown shaded), similar with the correlation of the Aertashi section (Fig. S8, brown shaded). For the Xiyu, conglomerate, which is the most typical deposit derived from unroofed mountain belts, reflects the start of rapid uplift of the Tibetan–Pamir Plateau and Tian Shan. We suggest the rapid uplift of the west Kunlun and deposition of the Xiyu formation occurring penecontemporaneously between Aertashi and Kekeya, so we correlate polarity zones of N17′ through N19′, N21′, N26′, N28′, and N29′ to C5Bn.2n through C5ADn, C5ABn, C5An.2n, C5r2n, and C5n.22n, respectively (Fig. S9, brown shaded).
For the first correlation, the onset of the loess from the Atux Formation is dated to about 22.6 Ma, while, for the second correlation, the age of the Atux Formation loess is about 26.7 Ma. Since the age of the Atux Formation loess can’t be entirely confirmed, we suggest that dust accumulation started at this area from about 22.6 Ma to 26.7 Ma, which means desertification in the Tarim Basin probably started in this time range.
Mazatagh Section.
The lithostratigraphy at the Mazatagh can be generally correlated to the basin margin sequences (6), except that the lithologies are generally finer and strata thicknesses are thinner. Previous magnetostratigraphic study suggested that the age range of the Mazatagh section is between ca. 10 Ma and 2.6 Ma. This correlation yielded a sediment accumulation rate of ∼1.51 cm/ky, close to the corresponding rates of the Aertashi (∼1.65 cm/ky) and the Kekeya (∼1.56 cm/ky) sections. Given the Aertashi and Kekeya sections’ location at the southern depression of the Tarim Basin, while the Mazatagh section is located in the in central uplift belt of the Tarim Basin and far away from the high topography of the West Kunlun, it is unlikely that their accumulation rates are almost the same. Therefore, based on stratigraphic correlation and our magnetostratigraphic studies at the southern margin of the Tarim Basin, we recorrelated the polarity zones to GTS2012 and yielded an age range from ca. 34 Ma to 17 Ma, with R2” correlating to C12r, N6” to C9n, N14” to C6Bn.2n through C6Bn.1n, and N16” to C6n (Fig. S10).
Sediment accumulation rate calculated by our new correlation is 0.63 cm/ky, much lower than the rates of the southern margin of the basin. However, this correlation implies that four short normal zones between C6Bn.1n and C6An.1n have not been detected. We argue that the low sedimentary rates and insufficient sampling resolution can explain the absence of these intervals. In addition, the basal age of this section is dated at ca. 34 Ma, close to the basal ages of the Wuqia Group of the Aertashi and Kekeya sections. We suggest that being located in the center uplift belt of the Tarim Basin, the Mazatagh area was under erosion after the Paratethys retreat; however, deposition started synchronously with the Aertashi and Kekeya sections at around ca. 34 Ma. Of great significance is the timing of the transition from fluvial deposit to eolian deposits, which is dated at ∼26 Ma and is roughly consistent with the occurrence of eolian siltstone at Kekeya section inferred from the second correlation. Two lines of evidences from the basin center and the basin margins indicate that the Taklimakan Desert is more likely to have come into existence at around 26–26.7 Ma. The dominant transition, which is bounded with a conglomerate layer, from reddish to the well-developed, yellowish, cross-bedded sandstone is dated at approximately Oligocene/Miocene time. This yellowish eolian sandstone unit suggests that a better-developed desert formed at that time, which is consistent with the records that show a significant increase of dust to the distal north Pacific Ocean (22–24) at about the same time.
Acknowledgments
The authors are grateful to the reviewers for their constructive comments. H.Z. is indebted to the late Prof. C. Powell and the late Prof. X. Wu for their contributions in the field work and analysis of sedimentological data. This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB03020300), the National Science Foundation of China (NSFC 40025207, 90211019, and 41021002), and the Priority Academic Program Development of Jiangsu Higher Education Institutions.
Footnotes
- ↵1To whom correspondence may be addressed. Email: zhenghb{at}njnu.edu.cn or xcwnju{at}gmail.com.
Author contributions: H.Z. designed research; H.Z., X.W., R.T., P.D.C., B.W., P.W., and M.H. performed research; X.W. and F.J. analyzed data; and H.Z., X.W., and F.J. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1424487112/-/DCSupplemental.
References
- ↵.
- Zhu ZD,
- Wu Z,
- Liu S,
- Di X
- ↵
- ↵
- ↵
- ↵.
- Sun J,
- Liu T
- ↵.
- Sun JM,
- Zhang ZQ,
- Zhang LY
- ↵
- ↵.
- Liu W, et al.
- ↵
- ↵
- ↵
- ↵.
- Zheng HB,
- Powell CM,
- An ZS,
- Zhou J,
- Dong GR
- ↵.
- Ke S, et al.
- ↵
- ↵.
- Li XH,
- Liu Y,
- Li QL,
- Guo CH,
- Chamberlain KR
- ↵
- ↵.
- Ludwig K
- ↵
- ↵.
- Gradstein FM,
- Ogg G,
- Schmitz M
- ↵
- ↵.
- Qiang X, et al.
- ↵.
- Pettke T,
- Halliday AN,
- Rea DK
- ↵.
- Rea DK,
- Leinen M,
- Janecek TR
- ↵
- ↵
- ↵.
- Harrison TM,
- Copeland P,
- Kidd WSF,
- Yin A
- ↵.
- Tapponnier P, et al.
- ↵.
- Royden LH,
- Burchfiel BC,
- van der Hilst RD
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵.
- Hendrix MS,
- Dumitru TA,
- Graham SA
- ↵.
- Liu H, et al.
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵.
- Cowgill E
- ↵
- ↵.
- Arrowsmith JR,
- Strecker MR
- ↵.
- Zhou ZY
- ↵
- ↵
- ↵.
- Renne PR,
- Balco G,
- Ludwig KR,
- Mundil R,
- Min K
- ↵
- ↵
Citation Manager Formats
Article Classifications
- Physical Sciences
- Earth, Atmospheric, and Planetary Sciences
This article has a Letter. Please see:
- Relationship between Research Article and Letter - October 01, 2015
See related content:
- More evidence for birth of the Taklimakan Desert- Oct 01, 2015
Sign up for Article Alerts
Jump to section
- Article
- Abstract
- Geological Setting and Lithostratigraphy
- 40Ar/39Ar Dating of Biotite from Volcanic Ash
- U–Pb Dating of Zircon from Volcanic Ash
- Birth of the Taklimakan Desert: When and How?
- SI Text
- Regional Geology and Stratigraphy of Cenozoic Sedimentary Sequences
- Volcanic Ash at Kekeya and Aertashi
- 40Ar/39Ar Dating of Biotite from Kekeya and Aertashi
- Paleomagnetic Analysis
- Correlations with GTS2012
- Acknowledgments
- Footnotes
- References
- Figures & SI
- Info & Metrics