Late Holocene Environment Change Based On High-resolution Pollen Records From Varved Lake Sediments In West China

Knowledge of climate change during the last two millennium can both improve our understanding of natural climate variability and help address the question of whether modern climate change is unprecedented in a long-term context, which also be of great significance in predicating future climate change. Well-dated, high-resolution records from different key regions in the world are crucial to reconstruct the climate change during the last two thousand years. A great number of fine-resolution paleoclimate studies from China using tree rings, ice cores, speleothem as well as historical documents have revealed significant climate variability over the last two millenniums. However, most of these high quality proxy climate records (such as tree rings and ice cores) have been collected at high elevations, which may show different climate respond to large-scale climate forcing compared to low elevations due to the topographic effects; moreover, historical document based reconstructions mainly came from East China and few studies are related to West China. Therefore, much more work is still needed in order to identify and understand the main patterns and modes of climate variability, their teleconection as well as driven and feedback mechanism in the past 2000 years. The northeastern Tibetan Plateau and western Loess Plateau are both located at the monsoonal boundary areas, so they are very sensitive to climate and environment changes. Unfortunately, high-resolution and well-dated climate records covering the last 2000 years in these two regions are still scarce.Annually laminated sediments are the ideal deposits for fine-resolution and multi-proxy studies, so they play a key role in last 2000 years environment reconstruction together with tree ring, ice core and speleothem. We selected the varved Sugan Lake (Qadam basin) and Tianchi Lake (West Loess Plateau) to reveal the environmental changes during the last 2000 years. The Sugan lake drainage is located at the remote areas with little human disturbance during the historical time, so the climate variation might be the main controller of the environment change. We conducted a 2700-year high resolution (ca.10-yr) pollen record extracted from Sugan Lake in order to reconstruct the regional vegetation change and climate variability; Moreover, multi-proxy analysis method were used together with the comparison with regional climate records to understand the wet-dry fluctuation mechanism in the last 2700 years in the Qadam Basin. The vegetation pattern on the Loess Plateau during the late Holocene, when vegetation was probably greatly affected by human activities, remains poorly understand. We present a high-resolution (ca.20-yr) pollen record together with charcoal, conifer stomata and Pediastrum data from Tianchi Lake in the western Loess Plateau, in order to reconstruct the paleovegetation change in the study areas and to investigate the human impacts on forest change by comparing with other regional paleoclimatic records from speleothems, archaeological records of human habitation and historical documents. We draw the following conclusions:1. Sugan Lake(1) Both the surface samples from lake sediment and the Sugan Basin suggested that pollen in the lake are mainly from the whole Sugan Basin, the pollen assemblages from the lake sediment could well represent the regional vegetation. The desert and steppe vegetation type could be distinguished by different A/C (Artemisia/Chenopodiaceae) ratio, and A/C ratio could be used as an effective moisture index in the paleclimate reconstruction in our study region.(2) Pollen data shows that the Sugan Basin was constantly covered by open desert-steppe vegetation dominated by Chenopodiaceae, Artemisia, Poaceae and Ephedra. However, large variations in Artemisia/Chenopodiaceae (A/C) ratios suggest regional moisture fluctuations over the last 2700 years, including a dry and relatively stable climate prior to 300 AD, relatively wet climate from 300-1200 AD with variability during 1100-1200 AD, and unstable climate since 1200 AD with relatively moister climate during 1250-1400 AD and 1700-1800 AD.(3) The pollen-based moisture change shows a generally opposite phase relationship with that of other proxies (varve thickness, low-salinity chironomid taxa percentage, oxygen isotope of precipitated carbonate) from Sugan Lake. Our hypothesis is that fresh water input from ice melting on the surrounding mountains might be responsible for low values of carbonate,δ18O and salinity, while low effective moisture inferred from A/C ratio in the Basin is probably due to the high evaporation. As a result, carbonate, chironomids, and isotopes in the lake convey a "moister" climate signal and contradict moisture patterns in the basin inferred from pollen A/C ratio, which reflects regional environment. This finding has great significance in paleolinimology studies in arid and semi-arid regions, where the melt water from mountain glacier is the main supply of the lakes.(4) The comparisons of pollen data and other regional paleoclimatic proxies suggest the contrast of moisture change in the Basin and surrounding mountains: dry-wet climate in the basin corresponded to the wet-dry climate in the high elevation, respectively. The regional topography may be important in controlling this regional moisture patterns as mediated by rising and subsiding air masses at regional scale.(5) Pollen A/C ratio shows a big shift at ca.1200 AD, from stablely moist to drier and variable conditions. This climate shift has the nature of widespread global occurrence; however, more evidence for this transition is needed.2. Tianchi Lake(1) The pollen record from the Tianchi Lake indicates that the vegetation changes from mixed deciduous-coniferous forest dominated by Betula, Quercus, Abies and Pinus at 3200-2200 cal yr BP, to forest steppe co-dominated by trees (Betula, Quercus) and herbs (mostly Artemisia) at 2200-1100 cal yr BP and to steppe-like vegetation in the present.(2) Distinct vegetation type shift occurred at ca.1100 cal yr BP, accompanied by the increase of anthropogenic indicators (cereal-type pollen, Humulus-type pollen and Pediastrum). We propose that a critical threshold of the forest changes must have been crossed in this region at ca.1100 cal yr BP. Climate is probably the main controlling factor on the vegetation before ca.1100 cal yr BP, while the vegetation is mainly affected by the anthropogenic activities since then. The enhanced human activities are also indicated by a generally increasing trend of the micro-charcoal abundance during the last 1000 years.(3) Our results reveal several distinct deforestation phases which show good correlation with historical human activities. The first significant deforestation occurred at 1100-750 cal yr BP, correlating with the late Tang-Song dynasty, a period with dense population, flourishing agriculture as well as plenty of border conflict. Forest expansion during 700-350 cal yr BP corresponds to the Yuan and the early Ming dynasty, with sharply decreasing population and many abandoned agriculture land. The final substantial forest clearance occurred at ca.350 cal yr BP. Vegetation history inferred from pollen record generally agrees with the deforestation records based on the historical documents in the Loess Plateau. 3. Comparsions between the two lakesOur study suggests that the environment change in the Sugan Lake drainage during the late Holocene is mainly caused by the natural variability, and the SE Asian Monsoon, Westerlies and regional topography paly a key role; while the environment change around the Tianchi Lake during the late Holocene is resulted from the combintion of climate change and human activities. Our results reveal two different environmental change patterns during the late Holocen in Western China, which indicates the temporal and spatial difference of the environmental change.