Reconstruction of the vegetation distribution of different topographic units of the Chinese Loess Plateau during the Holocene

Quaternary Science Reviews 173 (2017) 236e247

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Reconstruction of the vegetation distribution of different topographic units of the Chinese Loess Plateau during the Holocene Aizhi Sun a, b, *, Zhengtang Guo a, b, c, Haibin Wu a, b, Qin Li b, c, Yanyan Yu b, c, Yunli Luo d, Wenying Jiang b, Xiaoqiang Li a, e a

College of Earth Sciences, University of Chinese Academy of Sciences, 19A, Yuquan Road, Beijing 100049, China Key Laboratory of Cenozoic Geology and Environment, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing 100029, China CAS Center for Excellence in Tibetan Plateau Earth Sciences, Beijing 100101, China d State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China e Key Laboratory of Vertebrate Evolution and Human Origins of the Chinese Academy of Sciences, Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing 100044, China b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 January 2017 Received in revised form 31 July 2017 Accepted 7 August 2017

Soil erosion and related ecological restoration present a tremendous challenge to the socioeconomic development of the Chinese Loess Plateau (CLP). Although the Chinese government has addressed the problem of soil erosion via an afforestation programme, there have been several negative outcomes. One of the reasons for this is our incomplete understanding of the past natural vegetation distribution in the various topographic units of the CLP under different climate scenarios. Consequently, we used fossil pollen data from 41 sites from different topographic units, together with the biomization method, to reconstruct the Holocene vegetation distribution of the CLP. The results demonstrate significant differences in vegetation types between different topographic units: forest was distributed in mountainous areas, steppe was dominant in Yuan areas, and desert vegetation was distributed in the transition zone between loess and desert. The vegetation in the gully areas exhibited significant spatial differences during the mid-Holocene. In addition, the vegetation on the various topographic units was welldeveloped during the interval from 9 to 4 ka B.P., when regional moisture levels reached a maximum. This suggests that the East Asian Summer Monsoon was one of the main factors controlling the evolution of vegetation patterns during the Holocene. In addition, our results confirm that both topography and human activity were fundamental factors determining the vegetation distribution of the region. Against a background of ongoing global warming, we advocate a program of vegetation restoration including planting trees and shrubs in the mountainous areas, and promoting the growth of grasses in the Yuan areas and in the transitional zone between loess and desert. In the gully areas, the planting of trees and shrubs is appropriate for reducing soil erosion caused by human activities. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Biome Vegetation change Topographic units Holocene Chinese Loess Plateau

1. Introduction The Chinese Loess Plateau (CLP), occupying an area of ~4.4
105 km2 in north-central China, is a major archive of continental climatic change spanning at least the last 22 Ma (Guo et al., 2002; Liu, 1985). Increased soil erosion, low tree survival rate, severe water shortages and deep soil desiccation (Normile, 2007; Wang et al., 2007, 2009) present tremendous challenges to the continued socioeconomic development of this relatively impoverished region. A major reason for this is the lack of knowledge of * Corresponding author. College of Earth Sciences, University of Chinese Academy of Sciences, 19A, Yuquan Road, Shijingshan District, Beijing 100049, China. E-mail address: [email protected] (A. Sun). http://dx.doi.org/10.1016/j.quascirev.2017.08.006 0277-3791/© 2017 Elsevier Ltd. All rights reserved.

the past natural vegetation distribution on the CLP. In addition, it is especially important to understand the spatiotemporal patterns of vegetation change during the most recent geological epoch, i.e., the Holocene (the past ~12,000 years), because projected global changes associated with ongoing climatic warming will occur under similar natural boundary conditions. In additional, the midHolocene was a more significant Megathermal (warm and wet interval) than the present (Feng et al., 2004; Kaufman et al., 2004; Marcott et al., 2013; Shi et al., 1992), and a more comprehensive understanding the vegetation distribution on the CLP during this interval may suggest strategies for promoting vegetation recovery in a warmer future (Intergovernmental Panel on Climate Change, 2013).

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

The vegetation distribution on the CLP during the Holocene has been studied using various proxy records and archives (e.g., Guo et al., 1994; Jiang et al., 2013a; Li et al., 2003). However, the findings are controversial. For example, based on historical documents and archaeological materials, several researchers have suggested that relatively dense forest was present over a large area of the CLP prior to its destruction by human activity (He and Tang, 1999; Shi, 1981, 1991; Zhu, 1983, 1994). In contrast, others have argued that climatic conditions on the CLP since the last glacial were only suitable for grassland (e.g., Guo et al., 1994, 1998; Jiang and Ding, 2005; Li et al., 2003; Liu et al., 1996; Sun et al., 1997; Zhou et al., 2009). Recent studies have pointed out that the different topographic units should be considered when reconstructing the vegetation distribution of the CLP during the Holocene (e.g., Jiang and Ding, 2005; Jiang et al., 2013a, 2014; Lu et al., 2003; Shang and Li, 2010; Zhang and An, 1994). However, several uncertainties remain. For example, previous synthesis studies were based on a relatively small number of sites, many lacked reliable age control, and many of the sections were geographically located in the Yuan areas (one of the major topographic units in the CLP). Consequently, an increased number of reliable and welldated proxy records are required to provide an improved understanding of the spatial distribution of the vegetation of the CLP during the Holocene. Pollen analysis is an effective tool for reconstructing the vegetation and climate history of arid and semi-arid environments (Sun and Feng, 2013; Xiao et al., 2002; Zhao et al., 2011), including the CLP region (e.g., Jiang et al., 2013a; Li et al., 2003; Sun et al., 1997). In the present study, we used the quantitative biomization method to reconstruct Holocene vegetation of the CLP based on a synthesis of fossil pollen data. The aims of the study were: (1) to determine the vegetation types in different topographic units on the CLP during the Holocene, (2) to discuss the characteristics of the Holocene vegetation distribution and evolution in the various topographic units, (3) to determine the main factors controlling the Holocene vegetation distribution on the CLP, and (4) to provide a basis for determining the most appropriate strategies for promoting future vegetation recovery. 2. Regional setting The main body of the CLP is located in the middle reaches of the Yellow River (Fig. 1a). The CLP can be divided into three parts by the

237

Liupan Mountains and the Luliang Mountains: an eastern part (located east of the Luliang Mountains), a central part (located between the Liupan and Luliang Mountains), and a western part (located to the west of the Liupan Mountains). From the geomorphic units, the CLP is mainly divided into the following areas: mountains, loess “Yuan”, loess hill, and valley plain. In additional, the characteristics of the boundary area between the CLP and the northern sandy land are different from the geomorphic units listed above. Here, combined with the terrain and material composition of the underlying surface, the geomorphic units are divided into four main types: (i) Mountainous areas (comprising bare bedrock or bedrock with a thin loess cover), (ii) Yuan areas (flat-topped loess highlands, covered with thick loess deposits), (iii) Gullies (one or two river terraces), and (iv) the transitional zone between loess and desert. Because of the interaction between the winter and summer monsoon, the modern climate of the CLP exhibits a distinct SE-NW gradient. Both mean annual temperature and precipitation decrease gradually from southeast (13  C and 650 mm) to northwest (7  C and 250 mm) (Wan et al., 2014), whereas the aridity (ratio of evaporation to precipitation) increases from southeast (1.0) to northwest (3.0). The vegetation distribution closely follows the aridity gradient: broad-leaved deciduous forest occurs in the southeastern corner of the CLP, forest-steppe in the south-eastern part, steppe in the northwestern part, and desert-steppe in the northwestern corner (Zhang, 2007) (Fig. 1b). The tree species comprising the forest community mainly consist of Pinus taulaefor, Quercus liaotungensis, Platycladus sp., and Populus tremula. Shrubs exhibit a relatively high species diversity, and are dominated by Rosa hugonis, Hippophae rhamnoides, Prinsepia uniflora, and Ostryopsis davidiana. The natural steppe vegetation is dominated by Stipa bungeana, S. breviflora, S. grandis, S. kryocii, Thymus mongolicu, Artemisia gmelinii and A. frigida. 3. Material and methods 3.1. Pollen data and site selection Numerous Holocene pollen records with varying data quality are available for the region. Due to the many records with low temporal resolution and/or depositional hiatuses, we first conducted an assessment of the data quality in order to select the most suitable datasets. The following criteria were used for selecting

Fig. 1. Location of the Chinese Loess Plateau. (a) 41 studied pollen sites (dark brown squares) in the CLP. (b) The distribution of modern vegetation (Zhang, 2007) and mean annual isohyets (gray lines) across the Chinese Loess Plateau from Wan et al. (2014).

238

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

fossil pollen records: (1) The sites were from natural sections (nonarchaeological sites). (2) A reliable chronology, with a minimum of two age control points during the Holocene, was available. The age control points for the various sections were obtained using two different approaches: By direct measurement by radiocarbon accelerator mass spectrometry (AMS 14C), liquid scintillation radiocarbon (LSC 14C), optically stimulated luminescence (OSL), and thermoluminescence (TL). In the other approach, ages were obtained indirectly by identifying the stratigraphic contrasts between the loess-palaeosol intervals, and then dating the sequences indirectly via correlation with the Weinan section, for which detailed stratigraphic and direct dating were already available (Kang et al., 2013). Further details of age model development are given in section 2.2. (3) The loess-palaeosol sequences should have highresolution magnetic susceptibility or grain-size data. (4) The records should be continuous, spanning a minimum of 3000 years and without documented depositional hiatuses. A total of 41 sites passed these selection criteria and their locations are shown in Fig. 1a. 21 sites are from the Yuan areas, 8 sites are from gully areas, 7 sites are from mountainous areas, and 5 sites are from the transitional areas between loess and desert. More than 90% of our synthesis is based on research results since 2000. The number of dates or age control points for individual sections ranged from 2 to 23, with an average of 5. The number of pollen samples ranges from 10 to 250, with an average of 55. More detailed information about the sites is given in the Table in the Appendix. 3.2. Chronology As indicated in the previous section, direct dates were determined by AMS 14C, LSC 14C, OSL, and TL dating; the corresponding numbers of dates are 82, 25, 12, and 12, respectively. We used the original reference's depth-age model to obtain a chronology for the pollen data. Where direct dates were not available, we used indirect dating based on magnetic susceptibility (MSM) or grain-size method (GSM) methods. Correlations were made between the studied loess-palaeosol sequences and the Weinan section for which detailed stratigraphic and OSL dating (Kang et al., 2013) are already available, which yielded ages for the various layers (e.g., S0, L1-1). In addition, a chronology for the loess-palaeosol sequences was obtained using either the magnetic susceptibility method (Kukala and An, 1989) or the grain-size method (Porter and An, 1995). In this study, some of the dates (i.e., OSL, TL) are calendar dates and others (i.e., AMS 14C and LSC 14C) are uncalibrated 14C dates. In order to obtain a consistent chronology, all 14C ages (i.e., AMS 14C and LSC 14C) were calibrated to calendar years before the present (BP ¼ 1950 CE) using the CalPal2007 calibration curve (Weninger et al., 2007). Calibrated ages are used throughout (expressed as ka B.P.; 1 ka B.P. ¼ 1000 cal yr B.P.). We re-sampled each of the original time series at a 1000-year interval by averaging the original data in each time window. 3.3. Method of vegetation reconstruction Vegetation was reconstructed using a pollen-based quantitative method (biomization) (Prentice et al., 1996). The biomization method was developed based on knowledge of the biogeography and ecology of modern plants and associated pollen assemblages (Prentice et al., 1992; 1996). The method consists of four steps: (1) assignment of each pollen taxon to one plant functional type (PFT) or more PFTs according to its known ecology and biogeography, (2) assignment of characteristic PFTs to major vegetation types (biomes) according to their bioclimatic ranges, (3) calculation of the affinity scores for all pollen samples using an equation in which the score of a given biome is the sum of the square roots of the

percentage (e.g., above 0.5%) of each taxon present in the biome. The two matrices, including pollen taxa-PFTs and PFTs-Biomes, used in this study are based on the biomization procedure given in the Members of China Quaternary Pollen Data Base (2000). The biomization procedure was implemented using the PPPbase software (Guiot and Goeury, 1996). The vegetation type at a 1000-year time interval was reconstructed for each site. The spatial pattern of reconstructed vegetation was presented using the ArcGIS 10.0 software. 4. Results Based on the biomization reconstruction, the main Holocene vegetation (biome) types on the CLP are STEP (steppe), COMX (cool mixed forest), TEDE (temperate deciduous forest), and DESE (desert). The vegetation distribution on the CLP at 1000-year intervals is illustrated in Fig. 2. The results show that, broadly, steppe was dominant on the CLP, while forest (including COMX and TEDE) and desert vegetation continued to develop during the Holocene. During the interval from 12 to 9 ka B.P., steppe vegetation was dominant and a small area of forest and desert appeared; forest developed in the mountainous areas (e.g., the Qinling and Luliang Mountains) in the eastern and southern parts of the CLP, and desert emerged in the western and the northern part of the CLP, for example in the transitional region between the loess and desert areas. During the interval from 9 to 2 ka B.P., although steppe vegetation was dominant, the amount of forest increased significantly; 9 sites were occupied by forest and the distribution expanded to the central (e.g., the Liupan and Huanglong Mountains) and western (e.g., the Ziwuling Mountains) parts of the CLP. There was a slight increase in the number of sites occupied by desert vegetation, although its distribution was still located in the western and northern parts of the CLP during the interval from 7 to 6 ka B.P. After 2 ka B.P., steppe was numerically dominant in terms of the number of sites and was widely distributed over the CLP; the number of sites occupied by forest and desert vegetation decreased significantly, although there was some development of forest in the southern and eastern parts of the CLP, and desert developed in the western and central parts. To better illustrate the distribution and evolution of vegetation on the different geomorphic units of the CLP, we first generated frequency statistics of the numbers of the four vegetation types (i.e., STEP, COMX, TEDE, and DESE) occurring within a 1000-year interval, and calculated the percentage of each vegetation type. Secondly, we calculated the biome scores of the four vegetation types at a 1000-year interval. The results for the different topographic units of the CLP during the Holocene are illustrated in Fig. 3 and are summarized as follows: Mountainous areas (Fig. 3A) - The average proportion of forest (including TEDE and COMX) exceeds 60%, and the scores of the forest biome types are significantly higher than for STEP and DESE. The proportion and scores for forest vegetation were the highest from 9 to 3 ka B.P., and the proportion and scores for steppe vegetation decreased significantly during this interval. The proportion (<25%) and scores for desert are the lowest among the four vegetation types. Yuan areas (Fig. 3B) - The proportion (>80%) and scores for steppe are the highest among the four vegetation types. During the interval from 10 to 0 ka B.P., the average proportion of forest was less than 10%. COMX was present from 10 ka B.P. onwards, and TEDE from 8 ka B.P. Although the DESE score is higher, the proportion is lower (less than 10%) than for the other types. Gully areas (Fig. 3C) - The proportion of steppe is the highest (>50%), and the proportions for forest and desert are less than 25%. Forest vegetation continued to develop from 10 to 4 ka B.P.

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247 Fig. 2. Spatial pattern of vegetation distribution on the CLP at 1000-year intervals during the Holocene (0e12 ka B.P.). The vegetation types are: STEP, steppe; Forest, including COMX (cool mixed forest) and TEDE (temperate deciduous forest); and DESE, desert. 239

240

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

Fig. 3. Vegetation changes in different topographic units during the Holocene (0e12 ka B.P.). A: Mountainous areas, B: Yuan areas, C: Gullies, D: Transition zone between loess and desert. Vegetation types: COMX: cool mixed forest, TEDE: temperate deciduous forest, STEP: steppe, and DESE: desert.

Significant changes in vegetation scores occurred for this unit during the Holocene. From 12 to 10 ka B.P., DESE scored the highest, followed by STEP, and TEDE scored the lowest. Subsequently, the scores for STEP and DESE decreased significantly while those for TEDE and COMX increased. From 8 to 4 ka B.P., the score for TEDE changed the most, and subsequently the score for STEP increased and reached a maximum after 3 ka B.P. Loess/desert transitional areas (Fig. 3D) - The relative proportions of steppe and desert exhibit an alternating trend. The proportion of steppe was higher than that of desert during 11e10 ka B.P., 9e7 ka B.P., and 3e0 ka B.P., and the proportion of desert was higher than that of steppe from 5 to 3 ka B.P. The scores for STEP and DESE are significantly higher than that for forest, with the score for STEP higher than for DESE from 8 to 7 ka B.P. and from 2 to 0 ka B.P. The foregoing results clearly demonstrate that the different topographic units were characterized by significantly different vegetation types during the Holocene. Forest vegetation was

dominant in the mountainous areas, steppe vegetation completely dominated the Yuan and gully areas, while in the loess/desert transitional areas there were fluctuations between loess and desert vegetation. 5. Discussion 5.1. Vegetation distribution on the different topographic units of the CLP during the Holocene Our biome reconstruction results demonstrate significant differences in vegetation types on the different topographic units of the CLP during the Holocene. In the Yuan areas, steppe vegetation was completely dominant, although a small amount of forest was continuously present from 10 to 0 ka B.P. Various other vegetation proxy records from the loess-soil sequence of the Yuan areas, including soil micro-morphological evidence (Guo et al.,1994,1998),

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

phytoliths (Lu et al., 1996, 1999), stable carbon isotopes (Lin et al., 1991; Liu et al., 2011; Yao et al., 2011), and organic molecular compounds (Xie et al., 2002), also provide valuable information about the past vegetation distribution. These records reveal that steppe vegetation developed on the Yuan areas during the Last Interglacial (including the Holocene), in agreement with our findings. Our results indicate that forest located in mountainous areas reached its greatest proportion (>80%) from 7 to 3 ka B.P. However, the extent of the distribution of forest on the CLP during the Holocene is controversial. Studies based on the analysis of the other proxies listed above (soil micromorphology, phytoliths, stable carbon isotopes, and organic molecular compounds) conclude that the CLP was dominated by steppe during the Holocene, and that the distribution of forest was limited (Guo et al., 1998; Liu et al., 2011; Lu et al., 1999; Xie et al., 2002). However, based on documentary and archaeological evidence, Shi (1981) concluded that forest occupied more that 50% of the area of the CLP during the Zhou dynasty (771 BC-1046 AD). However, a detailed analysis of geological, archaeological and documentary evidence can lead to different conclusions. Firstly, the geological records were obtained mainly from the Yuan areas and may not be representative of all of the topographic units of the CLP. Secondly, many historical documents do not distinguish between the different topographic units, especially mountainous and Yuan areas. However, in the case of those historical documents which do refer to forest, it is mainly located on mountains (Wang, 1990). Thirdly, in Shi (1981)’s synthesis of documentary evidence, primary and secondary forest were not distinguished. These issues may potentially result in both over- and under-estimates of the extent of primary forest in the CLP. The mid-Holocene (8e5 ka B.P.) was the Holocene climatic optimum (warm and wet) on the CLP (Liu, 1985), and thus the maximum extent of forest would be expected to have occurred during this interval. The vegetation distribution during the interval from 8 to 5 ka B.P. is shown in Fig. 2. Forest (including TEDE and COMX) was mainly distributed in the mountainous areas, for example on Luliang, Qinling, Liupan, Huanglong, Ziwuling Mountains. Based on statistical analysis of the 151 plant-species from different topographic units recorded in “The Book of Songs”, He (1969) found that 70% of the forest was in mountainous areas and less than 25% was in low-lying wetland areas. Consequently, he concluded that forest mainly occurred in the mountains, in agreement with our results. Overall, our findings reveal that forest on the CLP during the Holocene was mainly located in the mountainous areas and it reached its greatest extent in the mid-Holocene. Generally, the gully areas have the most suitable hydrological conditions for tree growth on the CLP, and more trees and shrubs are able to grow here than in the Yuan areas (Zhang and An, 1994; Shang and Li, 2010). For example, the gullies in the mountainous areas probably have better hydrological conditions due to higher precipitation and greater groundwater seepage; the gullies in the Yuan areas probably have floodplain- or low terrace-related improvement in hydrogeomorphic conditions; and gullies in loess hill areas probably have floodplain and low terrace-related and groundwater seepage-related improvement of hydrogeomorphic conditions. However, the results shown in Fig. 2 reveal that the vegetation in the gully areas was dominated by steppe during the Holocene and that the proportion of forest was less than 30%. Fig. 2 also reveals spatial differences in vegetation type in the gully areas during the mid-Holocene (8e5 ka B.P.). In the river valleys in the western part of the CLP (e.g., the Zuli River, and upper Weihe River), mixed coniferous broad-leaved forest (e.g., COMX) was present during the mid-Holocene. Charcoal records from the Dadiwan and Xishanping archaeological sites in the valleys of the western part of the CLP also indicate that COMX developed during the Holocene

241

climatic optimum (Li et al., 2013). In contrast, in the river valleys in the eastern part of the CLP (e.g., the middle and lower Weihe River, Luohe River and Fenhe River), the vegetation was dominated by steppe (e.g., sparsely-wooded grassland or grassland). One of the reasons for the development of steppe in the eastern part of the CLP gully areas may have been the impact of human activity on the extent of forest. The number and size of settlements provide a basis for estimating the population size and the regional intensity of human activity (Li et al., 2009). Li et al. (2009) assembled data for 1909 archaeological sites and concluded that the numbers of known archaeological sites increased greatly from the Yangshao period (7e5 ka B.P.) onwards and that the density of sites remained high through to the time of the western Zhou Dynasty, suggesting that a high intensity of human activity was maintained from the Neolithic to the Iron Age in the Guanzhong Basin, which is located in the middle Weihe River. Using recent archaeological and environmental data, together with the development of a quantitative prehistoric land use model (PLUM), Yu et al. (2016) characterized spatiotemporal changes in the intensity of human activity in the Wei River valley. The land area used by humans increased significantly between 8 and 4 ka B.P., expanding from the gentle slopes along the lower reaches of the river to the middle and upper reaches, implying a significant impact on the vegetation of the gully areas. In addition, Yu et al. (2012) reconstructed the land use during the period from 8 to 4 ka B.P. in Yiluo river valley (located in the eastern part of the CLP), and the results reveal that about 2e9% of the land area in the valley was already used for human activities. From the foregoing, it is reasonable to conclude that the vegetation in the eastern part of the CLP river valleys during the midHolocene may have been impacted by human activity. However, its intensity and spatial extent needs to be more comprehensively addressed via synthesis studies in the future. 5.2. Vegetation evolution on the CLP during the Holocene The main trends of vegetation evolution on the different topographic units of the CLP during the Holocene are illustrated in Fig. 3. In the mountainous areas, the proportion and scores for forest were the highest from 9 to 3 ka B.P., and subsequently the proportion and scores of steppe increased significantly. This suggests that optimum conditions for forest development in the mountainous areas occurred during 9e3 ka B.P. In the gully areas, forest continued to develop from 10 to 4 ka B.P. Scores for forest increased significantly at 9 ka B.P., attained their highest values from 7 to 4 ka B.P., and subsequently the scores for steppe increased and reached their highest values. These findings suggest that, as in the mountainous areas, the conditions for forest development in the gully areas was optimal from 9 to 3 ka B.P. In the Yuan areas, although the proportions and scores for steppe are the highest of the various topographic units, the average proportion of forest appears from 10 ka B.P. and was maintained until the present, while the scores for forest also increased from 9 to 5 ka B.P. The changes in the proportions and biome scores for the four vegetation types on the CLP during the Holocene are illustrated in Fig. 4. The results show that the proportions and scores are highest for steppe, revealing that the vegetation on the CLP as a whole was dominated by this vegetation type. It is noteworthy that the proportion of forest began to increase from 10 ka B.P. onwards, reached its maximum from 6 to 5 ka B.P., and decreased significantly from 2 ka B.P. onwards; its proportion was greater than 20% from 9 to 2 ka B.P. Changes in the scores for forest are consistent with those of the changes in proportion, with the scores being higher from 9 to 4 ka B.P. The changes of the proportion and scores for steppe vegetation contrast with those for forest, suggesting that forest development on the CLP was optimal during the interval from 9 to 4 ka B.P.

242

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

The vegetation evolution revealed by our synthesis of pollen data from the CLP agrees well with other records from the region. For example, phytolith records indicate that forest steppe and sparse forest steppe dominated the southernmost part of the CLP during the mid-Holocene (Lu et al., 1999). In addition, d13C data from the Yuan areas indicate that the contribution of C4 biomass may have been up to ~50e60% in the most southeastern part of the CLP, and up to ~30e40% in the northwest, suggesting that C4 grasses (steppe) flourished on the Yuan area during the mid-Holocene (Yao et al., 2011). Both findings are consistent with our results. 5.3. Factors controlling vegetation distribution and evolution on the CLP during the Holocene Despite differences in vegetation types between different topographic units, our results indicate that forest (mainly in the mountainous areas) and steppe (mainly in the Yuan areas and in the transitional zone between loess and desert) were welldeveloped during 9e4 ka B.P., suggesting that precipitation levels on the CLP were at a maximum at that time. Comparison of our pollen-based precipitation record with other proxy climate records from the CLP and elsewhere are illustrated in Fig. 5. The records include magnetic susceptibility (Fig. 5b) (Peterse et al., 2011), which is widely used to indicate changes in summer monsoon strength (An et al., 1991; Liu and Liu, 1991), and the paleosol sediment density (Fig. 5c) (Wang et al., 2014). There is a high degree of consistency between our record and that of other records reflecting changes in precipitation conditions on the CLP. Our results are also consistent with climatic records from adjacent regions, including tree pollen percentages in the monsoon marginal regions of northern China, including Bayanchagan (Fig. 5d) (Jiang

et al., 2006) and Daihai (Xiao et al., 2002), sedimentary facies of aeolian deposits in the East Sandy land (Fig. 5e) (Li et al., 2014a), and a synthesized moisture record based on tree pollen percentages from the eastern Tibetan Plateau (Fig. 5f) (Zhao et al., 2011). Recently, Yang et al. (2015a) presented d13C records of bulk organic matter from 21 loess sections across the CLP since the Last Glacial Maximum (LGM), and the results indicate that the C4 biomass increased from the LGM to the mid-Holocene, meaning that precipitation delivered by the East Asian Summer Monsoon (EASM) increased during the mid-Holocene on the CLP. These records confirm that precipitation generally increased from the early Holocene, reached a maximum during the mid-Holocene, and then decreased in the late Holocene. The timing of the peak in precipitation varies between the different proxies, which may be the result of differences in their climatic responses. The results from numerical climate models indicate that, following the interval of peak insolation in the early Holocene, the low-latitude oceans continued to warm until the mid-Holocene (Jiang et al., 2013b). This resulted in the enhancement of the sea-level pressure gradient between the East Asian continent and adjacent oceans, causing a strengthening of the EASM during the mid-Holocene (Jiang et al., 2013b). The strengthened EASM resulted in increased precipitation transport from the ocean to the arid and semi-arid region in northern China (Maher, 2016), in turn resulting in soil formation and vegetation development (Li et al., 2014b; Yang et al., 2015a; Wang et al., 2014). Therefore, it can be concluded that the evolution of the EASM was of the main factors controlling vegetation evolution in the CLP during the Holocene. EASM changes were driven by three major forcing factors: (1) Earth orbitally-induced changes in solar insolation (An et al., 1990; Berger and Loutre, 1991; COHMAP, 1988; Kutzbach, 1981), (2)

Fig. 4. Changes in the four vegetation types on the CLP during the Holocene (0e12 ka B.P.), represented by percentages and biome scores. Vegetation types: COMX: cool mixed forest, TEDE: temperate deciduous forest, STEP: steppe, and DESE: desert.

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

243

Fig. 5. Comparison of effective moisture levels on the CLP indicated by the proportions and biome scores for forest vegetation (a) with various other proxy moisture records for north China the magnetic susceptibility record at Mangshan (Peterse et al., 2011) (b); probability density of palaeosol dates from the CLP (Wang et al., 2014) (c); tree pollen percentages at Daihai Lake (Xiao et al., 2002) and Bayanchagan Lake (Jiang et al., 2006) (d); sedimentary facies in the East Sandy Land (Li et al., 2014a,b) (e); synthesized moisture record based on tree pollen percentages from the Eastern Tibetan Plateau (Zhao et al., 2011) (f).

atmospheric CO2 concentration (Kripalani et al., 2007; Liu et al., 2009; Lu et al., 2013), and (3) changes in Arctic ice-sheets (Ding et al., 1995, 2005; Liu and Ding, 1992; Porter, 2001). Summer solar insolation at 30 N started to increase around 20 ka B.P. and reached its peak from 12 to 8 ka B.P. (Kutzbach and Gallimore, 1988). The insolation peak at 12e8 ka B.P. may have sufficiently warmed the low-latitude oceans to encourage the northward advance of the summer monsoon at 11e10 ka B.P. (Koutavas et al., 2006). However, a comparison of our reconstructed EASM precipitation record with summer insolation reveals that our record (Fig. 6d) significantly lags (by approximately 3e4 ka) changes in low latitude (30 N) Northern Hemisphere summer insolation (Fig. 6a). Similarly, our monsoon precipitation record also exhibits a lag of approximately 3e4 ka relative to the record of atmospheric CO2 concentration from Antarctic ice cores (Lüthi et al., 2008; Parrenin et al., 2013) (Fig. 6b). These discrepancies lead us to hypothesize that other mechanisms, and not northern hemisphere summer insolation or CO2 alone, drive EASM precipitation. The variations in EASM intensity reconstructed in our study correlate well with variations in Northern Hemisphere ice volume, which suggests that this may have been a major factor in controlling monsoon evolution in the CLP. Li et al. (2014b) compared variations in EASM precipitation since the LGM recorded in the desert of northern China with those of various proxies of potentially significant drivers: The Siberian High indicated by the EAWM (De Garidel-Thoron et al., 2001; Stevens et al., 2007; Sun et al., 2006), temperature gradients between low latitudes and Arctic regions,

sea surface temperature (SST) records in the western Pacific warm pool reconstructed from Mg/Ca ratios, and sea level. Their results indicate that EASM precipitation was correlated with increasing from 21 to 6 ka B.P., with the EASM maximum coincident with the highest sea level at approximately 6 ka B.P. In contrast, there is a clear discrepancy between the variation of EASM precipitation in the East Sandy Land and the intensity of the Siberian High, temperature gradients and SST records in the western Pacific warm pool: the latter three records did not achieve peak values until the early Holocene (Li et al., 2014b). A comparison of our reconstructed EASM precipitation record with sea level (Liu et al., 2014) and Northern Hemisphere ice-sheet area (Dyke, 2004) (Fig. 6c) indicated that rising sea level rise may have been an important forcing factor for the intensification of EASM precipitation from 12 to 6 ka B.P. The changes in precipitation after ~3 ka B.P., reflecting the weakening of the EASM, agrees with the coeval decrease in northern hemisphere summer insolation, suggesting that this became an important driver of EASM precipitation in northern China once sea level reached its modern level. The CLP is characterized by a varied topography which creates a mosaic of hydrological and soil microclimatic conditions (Zhang and An, 1994). In addition, very thick loess deposits (60e300 m) are present in the Yuan areas in the CLP; the loess is very fine in texture and mainly composed of loosely cemented silt (Liu, 1985; Yang and Ding, 2008), enabling the rapid infiltration of rainwater (Yang et al., 2012). Consequently, the water content of the surface soil is insufficient to maintain forests in the Yuan areas (Liu, 1985; Liu et al.,

244

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

Fig. 6. Comparison of the effective moisture levels (d) indicated by vegetation types in the CLP with other paleoclimatic records. (a) Summer insolation at 30 N (Berger and Loutre, 1991), the mean temperature from 30 to 90 N, and global temperature (Shakun et al., 2012). (b) Atmospheric carbon dioxide (CO2) concentration from Dome Concordia, Antarctica (Lüthi et al., 2008). (c) Northern Hemisphere ice-sheet area (Dyke, 2004).

1996; Shang and Li, 2010). However, in mountainous rocky areas, or in mountainous areas where the bedrock is covered by a thin loess layer, the bedrock itself can act as a water-resistant layer which allows the surface soil to retain the rainwater, enabling forest growth (Liu et al., 1996; Yang et al., 2015b; Zhang and An, 1994). In addition, although the hydrothermal conditions in the gully areas are generally suitable for the growth of trees and shrubs, the impact of human activities such as deforestation and agriculture, is significant. During the mid-Holocene, spatial differences in vegetation composition between the western and eastern part of the CLP river valley may have been affected by human activity. Therefore, human activity, in addition to topography, may have been an important factor in controlling the distribution of forest in the CLP. 6. Conclusions We have used fossil pollen data from 41 sites from different topographic units, together with the biomization method, to reconstruct the distribution and evolution of vegetation on the CLP during the Holocene. The results reveal significant differences in vegetation type between different topographic units: forest vegetation dominated the mountainous areas, steppe vegetation completely occupied the Yuan and gully areas, and desert vegetation was mainly developed in the transition zone between loess and desert. In spites of the differences in vegetation types in the different topographic units, the vegetation of the CLP was well-developed during 9e4 ka B.P., suggesting that precipitation levels were at a maximum during this interval. The strengthened EASM resulted in

increased moisture transport from the ocean to the arid and semiarid region in northern China, resulting in soil formation and vegetation development. Therefore, the development of the EASM was one of the main factors controlling the vegetation evolution in the CLP during the Holocene. In addition, both topography and human activity were major factors affecting the development of forest vegetation. Our results suggest that against a background future global warming, vegetation development will follow different pathways on the different topographic units of the CLP. Forest may expand in the mountainous areas, steppe may expand in the transition zone between loess and desert, and may also be dominant in the Yuan areas. However, if the destructive effects of human activity can be controlled and mitigated, forest vegetation is likely to develop in the gully areas. Therefore, in the context of ongoing climate change, an appropriate environmental management strategy for the region could include planting trees and shrubs in the mountainous areas, returning farmland to grassland in the Yuan areas, and planting trees and shrubs in the gully areas to reduce the destructive effects of human activity. Acknowledgments We are grateful to two anonymous reviewers for their useful suggestions. This research was funded by the National Natural Science Foundation of China (grant nos. 41430531, 41690114, and 41125011), the National Basic Research Program of China (Grant no. 2016YFA0600504), and the Bairen Programs of the Chinese Academy of Sciences.

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

245

The information about pollen data used in this study are presented as follows Appendix Table Characteristics of the fossil pollen sites selected from the Chinese Loess Plateau No Name Province LONG. LAT.

ALT. Data type

Archive type

Depth (cm)

Dating method

Num. of dating

Time span (ka Pollen data BP) points

Time resolution

Reference

1 2 3 4 5 6 7

BZC DX FX JCY QS XF XD

Shaanxi Gansu Shaanxi Shaanxi Shaanxi Shaanxi Shaanxi

605 1397 1188 1122 678 1400 621

Digit Digit Digit Digit Digit Digit Digit

Yuan Yuan Yuan Yuan Yuan Yuan Yuan

575 465 420 190 400 200 475

2 2 3 3 7 2 4

2.6e9.7 0.2e4.0 0e25.3 0e9.4 0e14.8 0e8.3 2.7e9.5

29 25 28 81 100 74 41

245 152 904 116 148 112 166

Shang and Li, 2010 Tang and An, 2007 Cheng, 2011 Gong, 2006 Han, 2000; Huang et al., 2002 Xu, 2006 Shang and Li, 2010

8

YGZ

Shaanxi 109.50 34.40 650

Digit

Yuan

1000

9

3.0e100.0

106

915

Sun et al., 1996

9 10 11 12 13 14 15 16 17 18 19 20

SGY YX PY PG WX LC PL HS FX JX XF WY

Shaanxi Shaanxi Shanxi Shanxi Shanxi Shaanxi Shaanxi Gansu Shaanxi Shanxi Shanxi Gansu

109.38 110.17 112.50 111.50 111.20 109.54 106.86 108.29 109.30 110.64 111.45 104.25

35.60 35.93 37.30 39.50 35.50 35.71 35.54 35.78 36.02 36.11 35.84 35.13

955 1556 1500 1400 1100 1068 1514 1475 1226 990 560 2061

Raw Digit Digit Digit Digit Digit Digit Digit Digit Digit Digit Digit

Yuan Yuan Yuan Yuan Yuan Yuan Yuan Yuan Yuan Yuan Yuan Yuan

320 392 350 450 400 1200 440 350 450 470 460 700

MSM AMS 14C GSM TL MSM TL OSL, AMS 14C AMS 14C, TL MSM TL MSM MSM OSL LSC 14C, TL GSM GSM GSM GSM GSM AMS 14C

3 2 4 3 5 5 3 3 3 3 3 4

0e23.0 0e12.0 0e21.6 0e30.0 0e43.9 0.1e83.1 6.0e23.0 2.8e24.5 1.0e32.0 6.0e25.0 5.5e25.0 0.2e26.0

32 67 34 45 37 62 12 12 28 12 12 17

719 179 635 667 1186 1339 1417 1808 1107 1583 1625 1518

21 22 23 24 25 26 27 28

HN DDW HY LYS LZ LT DST SJW

Gansu Gansu Ningxia Henan Gansu Shaanxi Shanxi Gansu

104.87 105.91 105.97 112.40 103.67 109.50 113.78 104.52

36.12 35.00 36.43 34.80 36.00 34.30 40.13 35.54

1911 1566 1701 251 1520 866 1247 1959

Digit Raw Raw Digit Raw Digit Digit Raw

Yuan Gully Gully Gully Gully Gully Gully Gully

800 500 850 580 260 1050 600 450

GSM AMS 14C AMS 14C LSC 14C LSC 14C LSC 14C LSC 14C AMS 14C

3 4 9 4 6 4 3 3

0.6e25.0 0e15.0 7.4e14.0 0e9.2 1.0e8.4 1.7e7.7 0e11.8 0e9.5

20 250 210 146 12 17 58 85

1220 60 31 63 617 353 203 112

Unpublished Li et al., 2003 Zhou et al., 2014 Zhou et al., 2014 Zhou et al., 2014 Li and Sun, 2004 Jiang et al., 2013a,b Jiang et al., 2013a,b Jiang et al., 2013a,b Jiang et al., 2013a,b Jiang et al., 2013a,b Yang et al., 2015a; Yang et al., 2015b Yang et al., 2015a An et al., 2003 Sun et al., 2007 Sun and Xia, 2005 Wang et al., 1991 Li and Sun, 2005 Fan et al., 2007 An et al., 2003; Sun et al., 2010

109.53 104.61 109.28 106.97 107.75 107.65 107.80

34.35 35.55 36.03 34.27 34.43 35.73 34.38

References An, C.B., Feng, Z.D., Tang, L.Y., 2003. Evidence of a humid mid-Holocene in the western part of Chinese Loess Plateau. Chin. Sci. Bull. 48 (22), 2472e2479. An, Z.S., Wu, X.H., Wang, P.X., Wang, S.M., Dong, G.R., Sun, X.J., Zhang, D.E., Lu, Y.C., Zheng, S.H., Zhao, S.L., 1991. The evolution in China paleomonsoon since 130 ka: Ⅱ. Paleomonsoon changes. Sci. China Ser. B 11, 1209e1215 (in Chinese). An, Z.S., Xiao, J.L., Zhang, J.Z., 1990. Monsoon and climate change in the Chinese loess plateau since 130,000 years. In: Liu, T.S. (Ed.), Loess. Quaternary Geology. Global Change (Part Two). Science Press, Beijing, pp. 108e114. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quat. Sci. Rev. 10, 297e317. Cheng, Y.F., 2011. Vegetation and Climate Changes in the Middle and Northern Loess Plateau over the Past 26000 Years. Master degree dissertation (in Chinese). COHMAP, 1988. Climatic changes of the last 18,000 years: observations and model simulations. Science 241, 1043e1052. De Garidel-Thoron, T., Beaufort, L., Linsley, B.K., Dannenmann, S., 2001. Millennialscale dynamics of the East Asian winter monsoon during the last 200,000 years. Paleoceanography 16, 491e502. Ding, Z.L., Derbyshire, E., Yang, S.L., Sun, J.M., Liu, T.S., 2005. Stepwise expansion of desert environment across northern China in the past 3.5 Ma and implications for monsoon evolution. Earth Planet. Sci. Lett. 237, 45e55. Ding, Z.L., Liu, T.S., Rutter, N.W., Yu, Z.W., Guo, Z.T., Zhu, R.X., 1995. Ice-volume forcing of East-Asian winter monsoon variations in the past 800,000 years. Quat. Res. 44, 149e159. Dyke, A.S., 2004. An outline of North American deglaciation with emphasis on central and northern Canada. Quat. Glaciat. Extent Chronology 2, 373e424. Fan, S.X., Liu, H.K., Zhou, J.Z., Tong, G.B., Bi, Z.W., 2007. Spore and pollen characteristics and environment evolution in Datong basin since Holocene. Acta Geoscientica Sinica 28 (6), 535e540 (in Chinese). Feng, Z.D., An, C.B., Tang, L.Y., Jull, A.J.T., 2004. Stratigraphic evidence of megahumid mid-Holocene climate in the western part of the Chinese Loess Plateau. Glob. Planet. Change 43, 145e155. Gong, X.Y., 2006. A high Resolution Holocene Palaeoventation Study on Jiacunyuan in Baoji, Shaanxi Province. Master degree thesis of Capital Normal University, China (in Chinese). Guiot, J., Goeury, C., 1996. PPPBASE, a software for statistical analysis of paleoecological and paleoclimatological data. Dendrochronologia 14, 295e300.

Guo, Z.T., Liu, T.S., An, Z.S., 1994. Paleosols of the last 0.15 Ma in the Weinan loess section and their paleoclimatic significance. Quaternanry Sci. 3, 256e269 (in Chinese). Guo, Z.T., Liu, T.S., Fedoroff, N., Wei, L.Y., Ding, Z.L., Wu, N.Q., Lu, H.Y., Jiang, W.Y., An, Z.S., 1998. Climate extremes in Loess of China coupled with the strength of deep-water formation in the North Atlantic. Glob. Planet. Change 18, 113e128. Guo, Z.T., Ruddiman, W.F., Hao, Q.Z., Wu, H.B., Qiao, Y.S., Zhu, R.X., Peng, S.Z., Wei, J.J., Yuan, B.Y., Liu, T.S., 2002. Onset of Asian desertification by 22 Myr ago inferred from loess deposits in China. Nature 416, 159e163. Han, Y.P., 2000. Vegetation and Environment Change during Holocene in Zhouyuan (Qishan) Region. Master degree thesis of Shaanxi Normal University, China (in Chinese). He, B.D., 1969. Loess and the Origin of Chinese Agriculture. The Chinese University Press, Hongkong, China, pp. 1e228 (in Chinese). He, X., Tang, K., 1999. Potentiality of vegetation construction on the loess plateau. Soil Watern Conservation China 204, 31e35 (in Chinese). Huang, C.C., Pang, J.L., Huang, P., Hou, C.H., Han, Y.P., 2002. Holocene climatic events on the Loess tableland in the western Guanzhong basin, China. Arid. Land Geogr. 25 (1), 10e15 (in Chinese). IPCC Working Group I-Twelfth session, 2013. The IPCC Fifth Assessment report (AR5), Climate Change 2013: the Physical Science Basis. IPCC, Stockholm, Sweden. Jiang, D.B., Lang, X.M., Tian, Z.P., Ju, L.X., 2013b. Mid-Holocene East Asian summer monsoon strengthening: insights from paleoclimate modelling intercomparison project (PMIP) simulations. Palaeogeogr. Palaeoclimatol. Palaeoecol. 369, 422e429. Jiang, H.C., Ding, Z.L., 2005. Temporal and spatial changes of vegetation cover on the Chinese Loess Plateau through the last glacial cycle: evidence from spore-pollen records. Rev. Palaeobot. Palynology 133, 23e37. Jiang, W.Y., Chen, Y.F., Yang, X.X., Yang, S.L., 2013a. Chinese loess plateau vegetation since the last glacial maximumu and its implications for vegetation restoration. J. Appl. Ecol. 50, 440e448. Jiang, W.Y., Guo, Z.T., Sun, X.J., Wu, H.B., Chu, G.Q., Yuan, B.Y., Hatte, C., Guiot, J., 2006. Reconstruction of climate and vegetation changes of lake bayanchagan (inner Mongolia): Holocene variability of the east Asian monsoon. Quat. Res. 65, 411e420. Jiang, W.Y., Yang, X.X., Cheng, Y.F., 2014. Spatial patterns of vegetation and climate on the Chinese loess plateau since the last glacial maximum. Quat. Int.

246

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247

334e335, 52e60. Kang, S.G., Wang, X.L., Lu, Y.C., 2013. Quartz OSL chronology and dust accumulation rate changes since the Last Glacial at Weinan on the southeastern Chinese Loess Plateau. Boreas 42, 815e829. Kaufman, D.S., Ager, T.A., Anderson, N.J., Anderson, J.T., Andrews, P.J., Bartlein, L.B., Coats, L.L., Cwynar, L.C., Duvall, M.L., Dyke, A.S., Edwards, M.E., Eisner, W.R., Gajewski, K., Geirsdottir, A., Hu, F.S., Jennings, A.E., Kaplan, M.R., Kerwin, M.W., Lozhkin, A.V., MacDonald, G.M., Miller, G.H., Mock, C.J., Oswald, W.W., OttoBliesner, B.L., Porinchu, D.F., Ruhland, K., Smol, J.P., Steig, E.J., Wolfe, B.B., 2004. Holocene thermal maximum in the western Arctic (0-180 W). Quat. Sci. Rev. 23 (5), 529e560. Koutavas, A., deMenocal, P.B., Olive, G.C., Lynch-Stieglitz, J., 2006. Mid-Holocene El Nino-Southern Oscillation (ENSO) attenuation revealed by individual foraminifera in eastern tropical Pacific sediments. Geology 34, 993e996. Kripalani, R.H., Oh, J.H., Chaudhari, H.S., 2007. Response of the East Asian summer monsoon to doubled atmospheric CO2: coupled climate model simulations and projections under IPCC AR4. Theor. Appl. Climatol. 87, 1e28. Kukala, G., An, Z.S., 1989. Loess stratigraphy in central China. Palaeogeogr. Palaeoclimatol. Palaeoeology 71, 203e225. Kutzbach, J., Gallimore, R.G., 1988. Sensitivity of a coupled atmosphere and mixed layer ocean model of changes in orbital forcing at ~9000 B.P. J. Geophys. Res. 93, 803e821. Kutzbach, J.E., 1981. Monsoon climate of the early Holocene: climate experiment with the Earth's orbital parameters for 9000 years ago. Science 214, 59e61. Li, B.C., Sun, J.Z., 2004. Vegetation and climate environment during the late Pleistocene in Loess Plateau, China. Geogr. Res. 23 (5), 641e648. Li, B.C., Sun, J.Z., 2005. Vegetation and climate environment during Holocene in Xi’an region of Loess Plateau, China. Mar. Geol. Quat. Geol. 25 (3), 125e132 (in Chinese). Li, Q., Wu, H.B., Yu, Y.Y., Sun, A.Z., Markovi
c, S.B., Guo, Z.T., 2014b. Reconstructed moisture evolution of the deserts in northern China since the last glacial maximum and its implications for the east Asian summer monsoon. Glob. Planet. Change 121, 101e112. Li, Q., Wu, H.B., Guo, Z.T., Yu, Y.Y., Ge, J.Y., Wu, J.Y., Zhao, D.A., Sun, A.Z., 2014a. Distribution and vegetation reconstruction of the deserts of northern China during the mid-Holocene. Geophys. Res. Lett. 41 http://dx.doi.org/10.1002/ 2014GL059952. Li, X.Q., Shang, X., Dodson, J., Zhou, X.Y., 2009. Holocene agriculture in the Guanzhong Basin in NW China indicated by pollen and charcoal evidence. Holocene 19, 1213e1220. Li, X.Q., Sun, N., Dodson, J., Zhou, X.Y., Zhao, K.L., 2013. Vegetation characteristics in the western Loess plateau between 5200 and 4300 cal. B.P. based on fossil charcoal records. Veg. Hist. Archaeobotany 22, 61e70. Li, X.Q., Zhou, J., Dodson, J., 2003. The vegetation characteristic of the “Yuan” area at Yaoxian on the Loess Plateau in China over the last 12,000 years. Rev. Palaeobot. Palynology 124, 1e7. Lin, B.H., Lui, R.M., An, Z.S., 1991. Preliminary research on stable isotopic composition of Chinese loess. In: Tungsheng, Liu (Ed.), Loess, Environment and Global Change. Science Press, Beijing, China, pp. 124e131 (in Chinese). Liu, T.S., 1985. Loess and the Environment. China Ocean Press, Beijing (in Chinese). Liu, T.S., Ding, Z.L., 1992. Stepwise coupling of monsoon circulation to continental ice volume variations during the past 2500,000 years. Quat. Sci. 1, 12e23 (in Chinese). Liu, T.S., Guo, Z.T., Wu, N.Q., Lu, H.Y., 1996. Prehistoric vegetation on the Loess Plateau: steppe or forest? J. Southeast Asian Earth Sci. 13 (3e5), 341e346. Liu, W.G., Yang, H., Sun, Y.B., Wang, X.L., 2011. d13C Values of loess total carbonate: a sensitive proxy for Asian summer monsoon in arid northwestern margin of the Chinese loess plateau. Chem. Geol. 284, 317e322. Liu, X.M., Liu, T.S., 1991. Chinese loess magnetic particle analysis and paleoclimatic significance. Sci. China Ser.B 6, 639e644 (in Chinese). Liu, Z.Y., Otto-Bliesner, B.L., He, F., Brady, E.C., Tomas, R., Clark, P.U., Carlson, A.E., Lynch- Stieglitz, J., Curry, W., Brook, E., Erickson, D., Jacob, R., Kutzbach, J., Cheng, J., 2009. Transient simulation of last deglaciation with a new mechanism for bolling-allerod warming. Science 325, 310e314. Liu, Z.Y., Wen, X.Y., Brady, E.C., Otto-Bliesner, B., Yu, G., Lu, H.Y., Cheng, H., Wang, Y.J., Zhang, W.P., Ding, Y.H., Edwards, R.L., Cheng, J., Liu, W., Yang, H., 2014. Chinese cave records and the east Asia summer monsoon. Quat. Sci. Rev. 83, 115e128. Lu, H.Y., Liu, T.S., Guo, Z.T., 2003. Natural vegetation of geological and historical periods in Loess Plateau. Chin. Sci. Bull. 48 (5), 411e416. Lu, H.Y., Liu, T.S., Wu, N.Q., Han, J.M., Guo, Z.T., 1999. Phytolith record of vegetation succession in the southern Loess Plateau since late Pleistocene. Quat. Sci. 4, 336e349 (in Chinese). Lu, H.Y., Wu, N.Q., Liu, T.S., Han, J.M., Qin, X.G., Sun, X.J., Wang, Y.J., 1996. Seasonal climatic variation recorded by phytolith assemblages from Baoji loess sequence in central China over the last 150000 a. Sci. China (Serise D) 26 (6), 629e639. Lu, H.Y., Yi, S.W., Liu, Z.Y., Mason, J.A., Jiang, D.B., Cheng, J., Stevens, T., Xu, Z.W., Zhang, E.L., Jin, L.Y., Zhang, Z.H., Guo, Z.T., Wang, Y., Otto-Bliesner, B., 2013. Variation of East Asian monsoon precipitation during the past 21 ky and potential CO2forcing. Geology 41, 1023e1026. Lüthi, D., Floch, M.L., Bereiter, B., Blunier, T., Barnola, J., Siegenthaler, U., Raynaud, D., Jouzel, J., Fischer, H., Kawamura, K., Stocker, T.F., 2008. High-resolution carbon dioxide concentration record 650,000e800,000 years before present. Nature 453, 379e382. Maher, B.A., 2016. Palaeoclimatic records of the loess/palaeosol sequences of the Chinese Loess Plateau. Quat. Sci. Rev. 154, 23e84.

Marcott, S.A., Shakun, J.D., Clark, P.U., Mix, A.C., 2013. A reconstruction of regional and global temperature for the past 11,300 years. Science 339 (6124), 1198e1201. MCQPDB (Members of China Quaternary Pollen Data Base), 2000. Pollen-based biome reconstruction at middle Holocene (6 ka B.P.) and Last Glacial Maximum (18 ka BP) in China. Acta Bot. Sin. 42 (11), 1201e1209 (in Chinese). Normile, D., 2007. Getting at the roots of killer dust storms. Science 317, 314e316. €hler, P., Raynaud, D., Paillard, D., Schwander, J., Parrenin, F., Masson-Delmotte, V., Ko Barbante, C., Landais, A., Wegner, A., Jouzel, J., 2013. Synchronous change of atmospheric CO2 and Antarctic temperature during the last deglacial warming. Science 339, 1060e1063. Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Monserud, R.A., Solomon, A.M., 1992. A global biome model based on plant physiology and dominance, soil properties and climate. J. Biogeogr. 19, 117e134. Peterse, F., Prins, M.A., Beets, C.J., Troelstra, S.R., Zheng, H.R., Gu, Z.Y., Schouten, S., Damste, J.S.S., 2011. Decoupled warming and monsoon precipitation in East Asia over the last deglaciation. Earth Planet. Sci. Lett. 301, 256e264. Porter, S.C., 2001. Chinese loess record of monsoon climate during the last glacialeinterglacial cycle. Earth-Science Rev. 54, 115e128. Porter, S.C., An, Z.S., 1995. Correlation between climate events in the North Atlantic and China during the last glaciation. Nature 375, 305e308. Prentice, I.C., Guiot, J., Huntley, B., Jolly, D., Cheddadi, R., 1996. Reconstructing biomes from palaeoecological data: a general method and its application to European pollen data at 0 and 6 ka. Clim. Dyn. 12, 185e194. Shakun, J.D., Clark, P.U., Marcott, S.A., Mix, A.C., Liu, Z.Y., Otto-Bliesner, B., Schmittner, A., Bard, E., 2012. Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation. Nature 484, 49e54. Shang, X., Li, X.Q., 2010. Holocene vegetation characteristics of the southern loess plateau in the Weihe River valley in China. Rev. Palaeobot. Palynology 160, 46e52. Shi, N.H., 1981. Heshan Ji (Part 2). Sanlian Bookstore, Beijing (in Chinese). Shi, N.H., 1991. Comments on the historical distribution and vicissitudes of the national vegetation of China. Collect. Essays Chin. Hist. Geogr. 3, 43e73 (in Chinese). Shi, Y.F., Kong, Z.C., Wang, S.M., Tang, L.Y., Wang, F.B., Yao, T.D., Zhao, X.T., Zhang, P.Y., Shi, S.H., 1992. The important climatic fluctuations and events in China during the Holocene Megathermal. Sci. China (Series B) 12, 1300e1308 (in Chinese). Stevens, T., Thomas, D.S., Armitage, S.J., Lunn, H.R., Lu, H., 2007. Reinterpreting climate proxy records from late Quaternary Chinese loess: a detailed OSL investigation. Earth-Science Rev. 80, 111e136. Sun, A.Z., Feng, Z.D., 2013. Holocene climatic reconstructions from the fossil pollen record at Qigai Nuur in the southern Mongolian Plateau. Holocene 23 (10), 1391e1402. Sun, A.Z., Feng, Z.D., Ma, Y.Z., 2010. Vegetation and environmental changes in western Chinese Loess Plateau since 13.0 ka BP. J. Geogr. Sci. 20 (2), 177e192. Sun, J.M., Li, S.H., Han, P., Chen, Y.Y., 2006. Holocene environmental changes in the central Inner Mongolia, based on single-aliquot-quartz optical dating and multi-proxy study of dune sands. Palaeogeogr. Palaeoclimatol. Palaeoecol. 233, 51e62. Sun, A.Z., Ma, Y.Z., Feng, Z.D., Li, F., Wu, H.N., 2007. Pollen-recorded climate changes between 13.0 and 7.0 14C ka BP in southern Ningxia, China. Chin. Sci. Bull. 52 (8), 1080e1088. Sun, X.J., Song, C.Q., Wang, B.Y., Sun, M.R., 1996. Vegetation history of the southern Loess Plateau of China during the last 100000 years based on pollen data. Acta Bot. Sin. 38 (12), 982e988. Sun, X.J., Song, C.Q., Wang, F.Y., Sun, M.R., 1997. Vegetation history of the Loess Plateau of China during the last 100,000 years based on pollen data. Quat. Int. 37, 25e36. Sun, X.W., Xia, Z.K., 2005. Paleoenvironment changes since mid-Holocene revealed by a palynological sequence from Sihenan profile in Luoyang, Henan Province. Acta Scicentiarum Nauralum Universitis Pekinesis 41 (2), 289e294 (in Chinese). Tang, L.Y., An, C.B., 2007. Pollen records of Holocene vegetation and climate changes in the Longzhong Basin of the Chinese Loess Plateau. Progr. Nat. Sci. 17 (12), 1445e1456 (in Chinese). Wang, S.L., Chen, F.H., Cao, J.X., 1991. The pollen assemblage and climatic fluctuations in Lanzhou river basin since 10000 a B.P. J. Glaciol. Geocryol. 13 (4), 307e314 (in Chinese). Wan, L., Zhang, X.P., Ma, Q., Zhang, J.J., Ma, T.Y., Sun, Y.P., 2014. Spatiotemporal characteristics of precipitation and extreme events on the Loess Plateau of China between 1957 and 2009. Hydrol. Process 28, 4971e4983. Wang, G., Innes, J.L., Lei, J., Dai, S., Wu, S.W., 2007. China's forestry reforms. Science 318, 1556e1557. Wang, H.P., Chen, J.H., Zhang, X.J., Chen, F.H., 2014. Palaeosol development in the Chinese Loess Plateau as an indicator of the strength of the East Asian summer monsoon: evidence for a mid-Holocene maximum. Quat. Int. 334e335, 155e164. Wang, S.C., 1990. On ancient loess plateau vegetation. Geogr. Res. 9 (4), 72e79 (in Chinese). Wang, Z.Q., Liu, B.Y., Liu, G., 2009. Soil water depleting depth by planted vegetation on the Loess Plateau. Sci. China (D) 52, 835e842. €ris, O., Danzeglocke, U., 2007. Cologne Radiocarbon Calibration & Weninger, B., Jo Paleoclimate Research Package CalPal. Available at: http://www.calpal.de/. Xiao, J.L., Nakamura, T., Lu, H.Y., Zhang, G.Y., 2002. Holocene climate changes over the desert/loess transition of north-central China. Earth Planet. Sci. Lett. 197, 11e18.

A. Sun et al. / Quaternary Science Reviews 173 (2017) 236e247 Xie, S.C., Wang, Z.Y., Wang, H.M., Chen, F.H., An, C.B., 2002. The occurrence of a grassy vegetation over the Chinese Loess Plateau since the last interglacier: the molecular fossil record. Sci. China 45 (1), 53e62. Xu, J., 2006. The Sequence of Vegetation Change by Pollen Records of Xifeng Area during the Holocene. Master degree dissertation (in Chinese). Yang, S.L., Ding, Z.L., 2008. Advance-retreat history of the East-Asian summer monsoon rainfall belt over Northern China during the last two glacialinterglacial cycles. Earth Planet. Sci. Lett. 274, 499e510. Yang, S.L., Ding, Z.L., Li, Y.Y., Wang, X., Jiang, W.Y., Huang, X.F., 2015a. Warminginduced northwestward migration of the East Asian monsoon rain belt from the Last Glacial Maximum to the mid-Holocene. Proc. Natl. Acad. Sci. U. S. A. 112 (43), 13178e13183. Yang, S.L., Ding, Z.L., Wang, X., Tang, Z.H., Gu, Z.Y., 2012. Negative d18O-d13C relationship of pedogenic carbonate from northern China indicates a strong response of C3/C4 biomass to the seasonality of Asian monsoon precipitation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 317e318, 32e40. Yang, X.X., Jiang, W.Y., Yang, S.L., Kong, Z.C., Luo, Y.L., 2015b. Vegetation and climate changes in the western Chinese loess plateau since the last glacial maximum. Quat. Int. 372, 58e65. Yao, Z., Wu, H., Liang, M., Shi, X., 2011. Spatial and temporal variations in C3 and C4 plant abundance over the Chinese Loess Plateau since the last glacial maximum. J. Arid Environ. 75, 881e889. Yu, Y.Y., Guo, Z.T., Wu, H.B., Finke, P.A., 2012. Reconstructing prehistoric land use change from archeological data: validation and application of a new model in

247

Yiluo valley, northern China. Agric. Ecosyst. Environ. 156, 99e107. Yu, Y.Y., Wu, H.B., Finke, P.A., Guo, Z.T., 2016. Spatial and temporal changes of prehistoric human land use in the Wei River valley, northern China. Holocene. http://dx.doi.org/10.1177/0959683616645943. Zhang, X.B., An, Z.S., 1994. Relationship between forests and loess thicknesses in the Loess Plateauregion. Bull. Soil Water Conservation 14 (6), 1e4 (in Chinese). Zhang, X.S., 2007. Vegetation Map of the People's Republic of China (1:100000). Geological Publishing House, Beijing (in Chinese). Zhao, Y., Yu, Z.C., Zhao, W.W., 2011. Holocene vegetation and climate histories in the eastern Tibetan Plateau: controls by insolation-driven temperature or monsoon-derived precipitation changes? Quat. Sci. Rev. 30, 1173e1184. Zhou, B., Shen, C.D., Zheng, H.B., Zhao, M.X., Sun, Y.M., 2009. Vegetation evolution on the central Chinese Loess Plateau since late Quaternary evidenced by elemental carbon isotopic composition. Chin. Sci. Bull. 54 http://dx.doi.org/ 10.1007/s11434-009-0084-8. Zhou, X.Y., Li, X.Q., Dodson, J., Yang, S.l., Long, H., Zhao, K.L., Sun, N., Yang, Q., Liu, H.B., Zhao, C., 2014. Zonal vegetation change in the Chinese Loess Plateau since MIS3. Palaeogeogr. Palaeoclimatol. Palaeoecol. 404, 89e96. Zhu, Z.C., 1983. The range of forest-steppe zone on the Loess Plateau of northern part in Shaanxi Province. Acta Phytoecol. Geobot. Sin. 7 (2), 122e131 (in Chinese). Zhu, Z.C., 1994. The basic characteristic of forest-grass in the Loess Plateau. Sci. Geogr. Sin. 14 (2), 152e156 (in Chinese).