Analysis of the Characteristics and Agglomeration Effect of the Rural Element Spatial Correlation Network in Northeast China

Abstract

In the face of the urgent need for the coordinated development of regional rural functions and the orderly and efficient integration of urban and rural areas, the problem of how to accurately identify the spatial correlation relationships and characteristics of rural elements among regions in Northeast China has become a key issue that urgently needs to be resolved. The results show the following: (1) The overall spatial correlation network (SCN) in the Northeast region from the perspective of rural element gravity has obvious differences. Each province has generated a strong connection center, and “strip-shaped” connection belts have been formed across provinces and cities. (2) From the perspective of the spatial pattern of the strong connection attributes of rural elements, Heilongjiang Province presents a polygonal “rhombus network”, Jilin Province presents a closed-loop “triangle network”, and Liaoning Province presents an irregular “trapezoid network”. (3) The connection relationships of rural element nodes within the provincial scope show that Yichun is an important hub connecting all directions within the province; Changchun and Siping have become the central nodes connecting the nodes on the northwest–southeast wings; Fuxin and Yingkou have become the central locations connecting the nodes on the southwest–northeast sides. (4) There are four sectors in the network, and the rural element transfer mechanism among the sectors shows that Block I and Block II are net spillover sectors, playing the role of “resource-based” sectors, and transmitting information to the net inflow Block IV through the broker Block III, presenting a “gradient” transmission mode.

1. Introduction

With the implementation of the rural revitalization strategy, China has entered a critical period for the transformation of urban–rural relations and rural development. For a long time, the problems of China’s urban–rural dual structure and the barriers to the flow of urban–rural elements have been prominent [1,2]. Urban–rural elements have experienced flows varying from small to large numbers, and from unidirectional to bidirectional, during the evolutionary stages of urban–rural relations, which have progressed from urban–rural differentiation, through opposition and integration to a more integrated phase [3,4,5]. Population, land, and capital in the composition of urban–rural elements are traditional core elements [6,7]. With the development of the social economy, other new elements such as services, technologies, and information have emerged. Compared with cities, the development of agriculture, rural areas, and farmers in China is relatively lagging behind. With the progress of reform and opening up as well as industrialization and urbanization, a large number of rural labor forces have been transferred to cities year by year. The phenomenon of “hollow villages” has been on the increase, and the phenomenon of rural decline has gradually emerged [8,9,10]. Academic research on the connection of rural element flows is mostly confined to the local element flows within the county scope, and the connection of rural elements across regions and its importance are highly neglected in the research of SCN. Rural areas in Northeast China have unique population, cultivated land, industrial base, and development trends. The research on the SCN of its elements is necessary for understanding the distribution laws and agglomeration directions of rural elements in Northeast China.

In addition, the extensive urban–rural spatial restructuring has a drastic impact on and poses challenges to rural areas [11]. The quality and quantity of rural cultivated land have gradually declined in the intense interaction of urban–rural element allocation. The realization of rural revitalization is inseparable from the inflow and agglomeration of various elements. Rural areas give birth to new business forms by gathering various elements and promoting the comprehensive economic and social development of rural areas [1,12]. The urban–rural division also leads to a positive correlation between income inequality and per capita consumption [13]. The precautionary savings motivation of rural residents is obviously stronger than that of urban residents [14]. The urban–rural income gap is the main indicator reflecting the urban–rural dual structure, and farmers’ income is a key variable affecting their consumption expenditure, living conditions, and welfare level [15]. The “push” of the interaction of elements (population, cultivated land, industries, and income) between urban and rural areas and the “pull” of the urban–rural income gap jointly become the actual obstacles to the coordinated development of urban and rural areas, and the overall trend is “urban-biased”. Currently, the models applied in the SCN generally fall into two categories. One is the gravity model based on the “static space” attribute. In 1942, Zipf first introduced the law of universal gravitation into the theory of spatial interaction. In the 1950s, foreign scholars conducted a large number of studies on inter-regional connections, trade flows, and central places [16,17]. Since then, domestic scholars have focused on research on the distance decay effect and spatial interaction and measured the spatial connection intensity through the economic connections, masses, and distances between regions [18,19,20]. The other is the data flow model based on the “dynamic space” attribute [21]. In recent years, it has become a new trend to explore the characteristics of urban network structures by using entity “flows”, including migration data flows, economic flows, trade flows, logistics flows, and so on [22,23,24,25]. He Renwei pointed out that in the “rural revitalization pentagon”, people, land, capital, and industries are the keys to rural revitalization. It is necessary to promote the coupled development of “people–land–capital–industries”. Research on rural population, cultivated land, industries, and income elements is the key to realizing rural revitalization [26].

As an important major grain-producing area in China, the rural development based on the black soil in the Northeast region has received widespread attention. The data from the Fifth National Population Census to the Seventh National Population Census indicate that the rural population in Northeast China has decreased by approximately 1.45 million in 20 years. Since the implementation of the cultivated land requisition–compensation balance policy in Northeast China, the cultivated land reserve has shown a continuous growth trend, increasing from 36.64% in 2000 to 36.97% in 2020 [27]. Meanwhile, the phenomena of reclaiming wetlands and forests for farming are relatively common, and the loss of a large amount of cultivated land reserves is severe [28]. The primary industry structure in Northeast China shows an unbalanced state. The total costs of planting wheat, corn, and soybeans in China are 203%, 53%, and 80% of those in the United States, respectively. The high cost of agricultural products is due to the high cost of production factors and low labor productivity, thus developing into a pattern of high yield, high inventory, and high import [29]. The per capita income of rural areas in Northeast China increased by 139% year-on-year from 2003 to 2023. Generally, the per capita income of rural areas in Heilongjiang Province and Liaoning Province is higher than that in Jilin Province, and the growth pattern has been stable over the years. The high agglomeration and high loss dynamic flows of rural elements in the Northeast region of China are particularly prominent and will affect the social and economic development of surrounding and other regions [30,31]. Despite the fact that research on the spatial correlation of various elements is increasingly garnering attention in the academic realm, research regarding the Northeast region still has numerous shortcomings. Past research has predominantly focused on the overall urban or regional economy, centering on the connections and cooperation between powerful entities in developed regions. Research on rural elements has usually been restricted to the flow of local elements within the scope of counties, overlooking the vital role of rural elements in cross-regional network linkages. Traditional research approaches have difficulty effectively capturing the intricate spatial correlation relationships among rural elements. They are incapable of precisely measuring the flow intensity and interaction mechanisms of different elements between regions, and as a result, it is arduous to put forward practical optimization strategies.

Therefore, the combination of the modified gravity model and social network analysis method used in this study effectively makes up for the deficiencies of existing research. By incorporating per capita GDP as a correction coefficient, the modified gravity model fully considers the influence of regional resource density differences on the connection strength, enabling it to more accurately measure the actual correlation strength among rural elements. From the multi-dimensional perspective of centrality, the social network analysis method systematically analyzes the structural characteristics of the rural element network, clearly revealing the status and role of each node in the network, as well as the transmission paths and agglomeration effects among elements, providing a comprehensive and in-depth perspective for a deeper understanding of the SCN of rural elements. These will strongly support the development of subsequent research and provide new ideas and methods for solving the rural development dilemma in Northeast China. Follow-up research will focus on how to formulate specific policy measures and action plans based on the network characteristics revealed in this study, such as optimizing the rural industrial layout, promoting the rational flow of elements, and strengthening regional coordinated development, with the aim of effectively improving the rural living environment in Northeast China and promoting the comprehensive revitalization of the rural economy and society.

2. Materials and Methods

2.1. Study Area

Northeast China comprises Heilongjiang, Jilin, and Liaoning provinces. It is in the northeast of China and the core area of Northeast Asia. It adjoins Russia and the DPRK in the east and north, links to Inner Mongolia in the west, connects to Hebei in the south, and faces the Shandong Peninsula across the sea (Figure 1). Its terrain mainly consists of plains, mountains, and hills, with a temperate monsoon climate. It is rich in natural resources like fertile black soil, vast forests, developed water systems, and diverse minerals. Economically, it is an important commercial grain base and heavy industry base, with a high level of agricultural mechanization and large grain output. It developed rapidly in the early days of New China and is still developing and transforming. Its geographical location offers an external expansion boundary for the rural element SCN and a convenient basis for the flow of elements within the region. Its internal and surrounding connections affect the flow of rural elements among provinces.

2.2. Rural Element Selection and Data Processing

The relevant rural elements mainly include population, cultivated land, industries, and income, among others [32,33]. Among them, as the most active entity element, the rural population’s density and scale can reflect the migrationtrend of the regional population [34]. The rural cultivated land element, as an important material basis for rural development, represents the natural resource endowment situation in rural areas and will directly affect agricultural production capacity and the potential for rural economic development [35,36]. The rural primary industry index, as the support for the rural economy and material support, can accelerate the urbanization process in China and expand domestic demand [37]. Rural per capita income can reflect the overall living standards and economic development status of residents in rural areas, and the level of rural per capita income directly determines the size of the urban–rural income gap [38]. Selecting rural population, rural cultivated land capacity, rural primary industry output value, and rural per capita income to construct an SCN and analyzing its gravitational intensity and distribution in the flow of regional elements can reflect the characteristics of rural areas in multiple aspects such as society, economy, and nature from different perspectives, understand the interaction intensity of rural elements among regions, and play a promoting role in the coordinated development of rural elements among provinces and within provinces, as well as in urban–rural integration.

Considering the timeliness of the data, this paper collects the statistical data of 2023. Panel data on rural population, rural cultivated land capacity, rural primary industry output value, and rural per capita income of 36 urban nodes in Northeast China (Heilongjiang Province, Jilin Province, and Liaoning Province) are obtained. The statistical data are sourced from the 2023 statistical yearbooks of each province (municipality). The website addresses are https://tjj.hlj.gov.cn/, http://tjj.jl.gov.cn, and https://tjj.ln.gov.cn, respectively, accessed on 1 June 2024.

2.3. Research Program

This research analyzes the rural factor gravity SCN of 36 research units in Northeast China. By using ArcGIS Pro 3.0 software to calculate the coordinates and relative distances of each region and applying the modified gravitational model, 4 × 1261 matrix data in Northeast China are obtained (Figure 2). Subsequently, through visualization operations such as density calculation, gravitational intensity, and XY line conversion, the spatial representation of the rural factor SCN in Northeast China is achieved. With the help of Ucinet 6.560 software, the social network analysis of the rural factor SCN in Northeast China is carried out, and four gravitational matrices of rural factor connection intensity are obtained: 36 × 36 (Northeast China), 13 × 13 (Heilongjiang Province), 14 × 14 (Liaoning Province), and 9 × 9 (Jilin Province). The matrices are binarized by taking the average of each gravitational value as the truncation value. Values greater than the truncation value are set to 1, and those less than it are set to 0. The central structure within the three provinces in Northeast China is characterized by centrality and other measures. The agglomeration effects among the three provinces are represented by cohesive subgroups and block models.

2.4. Research Methods

2.4.1. Modified Gravity Model

The traditional gravity model was developed by George Kingsley on the basis of Newton’s theory of universal gravitation and is a powerful tool for demonstrating the interaction among regions. Subsequently, Li Mengcheng, Tian Ye, and others referred to the research results of predecessors [39,40,41] and used the modified gravity model to measure the interaction among regions and obtain the connection intensity among regions. The connection intensity shown by the traditional gravity model ignores the influence of regional resource density. The differences in actual elements and levels among regions lead to an increase in the difference in connection intensity. Therefore, per capita GDP is incorporated into the formula as a correction coefficient to ensure that the difference in connection intensity among the research regions is reduced. In this paper, urban areas are taken as the research nodes, and its formula is as follows:

$$R_{ij}=K_{ij}{\displaystyle \frac{M_i{M}_j}{D_{ij}^2}}$$

(1)

$$K_{ij}=g_i/(g_i+g_j)={\displaystyle \frac{G_i/p_i}{G_i/p_i+G_j/p_j}}$$

(2)

In the above formula, R_ij represents the connection intensity between node i and node j. M_i and M_j represent the rural element index items of a certain node. D_ij represents the distance between node i and node j. K_ij is the correction coefficient. g represents the rural per capita GDP of a certain node. G represents the annual total rural GDP of a certain node. And P represents the annual rural population of a certain node.

2.4.2. Social Network Analysis Methods

Degree centrality (DC) is mainly used to measure the core position of a node within a specified area [42]. If the node has widespread connections with other nodes and presents a “star-shaped” expansion, that is, the node has the most connections in the gravitational network, then it can be regarded as the core position within the region [43,44]. In addition, degree centrality includes in-degree and out-degree. In-degree refers to the number of relationships where elements of other nodes flow into this node, while out-degree refers to the number of connections where elements of this node flow into other nodes, and its formula is as follows:

$$C_{p\mathrm{i}}=T_i/(N-1)$$

(3)

In the above formula, C_pi represents the degree centrality (DC), T_i represents the number of connections between a certain node and other nodes within the region, and N represents the maximum possible number of connections.

Closeness centrality (CC) is used to describe the influence intensity between a certain node and other nodes in the network, and the influence intensity is measured by using the sum of the path distances between nodes. If the node is closer to other nodes and the sum of the distances is smaller, it indicates that the mutual influence intensity between the two nodes is greater. That is, if the connection distances between this node and other nodes in the network are generally close, then this node is the center of gravity of the network [45]. Its formula is as follows:

$$C_{Ii}=(n-1)/{\displaystyle \sum }_{i=1}^n{d}_{ij}$$

(4)

In the above formula, C_Ii represents CC, n represents the total number of research units in the whole region, and d_ij represents the shortest distance from node i to node j.

Betweenness centrality (BC), also known as betweenness centrality or intermediary centrality, is used to measure the probability of a certain node being in the “middle” position in the network. If a node is passed through the most times during the process of contacting other nodes, then this node is in the middle position of the network, indicating that this node has a high BC. Its formula is as follows:

$$C_{B\mathrm{i}}={\displaystyle \frac{2{\displaystyle \sum _{j<k}}{g}_{jk(i)}/g_{jk}}{(g-1)(g-2)}}$$

(5)

In the above formula, C_Bi represents BC, g represents the total number of connections within the regional scope, g_jk(i) represents the number of connections passing through node i among all the connections from node j to node k within the region, and g_jk represents all the connections from node j to node k within the region, where j ≠ k.

2.4.3. Cohesive Subgroups

Cohesive subgroups are used to characterize the connection relationships among subnetworks in the overall network, that is, the connection relationships between subgroups. This method can measure the number of subgroups that the overall network can be divided into and analyze the characteristics of the relationships among subgroups and within subgroups. The characterization of the gravity of rural elements in node clusters can explore the structural relationships inside and outside subgroups and understand the network characteristics among provinces and within provinces.

2.4.4. Block Model

The block model theory is a theory for studying the characteristics of network positions. Its analysis method is the Concor method, which uses correlation coefficients as similarity measures to conduct cluster analysis on matrices. In view of the spillover characteristics shown by various positions in the network, it can be divided into four sectors: net spillover sector, net inflow sector, two-way spillover sector, and broker sector. Use the density matrix and image matrix of clustered sectors to analyze the interaction characteristics of each sector, and test the characteristics of the sectors according to the interaction mechanism [46].

3. Results

3.1. Network Characteristics and Agglomeration Effects of Integral Nodes

Based on the gravity model, the rural element SCN in Northeast China in 2023 was constructed (Figure 3). The ArcGIS natural breaks method was used to divide the connection strength into five connection attributes, namely weak, weaker, medium, stronger, and strong.

The SCN constructed by four elements—rural population, cultivated land, the primary industry economy, and per capita income—in Northeast China presents a diverse and complex pattern. Overall, the SCNs of various elements have significant regional differentiation characteristics. The southern region shows relatively strong comprehensive correlation advantages, with various elements closely connected and frequently interacting, forming a relatively mature and active spatial correlation subsystem. However, the northern region shows uniqueness in some element correlations. For example, the cultivated land SCN is relatively developed, but the synergy and comprehensiveness of the overall element correlations are weaker than those in the southern region (Figure 3a). The spatial correlations among various elements do not exist in isolation, but are intertwined and influence each other, with complex coupling relationships and feedback mechanisms.

In terms of the spatial correlation of the rural population, the southern region of Liaoning Province shows a trend of high agglomeration and strong connections, forming a population agglomeration network with several important nodal cities as the core (Figure 3b). Relying on its relatively superior economic development environment, abundant employment opportunities, and relatively complete infrastructure and public service system, this region has attracted a large number of rural population inflows and agglomerations and has constructed a close and complex SCN of the population. In contrast, the population spatial correlations in most areas of Heilongjiang and Jilin are relatively loose. Only in some individual areas, such as Jiamusi and Shuangyashan, are there relatively strong population connection belts. Overall, it shows the characteristics of low population connection intensity and small network density, reflecting the imbalance of the population agglomeration effect among regions and the profound impact of the differences in economic development levels on the population spatial distribution pattern. In terms of regional comparison, there is a sharp contrast in the spatial correlation of rural population between southern Liaoning and most areas of Heilongjiang and Jilin. This difference clearly reflects the impact of disparities in economic development levels, infrastructure construction, etc., among different regions on the spatial distribution pattern of the population. The situation in southern Liaoning indicates that favorable development conditions can generate a strong gravitational force for population agglomeration and build a close-knit SCN of the population. This highlights the crucial role of economic and other factors in the spatial mobility and agglomeration of the population.

The northern part of Northeast China dominates the rural cultivated land SCN, showing the characteristics of connection belts that are linearly extended and have obvious directions, relying on specific core cities (Figure 3c). The vast plain terrain in this region provides a unique natural foundation for the large-scale contiguous distribution of cultivated land. Coupled with the agricultural production traditions and farming patterns formed over a long period of time, the cultivated land resources show a high degree of continuity and correlation in space, constructing a cultivated land SCN that radiates and diffuses outward and is interconnected with core cities as the hubs. This model provides a broad and concentrated cultivated land spatial foundation for the large-scale development of the local primary industry, which is conducive to integrating resources and carrying out large-scale agricultural production activities. This cultivated land SCN pattern is significantly different from the population SCN pattern. It is mainly restricted and driven by natural geographical conditions and agricultural production methods, forming a relatively independent and stable spatial organization model. It has extremely important strategic significance for the large-scale and specialized development of the agricultural industry and the guarantee of food security in areas with obvious cultivated land agglomeration effects in the northern part of Northeast China.

The spatial correlation of the rural per capita income gravity intensity shows obvious characteristics of southern agglomeration and gradient decline (Figure 3d). The southern region concentrates a large number of areas with relatively high rural per capita income gravity intensity, forming high-value agglomeration areas with several cities as the core, and producing a significant siphon effect, which has a strong attraction for elements such as manpower and capital in surrounding areas. As the spatial location moves northward, the relatively strong and above rural per capita income gravity intensity shows a gradually weakening trend, and most of the connection directions with relatively high intensity show a tendency to extend from the southern core area to the northern area, that is, starting from the core area centered on Dalian and ending at the marginal area of Siping. This spatial correlation pattern is closely related to the spatial distribution of elements such as population and the economy of primary industry. It demonstrates the differences in economic development levels and industrial structures among regions, exerting a profound impact on the income status of rural residents. At the same time, it also significantly influences the flow direction of the rural population and the spatial distribution of factor resources.

In general comparison, the SCNs in Northeast China from the perspective of rural element gravity show obvious differences. Among them, the connection attributes of rural population and rural per capita income are generally the same, both showing the characteristic of “strong in the south and weak in the north”, and having a positive proportional relationship with node density, while the connection attributes of rural cultivated land are just the opposite. The connection attributes of rural primary industry economy are relatively balanced, and each province has generated a strong connection center. The distribution of node density of rural primary industry economy in various provinces in Northeast China is relatively scattered and balanced, and “strip-shaped” connection belts have been formed through cross-provincial and cross-city connections. Harbin and Changchun, as the central cities of the Harbin–Changchun urban agglomeration, have relatively frequent connections in multiple rural elements. It can be seen from the connection intensity that the strong connections of various rural elements are mostly with their spatially adjacent cities, which once again confirms the direct impact of spatial distance on the connection intensity.

Characteristics and Connection Relationships of Provincial Node Networks in Northeast China

In Heilongjiang Province (Figure 4), the spatial correlation intensity of rural elements is mostly distributed in a pattern of “local concentration and overall dispersion”. This characteristic indicates that the SCN of rural elements in Heilongjiang Province shows a disordered state and is less affected by rural elements in other regions. At the level of individual administrative regions, Harbin, Suihua, Jiamusi, and Shuangyashan have relatively strong element conduction, radiation, and control capabilities in the connection of rural elements, and should become the core demonstration areas for the development of agricultural modernization in the entire province. The core positions of Heihe and Qitaihe are not prominent, and they have not yet fully utilized their geographical advantages of linking the northern and southern central regions, and their radiation-driven effect on the surrounding node networks is relatively weak. Overall, Heilongjiang Province presents a polygonal “diamond network” at the spatial connection level.

The distribution area of the strong spatial connection of rural elements in Jilin Province is relatively dense, showing a “center-periphery” structure (Figure 5), presenting a closed-loop “triangle network”. Various rural elements are distributed along the Baicheng–Songyuan–Changchun–Jilin line in the northwest direction with Changchun as the central node. The densely spatially correlated areas of rural per capita income are mainly concentrated in the southwest, and there is no good spatial adaptability between other rural elements and rural per capita income, and the connection attributes are quite different from the former three. However, Changchun still becomes the “bridge” and “hub” of the surrounding regional connections by virtue of its economic strength, infrastructure, policy advantages, and population size.

The strong connection nodes of various rural elements in the Liaoning Province connection network show an obvious “mesh” structure, approximately an irregular “trapezoidal network”. Dalian and Yingkou, as the core nodes, are located on the southern edge of Liaoning Province. Although there are certain spatial obstacles in the spatial interaction with rural elements in other regions, due to their relatively high economic development level and proximity to the sea, they show a strong attraction to the rural population and industrial economy. As shown in Figure 6, Shenyang, Liaoyang, Panjin, Jinzhou, and Anshan, as the central nodes in Liaoning Province, are more equally and closely connected with each other, and rural elements can flow and spread quickly in the network, which is conducive to the cooperation and complementarity of resource configuration among regions.

The 13 × 13 (Heilongjiang Province), 14 × 14 (Liaoning Province), and 9 × 9 (Jilin Province) gravity matrices were binarized and imported into Ucinet software for calculation to obtain the rural element node connection relationships within each provincial region and were visualized through Netdraw. As shown in Figure 7, the node connection relationships are reflected by the size of the node identifiers and the density of the connection lines. It can be seen that there is a large difference between the node connection relationships of rural elements within the provincial region and the spatial connection intensity. Yichun has become the core connection node of the entire network system in Heilongjiang Province. It has a superior geographical location, adjacent to the provincial capital Harbin and close to the Russian border, and is located in the central part of Heilongjiang Province, serving as an important hub connecting all parts of the province from east to west and north to south. This geographical advantage enables Yichun to efficiently undertake the central node function of the provincial network system, providing convenience for the aggregation and allocation of resource information. Changchun and Siping are the core connection nodes of the network system in Jilin Province, located in the central part of the province and serving as important transfer stations for the comprehensive transportation network, such as railways and highways, providing strong support for the connections and resource flows within and outside the region, and becoming the central nodes connecting the nodes on the northwest–southeast wings. The core connection nodes of the network system in Liaoning Province are Fuxin and Yingkou. Fuxin is located in the west and is an important passage connecting Inner Mongolia and the hinterland of Northeast China, while Yingkou is located on the eastern coast and is an important foreign trade port, jointly constructing the inland–coastal linkage pattern in Liaoning Province and becoming the central location connecting the nodes on the north–south and northeast sides.

3.2. Network Characteristics and Agglomeration Effects of Individual Nodes

By calculating indicators such as degree centrality, closeness centrality, and betweenness centrality to analyze the centrality characteristics of associated individuals in the network, the status and agglomeration role of different provincial centers in the rural element SCN are revealed.

3.2.1. Degree Centrality Analysis

The roles of rural elements in various parts of Northeast China in the overall SCN show obvious differences (Figure 8). Among them, the nodes with spatial correlations mainly characterized by the inflow of rural elements from other regions are mainly concentrated in Liaoning Province. That is, the nodes with larger in-degrees are Liaoyang, Yingkou, Dalian, Dandong, Panjin, Benxi, Jinzhou, etc. This phenomenon reflects that Liaoning Province has unique advantages in attracting the inflow of rural elements. Its economic development level is relatively high, and its industrial structure is more diversified, being able to provide more employment opportunities and more complete infrastructure and public services.

The out-degree of a node is the out-degree of the node of the spatial correlation formed by the outflow of rural elements to other regions. The out-degrees of nodes such as Heihe, Benxi, Daqing, and Hegang are prominent. This indicates that these regions are relatively active in the output of rural elements, perhaps because they have relative weaknesses in some rural elements, and various rural elements are attracted by other nodes and cause outward flow.

The comprehensive centrality of Dalian, Yingkou, Shenyang, etc., is higher than that of other nodes, and they are all located in the central and southern parts of Northeast China. The stability of the overall network structure of the flow of rural elements in Northeast China is highly dependent on these nodes. They are the core nodes of the SCN and are in a central position. The SCNs of rural elements in other regions have a large number of relationships with these nodes. Among them, the centrality of Liaoyang is the highest, which indicates that Liaoyang is in a key position in the spatial connection of rural elements in Northeast China. This is related to the geographical location of Liaoyang. Liaoyang is in the center of Liaoning Province, connecting Shenyang upward and connecting Dalian downward through Anshan and Yingkou. Moreover, the in-degrees and closeness centralities of Yingkou and Anshan are the highest, respectively, playing a “bridge” role in the SCN.

3.2.2. Closeness Centrality and Betweenness Centrality Analysis

The average CC of the rural element SCN in Northeast China is 124. Among them, Anshan, Heihe, and Qiqihar are higher than the average value, indicating that these nodes can quickly connect with other nodes in the inter-provincial rural element SCN. The CC of Anshan is significantly higher than that of other nodes. Located in the central area of Liaoning Province, it benefits from its superior geographical location and relatively developed infrastructure such as transportation and communication. This enables elements such as rural labor, agricultural products, and agricultural technologies in surrounding areas to quickly gather here and spread to other regions, forming the “hub” of rural element connections.

The average BC of the rural element SCN in Northeast China is 22, indicating that rural elements in various regions in Northeast China can quickly and effectively establish correlation relationships. There are nine regions with values higher than the average, namely Benxi, Panjin, Dalian, Huludao, Shenyang, Tieling, Hegang, Yingkou, and Qitaihe. They are mainly distributed in the southern part of Northeast China. These nodes are in a central position in the SCN and have a relatively strong control ability over the spatial correlation of other regions. Moreover, as important port cities, they play an extremely important role in the import and export trade of agricultural products and the introduction and export of agricultural technologies and can effectively promote the circulation and exchange of rural elements in domestic and foreign markets.

Overall, there is not much difference in the change directions of DC, in-degree and out-degree. Most regions show a positive correlation, but there are individual regions with prominent differences. For example, Anshan has the highest CC and Benxi has the highest BC, with obvious differences in centrality degrees from other nodes. The fact that Anshan has the highest CC means that it has a super ability in the rapid connection and integration of rural elements. The highest BC of Benxi shows that it dominates the control of rural element circulation channels, which stems from its specific industrial layout and the status of a regional transportation hub, making it a key hub for the circulation of rural elements. It can be seen that on the basis of rural element connections, different nodes have a great disparity in control power in the overall SCN.

3.3. Cohesive Subgroup Analysis

The cohesive subgroup analysis of the rural element SCN in Northeast China was carried out with the help of Ucinet software, aiming to explore the network distribution pattern and the characteristics of the internal agglomeration structure, and to analyze the composition and relationships of each cohesive subgroup. The results showed that there were four cohesive subgroups in the network (Figure 9). Overall, the distribution of cohesive subgroups shows an obvious effect of geographical proximity. Most nodes appear in contiguous areas.

Cohesive Subgroup I has a relatively concentrated geographical space and frequent internal interactions, thus forming a cohesive subgroup with the closest SCN. The main reason for its obvious internal agglomeration effect is that these regions have a certain complementarity and synergy in cultivated land elements. They play an upstream role in the agricultural industrial structure. The central part of Northeast China is rich in forest land resources and has advantages in forestry supply and production links. The downstream areas, on the other hand, perform outstandingly in the processing or sales links, thereby promoting the frequent flow and close connection of elements.

Cohesive Subgroup II mainly consists of nodes within Heilongjiang Province and is located on both sides of Northeast China. This subgroup has a high degree of sharing and coordination in agricultural production resources (such as cultivated land resources) due to similar agricultural natural conditions and the long-formed agricultural production cooperation system within it, which can effectively reduce agricultural production costs and improve production efficiency. However, there may be industrial homogeneity within the subgroup, with an excessive focus on traditional agricultural planting or breeding, and a lack of diversified rural industrial development models.

The agglomeration effect of Cohesive Subgroup III is due to the proximity of the two provinces, and rural elements flow widely and are closely connected among these nodes. Its advantage lies in being able to fully utilize the geographical advantage to achieve the optimal allocation of rural elements between the northern nodes and the southern nodes. In terms of agricultural product circulation, it can shorten the transportation distance and reduce logistics costs. However, this subgroup also faces issues such as different policy orientations and development focuses in different regions, which may limit the efficiency of element flow and further affect the further optimization and expansion of the rural element SCN.

Cohesive Subgroup IV contains the most nodes, reflecting the complexity and diversity of the spatial correlation of rural elements in Liaoning Province. With the support of relatively high economic development levels in coastal nodes such as Dalian and Yingkou, there are more cooperation opportunities in the integrated development of rural industry and agriculture, and the expansion of rural service industries, which promotes the extensive flow and connection of rural elements. The extreme agglomeration effect may lead to intensified competition for rural elements within the region. In order to develop their own rural economies, each node may have excessive competition in attracting investment and competing for talents, which is not conducive to forming a synergistic force in the spatial correlation of rural elements within the region and may even cause resource waste and redundant construction.

3.4. Block Model Analysis

With the help of the Concer algorithm in Ucinet, by setting the maximum segmentation depth to 2 and the concentration standard to 0.2, the rural element SCN in Northeast China was divided into four blocks, and the block spillover effects were calculated (

Table 1. Block spillover effects of the rural element SCN in Northeast China.

Block	I	II	III	IV	Number of Nodes	Spillover Relations	Acceptance Relations	Expected Proportion of Internal Relations	Actual Proportion of Internal Relations	Block Type
I	26	33	50	103	10	186	13	25.71%	12.26%	Net spillover
II	1	3	17	89	10	107	52	25.71%	2.73%	Net spillover
III	8	12	16	48	5	68	96	11.43%	19.05%	Broker
IV	4	7	29	107	11	40	240	28.57%	72.79%	Net spill-in

).

As can be seen from Table 1, Block I has 186 spillover relations and 13 acceptance relations, with 10 nodes in the block. The expected proportion of internal relations is greater than the actual proportion of internal relations, indicating that the integration of elements within the block has not yet reached an ideal state, and there is a situation where resources have not been fully and effectively utilized. The number of external spillover relations is greater than the number of acceptance relations, and the radiation power of nodes is stronger than the attraction power, making it a net spillover block. This means that this block has a relatively strong output ability in terms of rural element correlations and has certain advantages in aspects such as rural population and agricultural products and exports them to the surrounding areas.

The connections among the number of spillover relations, the number of acceptance relations, the number of block nodes, the expected proportion of internal relations, and the actual proportion of internal relations in Block II are quite similar to those in Block I. This similarity shows that Block II and Block I have certain commonalities in rural element correlations, and the block type is also called net spillover. In the entire rural element SCN in Northeast China, the existence of multiple net spillover blocks may lead to intensified competition among regions. Maintaining their net spillover status among blocks may lead to conflicts with other blocks in the competition for resources.

Block III has 68 spillover relations and 96 acceptance relations, with five nodes in the block. The difference between the number of external spillover relations and the number of acceptance relations is not large. The block both spills relations to other blocks and receives relations from other blocks, and it is in a core position in the network, being a broker block. This special position enables it to play a crucial hub role in the network and can promote the exchange and interchange of rural elements among different blocks.

Block IV has 49 spillover relations and 240 acceptance relations, with 11 nodes in the block. The number of external spillover relations and the actual proportion of internal relations are much greater than the number of acceptance relations and the expected proportion of internal relations, making it a net spill-in block. Although the inflow of a large number of rural elements has injected vitality into its rural development to a certain extent, bringing rich agricultural production technologies and management experience and promoting the optimization and upgrading of the agricultural industrial structure, the number of acceptance relations far exceeding the number of spillover relations may lead to excessive dependence on external elements. Once there are fluctuations in the supply of external elements, the rural economy of Block IV will face a relatively large impact.

Through block model measurements, the block density matrix of the rural element SCN in Northeast China was obtained, with a total density of 0.427. Referring to the existing research [47], if the block density is greater than the overall network density of 0.427, it is assigned a value of 1; otherwise, it is assigned a value of 0. The value of 1 indicates the existence of a transmission relationship, and 0 indicates the absence of a transmission relationship. Then, the density matrix is transformed into an image matrix (

Table 2. Block density matrix and image matrix of the rural element SCN in Northeast China.

Block	Block Density Matrix				Block	Block Image Matrix
	I	II	III	IV		I	II	III	IV
I	0.267	0.33	1	0.936	I	0	0	1	1
II	0.01	0.033	0.34	0.809	II	0	0	0	1
III	0.16	0.24	0.65	0.873	III	0	0	1	1
IV	0.036	0.064	0.527	0.882	IV	0	0	1	1

), and the transmission relationships among the four blocks are drawn (Figure 10) to analyze the transmission paths of spillover and reception among the blocks.

It is easier to see the spatial correlation among the blocks from the statistics of the image matrix in Table 2. The density of Block IV itself has reached 0.882, reflecting a relatively high degree of reflexivity of the block. The corresponding values in the image matrix are all 1, revealing that the development connections among the internal members of the blocks are relatively close, and it has the characteristic of high aggregation of internal members. Figure 10 can more intuitively describe the transmission mechanism among the four blocks. Block I and Block II are net spillover blocks, assuming the role of “resource-based” blocks. We can also observe that Block I only spills over to Block III and Block IV, having no connection with Block II. As shown in Figure 9, spatially, Block I and Block III are adjacent, and Anshan, as a node of Block I, is embedded in Block IV; thus, an associated relationship is formed between them. In brief, Block I transmits information to the net-inflow Block IV via the broker Block III, presenting a “gradient” transmission pattern. Specifically, the net spillover Block I mainly establishes spillover relationships with the broker Block III and the net spill-in Block IV. However, the network density between it and the net spillover Block II is lower than the overall density, and there is no transmission relationship between the two blocks. This indicates that the central regional group in Northeast China has a relatively strong transmission force towards the southern region.

The transmission scope mainly covers the broker Block III and the net spill-in Block IV, and the connection intensity between the central region and the urban agglomerations on both sides of the central region is not obvious. The net spill-in Block IV mainly receives the spillover from the net spillover Block I and the net spillover Block II, and there is a one-way reception and transmission path. There is a mutual transmission path with the broker Block III. The net spill-in Block IV, as the key point of transmission, receives the spillover relations from the other three blocks. The broker Block III, as an intermediary block, plays the role of a “middleman”. Obviously, the correlative relationships among the blocks have constructed a complete transmission path.

4. Discussion

(1) Against the backdrop of the free flow of urban–rural elements, the flow of rural elements in Northeast China has presented a new trend of cross-regional, cross-provincial, and even cross-city movement. This phenomenon echoes the view in some previous studies that the flow of rural elements gradually breaks through the county-level restrictions. However, this study more deeply analyzes the significant differences in rural element flow and development between the northern and southern regions within Northeast China. Many previous studies have mostly focused on the analysis of rural elements within the county. This study, however, expands its perspective to the entire Northeast China and reveals the flow patterns of rural elements from a macro-regional level. For example, Liaoning Province in the southern part of Northeast China has a relatively high level of urbanization and economic development, which has attracted a large number of rural labor forces to flow in, effectively improving the rural population and income levels. However, the relative shortage of rural cultivated land elements has restricted agricultural development to some extent. The northern region of Northeast China is dominated by heavy industry, has abundant cultivated land resources and great agricultural development potential, and has attracted the injection of agricultural talents and capital. These differences in economic structure directly lead to the unbalanced development between the northern and southern regions, and this finding is a refinement and supplement to previous studies on the analysis of regional development differences.

(2) Regarding the research process, this study also has some shortcomings. Although the modified gravitational model and social network analysis method were used to analyze the rural element SCN, there are certain limitations in the time scale. One-year data may not objectively reflect the regional development dynamics. Research on the spatio-temporal evolution of rural elements can better analyze the phased changes. At the same time, although various elements were comprehensively considered, in-depth analysis has not been carried out on some complex factors, such as the short-term and long-term impacts of sudden policies on rural element flows, and the differential effects of different industrial policies on the flows of agricultural and industrial elements. There are also certain limitations in the research methods. Although the gravity model has some value in measuring the intensity of regional connections, it indeed cannot fully represent the actual connections between units. Limited by the length, future research can, on this basis, further improve the research methods, update the data, and deeply explore the complex relationship between rural elements and regional development, providing more solid theoretical support and practical guidance for rural development in Northeast China and even across the country.

(3) In terms of strengthening the radiating and driving role of core nodes, Heilongjiang Province can take Yichun as the core, upgrade the transportation network, build agricultural and forestry product processing industrial parks, and establish an element exchange platform to promote the efficient flow of rural elements across the province. Jilin Province constructs comprehensive rural industrial service centers in Changchun and Siping and sets up agricultural science and technology innovation bases to enhance the radiation to the wing nodes. Liaoning Province gives full play to the core-connecting advantages of Fuxin and Yingkou. It builds an inland agricultural product logistics hub in Fuxin and improves the import and export capacity of agricultural products at the port in Yingkou, promotes industrial collaboration between the two places, and constructs an inland–coastal linkage pattern. In terms of optimizing the SCN structure of rural elements, in view of the “strong in the south and weak in the north” pattern, Heilongjiang Province should increase investment in infrastructure and guide the gradient transfer of industries. In combination with the characteristics of the network structure of each province, demonstration areas for the agglomerated development of rural elements can be built in locally concentrated areas such as Harbin and Suihua, and the development achievements can be radiated across the province through e-commerce and logistics. Jilin Province can strengthen the industrial leadership of Changchun, create agricultural industrial clusters, and enhance industrial cooperation with peripheral cities. Liaoning Province can utilize the advantages of its “network-like” structure, establish a regional industrial cooperation alliance, and promote the integration of fishery production, processing, and sales. In terms of promoting the collaborative development of rural elements among regions, it strengthens the cooperation between the net-spillover regions and the net-absorption regions. Heilongjiang Province, with its cultivated land resources, and Jilin Province, relying on its advantages in agricultural product processing, can cooperate with the net-absorption region IV (mainly some areas in Liaoning Province) to build agricultural product production and processing bases, achieving complementary advantages.

5. Conclusions

This paper calculates the regional SCN from the perspective of rural element gravity through the modified gravity model and explores the network characteristics of its spatial connections by combining the social network analysis method, and draws the following conclusions:

(1) The overall network of spatial correlations in Northeast China, when viewed from the perspective of rural element gravity, shows obvious differences. Among them, the connection attributes of rural population and rural per capita income are generally the same, both showing the characteristic of “strong in the south and weak in the north”, and having a positive proportional relationship with node density, while the connection attributes of rural cultivated land are just the opposite; the connection attributes of the rural primary industry economy in Northeast China are relatively balanced, and each province has generated a strong connection center, and “strip-shaped” connection belts have been formed across provinces and cities.

(2) From the perspective of the distribution of strong connection attributes of rural elements, the number of nodes with strong connection attributes in Heilongjiang Province and Liaoning Province is relatively small, and the connection intensity is mostly distributed in a pattern of “local concentration and overall dispersion”; the distribution area of strong connections in Jilin Province is relatively dense, showing a “center–periphery” structure; the strong connection nodes of various rural elements in the Liaoning Province connection network show an obvious “mesh” structure. At the spatial level, Heilongjiang Province presents a polygonal “diamond network”; Jilin Province presents a closed-loop “triangle network”; and Liaoning Province presents an irregular “trapezoidal network”.

(3) The connection relationships of rural element nodes within the provincial scope show that Yichun has become the core connection node of the entire network system in Heilongjiang Province and is an important hub connecting all parts of the province from east to west and north to south; Changchun and Siping are important transfer stations in the network system of Jilin Province and have become the central nodes connecting the nodes on the northwest–southeast wings; the core connection nodes of the network system in Liaoning Province are Fuxin and Yingkou, jointly constructing the inland–coastal linkage pattern in Liaoning Province and becoming the central location connecting the nodes on the southwest–northeast sides.

(4) There are four cohesive subgroups in the network, corresponding to four blocks. The development connections among the members within the blocks are relatively close, and they have certain characteristics of internal member agglomeration. The rural element transfer mechanism among the four blocks shows that Block I and Block II are net spillover blocks, playing the role of “resource-based” blocks, and they transmit information to the net spill-in Block IV through the broker Block III, presenting a “gradient” transfer mode. The net spill-in Block IV mainly receives the spillover from the net spillover Block I and the net spillover Block II, and there is a one-way reception and transmission path. There is a mutual transmission path with the broker Block III. The net spill-in Block IV, as the key point of transmission, receives the spillover relations from the other three blocks. The broker Block III, as an intermediary block, plays the role of a “middleman”. Obviously, the correlative relationships among the blocks have constructed a complete transfer path.