Evolution of vegetation and climate variability on the Tibetan Plateau over the past 1.74 million years



Abstract

The Tibetan Plateau exerts a major influence on Asian climate, but its long-term environmental history remains largely unknown. We present a detailed record of vegetation and climate changes over the past 1.74 million years in a lake sediment core from the Zoige Basin, eastern Tibetan Plateau. Results show three intervals with different orbital- and millennial-scale features superimposed on a stepwise long-term cooling trend. The interval of 1.74–1.54 million years ago is characterized by an insolation-dominated mode with strong ~20,000-year cyclicity and quasi-absent millennial-scale signal. The interval of 1.54–0.62 million years ago represents a transitional insolation-ice mode marked by ~20,000- and ~40,000-year cycles, with superimposed millennial-scale oscillations. The past 620,000 years are characterized by an ice-driven mode with 100,000-year cyclicity and less frequent millennial-scale variability. A pronounced transition occurred 620,000 years ago, as glacial cycles intensified. These new findings reveal how the interaction of low-latitude insolation and high-latitude ice-volume forcing shaped the evolution of the Tibetan Plateau climate.


INTRODUCTION

The Tibetan Plateau has long been a focus of geoscientific studies because of its importance in global tectonics and Asian and global climate change across a wide range of time scales (1). However, with only few available paleoarchives of coarse resolution [>8 thousand years (ka)] (2, 3), little is known about its environmental history through the Quaternary ice ages. To understand the mode and tempo of changes and, ultimately, the underlying drivers during this period, we need long-term high-resolution records from the elevated plateau with well-constrained chronologies.

The Zoige Basin, occupied by a huge lake until the latest Pleistocene (3) and located on the eastern Tibetan Plateau within the South Asian monsoon zone (fig. S1), represents a potential site to fill this gap. Mean annual precipitation (MAP) at Zoige is ~600 to 650 mm, and the basin is primarily covered by alpine meadows, and the surrounding mountains have scattered forests up to ~4000 m above sea level (a.s.l.) (fig. S1). A sediment core extending over the past 0.9 million years (Ma) was previously recovered, but its analytical resolution and chronological reliability were insufficient to resolve orbital- and suborbital-scale changes (3).

CORE ACQUISITION AND CHRONOLOGY

New drilling (33°58.163′N, 102°19.855′E, 3434 m a.s.l.) was undertaken in 2013 in the central basin guided by a seismic survey. A 573.39-m core (ZB13-C2) was obtained with 96% recovery, mostly consisting of fine-grained freshwater lacustrine sediments. Only the upper 50 m contains two episodic fluvial sandy layers, 10.11 and 10.4 m thick, respectively.

Independent age control derived from magnetostratigraphy in combination with radiocarbon [accelerator mass spectrometry (AMS)] and luminescence [optically stimulated luminescence (OSL)] dating provides an initial chronological framework (Materials and Methods; figs. S2 and S3), according to which the ZB13-C2 core extends back to 1.74 Ma before present (BP) (Materials and Methods; fig. S3). Fluctuations in arboreal pollen abundances (AP%) based on an initial age model using a combination of 14C, OSL, and paleomagnetic control points (table S3) show clear ~100-, 40-, and 20-ka cyclicities (fig. S3), suggesting possible eccentricity (E), tilt (T), and precession (P) powers. The presence of astronomical frequencies in the Zoige Basin record is further supported by spectral analyses of AP% in the depth domain, which indicate the occurrence of ~34-, 15-, and 7.5-m cycles, whose ratios are close to those of 100:40:21; the ~7.5-m cycle appears to be stronger in the lower ~75 m, while the ~34-m cycle is stronger in the top ~200 m (Materials and Methods and fig. S3). On this basis, a more detailed age model was constructed by tuning the AP% record to an ETP record that is generated by normalizing and averaging variations in eccentricity, tilt, and reversed precession (4). As this approach may artificially introduce astronomical frequencies in our record, we compared the ETP age model against an age model constructed by aligning the Zoige AP% to the Chinese speleothem δ18Ocalcite record, an independently dated (U-Th) archive of changes in Asian monsoon intensity over the past 640 ka (5). Comparison of the ETP- and speleothem-based age models of Zoige reveals a close correspondence (fig. S3).

CLIMATE PROXIES

Pollen and sediment analyses (2787 and 3274 samples, respectively, with a mean sampling resolution of ~530 to 620 years), as well as x-ray fluorescence (XRF) scanning (~6-year resolution) data, yield a detailed multiproxy record back to 1.74 Ma BP.

The vegetation in the eastern Tibetan Plateau is strongly influenced by the Asian summer monsoon. A stronger monsoon with warmer and moister climate would cause an expansion of tree populations. However, in the alpine Zoige region, the density and elevational limits of forests are primarily controlled by temperature, as moisture availability is relatively plentiful (6, 7). This is confirmed by the distinct elevational distribution of modern vegetation (fig. S1), the good match of AP% and axis 1 of principal components analysis (PCA) on the pollen data (fig. S4), and the close relationship between summer temperature and PCA axis 1 in the significance test of the quantitative temperature reconstructions for core ZB13-C2 (fig. S4). Thus, on long time scales, variations in AP% in the core are mostly a reflection of changes in temperature, particularly summer temperature, but drought stress, for example, during weak monsoon or glacial intervals, would also have an impact on tree populations.

Pollen-based quantitative reconstruction of past climate variables was undertaken (Materials and Methods), providing the first independent paleotemperature history for the Tibetan Plateau because of the dominant control of temperature on local vegetation (figs. S1 and S4) (6, 7). The mean temperature of the coldest month (MTCM) and precipitation reconstructions failed the relevant statistical significance test and are thus less reliable (Materials and Methods).

Rubidium/strontium (Rb/Sr) ratios and carbonate content (Carb%) are used as supplementary proxies. Rb/Sr primarily reflects the chemical weathering intensity of the catchment or strength of summer monsoon and associated run-off (8). When weathering/run-off is stronger, there is greater Sr input into the lake, leading to a lower Rb/Sr ratio. However, Sr contained in carbonate can influence Rb/Sr ratios from bulk scanning data, rendering climate interpretations less reliable. We therefore measured Rb/Sr ratios on bulk samples after removing the carbonate content (Materials and Methods). This showed good agreement with the high-resolution XRF scanning data (fig. S5), supporting the view that the ZB13-C2 Rb/Sr signal derived from XRF scanning is largely independent of changes in carbonate content. Grain size changes could also distort the climate signal of Rb/Sr, but examination of core ZB13-C2 shows a weak correlation between Rb/Sr and grain size changes (fig. S5). The Zoige lake sediments are generally fine and have no large variations, except the two sandy layers near the top. The high correlation between Rb/Sr and the chemical index of alteration (CIA) at Zoige further support the weathering interpretation (fig. S5) that the proxies are sensitive to both summer precipitation and temperature conditions.

Previous studies from core RH nearby ZB13-C2 suggest that carbonate content mainly represents authigenic chemical precipitation, as both detrital carbonate content and shell carbonate are in trace amounts (3). The measured Carb% in core ZB-C2 should mainly reflect processes of chemical precipitation in the lake, which are largely related to summer temperature and precipitation. High temperature could increase the precipitation of carbonate through changing the precipitation-dissolution equilibrium and photosynthesis process. Precipitation could also enhance carbonate content by washing more Ca2+ and HCO3− into the lake through chemical weathering. Carb% therefore indicates warm and wet climate. Carbonate content from core RH shows good positive correspondence with hydrogen index, a proxy of the effect of lake water depth (3). Our loss-on-ignition (LOI) measurements of surface mud samples, taken along water-depth transects from four lakes near ZB13-C2, also reveal that Carb% generally increases with water depth in each lake (fig. S5). High Carb% from core ZB13-C2 likely agrees with high lake level, which depends on the balance between precipitation and evaporation.

The coherent variations in AP%, Rb/Sr, and Carb% (fig. S5), with low Rb/Sr and high Carb% corresponding to high AP%, therefore suggest that coupled changes of temperature and precipitation occurred over the past 1.74 Ma.

READ OR DOWNLOAD HERE Evolution of vegetation and climate variability on the Tibetan Plateau over the past 1.74 million years | Science Advances (sciencemag.org)

View ORCID ProfileYan Zhao1,2,*,
View ORCID ProfilePolychronis C. Tzedakis3,
View ORCID ProfileQuan Li1,
Feng Qin1,
View ORCID ProfileQiaoyu Cui1,
Chen Liang1,
View ORCID ProfileH. John B. Birks3,4,
Yaoliang Liu1,
Zhiyong Zhang1,5,
View ORCID ProfileJunyi Ge6,7,
Hui Zhao8,
View ORCID ProfileVivian A. Felde4,
View ORCID ProfileChenglong Deng9,
Maotang Cai1,
Huan Li10,
Weihe Ren1,
Haicheng Wei11,
View ORCID ProfileHanfei Yang1,
Jiawu Zhang12,
View ORCID ProfileZicheng Yu13,14 and
View ORCID ProfileZhengtang Guo2,9

See all authors and affiliations
Science Advances 06 May 2020:
Vol. 6, no. 19, eaay6193
DOI: 10.1126/sciadv.aay6193

Article
Figures & Data
Info & Metrics
eLetters
PDF