Front. Earth Sci. Frontiers in Earth Science Front. Earth Sci. 2296-6463 Frontiers Media S.A. 1137528 10.3389/feart.2023.1137528 Earth Science Original Research Human planting strategies and its relation to climate change during ∼4,800–3,900 BP in the mid-lower Hulu River Valley, northwest China Wei et al. 10.3389/feart.2023.1137528 Wei Wenyu 1 Ma Minmin 1 * † Chen Guoke 2 * † Dong Jiajia 1 Wu Zekun 1 Li Haiming 3 4 Li Xiaobin 5 1 Ministry of Education Key Laboratory of Western China’s Environmental Systems, College of Earth and Environmental Sciences, Lanzhou University, Lanzhou, China 2 Gansu Provincial Institute of Cultural Relics and Archaeology Research, Lanzhou, China 3 College of Humanities and Social Development, Nanjing Agricultural University, Nanjing, China 4 Institution of Chinese Agricultural Civilization, Nanjing Agricultural University, Nanjing, China 5 Zhuanglang County Museum, Pingliang, China

Edited by: Harry F. Lee, The Chinese University of Hong Kong, China

Reviewed by: Xianyong Cao, Chinese Academy of Sciences (CAS), China

Xin Wang, Wuhan University, China

Ray Liu, British Museum, United Kingdom

 *Correspondence: Minmin Ma, mamm@lzu.edu.cn; Guoke Chen, chenguoke1980@sina.com

These authors have contributed equally to this work

 06 07 2023 2023 11 1137528 04 01 2023 07 04 2023 Copyright © 2023 Wei, Ma, Chen, Dong, Wu, Li and Li. 2023 Wei, Ma, Chen, Dong, Wu, Li and Li

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

The response of agricultural societies to global climate events during the Neolithic (e.g., 4.2 ka event) is a scientific issue of general interest. In the mid-lower Hulu River Valley of northwest China, millet cultivation became the primary subsistence during the late Neolithic. Local paleoclimate studies have detected a notable decline in temperature and precipitation around 4,400 BP (Before Present), while the Qijia culture (4,200–3,600 BP) sites far outnumber those of the Lower Changshan culture (4,800–4,400 BP) in the area. Why the intensity of millet farming groups increased when climate was relatively cold and dry, however, has not been well understood. To explore the issue, we performed archaeobotanical analysis, grain size measurement, stable isotope analysis and radiocarbon dating in the excavated sites of the Zhongtianxingfucheng (ZTXFC) and Wangjiayangwan (WJYW), which were dated to between ∼4,800–4,400 BP and ∼4,200–3,900 BP, respectively. Our results demonstrate the overall declines in the proportion, grain sizes and carbon isotope values of millets from the WJYW site compared to ZTXFC. The nitrogen isotopes of millets from the two sites are similar [foxtail millet: 6.8‰ ± 1.9‰ (ZTXFC), 7.5‰ ± 1.5‰ (WJYW); broomcorn millet: 7.3‰ ± 2.0‰ (ZTXFC), 7.5‰ ± 1.2‰ (WJYW)]. These results suggest that the degree of field management during ∼4,200–3,900 BP was lower than ∼4,800–4,400 BP in the mid-lower Hulu River Valley. Instead of improving cultivation management or altering cropping patterns, Qijia millet farmers might have adopted a strategy of expanding cultivated lands to promote the social development under a relatively cold-dry climate.

 archaeobotanical analysis grain size carbon and nitrogen isotope analysis subsistence strategy mid-lower Hulu River Valley late Neolithic Grant No. 2018YFA0606402 Grant Nos. lzujbky-2021-77 lzujbky-2021-kb01 Grant No. BK20221027 Grant No. 41871076 National Key Research and Development Program of China10.13039/501100012166 Fundamental Research Funds for the Central Universities10.13039/501100012226 Natural Science Foundation of Jiangsu Province10.13039/501100004608 National Natural Science Foundation of China10.13039/501100001809 section-at-acceptance Quaternary Science, Geomorphology and Paleoenvironment 1 Introduction

The response of human societies to global climate events during the Neolithic era, such as the 8.2 ka event, the 5.5 ka event and the 4.2 ka event, is a widely studied multidisciplinary issue (e.g., Flohr et al., 2016; Goldsmith et al., 2017; Wu et al., 2018; Park et al., 2019; Zhao, 2020). With the development and expansion of agriculture across Eurasia (Liu X et al., 2019; Dong et al., 2022a), global population significantly increased with the extensive expansion of farmer habitats during the late Neolithic (Styring et al., 2017; Dong et al., 2022b), which resulted in the rise of survival stress and social vulnerability to climate change (Kelly et al., 2012). Therefore, the impact of the 4.2 ka event on social evolution in the Old World has been intensively discussed in recent decades (Weiss, 2016; Ran and Chen, 2019; Manning et al., 2020). The 4.2 ka event has been proposed as an important trigger for the collapse of ancient civilizations in Mesopotamia (Cullen et al., 2000; Weiss, 2017) and India (Staubwasser et al., 2003). Additionally, the 4.2 ka event is suggested to have transformed the human settlement patterns in late Neolithic China (Xiao et al., 2019; He et al., 2022), resulting in the expansion and shrinkage of areas settled by millet farming groups and rice farming groups (He et al., 2022).

The responses of millet farming groups to the climatic deterioration event around 4,200 BP in different areas of north China were diverse. For example, settlement intensity of millet farming groups in east Inner Mongolia evidently declined (Xiao et al., 2019) and increased in the Yi-Luo River Valley and Gansu-Qinghai region (Liu and Feng, 2012; Liu L et al., 2019; He et al., 2022). In the mid-lower Hulu River Valley (MLHRV) of the western Loess Plateau, northwest China, intensive rain-fed agriculture was focused on millet cultivation since ∼5,900 BP (Barton, 2009), and was the primary subsistence in the area during the late Neolithic period (Li et al., 2022a; Yang J et al., 2022; Yang Y et al., 2022). Paleoclimate studies from a lake in the nearby Liupan Mountains suggest that temperature and precipitation in the MLHRV declined around 4,400 BP and the overall climate trend was colder and dryer during ∼4,400–3,600 BP than ∼4,800–4,400 BP (Zhao et al., 2010; Chen et al., 2015). However, the third national archaeological survey in the MLHRV revealed that the site numbers of the Lower Changshan culture (4,800–4,400 BP) was much smaller than the Qijia culture (4,200–3,600 BP) (Figure 1; Li et al., 1993). This indicates that millet farming groups substantially expanded in the area under a relatively cold-dry climate. Previous studies have suggested that Neolithic groups might have successfully adapted to climate events in different ways, including the alteration of cropping patterns (Pokharia et al., 2017; Chen et al., 2020; Li R et al., 2020), the improvement of field management (Masi et al., 2014; Ren et al., 2021) and the enlargement of the cultivated land area (An et al., 2021). Nevertheless, differences in planting strategies between the Lower Changshan period and the Qijia period in the MLHRV remains unclear due to the absence of systematic archaeobotanical studies.

Distribution of the Lower Changshan and Qijia cultural sites in the mid-lower Hulu River Valley.

Recent archaeobotanical and stable isotope analysis of crops remains from investigated sites in the MLHRV provided a valuable dataset to facilitate understanding of cropping pattern variations and water and soil management for crops since ∼6,000 BP (Li et al., 2022b). However, data from the excavated sites of the Lower Changshan and Qijia cultures are still scarce. Moreover, the measurement of millets grain sizes from archaeological sites in the MLHRV has not been reported, which is valuable to evaluate cropping strategies in prehistory (Motuzaite-Matuzeviciute et al., 2012; Bao et al., 2018). Recent excavations at the Zhongtianxingfucheng (ZTXFC) site of the Lower Changshan culture and the Wangjiayangwan (WJYW) site of the Qijia culture provide a rare opportunity to study strategies for millet farming and their relation to climate change during these two periods. In this paper, we report the results of radiocarbon dating, identification of plant remains, grain size measurement and carbon/nitrogen isotope analysis of charred millets grains from ZTXFC and WJYW sites. Results are then integrated with published archaeological and paleoclimate data to explore how millet farmers in the MLHRV responded to climate change during the Lower Changshan and Qijia periods.

 2 Study areas

The Lower Changshan culture is named after the discovery of the lower cultural remains at the Changshan site in Zhenyuan, Gansu (Hu, 1981). This has elements of the Yangshao culture and features of the Qijia culture, indicating a transitional stage from Yangshao culture to Qijia culture (Hu, 1991; Li et al., 1993). The Lower Changshan culture is mainly distributed in the Loess Plateau, such as Longdong area and southern Ningxia, and is dated to 4,800–4,400 BP in the Gansu (Li et al., 1993). The Lower Changshan cultural potteries includes clay potteries and coarse potteries, which are primarily orange and reddish brown (Hu, 1981). These potteries were mainly decorated with basket pattern and pile pattern (Li et al., 1993). The burial forms include vertical pit graves and earth-caved tombs (Li et al., 1993; Wei, 2021), and the skeletal position is mainly flexed and lying on the side (Wei, 2021). The Qijia culture is widely distributed in the Gansu-Qinghai region and likely developed from the Lower Changshan culture through the Caiyuan type and was influenced by the Keshengzhuang culture (Wang, 2012). Qijia cultural potteries include clay red potteries, gray and red coarse potteries (Ye, 2014), decorated with cord-marking and vertical basket veins (Li et al., 1993; Underhill, 2013). The burial forms mainly include vertical pit graves, vertical and side chambers (Xie, 1986; Qian et al., 2014; Yang, 2017), and the skeletal position is mainly supine and extended and flexed and lying on the side (Chen, 2003; Underhill, 2013). The Qijia culture, dating to 4,600–3,500 BP, is mainly centered around 4,300–3,900 BP (Wang, 2012), and is dated between 4,200 and 3,600 BP in the mid-lower Hulu River Valley (Li et al., 2022b; Yang Y et al., 2022). Based on limited studies (Chen et al., 2020; Li et al., 2022b), the Lower Changshan and Qijia populations were mainly engaged in millet farming, supplemented by pig raising and hunting (Xie, 1975; The Institute of Archaeoloogy Chinese Academy of Social Sciences, 1999; Womack et al., 2021).

The Hulu River (34.72°–36.5°N, 105.5°–106.5°E) is located on the western Loess Plateau, with the Liupan Mountains to the east and the Qinling Mountains to the south (Han et al., 2020). The terrain gradually decreases from north to south and from east to west, and the vast majority of the basin is a loess hilly area with loose soil and sparse vegetation (Wang F et al., 2022). The study area is characterized by a temperate continental monsoon climate with relatively pronounced seasonal changes-hot and rainy in summer, rapidly cooling in autumn, and cold in winter (Han et al., 2020). According to the meteorological stations located on the mid-lower Hulu River (Jingning, Zhuanglang, and Qin’an), the mean annual temperature is 9.03°C and the mean annual precipitation is 444.3 mm (see http://data.cma.cn). The Hulu River is the largest tributary in the upper reaches of the Wei River. The basin is crisscrossed by ravines with a well-developed river composed of many tributaries, such as the Zhuanglang River, Shuiluo River, Qingshui River (Xin et al., 2016; Wang L et al., 2022).

The Zhongtianxingfucheng site (35.2°N, 106.0°E) is in the hilly and ravine region of the Loess Plateau, located on the terrace of the Shuiluo River in Zhuanglang county. In 2019, the Gansu Provincial Institute of Cultural Relics and Archaeology conducted rescue excavation of the ZTXFC site, with an excavated area of 2,100 m2. The main part of the ZTXFC site is the Lower Changshan Period, and items including pottery, stone implements, bones, and teeth were unearthed. There are a wide variety of vessel classifications, the most common types are sand monaural pots excavated in ash pits, which could likely be early relics of the ZTXFC site from the Lower Changshan culture (Yang, 2012).

The Wangjiayangwan site (35.3°N, 105.9°E) is located in Zhuanglang county, situated on the terrace of the Zhuanglang River. In 2019, the Gansu Provincial Institute of Cultural Relics and Archaeology also conducted rescue excavation of the WJYW site, and the excavation area of WJYW is 3,100 m2. The site includes two periods: the Qijia culture and the Qing Dynasty (1,644–1,912 AD). The later relics are located in the cultivated layer and come from the Qing Dynasty. Based on the distinctive convex shape with the white-grey floor as the building structure and the large double-ear pots unearthed in the house relics, it can be concluded that the early relics of the WJYW site are from the Qijia culture (Womack et al., 2017).

 3 Materials and methods

To illuminate cropping patterns changes in the mid-lower Hulu River Valley during ∼4,800–3,900 BP, we excavated two sites in 2019 within the study region shown in Figure 1. We used a targeted sampling strategy to sample the sediments in each layer of each unearthed relic unit. We sampled as far as possible to the middle of each layer to avoid inter-layer disturbance. Given the two sites were severely damaged, there were 84 samples collected from the WJYW site, all from house relics, and 18 samples were collected from the ZTXFC site and 16 samples of the samples were from ash pits and two of the samples were from house relics and kiln.

A total of 102 flotation soil samples were collected from two sites. The total amount of flotation soil collected from WJYW was 978.5 and 181 L from ZTXFC, with an average of 11.37 L per sample and a total of 1,159.5 L. The collected soil samples were all floated using the manual bucket flotation technique (Zhao, 2010). The floated objects with a specific gravity lighter than water, such as charcoal and charred plant seeds, floated upward. The objects were gathered with 0.2 mm aperture sieves, wrapped in gauze, and hung in a shady and cool area for desiccation. Then they were sorted through 0.35, 0.7, 1.2, and 4.0 mm mesh sieves. All seeds were selected using a ×40 stereo microscope (Zhao, 2010). The charred plant seeds were identified in the Environment Archaeology Laboratory, Lanzhou University.

One bone sample and three charred plant grain samples from the two sites were selected for accelerator mass spectrometry (AMS) radiocarbon dating (Table 1). One charred broomcorn millet seeds sample was measured at Beta Analytic in Miami, United States. The other samples were dated at the MOE Key Laboratory of Western China’s Environmental Systems, Lanzhou University. The IntCal 20 calibration curve (Reimer et al., 2020) and the Libby half-life of 5,568 years were used in the calculation of all dates and the calibration was performed using OxCal v.4.4.4 (Ramsey, 2021). All ages are reported as ‘cal. yr BP’.

Calibrated radiocarbon dates from ZTXFC and WJYW sites.

Site Lab number Sampling feature Dated material 14C age Calibrated age (cal yr BP) 2 σ Culture Dating method References
ZTXFC LZU20315 Ash pit Sheep bone 4,010 ± 30 4,567–4,414 Lower Changshan AMS Dong J et al. (2022)
ZTXFC LZU20320 Tomb Human bone 4,040 ± 30 4,612–4,418 Lower Changshan AMS Dong J et al. (2022)
ZTXFC LZU20321 Tomb Human bone 4,080 ± 20 4,795–4,446 Lower Changshan AMS This study
WJYW Beta567015 House relic Broomcorn millet 3,730 ± 30 4,221–3,981 Qijia AMS This study
WJYW LZU20157 House relic Pig bone 3,700 ± 20 4,144–3,976 Qijia AMS Dong J et al. (2022)
WJYW LZU21266 House relic Foxtail millet 3,650 ± 20 4,082–3,895 Qijia AMS This study
WJYW LZU21249 House relic Wheat — — Modern AMS This study

To analyze the structure of past agricultural activities in the study area, the number of different plant species was recorded. However, different crop plants can vary considerably in weight and behave differently during harvesting, utilization, and carbonization (Yang et al., 2011). Therefore, simple ratios produced between species may not accurately correspond to the actual proportion of different species in use. To counteract this, we used a modified method of the Weight Ratio Function for different crops, proposed by Zhou et al. (2016) and Sheng et al. (2018). The Weight Ratio Function takes the average weight of 1,000 grains of the two main crops as conversion factors for estimating the actual yield percentage. This calculation is based on the results of the statistical analysis of the flotation samples (Eq. 1). P S = N s × F s N 1 × F 1 + N 2 × F 2

Where N1 = number of foxtail millet grains, F1 = 2.6, N2 = number of broomcorn millet grains, F2 = 7.5, Ns = number of that certain crop, Fs = conversion factors of that certain crop, and P(S) = actual yield percentage of that certain crop.

The length, width, and thickness of the millet seeds were measured using a vernier caliper at the MOE Key Laboratory of Western China’s Environmental Systems, Lanzhou University, Gansu Province, China. 462 mature and intact charred millets were selected for particle size measurement. The millets were placed in the sand tray in a suitable position. The longest, widest, and thickest parts of the millet seeds were measured under a stereoscopic microscope (Olympus SZX16), respectively, and recorded these data.

Forty-three randomly selected charred millet seeds were subjected to isotopic analysis. The charred millets were put into a test tube with 0.5 mol/L hydrochloric acid and then placed in a pot of water at 80°C for 30 min. Next the millets were repeatedly rinsed with pure water several times to neutral and then dried in an oven at 70°. Finally, the samples were ground into powder with a mortar and put into tinfoil bags. The δ13C and δ15N values of millet samples were measured with an automated carbon and nitrogen analyzer coupled with a Thermo Finnigan Flash DELTAplus XL mass spectrometer (Finnigan, Germany) at the MOE Key Laboratory of Western China’s Environmental System at Lanzhou University. After ten samples, a standard (Graphite, δ13C: −16.0‰; Protein, δ15N: 5.94‰) was inserted into the sample list for calibration and stability monitoring. The analytical precision of the δ13C and δ15N values were ±0.2‰. All C and N isotopes were measured relative to Vienna Pee Dee Belemnite (V-PDB) and Ambient Inhalable Reservoir (AIR) standards, respectively.

In order to better use plant carbon isotope indicators to reflect ancient human management practices for agriculture, we used Farquhar et al. (1982) to air-correct δ13C and obtain Δ13C values (Eq. 2). Where δ13Cair represents the δ13Cair value in air at that time (Leuenberger et al., 1992; Francey et al., 1999; Ferrio et al., 2005), and δ13Cplant represents the δ13C value in plants. Δ 13 C = δ 13 C a i r − δ 13 C p l a n t 1 + δ 13 C p l a n t

4 Results 4.1 AMS radiocarbon dating

The calibrated 14C ages are given with 2σ age ranges in Table 1; Figure 6D. Three bone samples were dated to between 4,795 and 4,414 cal yr BP, corresponding to the Lower Changshan period. One bone sample and two charred plant remain samples were dated to between 4,221 and 3,895 cal yr BP, corresponding to the Qijia period. One charred wheat was dated to be modern in the WJYW site, and it may have been redeposited from the upper sediments (Table 1).

 4.2 Crop assemblages in different periods in the mid-lower Hulu River Valley during ∼4,800–3,900 BP

A total of 8,258 charred crop seeds were identified from 102 flotation samples collected from two sites (Supplementary Table S1), including: 6,935 foxtail millet seeds (Setaria italica; Figure 2A), 558 broomcorn millet seeds (Panicum miliaceum; Figure 2B), 11 wheat seeds (Triticum aestivum; Figure 2C), and 16 barley seeds (Hordeum vulgare; Figure 2D). The remaining 738 seeds were composed of uncultivated species or weed remains (such as Setaria viridis, Melilotus suaveolens, Atriplex patens, Salsolacollina pall, Avena fatua, Kochia scoparia, Digitaria sanguinalis, Carex tristachya, Rumex acetosa, Galium tricorne and so on.) (Figure 2E–P). The wheat/barley are modern seeds from house relics in WJYW site, and they may have been redeposited from younger cultural layers. Therefore, we will not discuss it here, which has no influence on the final results and discussion.

Charred plant seeds collected from ZTXFC and WJYW sites (scale bar: 1 mm): (A) Setaria italica, (B) Panicum miliaceum, (C) Triticum aestivum, (D) Hordeum vulgare, (E) Setaria viridis, (F) Astragalus membranaceus, (G) Melilotus suaveolens, (H) Atriplex patens, (I) Salsolacollina pall, (J) Avena fatua, (K) Kochia scoparia, (L) Digitaria sanguinalis, (M) Carex tristachya, (N) Rumex acetosa, (O) Perilla frutescens, (P) Galium tricorne.

Using the new additions to the archaeological record, linked both by agricultural practice and chronologically defined context, it is possible to chart agricultural development during ∼4,800–3,900 BP in the MLHRV (Supplementary Table S1; Figure 6). 2,599 foxtail millet seeds (77.21% of weight) and 266 broomcorn millet seeds (22.79% of weight) were identified from 18 samples (181 L of soil in total) in the ZTXFC site. This indicates the predominance of a millet-based agriculture during the 4,800–4,400 BP. Between 4,200 and 3,900 BP, a mixed farming practice gradually emerged. 4,336 foxtail millet seeds (83.73% of weight), and 292 broomcorn millet seeds (16.27% of weight) were collected from 84 samples totaling 978.5 L of soil. 11 wheat and 16 barley grains were identified from WJYW, which was probably due to a disturbed context given that similar issues were also encountered in previous studies (Table 1; e.g.; Dodson et al., 2013; Jia et al., 2013). It is worth noting that the dominant species (foxtail millet) continues to account for four-fifths of the total seeds in the assemblage (Figure 3).

Results of archaeobotanical analysis from the ZTXFC and WJYW sites.

 4.3 Size of foxtail and broomcorn millet remains in the mid-lower Hulu River Valley during ∼4,800–3,900 BP

A total of 326 foxtail millets and 136 broomcorn millets were measured and the mean values, ranges and standard deviations of length, width and thickness were calculated in this study (Table 2; Figure 4). The length, width, and thickness of the foxtail millets from ZTXFC site (4,800–4,400 BP) ranged from 0.90 to 1.62 mm (mean = 1.33 ± 0.11 mm), from 0.89 to 1.33 mm (mean = 1.12 ± 0.09 mm), and from 0.58 to 1.29 mm (mean = 0.91 ± 0.14 mm), respectively. For the broomcorn millets, the same measurements ranged from 1.38 to 1.96 mm (mean = 1.61 ± 0.13 mm), from 1.32 to 1.89 mm (mean = 1.53 ± 0.13 mm), and from 1.03 to 1.80 mm (mean = 1.32 ± 0.15 mm), respectively. While the length, width, and thickness of the foxtail millets from WJYW site (4,200–3,900 BP) ranged from 0.95 to 1.99 mm (mean = 1.36 ± 0.18 mm), from 0.74 to 1.77 mm (mean = 1.12 ± 0.14 mm), and from 0.50 to 1.55 mm (mean = 0.84 ± 0.18 mm), respectively. The broomcorn millets ranged from 1.20 to 2.20 mm (mean = 1.65 ± 0.20 mm), from 0.97 to 1.84 mm (mean = 1.44 ± 0.19 mm), and from 0.63 to 1.74 mm (mean = 1.24 ± 0.22 mm), respectively. Statistical analysis of the grain size data of foxtail millets and broomcorn millets by SPSS showed that the length and thickness of the foxtail millets were statistically significant. For broomcorn millets width and thickness were statistically significant.

Statistical results regarding length, width, and thickness of single millet from ZTXFC and WJYW sites.

Site Species Type Number Mean (mm) SD Range (mm)
ZTXFC Foxtail millet Length 186 1.33 0.11 0.90–1.62
Width 186 1.12 0.09 0.89–1.33
Thickness 186 0.91 0.14 0.58–1.29
ZTXFC Broomcorn millet Length 63 1.61 0.13 1.38–1.96
Width 63 1.53 0.13 1.32–1.89
Thickness 63 1.32 0.15 1.03–1.80
WJYW Foxtail millet Length 140 1.36 0.18 0.95–1.99
Width 140 1.12 0.14 0.74–1.77
Thickness 140 0.84 0.18 0.50–1.55
WJYW Broomcorn millet Length 73 1.65 0.20 1.20–2.20
Width 73 1.44 0.19 0.97–1.84
Thickness 73 1.24 0.22 0.63–1.74

Scatter diagram and box plot detailing the length, width, and thickness of foxtail and broomcorn millets from ZTXFC and WJYW sites. (A, B) represent the result of foxtail millets; (C, D) represent the results of broomcorn millets.

 4.4 Carbon and nitrogen isotopes of foxtail and broomcorn millet in the mid-lower Hulu River Valley during ∼4,800–3,900 BP

Stable isotopic analysis of the 22 foxtail millet samples and 21 broomcorn millet samples are presented in Table 3 and Figure 5. The broomcorn millet has a wider isotopic range than the foxtail millet in terms of δ13C values, while the δ15N values are opposite. During the Lower Changshan culture period (4,800–4,400 BP), the δ13C and Δ13C values of foxtail millet samples ranged from −9.6‰ to −8.8‰ (mean = −9.2‰ ± 0.2‰, n = 11) and 2.5‰–3.3‰ (mean = 3.0‰ ± 0.2‰). δ15N ranged from 4.9‰ to 12.1‰ (mean = 6.8‰ ± 1.9‰, n = 11). The δ13C and Δ13C values of broomcorn millet samples ranged from −10.6‰ to −9.3‰ (mean = −9.7‰ ± 0.3‰, n = 10) and 3.0‰–4.4‰ (mean = 3.4‰ ± 0.3‰), respectively. δ15N ranged from 5.0‰ to 13.0‰ (mean = 7.3‰ ± 2.0‰, n = 10). In the foxtail millet samples from the Qijia Culture period (4,200–3,900 BP), the δ13C, Δ13C, and δ15N values ranged from −9.8‰ to −8.8‰ (mean = −9.5‰ ± 0.3‰, n = 11), from 2.4‰ to 3.4‰ (mean = 3.1‰ ± 0.3‰), and from 5.7‰ to 11.5‰ (mean = 7.5‰ ± 1.5‰, n = 11), respectively. The δ13C, Δ13C, and δ15N values of broomcorn millet samples ranged from −11.7‰to −9.7‰ (mean = −10.2‰ ± 0.6‰, n = 11), from 3.3‰ to 5.3‰ (mean = 3.8‰ ± 0.6‰), and from 5.9‰ to 10.7‰ (mean = 7.5‰ ± 1.2‰, n = 11), respectively.

Summary results of δ13C and δ15N of millets from ZTXFC and WJYW sites.

Site Species Number δ13C (‰) δ15N (‰)
Mean SD Range Mean SD Range
ZTXFC Foxtail millet 11 −9.2 0.2 −9.6–−8.8 6.8 1.9 4.9–12.1
ZTXFC Broomcorn millet 10 −9.7 0.3 −10.6–−9.3 7.3 2.0 5.0–13.0
WJYW Foxtail millet 11 −9.5 0.3 −9.8–−8.8 7.5 1.5 5.7–11.5
WJYW Broomcorn millet 11 −10.2 0.6 −11.7–−9.7 7.5 1.2 5.9–10.7

Scatter diagram of carbon and nitrogen isotope values for millets from ZTXFC and WJYW sites.

 5 Discussion 5.1 Human planting strategies in the mid-lower Hulu River Valley during ∼4,800–3,900 BP

The results of archaeobotanical analysis from the ZTXFC and WJYW sites (Supplementary Table S1; Figure 3) show that foxtail millet was the dominant cultivated crop, and broomcorn millet acted as the auxiliary crop in the MLHRV during both ∼4,800–4,400 BP and ∼4,200–3,900 BP. Although the charred wheat was found in the sediment of the Xishanping site that dated was to 4,650 BP, the wheat was not directly dated (Li et al., 2007). Wheat remains excavated at the Gaozhuang site in Zhuanglang County were dated to 3,561–3,405 BP (Li, 2018), which is the earliest evidence for wheat appearing in the MLHRV. Though few charred grains of wheat and barley were identified from the WJYW site, the direct radiocarbon date of one charred wheat grain suggests that these West Asian domesticated crop remains were likely derived by disturbance from modern periods (Table 1). This phenomenon was also reported in archaeobotanical studies from other Neolithic sites of North China (Zhang et al., 2018). This indicates that humans may have not yet incorporated wheat and barley into their cropping patterns in the Hulu River Valley during this period. Nevertheless, the radiocarbon date of foxtail millet remains from the same flotation sample that yielded the wheat/barley (Table 1) reveals these remains of indigenous crop were utilized in Qijia period. In addition, other Qijia sites in the study area mainly grew foxtail and broomcorn millets, and only a very small amount of barleys were excavated from the Gaozhuang site (Li, 2018). Isotopic evidence also suggests that C4 crops were mainly consumed in the western Loess Plateau during 5,300–4,000 BP (Dong J et al., 2022). The number and mass percent of foxtail millet remains in total plant remains in the ZTXFC and WJYW sites account for 87.95%/77.21% and 82.18%/83.73%, respectively. The same values for broomcorn millet at these two sites were 9.00%/22.79% and 5.53%/16.27%, respectively. This suggests that cropping patterns in the MLHRV were roughly similar during the Lower Changshan and Qijia periods. The significance of foxtail millet in plant subsistence strategy was slightly higher in the Qijia period than in the Lower Changshan period.

The increased weight of foxtail millet relative to broomcorn millet as a cropping strategy in the MLHRV during the Qijia period in comparison to the Banpo-Miaodigou period (∼6,100–5,500 BP) and the late Yangshao period (∼5,500–4,800 BP) were reported in previous archaeobotanical studies (Li et al., 2022b; Yang Y et al., 2022). Our results from the ZTXFC site are the first reported archaeobotanical data of the Lower Changshan culture. The proportion of weed remains in the plant remains at the ZTXFC site is only 3.05%, while the WJYW site reaches ∼12.28%. To reduce sampling error (houses vs. ash pits), we compared the Lower Changshan data with all data derived from various Qijia residential contexts in the Western Loess Plateau. We found that the weed proportion of ash pits derived from other Qijia sites was 9%–16.8% (Yang, 2014; Li, 2018; Li et al., 2022b), which is closer to the Qijia result of this study (12.28%). It suggests that results from different residential contexts (houses vs. ash pits) may not be influenced in this case. These data likely indicate that human input to field management in the MLHRV during the Qijia period was less than in the Lower Changshan period (Figure 6E), though the cropping strategy were dominated by foxtail millet with broomcorn millet as an auxiliary during both periods.

Human planting strategies and its relation to climate change during ∼4,800–3,900 BP in the mid-lower Hulu River Valley, northwest China, compared with climate records, radiocarbon dates and ∆13C values of millets. (A) Summer temperature reconstruction on the Tibetan Plateau (Zhang et al., 2022). (B) Mean annual air temperature (MAAT/°C) record from the Agassiz ice cap (Lecavalier et al., 2019). (C) Pollen-based annual precipitation (PANN/mm) reconstructed from Tianchi Lake in the past 6,000 years (Zhao et al., 2010; Chen et al., 2015). (D) Radiocarbon dates from the ZTXFC and WJYW sites. (E) Crop number/Weed number, n is the sum of crops number and weeds number from two sites. (F) Frequency during ∼4,800–3,900 BP in the mid-lower Hulu River Valley. (G) Δ13C value of foxtail and broomcorn millets from ZTXFC and WJYW. VPDB, Vienna Pee Dee Belemnite.

Carbon and nitrogen isotopes of crop remains unearthed from archaeological sites were used to study ancient human management behaviors of cultivated land, for example, irrigation and fertilization (Wang et al., 2018; Li et al., 2022b; Dong Y et al., 2022; Liu et al., 2022). The δ13C value of foxtail and broomcorn millet remains from the ZTXFC and WJYW sites (Table 3; Figure 5) show that the δ13C values of millets from the WJYW site are more negative than in the ZTXFC site. The δ15N values of millets from the WJYW site overlap with and are slightly higher than in the ZTXFC site. The mean δ13C and δ15N values of foxtail millet remains are different from the broomcorn millet remains, which may be related to the physiological differences between the two species (An et al., 2015).

The δ13C values of plants might have been affected by varying atmospheric δ13C values during different periods of the Holocene (Cleveland, 1979; Leuenberger et al., 1992; Francey et al., 1999; Indermühle et al., 1999; Ferrio et al., 2005). Therefore, we air-corrected the carbon isotope values for all samples to obtain Δ13C values (Eq. (2); Farquhar et al., 1982). The Δ13C values of crops can be affected by the physiological properties of plants themselves and by external environmental factors (O’leary, 1981; O’leary, 1988; Farquhar et al., 1982; 1989; Farquhar and Richards, 1984; Wallace et al., 2013). Physiological properties of plants include photosynthesis pathway, plant species, and different parts of the same plant species (O’Leary, 1981; 1988; Farquhar et al., 1989; Hattersley and Watson, 1992; An et al., 2015). However, since foxtail and broomcorn millet are gramineous C4 plants, their Δ13C values from the ZTXFC and WJYW sites were mainly affected by environmental conditions rather than physiological difference.

Previous analysis of carbon isotopes of millet remains from the archaeological sites in the nearby Qin’an and Lixian counties suggest temperature differences are not the main factor affecting the δ13C values of millets (Ji, 2007). Therefore, water availability for the growth of millet crops in the study area is closely related to water conditions that may be affected by changes in precipitation or human irrigation behavior. Considering the long-run tradition of rain-fed farming for foxtail and broomcorn millet, even in modern North China, it is unlikely that human irrigated these drought-resistant millets during the late Neolithic. Moreover, the mean ∆13C values of millet remains from the WJYW and ZTXFC sites are lower than the mean ∆13C values of S. viridis samples grown under modern natural precipitation, indicating that artificial irrigation management was not applied during ∼4,800–3,900 BP (An et al., 2015; Li, 2018). Modern experiments on the Loess Plateau and elsewhere indicate that the Δ13C values of C4 grasses are negatively correlated with rainfall (Wang et al., 2005; Li, 2018; Sanborn et al., 2021; Dong Y et al., 2022). The Δ13C values of millet remains at the WJYW site higher than those at the ZTXFC site (Figure 6G), suggesting that water availability of millet crops in the WJYW was likely lower than ZTXFC site. The pollen percentage data from the Tianchi Lake sediment indicated that precipitation was lower during ∼4,200–3,800 BP than ∼4,800–4,400 BP (Figure 6C; Zhao et al., 2010; Chen et al., 2015). Due to the Tianchi Lake is 30 km away from these two sites, the precipitation reconstruction data derived from the Lake is a suitable and reliable record for human actives in the MLHRV. This is a further demonstration of the relatively low water availability in the Qijia period compared with Lower Changshan period, suggesting that the Δ13C values of millet remains from the ZTXFC and WJYW sites were primarily affected by the decreased precipitation (Figure 6C), instead of other factors, such as fertilization behavior.

The ranges of δ15N values for foxtail and broomcorn millet remains in the WJYW and ZTXFC sites are quite wide (Figure 5). However, no statistically significant differences of δ15N values for foxtail and broomcorn millet remains from these two sites are detected by the statistical analysis of variance (ANOVA, foxtail millet: p = 0.349; broomcorn millet: p = 0.808). This indicates the soil fertility characteristics for millet crop growth in the WJYW and ZTXFC sites were similar. The δ15N values for all samples were higher than the estimated moderate fertilization level (3.5‰, Li et al., 2022b), and significantly higher than the local vegetation baseline (1.9‰, Barton, 2009) and fertilization level from the Dadiwan site in the Yangshao period (6,500–4,800 BP, Yang et al., 2022). This suggested humans probably added organic fertilizers (animal manure, sewage, food waste, and so on) to the cultivated lands in the MLHRV during ∼4,800–4,400 BP and ∼4,200–3,900 BP, as were reported at Neolithic sites in Eurasia (Bol et al., 2005; Bogaard et al., 2007; Aguilera et al., 2008; Bogaard et al., 2013; Araus et al., 2014). However, the relationship between the δ15N values of crop remains and fertilization behaviors in the area still needs to be further examined by detailed and modern simulation experiments.

The measurement of crop remains grain size provides a perspective for studying human cultivating behavior of the major crops (Willcox, 2004; Motuzaite-Matuzeviciute et al., 2012; Bao et al., 2018). According to grain size data analysis from numerous sites in North China, foxtail and broomcorn millet grain sizes generally increased from the Neolithic to historical periods (Bao et al., 2018). The same diachronic change was also detected in the western Loess Plateau (Li, 2018), suggesting humans continuously performed breeding selection on these two indigenous crops in North China. However, we observe the opposite trend for millets grain size variation in the MLHRV from ∼4,800–4,400 BP to ∼4,200–3,900 BP which was based on the measurements of millets grain sizes at the ZTXFC and WJYW sites (Table 2; Figure 4). The difference of length and thickness of foxtail millet remians are statistically significant. The same notable divergence in broomcorn millet remains is reflected by the width and thickness. Therefore, the change of foxtail and broomcorn millet grain sizes from the Lower Changshan to Qijia culture can be evaluated by those parameters. The results show that grain sizes for both millet crops at the WJYW site were overall smaller than the ZTXFC site (Figure 4), suggesting human might have not engaged in artificial breeding of millet crops during the Qijia period. This is consistent with other investigated Qijia sites in the study region (Supplementary Table S2). Compared with the grain size of millets from the ZTXFC site (Lower Changshan culture), grain size of millets from all contexts of Qijia sites in this region were overall smaller (Supplementary Table S2).

Crop grain sizes may be influenced by multiple factors, such as domestication and hereditary properties (Li et al., 2011; Fuller et al., 2014), maturity (Motuzaite-Matuzeviciute et al., 2012) and the environmental conditions of crop growth (Araus et al., 1999; Willcox, 2004). Immaturity may be an important influencing factor for the smaller grain size of millets (Motuzaite-Matuzeviciute et al., 2012), which was mainly examined by the morphology of millets grains. According to the evaluative criteria proposed by Song et al. (2012), all measured millet remains from the WJYW and ZTXFC sites were mature millets, as the length-width ratio of foxtail and broomcorn millet remains are basically same at these two sites (Table 2). The primary factor affecting grain size of millet remains at the ZTXFC and WJYW sites may be the environmental conditions of millets growth, such as water stress and soil fertility (Araus et al., 1999; Willcox, 2004). As was mentioned above, the soil fertility for millets growth at the ZTXFC and WJYW sites were basically similar, while water availability for millets growth was lower at WJYW than ZTXFC. This further demonstrated that the decreased precipitation in the MLHRV was responsible for the reduction in the grain sizes of millet crops during the Qijia period in comparison to the Lower Changshan period.

 5.2 How millet farming groups responded to climate change in the mid-lower Hulu River Valley during ∼4,800–3,900 BP

Climate changes, especially rapid and extreme climate events, are proposed as an important factor for cultural evolution and the transformation of subsistence strategies in different corners of Eurasia during the Neolithic periods (Staubwasser et al., 2003; Wu and Liu, 2004; Weiss, 2017; Park et al., 2019; Ran and Chen, 2019; Zhang H et al., 2021), including northwest China (Dong et al., 2012; Zhou et al., 2016; Cao and Dong, 2020; Ren et al., 2021). While social resilience to climate change gradually increased with innovations and the dispersal of agricultural techniques across Eurasia, Neolithic groups could adopt different strategies to cope with climate change (Yang et al., 2019; Yang et al., 2020; Dong et al., 2021; Tao et al., 2022). In the Apulia region of southern Italy, humans altered cropping patterns as an adaptation response to the rapid dry-humid fluctuations during ∼8,500–5,700 BP (Fiorentino et al., 2013). At the Jinchankou site in Qinghai, human increased field management during ∼4,100–3,700 BP to cope with climate (Ren et al., 2021). In North China, the areas utilized for millet cultivation extensively expanded, facilitating the growth in population during and after the 4.2 ka climate event (He et al., 2022).

In the MLHRV, a climate event around 5,500 BP triggered the transition of the primary cultivated crop from broomcorn millet to the relatively high-yielding foxtail millet (Yang Y et al., 2022). This might have contributed to the development of millet-pig intensive agriculture in the area (Yang J et al., 2022). Recent studies of lake sediments in the Liupan Mountains, the closest high-resolution paleoclimate archives to the MLHRV, and temperature reconstruction from the Tibetan Plateau and Agassiz ice cap, reveal a rapid decline in temperature and precipitation at ∼4,400 BP (Figures 6A–C; Zhao et al., 2010; Chen et al., 2015; Lecavalier et al., 2019; Zhang et al., 2022). This generally corresponds with paleoclimate records in surrounding areas (Tan et al., 2020; Zhang C et al., 2021). The lowest temperature and precipitation of 4,800–3,800 BP in the MLHRV occurred between ∼4,400 and 4,200 BP, corresponding to the gap between the Lower Changshan and Qijia periods in the area (Figure 6). This suggests that the cold-dry climate during these two centuries probably led to the collapse of the Lower Changshan society in the MLHRV. This heavily relied on the production of frost-sensitive millet crops which was susceptible to climate change, especially substantial drops of temperature.

Climate turned warmer and wetter to some extent in the MLHRV around 4,200 BP, and Qijia groups settled widely in the area during the subsequent centuries (Figure 1). The number of Qijia sites in the MLHRV reached 381, which was much more than the Lower Changshan sites (8), while climate was colder and dryer during 4,200–3,800 BP than 4,800–4,400 BP. Our results of archaeobotanical analysis, grain size measurement, and stable isotope analysis at the ZTXFC and WJYW sites reveal that the cropping patterns remained consistent, while the level of farmland management was regressive from the Lower Changshan to Qijia periods. There is no evidence to suggest that humans strengthened the behavior of irrigation, fertilization or breeding during the Qijia period in comparison to the Lower Changshan period. Furthermore, the decline of precipitation during the Qijia period compared with Lower Changshan period resulted in the increase of water stress for millet growth and then the reduction of millet grains. However, both the intensity and space of human settlements in the MLHRV during the Qijia period were significantly larger than the Lower Changshan period (Figures 1, 6F). After excluding the extreme values, the ranges of δ15N values for foxtail millet and broomcorn millet remains during the Qijia period are significantly wider (ANOVA, foxtail millet: p = 0.003, broomcorn millet: p = 0.002) than the Lower Changshan period in the Western Loess Plateau (Supplementary Table S3; Yang, 2021; Li et al., 2022b). This indicated that Qijia populations cultivated millet in farmlands with different fertility conditions and expanded their farmlands. This suggests that Qijia groups may have adopted a strategy of expanding farmlands to promote social development, although climate was relatively cold and dry compared with Lower Changshan and Yangshao periods. This strategy of agricultural extensification was also witnessed in contemporaneous Yellow River Valleys (He et al., 2022). Additionally, massive emigration of millet farming groups occurred in some areas of North China, such as eastern Inner Mongolia, probably due to the spatial differences in social resilience and the amplitude of climate change influencing the growth of millet crops. Qijia population may have adopted new crops, domesticated animals, and metallurgy to cope with climate change in the later period. For instance, humans incorporated wheat and barley farming, animal herding, metallurgy, and luxury (e.g., jade) into their economic system in the Hexi Corridor (Chen, 2017; Yang et al., 2019; Cao, 2022; Ren et al., 2022) and western Loess Plateau during the Qijia period (Li, 2018; Ren et al., 2021; Cao, 2022; Wang L et al., 2022; Zhang, 2022). The diversified economic strategy enhanced Qijia social resilience to cope with climate change.

 6 Conclusion

Through archaeobotanical analysis, radiocarbon dating, grain size measurement, and stable isotope studies at the ZTXFC and WJYW sites, we reveal the characteristics of planting strategies in the MLHRV during the Lower Changshan and Qijia periods. The proportions of foxtail millet and broomcorn millet in plant remains account for 87.95%/9.00% and 82.18%/5.53%, while those of weeds are 3.05% and 12.28% in the ZTXFC and WJYW sites, respectively. This suggests that rain-fed agriculture served as the dominant plant utilization strategy, while human input to field management declined from ∼4,800–4,400 BP to ∼4,200–3,900 BP in the MLHRV. Measurements and stable isotope analysis results of carbonized millet grains demonstrates that the grain size of millets and water availability during the Qijia period was lower than in the Lower Changshan period, which was likely affected by the decline of temperature and precipitation. This pattern of planting strategies is supported from previous archaeobotanical studies in the MLHRV. These indicate that millet farming groups in the area might have neither altered cropping patterns nor improved their farmland management. Instead, they enlarged their cultivated lands to support the rapid growth of local populations during the Qijia period, when climate was relatively cold and dry in comparison to the Lower Changshan period. Our work provides a valuable case study to understand the pattern of human-environment interaction at a local scale in millet farming areas during the late Neolithic.

 Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

 Author contributions

The study was designed by WW and MM. GC, JD, and ZW conducted field works and sample collection. WW and HL completed experiments and data correction. WW and MM analyzed data and designed the figures. WW, MM, GC, JD, ZW, HL, and XL wrote the manuscript. All authors discussed the results and commented on the manuscript.

 Funding

This work was supported by the Second Tibetan Plateau Scientifific Expedition and Research Program (STEP) (Grant No. 2019QZKK0601), the National Key R&D Program of China (Grant No. 2018YFA0606402), the Fundamental Research Funds for the Central Universities (Grant No. lzujbky-2021-77), Natural Science Foundation of Jiangsu Province, China (Grant No. BK20221027), the Open Foundation of MOE Key Laboratory of Western China’s Environmental System, Lanzhou University and the Fundamental Research Funds for the Central Universities (Grant No. lzujbky-2021-kb01), the National Natural Science Foundation of China (Grant Nos. 41871076).

 Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/feart.2023.1137528/full#supplementary-material

References Aguilera M. Araus J. L. Voltas J. Rodríguez-Ariza M. O. Molina F. Rovira N. (2008). Stable carbon and nitrogen isotopes and quality traits of fossil cereal grains provide clues on sustainability at the beginnings of Mediterranean agriculture. Rapid Commun. Mass Spectrom. 22, 1653–1663. 10.1002/rcm.3501 An C. Dong W. Li H. Zhang P. Zhao Y. Zhao X. (2015). Variability of the stable carbon isotope ratio in modern and archaeological millets: Evidence from northern China. J. Archaeol. Sci. 53, 316–322. 10.1016/j.jas.2014.11.001 An J. Kirleis W. Jin G. (2021). Understanding the collapse of the longshan culture (4400–3800 BP) and the 4.2 ka event in the haidai region of China – from an agricultural perspective. Environ. Archaeol., 1–15. 10.1080/14614103.2021.2003583 Araus J. L. Febrero A. Catala M. Molist M. Voltas J. Romagosa I. (1999). Crop water availability in early agriculture: Evidence from carbon isotope discrimination of seeds from a tenth millennium BP site on the euphrates. Glob. Change Biol. 5, 201–212. 10.1046/j.1365-2486.1999.00213.x Araus J. L. Ferrio J. P. Voltas J. Aguilera M. Buxó R. (2014). Agronomic conditions and crop evolution in ancient Near East agriculture. Nat. Commun. 5, 3953. 10.1038/ncomms4953 Bao Y. Zhou X. Liu H. Hu S. Zhao K. Atahan P. (2018). Evolution of prehistoric dryland agriculture in the arid and semi-arid transition zone in northern China. PLoS One 13, e0198750. 10.1371/journal.pone.0198750 Barton L. W. (2009). Early food production in China's western Loess Plateau. [Davis (USA)]: University of California. Bogaard A. Fraser R. Heaton T. H. E. Wallace M. Vaiglova P. Charles M. (2013). Crop manuring and intensive land management by Europe’s first farmers. Proc. Natl. Acad. Sci. U. S. A. 110, 12589–12594. 10.1073/pnas.1305918110 Bogaard A. Heaton T. H. E. Poulton P. Merbach I. (2007). The impact of manuring on nitrogen isotope ratios in cereals: Archaeological implications for reconstruction of diet and crop management practices. J. Archaeol. Sci. 34, 335–343. 10.1016/j.jas.2006.04.009 Bol R. Eriksen J. Smith P. Garnett M. H. Coleman K. Christensen B. T. (2005). The natural abundance of 13C, 15N, 34S and 14C in archived (1923–2000) plant and soil samples from the Askov long‐term experiments on animal manure and mineral fertilizer. Rapid Commun. Mass Spectrom. 19, 3216–3226. 10.1002/rcm.2156 Cao F. (2022). Characteristics and processes of jade-using during the prehistoric period in the Gan-Qing area (in Chinese with English Abstract). Sichuan Cult. Relics 221, 43–59. Cao H. Dong G. (2020). Social development and living environment changes in the Northeast Tibetan Plateau and contiguous regions during the late prehistoric period. Reg. Sustain. 1, 59–67. 10.1016/j.regsus.2020.09.001 Chen F. Xu Q. Chen J. Birks H. J. B. Liu J. Zhang S. (2015). East Asian summer monsoon precipitation variability since the last deglaciation. Sci. Rep. 5, 11186. 10.1038/srep11186 Chen G. (2017). Metallurgical community of xichengyi-qijia culture: Early metal-making specialists in the hexi corridor and related issues (in Chinese with English abstract). Archaeol. Cult. Relics 05, 37–44. Chen H. (2003). Analysis on the posture of the human bones seen in the pre-historic graves in Gansu and Qinghai (in Chinese with English abstract). J. Anc. Civilizations 2, 138–153. Chen T. Qiu M. Liu R. Li H. Hou H. Howarth P. (2020). Human responses to climate change in the late prehistoric Western Loess Plateau, northwest China. Radiocarbon 62, 1193–1207. 10.1017/RDC.2020.32 Cleveland W. S. (1979). Robust locally weighted regression and smoothing scatterplots. J. Am. Stat. Assoc. 74, 829–836. 10.1080/01621459.1979.10481038 Cullen H. M. deMenocal P. B. Hemming S. Hemming G. Brown F. H. Guilderson T. (2000). Climate change and the collapse of the Akkadian empire: Evidence from the deep sea. Geology 28, 379–382. 10.1130/0091-7613(2000)028<0379:ccatco>2.3.co;2 Dodson J. R. Li X. Zhou X. Zhao K. Sun N. Atahan P. (2013). Origin and spread of wheat in China. Quat. Sci. Rev. 72, 108–111. 10.1016/j.quascirev.2013.04.021 Dong G. Du L. Wei W. (2021). The impact of early trans-Eurasian exchange on animal utilization in northern China during 5000–2500 BP. Holocene 31, 294–301. 10.1177/0959683620941169 Dong G. Du L. Yang L. Lu M. Qiu M. Li H. (2022a). Dispersal of crop-livestock and geographical-temporal variation of subsistence along the Steppe and Silk Roads across Eurasia in prehistory. Sci. China Earth Sci. 65, 1187–1210. 10.1007/s11430-021-9929-x Dong G. Jia X. An C. Chen F. Zhao Y. Tao S. (2012). Mid-Holocene climate change and its effect on prehistoric cultural evolution in eastern Qinghai Province, China. Quat. Res. 77, 23–30. 10.1016/j.yqres.2011.10.004 Dong G. Lu Y. Zhang S. Huang X. Ma M. (2022b). Spatiotemporal variation in human settlements and their interaction with living environments in Neolithic and Bronze Age China. Prog. Phys. Geogr. 46, 949–967. 10.1177/03091333221087992 Dong J. Wang S. Chen G. Wei W. Du L. Xu Y. (2022). Stable isotopic evidence for human and animal diets from the late neolithic to the ming dynasty in the middle-lower reaches of the Hulu River Valley, NW China. Front. Ecol. Evol. 10, 905371. 10.3389/fevo.2022.905371 Dong Y Y. Bi X. Wu R. Belfield E. J. Harberd N. P. Christensen B. T. (2022). The potential of stable carbon and nitrogen isotope analysis of foxtail and broomcorn millets for investigating ancient farming systems. Front. Plant Sci. 13, 1. 10.3389/fpls.2022.1018312 Farquhar G. D. Ehleringer J. R. Hubick K. T. (1989). Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 40, 503–537. 10.1146/annurev.pp.40.060189.002443 Farquhar G. D. Richards R. A. (1984). Isotopic composition of plant carbon correlates with water-use efficiency of wheat genotypes. Funct. Plant Biol. 11, 539–552. 10.1071/pp9840539 Farquhar G. O’Leary M. Berry J. (1982). On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Funct. Plant Biol. 9, 121. 10.1071/pp9820121 Ferrio J. P. Araus J. L. Buxó R. Voltas J. Bort J. (2005). Water management practices and climate in ancient agriculture: Inferences from the stable isotope composition of archaeobotanical remains. Veg. Hist. Archaeobot. 14, 510–517. 10.1007/s00334-005-0062-2 Fiorentino G. Caldara M. De Santis V. D’Oronzo C. Muntoni I. M. Simone O. (2013). Climate changes and human–environment interactions in the Apulia region of southeastern Italy during the Neolithic period. Holocene 23, 1297–1316. 10.1177/0959683613486942 Flohr P. Fleitmann D. Matthews R. Matthews W. Black S. (2016). Evidence of resilience to past climate change in Southwest Asia: Early farming communities and the 9.2 and 8.2 ka events. Quat. Sci. Rev. 136, 23–39. 10.1016/j.quascirev.2015.06.022 Francey R. J. Allison C. E. Etheridge D. M. Trudinger C. M. Enting I. G. Leuenberger M. (1999). A 1000-year high precision record of delta13C in atmospheric CO2. Tellus B 51, 170–193. 10.1034/j.1600-0889.1999.t01-1-00005.x Fuller D. Q. Denham T. Arroyo-Kalin M. Lucas L. Stevens C. J. Qin L. (2014). Convergent evolution and parallelism in plant domestication revealed by an expanding archaeological record. Proc. Natl. Acad. Sci. U. S. A. 111, 6147–6152. 10.1073/pnas.1308937110 Goldsmith Y. Broecker W. S. Xu H. Polissar P. J. deMenocal P. B. Porat N. (2017). Northward extent of East Asian monsoon covaries with intensity on orbital and millennial timescales. Proc. Natl. Acad. Sci. U. S. A. 114, 1817–1821. 10.1073/pnas.1616708114 Han H. Hou J. Huang M. Li Z. Xu K. Zhang D. (2020). Impact of soil and water conservation measures and precipitation on streamflow in the middle and lower reaches of the Hulu River Basin, China. Catena 195, 104792. 10.1016/j.catena.2020.104792 Hattersley P. W. Watson L. (1992). “Diversification of photosynthesis,” in Grass evolution and domestication. Editor Chapman G. P. (Cambridge: Cambridge University Press), 38–116. He K. Lu H. Jin G. Wang C. Zhang H. Zhang J. (2022). Antipodal pattern of millet and rice demography in response to 4.2 ka climate event in China. Quat. Sci. Rev. 295, 107786. 10.1016/j.quascirev.2022.107786 Hu Q. (1991). Answer to “A brief analysis of the lower changshan remains in zhenyuan Longdong” (in Chinese). Archaeology 03, 238–244. Hu Q. (1981). Brief report on excavation of changshan site in zhenyuan county, Longdong (in Chinese). Archaeology 03, 201–210. Indermühle A. Stocker T. F. Joos F. Fischer H. Smith H. J. Wahlen M. (1999). Holocene carbon-cycle dynamics based on CO2 trapped in ice at Taylor Dome, Antarctica. Nature 398, 121–126. 10.1038/18158 Ji D. (2007). Environmental archaeological perspective: Dispersal of anatomically modern human and origin of agriculture in northern China-A case study in Gansu and Ningxia. [Lanzhou (China)]: Lanzhou University. Jia X. Dong G. Li H. Brunson K. Chen F. Ma M. (2013). The development of agriculture and its impact on cultural expansion during the late Neolithic in the Western Loess Plateau, China. Holocene 23, 85–92. 10.1177/0959683612450203 Kelly R. L. Surovell T. A. Shuman B. N. Smith G. M. (2012). A continuous climatic impact on Holocene human population in the Rocky Mountains. Proc. Natl. Acad. Sci. U. S. A. 110, 443–447. 10.1073/pnas.1201341110 Lecavalier B. Fisher D. A. Milne G. A. Vinther B. M. Tarasov L. Huybrechts P. (2019). Data from: Oxygen isotope ratio and reconstructed temperature record from the Agassiz ice cap, Greenland. PANGAEA. (2019). 10.1594/PANGAEA.904113 Leuenberger M. Siegenthaler U. Langway C. (1992). Carbon isotope composition of atmospheric CO2 during the last ice age from an Antarctic ice core. Nature 357, 488–490. 10.1038/357488a0 Li F. Li S. Shui T. (1993). Ancient culture and ancient environment of Hulu River basin (in Chinese). Archaeology 9, 822–842. Li H. (2018). Ancient strategies of crop use on the western Chinese Loess Plateau from Neolithic to historical periods. [Lanzhou (China)]: Lanzhou University. Li H. Cui Y. James N. Ritchey M. Liu F. Zhang J. (2022a). Spatiotemporal variation of agricultural patterns in different geomorphologic and climatic environments in the eastern Loess Plateau, north-central China during the late Neolithic and Bronze Ages. Sci. China Earth Sci. 65, 934–948. 10.1007/s11430-021-9879-x Li H. Sun Y. Yang Y. Cui Y. Ren L. Li H. (2022b). Water and soil management strategies and the introduction of wheat and barley to northern China: An isotopic analysis of cultivation on the Loess Plateau. Antiquity 96, 1478–1494. 10.15184/aqy.2022.138 Li R. Lv F. Yang L. Liu F. Liu R. Dong G. (2020). Spatial–Temporal variation of cropping patterns in relation to climate change in neolithic China. Atmosphere 11, 677. 10.3390/atmos11070677 Li X. Q. Zhou X. Y. Zhou J. Dodson J. Zhang H. B. Shang X. (2007). The earliest agricultural diversification of China recorded by biological indicators at Xishanping site, Gansu (in Chinese). Sci. China Earth Sci. 37, 934–940. Li Y. Fan C. Xing Y. Jiang Y. Luo L. Sun L. (2011). Natural variation in GS5 plays an important role in regulating grain size and yield in rice. Nat. Genet. 43, 1266–1269. 10.1038/ng.977 Liu B. Lu Y. Yang Y. Wei W. Chen G. (2022). Evaluating water fertilizer coupling on the variations in millet chaff size during the late seventh century in northwest China: Morphological and carbon and nitrogen isotopic evidence from the chashancun cemetery. Sustainability 14, 3581. 10.3390/su14063581 Liu F. Feng Z. (2012). A dramatic climatic transition at ∼4000 cal. yr BP and its cultural responses in Chinese cultural domains. Holocene 22, 1181–1197. 10.1177/0959683612441839 Liu L. Chen X. Wright H. Xu H. Li Y. Chen G. (2019). Rise and fall of complex societies in the Yiluo region, North China: The spatial and temporal changes. Quat. Int. 521, 4–15. 10.1016/j.quaint.2019.05.025 Liu X. Jones P. J. Motuzaite Matuzeviciute G. Hunt H. V. Lister D. L. An T. (2019). From ecological opportunism to multi-cropping: Mapping food globalisation in prehistory. Quat. Sci. Rev. 206, 21–28. 10.1016/j.quascirev.2018.12.017 Manning S. W. Lorentzen B. Welton L. Batiuk S. Harrison T. P. (2020). Beyond megadrought and collapse in the Northern Levant: The chronology of Tell Tayinat and two historical inflection episodes, around 4.2ka BP, and following 3.2ka BP. PLoS One 15, e0240799. 10.1371/journal.pone.0240799 Masi A. Sadori L. Balossi Restelli F. Baneschi I. Zanchetta G. (2014). Stable carbon isotope analysis as a crop management indicator at arslantepe (malatya, Turkey) during the late chalcolithic and early bronze age. Veg. Hist. Archaeobot. 23, 751–760. 10.1007/s00334-013-0421-3 Motuzaite-Matuzeviciute G. Hunt H. V. Jones M. K. (2012). Experimental approaches to understanding variation in grain size in Panicum miliaceum (broomcorn millet) and its relevance for interpreting archaeobotanical assemblages. Veg. Hist. Archaeobot. 21, 69–77. 10.1007/s00334-011-0322-2 O’Leary M. H. (1981). Carbon isotope fractionation in plants. Phytochemistry 20, 553–567. 10.1016/0031-9422(81)85134-5 O’Leary M. H. (1988). Carbon isotopes in photosynthesis. BioScience 38, 328–336. 10.2307/1310735 Park J. Park J. Yi S. Cheul Kim J. Lee E. Choi J. (2019). Abrupt Holocene climate shifts in coastal East Asia, including the 8.2 ka, 4.2 ka, and 2.8 ka BP events, and societal responses on the Korean peninsula. Sci. Rep. 9, 10806. 10.1038/s41598-019-47264-8 Pokharia A. K. Agnihotri R. Sharma S. Bajpai S. Nath J. Kumaran R. N. (2017). Altered cropping pattern and cultural continuation with declined prosperity following abrupt and extreme arid event at ∼4,200 yrs BP: Evidence from an Indus archaeological site Khirsara, Gujarat, Western India. PLoS One 12, e0185684. 10.1371/journal.pone.0185684 Qian Y. Wang Y. Mao R. Xie Y. (2014). Excavation report of Qijia culture tomb in mogou cemetery excavation brief of Qijia culture tombs in mogou cemetery, lintan, Gansu province in 2009 (in Chinese). Wenwu Cult. Relics 697, 4–23. Ramsey B. C. (2021). OxCal. version. Available at: https://c14.arch.ox.ac.uk/oxcal.html (Accessed, 2021). Ran M. Chen L. (2019). The 4.2 ka BP climatic event and its cultural responses. Quat. Int. 521, 158–167. 10.1016/j.quaint.2019.05.030 Reimer P. J. Austin W. E. N. Bard E. Bayliss A. Blackwell P. G. Bronk Ramsey C. (2020). The IntCal20 northern hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon 62, 725–757. 10.1017/rdc.2020.41 Ren L. Yang Y. Qiu M. Brunson K. Chen G. Dong G. (2022). Direct dating of the earliest domesticated cattle and caprines in northwestern China reveals the history of pastoralism in the Gansu-Qinghai region. J. Archaeol. Sci. 144, 105627. 10.1016/j.jas.2022.105627 Ren L. Yang Y. Wang Q. Zhang S. Chen T. Cui Y. (2021). The transformation of cropping patterns from late neolithic to early iron age (5900–2100 BP) in the gansu–qinghai region of northwest China. Holocene 31, 183–193. 10.1177/0959683620941137 Sanborn L. H. Reid R. E. Bradley A. S. Liu X. (2021). The effect of water availability on the carbon and nitrogen isotope composition of a C4 plant (pearl millet, Pennisetum glaucum). J. Archaeol. Sci. Rep. 38, 103047. 10.1016/j.jasrep.2021.103047 Sheng P. Shang X. Sun Z. Yang L. Guo X. Jones M. K. (2018). North-south patterning of millet agriculture on the Loess Plateau: Late Neolithic adaptations to water stress, NW China. Holocene 28, 1554–1563. 10.1177/0959683618782610 Song J. Zhao Z. Fuller D. Q. (2012). The archaeobotanical significance of immature millet grains: An experimental case study of Chinese millet crop processing. Veg. Hist. Archaeobot. 22, 141–152. 10.1007/s00334-012-0366-y Staubwasser M. Sirocko F. Grootes P. M. Segl M. (2003). Climate change at the 4.2 ka BP termination of the Indus valley civilization and Holocene south Asian monsoon variability. Geophys. Res. Lett. 30, 1. 10.1029/2002gl016822 Styring A. K. Charles M. Fantone F. Hald M. M. McMahon A. Meadow R. H. (2017). Isotope evidence for agricultural extensification reveals how the world's first cities were fed. Nat. Plants 3, 17076–17111. 10.1038/nplants.2017.76 Tan L. Li Y. Wang X. Cai Y. Lin F. Cheng H. (2020). Holocene monsoon change and abrupt events on the western Chinese Loess Plateau as revealed by accurately dated stalagmites. Geophys. Res. Lett. 47. 10.1029/2020gl090273 Tao D. Zhang R. Xu J. Wu Q. Wei Q. Gu W. (2022). Agricultural extensification or intensification: Nitrogen isotopic investigation into late Yangshao agricultural strategies in the middle Yellow River area. J. Archaeol. Sci. Rep. 44, 103534. 10.1016/j.jasrep.2022.103534 The Institute of Archaeoloogy Chinese Academy of Social Sciences (1999). Shizhaocun and xishanping (in Chinese). Beijing: Encyclopedia of China Publishing House. Underhill A. (2013). A companion to Chinese archaeology. New Jersey: Wiley-Blackwell. Wallace M. Jones G. Charles M. Fraser R. Halstead P. Heaton T. H. E. (2013). Stable carbon isotope analysis as a direct means of inferring crop water status and water management practices. World Archaeol. 45, 388–409. 10.1080/00438243.2013.821671 Wang F. He J. Han F. Zhao Q. Zhang H. (2022). Quantitative allocation of soil and water conservation measures in Xihulu River basin of Liupan Mountain in loess hilly and gully region (in Chinese). Gansu Agric. 4, 58–63. 10.15979/j.cnki.cn62-1104/f.2022.04.019 Wang G. Han J. Zhou L. Xiong X. Tan M. Wu Z. (2005). Study on the stable carbon isotopic composition of C4 plants in loess region of northern China (in Chinese). Sci. China Earth Sci. 12, 1174–1179. Wang H. (2012). Sequence and pattern of archaeology culture in Neolithic-Bronze Age in Gansu-Qinghai province (in Chinese). Collect. Stud. Archaeol. 9, 210–243. Wang L. Mei J. Chen K. Mao R. Qian Y. Wang H. (2022). Budget-constraint admissible output consensus tracking for intermittent-interaction singular multiagent networks. Archaeology 658, 71–80. 10.1016/j.isatra.2021.10.021 Wang X. Fuller B. T. Zhang P. Hu S. Hu Y. Shang X. (2018). Millet manuring as a driving force for the Late Neolithic agricultural expansion of north China. Sci. Rep. 8, 5552. 10.1038/s41598-018-23315-4 Wei J. (2021). Research on the tombs of changshan lower culture. [Lanzhou (China)]: Northwest Normal University. Weiss H. (2016). Global megadrought, societal collapse and resilience at 4.2-3.9 ka BP across the Mediterranean and west Asia. Past. Glob. Change Mag. 24, 62–63. 10.22498/pages.24.2.62 Weiss H. (2017). Megadrought and collapse: From early agriculture to Angkor. Oxford: University Press. Willcox G. (2004). Measuring grain size and identifying near eastern cereal domestication: Evidence from the euphrates valley. J. Archaeol. Sci. 31, 145–150. 10.1016/j.jas.2003.07.003 Womack A. Flad R. Zhou J. Brunson K. Toro F. H. Su X. (2021). The majiayao to Qijia transition: Exploring the intersection of technological and social continuity and change. Asian Archaeol. 4, 95–120. 10.1007/s41826-021-00041-x Womack A. Jaffe Y. Zhou J. Hung L. Wang H. Li S. (2017). Mapping qijiaping: New work on the type-site of the Qijia culture (2300–1500 B.C.) in Gansu province, China. J. Field Archaeol. 42, 488–502. 10.1080/00934690.2017.1384669 Wu W. Liu T. (2004). Possible role of the “Holocene event 3” on the collapse of neolithic cultures around the central plain of China. Quat. Int. 117, 153–166. 10.1016/s1040-6182(03)00125-3 Wu W. Zheng H. Hou M. Ge Q. (2018). The 5.5 cal ka BP climate event, population growth, circumscription and the emergence of the earliest complex societies in China. Sci. China Earth Sci. 61, 134–148. 10.1007/s11430-017-9157-1 Xiao J. Zhang S. Fan J. Wen R. Xu Q. Inouchi Y. (2019). The 4.2 ka event and its resulting cultural interruption in the Daihai Lake basin at the East Asian summer monsoon margin. Quat. Int. 527, 87–93. 10.1016/j.quaint.2018.06.025 Xie D. (1986). A brief discussion on the tombs of Qijia culture (in Chinese). Archaeology 2, 147–161. Xie D. (1975). Qijia cultural cemetery of qinweijia in yongjing county, Gansu (in Chinese). Acta Archaeol. Sin. 2, 57–96. Xin H. Dai C. Cao W. Yang B. , and Yellow River water conservancy commission tianshui hydrological and water resources survey bureau. (2016). analysis of hydrological characteristics in Hulu River basin (in Chinese). Water Resour. Hydropower Northeast 34, 32. 10.14124/j.cnki.dbslsd22-1097.2016.05.014 Yang H. (2017). Study on the tombs of Qijia culture in Hexi area. [Lanzhou (China)]: Northwest Minzu University. Yang J. (2012). The studies of prehistoric culture in Longdong area. [Zhengzhou (China)]: Zhengzhou University. Yang J. Zhang D. Yang X. Wang W. Perry L. Fuller D. Q. (2022). Sustainable intensification of millet–pig agriculture in Neolithic North China. Nat. Sustain. 5, 780–786. 10.1038/s41893-022-00905-9 Yang L. Shi Z. Zhang S. Lee H. F. (2020). Climate change, geopolitics, and human settlements in the hexi corridor over the last 5,000 years. ACTA Geol. SIN-ENGL 94, 612–623. 10.1111/1755-6724.14529 Yang L. (2021). Strategies of crop use on the southern Chinese Loess Plateau from late neolithic to bronze age. [Lanzhou (China)]: Lanzhou University. Yang Q. Li X. Liu W. Zhou X. Zhao K. Sun N. (2011). Carbon isotope fractionation during low temperature carbonization of foxtail and common millets. Org. Geochem. 42, 713–719. 10.1016/j.orggeochem.2011.06.012 Yang Y. Ren L. Dong G. Cui Y. Liu R. Chen G. (2019). Economic change in the prehistoric Hexi corridor (4800–2200 bp), north‐west China. Archaeometry 61, 957–976. 10.1111/arcm.12464 Yang Y. (2014). The analysis of charred plant seeds at Jinchankou site and lijiaping site during Qijia Culture Period in the hehuang region. Lanzhou, China: Lanzhou University. Yang Y. Wang J. Li G. Dong J. Cao H. Ma M. (2022). Shift in subsistence crop dominance from broomcorn millet to foxtail millet around 5500 BP in the Western Loess Plateau. Front. Plant Sci. 13, 939340. 10.3389/fpls.2022.939340 Ye S. (2014). A comprehensive study of pottery in Qijia culture (in Chinese). Young Society (01), 342. Zhang C. Zhao C. Yu S.-Y. Yang X. Cheng J. Zhang X. (2022). Seasonal imprint of Holocene temperature reconstruction on the Tibetan Plateau. Earth-Sci. Rev. 226, 103927. 10.1016/j.earscirev.2022.103927 Zhang C. Zhao C. Zhou A. Zhang H. Liu W. Feng X. (2021). Quantification of temperature and precipitation changes in northern China during the “5000-year” Chinese History. Quat. Sci. Rev. 255, 106819. 10.1016/j.quascirev.2021.106819 Zhang H. Cheng H. Sinha A. Spötl C. Cai Y. Liu B. (2021). Collapse of the Liangzhu and other Neolithic cultures in the lower Yangtze region in response to climate change. Sci. Adv. 7, eabi9275. 10.1126/sciadv.abi9275 Zhang J. Chen Z. Lan W. Yang Y. Luo W. Yao L. (2018). The new progresses of the paleoethnobotanical studies of the jiahu site in wuyang, henan (in Chinese with English abstract). Archaeology 4, 100–110. Zhang S. (2022). Times new roman arrangement and research of jade unearthed from Qijia culture. [Dalian (China)]: Liaoning Normal University. Zhao C. (2020). The climate fluctuation of the 8.2 ka BP cooling event and the transition into neolithic lifeways in North China. Quaternary 3, 23. 10.3390/quat3030023 Zhao Y. Chen F. Zhou A. Yu Z. Zhang K. (2010). Vegetation history, climate change and human activities over the last 6200 years on the Liupan Mountains in the southwestern Loess Plateau in central China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 293, 197–205. 10.1016/j.palaeo.2010.05.020 Zhao Z. (2010). Paleoethnobotany: Theories, methods and practice (in Chinese). Beijing: Science Press. Zhou X. Li X. John D. Zhao K. (2016). Rapid agricultural transformation in the prehistoric Hexi corridor, China. Quat. Int. 426, 33–41. 10.1016/j.quaint.2016.04.021