New evidence on deglacial climatic variability from an alpine lacustrine record in northwestern Yunnan Province, southwestern China

Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

New evidence on deglacial climatic variability from an alpine lacustrine record in northwestern Yunnan Province, southwestern China Xiayun Xiao a,⁎, Simon G. Haberle b, Xiangdong Yang a, Ji Shen a,⁎⁎, Yong Han c, Sumin Wang a a b c

State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, 73 East Beijing Road, Nanjing 210008, China Department of Archaeology & Natural History, College of Asia and the Pacific, Australian National University, Canberra, Australian Capital Territory 0200, Australia School of Atmospheric Science, Nanjing University, Nanjing 210093, China

a r t i c l e

i n f o

Article history: Received 17 December 2013 Received in revised form 19 March 2014 Accepted 9 April 2014 Available online 26 April 2014 Keywords: Pollen analysis Abrupt climate events Southwest monsoon Last deglaciation Alpine lake Southwestern China

a b s t r a c t New deglacial pollen and conifer stoma records from Tiancai Lake, northwestern Yunnan Province, southwestern China, an alpine lake in the southwest monsoon region, are presented in this study. Based on these records, the lithology of core TCYL1, and PCA analysis of pollen data between ~21 and 11.5 ka BP (calibrated 14C years), the deglacial vegetation and climate changes are discussed in detail. The results show that Tiancai Lake was above the upper limit of Picea/Abies forest (a treeline in the study area) between ~21 and 11.5 ka BP, and the climate was colder and drier than that of the Holocene. During this period eight significant vegetation changes are recorded that are considered to be responses to changing temperatures and variations in the southwest monsoon in southwestern China. The Heinrich Event 1 (H1), the Bølling/Allerød warm period (BA) and the Younger Dryas cold event (YD) are all clearly detected in this record. In addition, this study finds that the initial late glacial warming in northwestern Yunnan Province was at ~18.7 ka BP, which is coincident with the climate records in monsoonal Central Asia, the Indian Ocean, the tropical and subtropical Pacific Ocean, and Antarctica, and is a response to solar insolation changes. A noted temperature increase between 15.8 and 14.4 ka BP occurred at the end of the H1 and before the BA, which indicates a strong pre-Bølling warming. Based on the study, we consider that the hypothesis about a slowdown of the ocean's thermohaline circulation is sufficient to explain these late glacial abrupt events. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The last deglaciation in the Northern Hemisphere is not a smooth transition from one climate state to another and is punctuated by some abrupt climate events. The most prominent deglacial events in the Northern Hemisphere are the Heinrich Event 1 (H1, ~ 17.5 to 16 ka) and the associated Oldest Dryas cold period (~ 18 to 14.7 ka), the Bølling/Allerød warm period (BA, ~14.7 to 12.9 ka) and the Younger Dryas cold event (YD, ~ 12.9 to 11.7 ka) (Alley and Clark, 1999). The global extent of these events is not yet clearly established (Clark et al., 2002) as the geographic distribution of high resolution last deglaciation records is largely focused on high-latitude regions such as Greenland, North Atlantic Ocean and Antarctica (Johnsen et al., 1972; Duplessy et al., 1992; Blunier et al., 1998; Stuiver and Grootes, 2000; Blunier and Brook, 2001; Jouzel et al., 2001; Waelbroeck et al., 2001; Knorr and Lohmann, 2003; EPICA Project Members, 2004; McManus et al., 2004; NGRIP members, 2004; Brook et al., 2005; EPICA Community Members, 2006; Gherardi et al, 2009; Barker et al., 2010; Thornalley et al., 2010; Pedro et al., 2011; Ritz et al., 2013). ⁎ Corresponding author. Tel.: +86 25 86882146; fax: +86 25 86882189. ⁎⁎ Corresponding author. Tel.: +86 25 86882005; fax: +86 25 86882189. E-mail addresses: [email protected] (X. Xiao), [email protected] (J. Shen).

http://dx.doi.org/10.1016/j.palaeo.2014.04.008 0031-0182/© 2014 Elsevier B.V. All rights reserved.

Southwestern China, especially northwestern Yunnan Province, is a region affected strongly by the southwest summer monsoon (Indian summer monsoon). At the same time, it is adjacent to the Tibetan Plateau, which makes southwestern China a key region for reconstruction of paleoclimate changes and past southwest monsoon dynamics. However, paleoenvironmental records spanning the whole last deglaciation are relatively few in southwestern China (Wu et al., 1991; Tang, 1992; Liu et al., 1995; Jiang et al., 1998; Hodell et al., 1999; Qin et al., 2001, 2004; Yin et al., 2002; Zhang et al., 2002, 2004a, 2004b; Zhang and Mischke, 2009; Yang et al., 2010; Cook et al., 2011). Because of low temporal sampling resolution, age uncertainties and/or regional sensitivity, there are large differences between the results of these published studies. Some studies suggest that the climate was cold and wet during the whole last deglaciation (Wu et al., 1991; Tang, 1992). A majority of studies detect the Last Glacial Maximum (LGM) and YD (Liu et al., 1995; Jiang et al., 1998; Hodell et al., 1999; Yin et al., 2002). Only a few studies detect H1 (Qin et al., 2004; Zhang et al., 2004a) and the BA (Kramer et al., 2010; Yang et al., 2010), and the timing and paleoclimatic interpretations vary across these records. Consequently, it is necessary to develop more precise, high-resolution studies in southwestern China, which can help clarify regional climate dynamics and detect possible teleconnections between the southwest monsoon and the global climate system.

10

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

High-elevation alpine lake sediments have been shown to be sensitive to climate change, recording changes in subalpine treeline vegetation since the late glacial and early Holocene (Pellatt and Mathewes, 1997; Barnett et al., 2001; Jiménez-Moreno et al., 2008). In many regions, alpine environments present the additional advantage for paleoclimate reconstruction by being less disturbed by humans than low elevation sites. Tiancai Lake is an alpine lake in the southwest monsoon region, little disturbed by human activities, which makes the lake more sensitive to natural processes. At the same time, sedimentation in Tiancai Lake is continuous and provides the possibility for a high-resolution study. Thus, Tiancai Lake is an ideal area to study paleoclimate changes in the southwest monsoon region (Xiao et al., 2014). Here we present a new centennial scale paleolimnological record from Tiancai Lake, which provides new evidence on the last deglacial regional climate variability and southwest monsoon dynamics. This study is an extension of the record presented in Xiao et al. (2014) that focused on vegetation and climate changes since 12.2 ka BP from Tiancai Lake. 2. Regional setting Tiancai Lake is an alpine ice-scoured lake in northwestern Yunnan Province, southwestern China. The lake is located on the southeast edge of the Qinghai–Tibet Plateau, a transition zone from the Hengduan Mountains to the Yunnan–Guizhou Plateau. The elevation of Tiancai Lake is 3898 m a.s.l. with a maximum water depth of 7 m, surrounded on three sides by mountains rising to 4100–4200 m a.s.l. with a stream outlet in the northeast of the lake (Xiao et al., 2014) (Fig. 1). The study region is characterized by a highland cold temperate humid monsoon climate. It is mainly determined by warm–humid airflow from the Indian Ocean and Bengal Bay in summer and by the south-branch westerly winds in winter and is also affected by the local climate of the Qinghai–Tibet Plateau. Due to the monsoon, the lake region is characterized by distinct wet (summer) and dry (winter) seasons. There is no weather station around the lake at present. The mean annual temperature and mean annual precipitation for the lake region have been obtained by interpolation using a gradient-plusinverse distance squared (GIDS) method, which combines multiple linear regression and distance weighting (Nalder and Wein, 1998) according to the climatic data of the 46 weather stations around the study area. The result shows that the mean annual temperature around

the lake is 2.5 °C and the average annual precipitation is about 910 mm (Xiao et al., 2011). Because of marked altitudinal gradients and complex topography in northwestern Yunnan Province, altitudinally controlled vegetation belts in the region are distinct (Xiao et al., 2008, 2010). Tiancai Lake is located about 200 m below the treeline. Primary forest around the lake appears to be undisturbed in the historic past, with cold-temperate old-growth conifer forest composed of Abies and Picea covering the catchment (Xiao et al., 2014). 3. Materials and methods 3.1. Coring and sampling Sediment of core TCK1 collected in October 2008 from Tiancai Lake using a UWITEC piston corer was found to be continuous and was analyzed at high-resolution (Xiao et al., 2014), though the core did not reach the basal sediments of the lake. In November 2010, another core (26°38′3.8″ N, 99°43′0.5″ E) was collected that reached the bottom of the sediments near core TCK1 using the same piston corer (Fig. 1). The second core (TCYL1) is an 865-cm-long sediment core, and was sectioned at 1 cm intervals. Samples were stored at 4 °C until analyzed. Core correlation between TCK1 and TCYL1 was carried out using surface scanning magnetic susceptibility (Anderson, 1986). There is a very good correlation between cores TCK1 and TCYL1 (Fig. 2), and by comparing peaks in the magnetic susceptibility records, it is shown that core TCYL1 is older than core TCK1 as the lowest peak at 820 cm in core TCK1 corresponds to the peak at 678 cm in core TCYL1. Thus, the date at 678 cm in core TCYL1 equals the date at 820 cm in core TCK1, which is 10,527 cal. yr BP (Table 1). Based on the date obtained by core correlation and 12 AMS 14C dates in the lower part of core TCYL1, the time series of the lower part of core TCYL1 is established (see Section 4.1 for detail). According to the ages of cores TCK1 and TCYL1, the stratum at 707 cm of core TCYL1 can be joined to the bottom of core TCK1. Consequently, the integrated core in Tiancai Lake consists of core TCK1 and the strata between 707 and 865 cm of core TCYL1. Core TCK1 has been discussed in detail in Xiao et al. (2014). In this study, we focus on the lower part of core TCYL1, namely, the sediment between 693 and 865 cm. The sediment is almost continuous lacustrine sediment except for 24-cm-long ground moraine with poor psephicity in the lowest part of core TCYL1 (841 to 865 cm). The lithology of the lower part of core TCYL1 consists of black fine detritus mud between

Fig. 1. (a) Location map of Tiancai Lake, northwestern Yunnan Province, China. Other sites mentioned in Section 5.2 are also shown. (b) Topography of Tiancai Lake surrounded on three sides by mountains.

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

11

Fig. 2. Correlation of cores TCK1 and TCYL1 using surface scanning magnetic susceptibility.

693 and 750 cm; green-gray mud between 750 and 774 cm; light greengray mud between 774 and 824 cm; and green-gray mud with fine sand between 824 and 841 cm (Fig. 4).

TCK1 and TCYL1. All dates were calibrated to calendar years before present (0 BP = 1950 AD) with the program CALIB 6.0 utilizing the IntCal09 calibration data set (Reimer et al., 2009; Table 1).

3.2. Dating methods

3.3. Laboratory analysis

Bulk organic carbon dating of 12 samples from Tiancai Lake (core TCYL1) was undertaken at the Poznań Radiocarbon Laboratory of the A. Mickiewicz University. In addition, one date at 678 cm (the depth of core TCYL1) is obtained by using the core correlation between cores

Pollen was analyzed at 1 or 2 cm intervals except for 2 samples in the lowest 24 cm ground moraine of core TCYL1. Pollen samples were processed using heavy liquid separation (Moore and Webb, 1978; Nakagawa et al., 1998). In brief, the preparation procedures involve treatment with 10% HCl to remove carbonates, 10% KOH to remove humic substances, heavy liquid floatation (a ZnCl2 solution was used having a specific gravity of between 2.0 and 2.2), acetolysis to remove cellulose, followed by mounting the sample in a glycerin jelly. Tablets containing a known quantity of Lycopodium spores were added to each sample to allow determination of pollen concentration (grains/cm3) or pollen accumulation rate (grains cm−2 yr−1). Pollen samples were examined using a Zeiss microscope at magnifications of 400 × and 630 ×. An average of 556 terrestrial pollen grains per sample in the sediment between 693 and 865 cm of core TCYL1 was counted. Algae and conifer stomata were found together with the pollen grains in the pollen residue and were also counted. Terrestrial pollen percentages were based on a sum of terrestrial pollen taxa, while pollen percentages of aquatic herbs, ferns, algae, and conifer stomata were calculated using a sum of all pollen and spore taxa. Pollen accumulation rate was estimated for each level by multiplying the pollen concentration by the sedimentation rate, determined using the age–depth model. The data are expressed both as percentage and pollen accumulation rate, and graphed using TiliaGraph (Grimm, 1993).

Table 1 AMS radiocarbon dates from Tiancai Lake (core TCYL1). Lab code

Depth of core TCYL1 (cm)

14 C date (yr BP)

Error (±yr)

Calibrated age (cal. yr BP, 2σ)

Mid-point (cal. yr BP)

One datea Poz-43227 Poz-43228 Poz-43230 Poz-40671 Poz-40672 Poz-40673 Poz-40675 Poz-43231 Poz-40676 Poz-40677 Poz-40678 Poz-40679

678 707 721 735 749 760 768 779 781 794 807 824 839

9160 10,590 10,900 11,410 12,850 12,870 14,140 12,050 13,050 14,980 16,450 16,230 16,380

50 60 60 50 60 60 70 70 70 70 100 130 180

10,405–10,649 12,406–12,655 12,616–12,938 13,149–13,397 14,964–15,891 14,971–15,937 16,916–17,540 13,741–14,084 15,179–16,399 17,978–18,541 19,404–19,910 18,913–19,584 19,215–20,058

10,527 12,531 12,777 13,273 15,428 15,454 17,228 13,913 15,789 18,260 19,657 19,249 19,637

a One date is obtained from correlation of cores TCK1 and TCYL1 using surface scanning magnetic susceptibility.

12

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

3.4. Numerical methods Stratigraphic zones were mainly aided by CONISS of the pollen data in the TILIA program (Grimm, 1987). All pollen taxa with N1% in at least two samples were included in the CONISS analysis for zonation of the integrated core in Tiancai Lake. The CONISS analysis used an algorithm of stratigraphically constrained chord-distance clustering and squareroot transformation of the pollen percentage data. The pollen data presented here are complex and multivariate. Ordination provides an effective way to identify and visualize the main directions of variance. It seeks to represent trends in a smaller number of dimensions, preferably two or three, so that the main directions of variance are displayed (Birks and Gordon, 1985; Maddy and Brew, 1995). In this study, we first used the percentages of 46 terrestrial pollen types with values greater than 1% in at least two samples to analyze numerically all pollen samples in the integrated core of Tiancai Lake using the CANOCO program version 4.5 (ter Braak and Šmilauer, 2002). Pinus pollen is considered to be transported from the lowest vegetation belt in the region or other distant regions, and its percentage values are highest (mean of 28.3%) in the pollen assemblage after 11.5 ka BP with no obvious phase change, so its weight is set to be 0.1 in the numerical analysis. In addition, because the lithology of core TCYL below 841 cm is ground moraine, and pollen accumulation rates of samples at 844 and 865 cm are very low, their pollen percentage assemblages may be misleading. At the same time, the sample at 719 cm is an abnormality with the highest Abies/Picea pollen percentage (42.3%). Thus, these three samples are deleted in the numerical analysis. To check the underlying linearity of the data, a detrended correspondence analysis (DCA) was carried out initially. The DCA results indicated a maximum gradient length of 2.037, which is less than 2.5 standard deviations, suggesting that most of the underlying responses are linear or at least monotonic to the underlying latent variables (ter Braak and Šmilauer, 2002; Zhao and Herzschuh, 2009). Therefore, principal component analysis (PCA) was used to analyze the pollen assemblages in relationship to environmental factors by using inter-species correlations and no transformation of pollen percentages. 4. Results 4.1. Chronology An age-model was constructed using the 12 AMS 14C dates between 707 and 839 cm of core TCYL1 and the date obtained by core correlation. Through the dated depths, age–depth models were drawn consisting of 39 1-cm thick sections of piece-wise linear accumulation, with the accumulation rate of each section depending somewhat on the previous section (an autoregressive process with a degree of “memory”) (Blaauw and Christen, 2011). Using an iterative Markov Chain Monte Carlo process, several thousands of independent likely age–depth models were constructed, each of them assigning calendar ages to all depths (dated or non-dated). From these models, calendar age distributions were constructed for all depths, and depicted as gray-scales where more likely calendar ages are depicted by darker gray levels (Fig. 3). According to the age-model, the study period of this study is presumed to be between ~ 21 and 11.5 ka BP, and the average temporal sampling resolution is ~ 100 years for the fossil pollen record. Ages based on this age model are applied to the fossil record. 4.2. Pollen record A total of 139 pollen and spore types were identified from 84 pollen samples between ~21 and 11.5 ka BP in Tiancai Lake. This pollen record differs from the pollen assemblage post-11.5 ka BP reported in Xiao et al. (2014), which is characterized by very high pollen percentages of trees dominated by Pinus, Tsuga, and evergreen oaks, very low pollen percentages of Betula and herbs such as Artemisia, and high pollen

accumulation rates (Figs. 4, 5). The pollen assemblage in this study is dominated by tree and herb pollen, averaging 52.2% and 36.1%, respectively (Fig. 4). Pollen percentages of shrubs are relatively low, averaging 11.7%. The dominant tree taxa are Pinus, Picea and Abies, Betula, deciduous oaks, and evergreen oaks. Other tree taxa such as Alnus, Ulmus, and Castanopsis/Lithocarpus (Cast./Lith.) have some percentages, while Tsuga, Carya, and Juglans only occur with minor quantities in the pollen assemblage. Shrub pollen is dominated by Moraceae, Rosaceae, and Theaceae. The main components of herb pollen are Artemisia, Scrophulariaceae, Lamiaceae, Gesneriaceae, Poaceae, and Ranunculaceae. Aquatic herbs only occur sporadically (not presented in the pollen diagram). Ferns are mainly represented by monolete spores. In addition, a few algae dominated by Pediastrum are found. There are no conifer stomata such as Abies and Sabina stomata occurring in the pollen assemblage between ~21 and 11.5 ka BP, but they are continuously present in the pollen assemblage after 11.5 ka BP (Xiao et al., 2014). Pollen accumulation rates are obviously lower than that after 11.5 ka BP. Under the background of these general pollen assemblage characteristics, the pollen assemblage between ~ 21 and 11.5 ka BP can be divided into eight zones based on the CONISS classification of the pollen data and visual inspection (Figs. 4, 5). The characteristics of these pollen zones are described from bottom to top as follows.

4.2.1. Zone TCYL-1 (865–842 cm, 21–19.9 ka BP) There are only two samples in the zone. Its lithology is ground moraine with poor psephicity, resulting in the lowest pollen accumulation rates. The total pollen accumulation rate (TPAR) averages only 48 grains cm− 2 yr− 1. Pollen percentages of trees are higher than herbs, which may be deceptive because of the lithology and very low pollen accumulation rate in the zone. Thus, the pollen percentage assemblage in the zone is not discussed further.

4.2.2. Zone TCYL-2 (842–801 cm, 19.9–18.7 ka BP) The lithology of the zone changes into green-gray mud with fine sand between 841 and 824 cm and light green-gray mud between 824 and 801 cm. The pollen accumulation rate increases substantially, and the TPAR averages 982 grains cm−2 yr−1. Pollen percentages of herbs are highest (mean of 56.8%) for the entire core and higher than that of trees (32.4%), dominated by the highest pollen percentages of Artemisia (16.9%), Lamiaceae (9.9%), and Gesneriaceae (8.2%) for the entire core. Tree pollen is dominated by Picea/Abies (11.2%), and pollen percentages of other tree taxa are relatively low.

4.2.3. Zone TCYL-3 (801–789 cm, 18.7–17.7 ka BP) The lithology of the zone is light green-gray mud. Compared with the former zone, tree pollen percentages (43.3%) are higher and herb pollen percentages (45.1%) are lower in the zone. The increase of tree pollen percentages is mainly caused by Picea/Abies (18.3%). The decrease of herb pollen percentages is mainly represented by Artemisia (10.9%), Lamiaceae (6.0%), and Gesneriaceae (3.7%), although they still have relatively high percentages. The pollen accumulation rate increases notably, and the TPAR averages 6547 grains cm−2 yr−1.

4.2.4. Zone TCYL-4 (789–785.5 cm, 17.7–17 ka BP) There is a clear peak in tree pollen percentages (58.9%) in the zone, represented by Pinus, Tsuga, and deciduous oaks with their pollen percentages averaging 17.6%, 1.9%, and 8.8% respectively. Additionally, Theaceae, Apiaceae, Solanaceae, and Liliaceae show small peaks in their pollen percentages, averaging 3.0%, 1.2%, 1.4%, and 2.5%, respectively, while Artemisia, Thalictrum, Scrophulariaceae, and Lamiaceae all show obvious troughs, averaging 3.9%, 0%, 1.4%, and 1.9%, respectively. The pollen accumulation rate is relatively low, and the TPAR averages 757 grains cm−2 yr−1.

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

13

Fig. 3. Age–depth model produced by BACON software (Blaauw and Christen, 2011) based on 12 AMS 14C dates and one date obtained from correlation of cores TCK1 and TCYL1 using surface scanning magnetic susceptibility. Age-models are based on 39 1-cm thick sections of linear accumulation, with a degree of memory between each section. Gray-scales indicate all likely age–depth models, and dotted lines indicate the 95% confidence ranges. Shapes show calendar age.

4.2.5. Zone TCYL-5 (785.5–766 cm, 17–15.8 ka BP) The zone is characterized by lower tree pollen percentages (45.8%), and higher herb pollen percentages (42.0%) dominated by Artemisia (12.0%), Lamiaceae (5.9%) and Gesneriaceae (5.8%), than Zone TCYL-4. The pollen accumulation rate increases notably again, and the TPAR averages 2503 grains cm−2 yr−1. 4.2.6. Zone TCYL-6 (766–748 cm, 15.8–14.4 ka BP) The zone is dominated by green-gray mud. In the zone, pollen percentages of trees increase gradually (48.7%), while herb pollen decreases gradually (38.5%). Pollen percentages of broadleaved trees are highest for the entire core, with highest Betula pollen percentages (13.3%), markedly increasing Ulmus pollen percentages (3.5%), and some percentages of Carya pollen (0.7%). Artemisia pollen (4.9%) declines markedly, while pollen percentages of Scrophulariaceae (6.1%) and Cyperaceae (2.7%) are slightly higher. The pollen accumulation rate is relatively high, and the TPAR averages 10,109 grains cm−2 yr−1. 4.2.7. Zone TCYL-7 (748–718 cm, 14.4–12.9 ka BP) The zone is characterized by high pollen percentages of trees (67.4%) and low pollen percentages of herbs (21.2%). Picea/Abies (21.3%) increases sharply to its highest values for the entire core, and Pinus (19.1%) is markedly higher. Pollen percentages of Betula are still relatively

high, but decline slowly (12.6%), and Ulmus pollen declines gradually from its highest value (mean of 3.8%, ranging from 2.0 to 7.9%). Pollen percentages of Artemisia are very low (1.4%). Apiaceae and Cyperaceae pollen has relatively high percentages. The pollen accumulation rate increases further, and the TPAR averages 15,548 grains cm−2 yr−1. 4.2.8. Zone TCYL-8 (718–693 cm, 12.9–11.5 ka BP) Tree pollen percentages (67.5%) and herb pollen percentages (20.7%) in the zone are almost the same as those of Zone TCYL-7. However, Pinus pollen (27.2%) is higher, while Picea/Abies (14.5%) pollen and Betula (10.9%) pollen decrease gradually through the zone. Pollen percentages of Artemisia (1.6%) are slightly higher than those in Zone TCYL-7. The pollen accumulation rate increases greatly, and the TPAR reaches almost the level achieved in the Holocene, averaging 53,513 grains cm−2 yr−1. 4.3. Ordination of pollen assemblages The PCA results of 46 terrestrial pollen types from 320 samples in the integrated core of Tiancai Lake reveal that the first two axes capture 74.5% of the variance in the pollen data (axis 1: 59.9%, axis 2: 14.6%) (Fig. 6a). Almost all herbs, Picea/Abies and some broadleaved trees growing in low altitude regions such as Betula, Ulmus, Carya, and

14 X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21 Fig. 4. Percentage diagram of selected pollen taxa and conifer stomata from Tiancai Lake. Gray parts in Tsuga and Artemisia percentage curves mean 5× exaggerations. Gray shading section has been discussed in the former article (Xiao et al., 2014).

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

Fig. 5. Pollen accumulation rate diagram of selected taxa from Tiancai Lake. Open curves are 5× exaggerations. Gray section has been discussed in the former article (Xiao et al., 2014).

15

16

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

Fig. 6. (a) PCA ordination with the pollen percentage data of all samples (except for the three excluded samples) in the integrated core from Tiancai Lake. (b) PCA ordination of all pollen species located on the positive side of the first axis in panel (a) and samples before 11.5 ka BP (except for the three excluded samples).

deciduous oaks are located on the positive side of PCA axis 1, while Tsuga, Juglans, evergreen oaks, Pinus, and Ericaceae are found on the negative side. Thus, the first axis may reflect the degree of openness in local vegetation communities, namely, vegetation communities change gradually from very open to closed from the positive side to the negative side of the first axis (Xiao et al., 2014). Pollen taxa arranged along PCA axis 2 are mainly located on the right side of the first axis, though their significance is not clear in the ordination. The PCA ordination of pollen samples shows that samples from Zones TCYL-2 to TCYL-8 can be basically divided and almost all situated on the right side of the dash line separating PCA axis 1, which means that vegetation communities before 11.5 ka BP were all very open, compared with those during the Holocene. In order to discuss further the meaning of last deglacial variation, all pollen species located on the right side of the first axis in Fig. 6(a) and samples from Zones TCYL-2 to TCYL-8 were selected for another PCA (Fig. 6b). The PCA results of 37 pollen types from 81 samples before 11.5 ka BP reveal that the first two axes capture 78.7% of the variance in the pollen data (axis 1: 61.6%, axis 2: 17.1%) (Fig. 6b). In the ordination, most herbs, especially some drought-tolerant herbs, such as Artemisia, Chenopodiaceae, Lamiaceae, Gesneriaceae, Scrophulariaceae, Poaceae (b 40 μm), and Ranunculaceae are located on the positive side of the first axis, while all tree pollen taxa such as Picea/Abies, Betula, Ulmus, deciduous oaks, and Carya are found on the negative side of PCA axis 1; at the same time, some herbs of relatively wet environments such as Cyperaceae, Apiaceae, and Poaceae (N 40 μm) are also negatively correlated with PCA axis 1. Thus, the first axis may reflect climatic changes from cold and dry conditions to relatively warm and humid conditions. The arrangement of pollen taxa along PCA axis 2 is difficult to explain. The PCA results of pollen samples show that Zones TCYL-2 to TCYL-8 are basically divided, and Zones TCYL-2, TCYL-5, TCYL-3, TCYL-6, TCYL-4, TCYL-8, and TCYL-7 are sequentially arranged from right to left along PCA axis 1, indicating that the period 19.9– 18.7 ka BP (TCYL-2) was the coldest and driest period between 19.9 and 11.5 ka BP, and temperature and humidity increased gradually from the period 18.7–17.7 ka BP (TCYL-3) to the period 17.7–17 ka BP (TCYL-4), then declined again during the period 17–15.8 ka BP (TCYL5). There were gradual increases of temperature and humidity from the period 15.8–14.4 ka BP (TCYL-6) to the period 14.4–12.9 ka BP (TCYL-7), then there was a slight decrease of temperature and humidity

during the period 12.9–11.5 ka BP (TCYL-8). These climate changes can be distinctly shown by the PCA axis 1 sample score of 37 pollen types from 81 samples before 11.5 ka BP (Fig. 7a). 5. Discussion 5.1. Last deglacial vegetation and climate changes inferred from the Tiancai Lake pollen record Tiancai Lake, at 3898 m a.s.l., is today located approximately 200 m below treeline. Thus, the altitudinal position of the lake makes it an important site for understanding vegetation changes and treeline movement. The conifer stomata record of the integrated core in Tiancai Lake (Figs. 4, 5) shows that Abies and Sabina stomata, especially Abies stomata, occurred continuously in the conifer stomata record after 11.5 ka BP (Xiao et al., 2014), while these conifer stomata didn't appear in the record before 11.5 ka BP, indicating that Tiancai Lake was continuously above the upper limit of Picea/Abies forest (treeline in the study area) between ~ 21 and 11.5 ka BP, while, in the Holocene (since 11.5 ka BP), Tiancai Lake was below the upper limit of Picea/Abies forest. The different pollen assemblage characteristics suggest that climate conditions before 11.5 ka BP were significantly different to those after 11.5 ka BP, with the climate between ~21 and 11.5 ka BP being markedly colder and drier than the climate after 11.5 ka BP. This suggests that the southwest monsoon between ~ 21 and 11.5 ka BP was markedly weaker than that after 11.5 ka BP. Under the background of the relatively cold and dry climate conditions during the period 21–11.5 ka BP, the pollen record in the study still reflects eight obvious vegetation changes and treeline fluctuations in response to environmental conditions around Tiancai Lake. Modern pollen rain can help to better understand fossil pollen records. The results of modern pollen rain studies in the region have been published (Xiao et al., 2009, 2011). The pollen results from moss posters and surface lake sediments show that high percentages of Pinus pollen are found at almost all elevations regardless of whether the taxon is present in the surrounding vegetation, and Picea, Abies, Larix and Tsuga pollen are not (or rarely) transported downhill. Modern pollen rain from moss polsters indicates that pollen assemblages in different vertical vegetation zones in northwestern Yunnan Province contain different representative components. Abies, Larix, Picea, and

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

17

Fig. 7. Regional and global correlations and forcings during the last deglaciation. (a) PCA axis 1 sample score of pollen data from 81 samples before 11.5 ka BP in Tiancai Lake (this study). (b) Stalagmite M1–5 δ 18O from Socotra Island, Yemen in the northwest Indian Ocean (Shakun et al., 2007). (c) The stable δ 18O record of the planktic foraminifera G. rubber for the NOIP905 core, western Arabian Sea (Huguet et al., 2006). (d) The mean effective moisture from the Asian monsoon margin, inferred from paleoclimate records (Herzschuh, 2006). (e) Principal component (PC) 1 for temperature in the Indian Ocean (black line); principal component (PC) 1 for temperature in global land (red line); Principal component (PC) 1 for precipitation in global land (green line) (Clark et al., 2012). (f) Dongge Cave δ 18O (blue line) (Dykoski et al., 2005) and Hulu Cave δ 18O (black line) (Wang et al., 2001a), China. (g) The oxygen isotope (δ 18O) records from Greenland Ice Sheet Project Two (GISP2) (Grootes and Stuiver, 1999). (h) The combined radiative forcing (black line) from CO2, CH4, and N2O relative to preindustrial levels (Clark et al., 2012); Average summer insolation for 25° N (red line) (Berger and Loutre, 1991). Black horizontal lines indicate boundaries of climatic phases for the paleoclimate record in Tiancai Lake. Red horizontal lines indicate boundaries of climatic phases for other climatic records. Gray bars indicate the widespread accepted beginning and ending times of climate events.

Tsuga pollen are indicative elements of pollen assemblages in the Abies, Larix, Picea, and Tsuga forests, respectively. Their peak values are about 13.6%, 4.7%, 39.4% and 11.5%, respectively. High percentages of Castanopsis/Lithocarpus pollen are characteristic of mid-montane humid evergreen broadleaved forest and high percentages of evergreen oak pollen are characteristic of conifer and broadleaved mixed forest (Xiao et al., 2009, 2011). Modern pollen rain from surface lake sediments shows that large amounts of arboreal pollen, such as Pinus, Picea and Abies, Betula, Juglans, deciduous oaks, evergreen oaks, and Rosaceae, from lower elevations are introduced into sub-alpine and alpine lakes by upslope winds. Fortunately, the pollen assemblages of surface lake sediments within different vegetation types can still be distinguished by their pollen spectra and indicator species (Xiao et al., 2011). The core lithology of ground moraine and very low pollen accumulation rate between 21 and 19.9 ka BP indicate that the lake basin was still covered by glaciers, and the climate was very cold. However, the pollen percentage assemblage during this period may be deceptive because of its lithology and very low pollen accumulation rate. During the early part of 19.9–18.7 ka BP, the core lithology changed to green-gray mud with fine sand, and these sediments would have been deposited in the course of the retreat of glacial ice from the catchment. The lithology then changed into light green-gray mud midway through the period, indicating that the glacier had retreated from the catchment, and that the lake gradually stabilized. These lacustrine deposits and the obvious increase of pollen accumulation rate imply that the pollen assemblage since 19.9 ka BP is reflecting vegetation and climate changes. During the period 19.9–18.7 ka BP, pollen percentages of herbs are higher than those of trees, and dominated by highest

pollen percentages of Artemisia, Lamiaceae, and Gesneriaceae for the entire core. This assemblage indicates that vegetation around Tiancai Lake was alpine meadow dominated by Artemisia, Lamiaceae, and Gesneriaceae, and that the treeline and forests were far from Tiancai Lake. It can be inferred that the climate between 19.9 and 18.7 ka BP had an upturn in terms of warming and humidity relative to the period 21.0–19.9 ka BP, but was still very cold and dry, suggesting that the influence of the southwest monsoon was still very weak. From 18.7 to 17.7 ka BP, the moderate decrease in representation of Artemisia, Lamiaceae, and Gesneriaceae, but still maintaining relatively high percentages of these herb taxa, the increase of tree pollen percentages, and obviously higher pollen accumulation rate indicate that alpine meadow retreated and the treeline rose slightly. We consider that temperature and humidity increased slightly, implying a slight strengthening of the southwest monsoon. Between 17.7 and 17 ka BP, pollen percentage peaks of Pinus, Tsuga, deciduous oaks, Theaceae, Apiaceae, Solanaceae, and Liliaceae, and obvious troughs of Artemisia, Thalictrum, Scrophulariaceae, and Lamiaceae indicate that alpine meadow was further reduced and the treeline rose, but the treeline was still located below Tiancai Lake. These vegetation changes indicate that the climate was milder and more humid than that during the previous period, suggesting that the southwest monsoon strengthened. During the period 17–15.8 ka BP, the lower tree pollen percentages and higher herb pollen percentages for Artemisia, Lamiaceae and Gesneriaceae, compared to the period 17.7–17 ka BP indicate that alpine meadow expanded markedly and the treeline declined. It can be inferred that the climate was colder and drier, and the southwest monsoon was weaker than the preceding period 17.7–17 ka BP.

18

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

Between 15.8 and 14.4 ka BP, the gradual increase of tree pollen percentages and gradual decrease of herb pollen percentages, the highest Betula pollen percentages, marked increase of Ulmus pollen, quick decline of Artemisia pollen, and relatively high pollen accumulation rate indicate that alpine meadow shrank markedly and the treeline shifted to higher altitudes. The temperature requirement of Betula forest is similar to that of sub-alpine cold-temperate conifer forest, but more resistant to dry climatic condition. It is also considered a pioneer vegetation community preceding the establishment of sub-alpine cold-temperate conifer forest (Shen et al., 1996). The steep rise of Betula pollen to the highest values during the period probably indicates that Betula forest, an early successional forest, had reached the highest altitude it could achieve since 21 ka BP. These vegetation changes may have been the result of temperature rising markedly and humidity increasing slightly, implying a relatively strong southwest monsoon. During the period 14.4–12.9 ka BP, Picea/Abies increases sharply to the highest values for the entire core, while Betula pollen percentages are still relatively high with a slowly decreasing tendency, which means that the early successional stands (Betula forest) were later gradually replaced by Picea/Abies forests. This vegetation succession is similar to the observation that Betula forms stands successional to the Abies forests in the modern vegetation in northwestern Sichuan (Taylor and Qin, 1988). At the same time, high pollen percentages of trees and low pollen percentages of herbs, obvious increase of Pinus pollen, gradual decline of Ulmus pollen, very low Artemisia pollen percentages, and further increase of the pollen accumulation rate during the period indicate that Pinus forest occurred in low altitudes and Picea/ Abies forest expanded upward. The treeline (the upper limit of Picea/ Abies forest) reached its highest altitude for the deglacial, while alpine meadow retreated to its minimum extent, but still occupied regions around and above Tiancai Lake. It can be inferred that temperature might have increased slightly and that humidity increased markedly between 14.4 and 12.9 ka BP, suggesting the influence of the southwest monsoon was at its maximum in this record. From 12.9 to 11.5 ka BP, a further increase of Pinus pollen, gradual decreases of Picea/Abies and Betula pollen, and a slight increase of Artemisia pollen suggest that the treeline dropped slightly and the alpine meadow expanded slightly. This may have been the result of slight declines in temperature and humidity because of a weakening of the southwest monsoon. The climate history revealed by the pollen record is consistent with the PCA results of 37 pollen types from 81 samples before 11.5 ka BP. These climate changes are clearly tracked by the PCA axis 1 sample score (Fig. 7a). Thus, the PCA axis 1 sample score serves as a climatic proxy representing the climate changes from the pollen record in Tiancai Lake to compare with regional or extra-regional climate histories (Fig. 7). 5.2. Comparison with regional climate histories The pollen record and core lithology of Tiancai Lake suggest that very cold climate conditions existed between 21 and 19.9 ka BP. This is followed by a slight upturn in climate conditions, but still very cold and dry climate conditions were maintained until 18.7 ka BP. The period 21–18.7 ka BP belongs to the end of the LGM, occurring globally from 26.5 to 19 ka BP, the most recent interval in Earth history when global ice sheets reached their maximum integrated volume (Clark et al., 2009). The paleoclimatic records from southwestern China and other regions affected by the southwest monsoon show that the regional characteristics and timing of the LGM remain open to debate. In these regions, most records suggest that cold and dry or cold and moderately dry conditions existed for the LGM. For example, the pollen record from DC section of Heihe Pasture in Zoige, the northeastern Tibetan Plateau shows a lack of pollen during the period from about 18–16 14C ka BP (20–18 cal. ka BP), indicating that the climate was cold and dry (Liu et al., 1995). The CaCO3, total carbon and nitrogen proxies from Xingyun

Lake and Qilu Lake in Yunnan Province suggest that the climate from 38 to 12 ka BP was cold and dry (Hodell et al., 1999). The oxygen isotope record of the stalagmite from Qixing cave, Guizhou Province indicates that the climate between 29.6 and 14.6 ka BP was also cold and dry (Peng et al., 2002). A combination of variables including pollen, charcoal, particle size, magnetic susceptibility and loss-on-ignition from Shudu Lake in southwestern China shows that climatic conditions between 22.6 and 17.7 ka BP were cold and dry (Cook et al., 2011). The late-Quaternary climate history of monsoonal Central Asia inferred from 75 paleoclimatic records suggests that the LGM at ~21 ka BP was characterized by dry or moderately dry climate conditions (Herzschuh, 2006). In addition, some workers have proposed that cold and humid or cool and humid conditions occurred during the LGM. For example, the multi-proxy sediment records from Heqing paleolake (Jiang et al., 1998) and Napahai Lake (Yin et al., 2002), northwestern Yunnan Province, indicate that the climate during the LGM (23.7–14.2 14C ka BP and 32–15 14C ka BP, respectively) was cold and humid. The pollen record from Menghai region, Yunnan Province shows that relatively high percentages of Dacrydium pollen occurred between 27 and 11.9 14C ka BP, indicative of cool and humid climate conditions (Tang, 1992). The pollen record from Cao 2 core, Dianchi Lake, Yunnan Province suggests that the climate was also cool and humid from 28.4 to 11.8 14C ka BP (Wu et al., 1991). The causes of these varying descriptions of climatic conditions and timing associated with the LGM in the regions influenced by the southwest monsoon are related to low temporal sampling resolution, age uncertainties, regional sensitivity, and even different interpretations of proxy data. The cold and dry climate disclosed by the pollen record and core lithology from this study supports the viewpoint suggested by most records from southwestern China and other regions affected by the southwest monsoon, that climate change is primarily affected by low solar insolation (Fig. 7h, red line) and combined radiative forcing from CO2, CH4, and N2O (Fig. 7h, black line). The pollen assemblage and PCA results of pollen data in this study indicate that the temperature and humidity began to increase from 18.7 ka BP, then relatively milder and more humid climate conditions were maintained between 17.7 and 17 ka BP. Most published paleoclimate records in southwestern China don't find this result. Only Cook et al. (2011) concluded that a phase of warmer and wetter climatic conditions was observed between c. 17.7 and 17.4 ka BP based on multiproxy analysis in Shudu Lake, Yunnan Province, China. Recently, Wang et al. (2014) considered that regional temperatures began to rise from ~ 18 ka BP based on the expansion of planktonic taxa (Stephanodiscus and Asterionella formosa) and increased %TOC from Lugu Lake lying at the boundary between Yunnan Province and Sichuan Province in southwestern China. In other regions, many records show that initial late glacial warming was at ~19 ka BP. For example, the δ 18O record of stalagmite M1–5 from Socotra Island, Yemen, in the northwest Indian Ocean suggests that Indian Ocean rainfall increased slightly between ~ 19 and ~ 17.4 ka BP (Fig. 7b) (Shakun et al., 2007). SST derived from δ 18O values from the planktic foraminifer Globigerinoides ruber of the NIOP905 core from the western Arabian Sea (Fig. 7c) (Huguet et al., 2006) and SSTs derived from Mg/Ca ratios and δ 18O values from the planktic foraminifer G. ruber of the SK218/1 core from the Bay of Bengal and the AAS9/21 core from the eastern Arabian Sea (Naidu and Govil, 2010) show that initial late glacial warming in the northern Indian Ocean began around 19 ka BP. The mean effective moisture in monsoonal Central Asia inferred from 75 paleoclimate records indicates that the climate ameliorated at ~19 ka BP following the LGM and a phase of stable and slightly wetter conditions was maintained between 18.5 and 17 ka BP in Asia (Fig. 7d) (Herzschuh, 2006). The principal components (PC1s) for the Indian Ocean and global land temperatures show that the initial temperature increases in the Indian Ocean and global land started at ~ 18.5 and ~ 18.8 ka BP, respectively (Fig. 7e, black and red lines) (Clark et al., 2012). Within the limits of absolute chronologies, all tropical and subtropical Pacific SST records show an onset of deglacial

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

warming at 19 ± 1 ka BP, coeval with the onset of the deglacial rise in sea level (Kiefer and Kienast, 2005). However, the timing of initial late glacial warming remains controversial. Dansgaard et al. (1971) first observed and Johnsen et al. (1992) and Grootes et al. (1993) confirmed that the first abrupt deglacial warming event over the North Atlantic and Greenland occurred 14.5 ± 0.15 ka BP (Fig. 7g). Consistent with this viewpoint, the δ 18O records from Dongge cave and Hulu cave stalagmites indicate that the East Asian Monsoon strengthened first at ~14.5 ka BP after the LGM (Fig. 7f) (Wang et al., 2001a; Dykoski et al., 2005). The other records from monsoonal Central Asia demonstrate also that the first strong intensification of the Asian monsoon circulation after the LGM, took place at the beginning of the BA (e.g. Zhou et al., 1999; Wang et al., 2001b). Deglacial warming ca. 19 ka BP in northwestern Yunnan Province from our study differs from the North Atlantic and Greenland temperature records, and the East Asian Monsoon records based on Stalagmites Dongge and Hulu caves, indicating that the initial warming is not primarily affected by ice volume in high latitudes of Northern Hemisphere. However, it is synchronous with the increase of the mean effective moisture in monsoonal Central Asia (Herzschuh, 2006), the initial late glacial temperature increases in the Indian Ocean (Naidu and Govil, 2010; Clark et al., 2012), the tropical and subtropical Pacific Ocean (Kiefer and Kienast, 2005), Antarctica (Petit et al., 1999), and global land (Clark et al., 2012), suggesting a strong connection in the propagation of climate signals among these regions. Average summer insolation for 25° N started to increase ca. 21 ka BP (Fig. 7h, red line), indicating that this increase led the initial late glacial warming. With an increase of summer insolation in low-latitude tropical regions, the tropical Indian Ocean received more heat energy, resulting in a strengthening of the southwest monsoon. The strengthening southwest monsoon affected southwestern China and caused it to warm at 18.7 ka BP. From 17 to 15.8 ka BP, our study indicates cold and dry climate conditions, which matches the H1 event, one of the Heinrich events first documented in the North Atlantic as anomalous occurrences of ice-rafted detritus (IRD) (Heinrich, 1988). There is good evidence for a global, or at least a Northern Hemisphere-wide, footprint for the H1 event (Broecker, 1994). For example, a late glacial pollen record from Naleng Lake, southeastern Tibetan Plateau, indicates low effective moisture between 17.7 and 14.8 ka BP (Kramer et al., 2010). The δ 18O record of stalagmite M1–5 from Socotra Island, Yemen, in the northwest Indian Ocean indicates an abrupt drying event at ~16.4 ka BP, perhaps related to the H1 (Fig. 7b) (Shakun et al., 2007). Low SSTs during H1 are indicated by the δ 18O record from the planktic foraminifer Globigerinoides ruber of the NIOP905 core from the western Arabian Sea (Fig. 7c) (Huguet et al., 2006) and the principal component (PC1) for the Indian Ocean temperatures (Fig. 7e, black line) (Clark et al., 2012). The reconstructed mean effective moisture indicates slightly drier conditions between 17.2 and 15.4 ka BP in monsoonal Central Asia, corresponding to H1 (Fig. 7d) (Herzschuh, 2006). The H1 event was clearly recorded in the δ 18O records from Dongge cave (southwestern China) (Dykoski et al., 2005), Hulu cave (Eastern China) (Fig. 7f) (Wang et al., 2001a), and Songjia cave (central China) (Zhou et al., 2008). Bond et al. (1993) and Broecker (1994) pointed out that the H1 event was also documented in ice cores (e.g. Fig. 7g). However, H1 was not found in all paleoclimate records. For example, lake sediment records from Shudu Lake (Cook et al., 2011) and Lugu Lake (Wang et al., 2014) both fail to capture the abrupt event, which might be due to low sensitivity to global climate changes or inadequate sampling resolution. The cause of H1 is still under debate. Currently, the leading hypothesis involves a slowdown of the ocean's thermohaline circulation. The large ice sheets that rimmed the North Atlantic during the last glacial began to melt by ca. 19 ka BP under the influence of an increase in Northern Hemisphere summer radiation beginning ca. 21 ka BP followed by a strengthening Atlantic meridional overturning circulation (AMOC) that transported heat energy to high-latitudes. Large amounts of freshwater pouring into the North Atlantic Ocean in the form of

19

water and icebergs reduced ocean salinity enough to slow deepwater formation and the thermohaline circulation. Since the thermohaline circulation plays an important role in transporting heat northward, a slowdown would cause the North Atlantic and even the Northern Hemisphere to cool (Denton et al., 2010; Xu et al., 2013). The pollen record of core TCYL1 shows that temperature rose markedly and humidity increased after 15.8 ka BP, and then the temperature and humidity reached the highest levels between 14.4 and 12.9 ka BP. This warm–wet maximum period is in good agreement in timing and duration with the BA between 14.6 and 12.8 ka BP shown in the GISP2 record (Fig. 7g). The BA is also mirrored in the δ 18O records from stalagmite M1–5 (Socotra Island, Yemen in the northwest Indian Ocean) (Fig. 7b), Dongge cave (southwestern China) (Dykoski et al., 2005), Hulu cave (Eastern China) (Fig. 7f), the δ 18O record from the planktic foraminifer Globigerinoides ruber of the NIOP905 core from the western Arabian Sea (Fig. 7c), the reconstructed mean effective moisture in monsoonal Central Asia (Fig. 7d), and the principal components (PC1s) for global land temperature and precipitation (Fig. 7e, red and green lines). An evident temperature increase between 15.8 and 14.4 ka BP indicated by this study occurred at the end of H1 and before the BA, which differs from the initial late glacial warming taking place at 18.7 ka BP. The finding is consistent with pre-Bølling warming exhibited by a growing number of records from around the globe (Hill et al., 2006). For example, δ 18O values of both planktonic and benthics from Santa Barbara Basin, California indicate that surface and intermediate waters began to warm at ~ 16.5 ka BP, some 2 ka prior to Termination IA (14.7 ka BP) (Hill et al., 2006). Based on data from core 74KL from the western Arabian Sea, Sirocko et al. (1996) suggest that an earlier event of monsoon intensification at ~ 16 ka BP occurred at the end of H1. The increase in global atmospheric methane concentrations recorded in the GRIP ice core started about 16 ka BP, indicating an early onset of environmental change in the tropics (Chappellaz et al., 1993). In addition, distinct warming or moisture increases at 16–15.4 ka BP are observed in the δ 18O record from the planktic foraminifer Globigerinoides ruber of the NIOP905 core from the western Arabian Sea (Fig. 7c), the reconstructed mean effective moisture in monsoonal Central Asia (Fig. 7d), the principal components (PC1s) for global land temperature and precipitation (Fig. 7e, red and green lines), the δ 18O records from Dongge and Hulu caves (Fig. 7f), and even the δ 18O record from GISP2 (Fig. 7g). The combined variations in radiative forcing due to greenhouse gases is dominated by CO2, but abrupt changes in CH4 and N2O modulate the overall structure, accentuating the rapid increase at 14.7 ka BP and the clear increase at 15.7 ka BP (Fig. 7h, black line). At the same time, because of the cold climate during the H1 event, the addition of freshwater from ice melt decreased, ocean salinity and deepwater formation increased, and the Atlantic meridional overturning circulation (AMOC) recovered. Thus, we consider that the pre-Bølling and BA warming transitions are caused by a combination of radiative forcing, the gradual recovery of AMOC from the H1, and an AMOC overshoot, which is also suggested by Liu et al.'s simulation (Liu et al., 2009). The slight declines of temperature and humidity from 12.9 to 11.5 ka BP displayed by the pollen record from Tiancai Lake correspond to the YD, a millennial-scale cold period (~1300 yr duration) between approximately 12.8 and 11.5 ka BP (Muscheler et al., 2008), reflected as drier climate conditions in areas from monsoonal Central Asia (Fig. 7d) and a cold event in Greenland (Fig. 7g). The termination and climatic characteristic of the YD in southwestern China and other regions affected by the southwest monsoon have been discussed by Xiao et al. (2014). Due to the basal age of the TCK1 core only reaching 12.2 ka BP, Xiao et al. (2014) did not deduce the age of the start of the YD. In this study the integrated core from Tiancai Lake (TCK1 and TCYL1) shows that 12.9 ka BP is the age of the beginning of the YD, which is synchronous with a widespread accepted age of the beginning of the YD. The cause of the YD is considered to be similar to that of H1 and the hypothesis about a slowdown of the ocean's thermohaline circulation is also likely to be involved in this period.

20

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21

6. Conclusions The pollen and conifer stoma records of an integrated core spanning the last 21 ka BP from Tiancai Lake show that Tiancai Lake was above the upper limit of Picea/Abies forest (a treeline in the study area) between ~ 21 and 11.5 ka BP while, in the Holocene (since 11.5 ka BP), Tiancai Lake was below the upper limit of Picea/Abies forest. The climate after 11.5 ka BP was markedly warmer and moister than the climate before 11.5 ka BP, and the southwest monsoon after 11.5 ka BP was obviously stronger than that before 11.5 ka BP. Under the background of cold and dry climate conditions during the period 21–11.5 ka BP, this study reveals eight significant shifts in the vegetation, climate and monsoon between ~21 and 11.5 ka BP based on the pollen record, core lithology, and PCA analysis of pollen data. Between 21 and 19.9 ka BP, the lake basin was covered by glaciers, indicating very cold climate conditions and a very weak southwest monsoon. During the period 19.9–18.7 ka BP the vegetation around Tiancai Lake was alpine meadow and forests were far from Tiancai Lake implying slightly warmer, but still very cold and dry, climate conditions and a still very weak southwest monsoon. From 18.7 to 17.7 ka BP, alpine meadow retreated upslope and the treeline rose slightly, suggesting that temperature and humidity increased slightly, and the southwest monsoon was beginning to strengthen. Between 17.7 and 17.0 ka BP, alpine meadow continued to retreat upslope and the treeline rose, indicating milder and more humid climate conditions and a continued strengthening of the southwest monsoon. During the period 17.0–15.8 ka BP, alpine meadow expanded markedly and the treeline declined, implying cold and dry climate conditions and a weakening southwest monsoon. From 15.8 to 14.4 ka BP, the extent of alpine meadow reduced, the treeline shifted to higher altitudes, and Betula forest extended to the highest altitude recorded since 21 ka BP, suggesting that temperature rose markedly, humidity increased slightly, and the southwest monsoon was relatively strong. During the period 14.4– 12.9 ka BP, the treeline attained its highest altitude and alpine meadow retreated to its minimum extent, indicating the warmest and wettest climate conditions and the strongest southwest monsoon influence for the deglacial period. From 12.9 to 11.5 ka BP, the treeline dropped slightly and alpine meadow expanded slightly, suggesting slight declines of temperature and humidity because of weakening of the southwest monsoon. The paleoclimate record in this study detects for the first time the three most prominent deglacial events (the H1, BA and YD) of the Northern Hemisphere from a lake sediment record in southwestern China. In addition, this study finds that the initial late glacial warming in northwestern Yunnan Province was at ~18.7 ka BP. The temperature increase between 15.8 and 14.4 ka BP occurred at the end of the H1 and before the BA, which is consistent with pre-Bølling warming. The cause of these late glacial abrupt events is still under debate. In this study, the hypothesis involving a slowdown of the ocean's thermohaline circulation is considered appropriate to explain the observed changes. Acknowledgments This research was financially supported by the National Basic Research Program of China (2010CB950201), the National Natural Science Foundation of China (41272188, 41072132), and the Nanjing Institute of Geography and Limnology, CAS (NIGLAS2011KXJ02). We thank two anonymous reviewers who gave us important advice that let us improve the quality of the manuscript. We also thank Professors Enlou Zhang and Yanling Li, and Dr. Qian Wang for doing the fieldwork. References Alley, R.B., Clark, P.U., 1999. The deglaciation of the northern hemisphere: a global perspective. Ann. Rev. Earth Planet. Sci. 27, 149–182.

Anderson, N.J., 1986. Diatom biostratigraphy and comparative core correlation within a small lake basin. Hydrobiologia 134, 105–112. Barker, S., Knorr, G., Vautravers, M.J., Diz, P., Skinner, L.C., 2010. Extreme deepening of the Atlantic circulation during deglaciation. Nat. Geosci. 3, 567–571. Barnett, C., Dumayne-Peaty, L., Matthews, J.A., 2001. Holocene climatic change and treeline response in Leirdalen, central Jotunheimen, south central Norway. Rev. Palaeobot. Palynol. 117, 119–137. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quat. Sci. Rev. 10, 297–317. Birks, H.J.B., Gordon, A.D., 1985. Numerical Methods in Quaternary Pollen Analysis. Academic Press, London. Blaauw, M., Christen, J.A., 2011. Flexible paleoclimate age–depth models using an autoregressive gamma process. Bayesian Anal. 6, 457–474. Blunier, T., Brook, E.J., 2001. Timing of millennial-scale climate change in Antarctica and Greenland during the last glacial period. Science 291, 109–112. Blunier, T., Chappellaz, J., Schwander, J., D¨allenbach,, A., Stauffer, B., Stocker, T.F., Raynaud, D., Jouzel, J., Clausen, H.B., Hammer, C.U., Johnsen, S.J., 1998. Asynchrony of Antarctic and Greenland climate change during the last glacial period. Nature 394, 739–743. Bond, G.C., Broecker, W.S., Johnsen, S., McManus, J.F., Labeyrie, L., Jouzel, J., Bonani, G., 1993. Correlation between climate records from North Atlantic sediments and Greenland ice. Nature 365, 143–147. Broecker, W.S., 1994. Massive iceberg discharges as triggers for global climate change. Nature 372, 421–424. Brook, E.J., White, J.W.C., Schilla, A.S.M., Bender, M.L., Barnett, B., Severinghaus, J.P., Taylor, K.C., Alley, R.B., Steig, E.J., 2005. Timing of millennial-scale climate change at Siple Dome, West Antarctica, during the last glacial period. Quat. Sci. Rev. 24, 1333–1343. Chappellaz, J., Bluniert, T., Raynaud, D., Barnola, J.M., Schwander, J., Stauffert, B., 1993. Synchronous changes in atmospheric CH4 and Greenland climate between 40 and 8 kyr BP. Nature 366, 443–445. Clark, P.U., Pisias, N.G., Stocker, T.F., Weavers, A.J., 2002. The role of the thermohaline circulation in abrupt climate change. Nature 415, 863–869. Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., Hostetler, S.W., Marshall McCabe, A., 2009. The Last Glacial Maximum. Science 325, 710–714. Clark, P.U., Shakun, J.D., Baker, P.A., Bartlein, P.J., Brewer, S., Brook, E., Carlson, A.E., Cheng, H., Kaufman, D.S., Liu, Z.Y., Marchitto, T.M., Mix, A.C., Morrill, C., Otto-Bliesner, B.L., Pahnke, K., Russell, J.M., Whitlock, C., Adkins, J.F., Blois, J.L., Clark, J., Colman, S.M., Curry, W.B., Flower, B.P., He, F., Johnson, T.C., Lynch-Stieglitz, J., Markgraf, V., McManus, J., Mitrovica, J.X., Moreno, P.I., Williams, J.W., 2012. Global climate evolution during the last deglaciation. Proc. Natl. Acad. Sci. U. S. A. 109 (19), E1134–E1142. Cook, C.G., Jones, R.T., Langdon, P.G., Leng, M.J., Zhang, E.L., 2011. New insights on Late Quaternary Asian palaeomonsoon variability and the timing of the Last Glacial Maximum in southwestern China. Quat. Sci. Rev. 30, 808–820. Dansgaard, W., Johnsen, S.J., Clausen, H.B., Langway Jr., C.C., 1971. Climatic record revealed by the Camp Century ice core. In: Turekian, K.K. (Ed.), The Late Cenozoic Glacial Ages Symposium 1969. Yale University Press, New Haven, Connecticut, pp. 37–56. Denton, G.H., Anderson, R.F., Toggweiler, J.R., Edwards, R.L., Schaefer, J.M., Putnam, A.E., 2010. The last glacial termination. Science 328, 1652–1656. Duplessy, J.C., Labeyrie, L., Arnold, M., Paterne, M., Duprat, J., van Weering, T.C.E., 1992. Changes in surface salinity of the North Atlantic Ocean during the last deglaciation. Nature 358, 485–488. Dykoski, C.A., Edwards, R.L., Cheng, H., Yuan, D.X., Cai, Y.J., Zhang, M.L., Lin, Y.S., Qing, J.M., An, Z.S., Revenaugh, J., 2005. A high-resolution, absolute-dated Holocene and deglacial Asian monsoon record from Dongge Cave, China. Earth Planet. Sci. Lett. 233, 71–86. EPICA Community Members, 2006. One-to-one coupling of glacial climate variability in Greenland and Antarctica. Nature 444, 195–198. EPICA Project Members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429, 623–628. Gherardi, J.M., Labeyrie, L., Nave, S., Francois, R., McManus, J.F., Cortijo, E., 2009. Glacial– interglacial circulation changes inferred from Pa-231/Th-230 sedimentary record in the North Atlantic region. Paleoceanography 24, PA2204. http://dx.doi.org/10.1029/ 2008PA001696. Grimm, E.C., 1987. CONISS: a Fortran 77 program for stratigraphically constrained cluster analysis by method of incremental sum of squares. Comput. Geosci. 13, 13–35. Grimm, E.C., 1993. TILIA v2.0 (Computer Software). Research and Collections CentreIllinois State Museum, Springfield, Ill. Grootes, P.M., Stuiver, M., 1999. GISP2 Oxygen Isotope Data. http://dx.doi.org/10.1594/ PANGAEA.56094. Grootes, P.M., Stuiver, M., White, J.W.C., Johnsen, S., Jouzel, J., 1993. Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366, 552–554. Heinrich, H., 1988. Origin and consequences of cyclic ice rafting in the northeast Atlantic Ocean during the past 130,000 years. Quat. Res. 29, 142–152. Herzschuh, U., 2006. Palaeo-moisture evolution at the margins of the Asian monsoon during the last 50 ka. Quat. Sci. Rev. 25, 163–178. Hill, T.M., Kennett, J.P., Pak, D.K., Behl, R.J., Robert, C., Beaufort, L., 2006. Pre-Bølling warming in Santa Barbara Basin, California: surface and intermediate water records of early deglacial warmth. Quat. Sci. Rev. 25, 2835–2845. Hodell, D.A., Brenner, M., Kanfoush, S.L., Curtis, J.H., Stoner, J.S., Song, X.L., Wu, Y., Whitmore, T.J., 1999. Paleoclimate of southwestern China for the past 50,000 yr inferred from lake sediment records. Quat. Res. 52, 369–380. Huguet, C., Kim, J.H., Damsté, J.S.S., Schouten, S., 2006. Reconstruction of sea surface temperature variations in the Arabian Sea over the last 23 kyr using organic proxies (TEX86 and UK′37). Paleoceanography 21, PA3003. Jiang, X.Z., Wang, S.M., Yang, X.D., 1998. Paleoclimatic and environmental changes over the last 30000 years in Heqing Basin, Yunnan Province. J. Lake Sci. 10 (2), 10–16 (in Chinese).

X. Xiao et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 406 (2014) 9–21 Jiménez-Moreno, G., Fawcett, P.J., Anderson, R.S., 2008. Millennial- and centennial scale vegetation and climate changes during the late Pleistocene and Holocene from northern New Mexico (USA). Quat. Sci. Rev. 27, 1442–1452. Johnsen, S.J., Dansgaard, W., Clausen, H.B., Langway, C.C., 1972. Oxygen isotope profiles through Antarctic and Greenland ice sheets. Nature 235, 429–434. Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C.U., Iversen, P., Jouzel, J., Stauffer, B., Steffensen, J.P., 1992. Irregular glacial interstadials recorded in a new Greenland ice core. Nature 359, 311–313. Jouzel, J., Masson, V., Cattani, O., Falourd, S., Stievenard, M., Stenni, B., Longinelli, A., Johnsen, S.J., Steffenssen, J.P., Petit, J.R., Schwander, J., Souchez, R., Barkov, N.I., 2001. A new 27 ky high resolution East Antarctic climate record. Geophys. Res. Lett. 28, 3199–3202. http://dx.doi.org/10.1029/2000GL012243. Kiefer, T., Kienast, M., 2005. Patterns of deglacial warming in the Pacific Ocean: a review with emphasis on the time interval of Heinrich Event 1. Quat. Sci. Rev. 24 (7–9), 1063–1081. Knorr, G., Lohmann, G., 2003. Southern Ocean origin for the resumption of Atlantic thermohaline circulation during deglaciation. Nature 424, 532–535. Kramer, A., Herzschuh, U., Mischke, S., Zhang, C.J., 2010. Late glacial vegetation and climate oscillations on the southeastern Tibetan Plateau inferred from the Lake Naleng pollen profile. Quat. Res. 73, 324–335. Liu, G.X., Shen, Y.P., Wang, R., Wang, S.M., 1995. The vegetation and climatic changes in Zoige during the last 20000 years determined by pollen records. J. Glaciol. Geocryol. 17 (2), 132–137 (in Chinese). Liu, Z., Otto-Bliesner, B.L., He, F., Brady, E.C., Tomas, R., Clark, P.U., Carlson, A.E., LynchStieglitz, J., Curry, W., Brook, E., Erickson, D., Jacob, R., Kutzbach, J., Cheng, J., 2009. Transient simulation of last deglaciation with a new mechanism for Bølling–Allerød warming. Science 327, 310–314. Maddy, D., Brew, J.S., 1995. Statistical modelling of quaternary science data technical guide. Quaternary Research Association, vol. 5. Cambridge University Press, Cambridge. McManus, J.F., Francois, R., Gherardi, J.-M., Keigwin, L.D., 2004. Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial climate changes. Nature 428, 834–837. Moore, P.D., Webb, J.A., 1978. An Illustrated Guide to Pollen Analysis. Hodder and Stoughton, London pp. 5–50. Muscheler, R., Kromer, B., Björck, S., Svensson, A., Friedrich, M., Kaiser, K.F., Southon, J., 2008. Tree rings and ice cores reveal 14C calibration uncertainties during the Younger Dryas. Nat. Geosci. 1 (4), 263–267. Naidu, P.D., Govil, P., 2010. New evidence on the sequence of deglacial warming in the tropical Indian Ocean. J. Quat. Sci. 25 (7), 1138–1143. Nakagawa, T., Brugiapaglia, E., Digerfeldt, G., Reille, M., de Beaulieu, J., Yasuda, Y., 1998. Dense-media separation as a more efficient pollen extraction method for use with organic sediment/deposit samples: comparison with the conventional method. Boreas 27, 15–24. Nalder, I.A., Wein, R.W., 1998. Spatial interpolation of climatic normals: test of a new method in the Canadian boreal forest. Agric. For. Meteorol. 92, 211–225. NGRIP members, 2004. High-resolution record of Northern Hemisphere climate extending into the last interglacial period. Nature 431, 147–151. Pedro, J.B., van Ommen, T.D., Rasmussen, S.O., Morgan, V.I., Chappellaz, J., Moy, A.D., Masson-Delmotte, V., Delmotte, M., 2011. The last deglaciation: timing the bipolar seesaw. Clim. Past 7, 671–683. Pellatt, M.G., Mathewes, R.W., 1997. Holocene tree line and climate change on the Queen Charlotte Islands, Canada. Quat. Res. 48, 88–99. Peng, Z.C., Zhang, Z.F., Cai, Y.J., Zhang, M.L., Lin, Y.S., Liu, G.J., 2002. The paleoclimatic records from the late Pleistocene stalagmite in Guizhou Qixing cave. Quat. Sci. 22 (3), 273–282 (in Chinese). Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, J., Delaygue, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V., Lorius, C., Pépin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok Ice Core, Antarctica. Nature 399, 429–436. Qin, J.M., Yuan, D.X., Lin, Y.S., Zhang, M.L., 2001. Short-term abrupt climate events of last glaciation: the evidences from stalagmite records in southwest China. Bull. Mineral. Petrol. Geochem. 20 (4), 391–393 (in Chinese). Qin, J.M., Yuan, D.X., Cheng, H., Lin, Y.S., Zhang, M.L., Wang, H., Feng, Y.M., Tu, L.L., 2004. A high resolution late Pleistocene climato-stratigraphy of 4 stalagmites from Qixing Cave, Duyun, Guizhou. Quat. Sci. 24 (3), 318–324 (in Chinese). Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51 (4), 11–50. Ritz, S.P., Stocker, T.F., Grimalt, J.O., Menviel, L., Timmermann, A., 2013. Estimated strength of the Atlantic overturning circulation during the last deglaciation. Nat. Geosci. 6, 208–212. Shakun, J.D., Burns, S.J., Fleitmann, D., Kramers, J., Matter, A., Al-Subary, A., 2007. A highresolution, absolute-dated deglacial speleothem record of Indian Ocean climate from Socotra Island, Yemen. Earth Planet. Sci. Lett. 259, 442–456. Shen, C.M., Tang, L.Y., Wang, S.M., 1996. Vegetation and climate during the last 22 000 years in Zoigê region. Acta Micropalaeontologica Sin. 13 (4), 401–406 (in Chinese).

21

Sirocko, F., Garbe-Schönberg, D., McIntyre, A., Molfino, B., 1996. Teleconnections between the subtropical monsoons and high-latitude climates during the last deglaciation. Science 272, 526–529. Stuiver, M., Grootes, P.M., 2000. GISP2 oxygen isotope ratios. Quat. Res. 53, 277–283. Tang, L.Y., 1992. Vegetation and climate history at Menghai, Yunnan during the past 42000 years. Acta Micropalaeontologica Sin. 9 (4), 433–455 (in Chinese). Taylor, A.H., Qin, Z., 1988. Tree replacement patterns in subalpine Abies-Betula forests, Wolong Natural Reserve, China. Plant Ecol. 78, 141–152. ter Braak, C.J.F., Šmilauer, P., 2002. CANOCO Reference Manual and Cano-Draw for Windows User's Guide: Software for Canonical Community Ordination (Version 4.5). Microcomputer Power, Ithaca. Thornalley, D.J.R., McCave, I.N., Elderfield, H., 2010. Freshwater input and abrupt deglacial climate change in the North Atlantic. Palaeoceanography 25, PA1201. http://dx.doi. org/10.1029/2009PA001772. Waelbroeck, C., Duplessy, J.-C., Michel, E., Labeyrie, L., Paillard, D., Duprat, J., 2001. The timing of the last deglaciation in North Atlantic climate records. Nature 412, 724–727. Wang, H., Liu, H., Cui, H., Abrahamsen, N., 2001a. Terminal Pleistocene/Holocene palaeoenvironmental changes revealed by mineral-magnetism measurements of lake sediments for Dali Nor area, southeastern Inner Mongolia Plateau, China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 170, 115–132. Wang, Y., Cheng, H., Edwards, R.L., An, Z., Wu, J., Shen, C., Dorale, J.A., 2001b. A high resolution absolute-dated late Pleistocene monsoon record from Hulu Cave, China. Science 294, 2345–2348. Wang, Q., Yang, X.D., Anderson, N.J., Zhang, E.L., Li, Y.L., 2014. Diatom response to climate forcing of a deep, alpine lake (Lugu Hu, Yunnan, SW China) during the Last Glacial Maximum and its implications for understanding regional monsoon variability. Quat. Sci. Rev. 86, 1–12. Wu, Y.S., Chen, Y.S., Xiao, J.Y., 1991. A preliminary study on vegetation and climate changes in Dianchi Lake area in the last 40000 years. Acta Bot. Sin. 33 (6), 450–458 (in Chinese). Xiao, X.Y., Shen, J., Wang, S.M., Xiao, H.F., Tong, G.B., 2008. Plant diversity and its relationship to palaeoenvironment revealed by pollen record since 2.78 Ma in the Heqing deep drilling core. Chin. Sci. Bull. 53 (23), 3686–3698. Xiao, X.Y., Tong, S.M., Shen, J., Wang, S.M., Yang, X.D., Tong, G.B., 2009. Altitudinal distribution of surface pollen and their relations to modern vegetation in the Yulong mountains, Yunnan Province. Quat. Sci. 29 (1), 80–88 (in Chinese). Xiao, X.Y., Shen, J., Wang, S.M., Xiao, H.F., Tong, G.B., 2010. The variation of the southwest monsoon from the high resolution pollen record in Heqing Basin, Yunnan Province, China for the last 2.78 Ma. Palaeogeogr. Palaeoclimatol. Palaeoecol. 287, 45–57. Xiao, X.Y., Shen, J., Wang, S.M., 2011. Spatial variation of modern pollen from surface lake sediments in Yunnan and southwestern Sichuan Province, China. Rev. Palaeobot. Palynol. 165, 224–234. Xiao, X.Y., Haberleb, S.G., Shen, J., Yang, X.D., Han, Y., Zhang, E.L., Wang, S.M., 2014. Latest Pleistocene and Holocene vegetation and climate history inferred from an alpine lacustrine record, northwestern Yunnan Province, southwestern China. Quat. Sci. Rev. 86, 35–48. Xu, D.K., Lu, H.Y., Wu, N.Q., Liu, Z.X., Li, T.G., Shen, C.M., Wang, L., 2013. Asynchronous marine-terrestrial signals of the last deglacial warming in East Asia associated with low- and high-latitude climate changes. Proc. Natl. Acad. Sci. U. S. A. 110 (24), 9657–9662. Yang, Y., Yuan, D.X., Cheng, H., Zhang, M.L., Qin, J.M., Lin, Y.S., Zhu, X.Y., Edwards, R.L., 2010. Precise dating of abrupt shifts in the Asian Monsoon during the last deglaciation based on stalagmite data from Yamen Cave, Guizhou Province, China. Sci. China Earth Sci. 53 (5), 633–641. Yin, Y., Fang, N.Q., Sheng, J.F., Hu, C.Y., Nie, H.G., 2002. Lacustrine records of environmental changes during the last 57 ka in the Napahai Lake, northwestern Yunnan, China. Mar. Geol. Quat. Geol. 22 (4), 99–105 (in Chinese). Zhang, C.J., Mischke, S., 2009. A Lateglacial and Holocene lake record from the Nianbaoyeze Mountains and inferences of lake, glacier and climate evolution on the eastern Tibetan Plateau. Quat. Sci. Rev. 28, 1970–1983. Zhang, M.L., Yuan, D.X., Lin, Y.S., Qin, J.M., 2002. The record of climatic change from stalagmite in Xiashuilong cave of Xuanwei city since 46 ka. Acta Sedimentol. Sin. 20 (1), 124–128 (in Chinese). Zhang, M.L., Cheng, H., Yuan, D.X., Lin, Y.S., Qin, J.M., Zhang, C., 2004a. The high resolution climate records from two stalagmites in Qixing cave of Guizhou and the Heinrich events of the last glacial period. Acta Geosci. Sin. 25 (3), 337–344 (in Chinese). Zhang, Z.H., Zhao, M.X., Yang, X.D., Wang, S.M., Jiang, X.Z., Oldfield, F., Eglintonm, G., 2004b. A hydrocarbon biomarker record for the last 40 kyr of plant input to Lake Heqing, southwestern China. Org. Geochem. 35, 595–613. Zhao, Y., Herzschuh, U., 2009. Modern pollen representation of source vegetation in the Qaidam Basin and surrounding mountains, north-eastern Tibetan Plateau. Veg. Hist. Archaeobotany 18, 245–260. Zhou, W., Head, M.J., Lu, X., An, Z., Jull, A.J.T., Donahue, D., 1999. Teleconnection of climatic events between East Asia and polar, high latitude areas during the last glaciation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 152, 163–172. Zhou, H.Y., Zhao, J.X., Feng, Y.X., Gagan, M.K., Zhou, G.Q., Yan, J., 2008. Distinct climate change synchronous with Heinrich event one, recorded by stable oxygen and carbon isotopic compositions in stalagmites from China. Quat. Res. 69, 306–315.