Regional Environmental Change

, Volume 18, Issue 6, pp 1721–1730 | Cite as

Famine relief, public order, and revolts: interaction between government and refugees as a result of drought/flood during 1790–1911 in the North China Plain

  • Lingbo Xiao
  • Xiuqi Fang
  • Wanyi Zhao
Original Article


As a direct result of climate change and extreme weather events, refugee migration has attracted worldwide concern. In this study, we select the North China Plain (NCP) as the case area, where the capital (Beijing) is located during the late Qing dynasty (1790–1911). Based on records of porridge charity, public order (inside Beijing), and revolt events (outside Beijing) in ancient official documents, the interaction between behavior of refugees and the governmental management after floods and droughts is analyzed. The main conclusions can be summarized as follows. (1) Flood/drought has become an important triggering factor of refugee migration during 1790–1911 besides the social factors (e.g., man-land contradiction, fiscal crisis). (2) A negative interaction has existed between the government and refugees throughout the late nineteenth century, which leads to social disorder in the mid-1890s in Beijing and large-scale revolt (the Boxers Movement) in 1900 in the NCP. (3) The negative interaction has been finally ended by the adjustment of migration policy at the beginning of the twentieth century, which officially permits the refugees from the NCP to emigrate to Manchuria and Mongolia outside the Great Wall. The historical case study may provide valuable lessons for the successful adaptation of human societies to future similar occurrences, and will enhance our understanding of the interactions between climate change and social vulnerability.


Flood/drought Refugee migration Governance mechanisms Late Qing dynasty (1790–1911) North China Plain 


The global climate system is suffering rapid turbulence. In the future, the adaptability of human society will be increasingly tested severely by aggregate extreme weather events and severe ecological crises related to climate change (Lal et al. 2012). Some countries and regions, such as Niger (Afifi 2011), and the Gulf Coast (Myers et al. 2008) have vulnerable ecological systems, or the livelihood of the people depends on natural resources, climate change, and hydro-meteorological disasters that have become a “trigger” for mass migration, and the forced migration appears to be an emerging issue that requires additional scrutiny from the governments (Lal et al. 2012). Meanwhile, the solution of some relative scientific problems, such as how climatic factors trigger displacement and migration, will benefit from the study on interactive mechanism between climate change and human society.

To date, majority of studies on migration related to climate change are based on real cases that occur in recent years (Adger et al. 2014). Situating a migration event in a long time series will enable researchers to obtain additional information, such as long-term change in refugee conditions (scale, behavior, or destination), results of emergency measures (relief and resettlement), and subsequent social impact on both origin and destination. Logically, an additional historical case study in the future is necessary to cover the shortage of recent cases in a spatio-temporal scale. In the preindustrial period, human society was vulnerable to environmental pressures. In many historical cases, the collapse of regional society and mass migration can be correlated to climate change and extreme disasters (Hsu 1998; Haug et al. 2003; Büntgen et al. 2011). However, the role of climate in migration and conflict remains to be clearly identified. Generally, the immediate impact of climate on the past human society has been concentrated in agricultural and livestock production as the foundation of food security (GECAFS 2006). From crop failure caused by bad weather to displacement of refugees, many intermediate stages exist, such as market regulation and governmental management, which will mitigate or even aggravate the negative impact of climatic factors. Particularly, the role played by government is a major consideration.

As a civilization with a history of over 4000 years, China has a socioeconomic system that has been severely affected by the East Asian Monsoon, especially extreme floods and droughts throughout history (Ge 2011). Numerous studies have investigated the relationship between long-term climate change and human dimensions, such as grain harvest (Yin et al. 2015), economic fluctuation (Wei et al. 2015), fiscal balance (Wei et al. 2014), population (Lee and Zhang 2010, 2013), war (Zhang et al. 2007), and ancient dynastic cycle (Zheng et al. 2014; Yin et al. 2016). Mass displacement and migration triggered by climatic factors and their consequences has also been discussed in some cases at large spatial (national) and temporal (centennial to millennial) scales (Hsu 1998; Fang et al. 2015). Typical regional cases and additional details involving the aforementioned intermediate stages are important to reconstruct a complete process that considers the generation of refugees, the governmental response, and the interaction between the two.

In this paper, we selected the North China Plain (NCP), a typical preindustrial, primarily agricultural society during the Qing dynasty (1644–1911), as the study area. Based on historical documents and existing research on historical climate, a group of proxy series describing the extreme disasters (flood/drought), behavior of refugees, governmental management, and consequent social impact have been quantitatively reconstructed and analyzed. Then, the interaction between refugees and the government during the late Qing dynasty (1790–1911) when refugees migrating towards Beijing have become uncontrollable gradually will be discussed.

Study area

The NCP is located between 113.4° E–120.2° E and 34.6° N–41.4° N. Twenty-two prefectures (“Fu”) and 198 counties (“Xian”) cover the majority of Zhili province (south of the Great Wall), northeastern Henan, and northwestern Shandong during the Qing dynasty, according to administrative divisions in 1820 (Fig. 1) (Tan 1987; Niu 1990). This region roughly covers modern Beijing and Tianjin, majority of Hebei, and part of Henan and Shandong.
Fig. 1

Administrative division of North China Plain in 1820

The NCP is located in a warm temperate zone with a strong impact of the East Asian Monsoon. Given the high variability of annual and inter-annual precipitation, severe flood/drought often occurs and has become a serious threat to agricultural production (Li 1990). With Beijing in its heartland, NCP is the political center of the Qing government. The disaster relief (tax cut, cash, or food relief) directly controlled by the central government is important to control the scale of refugees during the peak period of the Qing (Will 1990; Li 2007; Xiao et al. 2014). Meanwhile, for some refugees, the vast uncultivated land outside the Great Wall (Inner Mongolia and Manchuria) is suitable for resettlement and reclamation. Thus, interregional migration from NCP to Manchuria and Inner Mongolia during the Qing dynasty has been discussed as a spontaneous social response to climate change and extreme disasters (Ye et al. 2012; Xiao et al. 2013).

However, Beijing, although an important destination, has not attracted enough attention. In fact, for many disaster victims, short-term and short-distance migration into the capital to receive relief from the central government is an economical and rational choice. This centripetal refugee flow after an extreme disaster has ran throughout the whole Qing dynasty and has gradually become uncontrollable over time (Chi and Li 2007).

Sources and methods

Flood/drought index

Flood/drought in the NCP during 1790–1911 is identified with the data on the dryness/wetness grade of 10 stations (Fig. 1) in Yearly charts of dryness/wetness in China for the last 500-year period (classified into 5 grades: 1, very wet; 2, wet; 3, normal; 4, dry; and 5, very dry) (China Meteorological Administration 1981). Annual drought index (Pd) and flood index (Pf) are calculated using weighted average method according to Chen (1989).
$$ D=\frac{2A}{N} $$
$$ F=\frac{2B}{N} $$
$$ {D}^{\hbox{'}}=\frac{A1}{N} $$
$$ {F}^{\hbox{'}}=\frac{B1}{N} $$
$$ Pd=\frac{Di}{D\max}\times W1+\frac{D^{\hbox{'}}i}{D^{\hbox{'}}\max}\times W2 $$
$$ Pf=\frac{Fi}{F\max}\times W1+\frac{F^{\hbox{'}}i}{F^{\hbox{'}}\max}\times W2 $$

In the preceding equations, A is the sum of stations affected by drought (dryness grade = 4 or 5), N is the sum of stations (= 10), B is the sum of stations affected by flood (wetness grade = 1 or 2), A1 is the sum of stations severely affected by drought (dryness grade = 5), and B1 is the sum of stations severely affected by flood (wetness grade = 1). Di (Fi) represents the area affected by drought (flood), and D′i (F′i) represents the area severely affected by drought (flood) in a single year during 1790–1911. Dmax (Fmax, Dmax, Fmax) is the maximal value in the time series. W1 and W2 are the weights assigned as 0.4 and 0.6, respectively. Pd and Pf range from 0 to 1, which means that the intensity of disaster varies from slight to severe (Fig. 3(f, g)).

Compared with the method used in another study (Xiao et al. 2014) to calculate the flood/drought index, this method can be taken as the normalization style of the former to a certain extent. Considering that in the NCP, drought is a dangerous factor for food security and social order (Xiao et al. 2011), in this study, different standard levels are used to identify drought (a drought is defined as Pd ≥ 0.5 in a single year) and flood (Pf ≥ 0.6). A total of 10 drought events in Pd series (Fig. 3(f)) and 10 flood events in Pf series were recorded during 1790–1911 (Fig. 3(g)). The sum of Pd and Pf in a certain year is used to describe the total disaster intensity for that year.

Historical proxy indicator

In this study, the main data source of historical proxy indicator series is The veritable records of the Qing dynasty reprinted by the Zhonghua Book Company (1985), which is an important collection of official records (4433 volumes edited on a daily basis) in the Qing dynasty. The historical information mainly includes original records of porridge charity, public order (inside Beijing), and revolt events (in the NCP).

Porridge charity in Beijing

Porridge charity was an important component of the disaster relief system in the Qing dynasty. Some specialized institutions located in a city have generally managed it, and majority of them were organized or aided by the government (Li 1995). Regularly, they opened after autumn harvest every year until the next spring to provide simple food (mainly porridge) and camp for the poor and disaster victims from surrounding areas. Porridge charity in Beijing is directly supported by the central government at the largest scale, and with sufficient funds and efficient organization (Wang 2007). Records of porridge charity in Beijing have been kept in The veritable records of the Qing dynasty, with 475 pieces in total (1644–1911) that can be divided into three categories (Xiao et al. 2014).
  1. (1)

    “Regular” records (258 pieces) referred to the records of regular execution of porridge charity every year, such as the opening and closing of porridge stations on schedule,1 and a standard amount of grain and funds distributed to every station.2 The quantity and content of these records are relatively fixed in a certain period, which can reflect the normal implementation of porridge relief and official estimation of the population scale and residence time of the refugees in Beijing.

  2. (2)

    “Temporary” records (190) appeared in the cases of emergency, such as large numbers of refugees arriving in Beijing when some severe natural disasters occur, when the government flees from the convention and provides extra relief, such as extending the time limit,3 setting up provisional charity stations,4 and distributing additional grain and funds.5 This kind of records varies significantly among different years, and the quantity and content in a year usually reflected the seriousness of the refugee problem.

  3. (3)

    Other records (27) are relevant documents of porridge charity activities.

By counting the frequency (pieces/year) of temporary (Fig. 3(a)), regular (Fig. 3(b)), and summary (Figs. 2(b) and 3(e)) records, we can quantitatively analyze in a time series the policy of porridge charity as a governmental response to the migration of refugees affected by flood/drought.
Fig. 2

Frequency series of the records on public order (a), porridge charity (b) in Beijing, and revolt events in NCP (c) during 1644–1911 (in 10-year resolution)

Fig. 3

Multi-stage change of interaction among floods/droughts, policy management, and refugee behavior during1790–1911 in the NCP

Public order in Beijing

The influx of refugees has become a factor that leads to social instability in Beijing, especially when they do not have access to food relief and proper resettlement. Consequently, some refugees may resort to violence to survive, which will result in deterioration of security and public order in Beijing.

According to The veritable records of the Qing dynasty, a total of 567 pieces of record on public order occurred in Beijing during 1644–1911. The summary descriptions of the security situation show that majority of the records tackle actual cases, such as security accidents (crowds assembled for brawls and group gambling),6 strikes or petitions,7 criminal cases (theft, robbery, and murder),8 and rebellion,9 many of which are inextricably linked to immigrant refugees.

In this study, decadal and annual variations in frequency (pieces/year) of public order records have been identified to describe the tendency of public order in Beijing (Figs. 2(a) and 3(d)).

Revolt events in North China Plain

Revolt (including mass demonstrations, banditry incidents, and peasant uprisings) is the most extreme response to natural disasters, and refugees might participate in them due to desperation. The records of revolt events in the NCP (outside Beijing) have been kept in The veritable records of the Qing dynasty, including the location of outbreaks, affected regions, duration, and seriousness of the revolts. The frequency of revolt events in the NCP during 1644–1911 is calculated in units of “county-time” in which one county-time signifies that one county was affected by a revolt in 1 year (Xiao et al. 2011). As records on the Boxers Movement during 1899–1901 are incomplete because of war, the frequency of revolt events from 1899 to 1901 is counted based on Brief records on Boxer Movement (Li 1986).

Results and analysis

According to the decadal frequency series of records on public order (Fig. 2(a)), porridge charity (Fig. 2(b)) in Beijing, and revolt events in the NCP (Fig. 2(c)) during the Qing dynasty, by the late eighteenth century (1780s), low frequency and variability of all the three proxy series have lasted for over 140 years, and their interaction is insignificant. In the prosperous period of the Qing dynasty, with a healthy financial state, the government can efficiently control the scale of displacement in a disaster and the consequent migration in the NCP. Compared with the tax cut and “great relief” (direct and widespread allocation of food and funds to victims in disaster areas), the porridge charity in a city is not an important measure because majority of the victims decide to stay where they have received relief. Only a limited number of victims have decided to migrate in a relatively peaceful manner.

This paper shows that the beginning of the late Qing dynasty in the 1790s is the period when the frequency of records on porridge charity has increased significantly. The annual frequency series of records on public order, porridge charity in Beijing, and revolt events in the NCP are illustrated in Fig. 3 and compared with flood/drought events in the corresponding period.

Variation of proxy indicators

Floods/droughts in North China Plain (Fig. 3(f, g))

A total of eight flood/drought events occurred during 1871–1900, the highest-frequency period in the Qing dynasty. Subsequently, 1833–1846 was a peaceful period without flood/drought events. However, a severe drought occurred in 1877 (highest Pd) and a flood in 1890. Generally, increased pressure caused by natural disasters on society occurred in the NCP during the late nineteenth century more than in earlier periods.

Records on porridge charity

A low frequency of summarized records, which is 2.3 items/year occurred in 1790–1855, and after 1855, it sharply reached 5.3 (Fig. 3(e)). The frequency of regular records has remained in low variability from the 1790s to the early 1870s (varying in the range of 0–2 items in most years), which reflects a stable relief policy. However, the frequency has continued to increase from the late 1870s to the beginning of the twentieth century, with an average of 4.0 in 1878–1907 and up to the peak of 9 in 1897 (Fig. 3(b)), which means that the central government has further focused on the porridge charity. The frequency of temporary records, because of its close relationship with the intensity of natural disaster, has higher amplitude of variation (Fig. 3(a)) than regular ones with the peak of 13 in 1878. It sharply increased since the mid-1850s, with an average of 3.1 in 1856–1895, while the average for 1790–1855 was 0.6.

Records on public order in Beijing

The records have shown a remarkably fluctuant rising trend of frequency (Fig. 3(d)), especially after the 1870s (the average frequency in 1875–1900 is 13.9, while the average of the Qing dynasty was 2.1), and it has peaked in the mid-1890s (45 items in 1895).

Revolt events in the North China Plain

During the 1790s–1860s, the frequency of revolt events increased gradually and remained at a relatively low value in the 1870s–1890s (with an average of 3.5 county-times in 1870–1898), while two peak points at around 1860 and 1900 occurred, representing two large-scale armed revolts, respectively.

Multi-stage change of interaction among disaster, policy, and refugee crisis

Stage 1: 1790–1832 (preliminary strengthening of porridge charity)

The frequency of records on porridge charity began to increase at this stage (2.52 pieces/year on average; Fig. 3(e)), characterized by the concentration of temporary records in the years (or subsequent years) of flood/drought (e.g., in 1792, 1801–1802, 1823, and 1833; Fig. 3(a, f, g)), which means that natural disasters have significantly triggered the migration of refugees from the NCP to Beijing. During 1790–1832, the Pearson correlation coefficient between disaster intensity and frequency of temporary records is 0.309 (Table. 1). For example, during the drought in 1792, the number of refugees staying in Beijing temporarily exceeded 20,000,10 which shocked the emperor because in the first half of the eighteenth century, the number never exceeded 10,000.11
Table 1

Pearson correlations between proxy indicators in different stages during the late Qing dynasty












− 0.006





− 0.174

− 0.277



− 0.246

− 0.167

− 0.074







− 0.101







C1, Pearson correlation coefficient between disaster intensity (sum of Pd and Pf each year) and frequency of temporary records on porridge charity; C2, between disaster intensity and frequency of records on public order; C3, between disaster intensity and frequency of revolt events; C4, between frequency of records on porridge charity and public order; C5, between frequency of records on porridge charity and revolt events

*means correlation is significant at the 0.05 level; **means correlation is significant at the 0.01 level

As an increasing number of indigent peasants lost their livelihood and were displaced in a flood/drought disaster event, the importance of porridge charity has been emphasized in the famine relief system. Given the emerging fiscal crisis, the government continuously faced difficulties to afford “great relief” in disaster areas, such as during the early Qing dynasty. Thus, porridge charity was given additional attention for its economy and timeliness. The government wanted to improve the refugee settlement capacity of Beijing by strengthening porridge charity to ease the pressure of disaster relief in surrounding areas.

At this stage, the emergency measure effectively controlled the risk of large-scale violent conflict in the NCP because many refugees resettled in Beijing. The frequency of revolt events in the NCP was maintained at a relatively low level (1.6 county-times on average; Fig. 3(c)) except in 1813, when a premeditated peasant uprising erupted in the background of drought. Meanwhile, public order in Beijing was controlled (Fig. 3(d)) because additional relief had been supplied to the refugees.

Stage 2: 1833–1855 (transitory weakening of porridge charity)

A trend of reduction of relief pressure and social conflicts occurred in the NCP at this stage because flood/drought occurred less frequently (and without any extreme flood/drought in 1833–1846). The porridge charity transitorily weakened during this period, as reflected by fewer regular and temporary records (Fig. 3(a, b)). The frequencies of records on public order and revolt events also decreased around 1840 (Fig. 3(c, d)).

However, in this short peaceful period, the government seemed to lose its vigilance against disasters and refugees. When the drought of 1847 and flood of 1853 occurred, the disaster relief, including porridge charity, was severely inadequate (only two items of temporary records could be found in and after the two disasters). Consequently, the refugees who did not receive enough relief and proper resettlement became even more violent; the revolt events in the NCP sharply increased around 1850 (Fig. 3(c)) and public order in Beijing was also seriously undermined (Fig. 3(d)), which was visible as the turning point to a social upheaval.

Stage 3: 1856–1875 (re-strengthening of porridge charity)

This stage began with unprecedented upheaval. In the early 1860s, as a result of chronic accumulation of social contradictions, a series of armed uprisings erupted in the NCP (Fig. 3(c)). Under the dual pressures of disaster and disorder, numerous refugees swarmed into Beijing, which immediately affected the public order in the city (Fig. 3(d)).

The central government had no choice but to strengthen its porridge charity policy (by extending its duration, and adding grains and funds) to feed refugees. The temporary records occurred year by year and showed an apparent increase in frequency (Fig. 3(a)) compared with stage 1, where they are concentrated in disaster years. After the peak of social upheaval in the 1860s, large-scale revolt events in the NCP finally calmed down around 1870 (Fig. 3(c)) and many of the refugees were able to return home. Meanwhile, benefiting from the strengthening of porridge charity, the public order in Beijing also improved by the early 1870s (Fig. 3(d)).

Stage 4: 1876–1900 (continuous strengthening of porridge charity)

Induced by drought in 1877–1878, which was the most severe during the late Qing dynasty, temporary records peaked in 1878, suggesting an extraordinary number of refugees in Beijing (Fig. 3(a)). Thereafter, the remarkable increase of regular records (Fig. 3(b)) revealed a significant reformation of the famine relief policy. With the stations expanding and regulations improving successively, the porridge charity gradually became an important measure to relieve disaster victims instead of its subordinate position in the previous famine relief system. Nevertheless, during and after the disaster years, the routine porridge charity could not meet the food and relocation needs of the refugees, thereby forcing the government to supply extra food and funds. During 1876–1900, the frequency of temporary records remained high (2.72 pieces/year) and its correlation coefficient with disaster intensity was 0.451 (Table 1). This condition implies an increasing pressure from refugees especially after a disaster, which meant a large amount and long residence time.

At this stage, the interaction between the government and refugees became increasingly negative. Catering to a crowd of refugees, the central government strengthened the porridge charity to expand the emergency capacity of Beijing, which had attracted additional refugees from far and near the NCP because they did not receive any assistance from their local governments. However, many of these refugees could not be properly resettled and became a significant threat to the social order. The records on public order in Beijing increased to 13.9 items/year (Fig. 3(d)), tripling the average for 1790–1911, which suggests that the scale of refugees in Beijing had become uncontrollable and the situation had become increasingly violent.

At the cost of deterioration of public order in Beijing, the surrounding districts were relatively peaceful in the 1870s–1890s without large-scale revolts (Fig. 3(c)). However, after the social disorder in Beijing peaked in the mid-1890s, the negative interaction would be inevitably ended by the transition in refugee behavior. Against the background of severe drought, a large-scale armed uprising (the Boxers Movement) erupted again in the NCP in 1900.

Stage 5: 1901–1911 (the end of interaction)

After the uprising, social order in the NCP (both inside and outside Beijing) apparently improved (Fig. 3(c, d)), and records on porridge charity decreased (Fig. 3(e)), which had never happened in the last 100 years. On the one hand, no severe flood/drought had happened during this period (Fig. 3(f, g)); on the other hand, the central government dramatically adjusted its migration policy at the beginning of the twentieth century and formally terminated the quarantine policy for Manchuria and Mongolia, which had been implemented since the early Qing dynasty to keep the Hans out of traditional grazing fields of the Manchus and Mongols (Ye et al. 2012; Xiao et al. 2013). In the next several decades, millions of residents moved out of the NCP spontaneously or in an organized manner, and then resettled outside the Great Wall (Cao 1997). By then, the interaction between the government and refugees induced by flood/drought during the late Qing dynasty had finally ended.


Triggering mechanism of refugee migration

According to earlier research, in the late Qing dynasty (nineteenth century), large-scale refugee migration had become a social evil and caused serious social disorder in the NCP (Xiao et al. 2013, 2014). Considering that the climatic condition for agricultural production during this period had been much worse than in the eighteenth century (colder, extreme floods and droughts) (Ge 2011), the climate deterioration is speculated to be an important triggering factor for the refugee migration, especially the interregional migration from NCP to Manchuria and Inner Mongolia (Ye et al. 2012; Xiao et al. 2013).

In this paper, special attention was paid to short-term and short-distance migration from surrounding areas to Beijing, which also involved a large-scale population and lasted for over 100 years. According to Pearson correlation analysis, a significant positive correlation can be observed between disaster intensity (described with the sum of Pd and Pf) and the porridge charity (frequency of temporary records), public order (frequency of records), and revolt events (Table 1) during 1790–1911. In majority of the disaster events identified with Pd and Pf, remarkable migration from the disaster area to Beijing could be inferred with the increase of temporary records on the porridge charity (Fig. 3). The result could be a convincing evidence for the preceding inference that flood/drought had been an important triggering factor.

However, bad climate is not the only trigger. This close correlation has emerged on the premise of an intensified contradiction between humans and land. During the Qing dynasty, an unprecedented population growth occurred in NCP, which led to a continuous drop in the grain output per capita. By the beginning of the nineteenth century, grain output per capita in the NCP had decreased below 300 kg per capita (Fang et al. 2012), which is the modern subsistence level and is considered as a critical share of food security (Yin et al. 2009). Food insecurity has contributed to a sharp increase in the social sensitivity to extreme floods or droughts. Therefore, the refugee problem emerged in the late eighteenth century, and the drought in 1792 triggered an unprecedented refugee migration into Beijing.

Although the Pearson correlation coefficient between disaster intensity and frequency of the porridge charity is significantly positive in the entire period of 1790–1911, the importance of flood/drought seems to vary in different stages. Among the five stages, only stages 1 (1790–1832) and 4 (1876–1900) show a significant positive correlation (p < 0.05), and in the other three stages, the correlations are insignificant. In stage 2 (1833–1855), porridge charity’s weakening might limit the willingness of the refugees to migrate to Beijing (e.g., in the drought of 1847 and flood of 1853); in stage 3 (1856–1876), social upheaval is another trigger of refugee (e.g., in 1861 and 1867); and in stage 5 (1901–1911), Beijing is not the first destination of refugees because of the migration policy.

Thus, the triggering mechanism of refugee migration in the NCP during 1790–1911 is complex. The long-term food shortage and fiscal crisis left the regional society with a high risk due to natural disasters. Thus, flood/drought has become an important, but not the only trigger, for the movement of refugees. This historical case can support the criticism of the term “climate refugee” from many researchers because it simplifies the mechanism of refugee migration (McAdam 2011; Piguet 2013).

Policy responses in different periods

As discussed, in most of the late Qing dynasty, the refugee problem in the NCP increasingly became serious. To control risk, the central government adjusted its policy twice in the 1860s and 1900s after the two largest peasant uprisings.

The first adjustment focused on the famine relief. In the first half of the nineteenth century, the traditional relief system was still operating but in an inefficient way. Although the porridge charity in Beijing had been strengthened in the disaster years, it remained a supplementary measure to great relief. Particularly when the Taiping and Nien Rebellions erupted in the lower Yangtze and Yellow River Basin in the 1850s, the Grand Canal where grain and tax are transported from East and Central China to Beijing had been cut off (Fairbank 1978), and the grain storage in Beijing sharply dropped to one tenth in the early eighteenth century (Xiao et al. 2014). As the great relief was unsustainable, the victims could not be accommodated in their homeland. Faced with a high risk of social unrest caused by refugee movement, the government decided to strengthen the porridge charity in Beijing to provide temporary shelters to the refugees in winter, and hoped that they would return home in the next spring.

The second adjustment focused on the migration policy. In the last several decades of the nineteenth century, an increasing number of peasants became bankrupt due to extreme disaster and left their homeland permanently. These refugees tended to basically stay in Beijing because they had no home. When the porridge charity was overwhelmed by the endless demand of refugees, the social order both inside and outside Beijing finally collapsed. The result of Pearson correlation analysis shows a significant positive correlation between the frequency of records on porridge charity and public order, but an insignificant correlation between porridge charity and revolt events during 1790–1911 (Table 1), suggesting that in the long term, the strengthening of porridge charity will not help reduce the social unrest in the NCP, but will only worsen the public order in Beijing. After the Boxers Movement, the central government realized that the essential measure to resettle the refugees was to provide new jobs, specifically, arable lands. The quarantine policy was abolished and the vast northern frontier region was opened to the Han immigrants.

Although the term “climate refugee” is not scientifically accurate, the large-scale refugee migration as a direct result of climate change impacts has attracted worldwide attention. The importance of governance mechanisms for refugee resettlement is repeatedly emphasized (Mayer 2014). In this historical case, the interaction (especially negative interaction) between refugees and the government, and the subsequent policy adjustments, can provide a helpful reference for policy makers today.


The main conclusions can be summarized as follows:

The problem of refugee migration in the NCP significantly emerged in the 1790s, when the frequency of records on porridge charity in Beijing increased sharply. During 1790–1911, floods/droughts have become an important triggering factor of refugee migration, as reflected by the positive correlation between disaster intensity and porridge charity, public order, and revolt events. This close correlation has emerged on the premise of regional man-land contradiction and fiscal crisis that erupted at the turn of the nineteenth century.

In the nineteenth century, the main policy response of the central government to a crowd of refugees focused on famine relief. The porridge charity in Beijing, once a supplementary measure to great relief in the early Qing dynasty, was successively strengthened to expand the refugee management capacity, especially after the 1850s. However, the lack of regular famine relief has led to a severe negative interaction between the government and refugees throughout the late nineteenth century. In the long term, the strengthening of porridge charity will not help reduce the social unrest in the NCP, but will only worsen the public order in Beijing.

After the Boxers Movement in 1900, the negative interaction is finally ended with the adjustment of migration policy. The central government formally terminated the quarantine policy for Manchuria and Mongolia, and the refugees from the NCP can obtain a piece of arable land and resettle outside the Great Wall. Beijing has not been taken as the first destination any longer.

With the dramatic climate change and increasing extreme weather events in the future, the vulnerability and adaptability of human society will increasingly attract concern. Compared with the cases that have occurred in recent years, the historical case discussed in this paper enables us to analyze the occurrence mechanism of refugee migration (flood/drought as a triggering factor, and its relationship with other factors) and the development of refugee behavior (from peaceful to violent, and their change of destination) in a long time series (over 100 years). The results will enhance our understanding of the interactions between climate change and social vulnerability. The policy responses (especially relief and migration policy) and their interaction with the behavior of refugees during the late Qing dynasty will also provide valuable references for policymakers that will face similar occurrences in the future.


  1. 1.

    e.g., “the emperor orders the porridge stations in Beijing to open in this winter and next spring”. The veritable records of the Qing dynasty, vol. 28, pp. 388

  2. 2.

    e.g., “300 dan of millet was allocated to the Pujitang outside the Guang’an Gate”. The veritable records of the Qing dynasty, vol. 35, pp. 720

  3. 3.

    e.g., “close of the porridge charity in Beijing is delayed for a month because of spring drought”. The veritable records of the Qing dynasty, vol. 28, pp. 847

  4. 4.

    e.g., “a provisional porridge station has been set up in Dajing outside the Guangning Gate”. The veritable records of the Qing dynasty, vol. 29, pp. 264

  5. 5.

    e.g., “every porridge station in Beijing will get an extra dan of millet each day”. The veritable records of the Qing dynasty, vol. 43, pp. 671

  6. 6.

    e.g., “there was a fight with weapons between groups of people around Kwan-yin temple outside Zhengyang Gate on 3rd Feb”. The veritable records of the Qing dynasty, vol. 57, pp. 41

  7. 7.

    e.g., “in recent years, craftsmen and workers often gathered for raising salary, even struck when their requirement was not met”. The veritable records of the Qing dynasty, vol. 56, pp. 324

  8. 8.

    e.g., “disaster victims from surrounding area caused robbery to happen much more often in Beijing”. The veritable records of the Qing dynasty, vol. 56, pp. 298

  9. 9.

    e.g., “there were several rebels breaking into Cangzhen Gate of the Forbidden City and captured by eunuchs”. The veritable records of the Qing dynasty, vol. 31, pp. 719

  10. 10.

    The veritable records of the Qing dynasty, vol. 26, pp. 919

  11. 11.

    The veritable records of the Qing dynasty, vol. 11, pp. 631, 646


Funding information

Supported by National Natural Science Foundation of China (Grant No. 41501207, 41771572, 41701219).


  1. Afifi T (2011) Economic or environmental migration? The push factors in Niger. Int Migr 49:e95–e124. CrossRefGoogle Scholar
  2. Adger WN, Pulhin JM, Barnett J, Dabelko GD, Hovelsrud GK, Levy M, Oswald Spring Ú, Vogel CH, (2014) Human security. In: Field CB, Barros VR, Dokken DJ, Mach KJ, Mastrandrea MD, Bilir TE, Chatterjee M, Ebi KL, Estrada YO, Genova RC, Girma B, Kissel ES, Levy AN, MacCracken S, Mastrandrea PR, White LL (eds) Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, New York pp 755–791Google Scholar
  3. Büntgen U, Tegel W, Nicolussi K, McCormick M, Frank D, Trouet V, Kaplan JO, Herzig F, Heussner K-U, Wanner H, Luterbacher J, Esper J (2011) 2500 years of European climate variability and human susceptibility. Science 331:578–582. CrossRefGoogle Scholar
  4. Cao SJ (1997) Chinese migration history (Vol. 6: Qing Dynasty and Republic of China). Fujian People’s Publishing House, Fuzhou.Google Scholar
  5. Chen YQ (1989) A study of hazard indexes of drought and flood. J Catastrophology 12:10–13. CrossRefGoogle Scholar
  6. Chi ZH, Li HY (2007) Famine and refugees in Zhili at the turn of 19th and 20th century. In: Li WH, Xia MF (eds) Famine and society in Qing Dynasty. SDX Joint Publishing Company, Beijing, pp 82–96Google Scholar
  7. China Meteorological Administration (1981) Yearly charts of dryness/wetness in China for the last 500-year period. SinoMaps Press, BeijingGoogle Scholar
  8. Fairbank JK (1978) The Cambridge History of China: Vol. 10 Late Ch’ing, 1800–1911, Part I. Cambridge University Press, Cambridge, UK and New York, NY, USAGoogle Scholar
  9. Fang XQ, Xiao LB, Wei ZD (2012) Social impacts of the climatic shift around the turn of the nineteenth century on the North China Plain. Sci China (Earth Sci) 56:1044–1058. CrossRefGoogle Scholar
  10. Fang XQ, Su Y, Yin J, Teng JC (2015) Transmission of climate change impacts from temperature change to grain harvests, famines and peasant uprisings in the historical China. Sci China Earth Sci 58:1427–1439. CrossRefGoogle Scholar
  11. Ge QS (2011) Climate change in Chinese dynasties. Science Press, BeijingGoogle Scholar
  12. GECAFS (2006) A Set of Prototype Caribbean Scenarios for Research on Global Environmental Change and Regional Food Systems. GECAFS Report No.2, WallingfordGoogle Scholar
  13. Haug GH, Günther D, Peterson LC, Sigman DM, Hughen KA, Aeschlimann B (2003) Climate and the collapse of Maya civilization. Science 299:1731–1735. CrossRefGoogle Scholar
  14. Hsu KJ (1998) Sun, climate, hunger, and mass migration. Sci China Ser D 41:449–472. CrossRefGoogle Scholar
  15. Lal PN, Mitchell T, Aldunce P, Auld H, Mechler R, Miyan A, Romano LE, Zakaria S (2012) National systems for managing the risks from climate extremes and disasters. In: Field, CB, Barros V, Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner G-K, Allen SK, Tignor M, and Midgley PM (eds) Managing the risks of extreme events and disasters to advance climate change adaptation. A special report of working groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, New York, pp 339–392Google Scholar
  16. Lee HF, Zhang DD (2010) Changes in climate and secular population cycles in China, 1000 CE to 1911. Clim Res 42:235–246. CrossRefGoogle Scholar
  17. Lee HF, Zhang DD (2013) A tale of two population crises in recent Chinese history. Clim Chang 116:285–308. CrossRefGoogle Scholar
  18. Li KR (1990) Droughts and floods in the North China Plain. Science Press, BeijingGoogle Scholar
  19. Li LM (2007) Fighting famine in North China: state, market, and environmental decline, 1690s–1990s. Stanford University Press, StanfordGoogle Scholar
  20. Li WH (1986) Brief records on boxer movement. Qilu Press, JinanGoogle Scholar
  21. Li XJ (1995) Research on famine policy in the Qing dynasty. China Agricultural Press, BeijingGoogle Scholar
  22. McAdam J (2011) Refusing refuge in the Pacific: deconstructing climate-induced displacement in international law. In: Piguet É, Pécoud A, Guchteneire P (eds) Migration and climate change. United Nations Educational, Scientific and Cultural Organization (UNESCO), Cambridge University Press, Cambridge, UK and New York, NY, USA, pp 102–137Google Scholar
  23. Mayer B (2014) Climate migration governance. Handbook Climate Change Adaptation 1:1–14. CrossRefGoogle Scholar
  24. Myers CA, Slack T, Singelmann J (2008) Social vulnerability and migration in the wake of disaster: the case of Hurricanes Katrina and Rita. Population Environ 29:271–291. CrossRefGoogle Scholar
  25. Niu PH (1990) Comprehensive table of administrative region history in the Qing dynasty. SinoMaps Press, BeijingGoogle Scholar
  26. Piguet É (2013) From “primitive migration” to “climate refugees”—the curious fate of the natural environment in migration studies. Ann Assoc Am Geogr 103:148–162. CrossRefGoogle Scholar
  27. Tan QX (1987) The historical atlas of China (Vol. 8: Qing dynasty). SinoMaps Press, BeijingGoogle Scholar
  28. Wang L (2007) Summary of porridge charity stations in the Qing dynasty. Theory J 4:111–115. CrossRefGoogle Scholar
  29. Wei ZD, Fang XQ, Su Y (2014) Climate change and fiscal balance in China over the past two millennia. The Holocene 24:1771–1784. CrossRefGoogle Scholar
  30. Wei ZD, Fang XQ, Su Y (2015) A preliminary analysis of economic fluctuations and climate changes in China from BC 220 to AD 1910. Reg Environ Chang 15:1773–1785. CrossRefGoogle Scholar
  31. Will PE (1990) Bureaucracy and famine in eighteenth-century China. Stanford University Press, StanfordGoogle Scholar
  32. Xiao LB, Ye Y, Wei BY (2011) Revolts frequency during 1644–1911 in North China Plain and its relationship with climate. Adv Clim Change Res 2:218–224. CrossRefGoogle Scholar
  33. Xiao LB, Fang XQ, Ye Y (2013) Reclamation and revolt: social responses in Eastern Inner Mongolia to flood/drought-induced refugees from the North China Plain 1644-1911. J Arid Environ 88:9–16. CrossRefGoogle Scholar
  34. Xiao LB, Fang XQ, Zhang YJ, Ye Y, Huang H (2014) Multi-stage evolution of social response to flood/drought in the North China Plain during 1644–1911. Reg Environ Chang 14:583–595. CrossRefGoogle Scholar
  35. Ye Y, Fang XQ, Khan MAU (2012) Migration and reclamation in Northeast China in response to climatic disasters in North China over the past 300 years. Reg Environ Chang 12:193–206. CrossRefGoogle Scholar
  36. Yin J, Su Y, Fang XQ (2015) Relationships between temperature change and grain harvest fluctuations in China from 210 BC to 1910 AD. Quat Int 355:153–163. CrossRefGoogle Scholar
  37. Yin J, Su Y, Fang XQ (2016) Climate change and social vicissitudes in China over the past two millennia. Quat Res 86:133–143. CrossRefGoogle Scholar
  38. Yin PH, Fang XQ, Yun YR (2009) Regional differences of vulnerability of food security in China. J Geogr Sci 19:532–544. CrossRefGoogle Scholar
  39. Zhang DD, Zhang J, Lee HF, He YQ (2007) Climate change and war frequency in eastern China over the last millennium. Hum Ecol 35:403–414. CrossRefGoogle Scholar
  40. Zheng JY, Xiao LB, Fang XQ, Hao ZX, Ge QS, Li BB (2014) How climate change impacted the collapse of the Ming dynasty. Clim Chang 127:169–182. CrossRefGoogle Scholar
  41. Zhonghua Book Company (1985) The veritable records of the Qing dynasty. Zhonghua Book Company, BeijingGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.The Institute of Qing HistoryRenmin University of ChinaBeijingChina
  2. 2.School of Geography, Faculty of Geographical ScienceBeijing Normal UniversityBeijingChina

Personalised recommendations