Organic Carbon Isotope Records of Paleoclimatic Evolution since the Last Glacial Period in the Tangjia Region, Tibet

INTRODUCTION

In 1975, Stuiver firstly reported the relationship between the changes in δ¹³C within lake sediments and the ancient climate because of the varying utilization of carbon isotopes by different varieties of plants during photosynthesis process (Stuiver, 1975). Subse-quently, this relationship has allowed for the reconstruction of paleoclimates (Ning et al., 2008; Wynn, 2007; Zhou et al., 2006, 1997; Schidlowski, 2001; Kelly et al., 1998; Guillet et al., 1988).

Since then, many researchers have studied lake sediments in northern Tibet and organic carbon iso-topes found in pollen from southern Tibet to reconstruct probable paleoclimates for these regions (Wang and Zhu, 2007; Wu Y H et al., 2007; Wu Z H et al., 2006, 2004, 2003; Zheng et al., 2006; Chen et al., 2004; Li, 2004; Tang et al., 2004; Yu et al., 2004; Zhao et al., 2003, 2002; Sun et al., 1993). These data, along with tree-ring data, have revealed the scale of change in climate and vegetation within the region (Liu X H et al., 2003; Wu, 1990).

In recent years, the resolution scale for the paleo-climate research has been refined to 1 000-year or even 100-year intervals based on the findings from ice core research (Yao et al., 2006, 2000) which has played a significant role in the study of global climate change.

The Lhasa area is located in the central of south-ern region of Qinghai-Tibetan plateau. It exhibits a special terrain topography and environment setting that may have led to different paleoclimatic features compared with those found in other regions of Tibet.

Chen et al. (2008) chose the Lhasa River terraces as a paleoclimate study area and selected phytoliths as climatic indicators, thus establishing a sequence of past climate change within the Lhasa region over the past 5 000 years.

For this study, the Tangjia region in the Lhasa area was chosen for study to assess changes in climate and vegetation since the last glacial period by utilizing climate indicators such as δ¹³C, total organic carbon (TOC), and Rb/Rs ratios together with other dating information.

AN OVERVIEW OF THE STUDY AREA Geography

The study area, which is under the jurisdiction of Maizhokunggar County, Tibet Autonomous Region (Fig. 1), is located in a valley of the upper Lhasa River, approximately 80 km east of the city of Lhasa. The area is classified as a semi-arid plateau having a temperate monsoon climate with an average annual rain-fall of approximately 515.7 mm. Stratigraphic distribution was mainly Late Triassic and Quaternary (Fig. 1), with the Quaternary including pluvial-alluvial and alluvial sediments distributed mainly in the Lhasa and Mozhu rivers.

Figure  1.  Geology map for the Tangjia region in Lhasa, Tibet (after Xie et al., 2009). 1. Luobadui Group, limestone, marble, and biological clastic limestone; 2. Mailonggang Group, micrite and finely crystalline limestone; 3. Yeba Group, volcanic breccias; 4. Linbuzong Group, slate and siltstone; 5. Chumulong Group, sandstone and siltstone; 6. Dianzhong Group, andesite; 7. sand-gravel and sandsoil in the diluvium-alluvial; 8. alluvial-sand and gravel; 9. biotite adamellite from the Late Cretaceous; 10. fault; 11. geological bound-ary; 12. water system; 13. location of the section.

DownLoad: Full-Size Img PowerPoint

Section Description

The study section is located in the Tangjia region of Lhasa at an elevation of 3 850 m. The geographic coordinates were: 29°54'30.6"N, 91°48'58.5"E, 3 850 m height.

The surface of the first terrace is approximately 150-180 m wide and 135 cm above the river. The ter-race scarp was 50 cm from the river and the terraces are planted with barley. The lower layer is composed of small pebbles and thick gravels, the middle layer consists of cobblestones, and the upper layer with fine sandstone and gravel is covered by a brown-yellow silt (Fig. 2). Near the bottom of the gravels, water begins to appear.

Figure  2.  Section of the river terrace at Tangjia Village.

DownLoad: Full-Size Img PowerPoint

The second terrace is 4 m above the river, with a bottom layer composes of fine and coarse yellow-brown gravels, the middle layer with coarse sands and fine gravels, and the top layer with a brown-yellow silt layer that is 100 cm thick with interbed of a 0.5 cm muddy silts. Significantly, the thick sands little by little appear at the bottom of the gravels, whose structure and composition are approximate to the top layer of the third terrace. Moreover, it is difference to the gravels that are semi-weathered than the bottom of the second terrace.

The third terrace is 5 m above the river and has a composition similar to the first and second terraces; the most characteristics are the gravels that are subjected to weathering stronger in the upper layer than the other.

The authors survey and analyze three terraces from topography and sedimentary structures, the results show that the sections are more continuous and their sequences are clearly identified.

MATERIALS AND METHODS

This study focused on samples of the third terrace in the Tangjia region, although samples from all the terraces were collected.

The First Terrace and Flood Plain

Four samples were obtained from the base of the upper layer of the first terrace at 20 cm intervals; 12 samples were obtained from the top silt layer at 5 cm intervals. The flood plain consisted of a silt layer and is located at 30 cm above the river. Three samples were obtained from this location at 10 cm intervals.

The Second Terrace

The lower layer of the second terrace was com-posed of 300 cm thick coarse gravels; 6 samples were obtained at 50 cm intervals. The upper layer was composed of 100 cm thick aleurolite; 10 samples were obtained at 10 cm intervals.

The Third Terrace

The lower layer of the third terrace was com-posed of 400 cm thick coarse gravels. Eight samples were obtained from the lower layer at 50 cm intervals. The upper layer consisted of 100 cm thick aleurolite. Twelve samples were obtained from the base of the upper layer at 5 cm intervals. Four samples were ob-tained from the top of the upper layer at 10 cm intervals.

Fifty-nine samples were obtained. The δ¹³C, TOC, and Rb/Sr ratios were determined from these samples. δ¹³C contents in the samples were analyzed by the State Key Laboratory of Geological Processes and Mineral Resources at China University of Geosciences at Wuhan. Samples collected in the field were first dried at 80 ℃ and then pulverized to 100 meshes based on the carbon content of the samples. Samples of 5 g each were placed in a clean beaker and then mixed with an appropriate amount of hydrochloric acid. The reaction took 24 h to remove the carbonate from the samples.

The resulting samples were rinsed with distilled water and centrifuged at high speed and low tempera-ture, and any remaining clear liquid was drained off. This process was repeated abundant times for removing contaminants and neutralizing the samples, which were then dried in a vacuum oven at 40–60 ℃. A small amount of sample (typically 1–2 g) was loaded into a quartz reactor connected to a vacuum system. Pt was added as a catalyst. The gas mixture was slowly extracted by a mechanical pump to obtain a pressure under 2.0×10^-2 Pa. When the temperature of the quartz reactor reached 850 ℃, CO₂ produced was collected and purified with liquid nitrogen, which was analyzed using a MAT 251 mass spectrometer; the instrumental error was less than 0.2‰.

TOC in samples were analyzed by the SINOPEC Research Institute of Petroleum Geology Exploration and Development Test Center, Jianghan Oil Field Company. Samples were first air-dried and then pul-verized to less than 80 meshes. Then, 5% hydrochloric acid was added to 1–2 g of each sample. After 10 h, the sample mixtures were rinsed with distilled water, dried at temperatures less than 50 ℃, and then pul-verized to 150 meshes. About 1.8 mg of each sample was weighed out. TOC samples were assayed using CE440-type elemental analysis.

Rb and Sr contents were analyzed by the Super-vision and Testing Center at Wuhan, Ministry of Land and Mineral Resources. X-ray fluorescence spec-trometry (XRF) was used. The samples were first air-dried. About 10 g samples were trimmed with low-pressure polyethylene, and then slice was made by press-machine. A Philips PW2440 Wavelength Dispersive X-ray Fluorescence Spectrometer was used to test the Rb and Sr content; the measurement error was less than 1.8 μg.

The ¹⁴C ages of the samples were collected from this section and tested by the Accelerator Mass Spec-trometry Laboratory and the Laboratory of Quaternary Dating in Peking University. The half-life of the ¹⁴C used was 5 568 years. The tree-ring calibration curve was IntCa104 (1), and the procedure used was Ox-Cal3.10 (2).

RESULTS

Organic carbon isotope analysis showed that δ¹³C was between -18.82‰ and -22.89‰ (Table 1), and the average was -20.82‰. Perhaps influenced by modern vegetation roots, the maximum value of -18.82 ‰ ap-peared at the top of the first terrace, and the minimum value of -22.89‰ appeared in the upper silt and sand layer of the third terrace.

Table  1.  δ¹³C, TOC, and Rb/Sr analysis results

 | Show Table

DownLoad: CSV

The δ¹³C analysis showed that the soil organic carbon isotopes of this profile changed little from bottom to top. The result was higher in δ¹³C than modern topsoil from the Qinghai-Tibetan plateau (Lü et al., 2001).

The TOC analysis showed fluctuating values from 0.11% to 1.02%, and the average value was 0.33%, which was lower than that of other lakes in the Tibetan plateau (Wang et al., 2009; Wu Y H et al., 2006; Chen et al., 2004).

The sample size and organic carbon content were not positively correlated; thus, the results were not identical to those in previous studies of lakes (Chen et al., 2004).

The AMS¹⁴C dating results were corrected in accordance with the tree-ring age (Table 2). Based on sediment thickness, rate, and corresponding calculations on the age profile, construct the model of the depth-age (Fig. 3), the bottom ages of the first, second, and third terraces were 4 587, 8 361, and 13 746 kaB.P., respectively.

Figure  3.  Model of depth-age.

DownLoad: Full-Size Img PowerPoint

Table  2.  ¹⁴C dating results

 | Show Table

DownLoad: CSV

DISCUSSION

Identification of Sources of the Organic Carbon Profile

Terrestrial plants are the major source of soil or-ganic carbon. δ¹³C receives less digenesis in geologi-cal history, which is a good indicator of the relative proportions of C3 and C4 plants. Therefore, it plays a significant role in ancient ecological reconstruction (Cerling et al., 1989).

Studies have shown that many factors influence the growth of plant organic carbon isotopes, such as temperature change (Haag Kerwer et al., 1992), arid-ness (Cushman and Borland, 2002; Zotz and Winter, 1993; Winter and Ziegler, 1992), salt stress (Choi et al., 2005; Lissner et al., 1999; Bohnert and Jensen, 1996), hard light (Manuel et al., 1999; Earnshaw et al., 1987), and CO₂ concentration (Philip and Shoji, 1993). Therefore, using soil organic carbon isotopes to ana-lyze paleoclimates and paleoenvironments may give a variety of results. Thus, a comprehensive analysis from other indicators should be performed.

In this article, the upstream river terrace of the Lhasa River was selected for study. It was different from the lacustrine sediments. River alluvial terraces may be directly related to the new tectonic movement. It is also more open than a lake system, in which car-bon input and output differ greatly. Therefore, correct identification of the plant photosynthetic pathways is very important for paleovegetation reconstruction (Black and Osmond, 2003).

C3 plants live in cool environments with normal light. The δ¹³C in C3 plants is between -20‰ and -35‰ (average=-26‰). C4 plants thrive under higher temperatures with more intense sunlight. The δ¹³C of C4 plants is between -7‰ and -15‰ (average=-12‰). The δ¹³C of CAM plants is between -10‰ and -22‰ (average=-16 ‰) (Chen et al., 2002).

According to the section analysis shown in Table 1, the profile δ¹³C value for the Tangjia region in Lhasa is between -18.82‰ and -22.89‰. The types of CAM plants found in the region are Agave (Agave), Opuntia (cactus genus), and Clusia (pig rubber tree genus) (Black et al., 2003).

Pollen studies showed that vegetation types such as Artemisia, Gramineae, Cyperaceae, and coniferous species were found in the Hayden Lake area (Tang et al., 2004). Pollen types such as Pinus (Pinus), Betula (Betula), Artemisia (Artemisia), Ephedra (Ephedra), Quercus (Quercus), Chenopodiaceae, Gramineae, Po-lygonum (Polygonum), and Corylus (Corylus) were found in the Nam Co area (Zhao et al., 2003).

These types of plants show that C3 plants domi-nated ancient vegetation types. Therefore, the δ¹³C of this section should be dominated by C3 plants. There were also some C3-CAM facultative plants and a small number of CAM plants.

Change Features of δ¹³C, TOC, and Rb/Sr

From Fig. 4, the δ¹³C and TOC change with fluctuation features significantly from bottom to top. There are two relationships evident in both changes in the section: δ¹³C and TOC show a significant negative correlation in the bottom of the second and third ter-races, and there is a significant positive correlation in the first terrace as well as in the top of the second and third terraces. Similar relationship also appeared in Cuo'er Lake (Chen et al., 2004).

Figure  4.  Variations of δ¹³C, TOC, and Rb/Sr concentrations in the Quaternary section of the Lhasa Tangjia region.

DownLoad: Full-Size Img PowerPoint

When TOC and δ¹³C are positively correlated, the corresponding δ¹³C would be mostly positive as TOC increases. An increase in TOC indicates an in-crease in organic matter content in soil, so δ¹³C would be mostly positive, indicating a warm and humid cli-mate that is conducive to the growth of C4 plants. Conversely, as TOC decreases, δ¹³C becomes more negative, which indicates the dominance of C3 plants that grow in relatively arid climates. However, TOC and δ¹³C may not distinguish wet and dry climates clearly.

In supergene geochemical behavior, Sr is mostly in the isomorphic phase found in minerals containing Ca, carbonates, plagioclase, and pyroxene. It usually separates after weathering, causing the paleosol sediments to be rich in Sr (Chen et al., 1998).

The weathering products of the Rb/Sr ratio in-crease significantly with an increased degree of weathering. The Rb/Sr ratio and the degree of weath-ering are positively correlated (Dasch, 1969). When the Rb/Sr ratio increases, it means that the weathering rates are higher in warm climates with abundant rain-fall, so the Rb/Sr ratio is a good indicator of how wet or dry a climate is.

The Rb/Sr ratio in the section fluctuates between 1.48 and 0.83. There is also a positive correlation be-tween the Rb/Sr ratio and 1/Sr, suggesting that the Rb/Sr ratio could indicate the speed of weathering (Jin et al., 2005).

In Quaternary fluvial sediments, the role of extrinsic-geological in the basin shows long-term weathering and erosion, transportation, and deposition. Accumulation of loose pyroclasts in the fluvial facies and the abundance of Sr indicate an increase in the rate of chemical weathering in the basin. This suggests a relatively humid climate; therefore, the Rb/Sr ratio decreased. Conversely, a decrease in Sr indicates a slower rate of chemical weathering in the basin, sug-gesting a relatively cold and dry climate; therefore, Rb/Sr increases.

Ancient Climate Analysis

Through analysis of changes in climate charac-teristics with dating data, the climate change demon-strated in the profile could be divided into seven cli-mate zones (Fig. 4).

I. Climatic zone 1 065–835 cm (about 13.7–11.7 kaB.P.)

An increase in TOC, as shown in the larger peak (0.67%) in Fig. 4, indicates an increase in soil organic matter content. δ¹³C is partially negative, which indi-cates a warm climate. The Rb/Sr ratio shows three fluctuations, but the values are below average, indi-cating a relatively humid climate. The Hayden Co (Tang et al., 2004) pollen mainly contained Artemisia (40%), grass and sedge, as well as some conifer species, such as pine, fir, and spruce. These species be-long to alpine forest steppe vegetation, where the cli-mate was generally warm. Gasse et al. (1991), who did research in Sumuxi Co, thought that there were warm and wet weather events in 12.5 kaB.P. period. This result shows similar characteristics as the Rb/Sr ratio is revealed.

Comparing the timing and features of this stage in GRIP (Dansgaard et al., 1993) and NGRIP (Andersen et al., 2004), we found that δ¹⁸O has increased fluctuantly. The warm record of the north Tibet, Os-tracod fauna such as Candona neglecta, Cyprideis torosa, and Lim nncythere dubiosa are increased ob-servably (Li, 2004), suggesting that the warm episode was standing written in water. Significantly, in Chen Co core indication, this phase was not easy to distin-guish.

II. Climatic zone 835–685 cm (11.7–10.2 kaB.P.) The δ¹³C values are all positive with minor fluc-

tuations that are all above average, indicating a sig-nificant rise of C4 plants.

TOC values increase but not significantly. The Rb/Sr ratios show no major fluctuations and are all below average value, indicating a relatively rapid weathering rate. Climate was wetter than that of zone I. Vegetation types included forest, semi-forest, and semi-grassland vegetation.

When Sumuxi Co TOC (Gasse et al., 1991) is compared with ancient Guliya ice core δ¹⁸O (Thompson et al., 1997) (Fig. 5), it appears that the weather of Sumuxi Co turned cold from 12.51 kaB.P. whereas the ancient Guliya ice core indicated periods of warm climate in 11 ka, but the period of warm climate revealed by this profile may be longer.

Figure  5.  Comparisons of the δ¹³C curve in the Lhasa Tangjia region, the TOC curve in the Sumxi Co, and the δ¹⁸O curve in the Guliya ice-core.

DownLoad: Full-Size Img PowerPoint

Figure 5 shows the revelation presence of global cold events YD (Grootes and Stuiver, 1997; An et al., 1993; Mangerud et al., 1974). Although A1 fluctuates, the degree of fluctuation is small, whereas YD, indi-cated by A2 and A3, are in very significant fluctuation. In the study of ancient climates in the Lhasa area, there is currently no proper carrier for millennial-scale climate fluctuations records like YD events. YD events revealed by this section are not very clear. The main reason may be that the sediments were mainly coarse gravel, which greatly influenced the enrichment of organic carbon or the stored fine silt particles layer; as a result, δ¹³C values are all positive. Moreover, the most characteristic of YD event was revealed variously in Tibet (Gasse et al., 1996, 1991; Gu et al., 1993; Lister et al., 1991), and its continued period and resolution were controlled obviously by climatical proxy. Comparing with the GISP2 ice core records from Greenland (Stuiver et al., 1995), we found that the occurrence of YD in the Tibetan plateau differed from that of the GISP2 record. However, this does not rule out a period of warm and cool climate character-istics in the small watershed.

III. Climate zone 650–480 cm (about 10.2–8.1 kaB.P.)

This stage had three obvious climatic fluctuations. Based on the variations in δ¹³C and TOC, it can be sub-divided into three stages (Fig. 4).

Sub-stage 1 (685–600 cm): In this stage, δ¹³C was negative and at a minimum (-22.89‰), which means that grassy C3 plants had absolute superiority. TOC and δ¹³C had a synchronous reduction, indicating that the TOC input in the soil was sharply reduced. Both evidence indicated that the climate turned cold. The age of this stage was about 10.2–9.3 kaB.P. in the Early Holocene, when the climate had not recovered from the last glacial period. Moreover, the diminished precipitation of colder event was indicated by the δ¹⁸O fluctuations of the ice core from the North Atlantic and the Peat record from the Zoigê plateau (Zhou et al., 2002; Bond et al., 1997). The impact of this cold weather event in the Lhasa area may have exceeded the impact of the YD event. The δ¹⁸O in the ancient Guliya ice core in that period also showed negative fluctuations (Thompson et al., 1997) (Fig. 5), which confirms that there was a wet environment during that period.

Sub-stage 2 (595–560 cm): δ¹³C turns positive rapidly. TOC increases and peaks at 1.02%. The Rb/Sr ratio decreases with fluctuations, indicating that the weathering velocity had slowed down. This suggests that the climate was warm and dry in the Early Holo-cene. The C4 plants thrived, as did similar vegetation types in that period based on the δ¹³C record of Qing-hai Lake (Liu X Q et al., 2003). The vegetation type was forest-based forest, but steppe vegetation was auxiliary. The age of this stage was about 9.3–8.9 kaB.P.. In the Lhasa area, the climate had truly turned warm since 9.3 kaB.P..

Sun-stage 3 (560–480 cm): δ¹³C appeared to fluctuate. TOC decreased rapidly, which indicates that the soil organic carbon content had decreased rapidly and the climate turned cold. The increase of Rb/Sr ra-tio indicates that the climate was cold and wet, and the vegetation showed deforestation and an increase in grasslands. The period of this sub-stage was about 8.9–8.10 kaB.P.. Figure 3 shows that B1 and B2 had good correspondence, and this confirms the global cold event happened at about 8.2 kaB.P. (Alley et al., 1997).

IV. Climate zone 480–280 cm (about 8.1–6 kaB.P.)

The δ¹³C values were positive in this zone. The TOC values fluctuated but were above average value. The Rb/Sr ratio was larger than average with little variation. Comprehensive analysis indicates that the climatic characteristics in this stage were warm and a little dry, and the weathering rate increased more than ever before.

The vegetation was forest-based forest grassland. According to the record in the Sumuxi Co and ancient Guliya ice cores (Thompson et al., 1997; Gasse et al., 1991), this period in the Holocene was warm with abundant rainfall (Fig. 5). Organic carbon isotopes in Co er and Zigetang Co in northern Tibet (Wu Y H et al., 2007, 2006) showed that this period was warm, but the warm event was relatively longer.

V. Climate zone 280–165 cm (6.1–4.99 kaB.P.)

The δ¹³C had three climatic fluctuation cycles. TOC decreased rapidly, which indicates that the cli-mate turned cold. The Rb/Sr ratio increased rapidly, which shows that the weathering velocity had slowed down in this period. The cold event from the Ximencuo record between 5.9 and 5.5 cal kaB.P.. The δ¹³C and TOC were decreased at same time (Mischke and Zhang, 2010). In about 4.99 kaB.P., the climate turned warm and wet (Fig. 4), and vegetation varied with the climate and showed grassland-forest-grassland replacement. In Sumxi Co, the climate in 5.0 kaB.P. was relatively warm with some cool and dry periods (Gasse et al., 1991) (Fig. 5). In Chen Co core, the warm and dry situation became much stronger be-tween 5.5 and 5.0 kaB.P., and total pollen density de-creased, but Myriophyllum reached its highest value (Zhu et al., 2009).

In ancient ice cores, there is some indication of the climate turning cold (Fig. 5), which is very differ-ent from what this zone revealed.

VI. Climate zone 165–30 cm (about 4.99–1.9 kaB.P.)

This zone can be sub-divided into two climate zones.

Sub-stage 1 (165–85 cm): This sub-zone displayed a big-small-big trend. TOC decreased at a rela-tively fast rate, which shows that the climate had turned cold. The Rb/Sr ratio was at its lowest value (0.83), indicating that the weathering velocity in this period was at its highest. This sub-zone exhibited warm, cool, and wet climates. Vegetation was mainly grasslands; forested areas dropped sharply.

The period of this sub-zone was about 4.99–3.63 kaB.P., including D1, D2, and D3, as shown in Fig. 5. There were regional cooling events in 3.4 kaB.P., and it seems that our terrace suggests cold event as Xi-mencuo record (Mischke and Zhang, 2010). The phy-tolith combination in Lhasa Sangda was mainly the cold-type phytolith (Chen et al., 2008), and the climate was warm and cool.

Sub-stage 2 (85–30 cm): Both fluctuation of δ¹³C and TOC increased, which indicates that the climate had turned warm. The Rb/Sr ratio increased gradually and then decreased, which shows that the climate turned from wet to dry. In this period, grasslands dominated and forests showed signs of expansion.

The period was about 3.63–1.9 kaB.P.; earlier, it was warm and dry during the early stage, but ap-proaching 1.9 kaB.P. the climate gradually turned cold. The dry events in this period occurred in about 3.4 kaB.P., which was confirmed in Sumxi Co (Gasse et al., 1991) and Bangong Co (Fig. 1) (Gasse et al., 1996). Significantly, that the climate turned cold was demonstrated in the Lhasa Sangda profile phytolith combination III (Chen et al., 2008).

VII. Climate zone 30–0 cm

This zone is the accumulation of a modern flood plain, which is mainly composed of fine silts. The δ¹³C decreases rapidly. TOC declines and fluctuates from 0.13% to 0.20%, indicating an overall cold cli-mate. The Rb/Sr ratio appears to have a slight upward trend, which shows signs that the climate was turning from wet to dry. The vegetation was grassland.

Through comparison with the Lhasa Sangda pro-file (Chen et al., 2008), we concluded that the ther-moluminescent age of the floodplain at the bottom was about 1.9±20 kaB.P.. In about 2 kaB.P., China experienced widespread cooling (Wang et al., 2006), and there appeared to be many cold years from 1.28 to 1.82 kaB.P. (Xu, 1998). Over the past 100 years, global warming was probably an emerging trend against nature mostly caused by human activities in the natural cooling (Zhang et al., 2005). Although we try to control the use of fossil fuels and reduce greenhouse gas emissions, the millennial-scale climate shows that humankind will enter a new ice age, and the climate will be cold.

CONCLUSIONS

Through the analysis of δ¹³C, TOC, and the Rb/Sr ratio in the Lhasa Tangjia profile, we learned more about the variable features of the ancient climate since the last glacial period. We made the following conclusions.

There was an important reason for choosing δ¹³C, TOC, and the Rb/Sr ratio as indicators in our research on ancient climates. Ancient climatic evolution in the Lhasa Tangjia area can be divided into seven stages: 13.7–11.7 kaB.P., the climate was totally warm and humid; 11.7–10.2 kaB.P., the climate was cool and wet; 10.2–8.1 kaB.P., the climate was cold-warm-cold, and during 10.2–9.3 kaB.P., the climate was cold and dry; 9.3–8.9 kaB.P., the climate was warm and wet (this period may have been the warmest in the Holocene); 8.9–8.1 kaB.P., the climate was warm-cold-warm and turned cold again during global cooling in 8.2 kaB.P.; 8.1–6.1 kaB.P., the climate was warm and slightly dry; 6.1–4.99 kaB.P., the climate experienced three cycles of warm-cool fluctuations, and TOC and the Rb/Sr ra-tio showed that the climate in that period was overall warm with some cool and dry periods; and 4.99–1.9 kaB.P., the climate was warm and cool, and the early and late periods were wet. Since 1.9 kaB.P., the cli-mate has turned cold and dry.

This article shows climatic fluctuations on a mil-lennial-scale turbulence and indicates the cold event in 9.4, 8.2, 5.4, and 4.2 kaB.P.. In 10.2–8.1 and 4.99–1.9 kaB.P., climate resolutions were more obvious, showing that the climate had millennium variation law, and the study findings were similar with the Wang and Xie (2002).

It is important that we chose a river terrace as the indicator of the ancient climate in the Lhasa region. A major goal in ancient climate research is to better un-derstand climatic oscillation periods to predict climate variation in the future. The Lhasa River valley played an important role in Tibetan history because of its special topography geographic position. A unique cli-mate formed there in the environment of a small wa-tershed. Identification of the trends in regional climate change based on this river terrace was important in the study of Tibetan history and the prediction of abrupt climate change.