articles - lsuchinese science bulletin vol. 50 no. 6 march 2005 553 chinese science bulletin 2005...

ARTICLES

Chinese Science Bulletin Vol. 50 No. 6 March 2005 553

Chinese Science Bulletin 2005 Vol. 50 No. 6 553—562

Pollen records and time scale for the RM core of the Zoige Basin, northeastern Qinghai- Tibetan Plateau SHEN Caiming1,2, TANG Lingyu1, WANG Sumin3, LI Chunhai1,3 & LIU Kam-biu2 1. Nanjing Institute of Geology and Paleontology, Chinese Academy of

Sciences, Nanjing 210008, China; 2. Department of Geography and Anthropology, Louisiana State Univer-

sity, Louisiana 70803, USA; 3. Nanjing Institute of Limnology and Geography, Chinese Academy of

Sciences, Nanjing 210008, China Correspondence should be addressed to Tang Lingyu (e-mail: [email protected])

Abstract A continuous pollen record from the Zoige Basin in the northeastern Qinghai-Tibetan Plateau not only provides information on the vegetation and climate changes during the last two glacial/interglacial cycles, but also gives proof to establish the time scale of the upper 60 m of the RM core. Subalpine spruce-fir forests colonized the Zoige Basin during the interglacials and interstadials, implying warm and wet climate conditions. Alpine periglacial desert or dry desert may have existed during the penultimate glacial and the last glacial maxima, respectively. Alpine sedge meadow dominated the landscape during MIS 4. The MIS 3 is punc-tuated by a number of stadials similar to those documented in the Guliya and GISP2 ice cores, as indicated by repeated rise and fall of subalpine spruce-fir forests. Our pollen re-cord reveals a regional climate history similar to those from the neighboring sites, including the Arabian Sea and the Guliya ice core, and thus supports the notion that the Qing-hai-Tibetan Plateau acts as an important link between cli-matic events in the North Atlantic realm and the Asian mon-soon domain. Keywords: Asian monsoon, pollen, time scale, Qinghai-Tibetan Pla-teau.

DOI: 10.1360/04wd0209

The Qinghai-Tibetan Plateau plays a major role in the initiation and strengthening of the Asian monsoon system. However, little is known about the climate and environment changes in the Qinghai-Tibetan Plateau itself during the Quaternary. A few proxy records of the Late Pleistocene climate are available in the 1990s[1—9]. To date, the only available continuous climate records covering the last two glacial-interglacial cycles are from an ice core at Guliya[10], a 120-m-long RH core[11] and a 310.47-m-long RM core[12] in the Zoige Basin. The RM core (33°57′N, 102°21′E) is situated near the western limit of the influ-ence of the Southeast Asian summer monsoon, and also on the northern limit of the area influenced by the Southwest

summer monsoon[13]. Therefore, this region is likely to be sensitive to climate changes related to monsoon variations. Pollen and depositional environment from the upper 60 m of the RM core can provide some insights into the dy-namics of Asian monsoon system from the Late Pleisto-cene onward. Although some results have been published so far on the RM core[12,14,15], detailed interpretation of depositional environment and the chronology of the core need to be refined. Here we show a time framework of the upper 60 m segment of the RM core by comparing a pol-len record with δ 18O series of the Guliya and Greenland ice cores. 1 Environmental setting

The Zoige Basin is located in the northeastern mar-gin of the Qinghai-Tibetan Plateau (32°10′—34°10′N; 101°45′—103°25′E), with an altitude of ca. 3400 m. It is a tectonic basin controlled by WNW-, NE-, and NW- rend-ing faultlines. To the north, south, and east are mountains at elevations exceeding 4000 m a.s.l. The Zoige Basin is drained by the Yellow River (Huanghe River) to the west. However, a huge lake occupied the basin during most of the Pleistocene, until it was gradually drained at about 30—40 ka ago when the Yellow River cut through the mountain barrier to the east[11]. More than 300 m thick lacustrine sediments were deposited in the Zoige Basin during the Pleistocene.

Both temperature and precipitation exhibit a re-markable seasonality in the Zoige Basin. According to the meteorological data of the Maqu, Zoige, and Hongyuan stations, the mean annual temperature is 0.6—1.2 , with ℃

the lowest monthly temperature of −10.9℃ in January and the highest of 1l℃ in July, respectively. The annual mean precipitation is 622—827 mm. Most of which occur in summer months, delivered by the southwestern and southeastern Asian monsoons[16].

The Zoige Basin is vegetated mainly by sedge marshes including Carex muliensis, and Kobresia humilis. Subalpine meadows only occur on the hills and terraces within the basin and on generally sloping mountains near the basin at 3400—3800 m a.s.l. The sedge communities are dominated by Kobresia setchwanensis, and the grass communities by grasses such as Clinelymus nutans, Roegneria nutans, Poa partensis, and Festuca sinensis, together with some species of Compositae, Ranuncu laceae and Leguminosae. Scatted coniferous forests com-posed of Picea asperata, P. wilsonii, P. purpurea, and Abies faxoniana grow on north-facing slopes of some hills at an elevation of ca. 3400 m in this region. Alpine shrub-meadows cover the mountains above 3800 m a.s.1. The dominant components making up the alpine shrub-meadows are shrubs such as Rhododendron spp., Salix sclerophylla, Potentilla fruticosa, Caragana spp., Hippophae spp., and herbs like Kobresia pygmara, K.

ARTICLES

554 Chinese Science Bulletin Vol. 50 No. 6 March 2005

setchwanensis, Stipa calpillaceae, S. aliena, S. purpurea, Festuca ovila, F. rubra, Polygonum viviparum, P. sphaerostachyum, and Thalictrum spp.[16].

A distinct altitudinal vegetation gradient occurs in the mountains surrounding the Zoige Basin. In mountains south to the Zoige Basin, below 3000 m is the zone of montane conifer-hardwood mixed forest dominated by Tsuga chinensis, Pinus densata, P. armandii, Betula platyphylla, Populus davidiana, Quercus liaotungensis, together with Pinus tabulaeformis, Quercus baronii, Tilia intonsa, Fraxinus chinensis, Acer davidii, and Hippophae rhamnoides. From 3000 to 4000 m, fir (Abies faxoniana, A. recurvata, A. squamata A. ernestii) and spruce (Picea purpurea, P. likiangensis, P. asperata) dominate the vege-tation, together with Quercus spp. and Betula spp. Ever-green sclerophyllous oak forest (Quercus aquifolioides) is found on the south-facing slopes at the lower part of this zone (below 3800 m). The zone between 4000 and 4400 m is occupied by alpine shrub-meadow. Alpine periglacial desert appears from 4400 to 4600 m, which is composed of Saussurea medusa, S. lanicens, S. gnaphalodes, S. quercifolaia, S. obovallata, Androsace tapete, Eriophyton wallichii, Phyllophyton comlanatum, Rhodiola dumulosa, R. quadrifida, Corydalis trachycarpa, C. melanochlora, Fritillaris delavayi, and Logotis ramalana. The peaks above 4600 m are covered by ice and snow. Similar vege-tation zones can be found on mountains to the northeast of the Zoige Basin, although their altitudinal limits are gen-erally about 300 m lower[17]. 2 Materials and methods

The RM core, represented by a continuous 310.46- m-long sediment sequence, was drilled at 33°57′N and 102°21′E, 3400 m a.s.l. in the northern part of the Zoige Basin in the summer of 1993. Most part of the core was kept intact, but the top 4.25 m segment was disturbed during coring. This paper presents the results of pollen analysis for samples from the upper part (60—4.25 m) of the RM core.

The core was subsampled at 5 cm intervals from 4.25 to 46.1 m and at 10 cm intervals from 46.1 to 60 m. A total of 217 samples at irregular intervals of 15—40 cm were processed for pollen analyses. Samples were treated with HC1, NaOH, HF, and heavy liquid floating procedures. The weighing or aliquot method[18,19] was used to calculate pollen concentrations, in which we obtained dry weights of original samples and counted all pollen grains and spores in the weighed subsamples. Grimm’s (1983)[20] formula was used to adjust abnormally high pollen concentrations at a few levels. About 100—500 pollen grains were counted for most levels except those having very low pollen concentrations. The mean pollen sum of all the samples slightly exceeds 200 grains. Pollen percentages were calculated based on a sum of all arboreal and upland herb pollen grains.

Two palynological indicators are especially useful in reconstructing past climatic changes in the Zoige Ba-sin——total pollen concentrations and the percentages of arboreal pollen (AP). It is estimated that the ancient lake had an area of about 6300 km2[21] with a catchment of about 16200 km2 [11]. Therefore, more than 90% of pollen into the ancient lake should have been transported by wa-ter[22,23] and derived from the regional vegetation[24]. In the RM core, variations in total pollen concentrations should primarily reflect changes in the density or pollen produc-tivity of the vegetation, if they are not accompanied by corresponding changes in lithology or in sedimentation rate. Both vegetation density and pollen productivity are related to the precipitation. High precipitation should in-crease the vegetation cover and thus the pollen production in a region where moisture is a limiting factor. Moreover, high precipitation increases total pollen concentrations due to greater pollen input from surface runoff and streams[22,23]. So it is reasonable to use total pollen con-centration as a proxy for the regional precipitation when the observed changes in sedimentation rate are much smaller than the variations in the pollen concentration. On the other hand, in high alpine environments with distinct altitudinal vegetation gradients such as on the mountains surrounding the Zoige Basin, the density and elevational limits of forests are primarily controlled by air tempera-ture. Higher temperature or a warmer climate would probably cause an expansion of tree populations. Thus the percentages of arboreal pollen in the RM core can be used as a proxy for temperature as well. 3 Lithology and chronology

The sediments of the RM core are mainly composed of silt, clay and fine sand. Several lithological units are recognized in the upper 60 m sediments. Between 60 and 43 m is an interval of dark-brown grayer silt that consists of 60% silt, 30% clay, and 10% of fine sand. Organic mat-ter is low (ca. 0.4%). A layer of fine sand occurs at 48—46 m within this interval. The unit at 43— 20.2 m is composed predominantly of dark gray fine sand that con-tains less than 50% silt and clay, with interbedded layers of silty sand and silt. Thin peat layers and organic-rich layers also occur occasionally in this unit. From 20 to 8.5 m, the sediments consist of gray silt (ca. 60%), fine sand (ca. 25%) and clay. Organic matter (0.5%—1.0%) is somewhat higher than the previous units. The top part of the core, 8.5—4.25 m, is a unit of fluvial sediment[25], dominated by fine sand accounting for 60%—100% of sediments.

Three bulk samples from the RM core were dated using conventional radiocarbon method. Despite their relatively large error bar, the three dates (21600±1500, 33140±2350, and >40000 a BP at 6.3, 8.24 and 12.45 m, respectively) are consistent and comparable with 14C dates obtained from the nearby RH core and from the regional

ARTICLES


chronology[11]. Paleomagnetic measurements were carried out on oriented samples from the entire 310 m core[12]. The Brunches/Ivlatuyama boundary (780 ka BP) had been found to occur at 85.72 m. A number of jerks are observed in the upper 60 m. However, the origin of these jerks is unclear because absolute age constraints are needed to determine whether they may represent an excursion oc-curring in the Brunches Epoch[12]. Since megnetostrati-graphy alone cannot provide an age scale for the upper 60 m of the RM core, we have constructed a depth-age model based on our pollen record by “tuning” the pollen-based

climate history with the widely accepted ice-core chro-nology of Guliya[10] and the GISP2[26].

As mentioned above, total pollen concentration val-ues can be used as a proxy for the regional precipitation in the Zoige Basin. The pollen concentration record of the RM core shows striking similarities to the Guliya δ 18O record, and also a climatic proxy for the southwestern (Indian) summer monsoon (Fig. 1). Although our pollen concentration record has a lower resolution, the peaks and valleys in the curve can be correlated with those of the Guliya δ 18O record. Therefore major climatic events in

Fig. 1. (a) The GISP2 δ 18O record over the last 180 ka (Dansgaard et al., 1993). (b) The Guliya b180 record over the last 110 ka (Thompson et al., 1998). (c) Pollen concentration curve of the RM core. (d) Pollen percentages of Compositae, Artemisia and Chenopo-diaceae, three main components of desert. (e) Arboreal pollen percentages of the RM core indicating temperature variations. Circle num-bers are age control points established by cross correlation between the ice-core δ 18O curves and the RM pollen stratigraphy (see text for details).

ARTICLES


both records can be identified. We have thus obtained seven age control points (numbered with 3—9) by corre-lating the RM pollen concentration with the Guliya δ 18O record.

According to the present vegetational zonation on the mountains around the Zoige Basin, forests occur at a lower elevation than meadows and periglacial desert. Per-centages of arboreal pollen (AP) thus indicate temperature changes. Pollen spectra with higher AP percentages gen-erally reflect higher temperature. The only exceptions are in those samples with very low pollen concentrations (suggesting very sparse vegetation cover) accompanied by relatively high percentages of Artemisia, Chenopodiaceae, and Compositae (the main components of desert); in these cases the arboreal pollen is probably derived from long-distance transport. The highest AP percentages in the whole core, which are supported by high pollen concen-trations, occur at 43.1—39.2 m. Most remarkably, these intervals contain prominent pollen peaks of Betula and Corylus-thermophilous hardwood taxa that occur princi-pally on south-facing slopes and at the lower parts of the alpine forest zone today. In the RM pollen diagram, the maxima of these pollen taxa are confined to zones RM15 and RM1, the latter represents the present interglacial (MIS 1). Therefore, we interpret pollen zone RM15 as indicating a warm interglacial period corresponding to part of the last interglacial (MIS 5e). Accordingly, the cold and dry period at 53.5—43.1 m, indicated by extremely low pollen concentrations, high pollen percentage of de-sert components and the preponderance of long-distance transported arboreal pollen must correspond to the penul-timate glacial maximum. Since there is greater uncertainty in the Guliya chronology beyond 112 ka[10], we correlated the lower parts of the RM pollen record with the GISP2 δ 18O record. Two additional age control points (numbered with 10 and 11) are recognized (Fig. 2). By using these nine age control points and the two 14C dates, we have established the time scale for the upper 60 m of the RM core (Fig. 2). Based on this time scale, the negative jerks in magnetic inclinations[12] can be reasona-bly assigned to the excursions reported in the Brunhes Epoch (Fig. 3). The ages of these excursions are very close to those estimated by our depth-age model. While this procedure does not ensure the complete independence of our chronology, the fluctuation of climate shown by the pollen record is independent even though the time scale is tuned from the isotopic records of ice cores. 4 Pollen records and interpretations

The results of pollen analysis for the upper 60 m of the RM core are shown in Fig. 4. The pollen diagram is divided into 18 pollen zones by visual inspection on the basis of changes in pollen percentages and pollen concen-trations.

Zone RM18 (60.0—53.5 m; ca. 190000—169000 a

Fig. 2. Depth-age model of RM core based on the 14C dates and age control points derived from the correlation between the RM pollen re-cord and isotopic records of ice cores. BP). This zone shows consistently high percentages of Cyperaceae pollen (40%—60%), relatively high percent-ages of Gramineae pollen (8%—18%) and high pollen concentration values. Herbaceous taxa like Artemisia, Chenopodiaceae, Compositae, Ranunculaceae, Caryo-phyllaceae and Thalictrum are relatively common. Per-centages of arboreal pollen including Pinus, Picea, Abies. Betula, and Quercus are less than 20%. The pollen assem-blages of this zone represent the subalpine sedge meadow, widely found now on terraces within the basin and on gentle slopes of the surrounding mountains at 3400—3800 m a.s.l., indicating a climatic condition colder and wetter than that of the present. The depth-age curve shows that the sedimentation rate did not change greatly, especially at the lower part of the core (60—30 m). Relatively high sedimentation rate from 30 to 8.3 m may reflect hydrological changes caused by the Yellow River incision through the basin[27]. The sedimentation rate decreased significantly after 33000 a BP, corresponding to a change from lacustrine to fluvial environment as the ancient lake disappeared. Zone RM17 (53.5—50.0 m; ca. 169000—157000 a BP). A marked drop in pollen concentration, a decrease in Cyperaceae pollen, and a gradual increase in arboreal pollen and Chenopodiaceae pollen characterize zone RM17. This zone shows a much more open and sparse vegetation cover than Zone RM18, indicating a colder and drier period of alpine meadow development.

ARTICLES


Fig. 3. Magnetic inclination versus depth for the upper 60 m of the RM core and inferred excursions in the Brunhes Epoch. Age of excursions is referred to Beleil et a1. (1987), Vidar Valan et al. (1995), Nowaczyk and Antonow (1997), and Nowaczyk and Jochen (2000). Ages of the eleven control points (circle numbers) are marked by arrows on the left side of the magneto-stratigraphic column.

Zone RM16 (50.0—43.1 m; ca. 157000—134000 a

BP). Pollen concentration is very low (60 grains/g), even absent in two levels, indicating an extremely open land-scape with sparse vegetation. This zone is dominated by arboreal pollen, especially Pinus pollen (>20%), which is probably attributable to long-distance transport. Herba-ceous taxa, especially Chenopodiaceae, Cyperaceae, Gramineae, Compositae and Artemisia, are relatively common. If the arboreal pollen is excluded from the pol-len sum, percentages of Chenopodiaceae, Artemisia and Compositae will be close to 50%. Alpine periglacial desert probably existed under very harsh periglacial conditions during this period.

Zone RM15 (43.1—39.4 m; ca. 134000—120000 a BP). This zone, corresponding to the last interglacial, is characterized by the highest percentages of arboreal pollen

(40%—90%) in the core, accompanied by high pollen concentrations. Picea and Abies pollen occurs in promi-nent peaks in this zone. Pinus, Betula and Quercus pollen are relatively abundant in the lower part, and Corylus pol-len also attains maximum abundance in the upper part of this zone. The herbaceous taxa are dominated by Cyper-aceae and Gramineae. This zone clearly represents a major expansion of subalpine spruce-fir forest, which was also mixed with abundant themophilous hardwoods such as birch, oak, and hazel, under a climate warmer and wetter than that of today.

Zone RM14 (39.4—37.6 m; ca. 120000—111000 a BP). This zone is characterized by a decrease in arboreal pollen and a marked increase in Cyperaceae pollen. Pollen concentrations are high at the lower part of this zone but decrease non-linearly upward. In this period, the studied region was dominated by subalpine sedge meadow under a climatic condition with high precipitation but lower temperature.

Zone RM13 (37.6—33.5 m; ca. 111000—86000 a BP). Arboreal pollen, such as Picea, Abies, Pinus, Betula and Quercus, is much less than in zone RM15, al-though percentages of Picea in some levels reach more than 20%. Cyperaceae pollen is prominent, accounting for 32.7%—67.9% of total pollen. Pollen of Gramineae and Ranunculaceae increases slightly. Pollen concentrations vary a lot among different levels, with average of 580 grains/g. This zone suggests that subalpine sedge meadow and spruce-fir forest were the major vegetation types in this period and they expanded and contracted alternately as the climatic condition fluctuated.

Zone RM12 (33.5—30.5 m; ca. 86000—71000 a BP). In Zone RM12, pollen of Picea increases to form another strong peak with a maximum of 36.7%. Pollen of Abies is more abundant than in Zone RM 13. Cyperaceae remains prominent, and Gramineae and Ranunculaceae are still common. Pollen concentrations are high, reaching 860 grains/g on average. This zone represents another period of major spruce-fir forest expansion under warm and humid climatic conditions.

Zone RM11 (30.5—28.5 m; ca. 71000—60000 a BP). Apparent at first in this zone is a sharp drop in pol-len concentrations. Reduction and ultimate elimination of Picea and Abies are accompanied by the dominance of Cyperaceae. The frequencies of Compositae and Gramineae increase distinctly. This zone marks the disap-pearance of forest and its replacement by alpine meadow. Temperature and rainfall decreased significantly during this period compared with the last.

Zone RM10 (28.5—27.0 m; ca. 60000—57000 a BP). This zone shows an increase in pollen concentra-tions and the frequencies of arboreal pollen, and a decline in the frequencies of herbaceous pollen like Cyperaceae, Grarnineae and Compositae. The pollen spectra indicate

ARTICLES


that the landscape consisted of mosaics of subalpine spruce-fir forest and sedge meadow under ameliorated climatic conditions.

Zone RM9 (27.0—25.2 m; ca 57000—54000 a BP). The assemblage is reminiscent of Zone RM11 as reflected by very low frequencies of Picea and Abies. Pollen concentration (600 grains/g), nevertheless, is much higher than in Zone RM 11. Sedge meadow expanded and occupied most areas where spruce-fir forest had grown during the preceding period. The climatic condition probably was close to that in the period of pollen Zone RM11 although the precipitation was probably somewhat higher.

Zone RM8 (25.2—22.5 m; ca. 54000—49000 a BP). Small peaks of Abies, Picea, and Pinus occur, with a rela-tively high pollen concentration (370 grains/g). Pollen percentages of Cyperaceae remain high in this zone. The pollen assemblages of this zone represent a retreat of subalpine sedge meadow and an expansion of spruce-fir forest under climatic conditions similar to those of today.

Zone RM7 (22.5—20.2 m; ca. 49000—46000 a BP). Pollen percentages of Abies, Picea, and Pinus are lower than in Zone RM8, but the frequencies of Cramineae, Compositae and Cyperaceae increase. Pollen concentration is low, less than 100 grains/g for most sam-ples. The landscape may have been a mosaic of forest and sedge-grass meadow, probably indicating a cooler and drier climate relative to the above zone.

Zone RM6 (20.2—16.5 m; ca. 46000—43000 a BP). This zone is most distinctly characterized by the dominance of Cyperaceae, less arboreal pollen, and a rise in pollen concentration. Gramineae, Compositae and Ranunculaceae are common herbaceous taxa. The pollen assemblages of this zone show an expansion of alpine sedge meadow in regional vegetation.

Zone RM5 (16.5—12.0 m; ca. 43000—39000 a BP). Zone RM5 is marked by a sharp drop in the pollen percentages of Cyperaceae to less than 10%. The assem-blages are dominated by Picea, together with Pinus, Abies, Quercus, Gramineae, and Ranunculaceae. Pollen percent-ages of Gramineae steadily increase to 28.2%, and Gramineae becomes the dominant herbaceous taxon re-placing Cyperaceae. Pollen concentrations, however, fall to an average of 150 grains/g and less than 100 grains/g at the upper part and birch forest may have existed on the south-facing slopes. Grass-sedge meadow grew on the open land. These data suggest a sharp rise in temperature and rainfall.

Zone RMI (5.1—4.25 m; ca. 15000—7000 a BP). This zone is distinguished from Zone RM2 by higher pol-len concentrations, and higher percentages of arboreal pollen (45.4%—73.5%). The pollen percentages of Picea and Abies increase slightly, and the pollen frequencies of broadleaved hardwood trees and shrubs such as Betula,

Quercus, Corylus, Hippophae, and Rosaceae reach maxi-mum values in this zone. The pollen spectra show a major expansion of forests and a retreat of alpine meadows, re-flecting high temperature and precipitation levels equal to or exceeding those of today. 5 Discussion

(i) Vegetation and climate in the penultimate and last glacial maxima. During the penultimate and last glacial maxima (PGM and LGM), the region of the Zoige Basin was occupied by desert and/or periglacial desert. The cli-matic condition during the early part of the PGM (pollen zone RM 17) was similar to that in the LGM (pollen zone RM3), but more harsh during the later part of PGM (pol-len zone RM16) than that in LGM. All three zones (RM 17, 16, 3) are characterized by very low pollen concentra-tions and high pollen percentages of desert components. However, high percentages of arboreal pollen (over 50%), especially Pinus, in zone RM 16, coupled with very low total pollen concentrations, suggest that most of the arbo-real pollen are derived from long-distance transport, im-plying that the regional vegetation was much more sparse in this period than in the other two phases. This conclu-sion is consistent with glacial geologic data. The area of ice cover during the penultimate glaciation was found to be 45% larger than that during the last glaciation in the Goumshan Mountain 150 km west of the Zoige Basin[28].

Questions concerning the timing, extent, and charac-ter of Quaternary glaciation in the Qinghai-Tibetan Pla-teau are of vital importance to the global climate recon-struction and modeling[29,30, 31]. Kuhle[32,33] postulated that the Qinghai-Tibetan Plateau was covered by a uniform ice sheet during the last (Wisconsin-Würm) glacial maximum. His postulated existence of a large-scale Tibetan ice sheet has been criticized and challenged by geomorphological, geological, and glaciological observations[27,34,35]. Our pollen evidence from the RM core shows more sparse vegetation and colder and drier climate during the penultimate glaci-ation than during the last glaciation. Our data thus support the contention by some Chinese scientists[35—37] that glaci-ation on the Tibetan Plateau (including even the possibil-ity of a small ice sheet near the source area of the Yellow River) was more extensive during the penultimate glacia-tion than during the last glaciation.

(ii) Long interglacial-glacial transition. The Zoige Basin experienced the warmest climate condition during RM 15 (corresponding to the last interglacial) when spruce-fir forest widely grew along the shores of the an-cient lake and in the surrounding mountains. Mixed coni-fer-hardwood forest or even broad-leaved hardwood for-ests consisting of Betula and Corylus may have occurred in the early and late stages of this period. Both tempera-ture and precipitation were probably higher than those of today. The pollen data also show a relatively stable cli-mate within the last interglacial, similar to the reconstruc-

ARTICLES


tion based on Chinese loess records[11]. During RM 14—13 (120—86 ka), spruce-fir forest

and sedge meadow alternately expanded and contracted, suggesting frequent fluctuations of climate. The tempera-ture was lower in this period than in the last interglacial, with a distinct cold phase between 120 and 111 ka (RM 14). However, the precipitation during 120—111 ka was higher than that in other phases. Two low pollen concen-tration valleys located around 108 and 92 ka probably indicate two weaker summer monsoon phases during this period. High percentages of arboreal pollen and high pol-len concentration in RM 12 (ca. 86—71 ka) suggest that the climate during this period was nearly as warm and humid as the last interglacial. During this period, spruce-fir forest dominated this region. The pattern of climatic change during 120—71 ka reflected in our pollen record is consistent with the climatic stability documented from the proxy records of the Guliya ice core[10], Chinese

loess[11] and the GISP2 ice core[26], suggesting that our lacustrine pollen data may provide a record of global cli-mate teleconnection.

Low temperature and low precipitation existed from 71 to 60 ka (pollen zone RM 11). Alpine sedge meadow dominated the landscape during this period. The period from 60 to 32 ka (pollen zones RM 10—4) is punctuated by a sequence of stadial and interstadial events. The pre-cipitation has shown a nonlinear decrease since 60 ka suggested by a decline in pollen concentration, as well as a gradual increase in the pollen percentages of Chenopo-diaceae, Artemisia, and Compositae (Fig. 4).

It is also worth noting in our pollen record that sev-eral short cold events with very low AP percentages oc-curred during the last 80 ka. The timing of these events seems close to that of the Heinrich events. These events were also recorded in the grain-size time series from the Loess Plateau[38,39,41], implying that the Heinrich events

Fig. 4. Pollen percentage diagram for the upper 60 m of the RM core.

ARTICLES


Fig. 5. Comparison of paleoclimatic proxy records. (a) The Guliya δ 18O record over the last 112 ka (Thompson et al., 1998), (b) pollen concentration curve of the RM core, (c) monsoon stack from the Arabian Sea (Clemens, this issue), (d) arboreal pollen percentages of the RM core, (e) the GISP2 δ 18O record over 180 ka (Dansgaard et al., 1993). Dotted lines show the eleven control points.

may have left their signature not only in the Chinese loess record, but also in the pollen record of the Tibetan Plateau. Correlation with δ 18O record of ice cores and marine and terrestrial monsoon proxy records from the Arabian Sea and adjacent lands have shown that the Qinghai-Tibetan Plateau acts as a vital linkage between radiation forcing, glacial boundary conditions in the North Atlantic region, and strength of the southwestern Asian monsoon[39—41]. Overpeck et al.[41] hypothesized that during deglacial times, abrupt warming in the North Atlantic may have led to increased wanning in the Qinghai-Tibetan Plateau, thereby causing a reduced snow cover in the spring and summer. The lower albedo led to an intensification of the low pressure system over the Qinghai-Tibetan Plateau and thus a steepened land-sea pressure gradient, thus resulting in a stronger summer monsoon. Our glacial-interglacial pollen record from the Zoige Basin provides the vital data to test this hypothesis.

In Fig. 5 we plot the key pollen curves from the RM core alongside the oxygen-isotope curves from the Guliya ice core[10], the GISP2 ice core[26], and the summer mon-soon stack from the Arabian Sea (Clemens, this issue).

Comparison between the RM pollen concentration curve, which mainly reflects precipitation changes in the Zoige Basin, and the Guliya δ 18O record, which is a climatic proxy for the southwestern (Indian) summer monsoon, reveals striking similarities in the first order, especially for the interval between 110 and 60 ka. Notably, prominent peaks associated with MIS 5e, 5c and 5a, as well as sharp declines making their terminations are evident in both the RM pollen and Guliya δ 18O curves. The minima within MIS 5d, 5b, and MIS 4, as well as the frequent fluctua-tions within MIS 3 are equally distinct in both curves (Fig. 5). The agreements between the two curves are not unex-pected since our age model is tuned from the Guliya δ 18O record. However, the similarities in details, e.g. variations of MIS 5b and MIS 4, seem to further support our age model. The parallelism between the two records becomes less distinctive after ca. 60 ka, mainly because there is a long-term decline in pollen concentrations in the RM core and the amplitude of variations has become much lower. The loss of sensitivity in the pollen concentration curves is due to a significant increase in sedimentation rate in the Zoige Basin caused by the increasing fluvial influence of

ARTICLES


the Yellow River. Nevertheless, even in the uppermost 8.5 m of the RM core, where the sediment is alluvial sand, the pollen minimum in MIS 2 and its subsequent rise in MIS 1 are still broadly parallel to the trend exhibited in the Guliya record. The similarity between the RM pollen re-cord and the Guliya δ 18O record implies that the climate changes in the Zoige region primarily reflect the influence of the southwestern summer monsoon. Moreover, simi-larities between the RM pollen concentrations and the monsoon stack in the Arabian Sea also exist (Fig. 5), which seems to further support this conclusion. Discrep-ancies in the timing and amplitudes between the two curves do exist, for example, the 5a peak seems to occur earlier in the RM pollen record than in the Arabian Sea. However, these discrepancies do not seem to obscure the general trends of the major variations. In addition, the generally good match between the AP percentage curve of the RM core and the monsoon stack curve of the Arabian Sea suggests a link between Tibetan temperature and summer monsoon strength. High summer temperatures and the resulting strong low pressure system in the Tibetan Plateau are the main factors maintaining a strong summer monsoon in the Arabian Sea, as demonstrated by modem meteorological data[42].

Striking similarities are also evident between the RM arboreal pollen curve and the δ 18O curve of the GISP2 ice core[26]. Prior to MIS 5e, the two records cannot be di-rectly compared because the AP percentage at RM is dis-torted by exotic pollen input due to extremely low local pollen production. However, after ca. 130 ka, both curves exhibit prominent peaks during MIS 5e, followed by os-cillating declines during MIS 5d through 5a, major mil-lennial-scale fluctuations during MIS 4 and 3, low values during MIS 2, and a resurgence in MIS 1 (Fig. 5). Re-markably, millennial-scale climate instability reflected as the Dansggard-Oscheger cycle in the GISP2 core during the last 130 ka also seems to be registered in the pollen record of the RM core, even though its timing appears to be somewhat offset. The general parallelism between the RM AP percentage curve and the GISP2 δ18O record sug-gests that climate changes in the eastern Qinghai-Tibetan Plateau and the North Atlantic region were linked, as Overpeck et al.[41] have postulated. 6 Conclusions

(i) Pollen record from the upper 60 m of the RM core, the longest continuous pollen record from the Qing-hai-Tibetan Plateau, provides evidence of vegetation and climate changes during the last two glacial/interglacial cycles. Subalpine spruce-fir forests were present in the Zoige Basin during interglacial and interstadial times, suggesting that relatively warm and wet climatic condi-tions prevailed in these periods. Alpine periglacial desert and/or dry desert existed under cold and dry climatic con-ditions during the glacial maxima.

(ii) Major climatic trends and events during the long transition between the last interglacial and the last glacia-tion inferred from our pollen record are consistent with those recorded in the Chinese Loess Plateau, the Greenland and Guliya ice cores, and marine sediments in the Arabian Sea and the North Atlantic, suggesting that the Qinghai-Tibetan Plateau acts as an important link between the climatic variations in the North Atlantic region and the changing strengths of the southwestern Asian monsoon.

(iii) We obtain nine dates by comparing pollen record from the upper 60 m of the RM core with Guliya and GISP2 ice-core δ 18O records, and then construct a depth-age model. According to the time scale, the Brunches/Matuyama boundary can be ascertained and Brunches Epoch can be demarcated at 46 m of RM core. Acknowledgements The authors thank Steve Clemens for editorial assistance and excellent suggestions to improve their age model, and are grateful to Shouyun Hu for providing paleomagnetic data and Houyuan Lu for helpful discussion. Their gratitude is also extended to Dr. Shi-Yong Yu for improving the manuscript. This work was supported by 973 Project of China (Grant No. 998040810), the National Key Project for the Qinghai-Tibetan Plateau in the Eighth Five-year Project of China, the National Natural Science Foundation of China (Grant Nos. 49371068 and 49871078), and the U.S. National Science Foundation.

References 1. Thompson, L. G., Mosley-Thompson, E. M., Davis, M. E. et al.,

Glacial stage ice core records from the subtropical Dunde ice cap, China, Annals of Glaciology 1990, 14: 288—297.

2. Gasse, F., Arnold, M., Fontes, J. C. et al., A 13000 year climate re-cord from western Tibet, Nature 1991, 353: 742—745.

3. Gasse, F., Fontes, J. C., Van Campo, E. et al., Holocene environ-mental changes in Bangong Co Basin (western Tibet), Part 4. Dis-cussion and conclusions, Palaeogeography, Palaeoclimatology, Pa-laeoecology, 1996, 120: 79—92.

4. Van Campo, E., Gasse, F., Pollen- and diatom-inferred climatic and hydrological changes in Sumxi Co Basin (Western Tibet) since 13000 yr B.P., Qnatemary Research, 1993, 33: 303—313.

5. Fontes, J. C., Gasse, F., Gibert, E., Holocene environmental changes in Lake Bangong Basin (western Tibet), Part 1. Chronol-ogy and stable isotopes of carbonates of a Holocene lacustrine core, Palaeogeography, Palaeoclimatology, Palaeoecology, 1996, 120: 25—47.

6. Jarvis, D. E., Pollen evidence of changing Holocene monsoon cli-mate in Sichuan Province, China, Quaternary Research, 1993, 33: 325—337.

7. Liu, K. B., Yao, Z., Thompson, L. G., A pollen record of Holocene climatic changes from the Dunde ice cape, Qinghai-Tibetan Plateau, Geology, 1998, 26: 135—138.

8. Yan, G, Wang, F. B., Shi, G. R., Palynological and stable isotopic study of paleoenvironmental changes on the northeastern Tibetan Plateau in the last 30000 years, Paleogeography, Paleoclimatology, Paleoecology, 1999, 153: 147—159.

9. Lehmkuhl, F., Haselein, F., Quaternary paleoenvironmental change

ARTICLES


on the Tibetan Plateau and adjacent areas (Western China and Western Mongolia), Quaternary International, 2000, 65/66: 121—145.

10. Thompson, L. G., Yao, T., Davis, M. E. et al., Tropical climate in-stability: the last glacial cycle from a Qinghai-Tibetan ice core, Science, 1998, 276: 1821—1825.

11. Chen, F. H., Bloemendal, J., Zhang, P. Z. et al., An 800 ka proxy record of climate from lake sediments of the Zoige Basin, eastern Tibetan Plateau, Paleogeography, Paleoclimatology, Paleoecology, 1999, 151: 307—320.

12. Hu, S., Appel, E., Wang. S. et al., A preliminary magnetic study on lacustrine sediments from Zoige basin, eastern Tibetan Plateau, China: magnetostratigraphy and environmental implications, Phys-ics and Chemistry of the Earth (A), 1999, 24: 811—816.

13. Yie, D. Z., Gao, Y. X., The Meteorology of the Qinghai-Xizang Plateau, Beijing: Meteorology Press, 1979.

14. Wang, S. M., Xue, B., Environmental evolution of the Zoige Basin since 900 ka BP and comparison study with Loess Plateau, Science in China, Ser. D, 1997, 40(3): 329—336.

15. Wang, Y. F., Wang, S. M., Xia, W. L., Mineralogy and Paleoclimate Variety of the Sediments from the Zoige Basin in the Last 150 Thousand Years, Study on the Formation, Evolution, Environ-mental Changes and Ecosystem of the Qinghai-Tibetan Plateau, Beijing: Science Press, 1996, 118—122.

16. Editorial Board of Sichuan’s Vegetation, The Vegetation of Sichuan (in Chinese), Chengdu: Sichuan People’s Press, 1980.

17. Institute of Expedition and Planning Ministry of Forestry, Montane Forest in China, Beijing: Chinese Forestry Industry Press, 1979.

18. Jørgensen, S., A method of absolute pollen counting, New Phytolo-gist, 1967, 66: 489—493.

19. Moore, P. H., Webb, J. A., Collinson, M. E., Pollen Analysis (2nd ed.), London: Blackwell Scientific Publications, 1991.

20. Grimm, E. C., Chronology and dynamics of vegetation change in woodland region of southern Minnesota, U.S.A., New Phytologist, 1983, 93: 311—350.

21. Bureau of Geology and Mineral Resources of Sichuan Province, Areal Geology Records of Sichuan Province, Beijing, 1991.

22. Peck, R. M., Pollen budget studies in a small Yorkshire catchment, in Quaternary Plant Ecology (eds. Birks, H. J. B., West, R. G.), Oxford: Blackwell, 1973.

23. Bonny, A. P., Recruitment of pollen to the seston and sediment of some English Lake District lakes, Journal of Ecology, 1976, 64: 859—887.

24. Prentice, I. C., Pollen representation, source area, and basin size: toward a Unified theory of pollen analysis, Quaternary Research, 1985, 23: 76—86.

25. Wang, Y. F., Wang, S. M., Xue, B. et al., Sedimentological evidence of the piracy of ancient Zoige lake by the Yellow River, Chinese Science Bulletin, 1995, 40: 1539—1544.

26. Dansgaard, W., Johnsen, S. J., Clausen, H. B. et al., Evidence for

general instability of past climate from a 250 kyr ice-core record, Nature, 1993, 364: 218—220.

27. Lehmkuhl, F., Extent and spatial distribution of Pleistocene glaci-ations in Eastern Tibet, Quaternary International, 1998, 45/46: 123—134.

28. Shi, Y., Li, J., Li, B., Uplift and Environmental Changes of Qing-hai-Xizang (Tibetan) Plateau in the Late Cenozoic, Guangzhou: Guangdong Science and Technology Press, 1998.

29. Manabe, S., Broccoli, A. J., The influence of continental ice sheets on the climate of an ice age, Journal of Geophysical Research, 1985, 90: 2167—2190.

30. COHMAP members, Climatic changes of the last 18000 years: Observations and model simulations, Science, 1988, 241: 1043—1052.

31. Bush, A. B. G., A positive climatic feedback mechanism for Hima-layan glaciation, Quaternary International, 2000, 65/66: 3—13.

32. Kuhle, M., Geomorphological findings on the build-up of Pleisto-cene Glaciation in Southern Tibet and the problem of Inland Ice, Geo. Jour., 1988, 17(4): 457—512.

33. Kuhle, M., Reconstruction of the 2.4 million km2 late Pleistocene ice sheet on the Tibetan Plateau and its impact on the global climate, Quaternary International, 1998, 45-46: 71—108.

34. Derbyshire, E., Shi, Y., Li, J. et al., Quaternary glaciation of Tibet: the geological evidence, Quaternary Science Reviews, 1991, 10: 485—510.

35. Shi, Y., Zheng, B., Li, S., Last glaciation and maximum glaciation in Qinghai-Xizang Plateau, Journal of Glaciology and Geocryology, 1990, 12: 1—15.

36. Lehmkuhl, F., Extent and spatial distribution of Pleistocene glaci-ations in Eastern Tibet, Quaternary International 1998, 45/46: 123—134.

37. Lü, H. Y., Wang, S. M., Wu, N. Q. et al., A new pollen record of the last 2.8 Ma from the Co Ngoin, central Tibetan Plateau, Science in China, Ser. D, 2001, (Supp. 1): 292—300.

38. Porter, S. C., An, Z., Correlation between climate events in the North Atlantic and China during the last glaciation, Nature, 1995, 375: 305—308.

39. Clemens, S. C., Prell, W., Murray, D. et al., Forcing mechanisms of the Indian Ocean monsoon, Nature, 1991, 353: 720—725.

40. Sirocko, F., Samthein, M., Erlenkeuser, H. et al., Century-scale events in monsoonal climate over the past 24000 years, Nature, 1993, 364: 322—324.

41. Overpeck, J., Anderson, D., Trumbore, S. et al., The southwest In-dian Monsoon over the last 18000 years, Climate Dynamics, 1996, 12: 213—225.

42. Webster, P. J., Magana, V. O., Palmer, T. N. et al., Monsoons: proc-esses, predictability, and the prospects for prediction, Journal of Geophysical Research, 1998, 103: 14451—4510.

(Received August 4, 2004; accepted November 18, 2004)

articles - lsuchinese science bulletin vol. 50 no. 6 march 2005 553 chinese science bulletin 2005...

Documents