Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses

Abstract

The question of what the ecology communities of the Loess Plateau looked like before the extensive anthropogenic reshaping processes of the late-Holocene has stirred a long debate. A better understanding of these human-induced changes will not only help us understand the extent of paleoeconomic practices, but also inform future conservation actions on this ecologically fragile landscape. This paper presents a systematic study of wood charcoal from a series of archaeological sites, demonstrating that woody plants and woodlands were widely distributed and geographically diverse on the Loess Plateau in response to the East Asian Monsoon. In the Guanzhong Basin, oak (Quercus spp.) woodlands were associated with a few tropical-subtropical taxa, which changed to oak and pine (Pinus spp.) mixed forests on the eastern plateau, while on the northern plateau coniferous woodlands were dominant. On the western Loess Plateau, oak and elm (Ulmus spp.) woodland and spruce (Picea spp.) forests were widespread. The charcoal results suggest that human impacts on the dominant species might have begun as early as ca. 3500 cal yr BP, with oak replaced by Prunus as the dominant taxon, including many economically significant species, such as peaches (P. persica) and apricots (P. armeniaca). Furthermore, the charcoal data show that due to warm and wet climatic conditions in the mid-Holocene, the distribution of tropical-subtropical taxa shifted markedly northwards into the Guanzhong and Tianshui Basins, and the central part of the eastern Loess Plateau, which became characterized by high frequencies of Bamboo.

Keywords

Loess Plateau mid-Holocene paleoecology tropical-subtropical taxa wood charcoal woodland

Introduction

Located in northern China, the climate of the Loess Plateau is principally influenced by the East Asian Monsoon (EAM) and characterized by a fragile ecosystem (Liu, 1985). Current evidence shows that this region is part of the broad center of origin for both broomcorn (Panicum miliaceum) and foxtail millet (Setaria italica), with cultivation beginning as early as 8000 cal yr BP (Hu et al., 2009; Liu et al., 2004). The rapid expansion of a farming economy led to significant changes in vegetation communities across this area (Lu et al., 2003; Shu et al., 2006). Pollen and charcoal archives reveal that the frequency of human-induced fires increased significantly during this period, and agricultural activities reduced trees and promoted grasslands (Li et al., 2009; Tan et al., 2015; Zhou et al., 2011). Additionally, since the loess is loose, porous, homogeneous and easily soluble in water, the landscape here has experienced considerable soil erosion (Liu, 1985; Zhao et al., 2013). The National Development and Reform Commission et al. (2010) reported this region as one of the most erosion-prone areas in China, with a particularly susceptible area of 4.72 × 10⁵ km². Reconstructing the vegetation history of the Loess Plateau in an attempt to understand its natural woodland communities and vegetation responses to climate change, has become key to the future ecological conservation of this significant region (Liu et al., 1994; Lu et al., 1999, 2003, 2007). Also, the data and conclusions that we present here place East Asia into global-scale discussions of anthropogenic impacts on the environment. Understanding anthropogenic changes on the Loess Plateau is important to any discussion of human impacts on the globe, as this region saw some of the most rapid demographic expansions in human history and prehistory and continues to support high densities of human occupation (Hosner et al., 2016).

Over the past few decades, attempts to understand the natural vegetation of this plateau, using different climatic proxies have resulted in controversial conclusions. Based on historical records, Shi (1991) claimed that forests and forest-steppes occurred extensively, covering up to 53% of the plateau during the mid-Holocene. However, studies of paleoclimatic indicators such as pollen and phytoliths suggest a different vegetation history (Liu et al., 1994; Lu et al., 1999; Sun et al., 1995, 1998; Zhang and An, 1994). For example, Sun et al. (1995) demonstrated with palynological data that vegetation types throughout the Holocene were dominated by grasslands and some associated broadleaf trees. Relying primarily on phytolith archives, Lu et al. (1999) concluded that since the last interglacial period, grasslands and forest-steppes were widespread on the southern Loess Plateau, and that short-lived forest patches might have occurred only when the annual precipitation exceeded 800 mm during the Holocene Climate Optimum. At present, a growing corpus of studies have argued that prehistoric vegetation on the plateau is likely to have been highly spatially variable, and that it was always dominated by grasslands in the ‘Yuan’ areas (the tablelands), with no large areas of stable forest vegetation (Guo et al., 1994; Li et al., 2003; Lu et al., 2003; Xia et al., 1998; Xie et al., 2002). By contrast, in the valleys and mountains of the plateau, forests were likely well-developed (Lu et al., 2003).

However, due to limitations inherent in palynological methods, many questions remain unanswered regarding the composition and distribution of woodland patches, especially in valleys and protected areas away from heavy megafaunal herbivory. Some of these methodological issues include poor preservation of pollen in loess soils, relatively low resolution of pollen identification, and the fact that several of the main woody plant clades are insect-pollinated (e.g., Rosaceae) (Asouti et al., 2015; Kabukcu, 2017; Shen et al., 2018). Likewise, statistical studies are increasingly illustrating how unreliable temporal extrapolations can be when interpreting sediment cores, especially from lakes with high turbulence and uneven sedimentation (Carleton et al., 2014). Additionally, scarcity of good sediment achieves for pollen on the Loess Plateau, such as small lake/bog sites, has hampered the application of vegetation models.

Wood charcoal remains found in archaeological sites directly represent the use of wood as fuel and, therefore, comprise a combination of human selection choices and the composition of locally available wood (Kabukcu, 2018; Théry-Parisot et al., 2010). The carbonization process reduces microbial decomposition, and often leaves the anatomical features of the wood well-preserved, making it possible to precisely identify specimens that can be directly dated (Moskal-del Hoyo et al., 2010; Scott, 2010). Although many factors, such as wood features and cultural preference, can affect firewood collection, it is widely recognized that the availability of woody plants is one of the most important factors determining their ease of collection and thus their percentage abudannce in an assemblage can be regarded as a reflection of their proportions in local woodland (Asouti and Austin, 2005). In comparison to pollen, wood charcoal analyses allow for a relatively high taxonomic resolution, often to genus or sometimes species level (Cui et al., 2002). As a result, wood charcoal analyses are an effective tool for reconstructing the history of woodlands in semi-arid and arid regions, especially when incorporated into multiproxy interpretations also involving pollen and other lines of paleoclimatic data (Asouti and Hather, 2001; Asouti and Kabukcu, 2014).

Currently, wood charcoal research in East Asia is mainly restricted to northwest China (An et al., 2014; Li et al., 2013, 2016; Sun et al., 2013, 2016; Wang et al., 2011b). However, studies remain quite sporadic and are absent for most parts of the Loess Plateau. In this paper we present the results of anthracological analyses conducted on stratified charcoal assemblages from 13 archaeological sites on the plateau to explore its woodland histories.

Study area

The Loess Plateau (33°43′ to 40°16′ N, 100°45′ to 114°33′ E), with a total area of about 4.4 ×10⁵ km², is traversed by the Yellow River (Liu, 1985) (Figure 1). The terrain is high in the northwest and low in the southeast, with a range of 1000–2000 m above sea level (masl; Wang et al., 2014). It is characterized by a typically continental monsoonal climate, being cold and dry when under the control of the Siberian High during the winter, and hot and humid in the summer months, when the EAM transports water vapor from the tropical Pacific Ocean. The mean annual temperature (MAT) range is 0–13°C, and the mean annual precipitation (MAP) range falls between 150 and 750 mm (Li et al., 2008; Yu et al., 2002; Figure 2).

Figure 1.

Location of the Loess Plateau and 13 archaeological sites in this study. (1) Majiayao. (2) Shangguancun. (3) Liujiatai. (4) Panhe. (5) Ma’nan. (6) Dakou. (7) Zhaimao. (8) Shimao. (9) Xiaojiamao. (10) Lijiaya. (11) Guangcun. (12). Baiyan. (13). Youyao. (single-column).

Figure 2.

Vegetation types, mean annual temperature (MAT) and mean annual precipitation (MAP) for the Loess Plateau. (I) Forest. (II) Woodland. (III) Steppe. (IV) Desert steppe. (V) Desert (modified from Integrated investigation team of the Loess Plateau, CAS, 1991; Wang et al., 2011a). (single-column).

With the spatial variation in climatic conditions, the general trend in vegetation types progresses from forest, to forest-steppe, to steppe, to desert-steppe and to desert, from southeast to northwest (Shu et al., 2006; Figure 2). Forests are only found on the southeastern part where the vegetation community is dominated by Quercus variabilis, Q. aliena, and Q. acutissima. The forest-steppe is dominated by xerophytic, herbaceous plants and characterized by environmentally dwarfed trees. Representative trees include Prunus sibirica, Ulmus macrocarpa and U. glaucescens (Zhu, 1994). Steppe vegetation is usually represented by Artemisia leucophylla, Caragana korshinskii, C. microphylla, Stipa bungeana and Ziziphus jujuba. The desert steppe mainly consists of Artemisia ordosica, Haloxylon ammodendron, Stipa caucasica and S. breviflora. Desert vegetation is restricted to the northwestern part; typical arid-land plants include Ammopiptanthus mongolicus, Kalidium foliatum and Sympegma regelii.

Material and methods

We analyzed material from 13 archaeological sites including Shangguancun, Liujiatai, Panhe and Ma’nan on the Guanzhong Basin; Guangcun, Baiyan and Youyao on the eastern side of the plateau; Dakou, Zhaimao, Shimao, Xiaojiamao and Lijiaya on the northern plateau; and the Majiayao sites on the western plateau (Figure 1 and Table 1). These sites cover a timespan from the Yangshao Tradition (7000–5000 cal yr BP), to the Longshan Tradition (5000–4000 cal yr BP), to the Xia-Shang Period (4000–3000 cal yr BP) (Table 1). Of them, only one site (Majiayao) exhibits a continuous cultural sequence, with a stratigraphy 3.8 m in depth. Eight samples were collected from a stratigraphic column at Majiayao at 20 cm intervals, with a volume of soil for each sample of about 40 L. For the other sites, all the samples were acquired from ash pits, and each had a volume of roughly 20 L per sample.

Table 1.

Summary of geographic location, contextual attributes, charcoal samples & counts and archaeological periodisation for all sites included in the present study.

Site	Latitude (N)	Longitude (E)	Location	Sample origin	Volume of each soil sample	No of samples	Charcoal counts	Cultural period
Majiayao	35°18′20.4″	103°48′5.2″	Gansu	cultural layer	40 l	8	954	Yangshao
Ma’nan	34°28′18.2″	109°06′19.6″	Shaanxi	ash pit	20 l	6	108	Yangshao
Baiyan	37°26′39″	112°41′45.4″	Shanxi	ash pit	20 l	1	190	Longshan
Dakou	39°25′46.74″	111°12′5″	Inner Mongolia	ash pit	20 l	2	441	Longshan
Panhe	34°18′23.43″	107°23′39″	Shaanxi	ash pit	20 l	1	122	Longshan
Shimao	38° 33′34.17″	110°18′2.2″	Shaanxi	ash pit	20 l	6	457	Longshan
Xiaojiamao	38°08′18″	109°19′19″	Shaanxi	ash pit	20 l	1	11	Longshan
Youyao	38°22′41″	112°51′0.3″	Shanxi	ash pit	20 l	1	52	Longshan
Zhaimao	38° 49′13″	110°32′56″	Shaanxi	ash pit	20 l	1	115	Longshan
Guangcun	35°45′13″	111°10′2.8″	Shanxi	ash pit	20 l	1	114	Xia-Shang
Lijiaya	37°04′17″	110°26′0.4″	Shaanxi	ash pit	20 l	1	148	Xia-Shang
Liujiatai	34°21′57.7″	107°16′16.05″	Shaanxi	ash pit	20 l	1	169	Xia-Shang
Shangguancun	34°27′43″	107°11′43″	Shaanxi	ash pit	20 l	1	144	Xia-Shang

Archaeobotanical samples in this study were retrieved through manual tank flotation, with the mesh size of 0.3 mm. All wood charcoal specimens were collected from the float fraction only. Wood charcoal was botanically identified following standard protocols and microscopy work was focused on wood charcoal specimens >4 mm in order to minimize identification biases as suggested in previous studies (Asouti and Austin, 2005; Chabal et al., 1999; Kabukcu, 2018; Théry-Parisot et al., 2010). For wood charcoal samples containing more than 100 fragments >4 mm, we stopped counting once the number of taxa identified stabilized. If there were not sufficient charcoal fragments >4 mm, we tried to identify (depending on the preservation of sufficient anatomical features) charcoal from <4 mm. Each wood charcoal fragment was either hand- or pressure-fractured with a razor blade to acquire three fresh sections (including transverse, tangential and radial sections), and then examined under a reflected light microscopy at magnifications of ×50, ×l00, ×200, ×400 and ×500. Based on the anatomical characters evident from the three sections, and using atlases of modern wood anatomy (Cheng et al., 1992; Yao et al., 2002; Zhou et al., 1994) as a point of reference, charcoal fragments were identified into different wood genera or families. The main types of wood charcoal found on the Loess Plateau were also photographed using a scanning electron microscope (see Supplementary material 1).

Our modelling of past woodlands, using wood charcoal assemblages, was based on ecological studies of present-day vegetation distribution, woodland composition and individual species habit (Integrated investigation team of the Loess Plateau, CAS, 1991; Wu, 1991; Wu et al., 2013). Drawing on data compiled by the Flora of China (Wu et al., 2013), Plants of the World Online by Royal Botanical Garden, Kew (http://www.plantsoftheworldonline.org/), and Wu (1991), we classified each taxon into different bioclimatic regions before exploring the implications of this classification for reconstructing different woodland vegetation types.

Results

A chronology for each site was established using AMS ¹⁴C dating. In addition to ages of four sites, including Majiayao, Dakou, Shimao and Lijiaya, that had been previously dated, nine new AMS dates were generated for the present study. The results show that all 13 sites in this study fall within a period ranging between 5300 and 3000 cal yr BP (Table 2). The Majiayao and Ma’nan sites are relatively early, dating back to 5300–5000 cal yr BP, falling in the Yangshao cultural period. Five Longshang-Culture-affiliated sites, (Xiaojiamao, Baiyan, Panhe, Youyao and Dakou) were selected. Six sites date to the Xia-Shang Period. Of these, Zhaimao, Liujiatai and Guangcun fall within the 3900–3800 cal yr BP period, followed by Shimao (3800–3700 cal yr BP), Shangguancun (3500–3400 cal yr BP), and Lijiayao (3100–3000 cal yr BP).

Table 2.

AMS¹⁴C dates for the 13 archaeological sites discussed in this study.

Site	Sample type	Lab. no.	AMS¹⁴C age (yr BP)	calibrated age BP (95.4% probability)	calibrated age BP (68.2% probability)	Reference
Majiayao	wood charcoal	OZK422	4540 ±75	5300–5093	5305–5081	Zhou (2009)
Majiayao	wood charcoal	OZK423	4450 ±60	5213–4984	5215–4977	Zhou (2009)
Ma’nan	foxtail millet	OZM452	4549 ± 32	5306–5148	5310–5086	this study
Xiaojiamao	foxtail millet	Beta-490919	4350 ± 30	4955–4863	4960–4860	this study
Baiyan	pig bone	OZP080	4215 ± 35	4834–4824	4840–4727	this study
Baiyan	pig bone	OZP081	4220 ± 35	4836–4826	4840–4820	this study
Panhe	wood charcoal	Beta-490924	3900 ± 30	4406–4299	4413–4295	this study
Dakou	human bone	OZO945	3780 ± 40	4218–4100	4225–4094	Bao et al. (2018)
Dakou	human bone	OZO946	3720 ± 45	4089–4006	4141–3992	Bao et al. (2018)
Youyao	pig bone	OZO959	3770 ± 50	4152–4096	4221–4090	this study
Youyao	sheep bone	OZO960	3750 ± 45	4147–4090	4149–4087	this study
Guangcun	wood charcoal	Beta-490922	3560 ± 30	3869–3840	3881–3836	this study
Liujiatai	wood charcoal	Beta-490925	3570 ± 30	3880–3849	3890–3840	this study
Zhaimao	wood charcoal	Beta-490923	3620 ± 30	3963–3903	3972–3895	this study
Shimao	human bone	OZO955	3540 ± 45	3858–3830	3873–3734	Bao et al. (2018)
Shangguancun	wheat	Beta-490921	3280 ± 30	3556–3478	3560–3468	this study
Lijiaya	foxtail millet	OZP044	2930 ± 40	3156–3005	3207–2960	Bao et al. (2018)

In total, we identified and quantified 2911 fragments of wood charcoal from the 13 sites, that is, a mean of 224 fragments from each site. 23 taxa, including four coniferous trees and 19 broadleaved plants, were identified. The detailed information of wood charcoal identification from each site can be found in Supplementary material 2. All percentage fragment counts are based on the total number of identified charcoal fragments found at each site.

The Guanzhong Basin sites

Charcoal results from Ma’nan in the Yangshao Period show that four wood types were predominantly used including Ziziphus, Ulmus, Quercus and Tamarix. Ziziphus (>70%) was the most common type, followed by Ulmus at roughly 18% (Figure 3). Nine wood types from samples dated to the Longshan Period were identified, with Quercus being the most abundant taxon (>70%), followed by Ulmus (11%). Acer, Ailanthus, Bambusoideae, Celtis, Maloideae, other taxa were present with proportions of less than 6%.

Figure 3.

Wood charcoal results from the Guanzhong Basin (NISP= No of identified specimens and the same goes for Figures 4 –6). (2-column).

A total of 169 fragments of charcoal from the Xia-Shang Period were identified at the Liujiatai site, dominated by Quercus (75%). Ulmus fragments comprised 16% of the assemblage from this period. Other taxa such as Acer, Maloideae, Populus/Salix, Prunus and Rhus registered very low proportions of <5% each. Eight broadleaved taxa occur at the later site of Shangguancun where Prunus was dominant (about 60%), with Quercus dropping to about 10%. Ulmus fragments were also found in relatively large numbers, comprising about 13% of the total. Acer, Ailanthus, Bambusoideae, Maloideae and Populus/Salix represented <10% each.

The eastern Loess Plateau sites

Charcoal records from the Baiyan site for the Longshan Period suggest that 11 woody taxa appeared at that time (Figure 4). Platycladus wood fragments comprise the highest proportion of conifers at about 30%, followed by Pinus at 10%. Only a very small number of Abies specimens was recovered. Quercus is the most abundant broadleaved taxon, accounting for 25% of the assemblage. Prunus and Ulmus specimens account for nearly 9% each. Other trees include Ailanthus, Bambusoideae, Maloideae, Salix/Populus and Rhamnus (<10% each). Data from Youyao (later Longshan Period) suggest that Quercus and Acer were the most commonly used fuel woods, constituting nearly 40% and 25% of the charcoal assemblage, respectively, followed by Pinus. Abies, Populus/Salix, Prunus, Rhamnus and Ulmus wood fragments were all found in smaller proportions (<6% each).

Figure 4.

Wood charcoal results from the eastern Loess Plateau. (2-column).

Eight wood taxa were identified from the Xia-Shang Period at Guangcun. Abies and Pinus, with percentages of about 10% and 5% respectively, are the only coniferous trees. Quercus is still the most common tree during this period, comprising about 35%, followed by Ulmus (18%) and Ziziphus (13%). Acer, Bambusoideae and Prunus wood fragments account for <5% each.

The northern Loess Plateau sites

During the Longshan Period Pinus, Quercus, Prunus, Rhus and Ulmus fragments were present at Xiaojiamao (Figure 5). At Dakou, nine taxa were identified, including one coniferous tree and eight broadleaved taxa. Pinus fragments recorded the highest percentage, at about 40%. Broadleaved plants mainly consisted of Alnus and Prunus (35% and 12% respectively). Other taxa, including Betula, Celtis, Maloideae and Quercus, registered proportions of <10% each.

Figure 5.

Wood charcoal results from the northern Loess Plateau (“+” means the presence of wood taxon). (2-column).

For the Xia-Shang Period, data from Zhaimao show that Pinus is still the most abundant taxon, reaching nearly 60%. Abies accounts for 18% of the assemblage. Quercus is the main broadleaved taxon, with a percentage of 12%, followed by Celtis, Prunus, and Ulmus. At Shimao, Pinus fragments account for >60% of the assemblage, followed by Ulmus. The other identified genera include Abies, Prunus and Quercus, at roughly 5% each. The proportion of Pinus at the later Lijiaya site fell to 34%, with a smaller proportion of Abies (7%). Broadleaved taxa included Picrasma, Rhus and Ulmus, with percentages of between 10%–25%, followed by Maloideae, Populus/Salix, Quercus and Rhamnus.

The western Loess Plateau sites

In this region, we identified 954 fragments of charcoal from eight samples at Majiayao. The charcoal results attest to the presence of 13 taxa including 2 conifers and 11 broadleaves (Figure 6). For conifers, Picea fluctuates between 3% and 37%, with a mean value of about 20%, much higher than that of Abies. Quercus is the most abundant broadleaved taxon, with a mean proportion of about 25%, reaching nearly 80% in layer 6. Populus/Salix, which has a mean proportion of 17% is followed by Ulmus accounting for 10%. Other taxa, like Acer, Alnus, Betula, and Castanea are found only in small numbers. In the early phase (layers 6-8), Quercus fragments increase dramatically, from 6% to 78%. Conversely, Picea rapidly drops from 37% to < 4%.

Figure 6.

Wood charcoal results from Majiayao on the western Loess Plateau. (1.5-column).

Discussion

The geographic distributions of the 23 plants identified in this study were divided into different groups, such as north temperate and cosmopolitan (Table 3). Abies, Picea, and Pinus are usually restricted to temperate regions. Platycladus is widely distributed and typically found throughout East Asia. Of the broadleaved taxa, Ailanthus and Ziziphus are normally distributed in tropical and/or subtropical areas, as are Bambusoideae. However, considering that Ailanthus and Ziziphus are found almost throughout China, only Bambusoideae are classified as characteristic of tropical-subtropical vegetation communities in this study. Alnus, Betula, Castanea, Fraxinus, Quercus and Ulmus grow mostly in temperate regions. Celtis, Maloideae and Rhamnus, are widely distributed in both tropical and temperate regions.

Table 3.

Geographical distributions of 23 taxa found on the Loess Plateau.

Genera	Geographical distributions	Genera	Geographical distributions
Abies	North Temperate	Maloideae	Cosmopolitan
Picea	North Temperate
Pinus	North Temperate & Subtropics	Picrasma	Tropical and Temperate Asia, Tropical America
Platycladus	East Asia	Populus/Salix	North Temperate & Tropical Africa/ Subarctic & Temperature to Tropical Mountains
Acer	North Temperate	Prunus	Old World Temperate
Ailanthus	Tropical & Subtropical Asia to North Australia	Quercus	North Temperate
Alnus	North Temperate	Rhamnus	Cosmopolitan
Bambusoideae	Pantropic	Rhus	North Temperate & Tropical Asia
Betula	North Temperate	Tamarix	Old World Temperate & Tropical Asia
Castanea	North Temperate	Ulmus	North Temperate
Celtis	Cosmopolitan	Zanthoxylum	Pantropic & Temperate Asia
Fraxinus	North Temperate	Ziziphus	Tropics & Subtropics

According to a survey of present-day vegetation on the Loess Plateau conducted by the Integrated investigation team of the Loess Plateau CAS, (1991), Picea and Abies mostly prefer humid and cold conditions and their distributions are usually restricted to high-latitudes and subalpine areas. Conversely, Pinus and Platycladus are usually found in relatively warm and dry environments; indeed, they dominate coniferous forests distributed along the hills of the plateau. Today, Pinus tabuliformis is the most abundant pine species and is often associated with oak forests on the southeastern Loess Plateau, while it occurs as pure pine forests in arid mountainous regions. Quercus is the dominant tree found in temperate, deciduous, broadleaved forests. On the Loess Plateau, oak woodlands are restricted to the southern and eastern parts and can be replaced by pioneer Populus or Betula forests when destroyed by human activities, such as cutting. Birch forests comprise the second largest woodland type and are found in most of the Loess Plateau’s mountainous areas today.

Reconstruction of woodland composition

Drawing on our results and previously published anthracological studies from the Loess Plateau (An et al., 2014; Sun et al., 2013, 2014, 2016; Li et al., 2016; Wang et al., 2011b) (Table 4), we infer that woodlands were widespread and characterized by significant spatial differences in distribution (Figure 7). In the Guanzhong Basin, combining our findings with previous charcoal studies from Huiduipo, Beiniu and Xiahe (Sun et al., 2014, 2016), we find that a wide range of taxa occurred in this area (Table 4). During the Yangshao Period, Quercus was the dominant taxon and some tropical-subtropical taxa, such as Fargesia and Sassafras, also occurred. Charcoal records from the Longshan Period suggest that Quercus was still the most common taxon, associated with the presence of Bambusoideae. With regard to the early Xia-Shang period, oak woodlands dominated the landscape and the diversity of broadleaved taxa does not show clear changes compared to the Longshan period. Therefore, we infer that broadleaved forests, dominated by oak and associated with some tropical-subtropical taxa like Fargesia, were the main woodland type from the Yangshao to the early Xia-Shang periods. Phytolith evidence from the Weinan section also reveals the presence of Bambusoideae during this period (Lu et al., 1994), while pollen analysis from a loess section suggests that an Artemisia grassland dominated in the mid-Holocene, implying that vegetation in the ‘Yuan’ area was very different from that in the valley (Sun et al., 1995).

Table 4.

Anthracological data from previous studies on the Loess Plateau (N: number of samples; C: charcoal fragment counts; % C: percentage of charcoal fragment counts).

Huiduipo (6500–5600 cal BP) (N = 6)¹			Beiniu (6300–4600 cal BP) (N = 5)¹			Xiahe (5000–4800 cal BP) (N = 1)²			Taosi (4300–3900 cal BP) (N = 25)³
taxa	C	% C	taxa	C	% C	taxa	C	% C	taxa	C	% C
Abies	39	7.56	Carpinus	10	1.42	Abies	17	8.1	Acer	15	1.97
Berchemia	16	3.1	Castanopsis	5	0.71	Castanopsis	5	2.38	Celtis	1	0.13
Castanopsis	13	2.52	Cornus	14	1.99	Cyclobalanopsis	35	26.19	Cornus controversa	2	0.26
Clematis	7	1.36	Cyclobalanopsis	50	7.1	Dipteronia	1	0.48	Cotinus	3	0.39
Cotoneaster	9	1.74	Fargesia	15	2.13	Fagus	2	0.95	Ehretia thyrsiflora	2	0.26
Fargesia	69	13.37	Gleditsia	8	1.14	Hypericum	1	0.48	Madhuca	1	0.13
Hibiscus	18	3.49	Melia	22	3.13	Osyris	1	0.48	Pinus tabulaformis	154	20.26
Phellodendron	3	0.58	Myrica	31	4.4	Quercus	75	35.71	Platycladus orientalis	250	32.89
Prunus	13	2.52	Olea	156	22.16	Zelkova	24	11.43	Prunus padus	4	0.53
Quercus	77	14.92	Osyris	10	1.42	Ziziphus	29	13.81	Quercus	298	39.21
Rhamnus	6	1.16	Paliurus	68	9.66	Total	190	100	Rehderodendron	3	0.39
Salix	47	9.11	Pinus	64	9.09				Salix	17	2.24
Sassafras	51	9.88	Platycarya	28	3.98				Ulmus	2	0.26
Stewartia	8	1.55	Quercus	180	25.57				Ziziphus jujuba	8	1.05
Toxicodendron	25	4.84	Rhamnus	12	1.7				total	760	100
Ulmus	62	12.02	Zelkova	31	4.4
Vitis	34	6.59	Total	704	100
Wikstroemia	19	3.68
Total	516	100
Dadiwan (5300–4900 cal BP) (N = 17)⁴			Xishanping(4800–4300 cal BP) (N = 7)⁴			Shannashuzha (5300–4800 cal BP) (N = 73)⁵			DP, GZ, JJZ, LHZ and QC (4300-3800 cal BP) (N = 13)^6,7
taxa	C	% C	taxa	C	% C	taxa	C	% C	taxa	C	% C
Abies	46	1.99	Acer	65	8.04	Picea	555	24	Abies
Acer	246	10.66	Betula	69	8.54	Pinus	481	21	Picea		4.5
Alangium	15	0.65	Carpinus	61	7.55	Acer	20	0.9	Acanthopanax
Alnus	78	3.38	Castanea	28	3.47	Ailanthus	1	0	Acer		5.1
Betula	166	7.2	Castanopsis	14	1.73	Bambusoideae	755	33	Aesculus
Castanea	22	0.95	Cerasus	14	1.73	Betula	71	3.1	Alnus
Cercidiphyllum japonicum	55	2.38	Corylus	5	0.62	Cudrania	1	0	Bambusoideae		2.4
Corylus	65	2.82	Diospyros	15	1.86	Deutzia	2	0.1	Carpinus
Cyclobalanopsis	90	3.9	Eucommia ulmoides	7	0.87	Euonymus	6	0.3	Catalpa
Ehretia	48	2.08	Indocalamus	12	1.49	Fraxinus	3	0.1	Cerasus
Eucommia ulmoides	74	3.21	Liquidambar formosana	63	7.8	Hippophae	46	2	Cercidiphyllum japonicum
Fagus	5	0.22	Padus	20	2.48	Hydrangea	2	0.1	Corylus
Fargesia	34	1.47	Phyllostachys	56	6.93	Maloideae	8	0.3	Cotoneaster
Firmiana	4	0.17	Phyllostachys glauca	61	7.55	Padus	4	0.2	Euptelea
Gymnocladus chinensis	84	3.64	Picea	86	10.64	Philadelphus	3	0.1	Fraxinus
Indocalamus	6	0.26	Picrasma	3	0.37	Populus	79	3.4	Gymnocladus chinensis
Juglans	53	2.3	Pseudotsuga sinensis	15	1.86	Prunus	23	1	Hippophae
Liquidambar formosana	103	4.46	Quercus	61	7.55	Quercus	71	3.1	Liquidambar
Lonicera	10	0.43	Toxicodendron	48	5.94	Rhamnus	3	0.1	Ostrya
Osmanthus fragrans	12	0.52	Ulmus	105	13	Salix	98	4.3	Padus
Ostrya	118	5.11	Total	808	100	Sorbus	4	0.2	Prunus		3.8
Padus	97	4.2				Ulmus	5	0.2	Pyrus
Phyllostachys	33	1.43				unidentified	50	2.2	Quercus		42.9
Picea brachytyla	7	0.3				Total	2291	100	Rhamnus
Picrasma	2	0.09							Rhus
Prunus armeniaca	3	0.13							Salix
Quercus aliena	190	8.24							Sorbus
Quercus mongolica	138	5.98							Toona
Sorbus pohuashanensis	52	2.25							Ulmus		10.2
Tilia	42	1.82							Ziziphus
Toxicodendron	102	4.42							total	653
Ulmus	200	8.67
Vaccinium	20	0.87
Xylosma racemosum	87	3.77
Total	2307	100

Sun et al. (2016); ²Sun et al. (2014); ³Wang et al. (2011b); ⁴Sun et al. (2013); ⁵Li et al. (2016); ⁶Li (2013); ⁷An et al. (2014).

Figure 7.

Reconstruction of the woodland types found on the Loess Plateau. (I) Western plateau, Quercus and Ulmus woodland, and Picea forest. (II) Tianshui Basin, Quercus woodland associated with a few tropical-subtropical taxa, and Picea forest. (III) Guanzhong Basin, Quercus woodland associated with a few tropical-subtropical taxa. (IV) Eastern plateau, Quercus and Pinus mixed woodland. (V) Northern plateau, Pinus woodland. 1, Majiayao; 2, Shannashuzha; 3, Xishanping; 4, Dadiwan; 5, Gaozhuang; 6, Daping; 7, Jiangjiazui; 8, Qiaocun; 9, Shangguancun; 10, Liujiatai; 11, Panhe; 12, Huiduipo; 13, Ma’nan; 14, Beiniu; 15, Xiahe; 16, Guangcun; 17, Taosi; 18, Baiyan; 19, Youyao; 20, Lijiaya; 21, Xiaojiamao; 22, Shimao; 23, Zhaimao; 24, Dakou. (single-column).

Around 3500 years ago, changes in wood charcoal assemblage composition are evidenced. Notably, Quercus declines sharply to only 10%, while Prunus greatly increases to nearly 60% at Shangguancun. In natural vegetation, Prunus only occurs in small numbers and is much less abundant than Quercus. The high presence of Prunus at Shangguancun, thus, is likely to have resulted from anthropogenic selection. Additionally, nuts, notably acorns, were widely exploited in East Asia (e.g. by the Jomon and Liangzhu groups) before the transition to farming, supported by the occurrence of abundant nut remains and storage pits (Hosoya, 2011; Qin et al., 2010). Wild relatives of peaches, including Prunus persica, P. davidinia, P. armeniaca and P. mira are widespread across northern China (Lu and Bartholomew, 2003). All of these species would have been likely targets for early foragers and archaeobotanical evidence suggests that P. armeniaca and P. persica were probably cultivated or heavily collected by people in the lower Yangtze River valley by 4000 cal BP (Fuller and Stevens, 2019; Zheng et al., 2014). We hypothesize that these changes in charcoal assemblage composition were likely to be anthropogenic, as humans chose to harvest certain arboreal taxa as fuel and favored the preservation of economically significant fruit trees. This may also reflect shifts in subsistence practices, as archaeobotanical data indicate that pre-Neolithic cultures in China relied heavily on acorns as a food source; the use of acorns seems to have declined as farming became more prominent – coinciding with in the decline of oak forests. Similarly, increases in Rosaceae charcoal percentages during this period may indicate greater investment in and possible cultivation of fruit-producing trees.

On the eastern plateau, charcoal results from early Longshan Period sites suggest coniferous woods mainly consisted of Platycladus and Pinus. Quercus is the dominant broadleaved tree, and tropical-subtropical species are also present. During the late Longshan stage, Quercus is still the most common broadleaved taxon, and the coniferous charcoal taxa are principally Pinus, which is also supported by wood charcoal studies at Taosi (Wang et al., 2011b) (Table 4). Subsequently, Quercus and Ulmus account for more than half of the total species in the early Xia-Shang Period and coniferous trees, including Pinus and topical-subtropical taxa, are also present. Collectively, we infer that Quercus and Pinus mixed forests remained stable for a long period on the eastern plateau. A pollen record from Gonghai Lake also suggests that Pinus and Quercus were the dominant tree taxa during 5000–3000 cal yr BP (Xu et al., 2016).

On the northern plateau, our results show that Pinus was the main species identified, followed by Alnus and Prunus during the Longshan Period. Charcoal data for the Xia-Shang Period suggest that Pinus still remained the most abundant tree, accounting for >30% of the assemblage, and up to 60% at Shimao. We suggest that pine forests were a dominant land cover type, alongside sparse broadleaved forests, which is consistent with the pollen analysis from the Daihai Lake (Xiao et al., 2004).

On the western plateau, wood charcoal reveals that spruce and oak were dominant in the regional vegetation. Spruce forests probably occurred in the hinterlands of mountain ranges, which differed from woodlands (including Quercus, Picea and some tropical-subtropical trees) distributed on the Tianshui Basin (An et al., 2014; Sun et al., 2013) (Figure 7 and Table 4). At Majiayao, the woodland communities were dominated by oak, elm and spruce through all the site’s cultural layers. In the early stages, Quercus shows a rapid increase and Picea declines quickly, indicating that the site inhabitants might have consciously selected and utilized spruce wood for fuel, thus leading to a decline of the local spruce forest. The rapid decrease of Picea is also supported by the pollen record from the same section (Zhou, 2009).

Drawing on all these anthracological studies, oak forests were largely distributed across the southern plateau, while they only occurred sporadically in the northern parts, where drought-tolerant pine forests were dominant. The eastern part appears to be a transition zone, as Quercus and Pinus constitute the two most abundant woody taxa. The western plateau, where oak was the dominant taxon, is relatively different as spruce also occurred abundantly which is usually distributed in cold and humid environments. This spatial difference likely implies that climatic conditions gradually changed from warm and wet to dry and cold from the Guanzhong Basin to the eastern and northern areas, while the western plateau was relatively cold and wet. Additionally, the differences in woodland composition appear to correspond well with regional variance in precipitation and temperature regimes, indicating that climatic conditions exerted significant selective pressure upon local woody vegetation communities. As the EAM weakens, moving from the southeast to the northwest, hygrophilous plants gradually decline while drought-tolerant taxa increase.

The northern limits of tropical-subtropical taxa

Previous paleoclimatic reconstructions suggest that climatic conditions during the mid-Holocene were relatively warm and humid, and these exerted a significant impact on vegetation (Shi and Kong, 1992; Shi et al., 1992; Zhao and Yu, 2012). How the vegetation zonation shifted during this period is a subject of debate (Shi et al., 1993; Zhu, 1988). Studies of different archives, such as pollen records, phytoliths, soil microstructure and chemical analysis, wood charcoal and literary records have indicated that, tropical-subtropical species made a clear range shift at this time (Fan and Li, 2005; Li et al., 2015; Sun et al., 2013; Yu et al., 1998; Zhao, 2003). There still remains a debate about how far the climatic zone shifted, as well as the detailed process behind the response of tropical-subtropical trees to this shift. Here, we combined current charcoal research results to discuss these issues, focusing on the movement of tropical-subtropical taxa across the Loess Plateau.

On the Loess Plateau, numerous tropical-subtropical taxa such as Bambusoideae, Fargesia, Hibiscus, Liquidambar formosana, Melia, Phyllostachys, Sassafras and Stewartia, were identified in studies of wood charcoal remains from Beiniu and Huidui (Sun et al., 2016), Xiahe (Sun et al., 2014), Xishanping and Dadiwan (Sun et al., 2013), Taosi (Wang et al., 2011b), Shannashuzha (Li et al., 2016), Daping, Gaozhuang, Qiaocun and Jiangjiazui (An et al., 2014) (Figure 8 and Table 5). Of these taxa, Bambusoideae occur at 12 sites with a frequency of 50% and are the most abundant of the tropical-subtropical taxa. Some species, like Castanopsis, Cyclobalanopsis, Olea, Osyris and Toxicodendron, have been identified by previous studies but are excluded here due to issues related to taxonomy and the level of possible taxonomic certainty based on diagnostic anatomical features, as well as questions regarding the ancient distribution ranges of certain species that are rare in China today. A few issues that require further investigation are: (1) Castanopsis is anatomically indistinguishable from Quercus; (2) most Cyclobalanopsis has been clumped into the Quercus genus and only two species, namely Cyclobalanopsis hefleriana and C. sichourensis, which only occur in Southwest China, are accepted (http://www.theplantlist.org/; Huang et al., 1999); (3) Olea is mostly distributed in southern China and is anatomically similar to Osmanthus (Chang et al., 1996); (4) Osyris is rare in China, only occurring in Tibet, Sichuan, Yunnan, Guangxi and its presence in northern China needs further proof (Xia and Gilbert, 2003); (5) Toxicodendron is generally accepted as a member of the broader Rhus clade in Anacardiaceae (diverse in East Asia) (http://www.theplantlist.org/).

Figure 8.

Reconstruction of the distribution of tropical-subtropical woods and farming sites. (a) The northern boundary of t-s woods during the Middle Holocene. (b) The distribution of tropical-subtropical (t-s) woods during the Late-Holocene. (c) The distribution of farming sites during 8000–2000 cal yr BP (see Supplementary material 3). For site number coding in subfigure a and b see Figure 7. Site number in subfigure c: 1, Majiayao; 2, Shannashuzha; 3, Xishanping; 4, Dadiwan; 5, Changshan; 6, Qiaocun; 7, Nanzuo; 8, Yanggu; 9, Yuergu; 10, Zhouyuan; 11, Anban; 12, Wujiapo; 13, Gongjiachuan; 14, Yuhuazhai; 15, Xi’an Han Tomb; 16, Xinjie; 17, Nanshantou; 18, Muwanghe; 19, Xiahe; 20, Xinglefang; 21, Muzhuzhuliang; 22, Shenggedaliang; 23, Shimao; 24, Yangjiesha; 25, Taosi; 26, Didong; 27, Nanjiaokou; 28, Bancun.

Table 5.

Representation of tropical-subtropical charcoal taxa found on the Loess Plateau.

Site	Percentage of tropical-subtropical taxa (%)	Number of all identified charcoal fragments	Tropical-subtropical taxa
Panhe	0.82	122	Bambusoideae
Huiduipo¹	24.8	516	Fargesia, Sassafras, Stewartia
Beiniu¹	2.13	704	Fargesia
Xishanping²	24.64	808	Eucommia ulmoides, Indocalamus, Liquidambar formosana, Phyllostachys, Phyllostachys glauca
Dadiwan²	16.68	2307	Ehretia, Eucommia ulmoides, Fargesia, Indocalamus, Liquidambar formosana, Phyllostachys, Xylosma
Shannashuzha³	33	2291	Bambusoideae
Daping⁴	< 10	-	Bambusoideae
Gaozhuang⁴	< 10	-	Bambusoideae
Qiaocun⁴	< 20	-	Bambusoideae, Toona
Jiangjiazui⁴	< 10	-	Bambusoideae
Baiyan	3.16	190	Bambusoideae
Guangcun	2.63	114	Bambusoideae

Sun et al. (2016); ²Sun et al. (2013); ³Li et al. (2016); ⁴An et al. (2014).

Tropical-subtropical woods are mostly found in the Guanzhong and Tianshui Basins. In the Guanzhong Basin, tropical-subtropical taxa have been found in samples spanning the whole study period. Charcoal data from Huiduipo indicate the presence of three tropical-subtropical taxa including Fargesia, Sassafras and Stewartia, representing 24.8% of the charcoal assemblage (Table 5). During the Late-Holocene, Bambusoideae are only found in small numbers. On the eastern plateau, no tropical-subtropical taxa have been identified north of the site of Baiyan.

In the Tianshui Basin, charcoal results suggest that the diversity of arboreal taxa was quite high. At Dadiwan seven tropical-subtropical plants, such as Fargesia, Indocalamus and Liquidambar formosana were found. At the relatively late Xishanping site, tropical-subtropical taxa were included by Eucommia ulmoides, Indocalamus, Liquidambar formosana and Phyllostachys. A certain number of Bambusoideae taxa are also evident at ca. 4000 cal yr BP. However, at Majiayao in this study, there is no occurrence of subtropical or tropical taxa.

The range of the tropical-subtropical taxa on the Loess Plateau represented by Bambusoideae, Eucommia ulmoides, Fargesia, Liquidambar formosana and Phyllostachys shifted significantly northward during the mid-Holocene. Bambusoideae, with the highest frequencies, were likely the pioneer species during this northward movement. Our results also suggest that the northern margins of the distribution zone of this group fell approximately along the northern peripheries of the Tianshui Basin, the Guanzhong Basin, and the central part of the eastern Loess Plateau (Figure 8a). During the Late-Holocene, with the onset of cooler and more arid conditions, both the diversity and percentages of tropical-subtropical taxa showed a rapid reduction (Figure 8b).

In addition, it is worth noting that the northern periphery of the ancient rice-growing sites is quite close to that of the tropical-subtropical taxa (Figure 8c). Referring to previous studies related to early agricultural dispersal, rice-growing sites have mostly been found in regions located south of 37°N, which are also influenced by the EAM (He et al., 2017). On the plateau, archaeobotanical rice remains are restricted to the Tianshui and Guanzhong Basins and the central part of the eastern plateau, with almost no occurrence of rice having been reported north of this line (Figure 8c) (see Supplementary material 3). This coincidence might imply that early rice planting spread widely across northern China, but that its dispersal was most probably constrained by shifting climatic conditions. The northern margins of the tropical-subtropical forests may also, therefore, represent the northerly limits of the expansion of rice-based agriculture.

Conclusion

Collectively, the evidence presented in this paper feeds into a growing understanding of shifting vegetation communities across the Loess Plateau during the second half of the Holocene. Charcoal evidence demonstrates that woodlands were widespread and geographically diverse on the plateau, in response to the EAM. At about 3500 cal yr BP, human impact had led to the transformation of vegetation composition as the dominance of oak was replaced by Prunus spp. Tropical-subtropical plants shifted northward during the mid-Holocene due to improved climatic conditions. Most notably, the charcoal data we present here cannot be wholly explained by shifting monsoonal fronts. Humans undoubtedly played an increasingly important role in influencing the distribution of species since the Holocene.

Supplemental Material

Supplementary_material_1 – Supplemental material for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses

Supplemental material, Supplementary_material_1 for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses by Hui Shen, Xiaoqiang Li, Robert Spengler, Xinying Zhou and Keliang Zhao in The Holocene

Supplemental Material

Supplementary_material_2 – Supplemental material for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses

Supplemental material, Supplementary_material_2 for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses by Hui Shen, Xiaoqiang Li, Robert Spengler, Xinying Zhou and Keliang Zhao in The Holocene

Supplemental Material

Supplementary_material_3 – Supplemental material for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses

Supplemental material, Supplementary_material_3 for Forest cover and composition on the Loess Plateau during the Middle to Late-Holocene: Integrating wood charcoal analyses by Hui Shen, Xiaoqiang Li, Robert Spengler, Xinying Zhou and Keliang Zhao in The Holocene

Footnotes

Acknowledgements

We would like to thank Jian Wang and Guilin Zhang for their help in sample collection. Comments from two anonymous reviewers are gratefully acknowledged.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (Grant No. 41730319) and the Strategic Priority Research Program of CAS (Grant No. XDB26030401).

ORCID iDs

Hui Shen

Xinying Zhou

Data Availability

We state that all the wood charcoal data from 13 archaeological sites in this study have not been published yet. We have uploaded them in this paper and agree to archive them in appropriate public repository as required.

Supplemental material

Supplemental material for this article is available online.

References

Dong

, et al. (2014) How prehistoric humans use plant resources to adapt to environmental change: A case study in the western Chinese Loess Plateau during Qijia Period. The Holocene 24: 512–517.

Asouti

Austin

(2005) Reconstructing woodland vegetation and its exploitation by past societies, based on the analysis and interpretation of archaeological wood charcoal macroremains. Environmental Archaeology 10(1): 1–18.

Asouti

Hather

(2001) Charcoal analysis and the reconstruction of ancient woodland vegetation in the Konya Basin, south-central Anatolia, Turkey: Results from the Neolithic site of Çatalhöyük East. Vegetation History and Archaeobotany 10(1): 23–32.

Asouti

Kabukcu

(2014) Holocene semi-arid oak woodlands in the Irano-Anatolian region of Southwest Asia: Natural or anthropogenic? Quaternary Science Reviews 90: 158–182.

Asouti

Kabukcu

White

, et al. (2015) Early Holocene woodland vegetation and human impacts in the arid zone of the southern Levant. The Holocene 25(10): 1565–1580.

Bao

Zhou

Liu

, et al. (2018) Evolution of prehistoric dryland agriculture in the arid and semi-arid transition zone in northern China. PLoS ONE 13(8): e0198750.

Carleton

Campbell

Collard

(2014) A reassessment of the impact of drought cycles on the Classic Maya. Quaternary Science Reviews 105: 151–161.

Chabal

Fabre

Terral

J-F

, et al. (1999) L’anthracologie. In: Bourquin-Mignot

Brochier

Chabal

Crozat

Fabre

Guibal

Marinval

Richard

Terral

Théry-Parisot

(eds) La Botanique. Paris: Errance, pp. 43–104.

Chang

Qiu

Green

(1996) Oleaceae. In: Wu

Raven

Hong

(eds) Flora of China. Vol. 15. Beijing: Sciences Press and St Louis: Missouri Botanical Garden Press, pp. 272–319.

10.

Cheng

Yang

Liu

(1992) Atlas of Wood in China. Beijing: China Forestry Publishing House (in Chinese).

11.

Cui

, et al. (2002) Vegetation reconstruction of Bronze Age by using microscopic structure of charcoals. Chinese Science Bulletin 47(23): 2014–2017.

12.

Fan

(2005) Study on bamboo distribution in Yellow River drainage area in history. Scientia Silvae Sinicae 41(3): 75–81 (in Chinese).

13.

Fuller

Stevens

(2019) Between domestication and civilization: The role of agriculture and arboriculture in the emergence of the first urban societies. Vegetation History and Archaeobotany 28: 263–282.

14.

Guo

Liu

(1994) Paleosols of the last 0.15Ma in the Weinan loess section and their paleoclimatic significance. Quaternary Science 14(3): 256–269 (in Chinese).

15.

Zhang

, et al. (2017) Prehistoric evolution of the dualistic structure mixed rice and millet farming in China. The Holocene 27(12): 1801–1811.

16.

Hosner

Wagner

Tarasov

, et al. (2016) Spatiotemporal distribution patterns of archaeological sites in China during the Neolithic and bronze age: An overview. The Holocene 26(10): 1576–1593.

17.

Hosoya

(2011) Staple or famine food? ethnographic and archaeological approaches to nut processing in East Asian prehistory. Archaeological and Anthropological Sciences 3(1): 7–17.

18.

Wang

, et al. (2009) Assessment of genetic diversity in broomcorn millet (Panicum miliaceum L.) using SSR markers. Journal of Genetics & Genomics 36: 491–500.

19.

Huang

Zhang

Bartholomew

(1999) Fagaceae. In: Wu

Raven

Hong

(eds) Flora of China. Vol. 4. Beijing: Sciences Press and St Louis: Missouri Botanical Garden Press, pp. 314–400.

20.

Integrated investigation team of the Loess Plateau CAS (1991) Vegetation Resources and Their Rational Utilization in the Loess Plateau. Beijing: China Science and Technology Press (in Chinese).

21.

Kabukcu

(2017) Woodland vegetation history and human impacts in south-central Anatolia 16,000-6500 cal BP: Anthracological results from five prehistoric sites in the Konya plain. Quaternary Science Reviews 176: 85–100.

22.

Kabukcu

(2018) Wood charcoal analysis in archaeology. In: Pişkin

Marciniak

Bartkowiak

(eds) Environmental Archaeology: Current Theoretical and Methodological Approaches. Cham: Springer, pp.133–154.

23.

(2013) Research of macro-charcoal remains at archaeological sites of Qijia culture in eastern part of Gansu province. Master Thesis, University of Lanzhou, China (in Chinese).

24.

Dong

, et al. (2016) Woodland vegetation composition and prehistoric human fuel collection strategy at the Shannashuzha site, Gansu Province, northwest China, during the middle Holocene. Vegetation History and Archaeobotany 26(1-2): 1–9.

25.

Zhou

, et al. (2003) Characteristics of vegetation in the Loess Plateau area since Holocene. Marine Geology & Quaternary Geology 23(3): 109–114 (in Chinese).

26.

Sun

, et al. (2015) Paleontological evidence of the changes of the north boundary of southern mid-subtropical zone in eastern China during Holocene. Tropical Geography 35(2): 179–185 (in Chinese).

27.

Shang

Dodson

, et al. (2009) Holocene agriculture in the Guanzhong Basin in NW China indicated by pollen and charcoal evidence. The Holocene 19(8): 1213–1220.

28.

Sun

Dodson

, et al. (2013) Vegetation characteristics in the western Loess Plateau between 5200 and 4300 cal. B.P. based on fossil charcoal records. Vegetation History and Archaeobotany 22: 61–70.

29.

Wei

Wen

, et al. (2008) Analyses on climate variation characteristics over the Loess Plateau in recent 50 years. Journal of Arid Land Resources and Environment 22(3): 57–62 (in Chinese).

30.

Liu

(1985) Loess and the Environment. Beijing: Science Press (in Chinese).

31.

Liu

Guo

, et al. (1994) Natural vegetation in the prehistoric Loess Plateau, forest or steppe? Acta Geoscientia Sinica 3–4: 226–273 (in Chinese).

32.

Liu

Kong

Lang

(2004) Plant remains found in the Dadiwan site and its implication on past environment. Cultural Relics of Central China 4: 26–30 (in Chinese).

33.

Bartholomew

(2003) Amygdalus Linnaeus, Sp. Pl. 1: 472. 1753. In: Wu

Raven

Hong

(eds) Flora of China. Vol 9. Beijing: Sciences Press and St Louis: Missouri Botanical Garden Press, pp. 391–395.

34.

Liu

Guo

(2003) Studies on the past vegetation in geological and historical period of the Loess Plateau. Chinese Science Bulletin 48(1): 1–7 (in Chinese).

35.

Liu

, et al. (1999) Phytolith record of vegetation succession in the southern Loess Plateau since late Pleistocene. Quaternary Science 4: 336–349 (in Chinese).

36.

Song

Wang

, et al. (1994) The discovery and implication of bamboo phytolith on the loess section in the Guanzhong Basin. Chinese Science Bulletin 39(24): 2257–2259.

37.

Liu

, et al. (2007) Phytoliths as quantitative indicators for the reconstruction of past environmental conditions in China II: Palaeoenvironmental reconstruction in the Loess Plateau. Quaternary Science Reviews 26: 759–772.

38.

Moskal-del Hoyo

Wachowiak

Blanchette

(2010) Preservation of fungi in archaeological charcoal. Journal of Archaeological Science 37(9): 2106–2116.

39.

National Development and Reform Commission, Ministry of Water Resources, Ministry of Agriculture and State Forestry Administration, People’s Republic of China, 2010. Programming for Comprehensive Management of the Loess Plateau (2010–2030) 69:8–10 (in Chinese).

40.

Qin

Fuller

Zhang

(2010) Modelling wild food resource catchments amongst early farmers: case studies from the lower Yangtze and central China. Quaternary Sciences 30(2): 245–261 (in Chinese).

41.

Scott

(2010) Charcoal recognition, taphonomy and uses in palaeoenvironmental analysis. Palaeogeography, Palaeoclimatology, Palaeoecology 291(2): 11–39.

42.

Shen

Zhou

Zhao

, et al. (2018) Wood types and human impact between 4300 and 2400 yr BP in the Hexi Corridor, NW China, inferred from charcoal records. The Holocene 28(4): 629–639.

43.

Shi

(1991) The distribution and changes of vegetation in China during historical period. Journal of Chinese Historical Geography 3: 43–73 (in Chinese).

44.

Shi

Kong

(1992) The Climates and Environments of Holocene Megathermal in China. Beijing: China Ocean Press (in Chinese).

45.

Shi

Kong

Wang

, et al. (1992) Climatic variation and events during the Holocene Megathermal in China. Science in China (Series B) 12: 1300–1308 (in Chinese).

46.

Shi

Kong

Wang

, et al. (1993) Climate and environment during the Holocene Megathermal in China. Science in China (Series B) 23(8): 865–873 (in Chinese).

47.

Shu

Gao

Zhao

, et al. (2006) The discussion on further study of ecological zoning on the Loess Plateau. Agricultural Research in the Arid Areas 24(3): 143–148 (in Chinese).

48.

Sun

Wei

, et al. (1998) Vegetation and environment during the late Pleistocene in Loess Plateau, China. Journal of Geomechanics 4(4): 30–42 (in Chinese).

49.

Sun

Dodson

, et al. (2013) Plant diversity of the Tianshui Basin in the western Loess Plateau during the mid-Holocene – Charcoal records from archaeological sites. Quaternary International 308-309: 27–35.

50.

Sun

Dodson

, et al. (2016) The quantitative reconstruction of temperature and precipitation in the Guanzhong Basin of the southern Loess Plateau between 6200 BP and 5600 BP. The Holocene 26(8): 1200–1207.

51.

Sun

Shang

, et al. (2014) Vegetation characteristics and palaeoclimate of Xiahe site in the southern Loess Plateau during the mid-Holocene based on fossil charcoal records. Quaternary Sciences 34(1): 27–34 (in Chinese).

52.

Sun

Song

Wang

, et al. (1995) Vegetation in the southern Loess Plateau since the last 100 ka, based on pollen record from Weinan, Shaanxi. Chinese Science Bulletin 40(13): 1222–1224 (in Chinese).

53.

Tan

Han

Cao

, et al. (2015) Holocene wildfire history and human activity from high resolution charcoal and elemental black carbon records in the Guanzhong Basin of the Loess Plateau, China. Quaternary Science Reviews 109: 76–87.

54.

Théry-Parisot

Chabal

Chrzavzez

(2010) Anthracology and taphonomy, from wood gathering to charcoal analysis. A review of the taphonomic processes modifying charcoal assemblages, in archaeological contexts. Palaeogeography, Palaeoclimatology, Palaeoecology 291(1-2): 142–153.

55.

Wang

Chen

Zhang

, et al. (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. Quaternary International 334–335: 155–164.

56.

Wang

Fan

Wang

(2011a) Precipitation trends during 1961—2010 in the Loess Plateau region of China. Acta Ecologica Sinica 31(19): 5512–5523 (in Chinese).

57.

Wang

(2011b) Analysis of wood remains from the Taosi site. Archaeology 3: 91–96 (in Chinese).

58.

(1991) The areal-types of Chinese genera of seed plants. Acta Botanica Yunnanica IV: 1–139 (in Chinese).

59.

Raven

Hong

(2013) Flora of China. Beijing: Science Press and St. Louis: Missouri Botanical Garden Press.

60.

Xia

Gilbert

(2003) Santalaceae. In: Wu

Raven

Hong

(eds) Flora of China. Vol 5. Beijing: Sciences Press and St Louis: Missouri Botanical Garden Press, pp. 208–219.

61.

Xia

Chen

, et al. (1998) High-resolution record of vegetation and climate variations in Longxi Loess Plateau during Holocene. Journal of Lanzhou University (Natural Sciences) 34(1): 119–127 (in Chinese).

62.

Xiao

Nakamura

, et al. (2004) Holocene vegetation variation in the Daihai Lake region of north-central China: A direct indication of the Asian monsoon climatic history. Quaternary Science Reviews 23: 1669–1679.

63.

Xie

Wang

, et al. (2002) Grassland vegetation in the Loess Plateau since Last Interglacial, based on molecular fossil records. Science in China (Series D) 32(1): 28–35.

64.

Chen

Zhang

, et al. (2016) Vegetation succession and East Asian Summer Monsoon changes since the last deglaciation inferred from high-resolution pollen record in Gonghai Lake, Shanxi Province, China. The Holocene 27(6): 1–12.

65.

Yao

Fang

, et al. (2002) Bamboo Culm Anatomy of China. Beijing: Science Press (in Chinese).

66.

Min

(2002) The features of precipitation in the Loess Plateau and its possible impacts on vegetation distribution. Resources Science 24(6): 55–60 (in Chinese).

67.

Prentice

Harrison

, et al. (1998) Pollen-based biome reconstructions for China at 0 and 6000 years. Journal of Biogeography 25: 1055–1069.

68.

Zhang

(1994) Relationship between forests and loess thicknesses in the Loess Plateau region. Bulletin of Soil and Water Conservation 14(6): 194–197 (in Chinese).

69.

Zhao

(2003) Soil developed in the Holocene Magathermal and climatic migration in the Guanzhong area. Scientia Geographica Sinica 23(5): 554–559 (in Chinese).

70.

Zhao

Wen

, et al. (2013) Soil erosion, conservation, and eco-environment changes in the Loess Plateau of China. Land Degradation & Development 24: 499–510.

71.

Zhao

(2012) Vegetation response to Holocene climate change in East Asian. Earth-Science Reviews 113: 1–10.

72.

Zheng

Crawford

Chen

(2014) Archaeological evidence for peach (Prunus persica) cultivation and domestication in China. PLoS One 9(9): e106595.

73.

Zhou

(2009) Mid-Holocene environment changes, agricultural activity and human adaption in Gansu-Qinghai region. PhD Thesis, Institute of Earth Environment, CAS, China (in Chinese).

74.

Zhou

Jiang

(1994) Wood Anatomy and Ultrastructure of Gymnospermous Woods in China. Beijing: China Forestry Publishing House (in Chinese).

75.

Zhou

Zhao

, et al. (2011) Early agriculture and its impact on environment during Neolithic period in the Longdong area. Chinese Science Bulletin 56(4-5): 318–326 (in Chinese).

76.

Zhu

(1988) The distribution of natural vegetation during mid-Holocene in China. Journal of Chinese Historical Geography 1: 19–43 (in Chinese).

77.

Zhu

(1994) Basic features of forest steppe in the Loess Plateau. Scientia Geographica Sinica 14(2): 122–131 (in Chinese).

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

5.80 MB

0.14 MB