The Relationship Between Dry–Wet Change and the Manchu Rise in China

Abstract

Exploring the impact of dry–wet change on the Manchu rise has important implications for revealing the impact of climate change on ethnic dynamics. In this study, we used tree rings of Carya cathayensis and historical data to study this dynamic in Northeast Asia using function fitting and step-by-step elimination analysis. The results show a mean reconstructed scPDSI_4–10 of 0.822 from 1583 to 1644, which is 0.287 higher than the mean from 1548 to 2022 (0.535), and during 25 slightly wet years. This indicates that dry–wet change provided a favorable natural environment for the Manchu rise, under which the group’s area continued to expand and change shape in complex ways, and the population increased rapidly in the control region. However, in some years, the closer the scPDSI_4–10 was to the multi-year mean (0.774) of the deviation from the mean (0.535) of the scPDSI_4–10, the faster the control region expanded and the more the population increased. These results provide a reference for understanding the relationship between ethnic groups’ dynamics and climate change.

1. Introduction

Climate conditions can provide the foundation for different ethnic groups to form across the world [1,2,3]. The influence of dry–wet change on groups’ evolution is an important entry point for studying the mechanism of human–land relationships and is a research topic currently attracting significant interest [3,4,5,6]. In 221 BC-9 AD, China experienced a long and stable wet period that lent itself to agriculture and brought together the Qin and Han dynasties, which was important for the creation of the Han nationality [7]. The population sizes, living regions, and even customs of ethnic groups constantly change due to dry–wet change. Not only does dry–wet change influence ethnic groups in Northeast Asia, but this phenomenon also influences groups in the northwest of China. For example, nomadic ethnic groups gradually became the most populous by means of their habitual migration to wherever water and grass were available. This allowed them to better adapt to the expansion of grassland caused by increased precipitation in the Great Dunhuang region of China [8]. However, the intensification of dry–wet change leads to environmental degradation and food shortages. This sharpens internal fighting among different groups over reduced resources, resulting in a risk of extinction for some ethnic groups through population decrease [9]. With grassland expanding with increases in precipitation, northern nomadic tribes waged wars and invaded the Hexi Corridor of China to gain more grassland for further growth [10]. The excessively humid climate continued to intensify, and the Yi ethnic group adapted their settlement style to become more open in order to receive more sunlight [11]. Ethnic groups constantly change their habits in accordance with dry–wet change, but they also have the ability to maintain their original customs. Indigenous people in North America took advantage of wet conditions to engage in planting, but their hunting traditions were preserved in the postglacial period [12].

Different ethnic groups have varying sensitivities to temperature changes; Maori, Pacific, and Asian peoples have greater sensitivity compared to European groups [13]. Due to variations in economic strength, different ethnic groups have distinct abilities to cope with precipitation change in the coastal areas of Bangladesh. As precipitation variability rises, ethnic differences increase [14]. The resilience of the Taqi people’s livelihood to rainfall changes is better than that of the Paco in Central Vietnam [15]. Heat vulnerability is gradually increasing among different ethnic groups in the United States under the influence of global warming, and the non-Hispanic African American or Black group has the weakest tolerance for high temperatures, followed by the Hispanic or Latino group [16]. Previous studies have addressed the impact of climate change on the formation and evolution of ethnic groups; however, due to the challenges of accurately restructuring climate data and the dominance of qualitative descriptions of human data, quantitative analyses remain scarce, and the relationship between changes in ethnic groups and dry–wet change is not accurately described. Therefore, in the present work, we quantitatively studied the relationship between the Manchu rise and dry–wet change across Northeast Asia in Chinese history.

However, tree-ring technology is widely used in climate change research as it allows for the reconstruction of long-term climate data year by year. For instance, tree rings of fir (Abies pindrow) were used to reconstruct the Palmer Drought Severity Index (PDSI) (1820–1981) [17], summer temperatures (1370–2020) were obtained from tree rings of blue pine (Pinus wallichiana) [18], and annual precipitation was calculated from mixed forests [19]. Dry–wet change in North China over 480 years has previously been studied using tree rings of Carya cathayensis [20], suggesting that they may be utilized to restore the Self-calibrating Palmer Drought Severity Index (scPDSI) of the Manchu rise in Northeast Asia, providing accurate annual climate data for studying dry–wet change. In Northeast China, tree rings have also been applied to restore climate data for historical periods. May temperatures in Northeast China since 1797 were reconstructed using 51 sample tree-ring cores from 24 Pinus sylvestris var.mongolica. There were 31 warm and 43 cold years in the 224-year reconstructed temperature series, accounting for 13.8% and 19.2% of the total, respectively [21]. The precipitation from April to July in the Greater Khingan Range in Northeast China over 242 years (1776–2017) was restored based on tree rings of Pinus sylvestris var.mongolica [22]. Extreme dry and wet events accounted for 19.9% and 18.6% of a 156-year climate record (1859–2014), respectively, based on tree rings of Chinese pine (Pinus tabulaeformis Carr) in western Liaoning Province and the southern Horqin sand land, Northeast China [23]. Temperature in the growing season (May–July) across 211 years (1803–2013) was reconstructed based on the tree-ring width chronology of Korean pine (Pinus koraiensis) from the Zhangguangcai Mountains in Northeast China [24]. However, climate data have not been restored for the late Ming and early Qing dynasties in Northeast China.

These studies have addressed the relationship between ethnic group evolution and dry–wet change and have confirmed that the formation of these groups is closely related to climate change. However, quantitative research needs to be further strengthened; accurate and, especially, year-by-year data are needed for in-depth research on the impact of dry–wet change on specific ethnic groups based on climate data from particular historical periods. The scientific question of our study was whether and how dry–wet change influenced the Manchu rise. In this context, we studied the relationship between the Manchu rise and dry–wet change based on climate data from tree rings and historical material. The Manchu moved from Northeast Asia and then unified China, establishing the Qing Dynasty, which had a significant impact on both China and the rest of the world. The contributions of this work can be summarized as follows: (1) We reconstructed the Self-calibrating Palmer Drought Severity Index from April to October (scPDSI_4–10) for each year to reveal dry–wet change during the period of the Manchu rise (1583–1644). (2) We analyzed these changes with regard to areas of control and increased population size to recreate the progression of the Manchu’s rise. (3) We assessed the impact of dry–wet change on the Manchu rise using step-by-step elimination analysis. This study aims to provide a scientific reference for studying the relationship between people and the environment.

2. Materials and Methods

2.1. Study Area

Our study area is located in the easternmost part of the Changbai Mountains in Northeast China (Figure 1). The Great Black Mountain faces the northeast plain and is a typical agricultural and forestry ecotone [25,26]. The northeast plain is dominated by agriculture and is a major grain-producing area in China [27,28]. The Great Black Mountain is located on the edge of the forest region and is sensitive to climate change [29]. The annual temperatures are between 3.46 °C and 7.09 °C, and the annual precipitation has stayed within the range from 454 mm to 916 mm since 1960 (Figure 2a). The research area is cold in winter and hot in summer, with precipitation being concentrated in summer [30] (Figure 2b). Numerous Carya cathayensis are scattered across the mountains of the study area, with little human interference and old trees, making Carya cathayensis an ideal tree species for recreating climate data over a long period of time [31]. Therefore, based on tree rings of Carya cathayensis, the agroforestry ecotone in our study area is sensitive to climate change, meaning that the history of dry–wet change in Northeast Asia can be understood on a longer time scale.

2.2. Tree-Ring Sampling

From 1 May to 24 October 2023, ninety cores of Carya cathayensis were extracted from four plots in our study area (Figure 1) (

Table 1. Information on plots.

Plot	Longitude (E)	Latitude (N)	C/T	NCC	RP (Year)
1	126.9268	44.34423	20/10	16	1417–2022
2	126.0523	44.06951	16/8	12	1443–2022
3	125.7393	43.53871	23/13	21	1419–2022
4	124.9945	43.46301	22/12	18	1457–2022

C/T, NCC, and RP represent no. cores/trees, no. cores for the chronology, and recording period of Carya cathayensis, respectively.

). We used Haglof increment borers (5.15 mm) to collect two cores per tree, offset by as close to 90° as possible and located approximately 1.2 m above ground. A total of 86 cores were obtained from 43 large trees (diameter at breast height (DBH) > 30 cm). Five cores were removed as they were damaged during sampling, while all remaining cores were dried, mounted, and sanded using increasingly finer grits, until each annual ring was clearly visible. The ring width measurements were carried out using the Velmex ring measurement system, which has a resolution of 0.001 mm [32] (Table 1).

2.3. Chronology Development

The quality of the cross-dating was assessed using the COFECHA software, yielding a correlation coefficient (R) of 0.823 [33]. Cores with low correlation to the master series (R < 0.823) were removed from the final dataset prior to chronology development. The growth chronology of the tree ring was developed based on the remaining 67 cores using the ARSTAN software [34]. Negative exponential functions were applied to standardize the data, removing age-related growth trends [35]. Finally, the residual chronology (RES) was developed to reveal dry–wet change. The composite chronology had high statistical parameters, such as a mean sensitivity of 0.93, a standard deviation of 0.03, a first principal component (PCA1) of the RES that explained 42.48% of the variance [34], a mean series intercorrelation of 0.38, and a first-order autocorrelation of 0.49 [36], indicating a common environmental influence on the radial growth of trees. The chronology extended back to 1548, as the expressed population signal (EPS) exceeded the threshold of 0.85 with a sample depth of 20 cores [37] (Figure 3).

2.4. Statistical Analysis

2.4.1. Reconstruction of scPDSI_4–10

Despite the lack of accurate numerical records of annual dry–wet change during the historical period in question (1583–1644), we used tree rings of Carya cathayensis to restore the PDSI for each year from 1583 to 1644, providing accurate climate data to help us understand the relationship between dry–wet change and the Manchu rise. This study used data from four meteorological stations: Jiutai (125.84° E, 44.15° N), Shuangyang (125.65° E, 43.55° N), Yitong (125.28° E, 43.35° N), and Shulan (126.93° E, 44.39° N) (Figure 1). The scPDSI_4–10 data (1901–2022) were based on the global terrestrial scPDSI dataset, published by the Climate Research Unit (CRU) of the University of East Anglia from 1901 to 2022. The spatial resolution was 0.5° × 0.5° [38,39]. The four tree sampling plots were the closest to the four meteorological stations, and the scPDSI_4–10 data from the CRU also covered the study area exactly. Based on mutual verification between the station and CRU data, the spatial range of the climate information precisely covered the sampling points, which could best embody the impact of climate on ring width growth and accurately reconstruct the scPDSI_4–10. The average of the scPDSI_4–10 values was selected for all grid point data in the study area, and a linear equation was built, incorporating the scPDSI_4–10 values and the RES from 1901 to 2022. The scPDSI_4–10 value was taken as the dependent variable, and the ring width index of the RES was taken as the independent variable (Equation (1)). The equation passed leave-one-out model testing; the reduction in errors (RE) was 0.342 (

Table 2. Calibration results of the leave-one-out model (common period: 1901–2022).

	r	R2	R2adj	RE	ST	PMT	DW
Calibration	0.76 **	0.576	0.569
Verification	0.76 **	0.576	0.569	0.432	60+/62− *	3.744 *	1.810

r, R², R²_adj, RE, ST, PMT, and DW represent the correlation coefficient, explained variance, adjusted explained variance, reduction in error statistic, sign test, product mean test, and Durbin–Watson test, with ** p < 0.01 and * p < 0.05, respectively.

), the product mean test (PMT) was 3.744 (p < 0.05), and the sign test (ST) was 60+/62− (p < 0.05) (Table 2). The Durbin–Watson statistical coefficient (DW) was 1.810 (Table 2). These values were all greater than the test results of the leave-one-out model of the linear equations established by Li et al., 2011 (RE = 0.241; ST = 38+/17− (p < 0.05); DW = 1.645) [40], and Li et al., 2010 (PMT = 3.35) [41]. The r, R², and R²_adj of the equation were 0.76, 0.576, and 0.569, respectively, which were greater than the results of the linear equation established using tree rings by Shah et al. (r = 0.5075; R² = 0.2576; R²_adj = 0.2414) [42]. These results show that the equation could be used to reconstruct the scPDSI_4–10 using tree rings [43]. The scPDSI_4–10 values from 1548 to 1900 were thus computed using Equation (1) based on the RES as follows:

Y = −14.618 + 15.380x (r = 0.76, R² = 0.576, R²_adj = 0.569, p < 0.01)

(1)

Here, Y represents the scPDSI_4–10 and x represents the tree-ring width index of the RES. The annual values of the scPDSI_4–10 between 1548 and 1900 were calculated using Equation (1). Annual scPDSI_4–10 records were obtained from the CRU data for 1901–2022, which were then overlaid with the scPDSI_4–10 values between 1548 and 1900. Finally, a 474-year (from 1548 to 2022) series for the scPDSI_4–10 was built by merging the two sets (Figure 4). The mean of the reconstructed scPDSI_4–10 was 0.535, and when the index was centered around zero, this indicated that the region was an ecotone between arid and humid areas. Our study area was an ecotone between humid and semi-humid areas, and the scPDSI was predicted to be around 0.5 [44], further verifying the accuracy of our recovered data. Based on the reconstructed historical data from 1548 onwards, scPDSI_4–10 values from 1583 to 1644 were extracted, and the mean and standard deviation of these were calculated in order to reveal the dry–wet change in the 62-year period of the Manchu rise. According to pollen records, the northeast region was relatively humid during this period, while other parts of China experienced severe drought [45]. From the early 17th century to the mid-17th century, low temperatures in Northeast China weakened evaporation and increased surface moisture, based on records of peatlands in this region [46]. According to historical records, there was no evident extreme drought in Northeast China during the late Ming and early Qing dynasties, and the dry–wet conditions were relatively stable [47]. These research results are highly consistent with the dry–wet conditions we restored, further verifying the reliability of the restored scPDSI_4–10 from 1583 to 1644.

2.4.2. The Area, Shape, and Population of the Manchu Region of Control from 1583 to 1644

According to Yan’s historical works [48,49], and based on the Atlas of Chinese History [50] and the changes in Chinese territory in various dynasties (https://b23.tv/a7guNYj, accessed on 18 August 2024), the annual number of people and the control region of the Manchu from 1583 (when the Manchu leaders declared an uprising against Ming rule) to 1644 (when the group entered the southern Great Wall region from the north to complete the unification of China and establish the Qing Dynasty) were recorded. The area and shape index of this region were calculated, as was the population size for each year. The area of control is an important index of Manchu strength. Generally, the larger the area of control of an ethnic group, the more resources the ethnic group may utilize, and the stronger the ethnic group will be. Area of control is an important foundation of an ethnic group. Therefore, the area of control of the Manchu was taken as a symbol of the Manchu rise. The Manchu’s rise involved not only an increase in the control area but also the selection of expansion direction. If the Manchu expanded along an axis, indicating that the group was eager to quickly occupy new territories, the shape of the region of control would become more complex. If the Manchu expanded across an area, indicating a greater intention to strengthen the control region, this shape would become more regular and expansion would be relatively slow. Therefore, different territorial shapes indicate differences in expansion directions, reflecting the urgency of Manchu growth. Since we needed to address whether the differences in expansion directions were related to dry–wet change, the shape of the territory was separately listed for research. The shape index from 1583 to 1644 was calculated using Equation (2).

$$LSI={\sum }_{i=1}^n{\displaystyle \frac{0.25E}{\sqrt{A}}}$$

(2)

where LSI is the shape index, n is the patch number, E is the patch perimeter, and A is the patch area of the Manchu region of control. The calculation of the shape index is based on a square with a regular shape; that is, the ratio of the side length (0.25 of the perimeter) to the square root of the area is 1. If the shape is more complex, the index value is greater than 1. This index is commonly used to embody shape characteristics. The linear functions and the functions with the biggest R² per year were used as the independent variables and the data on the area of control, LSI, and population were used as the dependent variables to determine the changes in the area of control and population size during the Manchu rise.

2.4.3. The Relationship Between Dry–Wet Change and the Manchu Rise

We used segmented interpolation to supplement any missing annual data on the area of control and population from 1583 to 1644 and finally calculated the LSI of the Manchu region for each year. We subtracted the annual scPDSI_4–10 between 1853 and 1644 from the mean (0.535) of the scPDSI_4–10 values from 1548 to 2022 and then obtained the absolute value of the deviation from the mean (0.535) of the scPDSI_4–10 (DV) per year. The larger the absolute value was, the worse the dry or wet conditions (such as floods or droughts) for a certain year were. The following abbreviations are used: DV = deviation from the mean of the scPDSI_4–10; GA = growth rate of the area of control; GL = growth rate of the LSI; and GP = growth rate of the population of the region of control. The growth rates from 1583 to 1644 were calculated separately for the area, LSI, and population size of the region of control using the following formulas:

$$\mathrm{growth} \mathrm{rate} \mathrm{of} \mathrm{area} \mathrm{of} \mathrm{current} \mathrm{year} = {\displaystyle \frac{\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}-\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}{\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}}$$

(3)

$$\mathrm{growth} \mathrm{rate} \mathrm{of} \mathrm{LSI} \mathrm{of} \mathrm{current} \mathrm{year}={\displaystyle \frac{\mathrm{L}\mathrm{S}\mathrm{I} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}-\mathrm{L}\mathrm{S}\mathrm{I} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}{\mathrm{L}\mathrm{S}\mathrm{I} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}}$$

(4)

$$\mathrm{growth} \mathrm{rate} \mathrm{of} \mathrm{population} \mathrm{of} \mathrm{current} \mathrm{year}={\displaystyle \frac{\mathrm{p}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}-\mathrm{p}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}{\mathrm{p}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{o}\mathrm{f} \mathrm{p}\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{o}\mathrm{u}\mathrm{s} \mathrm{y}\mathrm{e}\mathrm{a}\mathrm{r}}}$$

(5)

The correlation coefficients between the DVs and growth rates of these three indices (the area, LSI, and population size) were calculated separately in order to reveal the impact of dry–wet change on the Manchu rise. If the significances of the correlation coefficients were greater than 0.05, this indicated that some years’ data had weak correlations in both corresponding data series. The step-by-step elimination analyses were used to sequentially identify the data with a low correlation with the year data. This method was used in correlation analysis, especially in revealing the relationship between two corresponding series, with the significances of the correlation coefficients being greater than 0.05 [51,52]. In this way, it was possible to comprehensively determine the role of dry–wet change for all years from 1583 to 1644. Then, the eliminated correlation coefficients for the DVs and the growth rates of the three indices were calculated by the means of step-by-step elimination analysis. This analysis was conducted using the DV and the growth rate of the area of control (GA) as examples. The calculation process was as follows: We took out the data for 1583 and calculated the correlation coefficient (R) between the DV and GA from 1584 to 1644. Then, we took out the data for 1584 and calculated the R between the DV and GA for the remaining 62 years; this was performed in the same way as for the rest of the data. When all 62 Rs were calculated, the first round of calculations was completed, and the largest R implied that the correlation between the DV and GA was the best when the data for a certain year were taken out. Therefore, the data for a certain year were eliminated in the first round, and the greatest R value was the eliminated correlation coefficient (ER) of that year. Then, the second round of elimination was carried out in the same way: the remaining 61 years were screened in turn, and so on. One data point per year was eliminated in each round to complete the screening. The greatest R was the ER of the year that was eliminated in every round.

3. Results

3.1. Dry–Wet Change

3.1.1. Trend and Variability

The mean of the reconstructed scPDSI_4–10 from 1583 to 1644 was 0.822 (Figure 5), which was 0.287 higher than the mean from 1548 to 2022 (0.535) (Figure 4), indicating that the climate in Northeast Asia was relatively humid during this period. The standard deviation of the scPDSI_4–10 from 1583 to 1644 was 0.881, which was 0.312 smaller than the standard deviation from 1548 to 2022 (1.193), indicating that the scPDSI_4–10 fluctuation range during the Manchu rise (1583–1644) was smaller than that of the entire period (1548–2022) in Northeast Asia. In addition, the slope value of the linear function was −0.002 with no significant trend in dry–wet change (Figure 5), but the values of the reconstructed scPDSI_4–10 ranged from −1 to 3, showing a change of significant amplitude. Therefore, the relative humidity and fluctuating dry–wet change was the natural basis for the Manchu rise.

3.1.2. Extreme Dry–Wet Change

Based on the method described by Song et al. [53] and Shi et al. [54], when the scPDSI_4–10 was between −1 and 1, it was taken to indicate a normal year; when it was between 1 and 2, it was taken to indicate a slightly wet year. Between 1583 and 1644, there were 34 normal years (scPDSI_4–10 from −1 to 1), accounting for 54.839% of the total of 62 years (1583–1644) (

Table 3. The number and proportion of slightly or moderately wet years from 1583 to 1644.

Dry or Wet Conditions	scPDSI4–10 Range	Number of Occurrences (Year)	Proportion (%)
Moderately wet	2–3	3	4.839
Slightly wet	1–2	25	40.323
Normal	−1–1	34	54.839

). The northeast region maintained normal conditions for most of this period, but the proportion of normal years was 4.951% lower than that in the region (59.790%) in the long term (1548–2022). The results indicate that some abnormal phenomena of dry–wet change, although not obvious, could be observed in the Northeast region. There were 25 slightly wet years from 1583 to 1644 (Table 3), accounting for 40.323%, which was 16.533% higher than the proportion of slightly wet years in the long term (1548–2022) (23.790%). The extreme dry–wet changes mainly exhibited slight wetting conditions, which may have promoted the growth of vegetation, increased the number of wild animals, and led to relatively abundant surface water in the Northeast region, which was conducive to the growth of fish populations. A slightly wet year was not likely to cause serious flooding. Therefore, the slight wetting conditions were favorable for the production of materials and for the life of the Manchu, who mainly relied on fishing and hunting.

3.2. Area and Population Changes in Manchu Region of Control

3.2.1. Expansion of Region of Control

The slope of the linear function between the area of control and the year was 5.462, and the R² was 0.842 (p < 0.01) (Figure 6a), indicating that the area of control continued to expand with the Manchu rise, which made full use of the favorable precipitation conditions to achieve regional expansion. A quadratic function provided the largest R² (R² = 0.984), with a sudden change point occurring around 1594 (Figure 6a), indicating that the expansion was relatively slow in the early stage; however, the expansion rate was very fast after 1594, rapidly expanding the area that was controlled by the Manchu. The relatively slow expansion in the early stage of the Manchu rise was mainly concentrated in the southern mountainous areas (Figure 6b), where the temperature was relatively high and the water conditions were the best, which was suitable for the Manchu way of fishing and hunting. Then, the Manchu expanded to the northern humid areas (Figure 6b), as the water conditions were good. They then expanded further into the western plain area (Figure 6b), where the water conditions were slightly poorer but still suitable for the Manchu way of life under slightly humid conditions, eventually reaching the Mongolian grassland (Figure 6b). Although the Mongolian grassland was a traditional habitat for nomadic peoples, the fact that their nomadic lifestyle was similar to that of the Manchu made the integration of these two ethnic groups relatively easy. Moreover, at this time, the Manchu were very strong, and could use their advantages to effectively control the Mongolian residential region. The Manchu took advantage of the region’s wet conditions to gradually move from their native areas into semi-humid zones and finally to a semi-arid region (a pastoral grassland region). Their territory expanded to 131 times its original size (1644/1588), completing Manchu regional control. As grassland ethnic groups (Mongolian) entered the Manchu-controlled areas, based on the Manchu people’s livelihood through fishing and hunting requiring teamwork, intensive pastoralism was adopted using the league–banner system which required a large amount of labor and led to the population flow to the Northeast region, increasing the Manchu-controlled areas [55].

3.2.2. The Complex Expansion of the Region of Control

The slope of the linear function between the LSI and the year was 0.036, and the R² was 0.679 (p < 0.05), indicating that as the controlled region gradually increased, its shape also became more complex. The ruling ability of the Manchu gradually strengthened as they expanded their territory, and their ability to control areas increased in the border regions. The trend of the sixth-order curve with the biggest R² (R² = 0.94, p < 0.01) (Figure 7) indicated that with the expansion of the region of control, its shape became more complex. From 1583 to 1594, the expansion of the region of control was relatively small, but the LSI changed significantly (Figure 7), indicating that in the early stage of expansion, the Manchu expanded to the region with the most similar geographical conditions to their original core area. Because the Manchu were mainly fishers and hunters, their region expanded along the river valley, resulting in a rapid increase in the complexity of its shape, and the initial expansion of the territory was promoted along this axis. From 1594 to 1609, as the expansion ability increased, the LSI did not change much (Figure 7). The expansion mode of the Manchu shifted from following an axis to a wider surface, indicating the strengthening of their ruling ability and the achievement of large-scale continuous expansion. From 1609 to 1629, the LSI increased again and the expansion region became more complex (Figure 7), indicating that the Manchu expanded along the same axis again. After 1629, the LSI decreased again (Figure 7), indicating that with the rapid improvement in expansion ability, the Manchu adopted the method of expanding across a wider surface again and expanded the Great Wall to rapidly increase their region of control.

3.2.3. The Population Growth in the Region of Control

According to historical records, the population size in 1643 (1.32 million) was 44 times that in 1588 (30 thousand) (Figure 8), demonstrating that the population increased rapidly over these 55 years. The slope of the linear function between population size and year was 1.911, and the R² was 0.809 (p < 0.05) (Figure 8), indicating that the population was continuously increasing, almost at the same rate as the expansion of the Manchu region of control. Not only did Manchu populations with similar lifestyles gather together, but people with other lifestyles also accepted that Manchu rule had promoted the rapid Manchu rise. The quadratic function achieved the largest R² (R² = 0.956, p < 0.01) at a sudden change point around 1600 (Figure 8), indicating that the population growth was relatively slow in the early stage in the traditional regions of the Manchu; however, the expansion rate was very fast after 1600, rapidly expanding the Manchu region of control. Before 1600, the traditional Manchu region in the eastern mountainous area expanded slowly due to the sparse population in the region. After 1600, the area of control reached the northeast plain and western grassland, and a large number of Han and Mongolian people gradually moved into the Manchu region of control, resulting in a rapid increase in population.

3.3. The Relationship Between the Rise of the Manchu and Dry–Wet Change

3.3.1. The Relationship Between Territorial Expansion and Dry–Wet Variability

The R between the DV and the GA was −0.18 (p = 0.17 > 0.05), showing that during the Manchu rise, there did not appear to be a close relationship between territorial expansion and dry–wet change. Therefore, step-by-step elimination was used to reveal this complex relationship. Elimination along the positive direction occurred in the 14th (p < 0.05) and the 18th round (p < 0.01), and along the negative direction, it occurred in the 2nd (p < 0.05) and the 5th round (p < 0.01) (

Table 4. The ERs along the positive and negative directions between the DV and GA.

Positive Direction				Negative Direction
Turn	Excluded Year	ER	p	Turn	Excluded Year	ER	p
1	1583	−0.14	0.271	1	1584	−0.22	0.087
2	1589	−0.11	0.405	2	1607	−0.25	0.048
3	1617	−0.07	0.583	3	1629	−0.29	0.026
4	1616	−0.03	0.829	4	1592	−0.33	0.013
5	1628	−0.01	0.930	5	1644	−0.36	0.006
6	1594	0.05	0.711
7	1587	0.09	0.506
8	1610	0.12	0.370
9	1585	0.17	0.263
10	1618	0.19	0.184
11	1624	0.22	0.124
12	1625	0.25	0.081
13	1621	0.28	0.053
14	1586	0.31	0.035
15	1593	0.33	0.022
16	1641	0.36	0.015
17	1611	0.38	0.010
18	1612	0.41	0.006

ER = eliminated correlation coefficient; DV = deviation from mean of scPDSI_4–10; GA = growth rate of area of region of control.

). Therefore, negative correlation elimination was selected to establish the relationship between the DV and GA. Except for during 5 years (Table 4) from 1583 to 1644, the smaller the dry–wet change was, the larger the expansion of the region of control was. The Manchu took advantage of the relatively stable and favorable environment with little dry–wet change, which provided increased and better resources for them to gradually expand their territory in most years.

3.3.2. Relationship Between Territorial Shape and Dry–Wet Variability

The R between the DV and the GL was −0.216 (p = 0.091 > 0.05), showing that during the Manchu rise, there was no strict adherence, with the change in the shape of the controlled region increasing as the DV decreased. Years that did not follow the above law were extracted via step-by-step elimination, which occurred in the 12th (p < 0.05) and 15th rounds (p < 0.01) along the positive direction and in the 1st (p < 0.05) and 6th rounds (p < 0.01) along the negative direction (

Table 5. The ERs along the positive and negative directions between the DV and GL.

Positive Direction				Negative Direction
Turn	Excluded Year	ER	p	Turn	Excluded Year	ER	p
1	1617	−0.14	0.276	1	1584	−0.26	0.042
2	1624	−0.07	0.622	2	1592	−0.29	0.023
3	1587	0.02	0.897	3	1609	−0.31	0.017
4	1585	0.06	0.661	4	1631	−0.32	0.013
5	1583	0.11	0.429	5	1595	−0.34	0.011
6	1586	0.15	0.256	6	1644	−0.35	0.007
7	1594	0.18	0.198	7
8	1632	0.20	0.149	8
9	1619	0.22	0.110	9
10	1644	0.25	0.079	10
11	1588	0.27	0.052	11
12	1595	0.30	0.033	12
13	1589	0.34	0.019	13
14	1600	0.37	0.010	14
15	1593	0.40	0.005	15

ER = eliminated correlation coefficient; DV = deviation from mean of scPDSI_4–10; GL = growth rate of LSI.

). Therefore, negative correlation elimination was selected to establish the relationship between the DV and GL. Except for 6 years (Table 4) in the period from 1583 to 1644, when the DV was small, the Manchu tended to expand along the river valley, forming an axial expansion feature, resulting in the area taking a complex shape. When the DV was large, the Manchu’s expansion was planar, which led to relatively less resistance and was therefore easier in most years.

3.3.3. The Relationship Between Population Growth in the Region of Control and Dry–Wet Variability

The R between the DV and the GP was 0.086 (p = 0.509 > 0.05), which demonstrates that during the Manchu rise, it was not strictly true that the greater the DV, the larger the population. Years that did not follow this law were extracted using step-by-step elimination, which occurred in the 4th (p < 0.05) and the 8th rounds (p < 0.01) along the positive direction and the 15th (p < 0.05) and the 18th rounds (p < 0.01) along the negative direction, respectively (

Table 6. The ERs along the positive and negative directions between the DV and GP.

Positive Direction				Negative Direction
Turn	Excluded Year	ER	p	Turn	Excluded Year	ER	p
1	1599	0.16	0.220	1	1621	0.03	0.826
2	1620	0.21	0.106	2	1587	0.00	0.970
3	1612	0.24	0.069	3	1613	−0.02	0.881
4	1643	0.27	0.042	4	1644	−0.04	0.738
5	1594	0.29	0.030	5	1595	−0.07	0.591
6	1632	0.31	0.020	6	1617	−0.10	0.481
7	1638	0.33	0.014	7	1639	−0.12	0.388
8	1584	0.35	0.009	8	1641	−0.14	0.324
				9	1589	−0.16	0.264
				10	1598	−0.18	0.214
				11	1625	−0.20	0.170
				12	1600	−0.22	0.130
				13	1583	−0.24	0.097
				14	1591	−0.26	0.069
				15	1602	−0.29	0.046
				16	1616	−0.32	0.030
				17	1586	−0.35	0.018
				18	1585	−0.38	0.010
				19	1614	−0.41	0.006

ER = eliminated correlation coefficient; DV = deviation from mean of scPDSI_4–10; GP = growth rate of population of controlled region.

). We used positive correlation elimination to establish the relationship between the DV and the GP. Except for the 8 years (Table 4) from 1583 to 1644, the smaller the dry–wet change was, the less the population increased. This result seems inconsistent with the results of the step-by-step elimination in the relationship between territorial expansion and dry–wet variability. The R between population size and the area of control was −0.48 (p < 0.05), indicating that in years with rapid territorial expansion, the increase in population size was very small. The territorial expansion could not immediately lead to a rapid increase in population, which may be related to the fact that the greater territorial expansion was, the larger the scale of war was and the greater the loss of population was. The 8 years that were excluded from the positive direction were 1599, 1620, 1612, 1643, 1594, 1632, 1638, and 1584 (Table 6); the 0.171 mean GP for these 8 years was greater than the mean GP of 0.069 from 1583 to 1644. This indicates that the population increased significantly in these 8 years, which happened to be 1–5 years after the years of significant territorial expansion: 1584 (GA = 0.58 > 0.11 (mean from 1583 to 1644) in 1583), 1594 (GA = 0.37 > 0.11 in 1593), 1599 (GA = 0.14 > 0.11 in 1597), 1612 (GA = 0.38 > 0.11 in 1610), 1620 (GA = 0.59 > 0.11 in 1616), 1632 (GA = 0.32 > 0.11 in 1629), 1638 (GA = 0.13 > 0.11 in 1635), and 1643 (GA = 0.13 > 0.11 in 1641). The mean DV was 0.589 for the scPDSI_4–10, which was smaller than the mean DV of 0.774 for the years 1583–1644. The scPDSI_4–10 values in these 8 years were 0.59, 0.73, 1.81, 1.53, −0.30, 0.44, 0.67, and −0.59, respectively; except for 2 slightly wet years, the other scPDSI_4–10 values were relatively small. These results show that in the years following rapid territorial expansion, the Manchu needed to engage in agricultural production in years with relatively small dry–wet changes, which required more labor. As labor increased, so did expected product returns, which not only promoted an increase in local birth rates but made the region more attractive for foreign immigrants, leading to a rapid increase in the population of Manchu-ruled region.

4. Discussion

4.1. Inconsistencies Between Regional Expansion and Dry–Wet Change

The results of our correlation analysis show that the year with the most significant regional expansion was not the year with the smallest dry–wet change. The mean DV in the 10 years with the highest GA (GA > 0.30) was 0.800, which was similar to the mean DV from 1583 to 1644 (0.774). These results indicate that the year with the most significant regional expansion was not the year with the smallest DV (0) but the year whose DV was closest to the mean DV from 1583 to 1644. The Manchu people who had lived in an area for a long time had strong adaptability to the mean dry–wet conditions. This conclusion is consistent with the finding that Yakuts have the strongest long-term adaptation to the average local climate over many years [56]. In addition, slight deviations in the dry–wet change had a certain impact on local human activities, causing a slight reduction in the number of production materials and a crisis of survival. However, it did not cause a significant amount of damage. Therefore, the Manchu were motivated to expand outward (by seizing territory to supplement the reduced number of materials) and had the strength for external growth (with certain material and population reserves) [57]. This strong drive led to the highest rate of outward expansion in the Manchu region. The scPDSI_4–10 values in these 10 years were −0.71 (1583), −0.88 (1589), 0.59, 0.22, 1.33, 0.30, 1.56, 1.68, 2.08, and 0.77, respectively; except for the early 2 dry years, the rest were wet years. The rapid territorial expansion of the Manchu was also carried out in these wet years, with the relatively abundant resources of fishing and hunting.

The mean GA of the 10 years with the lowest GA values was 0.05, much lower than the mean GA from 1583 to 1644 (0.11). The mean DV of these 10 years was 0.120, which was smaller than the mean DV from 1583 to 1644 (0.774). These results indicate that the dry–wet change was small, which was conducive to the progression of production and life in these years. Therefore, the Manchu returned to their already-occupied region to engage in production activities [58], and their motivation for external expansion decreased, resulting in the lowest expansion rate of the region of control. The scPDSI_4–10 values in these 10 years were −0.22, −0.16, −0.59, 1.33, 1.70, 0.73, 0.82, 0.73, 0.79, and 0.69, respectively; 3 years were normal dry years, 5 years were normal wet years, and only 2 years were slightly wet years among these 10 years. For further verification, in the years of small dry–wet change, territorial expansion was relatively weak, and normal wet years were favorable opportunities for the Manchu to rehabilitate.

In addition, 5 years were excluded from the negative direction using step-by-step elimination—1584, 1607, 1629, 1592, and 1644 (Table 4)—with a mean DV of 0.565 for the scPDSI_4–10, which was slightly smaller than the mean DV of 0.774 for the scPDSI_4–10 from 1583 to 1644. The mean GA was 0.285 for these 5 years, which was greater than the mean GA of 0.116 from 1583 to 1644. This result is similar to the results demonstrating that conditions close to the mean dry–wet conditions may have stimulated the greatest territorial expansion for the Manchu. The scPDSI_4–10 values in these 5 years were 0.59, 0.22, 0.30, 0.78, and 2.52, respectively, all of which were wet years, which is consistent with the results of wet years being conducive to territorial expansion. In the second year (1584) after the Manchu leaders declared the uprising, their strong motivation led to an increased rate of expansion. In 1592, the Manchu completed the unification of the core area (Jianzhou Jurchen), further increasing the rate of expansion [59]. In 1607, the capital of the Manchu-controlled region moved from the border area to the hinterland (Hetuala City) and expanded to the southern part of Changbai Mountain, rapidly increasing its territory [60]. Later, the Manchu expanded from the eastern mountains to the northeastern plain in 1629, with a significant increase in territory area. Territorial expansion was completed in 1644, and the region outside the Great Wall was as large as it could be.

4.2. Inconsistencies Between the Shape of the Region and Dry–Wet Change

The result of our correlation analysis showed that it was not smaller dry–wet change that led to the more complex change in shape during the Manchu rise. The mean GL (GL > 0.1) of the 9 years with the highest GL values was 0.203, which was higher than the mean GL (0.019) from 1583 to 1644. The scPDSI_4–10 values in these 9 years were −0.71, 0.59, −0.65, 1.73, 2.79, 1.33, 0.22, 1.33, and 2.08, respectively, which included 6 consecutive years from 1583 to 1588. The mean DV of these 9 years was 1.045, which was higher than the mean value from 1583 to 1644 (0.774). These results indicated that when the dry–wet change was significant, the unfavorable environment motivated the Manchu to grow outward, and the expansion direction followed the relatively favorable river valleys during the Manchu’s first rise with wet years, leading to an axial expansion characteristic [61]. The mean of the 9 years with the smallest GL values (GL < 0.02) was −0.122, which was lower than the mean GL (0.019) from 1583 to 1644. The scPDSI_4–10 values in these 9 years were 1.70, 0.73, 1.56, 1.36, 1.53, 1.68, 0.79, 1.66, and 0.43, respectively, and these were all wet years with large values. The mean DV of these 9 years was 0.761, which was almost the same as the mean of 0.774 from 1583 to 1644. These results indicated that the years with the most regular expansion of the Manchu had DVs closest to the mean value of several normal years and wet years. Due to the close connection between the expansion of the Manchu region and their original territory, regular growth was also the safest option [62], so when the dry–wet change was close to the mean conditions for several years, the Manchu adopted this.

The 6 years that were excluded from the negative direction were 1584, 1592, 1609, 1631, 1595, and 1644 (Table 5), with a mean DV of 0.927 for the scPDSI_4–10, which was slightly higher than the mean DV of 0.774 for the scPDSI_4–10 from 1583 to 1644. The scPDSI_4–10 values in these 6 years were 0.59, 0.22, 1.39, 1.66, 2.47, and 2.52, respectively, which were all wet years with large values. The mean GL for these 6 years was 0.070, which was higher than the mean GL of 0.019 from 1583 to 1644. These results showed that the relatively large dry–wet change and increased wetness over these 6 years were beneficial for the Manchu’s expansion further along the axis towards the core area of the region that they wanted to control. Data for the years 1584, 1592, and 1644 were consistent with the sudden increase in territory. The unification of the Manchu’s traditional region and the eastern mountains was completed the fastest in 1584 and 1592; in the mountains, the Manchu mainly resided in river valleys [63], so expansion here was mainly linear, resulting in significant change in the shape of the territory. In 1644, all the areas outside the Great Wall were occupied by the Manchu, and the unification of other scattered areas led to further important change in the shape of the territory. In 1629, the Manchu expanded from the eastern mountains to the northeastern plain, and in the following two years (1631), they extended along the Songhua River to enter the hinterland of the northeastern plain [64], changing the shape of the area again.

4.3. Inconsistencies Between Population Growth and Dry–Wet Change

The results of our correlation analysis show that greater dry–wet change did not cause more significant population increases during the Manchu rise. The mean DV of the 9 years with the highest GP values (GP > 0.15) (mean GP: 0.288) was 0.744, which was only slightly different from the mean DV (0.774) from 1583 to 1644. The scPDSI_4–10 values in these 9 years were 1.81, 1.53, 0.89, −0.30, 0.73, 1.18, 1.02, −0.25, and −0.59, respectively, which represented 6 wet years and 3 dry years. These results indicated that the years of the fastest population growth were not the years with the smallest DV (0) but the years, regardless of whether they were wet or dry, whose DV was closest to the mean value of several years. This result indirectly indicates that once an ethnic group living in a specific region adapts to its dry–wet change in the long term, the population will continue to increase, and the group may grow stronger [65].

The mean DV of the 9 years with the lowest GP values (<0.002) was 0.666, which was only 0.078 higher than the mean DV of the 9 years with the highest GP values (0.744). However, the standard deviation of the DV of the 9 years with the highest GP values was 0.338, while the standard deviation of the DV of the 9 years with the lowest GP values was 0.527, indicating a large difference. The scPDSI_4–10 values in these 9 years were 1.70, 0.73, 1.76, 1.68, 2.08, 0.75, 0.79, 044, and 0.67, respectively, which show big differences among the scPDSI_4–10 values. In the 9 years with the lowest GP values, there were 4 years with high DVs (mean DV: 1.273), during which the population was small. The remaining 5 years had low DVs (mean DV: 0.179); during these years, the population was also small. Therefore, the population growth was small regardless of whether the dry–wet change was too large or small. Significant dry–wet change meant that there was a trend of flooding, causing damage to social foundations and slowing or even reversing population growth [66]. In addition, if the deviation in dry–wet change was too small, the environment would be stable, and people may have focused on production, which may not have stimulated external expansion to increase the population. Only a dry–wet change with a certain, but not particularly large, deviation (that is, when it was just around the mean DV) would motivate the Manchu to regionally expand, leading to rapid population growth.

4.4. Spatial Heterogeneity and Manchu’s Expanding Territory

In the western grasslands of Northeast China, a traditional gathering region of the nomadic Mongolians, there was a six-year drought from 1628 to 1633 which severely damaged the grasslands. The Mongolian strength was greatly reduced [49], leading to a rapid decline in resisting Manchu expansion. The agricultural region in the central plain was affected by the Little Ice Age, which resulted in annual temperatures of about 2–3 °C lower than those of contemporary times. The 7-month freezing period made crop growth extremely vulnerable to frost disasters and reduced grain production, leading to a decline in the strength of the Han ethnic group, which mainly relied on agriculture [67], and they were unable to resist the Manchu invasion. Therefore, climate change, especially dry–wet change, increased the strength of the Manchu, while the strength of the Han and Mongolian groups declined. The spatial heterogeneity of dry–wet change in Northeast Asia also reduced resistance to the expansion of the Manchu. In addition, the region south of the Great Wall of China experienced severe drought, a sharp decline in population, frequent internal conflicts, and decreased control over border regions, especially in Northeast China [20]. At the same time, the Mongolian grasslands also experienced drought, resulting in decreases in grass production, population, and the strength of the Mongolians [68]. Therefore, the drought during the same period led to a decline in the ethnic governance of neighboring regions in Northeast Asia, providing strong external conditions supporting the Manchu rise.

4.5. The Impacts of Lower Temperatures

Affected by the global Little Ice Age, Northeast China also experienced a cold period during the late Ming Dynasty [69]. Low temperatures affected agricultural activities and led to a decline in the strength of the Han ethnic group [70]. However, the Manchu engaged in fishing and hunting activities, which had a relatively small impact [71]. Thus, the strength of the Manchu relatively increased, and the resistance to their expansion decreased. Some Han people in Northeast China were also forced to abandon agriculture and engage in hunting, resulting in their assimilation by the Manchu [72]. Therefore, the temperature drop in Northeast China during this period was also a favorable climatic condition for the rise of the Manchus.

5. Conclusions

Based on Carya cathayensis tree rings and historical data, we reconstructed scPDSI_4–10, geographical, and population data for the Manchu control region from 1583 to 1644. Function fitting and step-by-step elimination analysis were used to study the relationships among dry–wet change, areas of control, and population changes in this region and reveal the impact of the dry–wet change on the Manchu rise. Our research results are as follows: (1) The mean was 0.822 and the standard deviation was 0.881 for the reconstructed scPDSI_4–10 from 1583 to 1644. Dry–wet change provided favorable climatic conditions and a reliable foundation for the rise of the Manchu, who were mainly engaged in fishing and hunting. (2) The highest R² for the quadratic function (R² = 0.984; p < 0.01) was observed between the area of control and the year, indicating that during their rise, the Manchu continuously advanced from mountains to pastoral regions, taking advantage of the favorable conditions that were brought about by the dry–wet change. (3) The step-by-step elimination analysis showed that dry–wet change played a significant role in promoting the Manchu rise. However, regarding the expansion of the area and the increase in the population, the fastest developments did not occur in the years with the smallest dry–wet change but in the years with slight fluctuations.