Introduction
When and how prehistoric humans colonized the Tibetan Plateau has been intensively discussed during the last ten years (
Brantingham and Gao, 2006;
Zhao et al., 2009;
Aldenderfer, 2011;
Brantingham et al., 2013;
Dong et al., 2013; Rhode et al., 2014;
Chen et al., 2015;
d’Alpoim Guedes et al., 2015). Recent research indicated that prehistoric humans had extensively and permanently settled on the high areas of the Tibetan Plateau since 3600 BP, which was primarily promoted by agricultural development and diffusion across Eurasia during late prehistoric times (
Chen et al., 2015). However, human settlement and its influencing factors during the historic period in the high Plateau remain enigmatic. The absence of reliable dates from historic sites and documents analysis, and their comparison with high resolution paleoclimatic records, obstruct our understanding of the human-environment relationship in the Tibetan Plateau during the past two millennia.
Vast amounts of information have been recorded about historical cultures in ancient cities (such as politics, the military, religion, economics, etc.), and are valuable for studying the human-land relationship during late Holocene, especially in arid areas (
Gao et al., 2013). Numerous ancient cities were discovered through archaeological surveys in east Qinghai Province, China, located on the northeastern border of the Tibetan Plateau (
Bureau of National Cultural Relics, 1996). However, the ages of these cities were preliminarily estimated through the characteristics of the structures and archaeological remains (such as ceramic chips, bricks, etc.), and the fragmentary historical documents relating to this area. Radiocarbon dating has been found to be a more reliable method for determining the absolute ages of ancient cities (
Lu et al., 2010) than using characteristics of artifacts found in the Tibetan Plateau (
Sun et al., 2010;
Hudson et al., 2014). In this study, we investigate ancient cities in east Qinghai Province by collecting reliable dating materials, using the accelerator mass spectrometry (AMS) radiocarbon dating method to determine the ages of these cities, and thus providing a valuable dataset for the reconstruction of human settlement history during the historical period. We also compare the radiocarbon dates with the historical documents and high resolution paleoclimate records to further explore the potential causes behind the emergence of these ancient cities.
Study area
The study area (35°20′‒37°32′N, 96°23′‒103°04′E) is located in the northeastern Tibetan Plateau, northwest China. The investigated ancient cities are distributed in the eastern Qaidam Basin, the Qinghai Lake Basin, the Huangshui River Valley, the upper Yellow River Valley, and the Datong River Valley (Fig. 1), stretching about 500 km from Dulan County in the west to Minhe County in the east. The mean annual temperature ranges from-0.6°C ‒ 8.1°C, and the mean annual precipitation ranges from 37.9 ‒ 400 mm. The altitude gradually declines from west to east: the highest elevation is 5,536 m a.s.l. in Dulan County and the lowest is 1,650 m a.s.l. in Minhe County.
Materials and methods
We identified 47 ancient cities by remotely sensed imagery (Fig. 2), conducting an archaeological survey in east Qinghai Province (Fig. 1), and collecting 54 radiocarbon dating samples from these ancient cities, including two samples from 7 cities and one sample from the remaining 40. The dated samples included charred charcoal, grass, and sheep droppings from the cultural layers within the cities (Fig. 3(a), 3(b)), and dead wood or tree bark found in the walls around the cities, also used for their construction (Fig. 3(c)‒3(f)). In order to minimize the “old wood effect” of radiocarbon dating of charcoal (
McFadgen, 1982;
Schiffer, 1986;
Gavin, 2001;
Dong et al., 2014), we selected the outermost parts of the wood for dating.
All 54 samples were dated using AMS at Peking University, China. Ages were calibrated using Calib (v.7.0.2) (Stuiver and Reimer, 1993) and the INTCAL13 Calibration curve (
Reimer et al., 2013). All ages were reported as “Cal AD” (“Calendar years after AD 1”) or “Cal BC” (“Calendar years before AD 1”).
The probability distribution of the 54 radiocarbon dates was summed using Calib (v.7.0.2) (
Stuiver and Reimer, 1993). The summed probability plots of radiocarbon data have been widely used in archaeological research to infer prehistoric population fluctuations (
Gamble et al., 2005; Shennan and Edinborough, 2007; Bamforth and Grund, 2012) and occupational intensity (
Barton et al., 2007); however, caution must be used by considering the distorting effects of site sampling, sample size, radiocarbon calibration, and taphonomic bias (Williams, 2012). In this paper, the summed probability plot of radiocarbon dates is used as an auxiliary proxy to deduce when those ancient cities were intensively built or repaired.
Results
The radiocarbon dating results are shown in (Table 1). The oldest calibrated date is from Tawa in Guinan County and ranges between 474 and 390 Cal BC, corresponding to the Spring and Autumn Period (770 BC‒476 BC). Four calibrated dates from Hongxin in Dulan County, Douhouzong in Tongde County, Xiatan in Xinghai County, and Duobajiu in Huangzhong County, range between 54 Cal BC and 214 Cal AD, correspond to the “Han Dynasty” (202 BC‒AD 220). One calibrated date from Langshetou in Qilian County ranges between 430 and 584 Cal AD, corresponding to the “Wei Jin Southern and Northern Dynasties” (AD 220‒AD 581).
Ten calibrated dates range between 618 and 886 Cal AD, corresponding to the Tang Dynasty (AD 618‒AD 907). Eighteen calibrated dates range between 960 and 1218 Cal AD, corresponding to the Song Dynasty (AD 960‒AD 1279). Ten calibrated dates range between 882 and 1158 Cal AD, corresponding to the transitional period that partly overlaps the late Tang Dynasty, the Five Dynasties and Ten Kingdoms period (AD 907‒AD 960), and the Song Dynasty.
Two calibrated dates from Tadong and Longqu in Xinghai County range between 1298 and 1410 Cal AD, corresponding to the transitional period between the Yuan Dynasty (AD 1271‒AD 1368) and Ming Dynasty (AD 1368‒AD 1644). Four calibrated dates, from Shancheng in Minhe County, Fuxi in Gonghe County, Longqu in Xinghai County, and Hacheng in Huangyuan County, range between 1436 and 1635 Cal AD, corresponding to the Ming Dynasty. Three calibrated dates, from Heigu in Guide County and Shangtamai and Xiazhatan in Gonghe County range between 1522 and 1654 Cal AD, corresponding to the transitional period between the Ming Dynasty and Qing Dynasty (AD 1644‒AD 1911). One calibrated date from Duobaxin in Huangzhong County ranges between 1697 and 1922 Cal AD, which basically belongs to the Qing Dynasty.
Discussion
When human built or repaired ancient cities in northeastern Tibetan Plateau?
Radiocarbon dates from ancient cities can be used to determine when they were built, repaired, or occupied. Wood charcoals sampled from the rammed earth layer inside walls of ancient cities were the standard dating materials used in this study (Fig. 3, Table 1). The dating results may represent initial construction or maintenance of these cities.
The ancient city of Tawa in Guinan County dates back to the Spring and Autumn Period, a turbulent period in Chinese history. However, that date may be problematic because at that time, east Qinghai Province was occupied by aboriginal nomadic tribes, collectively called “
Qiang” (
Cui et al., 1999), and no ancient city built by these tribes has ever been reported or recorded. Furthermore, the date of Tawa is much older than the dates of the other four ancient cities in Guinan County, all of which date to the Song Dynasty (Table 1). In addition, a tile decorated with a lotus flower, prevalent during the Tang dynasty, was found in one of the eaves of a structure in Tawa and sampled (Supplementary file). This evidence suggests that the radiocarbon date may be older than the period of occupation of Tawa. This could possibly be due to: 1) the “old wood” effect of radiocarbon dating of charcoal, which can range from decades to centuries in northwest China (
Dong et al., 2014), or 2) the reuse of old materials (such as charcoal) from early human occupation for the purpose of building a wall. Artifacts from the Kayue culture (1600‒600 BC) were also found in the Tawa site during an archaeological survey (
Bureau of National Cultural Relics, 1996).
Four ancient cities were dated to the Han Dynasty (Table 1), when the Chinese central government firstly affected and controlled east Qinghai Province. According to historic documents (Hou Hanshu, Xiqiang Zhuan and Hanshu, Zhaodi Ji), Emperor Hanwu (156 BC‒87 BC), detached general Li Xi, and a 100,000-strong army suppressed the Qiang in the Huangshui River Valley and recaptured the area west to Qinghai Lake and Chaka Salt Lake in 111 BC. Subsequently, the government of the western Han Dynasty (202 BC‒AD 9) began to establish prefectures and counties to control the area. According to historic texts (Hanshu, Wangmang Zhuan), Wangmang, who controlled Han Dynasty during AD 8‒AD 23, built the ancient cities of Xihaijun in the Qinghai Lake Basin and Zhidonglajia in Xinghai County. The Eastern Han Dynasty government (AD 9‒AD 220) set up Xiping Jun (Jun is an administrative district in Chinese history), the predecessor of modern Xining. Since Douhouzong, Xiatan and Duobajiu are located in the areas (Xinghai, Huangzhong and Tongde counties) controlled by the Han Dynasty, their radiocarbon dates fit well with this historical evidence. The date of Hongxin in Dulan County is more open to debate, as this area was controlled by the Tuyuhun kingdom that was established in the area around AD 330, and collapsed during the Tang Dynasty. We were unable to find any records for the existence of ancient cities in Dulan County during the Han Dynasty in historic documents.
Only Langshetou in Qilian County was dated to the “Wei Jin Southern and Northern Dynasties” period, another turbulent phase of Chinese history. According to the “Jin Shu” text, this period covered the years between AD 397 and AD 414. East Qinghai Province was essentially controlled by both the Nan Liang kingdom, located in the Huangshui River Valley, and the Tuyuhun kingdom, located in the Qaidam Basin. Qilian County is located in the Datong River Valley, an important channel connecting the Huangshui River Valley and the Hexi Corridor. This area was controlled by the Hou Liang kingdom between AD 386 and AD 403. Frequent wars occurred between the Nan Liang and Hou Liang kingdoms. Two dates were obtained from Langshetou, one of which dated to AD 430‒AD 560 (BA120346) and the other at AD 685‒AD 780 (BA120345). These dates suggest that Langshetou may have been built during the Wei Jin Southern and Northern Dynasties and repaired during Tang Dynasty.
The number of ancient cities in east Qinghai Province noticeably increased during the Tang Dynasty when compared to earlier periods. The summed probability distribution of radiocarbon dates is much higher during this period than during former periods (Fig. 4). This suggests that in this area, ancient cities might have been built or repaired in large numbers during the Tang Dynasty. According to the dating results (Table 1), the ten ancient cities of Tang Dynasty are primarily located in the Huangshui and Datong River Valleys, and constructed between 600 Cal AD and 800 Cal AD (Table 1). These results are in agreement with the records found in historical documents. The
Tubo kingdom prospered during the middle of the sixth century, and unified the entire Tibetan area by AD 644, before expanding toward east Qinghai Province. The
Tubo kingdom defeated the
Tuyuhun kingdom in AD 663, which seriously threatened the border of the Tang Dynasty lands. To defend against the expansion of the
Tubo kingdom in east Qinghai Province, the Tang central government established a military defense system in the Huangshui River Valley, and set up the
Dudufu (a high-class military command authority) in
Shan Zhou (Xining area). In AD 677, the number of garrison troops in the area reached 14,000 (
Cui et al., 1999). In addition, the population in Qinghai Province increased rapidly from 219,000 in AD 609 to 421,000 in AD 752 (
Zhao and Xie, 1998) (Fig. 4), according to an analysis of historical records. The military confrontation between the Tang and
Tubo governments might have prompted the construction of cities in east Qinghai Province, with Kaoxiaotu in Dulan County likely controlled by the
Tubo government, while the other nine ancient cities east to the Qaidam Basin were likely governed by the Tang government.
According to the dating results, the number of ancient cities reached a maximum during the transitional period between the late Tang Dynasty, the Five Dynasties and Ten Kingdoms period, and the early Song Dynasty, especially between 1000‒1100 Cal AD, the period with the highest summed probability distributions of radiocarbon dates (Fig. 4(g)). According to historical records, east Qinghai Province was controlled by the
Qingtang kingdom during that period, which allied with the Song government to fight the
Xixia kingdom (AD 1038‒AD 1227), a strong regime built by
Li Yuanhao in AD 1038. Many major battles were fought between the troops of the
Xixia and
Qingtang kingdoms in the eleventh century in east Qinghai Province. For example,
Su Nuer, a
Xixia general, led 25,000 troops against the
Qingtang kingdom in AD l035, ending in defeat in the ancient city of Maoniu in Datong County, according to
Songshi, Xiaguo Zhuan. The
Xixia kingdom continually sent troops to the Huangshui and Upper Yellow River Valleys between AD 1035‒AD 1062, but failed to take the area due to the strong resistance of the
Qingtang kingdom. Radiocarbon dates are highly concentrated during that period, suggesting that the strong need for military defense contributed to the building and/or repair of ancient cities in the area. The results of radiocarbon dating may accurately reflect the true ages of the cities of that period. The population in Qinghai Province declined between the late Tang Dynasty and early Song Dynasty (
Zhao and Xie, 1998) (Fig. 4), likely induced by frequent wars during the early Song Dynasty in the area. The climate was warm and humid during that time (Fig. 4), but was not likely the cause for the decrease in population.
The summed probability distribution of radiocarbon dates declined post Song Dynasty period. Only ten ancient cities were dated to that period, most of which fell into the Ming Dynasty period (Fig. 4, Table 1). Two dates from Longqu and Tadong in Xinghai County correspond well with the Yuan Dynasty. According to the records in
Yuan Shi, the primary administrative center of the Yuan government on the south bank of the Yellow River was located in Guide Prefecture, adjacent to Xinghai County. Seven ancient cities are dated to the Ming Dynasty period. The Northern
Yuan (AD 1368‒AD 1402) was the primary threat to the central government of early Ming Dynasty. Abundant remains of the Great Wall dated to the Ming Dynasty date were found by the archaeological survey (
Bureau of National Cultural Relics, 1996), probably due to the defense against the invasion by Mongolian tribes, such as the Northern
Yuan and the later
Datan. The Ming government also set up many military districts in east Qinghai Province, covering the areas of the Qinghai Lake Basin, the Gonghe Basin, and the Huangshui River Valley. All six cities that were dated to the Ming Dynasty are located in these regions. The population in Qinghai Province was relatively high during the Ming Dynasty (
Zhao and Xie, 1998) (Fig. 4), which may suggest that these ancient cities might also have promoted human settlement during that period.
Only the Duobaxin ancient city in Huangzhong county was dated to the Qing Dynasty. However, according to the results of the second national archaeological survey (
Bureau of National Cultural Relics, 1996), the number of sites dating to the Qing Dynasty is higher than for any other dynasty, and the population of Qinghai Province was also high during this period (
Zhao and Xie, 1998). This suggests that the rarity of ancient cities dated to the Qing period, may not indicate that there were only a small number of ancient cities in this period. Rather, the absence of radiocarbon dates from this period may relate to the destruction of these ancient cities during recent decades, as modern cities very likely developed in the places that the Qing ancient cities are located.
Comparing results from those 7 ancient cities from which two dates were obtained, the difference between the two dates at most of these sites is less than two hundred years (Table 1), while the difference between the two radiocarbon dates from Douhouzong in Tongde county is around one thousand years. Although this site is probably influenced by the “old wood” effect of radiocarbon dating, multiple or sustained occupation of the same ancient cities could also affect the dating results. For example, the Fuxi ancient city had once been used as the capital of the Tuyuhun kingdom during the “Wei Jin Southern and Northern Dynasties” period, but was dated to Ming Dynasty in this work. This indicates that the ancient city had been occupied during different historical periods.
The influencing factors behind the construction of ancient cities in northeastern Tibetan Plateau.
Construction of ancient cities in northeastern Tibetan Plateau is probably strongly linked to the geopolitical situations, for the area is an important boundary area between nomadic tribes and agricultural Han people during historical periods. While aggression wars between agriculturalist and pastoralist empires are closely related to climate change, especially the evident decline of temperature and precipitation, which resulted in the ecological deterioration and social instability (
Zhang et al., 2010;
Zhang et al., 2015). Based on this premise, the dating results of ancient cities in this study is compared with historical documentary records and high-resolution paleo-climate reconstructions, to examine how geopolitical variables and climate change affected the construction of ancient cities in northeastern Tibetan Plateau.
The comparison between proxies from tree rings and ice core paleoclimate records in northeastern Tibetan Plateau (
Yang et al., 2003;
Zhang et al., 2003;
Liu et al., 2009;
Zhang et al., 2011), war number in the area and the summed probability plot of 54 radiocarbon dates in this text is shown in Fig. 4. Most ancient cites in the area were built during Tang, early Song and Ming Dyansties according to the dating results, when war frequency was relatively high due to the conflicts and confrontations of different powerful politics, including Tang and
Tubo empires, Northern Song,
Xixia and
Qingtang regimes, Ming and Northern Yuan regimes. War frequency was the highest in northeastern Tibetan Plateau during late Wei Jin Southern and Northern Dynasties (Fig. 4(e)), however, only one ancient city is dated to that period. According to the records of historical documents, more than 16 ancient cities in northeastern Tibetan Plateau were firstly built in Wei Jin Southern and Northern Dynasties (such as Shenna, Qunke, Shangtamai, Beigu), while well-preserved ancient cities were concentratedly dated between Tang and Song Dynasties, due to the repair and reinforcement of old cities during late periods (
Zhao, 1986;
Li, 1995). The control power of east Qinghai Province was struggled by different local regimes in Wei Jin Southern and Northern Dynasties, resulting in the frequent wars during the period. The positive correlation between the ages of ancient cities and the frequency of wars (Fig. 4(e) and 4(g)) suggests military confrontation of different geopolitics possibly led to the active demand of defense, and the construction and repair of ancient cities in northeastern Tibetan Plateau during historic times.
The multi-decadal severe decrease in temperature and precipitation could be the cause for the breakdown of food production, locust plagues, and uncontrollable crop prices, followed by social conflicts and frequent wars during the last two millennia (
Zhang et al., 2010;
Kennett et al., 2012;
Zhang et al., 2015). Conversely, the multi-centennial warm and humid climate potentially promoted the thriving of powerful dynasties (such as Han, Tang, and Ming) in Chinese history (
Yancheva et al., 2007;
Zhang et al., 2008). The periods of temperature decline in northeastern Tibetan Plateau correspond well with the collapse of the Han, Tang, Song, and Ming dynasties, further suggesting that climate change was an influencing factor for alternation of Chinese dynasties. However, neither the variabilities of temperature nor precipitation were related to the summed probability of radiocarbon dates from the 47 ancient cities in northeastern Tibetan Plateau, indicating that the potential impact of climate change on the building of those cities is complex.
The probability distribution of the 54 radiocarbon dates is relatively low before the Tang Dynasty and evidently increased during the early Tang Dynasty when temperature and precipitation decreased and war frequency was high. These data suggest that climate deterioration may have led to an increase in military struggle for resources between the Tang and
Tubo empires, and promoted the construction or repair of ancient cites. The probability increases again during the Five Dynasties and Ten Kingdoms period, a splitting and unrest period in central China that possibly related to climate deterioration (
Zhang et al., 2008). War frequency was low during that period in northeastern Tibetan Plateau (Fig. 4(e)), likely due to the collapse of both Tang and
Tubo empires, while frequent wars in nearby agriculturalist regimes may have promoted the reinforcement of cities in the area.
The summed probability of radiocarbon dates reached the maximum, while population declined, during early Song Dynasty when climate was warm and wet (Fig. 4). Favorable climate during the period likely promoted the rise of the Qingtang regime in the northeastern Tibetan Plateau and the Xixia and Northern Song empires in neighboring areas. These conditions resulted in the complex geopolitical structures and military confrontations in the area, followed by city construction and repair. The summed probability of radiocarbon dates was relatively low, and an increase in population was observed during the Ming and Qing Dynasties when the climate was typically warm and wet (Fig. 4) suggesting that the diminished occurrence of wars reduced the construction of ancient cities. Alternatively, favorable climate promoted human settlements in the northeastern Tibetan Plateau during these periods.
Our work suggests that the construction and repair of ancient cities in northeastern Tibetan Plateau are closely related to military conflict and confrontation, which were primarily affected by the emergence of different regimes in and around the area. Climate change may affect city construction and human settlements through its impact on the geopolitical situation. Climate change and geopolitical variables also influenced human settlement intensity in other areas of northwest China during historical periods, such as ancient Juyan Oasis (
Hu and Li, 2014), ancient Minqin oasis (
Xie et al., 2009) and Zhuanglang County (
Dong et al., 2012).
Conclusions
Radiocarbon dates from ancient cities in the northeastern Tibetan Plateau suggest that most ancient cities were built or repaired during the Han Dynasty (202 BC‒AD 220), the Tang Dynasty (AD 618‒AD 907), the Five Dynasties and Ten Kingdoms period (AD 907‒AD 960), the Song dynasty (AD 960‒AD 1279), and the Ming Dynasty (AD 1368‒AD 1644). This study contributed to our understanding of human settlement history in the Tibetan Plateau, given historical documentary records are deficient. Our work suggests that the number of ancient cities in east Qinghai Province during different historical periods relates closely to the military confrontations between strong regimes in the area, especially between Tang Dynasty and early Song Dynasty. The primary factor behind the building of these ancient cities is based on the geopolitical circumstances of east Qinghai Province and its surrounding areas, even though climate change may have also have been a contributing factor.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(3708KB)
