**3. Results**

#### *3.1. Regional Interactive Response and Characteristics*

#### 3.1.1. Spatial Characteristics of Famine-Related Migrations

Historically, Chinese farmers have had a tied connection to their homeland and farmland. Large-scale population migration is temporary and spontaneous in the period of extreme climate events. The migration is often driven by famine, plague and other events triggered by major meteorological disasters [6]. The transition from being starving to homeless is the result of the local social system losing its ability to adapt and the failure of individual survival strategies.

Famine victims tended to migrate from the hardest-hit areas to the nearest slightly-impacted areas, and non-disaster areas (Figure 4). In 1877, migrants left from Shandong and Anhui to the south of Jiangsu. Some people left from Shandong to Liaoning, while others moved from Shanxi, Henan, and Hebei to Beijing. The victims migrated between 130 and 766 km (straight line distance, the same below), with an average migration distance of approximately 429 km. In 1878, famished people in Henan moved to Beijing and Anhui. Some people in Shandong, Hebei, and Shaanxi moved to Beijing, and others in Shanxi moved to Hebei, Beijing and Anhui, as well as to the Daqingguan and Tongguan in Shaanxi. Famine victims migrated between 130 and 1080 km, with an average migration distance of approximately 460 km.

In this historical period, limited by poor tra ffic conditions and the physical weakness, the spontaneous famine-related migration flows just spread from the disaster areas to the nearby areas. In 1877–1878, the harvest rate of grains in Anhui was only about 50%, while in Jiangxi, Hubei, and Hunan, which are farther away from the disaster areas, the harvest rate reached over 60% or 70% (Table 3). Although with relatively low harvest rate, Anhui was still one of the major destinations in the famine-related migration due to its proximity to the disaster areas. Meanwhile, the migration directions of the famished people appeared to be relatively fixed and stable, indicating the influences of regional politics, dissemination of the relief information, and customs and traditions [47].


**Table 3.** Harvest rate of grains in di fferent provinces [48].

There was a visible di fference in the economic conditions between towns and villages in China during the Qing dynasties. Under the circumstance of the rural economic decline, cities and towns became the destinations of the migrants. Soup kitchens and shelters there o ffered more chances of survival. For example, the capital city at that time, Beijing, is close to the a ffected provinces, and it had become a popular destination for the starving migrants. Besides, some other places attracted migrants due to their advantageous geographical location and sophisticated tra ffic system. For instance, in 1878, the whole area of Shaanxi Province su ffered a severe drought, but its two towns, Tongguan and Daqingguan, had experienced the entry of famished people from the Shanxi Province on the east. Lying close to the shared borders of Shaanxi, Shanxi and Henan provinces, these two towns were distribution centers of relief goods and had been of strategic importance since ancient times.

#### 3.1.2. Spatial Characteristics of Money and Grain Allocations

In China, a state unified by centralized political power, the allocation of money and grain is a state administrative action. It is normal that the central governmen<sup>t</sup> provides emergency financial assistance and food aid to disaster areas. The relief silver and grain are mainly from state banks and granaries, which are supplied by every locality on a regular basis. Besides, some resources in the non-disaster areas could be requisitioned as emergency relief by the central governmen<sup>t</sup> for the duration of the famine.

In 1877, a total of 17 provinces were involved in the allocation of money and grain (Figure 5). The main drought-stricken areas, Shanxi, Henan, Shandong, Shaanxi, and Hebei, received the most amount of silver and grain. From the perspective of spatial pattern, the allocation is featured with a core-ring structure: (1) Shanxi, lying in the core, received the most amount of silver and grain; (2) Shandong, Henan, and Hebei, lying on the second layer, received the relief from other provinces and also supplied resources for Shanxi; (3) Hunan, Anhui, Jiangxi, Jiangsu and other provinces in the Middle–Lower Yangtze Plain, lying on the most peripheral ring, experienced mainly the outflow of silver and grain. In 1877, the straight-line distances of silver and grain allocations were between 270 and 1635 km, and the average transfer distance was approximately 800 km.

In 1878, also 17 provinces were involved in the relief allocations, but the spatial structure appeared to be a complex network (Figure 5). The amounts of the relief grain being allocated to Shanxi, Henan and Hebei accounted respectively for 34%, 39% and 27% of the total number of grain transfers. The distribution appeared to be more balanced than that in 1877. Shanxi, Shaanxi, Henan, and Hebei were the main receivers of the relief silver, accounting for 96% of the total silver transfers. Zhejiang, Hubei, and Hunan in the Yangtze River Basin remained as the suppliers. Apart from them, Fujian, Guangdong and Sichuan in the farther south also became the main relief suppliers. In 1878, the straight-line distances of silver and grain allocations were 132–1635 km, with an average of approximately 860 km.

Unlike the spontaneous migrations of famine victims, the allocation of money and grain is governmen<sup>t</sup> action. It was at a larger spatial scale and with more frequent transfers. Moreover, the spatial pattern of the money and grain allocations was relatively more complex, as it would vary according to the severity of the disasters and national relief policies.

3.1.3. Temporal Characteristics of Famine-Related Migrations and Money and Grain Allocations

The records of the money and grain allocations mainly came from o fficial documents, so the time of the documents was the time of the allocations. While the records of famine-related migrations might only appear when there were many migrations, they can still roughly indicate the period of mass famine-related migration.

According to the records that contain clear time information of the events in 1877–1878 (Figure 6), the relief allocations and famine-related migrations appeared to be seasonal and temporary, and shared with a similar peak period, which was from October 1877 to May 1878. However, the duration of money and grain allocations was longer than that of famine-related migrations.

**Figure 6.** The number of the records of famine-related migrations and relief allocations in 1877–1878.

In North China, during the Qing dynasty, the summer harvest season started from May to June, and the autumn harvest season was from September to October. For example, in Shandong Province, wheat was harvested in the fifth lunar month (June in the Gregorian calendar), while sorghum and millet were harvested in the eighth lunar month (September in the Gregorian calendar) [49].

A small number of records of migrations before the summer harvest in 1877 indicated the occurrence of drought, but it also suggested that the situation was not beyond control. However, a large number of records were found of the time period from October 1877 to May 1878. The autumn of 1877 experienced a severe harvest failure as the consequence of persistent extreme droughts in the previous months. In the spring and summer of 1878, as the secondary disaster, plagues began to spread in the disaster areas. These two might be the main causes of the significant rise of the number of records, which suggested that the impacts of the extreme droughts had exceeded the response capability of a single region. In this stage, regional interactions were necessary for the mitigation of disaster impacts.
