Late-Holocene vegetation change reveals the environment of ancient people and the origin of Huashan cliff paintings in Guangxi,southwestern China

Abstract

Human activities in the historical period were determined by the natural environment and social characteristics. The Zuojiang Huashan Rock Art Cultural Landscape in Guangxi is the most important World Heritage for prehistorical culture in southwestern China. However, knowledge about the natural environment and associated historic human–climate interactions during the period of cliff painting creation remains scarce, which has been an obstacle to understanding the process of regional civilization. Here, we report a pollen record from the core distributing region of Huashan cliff paintings to interpret the natural environment and climate changes, and further to discuss how and why ancient people create these giant cliff paintings based on the relationships of human–climate interactions. Pollen results imply an existence of a mixed broad-leaf and conifer forest in the woodland landscape at Huashan region before 1860 cal. yr BP, after which pine forests were more prevalent. During the cliff painting period from 2370 to 2115 cal. yr BP, broad-leaf and conifer trees dominated the tree layer and provided the key source materials for painting. It can be reasonably assumed that people used stake or ladder to climb the cliffs and created the cliff paintings during a humid climate condition from 2560 to 2000 cal. yr BP. Furthermore, the humid climate during 2700 and 2400 cal. yr BP had been interrupted by short-term drought episodes. These turbulent climate episodes probably played the role of disasters and impacted on the human activities, leading to the social–political tensions and the out-break of war. During the Qin and Han Dynasties, social–political changes relieved the pressure from natural climate changes and led to a more stable society, in which the creating of cliff paintings was gradually fading. Our results manifest the necessity to analyze the natural environment and social–political background in combination in this study on ancient human activities.

Keywords

creating motivation Dahu Lake Huashan cliff painting human–climate interaction living environment pollen analysis

Introduction

The Huashan cliff paintings are widely distributed in the watersheds of Zuojiang River and its tributary Mingjiang River in Chongzuo region, mainly in four counties of Ningming, Longzhou, Jiangzhou, and Fusui in the southwest of Guangxi Zhuang Autonomous Region, southwestern China, where is a border region between China and Vietnam (Figure 1). These red ocher/auburn cliff paintings, located in sheer cliffs of river bends, illustrated the life and sacrificial rituals of the ancient Luoyue people in more than 2000 years ago during the 5th century BC to the 2nd century AD (UNESCO, 2017). The Zuojiang Huashan Rock Art Cultural Landscape, including the best preserved 38 sites of rock art and their associated karst, riverine, and tableland landscape, has become the World Heritage in 2016. Such pictographs depicting drums and related elements are symbolic records directly associated with the bronze drum culture once widespread in southern China (UNESCO, 2017).

Figure 1.

(a) Locations of main sites of the Huashan cliff painting (UNESCO, 2017) and the Dahu Lake and (b) locations of Huashan cliff painting in Guangxi, southwestern China.

Many past and recent studies have definitely confirmed that the cliff paintings were painted by the dyestuff made by the hematite (China National Institute of Cultural Property (CNICP), 1986; Long, 1986; Qiu et al., 1990; Sun et al., 2012; Tang, 1996). The binding material of the painting to the limestone, which is widely distributed in the tropical and subtropical karst terrain of SW China, was thought in the 1980s to be a mixture of animal milk, blood, and skin (Long, 1986; Qin, 1987), or a kind of plant material, such as the lignin coming from the coniferyl alcohol of trees (Qiu et al., 1990). However, new techniques such as the 3D microscope, infrared and micro-Raman spectroscopy, and gas chromatography–mass spectrometry have revealed that the binding material should be mixed with the plant sap (Guo et al., 2005), and have recently confirmed that it was consist of whewellite (Tang, 1996) or calcium oxalate (Sun et al., 2012), all from plants. This inferred that the binding material should be the oxalic acid (H₂C₂O₄) from plant flowers, fruits, and stems, which can chemically react with the calcium carbonate (CaCO₃) from limestone to produce calcium oxalate (CaC₂O₄) to cement the dyestuff with the rock (Sun et al., 2012). However, it was unknown that from which plants the oxalic acid came from.

Calibrated ¹⁴C dating of three stalactite samples showed that the cliff paintings were completed in 2370 to 2115 cal. yr BP (Yuan et al., 1986). The MC–ICPMS measurements of 40 stalactite samples indicated that the cliff paintings were done between 1921 and 1793 cal. yr BP, but the early paintings started likely around 2668 cal. yr BP (Shao, 2016). The UNESCO (2017) report shows that it was from 2425 to 1730 cal. yr BP covering the Warring Dynasty (475–221 BC) to the East Han Dynasty (AD 9–220). These three chronologies inferred a coincident time period. During this old time period, however, what kinds of living environments for the ancient people existed? What kinds of vegetation and climate occurred? What kinds of plants, gymnosperms or angiosperms, trees or herbs, the ancient people utilized for making the binding material? Why and how they created this giant cliff paintings? All of these questions have so far not been fully answered.

Pollen record from lake sediments is an effective proxy to indicate past environmental changes such as the vegetation and climates (Lu et al., 2019; Ni et al., 2014; Prentice et al., 1991; Xiao et al., 2014). In this study, we use pollen and grain size data from a lake in the distributional area of Huashan cliff paintings to explore the paleoenvironmental change during the late Holocene, and attempt to reveal the source of making the blinding material, the way of painting on cliffs, and to answer why ancient people created these cliff paintings.

Study area

Dahu Lake (means ‘big lake’ in English) is administratively located in the north of Ningming County, southwestern Guangxi Zhuang Autonomous Region. It is actually in the central point among three counties of Ningming, Longzhou, and Chongzuo (Figure 1). The region is a northern tropical exposed karst region featured by a plurality of mountain peaks and valleys. It is significantly affected by the tropical monsoon climate. Meteorological records from two weather stations, Longzhou (22°20′N, 106°51′E, 128.8 m a.s.l., 1953–2014) and Ningming (22°07′N, 107°04′E, 162.6 m a.s.l., 1956–2017), show that the mean annual temperature is 22.2°C, and the mean temperature of January is 13.8°C and of July is 28.2°C, respectively. The mean annual precipitation is 1226.2 mm (690–1879 mm), along with 80% falling between April and September. The relative humidity is 79.8%. The Dahu Lake, mainly fed by rainfall, is usually dry up from December to next February and has a large lake area of ca. 0.5 km² with the water depth of 3 to 5 m during the rainy season. The lake discharges into the Zuojiang River through a big sinkhole.

Zonal vegetation in the southwestern Guangxi is dominated by the northern tropical karst seasonal rainforest, which now displays patchy distribution (Editorial Committee of Vegetation Map of the People’s Republic of China (ECVMP), Chinese Academy of Sciences, 2007). It is obvious that the vegetation has been disturbed by anthropogenic activities leading to the domination of secondary forest, shrub and grassland, as well as cultivated vegetation (Figure 2). The vegetation around the Dahu Lake is tropical and subtropical shrubs, tropical monsoonal rainforests and cultivated vegetation (Figure 2c). The tropical and subtropical shrubs, comprising Dracaena cambodiana and Psidium guajava, appears around the lake. The tropical monsoonal rainforest is on the east of the lake, mainly composed of Bombax ceiba and Albizia chinensis. Cultivated vegetation, the sugarcane field, lays north and east of the lake. Tropical rainforests (Burretiodendron hsienmu, Garcinia paucinervis, and Cephalomappa sinensis), subtropical coniferous forests (mainly Pinus massoniana) and tropical and subtropical grasslands (Miscanthus, Arundinella anomala, and Arundinella setosa) at flood plains are also widely distributed around the study area. During the field investigation, it can be found that the lake faces the shrubland and little tree forest on hills in the west, mainly comprising Eucalyptus, Celtis, Melia, Broussonetia, Prunus, Caesalpinia, and Alangium, and the sugarcane field in the east. Vegetation on the lakebed that has dried up is the grasses-dominated meadow-like plant community (Figure 2).

Figure 2.

(a) The sampling site, (b) the sediment profile of Dahu Lake, and (c) modern vegetation map (ECVMP, Chinese Academy of Sciences, 2007) around the lake.

Regional vegetation in southern China has changed greatly during the Holocene. Evidences of palynological assemblages from the southern China recorded that subtropical monsoonal rainforest and subtropical evergreen broad-leaf forest, mainly comprising Castanopsis, Ficus, evergreen Quercus, Lithocarpus, Palmae, and ferns, distributed at lowland areas during the mid-Holocene (Wang et al., 2007; Zheng et al., 2003; Zhouet al., 2004, 2015). Those tropical and subtropical evergreen taxa together with mangrove pollen are abundant in the northern Vietnam (Li et al., 2006). Because of the climate changes and effects of human activities, the evergreen broad-leaf forest and subtropical coniferous forest were prevail during the late Holocene, accompanied with the decrease of tropical tree taxa such as Castanopsis, but with the increases of Pinus, herbaceous plants, and some pioneer ferns such as Dicranopteris (Li et al., 2006; Wang et al., 2007; Zheng et al., 2003; Zhou et al., 2004, 2015). In the northern Guangxi region, subtropical evergreen forest mainly comprising evergreen Quercus dominated during 5390 to 1990 cal. Yr BP, after which Poaceae and Pinus increased obviously (Zhou et al., 2015). In the middle Guangxi region, the shrink of Castanopsis, Quercus, and Altingia and the expansion of coniferous taxa, herbaceous taxa and pioneer ferns manifest that zonal vegetation undergone dramatic anthropogenic disturbances during the past 1600 years (Zhao, 2013). In the northern Vietnam, the increase of wild and upland herbs along with the decrease of tropical and subtropical taxa also indicates an intensified human activity in the late Holocene (Li et al., 2006, 2009).

Guangxi, the birthplace of Zhuang culture, is characterized by a continuous cultural sequence from Paleolithic Age to late Neolithic Age (Guangxi Institute of Cultural Relics and Archaeology (GICRA), 2010; Jiang, 1998). The Neolithic Culture mainly occurred at the southern Guangxi during 10 to 4 ka BP, including the important archeological sites of Dingsishan, Hecun, Gantuoyan, and Dalongtan (GICRA, 2010; Jiang, 1998; Wang et al., 2014). Abundant chipped stones and animal remains from the cultural layers at above sites, mostly along the Zuojiang, Youjiang, and Yongjiang Rivers, indicated that ancient people from this region was reliant on gathering, fishing, and hunting in the early and middle Neolithic at 10 to 4.5 ka BP (Chi and Hung, 2015; Fu et al., 1998; GICRA, 2010; He et al., 2009). The earliest agricultural evidence has been recorded from the rice phytoliths during 4.5 to 4 ka BP (Chi and Hung, 2015; Fu et al., 1998; Zhao et al., 2005). Later around 3.8 to 2.8 ka BP, contemporary with the late Shang Dynasty, the carbonized rice, millet grains, shipped stones, pottery, and animal remains have indicated the coexistence of both agriculture and gathering–hunting (GICRA, 2010; Jiang, 1998). The Bronze Culture in Guangxi originated in the late Shang and early Zhou Dynasties, and reached its peak during the Warring Dynasty to Han Dynasty (Jiang, 1998; Zheng, 2004). The abundant bronze weapons discovered from the burials sites in southern Guangxi reflected a conflict history during this period (GICRA, 2010; Zheng, 2004).

Documentations suggested that strong influence of human activities in Guangxi started from Qin (221–207 BC) and Han Dynasties (202 BC–AD 220) when population from northern China migrated here significantly (Li, 1996). At 214 BC, the Qin Dynasty set up three counties of Nanhai, Xiang, and Guilin to govern this region, after a 5-year war between the army of Qin and the local Baiyue people (Huang, 2015). Since then, social form and ruling system from the central China gradually infiltrated and dominated Guangxi region, although with some short-term warfare before the East Han Dynasty (Huang, 2015; Li, 1996).

Materials and methods

Sampling

A 200-cm long sediment profile (22°19′10.58N, 107°07′06.52″E, 134 m a.s.l.) was excavated in March 2015 from the dried-up bottom of the Dahu Lake, with approximately 20 m distance from the sinkhole (Figure 2a). The lake sediment (Figure 2b) consists of black silty sand with abundant herbaceous roots in depth of 0 to 6 cm, yellow silty sand along with small shells in depth of 6 to 67 cm, brown–yellow clay loam in depth of 67 to 93 cm, brown clay sand in depth of 93 to 150 cm, dark brown clay sand in depth of 150 to 160 cm, dark brown clay in depth of 160 to 194 cm and dark brown clay with coarse sand in depth of 194 to 200 cm. The sediment at the top 67 cm is very different from other layers in color (light yellow) and texture (sandy). It was told by the local farmers that this was because of the mining activity in the upper mountains around the lake. This layer is therefore unusable. In total, 100 samples at a 2-cm interval were taken in the field and transferred to the laboratory for further analyses.

Dating

Five sediment samples, at depth of 6 to 8 cm, 68 to 70 cm, 100 to 102 cm, 158 to 160 cm, and 198 to 200 cm, respectively, were chosen for the accelerator mass spectrometry (AMS) ¹⁴C dating based on the total organic carbon in the Beta Radiocarbon Laboratory (Beta Analytic Inc., Miami, USA). The calendar ages were calibrated using the CALIB 6.0 with an IntCal13 calibration dataset (Reimer et al., 2013). Age–depth model was constructed with the Bacon 2.2 work package in R 3.2 (R Core Team, 2016) using the Bayesian statistics to reconstruct the accumulation history of the deposition (Figure 3). The model was performed using the default settings for lake sediments with 2-cm resolution.

Figure 3.

Age–depth model calibrated by the Bayesian statistics.

Pollen analysis

The 66 samples for pollen analysis were treated following the standard laboratory method (Faegri et al., 1989), including sieving to remove the fine and coarse fractions, treatment with HCl (10%) and HF (50%) to remove carbonate and silicate, boiling in KOH (10%) to remove humic acid, and mounting in silicone oil. To calculate concentrations of pollen, tablets containing a known quantity of Lycopodium spores (ca. 18,583 grains per tablet) were added to each sample. At least, 194 terrestrial pollen and spore grains per sample were counted. The pollen percentages of every terrestrial specie were calculated based on the total grains of terrestrial pollen, excluding the fern and aquatic herb. But the pollen percentages of terrestrial tree, shrub and herb, as well as fern and aquatic herb were calculated based on the total grains of pollen and spore. Pollen diagram was divided into biostratigraphic zones based on the constrained incremental sum of squares (CONISS) using the Tilia 2.0.60 program (Grimm, 1987, 1993).

Grain size analysis

In total, 66 samples from the depth of 68 to 200 cm were pre-treated with 10% H₂O₂ and then 10% HCl, over 24 h, respectively, to remove organism and carbonate. Approximately, 10 ml of 10% Na₂P₂O₇ was then added to the remaining sample, and kept in the sonic oscillation for 15 min. The grain size composition was determined using a Malvern Instruments Mastersizer-2000 laser diffraction particle size analyzer (Malvern Instruments Ltd., UK).

Numerical analyses

Only terrestrial pollen taxa accounting for >1% of the total terrestrial pollen in at least three samples were included in the numeric analyses. The pollen data were square-root transformed at first to stabilize variances and optimize the signal-to-noise ratio (Prentice, 1980). Detrended correspondence analysis (DCA) had been first applied and yielded a gradient of 1.519 standard deviations for selected pollen data, indicating that linear-based methods are appropriate for this data set. Redundancy analysis (RDA) had been applied to extract and summarize the variation in our pollen assemblages that can be explained by the selected environmental variables. Three environmental variables, that is, the fine silt fraction (2–20 µm) in the Dahu Lake profile representing the local hydrological condition, the stalagmite δ¹⁸O from Dongge Cave on behalf of monsoonal variability (Dykoski et al., 2005), and the summed probability distribution of calibrated radiocarbon dates (sPDC in the following text) in Guangxi reflecting the intensity of human activities (Wang et al., 2014), were selected to perform the RDA with the R package ‘vegan’ version 2.5-6 (Oksanen et al., 2019). The stalagmite δ¹⁸O data were linear interpolated and then rarefied to match the age of fine-silt fraction. Constrained clustering was further applied to partition the pollen data according to the environmental variables using the multivariate regression tree (MRT) analysis with the R package ‘mvpart’ version 1.2-6 (De’ath, 2007). The determination of tree size was guided by cross-validation. To focus our research questions and to avoid overinterpreting data, both the RDA and the MRT analysis only focused on lower part of Dahu Lake profile deposited before 1860 cal. yr BP.

Results

Chronology

The AMS ¹⁴C results (Table 1) show that sediments in the upper 67 cm are obviously exceptional young, confirming that the upper sediments are related to the modern human disturbance, that is, the mining activity in the upper reaches of the watershed. The modern sediment has therefore been proved to be at the depth of 68 to 70 cm. In combination with the lithology, the profile from 68 to 200 cm should be continuous and can be used for further investigations.

Table 1.

AMS ¹⁴C dates of the Dahu Lake sediments.

Sam. ID	Depth (cm)	δ¹³C (‰)	¹⁴C age (yr BP)	2 sigma calibration (cal. yr BP)	BC/AD
NM-6	6–8	–22.8	2070 ± 30	2046 ± 78	174–18 BC
NM-68	68–70	–21.3	106.4 ± 0.3 pMC	–	–
NM-100	100–102	–19.2	1590 ± 30	1477 ± 68	AD 405–541
NM-158	158–160	–17.1	2180 ± 30	2214 ± 95	360–169 BC
NM-198	198–200	–17.8	2480 ± 30	2577 ± 145	773–482 BC

pMC: percent modern carbon.

Dating material is organic sediment. The reported result indicates an age of post 0 BP and has been reported as a percentage of the modern reference standard, indicating the material was living about the last 60 years or so.

According to the age–depth model, the basal sediment of Dahu Lake profile was deposited from 2806 to 2580 cal. yr BP, with a median age of 2699 cal. yr BP (Figure 3). Although the chronology is of low resolution, the 2σ uncertainty was quite homogeneous, of ca. ± 147.9 year in average from 2699 to 1420 cal. yr BP. The profile was characterized by a relatively high sedimentation rate of 0.8 cm per decade during this period. After 1420 cal. yr BP, the 2σ uncertainty in age–depth model increased to ca. ± 324.3 year in average. And the sedimentation rate decreased to 0.2 cm per decade. For the convenience of discussion, the median age in the model has been selected to interpret our results. But one must keep in mind that the age-depth model is only based on three anchor points and the age is with variable error ranges (Figure 3).

Pollen assemblages

A total of 109 pollen and spores have been identified from the sediment, including 61 arboreal taxa, 22 herbaceous taxa, 6 aquatic herbaceous taxa, and 20 fern taxa, as well as 5 algae (Figure 4). Pollen concentration exhibits obvious range from 15,940 to 1,484,000 grains/g. Fern spores mainly comprising Diplopterygium, Lygodium, and Pteris, dominate the profile, followed by the coniferous pollen, terrestrial herb, deciduous tree, shrubs, evergreen tree, and aquatic herb. In arboreal pollen, Pinus, Podocarpus, Cupressaceae, evergreen Quercus, Ginko, deciduous Quercus, and Juglans are abundant. Rosaceae and Lonicera are dominant in shrubs. Terrestrial herb mainly comprises Amaranthaceae, Poaceae, Artemisia, Polygonum, and Asteraceae. Algae mainly consist of Zygnma, Pediasyrum, Concentricystes, and Mougeotia. According to the pollen composition variations, pollen spectra can subdivide into four zones (Figure 5).

Figure 4.

The pollen composition changes in Dahu Lake profile. The pollen and spore percentages were calculated based on the total sum of pollen and spores. The hollow lines representing the magnification of 10 times. Light blue area representing the periods of cliff paintings (UNESCO, 2017), while the pink-filled area implying the radiocarbon date from the cliff paintings (Yuan et al., 1986).

Figure 5.

Diagram showing the selected pollen taxa and grain size compositions in Dahu Lake profile. The terrestrial pollen percentages were calculated based on the total sum of terrestrial pollen. The spore percentages were calculated based on the total sum of pollen and spores. The hollow lines representing the magnification of 10 times. Light blue area representing the periods of cliff paintings (UNESCO, 2017), while the pink-filled area implying the radiocarbon date from the cliff paintings (Yuan et al., 1986).

Zone I-1 (200–186 cm, 2699–2560 cal. yr BP)

At this zone, pollen concentration was the least (only 31,470 grains/g in average). Trees and shrubs were less abundant, but it was rich in herbs, and especially ferns. Pinus, Magnolia, and evergreen Quercus, and deciduous Quercus and Ginkgo were dominant the coniferous plant, evergreen tree, and deciduous tree, respectively. The terrestrial herbaceous plants, comprising Amaranthaceae, Artemisia, Polygonum, Poaceae, and Asteraceae, reached the highest content in the whole profile and continuously occurred during this period. The ferns that with of Hicriopteris, Lygodium, Pteris, and Osmunda dominated the vegetation. The abundant forbs and less woody plants indicated that vegetation during this period was meadow or moist grassland. Sparse pines, oaks, and Ginkgo grew here or in the surrounding region. The absence of algae but relative high content of Sparganium implied a non-aquatic environment in the sampling site.

Zone I-2 (186–130 cm, 2560–1860 cal. yr BP)

Pollen concentration increased to an average of 61,580 grains/g. Coniferous tree showed slight increase in this zone, mainly contributed by the increased Pinus, Podocarpus, and Cupressaceae. Broad-leaf trees remained stable, characterized by decreasing in Magnolia, Ginkgo, and deciduous Quercus, whereas increasing in evergreen Quercus, Juglans, Carya, Liquidambar, and Carpinus. Shrubs increased as well, mainly comprising Rosaceae and Lonicera. Terrestrial herbs decreased, with dramatically reduced Poaceae, Polygonum, Artemisia, and Amaranthaceae. The ferns reached the maximum value around 2100 cal. yr BP. Diplopterygium, Lygodium, Alsophia, and Polypodium all increased. Aquatic herbs decreased but floating herbs such as Nelumbo and Trapa appeared in this zone. Algae increased at the same time. Pollen spectra changes in this zone indicate an expansion of arboreal forest.

Zone II-1 (130–96 cm, 1860–1420 cal. yr BP)

In this zone, pollen concentration increased to an average of 121,120 grains/g, with increasing coniferous trees and decreasing ferns. Pinus increased obviously and contributed most to the coniferous expansion. Other coniferous trees decreased, such as Podocarpus and Cupressaceae. Broad-leaf trees, shrubs, and herbs fluctuated and had no obvious changes. Hicriopteris and Lygodium still dominated the understory, but the decreases in Pteris, Alsophila, and other ferns were obvious. Aquatic plants were scarce, but Zygnma and Mougeotia were rich at some time slices. The increased aboral pollen probably indicated an expansion of subtropical forest in a gradually cooling climate. At the same time, the absence of aquatic herbs and decreased ferns probably implied a drought condition.

Zone II-2 (96–68 cm, 1420 cal. yr BP to present)

The concentration of pollen and spore kept increasing to 290,000 grains/g in average. Coniferous trees increased significantly, whereas ferns decreased gradually. Pinus pollen reached its highest percentage of ca. 90.37% among the terrestrial pollens in recent samples. Podocarpus and Cunninghamia also showed increasing trends. Deciduous trees mainly consisted by Ginkgo and Juglans, evergreen trees mainly comprising evergreen Quercus, and shrubs fluctuated in this zone but decreased significantly at the topmost samples. Amaranthaceae and Poaceae continuously occurred during this zone. Diplopterygium decreased with fluctuations. The decreasing trend in ferns and increasing in coniferous trees suggest a replacement of lowland forest by secondary pine forest. Zygnma, Concentricystes, and Pediastrum had higher concentration in this period than in any previous periods.

Grain size compositions

The sediments of Dahu Lake mainly consist of coarse silt (20–64 µm) and sand (>64 µm) ranging from 34.7% to 94.4%, then fine silt (2–20 µm) and clay (<2 µm) ranging from 5.6% to 65.3% (Figure 5). Coarse fractions were dominant, fluctuated but with slight decreasing trend during 2699 and 2000 cal. yr BP. On the contrary, fine silt was characterized by an increasing trend at the same time. It is noticeable that the coarse fractions peaked around 2650, 2500, and 2420 cal. yr BP, after which they remained in a relatively stable state until 2000 cal. yr BP. After 2000 cal. yr BP, coarse silt and sand fractions increased significantly and peaked with a value of 94.4% around 1950 cal. yr BP, and fine fractions reached the minimum value of 5.6% simultaneously. Coarse silt and sand fraction remained in high content during 1950 and 1800 cal. yr BP, after which both the fine and coarse fractions showed remarkable fluctuations.

Numeric analyses

The first two axes of RDA only capture about 7.6% (axis 1: 5.3%; axis 2: 2.3%) of the total variance, and only 8.9% constrained variance in total variance has been explained by the three environmental variables, indicating that the pollen spectra of Dahu Lake profile are impacted by several potential variables (Table S1 and Figure S1, available online). The forward selection option within RDA indicated that only the sPDC was above the significant level (p < 0.05). The MRT analysis indicated a four-node tree constrained by the sPDC and stalagmite δ¹⁸O (Table S2, available online and Figure 6). The first two nodes are characterized by higher sPDC and varied δ¹⁸O values, indicting the joint influence of human activity and climate change on pollen spectra. The second two nodes mainly split by the sPDC imply the contribution of human activity to pollen spectra. Pinus is always dominant in the four histograms showing the relative abundance of 20 pollen taxa at tip of the branches. Considering that the tree only explains 17% of the total species variance and the cross-validated error is relatively high (0.99), the environmental thresholds may be not with ecological significance. But the result still stressed the contribution of human activity on the pollen spectra.

Figure 6.

MRT analysis of the pollen data explained by their environmental variables. The relative abundances of the 20 pollen taxa are shown in histograms at the tips of the branches, with the taxa in the same order as in the legend. Under each histogram, n is the number of samples in the pollen group.

Discussion

The distance between Dahu Lake and the best preserved Ningming Huashan cliff paintings ranges from 7 km to 15 km, and the far most painting site of Wanrendongshan is about 75 km away from the lake (Figure 1). Therefore, it is reasonable to consider that they underwent most similar vegetation and climate changes during the painting period. Multiple evidences have shown that the upper and lower limits of the reliable cliff painting date were between 2370 and 2135 cal. yr BP, that is, the cliff paintings in the Zuojiang River watershed were made between the early Warring and East Han Dynasties (Qin, 1987; Shao, 2016; UNESCO, 2017; Yuan et al., 1986). The vegetation zone I-2 in the depth of 198 to 130 cm in Dahu Lake profile, deposited from 2560 to 1860 cal. yr BP, exactly covers the period of cliff painting created, and further can be a robust indicator of past environmental change in this region.

Past vegetation change and the source of painting materials

From the entire profile point of view, the vegetation surrounding the Dahu Lake during the past 2700 years was a mixed broad-leaf and conifer forest, with spares firs, cypresses, evergreen and deciduous broad-leaf trees, shrubs, forbs, and grasses, but with very abundant and dense ferns in the forest understory (Figure 4). The equal contribution of both evergreen and deciduous broad-leaf trees and shrubs in this northern tropical region inferred the effect of parent material of limestone on vegetation, and therefore, it was a typical karst landscape. Dahu Lake is with a relatively small area, in which the fern spores should be transferred from the basin by runoff or surface flow and have high representation in pollen and spore spectra. The unordinary abundant fern spores, in-washed from the catchment into lake sediment, can have the potential to distort the accuracy of the fossil record in vegetation reconstructions (Wilmshurst and McGlone, 2005). But the very rich ferns indicated the existence of very abundant fern in forest understory around the Dahu Lake. The appearance of some aquatic plants, such as Typha, Sparganium, Potamogeton, Nelumbo, Trapa, Nymphaea, and Salvinia indicated that the Dahu Lake was large and existed over a long period of time.

Intuitively, the vegetation around the Dahu Lake could be divided into two big zones by the depth of 130 cm (1860 cal. yr BP). Zone I was characterized by the small range of fluctuation of Pinus, Podocarpus, and Cunninghamia with a lower percentage of less than 45%. Pinus pollen is over represented in almost all pollen assemblages because of its high production and airborne capability (Jackson and Kearsley, 1998; Xiao et al., 2011; Xu et al., 2007). In Pinus forest, Pinus pollen is more than 90% of total tree pollen (Li and Yao, 1990), while in non-pine area, it is usually less than 30% (Li and Yao, 1990; Xu et al., 2007). The Pinus pollen in terrestrial pollen sum was 17% to 80% (39% in average), indicating an existence of pine forest in Dahu region before1860 cal. yr BP. The tree and shrub layers were dense, and the understory was lush with very rich ferns. It was likely a woodland landscape with mixed broad-leaf and conifer trees dominant the canopy. The feature of zone II was the fast and substantial increase of pines with a higher value of 40% to 90% (66% in average), and the increase of Podocarpus during the later stage, suggesting the expansion of pine forest. Ferns were still very rich, but in reduced percentages. The landscape gradually turned to open woodland during this period.

Anthropogenic disturbance on natural vegetation around the Dahu Lake exist during the past 2700 years. Pinus has been reported as a pioneer specie in the secondary forest disturbed by human activities (Huang et al., 2014; Li et al., 2008; Sun et al., 1986; Zhou et al., 2004, 2015). In pollen zone I, Pinus pollen fluctuated but with slight increase trend (Figure 4). Numeric analyses also suggest that the pollen spectra in this zone was impacted by the sPDC which is a proxy of human activity (Wang et al., 2014; Xu et al., 2019). Therefore, the human activity had been impacting the vegetation composition during the pollen zone I. The strong impact of human activity on natural vegetation should exist in pollen zone II, where Pinus pollen increased significantly. This is consistent with regional vegetation change in middle Guangxi region where dramatic anthropogenic disturbances led to expansion of coniferous taxa, herbaceous taxa, and pioneer ferns during the past 1600 years (Zhao, 2013).

The cliff painting period lasted ca. 600 to 700 years from the Warring to the East Han Dynasties (UNESCO, 2017; Yuan et al., 1986). Pinus were not in the highest percentages of the entire profile during 2699 to 1860 cal. yr BP, but Podocarpus, especially Cunninghamia reached their highest percentages during this period (Figure 5). Besides, the increase of evergreen trees and shrubs during this period, indicates an expansion of arboreal forest. Rich and dense ferns occupied the herb layer. As we mentioned in the ‘Introduction’ section, the binding material of cliff painting is the oxalic acid from plants (Guo et al., 2005; Qiu et al., 1990; Sun et al., 2012). Such abundant gymnosperm and pteridophyte should be the key source of the oxalic acid and oxalates which are present in leaves, roots stems, fruits, and seeds of many plants (Prasad and Shivay, 2017).

Plentiful tropical plant leaves, stems, and roots were treated by maceration, boil, and extraction to produce the sticky slurry with much oxalic acid or oxalate. These slurries blended with the major dyestuff of iron oxides to further form the sticky mixture of dyestuff and binder. The ancient local people used such material to paint and glue on the cliff of limestone, which is the most common rock in this karst region. The oxalic acid from the dyestuff and binder mixture infiltrated through rock fractures and chemically reacted with the calcium carbonate from the limestone to produce calcium oxalate. This is why the calcium carbonate was found in the dyestuff lumps and between the dyestuff and rock layer (Sun et al., 2012).

Past climate change and the ways of painting on the cliff

Grain size has been widely applied to study the past climate and environment changes, since it is with great significance on the interpretation for lake water dynamics, geomorphology, and sedimentology (Dietze et al., 2012; Liu et al., 2009; Xiao et al., 2015). During the transportation to the lake center, sorting process makes the gradual grain size refinement. In general, the deep water of the off-shore zone possesses lower hydraulic energy than the shallow water of the near-shore zone, and the clastic deposits in the lake center become finer with lower hydraulic energy (Xiao et al., 2015). Dahu Lake is a closed lake in geomorphology, although a throughput lake with a doline in lakebed. High contents of clay and silt fractions in the sediment therefore should occur in a larger lake body under humid climate. On the contrary, increases of coarse silt and sand fractions should indicate a decrease of lake water level in a dry condition. The negative correlation (r = –0.34, p = 0.005, n = 66) between the fine-silt fraction and stalagmite δ¹⁸O further verified this explanation (Figure S2, available online).

Pollen-based climate reconstructions have been widely reported (Chen et al., 2014; Huang et al., 2018; Prentice et al., 1998; Xiao et al., 2014), based on the dynamic equilibrium between climate and vegetation (Prentice et al., 1991). Modern pollen data analysis indicates that the mean annual precipitation is the most significant dominant climate variable to vegetation in China (Lü et al., 2011; Lu et al., 2019). However, our results only imply weak correlation between the stalagmite δ¹⁸O and the pollen spectra from Dahu Lake profile (Table S1, available online). The reasons beneath are probably the absence of more climate proxies can be used. In addition, the resistance of natural vegetation to climate change (Prentice et al., 1991) and the disturbance of human activity would complicate the relationship between vegetation and climate conditions. Alternatively, we discuss the regional climate change mainly based on the grain size composition variation from Dahu Lake profile.

The slight decrease of fine-silt fraction in grain size composition suggest that Dahu Lake region underwent short-term dry condition during 2699 to 2560 cal. yr BP. The maximum percentage of herbaceous taxa and rare plankton algae indicated the expansion of grassland patches and the more openness of vegetation around the lake under this drought episode (Figure 4). The short-term drought revealed in Dahu Lake profile should relate to the decreased monsoonal precipitation around 2700 cal. yr BP (Dykoski et al., 2005; Wang et al., 2005). Because of the shrink of monsoon strength, the 400 mm isohyet in the western China moved eastern-ward (Shen et al., 2008), and the tree taxa in pollen record dramatically decreased in Hainan Island during this episode (Zheng et al., 2003). This drought maybe probably coincident with the 2.7 ka BP ice-raft event in North Atlantic, in which the climate turn to dry and cold in southeast Asia (Shen et al., 2008; Xiao et al., 2014; Zheng et al., 2003; Zhong et al., 2010).

During 2560 and 1860 cal. yr BP, a humid climate was dominant in this region. Fine fractions in the sediments increased, manifesting the expansion of lake water area (Figure 5). The extremely rich of ferns in the understory and the increases in plankton also indicated a moist environment (Figure 4). However, two drought episodes around 2500 and 2420 cal. yr BP have been revealed by the grain size compositions with high content of coarse fractions (Figure 5). Grain size compositions changed obviously around 2000 cal. yr BP, with dramatically increasing of coarse silt and sand fractions (Figure 5). This variation in grain size should relate to the decrease of lake water level. After 1900 cal. yr BP, ferns began to decrease (Figure 4), indicating a drought tendency.

Since 1860 cal. yr BP, climate was characterized by decreased humidity. Coarse silt and sand fractions dominated the grain size composition (Figure 5), indicating a shallow lake status. The decrease trend in ferns was obvious, with the shrinking of ferns in the forest understory. Conifer forest expanded significantly, especially the Pinus forest. We suggest that the vegetation around Dahu Lake during this period must be subject to strong influence from the anthropogenic disturbance. But the variations of grain size composition certainly implied the drought tendency in Dahu Lake region.

The high-stand water level during 2560 to 2000 cal. yr BP provided a possibility for cliff painting. The rock art is generally located high up on the river cliffs (UNESCO, 2017), with 15 to 150 m above the modern water levels. The way approaching to the painting site on cliffs includes climbing upward to the cliff or suspending downward with ropes (UNESCO, 2017). The former is with more possibility, considering the stake founded in front of a cliff painting. More interestingly, the stake has been dated back to 2680 cal. yr BP (Yuan et al., 1986), just the initial stage of the cliff paintings. The fact is that all the rock paintings locate on river cliff at the outer bank and facing the open flood plain at the inner bank. Generally, the outer banks around river cliff are with higher water level when water gushing occurs within increased river runoff during periods with heavily precipitation. The existence of conifer forest provided the necessary trunk to make stake or ladder, and the higher water level of the river made it easier to approach the painting level. Therefore, a reasonable painting way should be that ancient people used stake or ladder to climb the cliffs and created the cliff paintings when the river water level raised during 2560 to 2000 cal. yr BP. From this view, the gradually fading of cliff painting activity after 2115 cal. yr BP may probably relate to the downdrawn of river water level after 2000 cal. yr BP at some degree.

Society development and the motivation of cliff painting

As discussed above, the cliff paintings were probably created with a higher river water level, that is, a humid climate condition. It seems contradictory that ancient people created this giant cliff paintings in the cold and dry late Holocene, but not in the warm–humid early and middle Holocene in the southern China (Dykoski et al., 2005; Sheng et al., 2017; Wang et al., 2005; Zhang et al., 2017). The fact should be that the motivation of the cliff paintings rooted in the social development and the relationship between human and environment.

The recovery of archeological sites dated back to the late Neolithic Age implied that the lifestyle in this region was still gathering and hunting (Fu et al., 1998; Jiang, 1998; Zhao et al., 2005), but with popularization and development of rice cultivation during 5000 to 4000 cal. yr BP (Chen, 2016; Chi and Hung, 2015; Zhang et al., 2006; Zhao et al., 2005). The repayable resources from gathering and hunting gradually reduced with the climate changing from middle to late Holocene, on the contrary the activities of rice cultivation increased (Zhao et al., 2005). The rice agriculture further developed, especially after the usage of bronze implements, that is, the introduction of Bronze Culture from central China as early as late Shang Dynasty (Jiang, 2004; Li, 1996; Zheng, 2003, 2004). Both the development of rice-cultivation and the usage of bronze implements should promote the population growth, based on which the tribes and social classes formed. Over the past decades, a lot of archeological sites have been investigated and dated in Guangxi region (GICRA, 2010; Huang, 2015; Jiang, 1998; Wang et al., 2014). During 3000 to 2000 cal. yr BP, the sPDC were obviously higher than the former period (Wang et al., 2014), suggesting that the population in Guangxi increased significantly after 3000 cal. yr BP (Figure 7a). The population growth will inevitably increase the demands for living resources, including food, space, and other resources. These demands formed the potentially motivation for human activities to disturb the natural vegetation (Figure 6) and fight with competitors. Actually, two major local tribes, that is, Luoyue and Ouyue, existed in Guangxi region 2000 years ago, and both of them were combative and fight each other frequently (Jiang, 1998; Li, 1996; Zheng, 2003). The most popular bronze ware they used were weapons, including bronze dagger and bronze battleax (Jiang, 2004), also indicating the frequency of war.

Figure 7.

Natural and social–political environments of the Huashan cliff paintings, including the information about the Chinese dynasties during Zhou and East Han Dynasties, (a) the summed probability distribution of calibrated radiocarbon dates in Guangxi over the last 3000 years, (b) the stalagmite δ¹⁸O from Dongge Cave as the Asian Summer Monsoon proxy (Wang et al., 2005), (c) the fine silt fraction and Pinus pollen percentage (d) from Dahu Lake profile. The summed radiocarbon probability distribution is calculated in Calib 7.1 with an IntCal13 calibration data set (Reimer et al., 2013), based on the radiocarbon data from Wang et al. (2014). Light blue area representing the periods of cliff paintings (UNESCO, 2017), while the pink-filled area implying the radiocarbon date from the cliff paintings (Yuan et al., 1986). The red dash line corresponding to the date that Qin Dynasty unified the Guangxi region.

The society in the past was extremely fragile in front of major climate change episodes and their broader consequences (Kaniewski et al., 2019; Zhang et al., 2008). Fundamentally, the rice agriculture was vulnerable to natural disasters, especially to drought or flood which can result in lower yields or less cultivated acreages. A rapid climate shift triggered the civilization collapse in western Asia around 3000 cal. yr BP (Kaniewski et al., 2019), and also the variability of Asian Summer Monsoon correlated well with Chinese cultural changes during the past two millennia (Zhang et al., 2008). The remarkable weak monsoon events occurred around 2800 cal. yr BP, with the second weak monsoon event around 2400 cal. yr BP (Dykoski et al., 2005; Wang et al., 2005), implying an extraordinarily high monsoonal variabilities during these two periods (Figure 7b). Actually, highly turbulent climate, with decreased temperature but strengthening of seasonal precipitation has been recorded in subtropical China during 2700 to 2400 cal. yr BP (Ran and Feng, 2013; Shen et al., 2013; Sheng et al., 2017; Zheng et al., 2003; Zhong et al., 2010). In Dahu Lake, the changes of grain size compositions also implied the frequent occurrence of drought episodes during 2700 and 2400 cal. yr BP (Figure 7c). The turbulent climate episodes probably played this role as the disaster and impact on the human activities and lead to the social–political tense, and finally, the outbreak of war.

The scenes on the Zuojiang Huashan cliff paintings illustrated the life and rituals of the Luoyue people (UNESCO, 2017). In a surrounding landscape of karst, rivers, and plateau, they depict ceremonies that have been interpreted as portraying the bronze drum culture once prevalent across southern China. Sacrifice and weapons are the major subject matter of cliff paintings. They used bronze drum in the sacrificial ceremony to pray a pleasant climate and/or the victory in the battle with the other tribe. These should be their motivation for creating the cliff paintings during this period.

The regional climate became drier after 2100 cal. yr BP, and the creating of cliff paintings was gradually fading (Figure 7). During this stage, the culture from central China must have been impacting the Guangxi region (Huang, 2015; Li, 1996). After the army of Qin Dynasty unified the southern China and set counties in Guangxi at 214 BC, the social–political forms from central China gradually became dominant in the social development (Huang, 2015; Li, 1996). More and more people migrated into Guangxi, and they were not allowed in private fight but demanded to devote themselves to agricultural production. What was more important, the immigrant brought advanced iron implements and dryland cereal crops, for example, millet and wheat (Gong, 2018). All of these social–political changes in Guangxi during Qin and Han Dynasties relieved the pressure from natural climate changes to rice agriculture and led to a more stable society than before when facing a major climate shift. This should be the reason why the creating of cliff paintings was gradually fading during the Han Dynasty in which climate became worse.

The expanded population in Guangxi during Qin and Han Dynasties would inevitably result in the destruction on natural vegetation. The remarkable increase of Pinus pollen in Dahu Lake after 1860 cal. yr BP, manifested that the human activities began to have strong influence on regional environment since the middle East Han Dynasty. Although the rock paintings stopped depicting the history of human activities, the vegetation and lacustrine sediment around Huajiang cliff painting region still record over the past 2000 years with their peculiar way.

Supplemental Material

sj-pdf-1-hol-10.1177_0959683620919980 - Supplemental material for Late-Holocene vegetation change reveals the environment of ancient people and the origin of Huashan cliff paintings in Guangxi, southwestern China

Supplemental material, sj-pdf-1-hol-10.1177_0959683620919980 for Late-Holocene vegetation change reveals the environment of ancient people and the origin of Huashan cliff paintings in Guangxi, southwestern China by Kai Li, Yun Zhang, Mengna Liao, Jian Ni and Yunfa Chen in The Holocene

Footnotes

Acknowledgements

The authors would like to thank Libin Liu, Gang Hu, Qiuping Zhu, Xiaowen Liao, Shichu Zhou, and Zhiming Xie for their help during the field work. They also appreciate the two anonymous reviewers for their constructive suggestions.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article. This study was supported by the National Key R&D Program of China (grant number 2016YFC0502101); the National Natural Science Foundation of China (grant number 41662002); the Natural History Museum of Guangxi, China; and the China Scholarship Council.

ORCID iD

Kai Li

Supplemental material

Supplemental material for this article is available online.

References

Chen

, et al. (2014) Holocene vegetation history, precipitation changes and Indian Summer Monsoon evolution documented from sediments of Xingyun Lake, south-west China. Journal of Quaternary Science 29: 661–674.

Chen

(2016) Study on the Time-Space Framework and Livelihood Mode of Neolithic Culture in Lingnan Area. Changchun, China: College of Literature, Jilin University, pp. 197–198.

Chi

Hung

(2015) The emergence of agriculture in southern China. Antiquity 84: 11–25.

China National Institute of Cultural Property (CNICP) (1986) Qualitative identification of the red pigment for the petroglyphs in Guangxi. Guangxi Cultural Relics 2: 72.

De’ath

(2007) mvpart: Multivariate partitioning (R package version 1.2-4). Available at: https://cran.r-project.org/web/packages/mvpart/index.html.

Dietze

Hartmann

Diekmann

, et al. (2012) An end-member algorithm for deciphering modern detrital processes from lake sediments of Lake Donggi Cona, NE Tibetan Plateau, China. Sedimentary Geology 243–244: 169–180.

Dykoski

Edwards

Cheng

, et al. (2005) A high-resolution, absolute-dated Holocene and deglacial Asian monsoon record from Dongge Cave, China. Earth and Planetary Science Letters 233: 71–86.

Editorial Committee of Vegetation Map of the People’s Republic of China (ECVMP), Chinese Academy of Sciences (2007) Vegetation Map of the People’s Republic of China (1:1 000 000). Xi’an, China: Geology Press.

Faegri

Kaland

Krzywinski

(1989) Textbook of Pollen Analysis. Hoboken, NJ: John Wiley & Sons.

10.

, et al. (1998) Excavation of Dingsishan Ruins in Yongning county, Guangxi. Archaeology 11: 11–33.

11.

Gong

(2018) Grain crops growing in Lingnan area during the Qin and Han Dynasties and the relevant problems. Agricultural Archaeology 2018: 33–39.

12.

Grimm

(1987) CONISS: A FORTRAN 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares ⋆. Computers and Geosciences 13: 13–35.

13.

Grimm

(1993) TILIA v2.0 (Computer Software). Springfield, IL: Research and Collections Centre, Illinois State Museum.

14.

Guangxi Institute of Cultural Relics and Archaeology (GICRA) (2010) Guangxi Archaeological Collection (Series 4). Beijing, China: Science Press.

15.

Guo

Han

Zhao

, et al. (2005) Pigment and the prevention of its fading on Petroglyphs of the Flower Mountain. Science of Conservation and Archaeology 17: 7–14.

16.

Yang

Ning

(2009) New discovery and preliminary understanding on beiqiu archaeological sites from Zuojiang river basin, Guangxi. Chinese Cultural Relics 5: 4–11.

17.

Huang

Zheng

Liao

, et al. (2014) Reconstructing late Holocene vegetation and fire histories in monsoonal region of southeastern China. Palaeogeography, Palaeoclimatology, Palaeoecology 393: 102–110.

18.

Huang

Peng

Rudaya

, et al. (2018) Holocene vegetation and climate dynamics in the Altai Mountains and surrounding areas. Geophysical Research Letters 45: 6628–6636.

19.

Huang

(2015) Archaeological Research on Nanyue Kingdom. Beijing, China: China Social Sciences Press.

20.

Jackson

Kearsley

(1998) Quantitative representation of local forest composition in forest-floor pollen assemblages. Journal of Ecology 86: 474–490.

21.

Jiang

(1998) Forty years of Guangxi archaeology. Archaeology 11: 961–970.

22.

Jiang

(2004) The bronze smelting and founding profession in Lingnan of the history of the Xianqin Dynasty and the Lianghan Dynasty. Journal of Guangxi University for Nationalities (Natural Science Edition) 10: 34–39.

23.

Kaniewski

Marriner

Cheddadi

, et al. (2019) Cold and dry outbreaks in the eastern Mediterranean 3200 years ago. Geology 47: 933–937.

24.

(1996) The southern spread of culture in Qin and Han dynasties and its influence on Ouyue and Luoyue cultures. Guangxi National Research 2: 62–66.

25.

Yao

(1990) A study on the quantitative relationship between Pinus pollen in surface sample and Pinus vegetation. Acta Botanica Sinica 32: 943–950.

26.

Zhou

Cui

(2008) Palyological indicators of human activity. Chinese Science Bulletin 53: 991–1002.

27.

Saito

Dang

, et al. (2009) Warfare rather than agriculture as a critical influence on fires in the late Holocene, inferred from northern Vietnam. Proceedings of the National Academy of Sciences 106: 11490–11495.

28.

Saito

Matsumoto

, et al. (2006) Climate change and human impact on the Song Hong (Red River) Delta, Vietnam, during the Holocene. Quaternary International 144: 4–28.

29.

Liu

Dong

Yang

, et al. (2009) Late Holocene forcing of the Asian winter and summer monsoon as evidenced by proxy records from the northern Qinghai–Tibetan Plateau. Earth and Planetary Science Letters 280: 276–284.

30.

Long

(1986) Deduction on the rock composition of the petroglyphs in the Zuo River drainage area. Guangxi National Research 4: 34–42.

31.

Zhu

, et al. (2019) Variability of East Asian summer monsoon precipitation during the Holocene and possible forcing mechanisms. Climate Dynamics 52: 969–989.

32.

Lü

Liu

K-B

, et al. (2011) Modern pollen distributions in Qinghai-Tibetan Plateau and the development of transfer functions for reconstructing Holocene environmental changes. Quaternary Science Reviews 30: 947–966.

33.

Cao

Jeltsch

, et al. (2014) Biome distribution over the last 22,000 yr in China. Palaeogeography, Palaeoclimatology, Palaeoecology 409: 33–47.

34.

Oksanen

Blanchet

Friendly

, et al. (2019) Package ‘vegan’.2.5-6 ed., Ordination methods, diversity analysis and other functions for community and vegetation ecologists. Available at: https://cran.r-project.org/web/packages/vegan/vegan.pdf.

35.

Prasad

Shivay

(2017) Oxalic acid/oxalates in plants: From self-defence to phytoremediation. Current Science 112: 1665–1667.

36.

Prentice

(1980) Multidimensional scaling as a research tool in quaternary palynology: A review of theory and methods. Review of Palaeobotany and Palynology 31: 71–104.

37.

Prentice

Bartlein

Webb

T III

(1991) Vegetation and climate change in Eastern North American since the Last Glacial Maximum. Ecology 72: 2038–2056.

38.

Prentice

Harrison

Jolly

, et al. (1998) The climate and biomes of Europe at 6000yr BP: Comparison of model simulations and pollen-based reconstructions. Quaternary Science Reviews 17: 659–668.

39.

Qin

(1987) Investigation and Research of Petroglyphs in the Zuo River Field, Guangxi. Nanning, China: Guangxi Nationality Publishing House.

40.

Qiu

Pan

, et al. (1990) Initial research of the pigment and adhesive of the petroglyphs on the Flower Mountain. Cultural Relics 1: 85–87.

41.

R Core Team (2016) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. Available at: https://www.R-project.org/.

42.

Ran

Feng

(2013) Holocene moisture variations across China and driving mechanisms: A synthesis of climatic records. Quaternary International 313: 179–193.

43.

Reimer

Bard

Bayliss

, et al. (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0-50,000 years cal BP. Radiocarbon 55: 1869–1887.

44.

Shao

(2016) Research on the U-series chronology of the Huashan cliff paintings in Zuojiang. China Cultural Heritage 4: 66–69.

45.

Shen

Liu

K-B

Morrill

, et al. (2008) Ecotone shift and major droughts during the mid-late Holocene in the central Tibetan Plateau. Ecology 89: 1079–1088.

46.

Shen

Zhang

, et al. (2013) Ti content in Huguangyan Maar Lake sediment as a proxy for monsoon-induced vegetation density in the Holocene. Geophysical Research Letters 40: 5757–5763.

47.

Sheng

Wang

Zhang

, et al. (2017) A 20,000-year high-resolution pollen record from Huguangyan Maar Lake in tropical–Subtropical South China. Palaeogeography, Palaeoclimatology, Palaeoecology 472: 83–92.

48.

Sun

Qiao

, et al. (1986) Late Pleistocene and Holocene vegetation history at Kunming, Yunnan Province, Southwest China. Journal of Biogeography 13: 441–476.

49.

Sun

Wang

Wei

(2012) Research on the pigment cementing materials of rock paintings on Ningming Flower Mountain, Guangxi. Sciences of Conservation and Archaeology 24: 38–43.

50.

Tang

(1996) Study and discussion on the composition of Huashan murals in Ningming County, Guangxi. Guangxi Geology 9: 83–87.

51.

UNESCO (2017) Zuojiang Huashan Rock Art cultural landscape. Available at: http://whc.unesco.org/en/list/1508.

52.

Wang

Zhang

, et al. (2014) Prehistoric demographic fluctuations in China inferred from radiocarbon data and their linkage with climate change over the past 50,000 years. Quaternary Science Reviews 98: 45–59.

53.

Wang

Lü

Liu

, et al. (2007) The early Holocene optimum inferred from a high-resolution pollen record of Huguangyan Maar Lake in southern China. Chinese Science Bulletin 52: 2829–2836.

54.

Wang

Cheng

Edwards

, et al. (2005) The Holocene Asian monsoon: Links to solar changes and North Atlantic climate. Science 308: 854–857.

55.

Wilmshurst

McGlone

(2005) Origin of pollen and spores in surface lake sediments: Comparison of modern palynomorph assemblages in moss cushions, surface soils and surface lake sediments. Review of Palaeobotany and Palynology 136: 1–15.

56.

Xiao

Fan

Zhai

, et al. (2015) Testing the model for linking grain-size component to lake level status of modern clastic lakes. Quaternary International 355: 34–43.

57.

Xiao

Haberle

Shen

, et al. (2014) Latest Pleistocene and Holocene vegetation and climate history inferred from an alpine lacustrine record, northwestern Yunnan Province, southwestern China. Quaternary Science Reviews 86: 35–48.

58.

Xiao

Shen

Wang

(2011) Spatial variation of modern pollen from surface lake sediments in Yunnan and southwestern Sichuan Province, China. Review of Palaeobotany and Palynology 165: 224–234.

59.

Chu

, et al. (2019) Synchronous 500-year oscillations of monsoon climate and human activity in Northeast Asia. Nature Communications 10: 4105.

60.

Yang

, et al. (2007) Quantitative relationship between pollen and vegetation in northern China. Science in China Series D: Earth Sciences 50: 582–599.

61.

Yuan

Chen

(1986) Research on ¹⁴C chronology of the petroglyphs in Huashan, Ningming County, Guangxi. Guangxi National Research 4: 27–34.

62.

Zhang

Chang

Cao

, et al. (2017) Holocene high-resolution quantitative summer temperature reconstruction based on subfossil chironomids from the southeast margin of the Qinghai-Tibetan Plateau. Quaternary Science Reviews 165: 1–12.

63.

Zhang

Cheng

Edwards

, et al. (2008) A test of climate, sun, and culture relationships from an 1810-year Chinese cave record. Science 322: 940–942.

64.

Zhang

Xiang

Qiu

, et al. (2006) Ancient rice from Shixia Ruins at Maba of Qujiang in Guangdong Province. Acta Agronomica Sinica 32: 1695–1698.

65.

Zhao

(2013) A Study on Human Activities Changes and Its Impact on Environment during the past 1600 Years in Guangxi and Chongqing, Based on Pollen-Spore and Charcoal Data. Chongqing, China: Southwest University.

66.

Zhao

Lü

(2005) Research on the phytoliths from Dingsishan site in Yongning County, Guangxi. Archaeology 11: 76–85.

67.

Zheng

(2003) From Gu kingdom to Fang kingdom – A new thinking on the origin of Zhuang civilization. Guangxi National Research 4: 60–68.

68.

Zheng

(2004) A Study on the Bronze Wares of Zhou Dynasty in the Regions of Wuyue and Baiyue. Changchun, China: Jilin University.

69.

Zheng

Wang

, et al. (2003) High resolution Holocene environmental record in Shuangchi Marr Lake, Hainan Island. Chinese Science Bulletin 48: 282–286.

70.

Zhong

Xue

Zheng

, et al. (2010) Climatic changes since the last deglaciation inferred from a lacustrine sedimentary sequence in the eastern Nanling Mountains, south China. Journal of Quaternary Science 25: 975–984.

71.

Zhou

Qin

Zhang

, et al. (2015) Vegetation, climate and depositional environment changes since the middle Holocene in the karst area of Guilin, Guangxi. Chinese Science Bulletin 60: 1197–1206.

72.

Zhou

Jull

AJT

, et al. (2004) High-resolution evidence from southern China of an early Holocene optimum and a mid-Holocene dry event during the past 18,000 Years. Quaternary Research 62: 39–48.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB