Stable isotopes in fossil mammals, fish and shells from Kunlun Pass Basin, Tibetan Plateau: Paleo-climatic and paleo-elevation implications

Earth and Planetary Science Letters 270 (2008) 73–85

Contents lists available at ScienceDirect

Earth and Planetary Science Letters j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e p s l

Stable isotopes in fossil mammals, fish and shells from Kunlun Pass Basin, Tibetan Plateau: Paleo-climatic and paleo-elevation implications Yang Wang a,⁎, Xiaoming Wang b, Yingfeng Xu a, Chunfu Zhang a, Qiang Li c, Zhijie Jack Tseng b, Gary Takeuchi b, Tao Deng c a b c

Department of Geological Sciences, Florida State University and National High Magnetic Field Laboratory, Tallahassee, Florida 32306-4100, USA Department of Vertebrate Paleontology, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007, USA Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing, 100044, PR China

A R T I C L E

I N F O

Article history: Received 10 December 2007 Received in revised form 25 February 2008 Accepted 3 March 2008 Available online 13 March 2008 Editor: H. Elderfield Keywords: stable isotopes fossils paleo-climate paleo-elevation Tibetan Plateau

A B S T R A C T We report the results of a stable isotope study of a late Pliocene fauna recently discovered in the Kunlun Mountain Pass area (∼4700 m above sea level) on the northern Tibetan Plateau. The δ13C values of enamel samples from modern herbivores from the Kunlun Pass Basin range from −14.8 to −10.6‰, with a mean of −12.0 ± 0.7‰, indicating pure C3 diets consistent with the current dominance of C3 vegetation in the area. In contrast, enamel samples from fossil herbivores yielded δ13C values of −5.4‰ to −10.2‰ (with a mean of −7.9± 1.3‰), significantly higher than those of modern herbivores in the area. The higher δ13C values indicate that these ancient herbivores, unlike their modern counterparts, had a variety of diets ranging from pure C3 to mixed C3/C4 vegetation. The local ecosystems in the Kunlun Pass area in the late Pliocene likely included grasslands that had small amounts of C4 grasses. The δ18O values of enamel from large herbivores shifted to higher values after the late Pliocene, indicating a significant change in the δ18O of local meteoric water. We estimate that there has been approximately 3.2‰ increase in annual δ18O values of meteoric water since ∼2–3 Ma, most likely driven by changes in the regional hydrological cycle possibly as a result of tectonic and climate change. The δ18O values of fossil fish teeth/bones and gastropod shells, along with abundance of aquatic plants and other invertebrate fossils, clearly indicate that the Kunlun Pass Basin once had plenty of water and was occupied by a freshwater lake in the late Pliocene. Our isotope data from both terrestrial and aquatic fossils suggest that the Kunlun Pass Basin was a hospitable place with a much warmer and wetter climate in the late Pliocene, very different from today's rock desert and cold steppe environments. The mean annual temperature in the late Pliocene estimated from the δ18O of fossil bone carbonate and paleo-water was about 10 ± 8 °C, much higher than the present-day mean annual temperature in the basin. If valid, the estimated temperature change would imply that the elevation of the basin has increased by ∼2700 ± 1600 m since ∼2–3 Ma. © 2008 Elsevier B.V. All rights reserved.

1. Introduction As the world's largest highland, the Himalayan–Tibetan Plateau plays an important role in driving the Asian Monsoons and global climate. However, the timing history of the uplift of the plateau remains a matter of considerable debate because there are few direct indicators of paleo-topography in geological record. Reconstructing the paleo-environment and the elevation history of the plateau can improve our understanding of the linkage between tectonics and long-term climate change. A late Pliocene fauna was recently discovered in the Kunlun Mountain Pass area on the northern Tibetan Plateau, at an elevation of about 4700 m above sea level (Wang et al., 2006). These fossil materials provide a unique window that allows us to examine the biotic and climatic consequences of the uplift of the Tibetan Plateau. ⁎ Corresponding author. Tel.: +1 850 644 1121; fax: + 1 850 644 0827. E-mail address: [email protected] (Y. Wang). 0012-821X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2008.03.006

Stable carbon and oxygen isotope analysis of these fossils allow reconstruction of certain aspects of the paleo-environment. Teeth and bones in living animals are primarily composed of hydroxyapatite [Ca10(PO4)6(OH)2], which contains a small amount of structural carbonate substituting for phosphate and hydroxyl ions. Because mammals maintain a constant body temperature, the stable carbon and oxygen isotope compositions of their tooth/bone apatite are determined by the isotopic compositions of diet and drinking water, independent of environmental temperature. Tooth enamel is resistant to isotopic exchange and tends to retain its original isotopic signal, reflecting dietary (δ13C) and, in obligate drinkers, local meteoric water (δ18O) compositions (e.g., Longinelli, 1984; Quade et al., 1992; Wang and Cerling, 1994; Delgado Huertas et al., 1995; Bryant and Froelich, 1995; Kohn and Cerling, 2002). Enamel is also preferred for isotopic analysis because it mineralizes progressively along the length of the tooth, recording seasonal variations in diet and climate (e.g., Koch et al., 1995; Fricke and O'Neil, 1996; Wang et al., 2008). Bones on the other hand are easily altered during diagenesis due to their porous

74

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

nature. The stable isotopic composition of structural carbonate in fossil bone is thought to reset completely to early diagenetic composition, comparable to palaeosol carbonate, which reflects the δ18O of local water as well as local mean annual temperature (Kohn and Law, 2006; Zanazzi et al., 2007). Thus, the δ18O of fossil bone carbonate may be used as a paleo-thermometer if the δ18O of local water could be determined independently from the δ18O of fossil tooth enamel (Zanazzi et al., 2007). Oxygen isotopic analyses of fish tooth/bone bioapatite and mollusk/gastropod shells can provide valuable information about aquatic environments in which fish or shells grew (e.g., Longinelli and Nuti, 1973; Kolodny et al., 1983; Dettman et al., 2001; Zazzo et al., 2006), because the δ18O values of fish bioapatite and of shell carbonate are determined by the δ18O of water as well as the water temperature. Although studies have shown that bioapatite in fish and authigenic lacustrine carbonates precipitate in isotopic equilibrium with lake water (Kolodny et al., 1983; Turner et al., 1983; Gasse and Fontes, 1987; Fritz et al., 1987; Talbot, 1990; Dutkiewicz et al., 2000), the δ18O of lake water can differ significantly from that of the precipitation feeding a given lake system due to the influences of regional climatic and hydrologic factors, such as evaporation and groundwater inflow.

Therefore, the δ18O values of fish bones/teeth and mollusk/gastropod shells, if unaltered, provide a record of changes in regional climate and hydrology that control the δ18O of lake water and water temperature. Previous studies of paleo-environments of northern Tibetan Plateau primarily rely on palynologic, invertebrate faunal and sedimentologic evidence, and attribute much of the environmental change to dramatic increases in basin elevation (Pang, 1982; Kong et al., 1982; Yin et al., 1996; Wu et al., 2001). Here we present a stable carbon and oxygen isotope record based on analyses of fossil herbivores, fish and gastropods from the Kunlun Pass Basin that shows a significant change in local habitats and regional hydrological cycle after the late Pliocene. 2. Study site Our fossils were collected from the Kunlun Pass Basin located in the Kunlun Pass area of the East Kunlun Mountains on the northern Tibetan Plateau (Fig. 1). The elevation of the basin is about 4600–5300 m above sea level (a.s.l.). The highest peak of the East Kunlun Mountains is the Yuzhufeng with an elevation of 6178 m a.s.l. Modern glacier tongues from the high mountains extend to about 4500 m on the north slope and 4900 m on the south (Wu et al., 2001). The high mountain area and the

Fig. 1. A map showing the location of the study area. The white dots mark the major vertebrate fossil localities that produced materials used in this study.

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85

Kunlun Pass basin are in the permafrost zone (Song et al., 2005). The mean annual temperature is around −6° to −7 °C and the annual rainfall is about 276 mm in the Kunlun Pass area (Kong et al., 1982; Pang, 1982; Liu et al., 2006). The day-time high temperature in the Pass basin in July is less than 10 °C. The temperature often falls to −35 °C or lower in winter. Because of the extreme climatic and topographic conditions, the desert or, at best, cold meadow or steppe environments prevail throughout the Kunlun Mountains, inhibiting development of vegetation. Much of the terrain consists of deserts. Occasional stagnant water pools and associated meadows and streams derived from glacial melts provide browsing and water for several wild ungulates, such as the Tibetan gazelle (Procapra picticaudata) and Tibetan antelope (Pantholops hodgsonii), along with large herds of wild asses (Equus kiang) and clusters of wild yaks (Bos grunniens).

75

The Kunlun Pass Basin (Fig. 1) is an asymmetric pull-apart basin bounded by major faults (Kidd and Molnar, 1988; Song et al., 2005; Lin et al., 2002). The basin is filled with Pliocene and Quaternary alluvial, lacustrine and glacial deposits, which dip ∼ 13° southwest and unconformably overlie Triassic metamorphic basement rocks (Wu et al., 2001; Song et al., 2005). The deposits in the basin have been divided into Kunlun Formation, Qiangtang Formation and Wangkun Till (Fig. 2). The Kunlun Formation consists primarily of conglomerates and sandy conglomerates. The Qiangtang Formation is mainly composed of siltstone and mudstone of lacustrine and fan delta deposits. The laminated organic-rich lacustrine siltstones/mudstones of the lower Qiangtang Formation contain abundant plant remains and ostracod and mollusk/gastropod shells. Studies of fossil plants, pollens and shells suggest that the basin was occupied by a shallow

Fig. 2. The lithostratigraphy and magnetic stratigraphy of the Kunlun Pass Basin, northern Tibetan Plateau (adapted from Song et al., 2005). The ages of the fossil localities are estimated on the basis of paleo-magnetic time scale and stratigraphic positions.

76

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

freshwater lake with abundant aquatic plants such as Pediastrum boryanum, Potamogeton, Sparanium, Nelumbo, Phragmites and Ligustium during that time (Kong et al., 1982; Yin et al., 1996). The upper most part of the sequence — the Wangkun Till is a sequence of glacial deposits, which is mixed breccia composed of angular, poorly sorted gravels (2–200 cm in size) of slate, meta-sandstones and mudstones, granite, granite gneiss and pyroxenolite, and overlies unconformably the Qiangtang Formation (Song et al., 2005). 3. Materials and methods We obtained 70 bulk and serial enamel samples from 9 modern herbivore teeth and 90 enamel samples from 31 fossil teeth and tooth fragments for carbon and oxygen isotope analyses. We also analyzed the oxygen isotopic compositions of fossil fish teeth and bones and fossil mammalian bones as well as gastropod shells collected from the fossil localities in the basin. Modern herbivores analyzed include six horse teeth (five from wild Tibetan asses E. kiang, one possibly from a domesticated horse Equus caballus), one yak tooth (B. runniens) and five small mammal teeth (Arvicolidae and Ochotona) from the Kunlun Pass Basin at an elevation of about 4700–4800 m a.s.l, and also two goat teeth from Xi-Da-Tan, north of the Kunlun Pass Basin, at ∼ 4100 m elevation (Fig. 1). Fossil materials are from Neogene fossil localities in the Kunlun Pass Basin (Fig. 1). These fossils were found in and around two layers of grayish black, organic-rich mudstones in the lower Qiangtang Formation, including hipparionine horse, bovids, rhino, and other unidentified mammalian herbivores (Fig. 2). A recent paleomagnetic study (Song et al., 2005) suggests that the layers containing the vertebrate fossils were deposited at about 2.0–2.5 Ma (Fig. 2), consistent with age estimates based on small mammalian fossils (Wang et al., 2006). In addition, we collected stream-water and rainwater samples and various plants in the Kunlun Pass area for oxygen, hydrogen and carbon isotopic analyses, and the plant isotope data were reported in Wang et al. (2008). For bulk enamel samples, we drilled enamel powder along the growth axis to ensure that samples reflected an average composition for that individual. We also collected serial enamel samples (that represent a time-series from six fossil teeth or tooth fragments and six modern teeth) by drilling in bands perpendicular to the growth axis of each tooth. The enamel samples were prepared following a treatment procedure described in Wang and Deng (2005). Fossil shells were cleaned in distilled water in an ultra-sonic bath to remove sediments adhered to their surfaces, dried and ground into powder. Sediment samples were also ground into fine powder. The treated enamel samples and the powdered shell/sediment samples were reacted with 100% phosphoric acid at 25 °C (over three nights for enamel samples and overnight for other carbonates) and the carbon and oxygen isotopic ratios of the CO2 produced were analyzed using a Gas Bench II Auto-carbonate device connected to a Finnigan MAT Delta Plus XP stable isotope ratio mass spectrometer (IRMS) at the Florida State University (FSU). For selected samples, we also measured the oxygen isotopic composition of enamel phosphate to check for diagenetic alteration. Enamel phosphate sample was converted to Ag3PO4 (O'Neil et al., 1992) and its oxygen isotopic composition was then analyzed using a TC/EA (High Temperature Conversion Elemental Analyzer) connected to the IRMS at FSU. Water samples were analyzed using the equilibration methods (Thermo Finnigan Operating Manual) as described in Wang et al. (2008). Isotope data are reported in the standard notation as δ13C and δ18O in reference to the international carbonate standard VPDB (Pee Dee Belemnite) for plant and enamel carbonate and to the international standard VSMOW (Vienna Standard Mean Ocean Water) for enamel phosphate and water. The analytical precision (based on replicate analyses of lab standards processed with each batch of samples) is ±0.1‰ or better for both δ13C and δ18O, and ±0.3‰ for enamel phosphate.

4. Results and discussion 4.1. Assessment of diagenetic alteration of fossil bioapatite and gastropod shells Tooth enamel is considered the most suitable material for paleoclimate study using stable isotopes because apatite crystals that make up tooth enamel are large and densely packed and are more resistant to diagenetic alteration (Kolodny et al., 1983; Shemesh et al., 1988; Quade et al., 1992; Wang and Cerling, 1994; Lecuyer et al., 1999). There is about 8–9‰ oxygen isotopic fractionation between coexisting phosphate and structural carbonate in enamel for modern samples (Bryant et al., 1996; Iacumin et al., 1996). It is generally believed that carbonate in enamel is more likely than phosphate to undergo isotopic exchange with fluids during diagenesis (Lee-Thorpe and Van der Merwe, 1991; Ayliffe et al., 1994). Thus, measurements of δ18O (PO3− 4 ) and δ18O (CO2− 3 ) of biogenic apatite can be used to evaluate preservation of original isotopic signatures in fossil teeth (Fricke et al., 1998). Modern enamel samples from the Kunlun Pass Basin display a 18 3− difference of 8.6–9.1‰ between δ18O (CO2− 3 ) and δ O (PO4 ) values (Fig. 3), consistent with predicted values for their formation from the same body water (Iacumin et al., 1996; Bryant et al., 1996). The δ18O (CO2− 3 ) and δ18O (PO3− 4 ) values of fossil enamel samples are also plotted on or close to the equilibrium line (Fig. 3), suggesting little or no alteration of the isotopic ratios of either phase in the samples. XRD analyses of selected shells and sediment show that the sediment matrix contains quartz, calcite and at least one other phase that is possibly clay mineral kaolinite and/or montmorillonite. The shells, on the other hand, are either entirely aragonite or almost entirely aragonite (with a small trace of calcite), suggesting that diagenetic alteration, if any, has been minimal and these shells are likely retaining their original isotopic signatures. 4.2. Carbon isotopes, diets and habitats of modern and fossil herbivores Our recent study of modern plants on the Tibetan Plateau (Wang et al., 2008) reveals that all grasses found in the Kunlun Pass Basin and surrounding areas (including the Qaidam Basin) are C3 plants and have

Fig. 3. The oxygen isotopic compositions of coexisting phosphate (δ18O–PO3− 4 ) and carbonate (δ18O–CO2− 3 ) in tooth enamel from the Kunlun Pass Basin. The solid and dashed lines represent equilibrium relationships between the two phases (Longinelli and Nuti, 1973; Iacumin et al., 1996; Bryant et al., 1996).

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85

77

Fig. 4. Variations in the δ13C (a) and δ18O (b) values of herbivore tooth enamel (including both bulk and serial enamel samples) over time in the Kunlun Pass Basin, northern Tibetan Plateau.

δ13C values ranging from −23.2‰ to −28.3‰, with a mean of −25.5 ±1.1‰ (n = 31). The δ13C values of both serial and bulk enamel samples from modern horses and yaks from the Kunlun Pass Basin range from −10.6 to −14.8‰, with an average of −12.0 ±0.7‰ (n =69 enamel samples from 7 teeth), indicating pure C3 diets for modern herbivores in the area (Figs. 4a and 5a). Tooth-enamel samples from modern goats from Xi-DaTan (at a lower elevation) north of the Kunlun Pass (Fig. 1) have δ13C values ranging from −9.9 to −11.9‰ (Wang et al., 2008), with an average of − 10.9 ± 0.6‰ (n = 21), also indicating a C3-based diet. The δ13C values of tooth enamel from modern small mammals Arvicolidae and Ochotona in the area range from − 9.0 to − 11.5‰, with a mean of − 10.2 ± 1.2‰ (n = 5), indicating that modern small mammals also consumed C3 vegetation. The C3 diets of these various modern mammals are consistent with the current C3 dominance in high elevation ecosystems of the Tibetan Plateau (Wang et al., 2008). The slightly higher δ13C values of goat teeth compared to horse/yak teeth are caused by feeding on plants experiencing water stress (Wang et al., 2008). In contrast, enamel samples from fossil herbivores in the Kunlun Pass Basin yielded δ13C values of −5.4‰ to −10.2‰ (Fig. 4a), averaging −7.9± 1.3‰ (n=90 enamel samples from 31 specimens). These δ13C values are significantly higher than those from modern herbivores in the area (t-test, t=24.365, d.f.=157, M.D.=4.07, Pb 0.0001). Even after accounting for changes in the δ13C of atmospheric CO2 due to burning of fossil fuels, the difference in the mean δ13C values of enamel between modern and fossil herbivores in the Kunlun Pass Basin is still significant (t-test, t=18.377, d.f.=157, M.D.=3.07, Pb 0.0001). If we assume that C3 and C4 end-member enamel δ13C values are −12‰ and +2‰, respectively, the higher δ13C values for fossil teeth would suggest that these ancient herbivores consumed both C3 and C4 plants with C4 grasses accounting for ∼10–45% of their diets (Appendix A). However, the present-day environment on the Tibetan Plateau is mostly water-stressed. The high Himalayan mountain ranges serve as a topographic barrier preventing moist monsoonal airs from the Indian Ocean and the Bay of Bengal from entering the vast region on the north side of the high mountains. Our recent study of modern herbivores from the Tibetan Plateau shows that the “cut-off” enamel-δ13C value for a pure C3 diet within the Tibetan Plateau is −8‰ for modern herbivores due to the prevailing water-stressed conditions in the region (Wang et al., 2008). If water-stressed conditions

had existed in the area in the late Pliocene, the “cut-off” enamel-δ13C value for a pure C3 diet could be −7‰ for fossil herbivores after accounting for changes in the δ13C of atmospheric CO2 due to addition of 13C-depleted CO2 from burning of fossil fuels (Cerling et al., 1997; Wang et al., 2008). Using −7‰ as the end-member for a pure C3 diet, the fossil enamel-δ13C values would suggest that these ancient herbivores mostly had a C3-based diet and some individuals consumed a small amount of C4 plants with C4 grasses comprising less than 20% of their diets (Appendix A). Thus, uncertainties exist in using enamel-δ13C values to reconstruct the proportion of C3 and C4 plants in the diets of ancient herbivores depending on whether the Tibetan Plateau was as arid in the late Pliocene as it is today. The estimated amount of C4 plants in the diet of late Pliocene herbivores in the Kunlun Pass Basin ranges from as high as 10–45% to as low as 0–17% depending on the end-member enamel-δ13C value (for a pure C3 diet) used in the calculation (Appendix A). Although only limited fossil enamel samples yielded δ13C values higher than −7‰ (Fig. 4a), which are unambiguous evidence for mixed C3–C4 diets, there must have been enough C4 grasses in the Kunlun Pass Basin or nearby regions to support these animals with grazing (C4) adaptations around 2–2.5 Ma. As discussed below, the δ18O data suggest a much wetter and warmer environment in the basin in the late Pliocene. Therefore, the enamel end-member δ13C value for a pure C3 diet in the late Pliocene was most likely lower than −7‰. That is, the above estimates likely represent the upper and lower limits of the amount of C4 grasses in these ancient herbivores' diets. While it appears that C4 grasses may have existed in the local ecosystems, it should be noted that some of the fossil taxa (i.e., antelope and rhino) are represented by only one or two specimens (Appendix A). Until a larger number of specimens are analyzed, we cannot determine with absolute certainty whether the presence of C4 in the diet reflects the existence of C4 in local ecosystems or was due to migration of animals from other habitats where C4 grasses were present. Multiple teeth for each species or samples of non-migratory species would be needed in order to resolve the uncertainty. 4.3. Stable oxygen isotopes in tooth enamel from modern and fossil herbivores The δ18O values of enamel from modern horses and yaks from the Kunlun Pass Basin range from − 4.1 to − 11.4‰ (Fig. 4b), averaging

78

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

Fig. 5. Variations in the mean enamel-δ13C (a), enamel-δ18O (b) and local water δ18O (c) over time in the Kunlun Pass Basin. The estimated δ18O values of water were calculated from the δ18O values of enamel carbonate from large mammals including horse, yak and rhino using the equation given in Kohn and Cerling (2002).

−7.9 ± 1.3‰ (n = 69 enamel samples from 7 teeth). Small modern mammals (Arvicolidae and Ochotona) yielded a mean δ18O value of −7.4 ± 1.8‰ (n = 5). Enamel samples from modern goats from Xi-Da-Tan, north of the Kunlun Pass Basin, have δ18O values of −3.2‰ to −10.8‰, with an average of −6.5 ± 2.7‰ (n = 21 enamel samples from 2 teeth). In comparison, fossil enamel samples from large herbivores (i.e., horse and rhino) in the Kunlun Pass Basin have a δ18O range of −5.9‰ to −11.7‰, which is smaller than the observed δ18O range for modern large herbivores in the basin (Fig. 4b). Enamel from fossil bovids is generally enriched in 18O compared to contemporary horse and rhino, whereas fossil horse teeth are more depleted in 13C compared to other contemporary species (Fig. 4). The isotopic differences among species may suggest resource partitioning. The mean δ18O of enamel carbonate from fossil large herbivores (i.e., horses and rhinos) is −9.7± 1.8‰ (n = 50), significantly lower than the mean δ18O value of their modern counterparts (t-test, t = 6.0915, d.f. =117, M.D. = 1.7, P b 0.0001). Studies have shown that the δ18O of enamel from an obligate drinker generally tracks the δ18O of local water (e.g., Bryant and Froelich,1995; Delgado Huertas et al.,1995; Kohn and Cerling, 2002; Wang et al., 2008). Enamel-δ18O values of non-obligate drinkers such as goats are strongly affected by the δ18O of food plants and do not show a

strong relationship with the δ18O of water in a water-stressed environment such as the Tibetan Plateau (Wang et al., 2008). Thus, the differences in enamel-δ18O values between the late Pliocene and modern large herbivores or obligate drinkers in the Kunlun Pass Basin (Fig. 5b) most likely reflect differences in the δ18O of local meteoric water (e.g., Longinelli, 1984; Luz et al., 1990; Fricke et al., 1995; Kohn and Cerling, 2002; Wang et al., 2008). Stream and rain/hail samples collected in the summers of 2005, 2006 and 2007 in the Kunlun Pass area have δ18O values ranging from −7.7 to −11.9‰, with a mean of −10.0 ± 1.1‰, and δD values of −41.7 to −83.3‰, averaging −68.2± 10.2‰ (Table 1). These values fall on or close to the Global Meteoric Water Line, suggesting they had not been affected significantly by evaporation. However, the difference in precipitation δ18O between warm and cold months at the Kunlun Pass is currently unknown. Stream water, which is mostly derived from glacial or snow melts, is likely a good approximation of the annual precipitation in the area. By using the relationship given in Kohn and Cerling (2002), we calculated the δ18O values of local meteoric water from the δ18O of enamel (Fig. 5c). The mean water-δ18O value calculated from enamelδ18O values of modern horse and yak teeth is −10.2 ± 1.5‰, which is statistically the same as the mean δ18O value of −10.0 ± 1.1‰ for present-

Table 1 δ18O and δD values of water in the Kunlun Pass area Sample

QD-W-6 QD-W-5 QD-W-11 QD-W-2 QD-W-9 QD-W-3 QD-W-14 QD-W-14 QD-W-8 TB-W06-5 TB-W06-2 TB-W06-1 TB-W06-4 W07-6 W07-7 W07-8

δ18OVSMOW

δDVSMOW

Elevation

(%)

(%)

(m)

− 10.4 − 10.3 −9.2 −11.9 −9.8 − 10.4 −11.3 −9.4 −8.9 −9.6 −9.5 − 10.1 − 10.3 −11.4 −7.7

− 77.6 −71.6 − 58.5 − 78.1 − 67.9 −68.1 −71.2 − 70.7 − 78.4 − 60.6 −64.7 − 62.6 − 70.6 − 65.5 − 83.3 −41.7

4617 4546 4100 4021 4100 3458 4586 4586 4872 4691 4825 4773 4756 3556 4800 4681

Location

Sampling date

Sample type

N35°39′12.01″/E94°03′28.8″ N35°39′12.01″/E94°03′28.8″ N35°44.581′/E94°18.650′ N35°44.581′/E94°18.650′ N35°44.581′/E94°18.650′

7/16/05 7/16/05 7/17/05 7/18/05 7/15/05 7/19/05 7/19/05 7/19/05 7/20/05 8/25/06 8/26/06 8/26/06 8/27/06 8/6/07 8/7/07 8/7/07

Rain Stream Rain Stream Stream Stream Stream Stream Stream Stream Stream Stream Puddle Spring Hail Stream

N35°39′12.01″/E94°03′28.8″ N35°39′12.01″/E94°03′28.8″ N35°39′12.01″/E94°03′28.8″ N35°39′50.7″/E94°03′06.5″ N35°38′41.9″/E94°06′19.8″ N35°39′41.5″/E94°04′29.3″ N35°37′23.2″/E94°04′18.0″ N35°52′27.8″/E94°34′05.00″

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85

79

Fig. 6. Intra-tooth δ13C and δ18O variations in modern herbivore teeth from the Kunlun Pass Basin and nearby Xi-Da-Tan.

day water in the area collected in the summers of 2005, 2006 and 2007 (Table 1). The estimated mean water-δ18O value for the late Pliocene based on the oldest large mammals is −13.4 ± 0.9‰, significantly lower than that of modern waters (Fig. 5c). The uncertainty in paleo-water δ18O estimate corresponds to the 1 sigma (1σ) standard deviation in δ18O of enamel. Although serial samples of modern herbivore teeth from the Kunlun Pass Basin and Xi-Da-Tan in general show relatively small intra-tooth δ13C variations, the δ18O values of these samples display large intra-tooth variations within individual teeth (Fig. 6), reflecting

large seasonal variations in the δ18O values of ingested water (from streams, puddles, lakes, and water in plants). The intra-tooth δ18O variations in the fossil bovid and antelope teeth display the same pattern as observed in the modern goat tooth KLP-6 (Figs. 6 and 7), suggesting that the ancient antelope and other bovid examined here may have very similar physiology and diet/drinking behavior as the modern goat. The amplitude of seasonal δ18O variations in the modern goat tooth (7.3‰) is larger than observed in the modern horse teeth (2.2– 4.4‰) (Fig. 6). Similarly, the fossil bovid and antelope teeth display a

80

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

Fig. 7. Intra-tooth δ13C and δ18O variations in fossil teeth from the Kunlun Pass Basin.

much larger amplitude of intra-tooth δ18O variations (6.9–7.1‰) than the fossil horse and rhino teeth (1.6–2.2‰) (Fig. 7). These isotopic differences can be explained by differences in dietary and drinking behavior of these different animals. Large mammals such as horse and rhino are obligate drinkers and obtain a larger proportion of oxygen from drinking water than from plants that they eat compared to goats. Plants take up water from soils. The δ18O of soil water is highly variable depending on the δ18O of meteoric water input and evaporation. Leaf water is normally enriched in 18O compared to soil water due to evapotranspiration (Flanagan et al., 1991; Yakir and Yechieli, 1995). This enrichment effect increases with increasing aridity

(Dongmann et al., 1974; Flanagan et al., 1991). Thus, leaf water δ18O may have a larger range of variation than local rainwater and is controlled by local relative humidity as well as temperature. On the other hand, streams and lakes have δ18O values reflecting an average δ18O of local precipitation in the catchment area modified by other processes such as evaporation and mixing with groundwater, and therefore may have smaller seasonal δ18O variations than local precipitation (Fritz, 1981; Gonfiantini et al., 1998). Thus, animals that drink less or obtain a larger proportion of ingested water from plant leaves would have δ18O values that are influenced by local relative humidity and temperature (Ayliffe and Chivas, 1990; Luz et al., 1990).

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85

Tooth-enamel δ18O values from such animals would show larger or amplified seasonal δ18O variations than local precipitation; whereas enamel δ18O records from obligate drinkers would more closely reflect seasonal variations in the δ18O of meteoric water in the catchment. The difference in the amplitudes of intra-tooth δ18O variations between large and smaller mammals observed in this study is consistent with predictions based on physiological models (Bryant and Froelich, 1995; Kohn, 1996). In addition to diet and drinking, migration may also contribute to δ18O differences among species. In the modern northern Tibetan environment, the herds move to high alpine meadows during the summer, and in the winter the lower ranges are used for winter pasture. Migration could reduce the amplitude of seasonal δ18O variation recorded in a tooth. These behavioral factors are difficult to access. Our limited samples show that the amplitude of seasonal δ18O variations recorded in the fossil horse and rhino teeth (1.6–2.2‰) is smaller than that in modern horse teeth (2.2–4.4‰). The fossil bovid teeth also recorded a smaller seasonal signal (6.6–7.1‰) compared to that in the modern goat tooth (7.3‰) although the modern goat tooth was collected from an elevation about 500 m lower than the current elevation of the fossil locality. Goat teeth from higher elevations may display even larger seasonal variation in δ18O values. The larger amplitude of intra-tooth δ18O variations in modern horses and goats compared to their fossil counterparts, if confirmed by analysis of more samples, would suggest a stronger seasonality today compared to the late Pliocene. 4.4. Oxygen isotopes in fossil fish bones and teeth and gastropod shells The lake sediments in the Lower Qiangtang Formation at Locality KL0607 (2–2.5 Ma) yielded many fossil cyprinid (of probably a single species) fish bones and pharyngeal teeth whereas the sediments at Locality KL0402 (2–2.3 Ma) contain abundant ostracods and mollusk/ gastropod shells. The oxygen isotopic ratios of fish tooth/bone bioapatite and of mollusk/gastropod shells contain valuable information about the δ18O of lake water and water temperature, which are controlled by regional climate and hydrology. The δ18O (PO3− 4 ) values of bioapatite in fossil fish teeth and bones range from 16.1‰ to 18.9‰, with a mean of 17.5 ± 0.8‰ (n = 13). The

81

δ13C values of fossil gastropod shells range from −1.1 to 1.9‰, with a mean of 0.2 ± 1.1‰. The δ13C variations in gastropod shells primarily reflect variations in the δ13C of dissolved inorganic carbon in the lake water. The δ18O values of fossil shells have a mean of −5.6 ± 2.4‰ (n = 8) and range from −8.6 to −0.6‰. Bulk sediments associated with the shells yielded a mean δ13C value of 1.6 ± 3.6‰, ranging from −3.7 to 4.0‰, and a mean δ18O of −7.6 ± 3.6‰ (n = 5) with a range of −11.8 to −3.4‰. The δ18O values of carbonate in sediment matrix are generally lower than those of associated shells (Appendix A), likely due to isotopic re-equilibration with diagenetic fluid (Dickson and Coleman, 1980). Although we do not know the exact temperatures of the paleolake, abundant aquatic plant remains, mollusk/gastropod shells and fish bones/teeth suggest that the Kunlun Pass Basin, unlike today, was once a hospitable place for life and the lake water temperatures must have been above freezing (at least below surface) in order to support fish and other aquatic life. Comparison of ostracods and mollusk shells in the Qiangtang Formation with their modern analogues suggests that the water temperatures may be as high as 17 °C (Pang, 1982; Wu et al., 2001). Modern lakes on the Tibetan Plateau are mostly saline (Wei and Gasse, 1999). The largest lake in the region today is the Qinghai Lake located on the northeastern Tibetan Plateau at ∼ 3000 m a.s.l. The average temperature of the Qinghai Lake is about 10 °C for surface water and ∼ 4 °C for bottom water in the summer, and the lake is mostly frozen between December and March (Xu et al., 2006). If we assume that the paleo-lake in the Kunlun Pass Basin had a temperature range of 1 °C to 15 °C, the δ18O value of paleo-lake water can be calculated from the δ18O values of fossil fish bioapatite and aragonite shells (Fig. 8). The estimated δ18O values of paleo-lake water, based on the mean 18 δ O values of fish bioapatite and mollusk shells, range from about −5‰ to −8‰ and from −6‰ to −9‰, respectively, for the water temperature of 1–15 °C, with more negative water-δ18O values corresponding to lower temperatures (Fig. 8). These estimated water-δ18O values for the paleolake are much more negative than seawater, and the water in the Qinghai Lake and other lakes on the present-day northern Tibetan Plateau (Wei and Gasse, 1999; Wang et al., 2008), which suggests a freshwater lake environment, consistent with the botanical and invertebrate fossil evidence (Kong et al., 1982; Yin et al., 1996). The estimated

Fig. 8. δ18O relationships (a) between fish bioapatite and water based on the fractionation factor vs. temperature relationship given in Friedman and O'Neil (1977) modified from Longinelli and Nuti (1973) and (b) between aragonite shells and water (Kim et al., 2007), assuming that both fish tooth/bone bioapatite and shell aragonite were formed in isotopic equilibrium with water. The thick red line in each diagram represents the mean δ18O value of fish bioapatite (a) or shells (b) from the Lower Qiangtang Formation in the Kunlun Pass Basin and the width of the shaded area corresponds to one standard deviation from the mean δ18O of all samples. The arrows delineate the range of the paleo-lake water δ18O values estimated from the δ18O of fossil fish teeth/bones and mollusk shells, assuming lake water temperature ranged from 1 °C to 15 °C.

82

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

δ18O values for the paleo-lake water are higher than the estimated δ18O values for meteoric water based on the enamel-δ18O values of large mammals (Fig. 5c), reflecting evaporative enrichment of 18O in lake water (Gonfiantini, 1986). 4.5. δ 18 O of fossil mammalian bones and paleo-temperature reconstruction Mammalian bones are known to be very susceptible to diagenetic alteration due to their porous nature and have normally been considered unsuitable for paleo-climate study using stable carbon and oxygen isotopes (e.g., Wang and Cerling, 1994). Recent studies (Kohn and Law, 2006; Zanazzi et al., 2007) suggest that the δ18O of fossil bone carbonate is completely reset on timescales of tens of thousands of years to early diagenetic composition (comparable to palaeosol carbonate), reflecting the δ18O of local water and local temperature. Therefore, if the δ18O of local water could be estimated from the δ18O of tooth enamel, mean annual temperature may be calculated from the δ18O of local water, measured bone δ18O(CO2− 3 ), and the fractionation factor for CaCO3water. That is, it may be possible to distinguish changes in water composition from changes in temperature by measuring the δ18O values of both fossil tooth enamel and bone carbonate (Zanazzi et al., 2007). The δ18O(CO2− 3 ) values of fossil mammalian bones collected from our fossil localities are −8.5 to −10.6‰, with a mean of −9.8 ± 0.6‰ (n = 10). Using the approach of Zanazzi et al. (2007), we estimated the paleo18 temperatures from the δ18O(CO2− 3 ) values of fossil bones and the δ O of paleo-meteoric water derived from the δ18O of fossil tooth enamel from large herbivores (Fig. 9a). The estimated mean annual temperature for the late Pliocene is about 10 ± 8 °C, which is significantly higher than the present-day mean annual temperature of −6 °C to −7 °C in the area. It is important to note that this approach of using the δ18O of fossil bone carbonates as a paleo-thermometer (Zanazzi et al., 2007) assumes that the δ18O(CO2− 3 ) values of our fossil mammalian bones record early diagenetic conditions in near surface environment, reflecting local temperature as well as the δ18O of local meteoric water. Thus, the reliability of the paleo-temperature estimates depends on the validity of this underlying assumption. 4.6. Climatic and tectonic implications Despite the limited specimens, the enamel-δ18O values of large herbivores from the Kunlun Pass Basin display a significant shift to less

negative values (t-test, t=10.662, d.f.=93, M.D.=2.82, Pb 0.0001) while enamel-δ13C shifted to more negative values (t-test, t = 32.333, d.f. = 93, M.D. = 4.9, P b 0.0001) after ∼2–3 Ma (Fig. 5a,b). These carbon and oxygen isotopic shifts in tooth enamel indicate a significant change in local flora and climate after the late Pliocene (Fig. 5). Since the δ18O values of enamel from large mammals that are obligate drinkers are strongly correlated with the δ18O of local meteoric water (e.g., Kohn and Cerling, 2002; Wang et al., 2008), the δ18O difference between modern and fossil enamel most likely reflects the difference in the δ18O of local meteoric water. The mean δ18O value of meteoric water in the late Pliocene estimated from the enamel-δ18O values of the oldest large herbivores in the Kunlun Pass Basin is ∼3.2‰ more negative than that of the present-day water in the area (Fig. 5c), which cannot be explained by either elevation or temperature change alone as discussed below. The present-day Kunlun Pass Basin is essentially a desert within the permafrost zone, with a mean annual temperature of about −6 °C to −7 °C (Kong et al., 1982; Pang, 1982; Song et al., 2005; Liu et al., 2006). The more negative δ18O value of paleo-meteoric water and the presence of a freshwater lake with abundant aquatic life in the Kunlun Pass Basin in the late Pliocene clearly indicate a drastic change in local environment since then. If we assumed that the modern water-δ18O vs. elevation relationships applied to the past, this change in water δ18O would correspond to a decrease in elevation of 1067 ± 300 m since the late Pliocene using the average rate of −0.3‰/100 m observed in modern world (Poage and Chamberlain, 2001). Assuming the temperature gradient of −5 °C/km determined from the presentday elevations and mean annual temperatures of the Kunlun Pass and Linxia Basin (located at about the same latitude) applied to the past, this inferred elevation change would represent an increase in temperature of ∼ 5 °C in the Kunlun Pass Basin since the late Pliocene, which would suggest a mean annual temperature of −11 °C to −12 °C, well below freezing, in the basin in the late Pliocene, inconsistent with geological and botanical evidence (e.g., Kong et al., 1982; Yin et al., 1996) and our δ13C data from herbivores. If the elevation of the area had been constant since the late Pliocene, the δ18O difference between the late Pliocene and present-day water would imply that the temperatures in the basin were ∼6 °C lower in the late Pliocene than today using the rate of 0.58‰/°C observed for present-day precipitation at mid to high latitudes (Rozanski et al., 1993), also in conflict with our enamel-δ13C data and other evidence for a warmer climate in the Pliocene (e.g., Kong et al., 1982; Yin et al., 1996). Clearly,

Fig. 9. Mean annual temperature and paleo-elevation of the Kunlun Pass basin estimated on the basis of the δ18O values of fossil bone carbonate and water, assuming that the δ18O of fossil bone carbonate is completely reset to reflect early diagenetic conditions in near surface environment (Zanazzi et al., 2007) and the present-day temperature gradient of − 5 °C/ km in the region applied to the past. The δ18O of water was estimated from the δ18O of tooth enamel. The lines in (a) represent equilibrium δ18O relationships between calcite and water at various temperatures based on the fractionation factor vs. temperature relationship given in Kim and O'Neil (1997).

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85

the modern relationships between the δ18O of precipitation and elevation or temperature do not apply to the distant past because other factors (e.g., precipitation amount, source and rainout history of atmospheric moisture, and monsoon strength) can also have a local influence on the δ18O of precipitation (Wang et al., 2008). The shift in water-δ18O in the basin since the late Pliocene is most likely the result of significant changes in the regional hydrological cycle such as increased aridity and/or changes in source and rainout history of atmospheric moisture. The present-day climate in the Tibetan region is influenced by Asian monsoons (Araguas-Araguas et al., 1998; Thompson et al., 2000). The differential heating between the Indian and Pacific Oceans and the high plateau drives the intense monsoon circulation and strongly influences the global circulation patterns (Webster, 1987). The Indian and East Asian summer monsoons are the main sources of moisture for the Himalayan–Tibetan region and a significant factor influencing the δ18O of meteoric water (Araguas-Araguas et al., 1998; Johnson and Ingram, 2004; Vuille et al., 2005; Tian et al., 2007). In the winter, moisture can be carried by the westerly winds, with the moisture most likely originating in the northern Atlantic Ocean and augmented by evaporation from the Mediterranean Sea (Thompson et al., 2000). Condensation preferentially removes heavy isotope 18O from vapor, resulting in a progressive 18O-depletion in the remaining vapor and subsequent precipitation as the air mass moves away from its source area. Recycling of moisture can also affect the δ18O of precipitation (Araguas-Araguas et al., 1998; Tian et al., 2007). The present-day high Himalayan Mountain ranges block most of the moisture from the Indian Ocean and the Bay of Bengal. Limited precipitation δ18O data from the Tibetan Plateau show that the present-day northern limit of the Indian monsoon influence appears to be located near Yushu, south of the Kunlun Mountain Pass, and therefore the Indian monsoon is not likely a major source of moisture for the Kunlun Pass Basin on the northern Tibetan Plateau (AraguasAraguas et al., 1998; Tian et al., 2001, 2007). However, if the Himalayan–Tibetan Plateau was not as formidable a barrier in the late Pliocene as it is today, a greater amount of moisture derived from the Indian Ocean could have been carried by the Indian monsoon farther inland into the Kunlun Pass Basin, which could explain the more negative water-δ18O values. The presence of a freshwater lake and the abundance of fossil vegetation in the Qiangtang Formation also imply that water was plentiful in the basin during the late Pliocene. Serial samples from 6 fossil teeth or fragments show relatively small intra-tooth δ13C variations (b2‰) but large intra-tooth δ18O variations of up to 7.1‰, reflecting seasonal variations in the isotopic compositions of diet and water. There exists an anti-correlation between δ13C and δ18O values with higher δ13C values corresponding to lower δ18O values in individual fossil teeth except rhino teeth (i.e., R2 = 0.49 for bovid tooth KLP-17, R2 = 0.81 for antelope tooth KL-XW-3, R2 = 0.66 for horse tooth KL-XW-1, R2 = 0.0.1–0.2 for rhinos KL-XW-2 and KL-YWCF-3) (Fig. 7). Such anti-correlation observed in individual fossil teeth (Fig. 7) is characteristic of Asian summer monsoon regions where C4 grasses grow. In regions that are strongly influenced by the East Asian summer monsoon and the Indian monsoon, summer precipitation has lower δ18O values than winter precipitation (Araguas-Araguas et al., 1998; Johnson and Ingram, 2004), resulting in lower δ18O values in enamels formed during summer months compared to enamels formed in winter months. Since C4 grasses are warm season grasses or summer grasses, the higher enamel-δ13C values also represent summer months where C4 grasses were available for consumption. Thus, the intra-tooth δ13C variations observed in the fossil teeth likely reflect seasonal variations in the availability/abundance of C4 grasses in the basin or nearby habitats. Our carbon isotope data from both serial and bulk enamel samples from fossil teeth suggest that C4 grasses were likely present in local or nearby ecosystems at the end of the Pliocene, around 2.0–2.5 Ma (Fig. 4a), consistent with grassland and forest mosaics represented in pollen

83

analysis (Yin et al.,1996). The isotopic variations among different species indicate mixed habitats (i.e., grassland and woodland) occupied and partitioned by different species, different from present-day habitats. The anti-correlation between δ13C and δ18O values observed in the fossil teeth suggests that summer monsoons were a major source of moisture for the basin in the late Pliocene. Our results have important implications for the tectonic evolution of the Tibetan Plateau and its role in controlling regional and global climate. Our carbon isotope data suggest that C4 grasses likely existed in local or nearby habitats in the late Pliocene, implying a warmer climate then, quite different from modern conditions. The presence of a shallow freshwater lake with abundant fish and other aquatic life in the late Pliocene, as suggested by our δ18O data (Fig. 8) and fossil evidence (Kong et al., 1982; Yin et al., 1996), also implicates that the temperatures in the basin must have been mostly above freezing ∼2– 3 Ma (i.e., warmer than today). Studies of pollens, ostracods and mollusk/gastropod shells in the Qiangtang Formation suggest that the lake water temperatures were about 10 °C and may be as high as 17 °C (Pang, 1982; Wu et al., 2001). As discussed in the previous section, we also estimated the paleo18 temperatures from the δ18O(CO2− 3 ) of fossil bones and the δ O of paleometeoric water derived from the δ18O of fossil tooth enamel from large herbivores (Fig. 9a), using the approach of Zanazzi et al. (2007). The estimated mean annual temperature for the late Pliocene is about 10 ± 8 °C, which is about 16–17 °C (±8 °C) higher than the present-day mean annual temperature in the Kunlun Pass Basin (Fig. 9b). This provides further evidence for a much warmer climate in the area in the late Pliocene. Although reliance on the δ18O of fossil bone carbonate as a paleo-thermometer entails an assumption (see previous section) that has yet to be validated, our paleo-temperature estimates (Fig. 9a) are broadly compatible with those from aquatic plants, ostracods and mollusk shells (∼10–17 °C). Assuming that (1) the temperature gradient of −5 °C/km determined from the present-day conditions of the Linxia and Kunlun Pass Basins applied to the past and (2) a temperature drop of 3 °C in the area was due to global cooling since the Pliocene (Ravelo et al., 2004), the estimated temperature change in the basin would correspond to an elevation change of ∼2700 ± 1600 m since the late Pliocene. This would imply that the elevation of the Kunlun Pass Basin in the late Pliocene was ∼2011 ± 1600 m a.s.l., much lower than its present-day elevation (Fig. 9b). 5. Conclusions Stable carbon and oxygen isotope analyses of both terrestrial and aquatic fossils reveal a drastic change in habitat and hydrological regime in the Kunlun Pass Basin since the late Pliocene. The δ13C values of both serial and bulk enamel samples from fossil herbivore teeth suggest that C4 grasses (i.e., warm climate grasses) were likely present in local ecosystems at the end of the Pliocene, around 2.0– 2.5 Ma. The carbon isotopic variations among different species indicate mix habitats, including grasslands and wooded grasslands, occupied and partitioned by different species, consistent with palynological evidence. The anti-correlation between δ13C and δ18O values observed in the fossil teeth suggests that summer monsoons (i.e., the East Asian summer monsoon or the Indian monsoon, or both) were a major source of moisture for the area in the late Pliocene. The more negative enamel-δ18O values of large herbivores in the late Pliocene suggest that paleo-meteoric water then was more depleted in 18 O compared to the present-day meteoric water in the basin. The most likely cause for this δ18O shift in tooth enamel or water after the late Pliocene is a drastic change in the regional hydrological cycle (e.g., change in source and rainout history of atmospheric moisture or atmospheric circulation pattern, increasing aridity, and etc.) possibly due to tectonic and climate change. Our carbon and oxygen isotope data, in conjunction with geological/fossil evidence, suggest that the Kunlun Pass Basin had a much warmer and wetter climate in the late

84

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73-85

Pliocene, quite different from modern conditions. The paleo-temperature estimates based on the δ18O values of fossil bones and paleometeoric water, if valid, would imply that the present-day high elevation of the basin was established after 2–3 Ma. Acknowledgments This study was funded by the U.S. National Science Foundation (EAR-0444073), Chinese Academy of Sciences (KZCX2-YW-120) and Chinese National Science Foundation (NSFC 40730210). All isotope analyses were performed at the Florida State University Stable Isotope Laboratory supported by grants from the U.S. National Science Foundation (EAR-0517806 and EAR-0236357). We thank Dr. Eric Lochner for help with XRD analyses. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.epsl.2008.03.006. References Araguas-Araguas, L., Froehlich, K., Rozanski, K., 1998. Stable isotope composition of precipitation over southeast Asia. J. Geophys. Res. 103, 28721–28742. Ayliffe, L.K., Chivas, A.R., 1990. Oxygen isotope composition of the bone phosphate of Australian kangaroos: potential as a palaeoenvironmental recorder. Geochim. Cosmochim. Acta 54, 2603–2609. Ayliffe, L., Chivas, A., Leakey, M., 1994. The retention of primary oxygen isotope compositions of fossil elephant skeletal phosphate. Geochim. Cosmochim. Acta 58, 5291–5298. Bryant, D., Froelich, P., 1995. A model of oxygen isotope fractionation in body water of large mammals. Geochim. Cosmochim. Acta 59, 4523–4537. Bryant, D., Koch, P., Froelich, P., Showers, W., Genna, B., 1996. Oxygen isotope partitioning between phosphate and carbonate in mammalian apatite. Geochim. Cosmochim. Acta 60, 5145–5148. Cerling, T.E., Harris, J., MacFadden, B., Leakey, M., Quade, J., Eisenmann, V., Ehleringer, J., 1997. Global vegetation change through the Miocene/Pliocene boundary. Nature 389, 153–158. Delgado Huertas, A., Iacumin, P., Stenni, B., Sanchez Chillon, B., Longinelli, A., 1995. Oxygen isotope variations of phosphate in mammalian bone and tooth enamel. Geochim. Cosmochim. Acta 59, 4299–4305. Dettman, D., Kohn, M., Quade, J., Ryerson, F., Ojha, T., Hamidullah, S., 2001. Seasonal stable isotope evidence for a strong Asian monsoon throughout the past 10.7 m.y. Geology 29, 31–34. Dickson, J., Coleman, M., 1980. Changes in carbon and oxygen isotope composition during limestone diagenesis. Sedimentology 27 (1), 107–118. Dongmann, G., Nurnberg, H., Forstel, H., Wagener, K., 1974. On the enrichment of H18 2 O in the leaves of transpiring plants. Radiat. Environ. Biophys 11, 41–52. Dutkiewicz, A., Herczeg, A., Dighton, J., 2000. Past changes to isotopic and solute balances of lacustrine carbonates. Chem. Geol. 165, 309–329. Flanagan, L., Comstock, J., Ehleringer, J., 1991. Comparison of modeled and observed environmental influences on the stable oxygen and hydrogen isotope composition of leaf water in Phaseolus vulgaris L. Plant Physiol. 96, 588–596. Friedman, I., O'Neil, J., 1977. Compilation of stable isotope fractionation factors of geochemical interest, In: Fleischer, M. (Ed.), Data of Geochemistry, sixth ed. U.S. Geol. Surv. Prof. Paper 440-KK. Chapter KK 12 pp. and 49 figures. Fricke, H., O'Neil, J., 1996. Inter- and intra-tooth variation in the oxygen isotope composition of mammalian tooth enamel phosphate: implications for palaeoclimatological and palaeobiological research. Palaeogeogr. Palaeoclimatol. Palaeoecol. 126, 91–99. Fricke, H.C., O'Neil, J.R., Lynnerup, N., 1995. Oxygen isotope composition of human tooth enamel from medieval Greenland: linking climate and society. Geology 23, 869–872. Fricke, H., Clyde, W., O'Neil, J., 1998. Intra-tooth variations in δ18O (PO4) of mammalian tooth enamel as a record of seasonal variations in continental climate variables. Geochim. Cosmochim. Acta 62, 1839–1850. Fritz, P., 1981. River waters. In: Gat, J., Gonfiantini, R. (Eds.), Stable Isotope Hydrology, Deuterium and Oxygen-18 in the Water Cycle. IAEA Tech. Rep. Series, vol. 210, pp. 177–201. Fritz, P., Morgan, A., Eicher, U., McAndrews, J., 1987. Stable isotope, fossil coleoptera and pollen stratigraphy in Late Quaternary sediments from Ontario and New York state. Palaeogeogr. Palaeoclimatol. Palaeoecol. 58, 183–202. Gasse, F., Fontes, J., 1987. Paleoenvironments and paleohydrology of a tropical closed lake (Lake Asal, Djibouti) since 10,000 yr BP. Palaeogeogr. Palaeoclimatol. Palaeoecol. 69, 67–102. Gonfiantini, R., 1986. Environmental isotopes in lakes studies. In: Fritz, P., Fontes, J. (Eds.), Handbook of Environmental Isotope Geochemistry: The Terrestrial Environment. Elsevier, Amsterdam, pp. 113–168. Gonfiantini, R., Frohlich, K., Araguas-Araguas, L., Rozanski, K., 1998. Isotopes in groundwater hydrology. In: Kendall, C., McDonnell, J. (Eds.), Isotope Tracers in Catchment Hydrology. Elserier, Amsterdam, pp. 203–246.

Iacumin, P., Bocherrens, H., Mariotti, A., Longinelli, A., 1996. Oxygen isotope analyses of co-existing carbonate and phosphate in biogenic apatite: a way to monitor diagenetic alteration of bone phosphate? Earth Planet. Sci. Lett. 142, 1–6. Johnson, K.R., Ingram, B.L., 2004. Spatial and temporal variability in the stable isotope systematics of modern precipitation in China: implications for paleoclimate reconstructions. Earth Planet. Sci. Lett. 220, 365–377. Kidd, W., Molnar, P., 1988. Quaternary and active faulting observed on the 1985 Academia Sinica-Royal Society Geotraverse of Tibet. Phil. Trans. R. Soc. Lond. A 327, 337–363. Kim, S.T., O'Neil, J.R., 1997. Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochim. Cosmochim. Acta 61, 3461–3475. Kim, S.T., O'Neil, J.R., Hillaire-Marcel, C., Mucci, A., 2007. Oxygen isotope fractionation between synthetic aragonite and water: influence of temperature and Mg2+ concentration. Geochim. Cosmochim. Acta 71, 4704–4715. Koch, P., Heisinger, J., Moss, C., Carlson, R., Fogel, M., Behrensmeyer, A., 1995. Isotopic tracking of change in diet and habitat use in African elephants. Science 267, 1340–1343. Kohn, M.J., 1996. Predicting animal δ18O: accounting for diet and physiological adaptation. Geochim. Cosmochim. Acta 60, 4811–4829. Kohn, M., Cerling, T.E., 2002. Stable isotope compositions of biological apatite. In: Kohn, M., Rakovan, J., Hughes, J. (Eds.), Phosphates — Geochemical, Geobiological, and Materials Importance. Reviews in Mineralogy & Geochemistry, vol. 48. Mineralogical Society of America, Washington D.C., pp. 455–488. Kohn, M., Law, J., 2006. Stable isotope chemistry of fossil bone as a new paleoclimate indicator. Geochim. Cosmochim. Acta 70, 931–946. Kolodny, Y., Luz, B., Navon, O., 1983. Oxygen isotope variations in phosphate of biogenic apatites, I. Fish bone apatite — rechecking the rules of the game. Earth Planet. Sci. Lett. 64, 398–404. Kong, Z.C., Liu, L.S., Du, N.Q., 1982. Neogene–Quaternary Palynoflora from the Kunlun to Tanggula Ranges and the uplift of the Qinghai–Xizang Plateau, in: study on the timing, magnitude and Type of the Uplift of the Qinghai–Xizang Plateau (ed., The Comprehensive Scientific Expedition to the Qinghai–Xizang Plateau, Academia Sinica) (in Chinese), Beijing, Science Press, p. 78–88. Lecuyer, C., Grandjean, P., Sherppard, S., 1999. Oxygen isotope exchange between dissolved phosphate and water at temperatures b135 °C: inorganic versus biological fractionations. Geochim. Cosmochim. Acta 63, 855–862. Lee-Thorpe, J., Van der Merwe, N., 1991. Aspects of the chemistry of modern and fossil biogenic apatites. J. Archaeol. Sci. 18, 343–354. Lin, A.-m., Fu, B., Guo, J., Zeng, Q., Dang, G., He, W., Zhao, Y., 2002. Co-seismic strike–slip and rupture length produced by the 2001 Ms 8.1 Central Kunlun earthquake. Science 296, 2015–2017. Liu, G.N., Zhang, X.Y., Cui, Z.C., Wu, Y.Q., Ju, Y.J., 2006. A review of glacial sequences of the Kunlun Pass, northern Tibetan Plateau. Quat. Int. 154–155, 63–72. Longinelli, A., 1984. Oxygen isotopes in mammal bone phosphate: a new tool for paleohydrological and paleoclimatological research? Geochim. Cosmochim. Acta 48, 385–390. Longinelli, A., Nuti, S., 1973. Revised phosphate-water isotopic temperature scale. Earth Planet. Sci. Lett. 19, 373–376. Luz, B., Cormie, A., Schwarcz, H., 1990. Oxygen isotope variations in phosphate of deer bones. Geochim. Cosmochim. Acta 54, 1723–1728. O'Neil, J., Roe, L., Reindhard, E., Blake, R., 1992. A rapid and precise method of oxygen isotope analysis of biogenic phosphate. Isr. J. Earth Sci. 43, 203–212. Pang, Q., 1982. The geological significance of the ostracoda from the Quaternary Qiangtang Formation at the mouth of Kunlun Mountains on the Tibet Plateau, in: Contribution to the Geology of the Tibet Plateau (ed., The Editorial Committee on the Tibet Plateau Geological Papers) (in Chinese), Beijin, Geological Publishing House, p. 151–165. Poage, M., Chamberlain, C.P., 2001. Empirical relationships between elevation and the stable isotope composition of precipitation and surface waters: considerations for studies of paleoelevation change. Am. J. Sci. 301, 1–15. Quade, J., Cerling, T., Barry, J., Morgan, M., Pilbeam, D., Chivas, A., Lee-Thorp, J., van der Merwe, N., 1992. A 16-Ma record of paleodiet using carbon and oxygen isotopes in fossil teeth from Pakistan. Chem. Geol. 94, 183–192. Ravelo, A., Andreasen, D., Lyle, M., Lyle, A., Wara, M., 2004. Regional climate shifts caused by gradual global cooling in the Pliocene epoch. Nature 429, 263–267. Rozanski, K., Araguas-Araguas, L., Gonfiantini, R., 1993. Isotopic patterns in modern global precipitation. In: Swart, P., Lohmann, K., McKenzie, J., Savin, S. (Eds.), Climate Change in Continental Isotopic Records. Geophysical Monograph, vol. 78. American Geophysical Union, Washington, DC, pp. 1–36. Shemesh, A., Kolodny, Y., Luz, B., 1988. Isotope geochemistry of oxygen and carbon in phosphate and carbonate of phosphorite francolite. Geochim. Cosmochim. Acta 52, 2565–2572. Song, C., Gao, D., Fang, X., Cui, Z., Li, J., Yang, S., Jin, H., Burbank, D., Kirschvink, J., 2005. Late Cenozoic high-resolution magnetostratigraphy in the Kunlun Pass Basin and its implications for the uplift of the northern Tibetan Plateau. Chin. Sci. Bull. 50,1912–1922. Talbot, M., 1990. A review of the paleohydrological interpretation of carbon and oxygen isotopic ratios in primary lacustrine carbonates. Chem. Geol. 80, 261–279. Tian, L., Masson-Delmotte, V., Stievenard, M., Yao, T., Jouzel, J., 2001. Qinghai–Tibetan Plateau summer monsoon northward extent revealed by measurements of water stable isotopes. J. Geophys. Res. 106, 28081–28088. Tian, L., Yao, T., MacClune, White, J., Schilla, A., Vaughn, B., Vachon, R., Ichiyanagi, K., 2007. Stable isotopic variations in west China: a consideration of moisture sources. J. Geophys. Res. 112, D10112. doi:10.1029/2006JD007718. Thompson, L., Yao, T., Mosley-Thompson, E., Davis, M., Henderson, K., Lin, P., 2000. A high-resolution millennial record of the South Asian monsoon from Himalayan ice cores. Science 289, 1916–1919.

Y. Wang et al. / Earth and Planetary Science Letters 270 (2008) 73–85 Turner, J., Fritz, P., Karrow, P., Warner, B., 1983. Isotopic and geochemical composition of marl lake waters and implications for radiocarbon dating of marl lake sediments. Can. J. Earth Sci. 20, 599–615. Vuille, M., Werner, M., Bradley, R., Keimig, F., 2005. Stable isotopes in precipitation in the Asian monsoon region. J. Geophys. Res. 110. doi:10.1029/2005JD006022. Wang, Y., Cerling, T.E., 1994. A model of fossil tooth and bone diagenesis: implications for paleodiet reconstruction from stable isotopes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107, 281–289. Wang, Y., Deng, T., 2005. A 25 m.y. isotopic record of paleodiet and environmental change from fossil mammals and paleosols from the NE margin of the Tibetan Plateau. Earth Planet. Sci. Lett. 236, 322–338. Wang, X., Qiu, Z., Li, Q., Wang, Y., Tseng, J., 2006. A new vertebrate fauna in late Pliocene of Kunlun Mountain Pass, northern Tibetan Plateau and its paleoenvironmental implications. J. Vertebr. Paleontol. 26 (3, suppl.), 136A. Wang, Y., Kromhout, E., Zhang, C., Xu, Y., Parker, W., Deng, T., Qiu, Z., 2008. Stable isotopic variations in modern herbivore tooth enamel, plants and water on the Tibetan Plateau: Implications for paleoclimate and paleoelevation reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology 260, 359–374. Webster, P., 1987. The elementary monsoon. In: Fein, J. (Ed.), Monsoons. Wiley, New York, pp. 3–32.

85

Wei, K., Gasse, 1999. Oxygen isotopes in lacustrine carbonates of West China revisited: implications for post glacial changes in summer monsoon circulation. Quat. Sci. Rev. 18, 1315–1334. Wu, Y., Cui, Z., Liu, G., Ge, G., Yin, J., Xu, Q., Pang, Q., 2001. Quaternary geomorphological evolution of the Kunlun Pass area and uplift of the Qinghai–Xizang (Tibet) Plateau. Geomorphology 36, 203–216. Xu, H., Ai, L., Tan, L., An, Z., 2006. Stable isotopes in bulk carbonates and organic matter in recent sediments of Lake Qinghai and their climatic implications. Chem. Geol. 235, 262–275. Yakir, D., Yechieli, Y., 1995. Plant invasion of newly exposed hypersaline Dead Sea shores. Nature 374, 803–805. Yin, J., Cui, Z., Ge, D., Liu, D., Wu, Y., 1996. Paleoecologial analysis of Quaternary fossil assemblages from the Kunlun Pass area, and geological significance for Kunlun Mountain rising. Earth Sci.- J. China Univ. Geosci. (in Chinese) 21 (3), 243–248. Zazzo, A., Smith, G., Patterson, W., Dufour, E., 2006. Life history reconstruction of modern and fossil sockeye salmon (Oncorhynchus nerka) by oxygen isotopic analysis of otoliths, vertebrae, and teeth: implication for paleoenvironmental reconstructions. Earth Planet. Sci. Lett. 249, 200–215. Zanazzi, A., Kohn, M., MacFadden, B., Dennis, O., Terry, D., 2007. Large temperature drop across the Eocene–Oligocene transition in central North America. Nature 445, 639–642.