Abstract

Historical records for the Mu Us Desert margin during the Ming dynasty (1368–1644) and corresponding high-resolution climate proxy records have prompted studies on societal responses to climatic changes in this region. The Mu Us Desert margin is highly sensitive to changes in desertification and biological productivity controlled in part by Asian monsoon variations. Here the existing historical temperature and precipitation records are examined to understand spatiotemporal climate variations and to identify potential mechanisms that have driven desertification in the region over the past 500 years. The focus here is on three severe desertification events that occurred in 1529–46, the 1570s, and 1601–50. The relationships among temperature, precipitation, and desertification indicate that a cold/drought-prone climate drives the desertification process. During the Ming dynasty, this region was one of nine important military districts, where the frontier wall (the Great Wall) and other fortifications were constructed. To maintain the defense system, military officers made a valiant effort to decrease the influence of desertification. However, the human-waged war against nature was largely futile, and local rebellions in the stricken region were spawned by the inability of the government to cope with the severe environmental stresses associated with rapid desertification.

1. Introduction

Strong correlations between past changes in climate and societal crises have been recognized by a number of recent studies (deMenocal 2001; Weiss and Bradley 2001; Zhang et al. 2006, 2011a,b). According to Chinese history, the Ming dynasty (1368–1644) was subjected to intensive environmental and economic crises that accompanied the unfavorable climate of the Little Ice Age (LIA). This situation likely led to the collapse of the Ming dynasty (Cook et al. 2010; Dardess 2012; Feng et al. 2013; Wang et al. 2010; Zhang et al. 2008; Zheng et al. 2014). In particular, Zhang et al. (2006, 2011b) have shown an influence of climate change on the outbreak of war and population decline during the preindustrial era at large temporal and spatial scales. Other research indicates that persistent relationships exist between the rise and decline of dynasties and historical cycles of desertification and biological productivity (Wang et al. 2010). During this period, central China experienced frequent periods of increased desertification and decreased biological productivity, which limited the expansion of the Ming domain northward and the dynasty’s control over western China. These studies provide valuable background information on the history of China, which has been significantly impacted by climatic changes at the macroscale. However, details on the societal responses to climate changes are vague, and more studies are needed to examine how humans and their institutions have dealt with the challenges posed by harsh climate conditions. Historical records have great potential to recover a firsthand detailed picture of societal responses to climatic change. This requires interdisciplinary cooperation between historians and climatologists. At the outset it should be understood also that differences between national and regional levels make it unrealistic to regard the whole of China as a unit through time. More case studies will provide the chance to check if there were different strategies employed in different areas to cope with climate change during the same period (Fan 2015).

From this perspective, we present a case study of the Mu Us Desert to demonstrate the influence of climate and desertification during the Ming dynasty. The Ming–nomadic margin was chosen because it comprised an important military area that was well documented. Thus, historical records are available to extract environmental information and to follow societal responses to environmental stress. In addition, this region is at the northern extent of the influence of the Asian summer monsoon, making the ecosystem here especially fragile and susceptible to frequent drought events that led to desertification and agricultural failures, as well as a range of other environmental issues (Yang et al. 2014). Hence, the region is ideal for our current research.

Evidence indicates that severe desertification occurred during the Ming dynasty. However, previous authors have not provided a clear and comprehensive picture of desertification of this region during the Ming dynasty. We attempt to improve on this aspect of the research by compiling all available historical data on desertification and land use, with as much detail as sources permit. With this historical perspective, we reconstruct the regional desertification process at both temporal and spatial scales. In addition, the potential climatic drivers of desertification are analyzed here.

To date, little attention has been paid to human responses or strategic solutions that may have been attempted but failed. An understanding of the responses of human societies in marginal environments to environmental change is critical because, in marginal ecological areas, societies must maintain careful balances with the environment to ensure survival. Societal responses to the degradation of the environment are analyzed from the perspective of military defense. In this study, we have extracted information on human activities, such as the effort to mitigate the deterioration of the natural environment and the forced struggle against nature. Moreover, this paper examines the impact of climate change and desertification in the context of the life of the local people and their economy.

In summary, this study integrates the methodologies of historical and environmental sciences at the regional scale with frequent long-term observations. From this perspective, the results contribute to our understanding of the role of human–environment–climate interactions during the Ming dynasty at the active margin of the desertification region.

2. Data and methods

a. Study area

The Mu Us Desert is located in the southeastern portion of the Ordos Plateau, which is encircled by the Yellow River’s Great Bend (YRGB) in the west, north, and east (Fig. 1). The Mu Us Desert stretches between 37°30′–39°20′N and 107°20′–111°30′E, covering 40 000 km2. The Great Wall of China crosses the southeastern edge of the desert. The Wuding River drains the area and flows into the Yellow River. The elevation ranges from 1000 to 1300 m in general, but it is as low as 950 m in many southeastern valleys and as high as 1400–1600 m in the northwest. The region has a typical arid and semiarid continental monsoonal climate, with mean annual temperatures of 6.0°–9.0°C and a mean annual precipitation of 200–400 mm, with 70% of the rainfall in summer. Northwesterly winds dominate this area, and the mean annual evaporation and aridity index are 1800–2500 mm and 1.0–2.5, respectively (Li et al. 2014). Spring is particularly prone to sandstorms from the northwest.

Fig. 1.

Geographic setting of the Yansui Garrison. (top) A sketch map of China; the purple rectangular area shows the relative location of the Ordos Plateau to China. (bottom) The military defense system of the Yellow River Loop (YRL) during the Ming dynasty; the black line indicates the frontier wall (the Great Wall), blue lines represent the rivers, black circles indicate the fortresses of the Ningxia Garrison, red circles show the relative location of old (north) and new (far away from the wall) Anbian Ying mentioned in the text; and green circles show the locations of fortresses belonging to the Yansui Garrison. The red rectangles along the border wall are (from east to west) the cities of Fugu, Shenmu, Yulin, Hengshan, Jingbian, and Wuqi. The city along the Wuding River was Mizhi (mentioned in Table 1). The digital elevation model (DEM) data in the background show the geomorphological features of the region.

Fig. 1.

Geographic setting of the Yansui Garrison. (top) A sketch map of China; the purple rectangular area shows the relative location of the Ordos Plateau to China. (bottom) The military defense system of the Yellow River Loop (YRL) during the Ming dynasty; the black line indicates the frontier wall (the Great Wall), blue lines represent the rivers, black circles indicate the fortresses of the Ningxia Garrison, red circles show the relative location of old (north) and new (far away from the wall) Anbian Ying mentioned in the text; and green circles show the locations of fortresses belonging to the Yansui Garrison. The red rectangles along the border wall are (from east to west) the cities of Fugu, Shenmu, Yulin, Hengshan, Jingbian, and Wuqi. The city along the Wuding River was Mizhi (mentioned in Table 1). The digital elevation model (DEM) data in the background show the geomorphological features of the region.

The study area is located at the southern margin of the Mu Us Desert, which is especially sensitive to climate variability. According to contemporary political divisions, this region is within the northern Shaanxi Province, which comprises 12 counties, six of which—Fugu, Shenmu, Yulin, Hengshan, Jingbian, and Wuqi—contain portions of the Great Wall. During the Ming dynasty, this region was called the Yansui Garrison or Yulin Garrison (Zhen), and it was one of the nine most vital military districts of China’s northern frontier. To protect against the invasion of nomadic tribes, particularly the Mongols, the Ming developed their defenses along the northern frontier. The term “Great Wall” was not used in the Ming period; however, the term jiubian, or “Nine Defense Area,” was used to refer to the primary garrison towns distributed along the frontier that, along with the walls, the bao (fortresses), signal stations, military towers, military farms, and other structures, comprised the entire defensive system (Dardess 2012). The protection of China’s northern border was once the main task of the city of Yulin during the Ming dynasty.

In the year 1437, the military commander Wang Zhen of the Yansui Garrison constructed 25 fortresses spanning an area that extended from Qingshui Ying (near Qingshui, Inner Mongolia) to Dingbian Ying (at present, Dingbian County) in the west (Fig. 1). Until the time of Emperor Chenghua’s reign (1465–87), the frontier wall (the Great Wall) of this region was constructed under the supervision of Yu Zijun, a famous Ming governor. Along with the construction of the wall, 36 fortresses were established on the existing foundation (Han 2003). These fortresses dotted the frontier wall from Huangfuchuan Bao in the east to Dingbian Ying in the west over a distance of 1700 li (one li equals 0.5 km); however, the construction was never completed. During the reign of Emperor Wanli (1573–1620), the number of fortresses increased to 39 (Zhang 2010). The distance between every two fortresses was no greater than 100 li, which indicates that their distribution was well planned.

In the early years of Emperor Zhengtong (1436–49), the Ming court began to construct fortresses along the northern border; Yulin contained only one such fortress, as it was a small component of a much larger-scale plan. By 1471, the military defenses of Yulin had been established, and the city acted as the administrative center of the Yansui Garrison until 1473. After the garrison moved to Yulin, the city gained an even more critical position in the defensive network. Along with the construction of more fortresses, the frontier wall was extended to the city of Yulin. After the completion of this extension, the wall experienced several periods of expansion and redevelopment. In summary, Yulin served as a crucial station and a patrol base for the massive defense system for a long period during the Ming dynasty.

b. Data and methods

Because of its vital military position, the Yansui Garrison was well documented during the Ming period, and a portion of this documentation can be used to reconstruct the changes in the historical landscape, particularly with respect to the dynamic process of regional desertification at an intermediate temporal scale. Compared with physical proxy records, historical documents provide significantly more accurate information in terms of both environmental conditions and human strategies. The analysis of historical documents in this study has two applications. First, these documents can be used to reconstruct the process of desertification. Second, they provide a record of societal responses to cold/drought events and the consequent increase in desertification.

To identify the role of climatic conditions in the process of desertification, we compiled six high-resolution and well-dated datasets that span one-half of a millennium to two millennia. These datasets can be divided into two groups that reflect either temperature or precipitation. The first dataset comprises a 2000-yr-long temperature reconstruction of China based on multiple temperature proxy records from five regions; it was reconstructed using a principal-component regression method (Ge et al. 2013). The second dataset is an annual temperature reconstruction for the last 1000 years based on a network of 415 well-distributed and accurately dated climate-proxy time series (Shi et al. 2012). These two datasets permit preliminary and macroscale research within China. A third important source of data is a valuable temperature synthesis for West Qinling (which includes our study area) based on high-resolution tree-ring records that cover the last 500 years (Yang et al. 2013). This dataset may include a more regional temperature signal.

Since the Neolithic period, the landscape dynamics, in terms of soil formation and aeolian sedimentation, have been largely attributed to changes in the East Asian summer monsoon (EASM) in semiarid areas. In the YRGB region, a lack of precipitation is recognized as one of the main factors driving the process of desertification and hampering recovery efforts. Three precipitation time series are cited here. The first is a stalagmite δ18O record from Wanxiang Cave, China, that characterizes the history of the EASM over the past 1810 years (Zhang et al. 2008). In monsoonal regions, speleothem δ18O values are considered to be a proxy of the amount of paleoprecipitation and the summer monsoon strength, in which lighter speleothem δ18O values indicate a stronger summer monsoon and higher rainfall (Tan et al. 2010; Wang et al. 2005; Zhang et al. 2008). The second dataset is based on a dryness/wetness index (DWI) at 120 locations in China from 1470 to 1999 (CAMS 1981; Zhang et al. 2003). The dryness/wetness data were derived from written descriptions of past conditions in local gazettes and other historical documents. These texts were scrutinized for their accuracy and consistency and then transformed into a quantitative DWI (Feng et al. 2013). An annual dryness/wetness grade or intensity value between 1 and 5 (representing very wet, wet, normal, dry, and very dry conditions) was assigned to each meteorological station in this region. These records were calibrated with instrument-recorded precipitation values. To understand the dry/wet variations at a local scale, we used the data for the Yulin station only to reconstruct the DWI records. For a more robust climate reconstruction and comparison on a larger scale, a third dataset that includes May–September precipitation data in Asia over the past five centuries (1470–1999) is used. This dataset was reconstructed on the basis of tree-ring data, historical documentation, ice-core records, and a few long-term instrumental time series data available for the region (Yi et al. 2012).

3. Results

a. Climate change during the Ming dynasty

1) Temperature variability during the Ming dynasty

Two composite temperature datasets help to place climatic events into the long-term context of historical temperature variations (Ge et al. 2013; Shi et al. 2012); thus, the temperature oscillations and their relationship to desertification during the Ming dynasty can be evaluated. An additional regional temperature record that covers our study area was also considered. Figure 2a shows that the two temperature time series in China reconstructed by Ge et al. (2013) and Shi et al. (2012) are in good agreement. Both time series capture the distinctly colder LIA; the start and end times of the LIA, which began in 1321, were nearly identical. Although the mean temperature was lower during the LIA, high temperature variability still occurred during this period, and several extremely cold events are evident in the records. During the Ming dynasty (1368–1644), four cooling periods were observed (1385–1465, 1515–35, 1565–75, and 1595–1695) from the curve of Ge et al. (2013), with temperature minima in 1415, 1535, 1565, and 1655. The record of Shi et al. (2012) shows considerably higher variability than that of Ge et al. (2013) during the past millennium. This discrepancy can be partially explained by the difference in data sources and the difference in reconstruction methods. Six cold events can be recognized in the series of Shi et al. (2012): 1439–56, 1501–31, 1554–64, 1593–98, 1618–32, and 1639–96. The coldest years are 1449, 1510, 1559, 1595, 1623, and 1670 (Fig. 2a). Despite the differences in the timing of the coldest temperatures, the cold phases of these two records are consistent. The cold periods of 1439–56 and 1501–31 correspond to the periods of 1385–1465 and 1515–35, respectively. The last three cold periods (1593–98, 1618–32, and 1639–96) identified by Shi et al. (2012) correspond to the cold period of 1595–1695 reconstructed by Ge et al. (2013). This level of agreement suggests consistency between the climate patterns of these two time series. Figure 2b depicts the reconstructed regional temperature anomalies (with respect to 1961–90) over the past 500 years, which are pertinent to our study area. The temperature reconstruction for this region captures three distinct cold periods, approximately 1520–35, 1560–75, and 1610–20, during the Ming dynasty (Yang et al. 2013). The dates of the first two cold periods are very similar to those identified from the record for the past 2000 years. However, the duration of the last cold period is significantly shorter than those of the other two composite time series for China. This difference in duration suggests a regional nature of this climatic feature. In conclusion, three phases of cold events can be identified during the Ming dynasty based on the three proxy records. However, because of the limited time span of the third record, we cannot verify the occurrence of the cold event from 1385 to 1465.

Fig. 2.

Composite temperature time series in China and West Qinling. (a) The blue curve is the temperature time series for the past two millennia (Ge et al. 2013), and the purple curve is the temperature reconstruction for the past millennium (Shi et al. 2012). (b) The regional temperature record for West Qinling that covers our study area over the past 500 years (Yang et al. 2013).

Fig. 2.

Composite temperature time series in China and West Qinling. (a) The blue curve is the temperature time series for the past two millennia (Ge et al. 2013), and the purple curve is the temperature reconstruction for the past millennium (Shi et al. 2012). (b) The regional temperature record for West Qinling that covers our study area over the past 500 years (Yang et al. 2013).

2) Precipitation and drought record during the Ming dynasty

Three precipitation datasets are cited here. Figure 3 shows a period characterized by generally weak monsoons from the fourteenth to the nineteenth centuries punctuated by four severe droughts during 1476–1502, 1509–37, 1577–90, and 1604–53. The minima among them are centered on 1483–87, 1528–34, 1583–87, and 1633–37. These drought events are evident in all three time series, indicating that the climatic trends were regional rather than local. These dry periods are correlated with reduced summer insolation in the Northern Hemisphere, a southward displacement of the intertropical convergence zone (ITCZ), and a weak EASM. Many studies have also recorded these drought events, particularly the drought during the late Ming dynasty of 1637–43, which is recognized in several historical records as the most severe drought in China over the past five centuries (Shen et al. 2007).

Fig. 3.

Precipitation time series for northern China. (a) Speleothem δ18O record for the Wanxiang Cave for the past 1810 years (Zhang et al. 2008). (b) A subset of the Wanxiang Cave δ18O records dating back to the year 1470 for comparison (Zhang et al. 2008). (c) Reconstruction of the May–September precipitation for the period 1470–1900 (Yi et al. 2012). (d) DWI record for the Yulin station for 1470–1999 (Chinese Academy of Meteorological Sciences 1981). The lines represent 10-yr intervals. The four vertical blue bars denote occurrences of the four drought events.

Fig. 3.

Precipitation time series for northern China. (a) Speleothem δ18O record for the Wanxiang Cave for the past 1810 years (Zhang et al. 2008). (b) A subset of the Wanxiang Cave δ18O records dating back to the year 1470 for comparison (Zhang et al. 2008). (c) Reconstruction of the May–September precipitation for the period 1470–1900 (Yi et al. 2012). (d) DWI record for the Yulin station for 1470–1999 (Chinese Academy of Meteorological Sciences 1981). The lines represent 10-yr intervals. The four vertical blue bars denote occurrences of the four drought events.

b. Climate drivers of land degradation and desertification

1) Land use and desertification records based on historical documents

Fig. 4 and Table 1 describe all of the evidence of land use and severe desertification events extracted from the historical documents. In the early years of the Ming dynasty, desirable land was available. In 1435, the frontier officer built up a city named Anbian Ying with lush grasslands and fertile soil. Because of the crucial military position of Yulin after the 1460s, increasing attention was devoted to the environmental changes along the Mu Us Desert margin by frontier officers of the Yansui Garrison. During this period, an officer named Wang Hong, who was assigned as minister of war to manage the frontier affairs said, “In Yansui and Qingyang [near Yansui], there is a very broad area of land very suitable for farming” (see Table 1, 1460s, for more information). This suggests that land resources were plentiful and that no large-scale desertification phenomena had occurred at that time (Table 1). During the 1470s, the local region of the Yansui Garrison experienced slight desertification. The book Huang Ming Fu Shi Bian (see Table 1, 1470s) noted that

[d]uring the early years of Emperor Zhengtong’s reign (about 1435), two vital military fortresses Dingbian Ying and Anbian Ying were built in the western part of the Yansui Garrison. However, to the years of Emperor Chenghua (1470s), this region has been surrounded by desert. At that time, when the provincial governor Yu Zijun wanted to build the new border wall, he found that there was no place here suitable for the project. Therefore, he just added a small fortress named Yongji east of the Dingbian Ying.

In spite of that, in 1473, provincial governor Ma Wensheng said that

[t]o reclaim wasteland is one of the important policies of frontier defenses. On the south of Yulin City, there is a large area of land that can be planted. In the recent year, officers advocated people to grow crops (Table 1, year 1473).

This shows that there was still a considerable amount of land available during the 1470s.

Fig. 4.

Spatial and temporal development of desertification based on historical documents and climate data. (top left) The land use conditions before the 1470s. During this period, there were still large amounts of land available except for the desertification around Dingbian Ying and Anbian Ying. (top right) The first desertification occurred in 1529–46. Yulin and the 1000-li frontier wall of the Yansui Garrison suffered from severe desertification. (bottom left) The second severe desertification likely occurred around the 1570s (1568–74), which was characterized by decreased land resources and severely desertified fortresses and desertified cities. (bottom right) The last stage of desertification, which occurred in 1601–50. At this time the degree of desertification reached its peak. The numbers of desertified fortresses increased dramatically and the town city of Yulin almost lost its function of defense because of the severe desertification.

Fig. 4.

Spatial and temporal development of desertification based on historical documents and climate data. (top left) The land use conditions before the 1470s. During this period, there were still large amounts of land available except for the desertification around Dingbian Ying and Anbian Ying. (top right) The first desertification occurred in 1529–46. Yulin and the 1000-li frontier wall of the Yansui Garrison suffered from severe desertification. (bottom left) The second severe desertification likely occurred around the 1570s (1568–74), which was characterized by decreased land resources and severely desertified fortresses and desertified cities. (bottom right) The last stage of desertification, which occurred in 1601–50. At this time the degree of desertification reached its peak. The numbers of desertified fortresses increased dramatically and the town city of Yulin almost lost its function of defense because of the severe desertification.

Table 1.

Land use and Desertification records based on historical documents

Land use and Desertification records based on historical documents
Land use and Desertification records based on historical documents

According to the degree of desertification, three periods of desertification are evident (Fig. 4). The first phase was 1529–46, when Yulin and the 1000-li frontier wall of the Yansui Garrison suffered from severe desertification. The second phase occurred around the1570s. During this time, the extent of arable land decreased and only occupied 30%–40% of the total area; by 1574, the local desertification of DingbianYing had increased in intensity. The last and severest phase occurred in 1601–16; the extent of desertification threatened the human population because of the large decreases in arable land area and water resources. Meanwhile, the defense function of the buried frontier wall and fortresses was diminished almost entirely.

2) Historical climate-driven desertification

In recent decades, significant attention has been focused on the historical desertification of the Mu Us Desert. What factors contributed to the environmental deterioration (desertification intensity) of the Mu Us Desert? Hou (1985) claimed that the deterioration of the environment should be mainly attributed to overuse of land for agriculture. After this, other researchers arrived at similar conclusions (Deng et al. 2001; Zhu and Wang 1992). However, some authors found that periodic cold and drought events had a more important role on desertification intensity (Han 2003; Zhao 1981, 1990). At present a consensus has yet to be reached on this question. The previous method of environmental analysis that depended on historical documents cannot resolve this problem. Modern monitoring data suggest that recent desertification in China has resulted primarily from climate change rather than human impacts (Wang et al. 2006, 2008, 2010; Yang et al. 2010). Recently, several multidisciplinary integration studies that included both historical documents and high-resolution paleoclimatological records also showed that climate was the primary driver of historical desertification of the Mu Us Desert (Cui and Chang 2013; Huang et al. 2009).

Figure 5 includes a comparison of temperature, precipitation, and desertification records. As demonstrated in previous studies, a significant positive correlation exists between temperature and precipitation at centennial to multidecadal scales. However, there is no significant correlation between the temperature and precipitation series cited here at the decadal time scale (r = 0.051, p = 0.253). Cold–dry, warm–wet, cold–wet, and warm–dry combinations are all possible at this time scale. Despite that, the three intense desertification events identified from historical documents were observed to occur during cold–dry periods, so these conditions may enhance the desertification process. Persistent high-frequency, low-temperature drought events caused environmental crises associated with severe desertification, land degradation, and reduced water resources. During the late Ming dynasty, an exceptionally cold, long-lasting drought (1604–50) occurred in here, although historical records of desertification extend only to 1616 (Table 1). One possible explanation for this abrupt end in record keeping was the relocation of the military defense center in 1616 (through the end of the Ming dynasty in 1644). Several pertinent factors related to the move include 1) an improved Sino–nomadic relationship with the Yansui Garrison after 1570, 2) the arrival of Manchurians from northeastern China that began in 1615, and 3) the relocation of the Ming government forces to that region. In addition, resisting the peasant uprising due to the megadrought and associated famine became the primary duty of government troops, and the fortresses and frontier wall became a lower priority between 1616 and 1644. Consequently, no historical documents are available for the purpose of extracting a desertification record in these years. Despite the lack of a record of desertification after 1616, the long duration of cold/drought climatic conditions indicates that the third desertification process would have endured at least until the end of the Ming dynasty. We support this view on the basis of the compiled data (Fig. 5). Historical records of desertification at the start of the Ming dynasty (1368) to 1400 also remain undocumented, and the available regional high-resolution proxy records cover only the past 500 years. An analysis of additional historical documents and longer proxy records related to our study area is the next research goal for studying desertification in the early period of the Ming dynasty.

Fig. 5.

Comparison between (a) the temperature time series for West Qinling and (b) the DWI record from the Yulin station at 10-yr resolution. The light blue curves represent the original unfiltered data. The three vertical yellow bars denote the timing of severe desertification events.

Fig. 5.

Comparison between (a) the temperature time series for West Qinling and (b) the DWI record from the Yulin station at 10-yr resolution. The light blue curves represent the original unfiltered data. The three vertical yellow bars denote the timing of severe desertification events.

4. Discussion

a. Societal responses to desertification: A case study of the frontier wall and fortresses

1) The controversy regarding the construction of the frontier wall

In the early Ming dynasty, the emperors attempted to create buffer zones around their borders by making treaties with friendly tribes and posting garrisons and watchtowers hundreds of kilometers into the steppe; however, after 1449, Mongol tribes moved into the Ordos region, settled around the YRGB, and repeatedly encroached upon China. Thus, the military defense of this region became a high priority. Yu Zijun was a hardworking officer devoted to improving conditions in the northwestern region under his supervision. In 1471, Yu Zijun recommended that the emperor construct a wall between the Chinese settlements and the Ordos to protect the local people. The minister of war was hesitant to comply but wanted to drive the Mongols out of the area. One argument opposing the construction of the wall was the high percentage of sand in the local soil: even if a wall was built, it would eventually collapse (Liu 1962). The Beijing administration sent General Wang Yue to force the Mongols out of the Ordos. However, this great enterprise was not completed because of the drought, poor support, and insufficient forces. Lacking better alternatives, the court agreed to build the wall (Liu 1962). Finally, in 1473 a wall stretching from west to east with a length of approximately 1700 li was constructed (Shu 2012). In summary, the main barrier to the construction of the wall was not the sandy soil but rather frontier policy. During this period, a large extent of arable land was available, and only low-intensity local desertification occurred at two fortresses in the western Yansui Garrison (Table 1; Fig. 1).

2) Balance between military defense and combating desertification: Movement of fortress locations and the frontier wall

In addition to the 1700-li frontier wall, Yu Zijun also constructed 12 fortresses, 15 border towers, 78 small towers, and 719 cliff-top stockades. One fortress was Changle Bao, which was located to the east of Yulin (Fig. 4). In 1489, because of the extent of desertification and lack of water, the provincial governor, Lu Xiang, relocated Changle Bao 20 li to the north (Fig. 4). Changle Bao was relocated to the north rather than south for two reasons. First, the desert was not continuous, but local desertification existed; hence, hospitable land was available to construct a new fortress to the north. Second, a move north was beneficial in the defense against enemies from the northern steppe. Earlier, Yu Zijun had also implemented steps to decrease the impact of local desertification. He moved Anbian Ying from the desert margin to the loess region, which was approximately 60 li from the frontier wall (Fig. 4). Later, another officer named Wang Lun moved Anbian Ying back to its original location, which was closer to the frontier wall. Thus, military defense was ultimately given priority over combating environmental problems.

3) Combating desertification

The Yansui Garrison continually suffered from environmental problems because of its proximity to the desert margin and the intensified desertification under the cold/drought climatic conditions. During the late Ming period, the wall and several fortresses near Yulin were so completely covered by sand that one could gallop over them on horseback. Hence, the intensity of desertification had endangered the safety of this frontier region. Substantial effort was devoted to protecting the defense system against erosion and blowing sand during the Ming dynasty. For example, two provincial officers (Sun Weicheng and Tu Zongjun) attempted to remove the sand in 1601 and 1610, respectively (Table 1). The attempt by Sun Weicheng was not completed because of a revolt by the soldiers (Table 1), while Tu Zongjun successfully led activities that removed the sand from around the wall. Although the landscape was temporarily modified by the project, Tu Zongjun was also conscious of the ineptitude of any long-term attempt to control nature (Table 1). Subsequent records demonstrate that the efforts by Tu Zongjun were performed in vain. A book published in 1616 (see Table 1, year 1616) stated that the landscape suffered from severe desertification. Ultimately, the human-waged war against nature failed.

b. Climate change, desertification, and the economic crisis of Yansui Garrison

The impact of climate change and desertification on the economic crisis of Yansui Garrison is reflected in the following aspects.

First, when the cold/drought climate led to desertification, the subsequent loss of arable land caused a collapse of the military farm system. The farm system had been the main source for the provisioning of government troops on the northern frontiers. Up to 1568, one office stated that the Yansui Garrison stretched along the frontier wall from east to west for a length of 1500 li. The arable land only occupied 30%–40% of the total area (Table 1). From 1566 to 1620, the yields on the military farms on the northern frontier decreased by 60%–70%. This drop was attributed to a sharp increase in expenditures and was exacerbated by fiscal deterioration during the late Ming dynasty (Zheng et al. 2014). Payments to soldiers in the area were often delinquent, and the life of a soldier incurred many hardships, including the danger of hunger. In extreme cases, some even fled or rebelled. A large number of soldiers joined the ranks of desperate refugees and became an integral component of the peasant uprisings.

Second, serious desertification increased the difficulty of maintaining a sound defense of the frontier areas (Table 1; refer to the years 1601, 1608, and 1610). Although the minority invasion eased after 1570, Yansui town remained an important node of the northern defense system. Severe desertification meant that the edge of the wall and the fort would continue to be buried, in extreme cases completely undermining their defense function (e.g., Table 1, years 1601, 1608, and 1610). Local conservation would have to spend significant manpower and material resources to maintain the defense system. From the historical literature, the strategy appears to have been to clean the sand, cultivate plants that grow well in sand, and maintain high walls (Zhang 2010). However, human efforts could only achieve temporary success, and in the long term human efforts were unable to combat the rapid development of desertification. Especially in the late Ming dynasty, the frontier defense system had fallen into a vicious cycle of serious shortages of funds and serious degradation of the ecological environment (Table 1, refer to the year 1616).

Third, drought and degraded land contributed to peasant uprisings and the development of peasantry troops. The drought event in the late Ming was the most extensive one over the last five centuries (Shen et al. 2007). The climatic deterioration in the late Ming dynasty (cooling, aridification) led directly to desertification and a sharp decrease in arable land. Along the margin of the Mu Us Desert, a large decrease in crop yields and a food crisis developed during the late Ming dynasty. At the same time, the people were enduring a high tax burden because of the military pressure from the northeast frontier. In contrast, governmental efforts toward disaster relief in the late Ming dynasty were insufficient, and the governmental capacity to support disaster relief continued to decrease in the early seventeenth century (Xiao et al. 2015). There was severe famine in 1628 in the Shaanxi area, especially in northern Shaanxi (Hui 1802; Liu and Shen 2012). A local officer appealed to the government of Emperor Chongzhen for a tax reduction and relief (Ma 2012), but the relief took two years to arrive (Guo 2014). Even worse, the unprecedentedly severe drought continued and exacerbated the economic crisis of the region. Hence, local efforts to find relief were severely limited. The peasants had to turn to foods such as Conyza canadensis (horseweed) and tree bark that are not eaten in a normal year. When running out of food, they even ended their hunger by eating pieces of stone, which resulted in death. When the famine was at its worst, cannibalism was reported (Ma 2012). Continuous drought, difficult transport from other regions, and the inability of government relief led to high grain prices (Guo 2014; Lai 2010). A large number of people lost their lives in the famine and the population declined sharply. When food/arable land becomes a scarcity, those who survive become uprooted and go on the move. A large number of refugees and hungry soldiers together formed an army and triggered massive riots (Li 1948). The first large-scale peasant uprising took place here, testifying to the severe societal stresses that were concentrated in the area. A peasant soldier named Li Zicheng (1606–45) mutinied with his fellow soldiers in northern Shaanxi (Mu Us Desert margin) in the early 1630s after the government failed to send supplies. Next, people rebelled in a number of localities, and they began to band together. The Ming troops were dispirited and perhaps underfed (Li 1948). In 1644, Li Zicheng’s troops entered Beijing. The last Ming emperor hung himself on a tree, ushering in the final decline of the Ming dynasty.

5. Conclusions

During the Ming dynasty, the Yansui Garrison, which was located at the southern margin of the Mu Us Desert, continually suffered from environmental problems caused by intensive desertification. Three periods of severe desertification were reconstructed based on historical documents and climate data. A comparative analysis indicates that these periods of intense desertification were correlated with cold/drought climate conditions.

At the centennial scale, a long-lasting weak EASM and low temperatures dominated the climate during the Ming dynasty; this climate was associated with drought and intense desertification. Our composite proxy records also characterize the spatiotemporal variations in precipitation and temperature over the past 500 years at a decadal time scale. The internal structure of the different temperature and precipitation time series records show good agreement, but significant spatiotemporal disparities occur between temperature and precipitation records.

During the early years of the Ming dynasty, a complex defense system composed of a frontier wall, fortresses, and border towers was constructed despite increasing desertification. The safety of the frontier region was the main concern of the state at that time. Although numerous activities were pursued by local officers to reduce the influence of desertification, their efforts were unable to cope with the pervasive deterioration of the natural environment.

The passive defense policy employed during the majority of the Ming dynasty was associated with great expenditures to support the forces and to repair the defense fortifications. During the late Ming dynasty, the state suffered from a severe economic crisis. The impact of this economic crisis on the fragile Yansui Garrison was particularly serious. The impacts of drought, famine, desertification, and land degradation, in combination with the other economic problems of the time, led to the famous peasant rebellions, which was one of the reasons for the eventual fall of the Ming dynasty.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (41571190), Key Research Program of the Chinese Academy of Sciences (KZZD-EW-04), the Fundamental Research Funds for the Central Universities (Grant 10SZYB10), and the Major project of National Social Science Fund (14ZDB031). The authors declare there are no conflicts of interest regarding the publication of this paper. Our deepest gratitude goes to the anonymous reviewers for their careful work and thoughtful suggestions that have helped improve this paper substantially.

REFERENCES

REFERENCES
Chinese Academy of Meteorological Sciences
,
1981
: Atlas of Flood and Drought in China in the Last 500 Years. Cartological Press, 332 pp.
Cook
,
E. R.
,
K. J.
Anchukaitis
,
B. M.
Buckley
,
R. D.
D’Arrigo
,
G. C.
Jacoby
, and
W. E.
Wright
,
2010
:
Asian monsoon failure and megadrought during the last millennium
.
Science
,
328
,
486
489
, doi:.
Cui
,
J. X.
, and
H.
Chang
,
2013
:
The possible climate impact on the collapse of an ancient urban city in Mu Us Desert, China
.
Reg. Environ. Change
,
13
,
353
364
, doi:.
Dardess
,
J. W.
,
2012
: Ming China, 1368–1644: A Concise History of a Resilient Empire. Rowman and Littlefield, 172 pp.
deMenocal
,
P. B.
,
2001
:
Cultural responses to climate change during the Late Holocene
.
Science
,
292
,
667
673
, doi:.
Deng
,
H.
,
Z.
Xia
, and
B.
Wang
,
2001
:
Human activities’ influence on the area of weak ecological environment in terms of the rise and fall of Tongwan City (in Chinese)
.
Collect. Essays Chin. Hist. Geogr.
,
16
,
104
113
.
Fan
,
K.-W.
,
2015
:
Climate change and Chinese history: A review of trends, topics, and methods
.
Wiley Interdiscip. Rev.: Climate Change
,
6
,
225
238
, doi:.
Feng
,
S.
,
Q.
Hu
,
Q. R.
Wu
, and
M. E.
Mann
,
2013
:
A gridded reconstruction of warm season precipitation for Asia spanning the past half millennium
.
J. Climate
,
26
,
2192
2204
, doi:.
Ge
,
Q. S.
,
Z. X.
Hao
,
J. Y.
Zheng
, and
X. M.
Shao
,
2013
:
Temperature changes over the past 2000 yr in China and comparison with the Northern Hemisphere
.
Climate Past
,
9
,
1153
1160
, doi:.
Guo
,
C.
,
2014
: Drought, Famine and Societal Responses of North Shaanxi during the Emperor Chongzhen Ming Dynasty (in Chinese). Shaanxi Normal University, 191 pp.
Han
,
Z. Q.
,
2003
:
The evolution of the Maowusu Desert and the reclamation in the adjacent areas in the Ming Dynasty (in Chinese)
.
Soc. Sci. China
,
5
,
191
204
.
Hou
,
R.-Z.
,
1985
:
Ancient city ruins in the deserts of the Inner Mongolia Autonomous Region of China
.
J. Hist. Geogr.
,
11
,
241
252
, doi:.
Huang
,
Y. Z.
,
N. A.
Wang
,
T. H.
He
,
H. Y.
Chen
, and
L. Q.
Zhao
,
2009
:
Historical desertification of the Mu Us Desert, northern China: A multidisciplinary study
.
Geomorphology
,
110
,
108
117
, doi:.
Hui
,
H.
,
1802
: Jiaqing Yan'an Fu Zhi. Vol 6, Shanghai Classics Publishing House, 559 pp.
Lai
,
J.
,
2010
: Provisions and funds for frontier troops and the economic crisis of late Ming, 1531–1602 (in Chinese). Zhejiang University Press, 464 pp.
Li
,
Q.
,
H. B.
Wu
,
Y. Y.
Yu
,
A. Z.
Sun
,
S. B.
Marković
, and
Z. T.
Guo
,
2014
:
Reconstructed moisture evolution of the deserts in northern China since the Last Glacial Maximum and its implications for the East Asian summer monsoon
.
Global Planet. Change
,
121
,
101
112
, doi:.
Li
,
W. Z.
,
1948
: Wan Ming Min Bian. Commerical Press, 225 pp.
Liu
,
J.
,
1962
: Ming Xian Zong Shi Lu. Vol. 93, Institute of History and Philology of Academia Sinica, 102 pp.
Liu
,
Y. Y.
, and
Q. Y.
Shen
,
2012
: Qing Yong Zheng Chi Xiu Shaan Xi Tong Zhi. Vol 47, Sanqin Press, 84 pp.
Ma
,
M. C.
,
2012
: Bei Chen Da Ji Shu. Qing Yong Zheng Chi Xiu Shaan Xi Tong Zhi, Vol 86, Y. Y. Liu, and Q. Y. Shen, Eds., Sanqin Press, 51–54.
Shen
,
C. M.
,
W. C.
Wang
,
Z. X.
Hao
, and
W.
Gong
,
2007
:
Exceptional drought events over eastern China during the last five centuries
.
Climatic Change
,
85
,
453
471
, doi:.
Shi
,
F.
,
B.
Yang
, and
L.
Gunten
,
2012
:
Preliminary multiproxy surface air temperature field reconstruction for China over the past millennium
.
Sci. China Earth Sci.
,
55
,
2058
2067
, doi:.
Shu
,
S. G.
,
2012
:
The construction time and process of the “Erbian” Great Wall of the Yansui Precinct in the Ming Dynasty under the supervision of Yu Zijun
.
J. Yanan Univ.
,
34
,
102
109
.
Tan
,
L. C.
,
Y. J.
Cai
,
Z. S.
An
,
R. L.
Edwards
,
H.
Cheng
,
C. C.
Shen
, and
H. W.
Zhang
,
2010
:
Centennial- to decadal-scale monsoon precipitation variability in the semihumid region, northern China during the last 1860 years: Records from stalagmites in Huangye Cave
.
Holocene
,
21
,
287
296
, doi:.
Wang
,
X. M.
,
F. H.
Chen
, and
Z. B.
Dong
,
2006
:
The relative role of climatic and human factors in desertification in semiarid China
.
Global Environ. Change
,
16
,
48
57
, doi:.
Wang
,
X. M.
,
F. H.
Chen
,
E.
Hasi
, and
J. C.
Li
,
2008
:
Desertification in China: An assessment
.
Earth Sci. Rev.
,
88
,
188
206
, doi:.
Wang
,
X. M.
,
F. H.
Chen
,
J.
Zhang
,
Y.
Yang
,
J.
Li
,
E.
Hasi
,
C.
Zhang
, and
D.
Xia
,
2010
:
Climate, desertification, and the rise and collapse of China’s historical dynasties
.
Hum. Ecol.
,
38
,
157
172
, doi:.
Wang
,
Y. J.
, and Coauthors
,
2005
:
The Holocene Asian monsoon: Links to solar changes and North Atlantic climate
.
Science
,
308
,
854
857
, doi:.
Weiss
,
H.
, and
R. S.
Bradley
,
2001
:
What drives societal collapse?
Science
,
291
,
609
610
, doi:.
Xiao
,
L. B.
,
X. Q.
Fang
,
J. Y.
Zheng
, and
W. Y.
Zhao
,
2015
:
Famine, migration and war: Comparison of climate change impacts and social responses in North China between the late Ming and late Qing dynasties
.
Holocene
,
25
,
900
910
, doi:.
Yang
,
B.
,
S. Y.
Kang
,
F. C.
Ljungqvist
,
M. H.
He
,
Y.
Zhao
, and
C.
Qin
,
2014
:
Drought variability at the northern fringe of the Asian summer monsoon region over the past millennia
.
Climate Dyn.
,
43
,
845
859
, doi:.
Yang
,
F. M.
,
N. A.
Wang
,
F.
Shi
,
F. C.
Ljungqvist
,
S. G.
Wang
,
Z. X.
Fan
, and
J. W.
Lu
,
2013
:
Multi-proxy temperature reconstruction from the West Qinling Mountains, China for the past 500 years
.
PLoS One
,
8
,
e57638
, doi:.
Yang
,
Y. M.
,
Z. L.
Guo
, and
G. H.
Yang
,
2010
:
Coupling effect of natural and artificial factors on desertification in Mu Us sandland
.
J. Anhui Agric. Sci.
,
25
,
13 934
14 010
.
Yi
,
L.
,
H. J.
Yu
,
J. Y.
Ge
,
Z. P.
Lai
,
X. Y.
Xu
,
L.
Qin
, and
S. Z.
Peng
,
2012
:
Reconstructions of annual summer precipitation and temperature in north-central China since 1470 AD based on drought/flood index and tree-ring records
.
Climatic Change
,
110
,
469
498
, doi:.
Zhang
,
D. D.
,
C. Y.
Jim
,
G. C. S.
Lin
,
Y. Q.
He
,
J. J.
Wang
, and
H. F.
Lee
,
2006
:
Climatic change, wars and dynastic cycles in China over the last millennium
.
Climatic Change
,
76
,
459
477
, doi:.
Zhang
,
D. D.
,
H. F.
Lee
,
C.
Wang
,
B. S.
Li
,
Q.
Pei
,
J.
Zhang
, and
Y. L.
An
,
2011a
:
The causality analysis of climate change and large-scale human crisis
.
Proc. Natl. Acad. Sci. USA
,
108
,
17 296
17 301
, doi:.
Zhang
,
D. D.
,
H. F.
Lee
,
C.
Wang
,
B. S.
Li
,
J.
Zhang
,
Q.
Pei
, and
J. A.
Chen
,
2011b
:
Climate change and large-scale human population collapses in the pre-industrial era
.
Global Ecol. Biogeogr.
,
20
,
520
531
, doi:.
Zhang
,
D. E.
,
X. Q.
Li
, and
Y. Y.
Liang
,
2003
:
Supplement to the atlas of flood and drought in China in the last 500 years (1993–2000) (in Chinese)
.
J. Appl. Meteor. Sci.
,
14
,
379
388
.
Zhang
,
P.
,
2010
:
Rise and fall: Deciphering urban sites and their environment in overlapping agrarian/pastoral regions (in Chinese)
.
Soc. Sci. China
,
31
,
84
102
, doi:.
Zhang
,
P. Z.
, and Coauthors
,
2008
:
A test of climate, sun, and culture relationships from an 1810-year Chinese cave record
.
Science
,
322
,
940
942
, doi:.
Zhao
,
Y. F.
,
1981
:
Problems about the change of the Mu Us Desert during the history (in Chinese)
.
Hist. Geogr.
,
1
,
34
47
.
Zhao
,
Y. F.
,
1990
:
Revisiting the change of the Mu Us Desert during the history (in Chinese)
.
Hist. Geogr.
,
7
,
171
180
.
Zheng
,
J. Y.
,
L. B.
Xiao
,
X. Q.
Fang
,
Z. X.
Hao
,
Q. S.
Ge
, and
B. B.
Li
,
2014
:
How climate change impacted the collapse of the Ming dynasty
.
Climatic Change
,
127
,
169
182
, doi:.
Zhu
,
Z. D.
, and
T.
Wang
,
1992
:
Theory and practice on sandy desertification in China (in Chinese)
.
Quat. Sci.
,
1992-02
,
97
106
.