The Spatial-Temporal Evolution and Urban–Rural Differences of County Resilience in China

Abstract

Under the background of urban–rural dual structure, integrating urban and rural resilience construction is inevitable to realize high-quality development. This study constructs a unified urban–rural resilience index system, evaluates county-level resilience of China, and reveals its spatial-temporal evolution and the characteristics of urban–rural differentiation. The results show that the development level of county resilience in China generally raised from 2010 to 2021, but regional differences are obvious. Development level of county resilience shows prominent urban–rural differences, with higher resilience in urban areas and lower in countryside, mainly due to the urban–rural differences in economic development and population. The urban–rural resilience difference in the southeast is significantly higher than that in the northwest China. These results suggest that there is a coexistence of regional and urban–rural divergence in the process of county-level resilience development, and attention needs to be paid to this complex feature when formulating development strategies.

1. Introduction

In recent years, the convergence of public health emergencies, geopolitical risks, and international financial crises has ushered the world into a period of turbulence and transformation, characterized by increasingly frequent external uncertainty shocks [1,2]. Amid this evolving global landscape, China’s development currently resides within a critical strategic opportunity period, yet it faces multiple threats stemming from both exogenous shocks and endogenous structural shifts [3]. Enhancing societal resilience against these risks has thus become imperative [4]. In response, building resilience has garnered significant attention from the state, the academic community, and policymakers as a key issue in social governance [5].

Derived from the Latin term “resilio”, meaning to rebound or spring back, the concept of resilience was initially adopted by physicists to describe the stability of materials and their resistance to external impacts [6]. With the rise of systems thinking, resilience was progressively extended to disciplines such as engineering and ecology, where its definition diversified according to the field of study. In engineering, resilience is defined as “engineering resilience”—the ability of a system to return to a state of equilibrium or stability following a disturbance [7]. In ecology, it is termed “ecological resilience”, referring to the magnitude of disturbance a system can absorb without undergoing structural change [6]. Subsequently, resilience research permeated socio-ecological domains, with evolutionary resilience framing it as the capacity of complex social-ecological systems to adapt, adjust, and transform in response to pressures [8]. Building on this conceptual evolution, studies on resilience have since expanded to encompass economic, technological, cultural, and infrastructural dimensions [9], giving rise in the field of social governance to prominent research themes such as urban resilience [1], community resilience [10], and disaster resilience [11]. Nevertheless, the applicability of these global frameworks often depends on national institutional contexts, which shape how resilience is manifested and operationalized in practice. In this regard, China offers a distinctive case.

China’s distinctive urban–rural dual structure offers a critical socio-institutional context for interrogating resilience theory. Since the mid-twentieth century, institutional mechanisms-including the household registration (hukou) system, differentiated land management regimes, and fiscal decentralization-have sustained systemic disparities in resource allocation, population mobility, and public service provision [12]. These institutional legacies have not only directed socio-economic development pathways but also configured the spatial distribution of environmental risks and adaptive capacities [13]. Consequently, resilience in China is better interpreted not merely as systemic adaptation, but as a process of coordination and rebalancing across structurally differentiated subsystems. Grounding global resilience theory within this context shifts the concept from abstract systems thinking toward a concrete analytical framework, capturing how resilience is shaped by pronounced spatial and institutional heterogeneity.

However, it has rarely been examined that, within China’s urban–rural dichotomy, substantial disparities exist in resilience development between urban and rural areas [14]. Top-down, one-size-fits-all resilience strategies have often proven ineffective, underscoring the urgent need for more flexible and differentiated approaches that account for these distinct contextual characteristics [15]. Although a limited number of studies [16,17,18] have identified significant differences in resilience levels between urban and rural settings, there is no consensus regarding which demonstrates higher resilience. For instance, Shi et al. [19] observed a non-linear variation in community resilience with increasing distance from the city center in the Beijing–Tianjin–Hebei region, yet their assessment relied solely on net primary productivity (NPP) and population density—indicators that may be considered insufficient. Similarly, Wu et al. [14] reported divergent growth patterns of urban and rural resilience in the Yangtze River Delta, but these findings were derived from separate measurement systems rather than a unified framework, limiting rigorous comparison and interpretation of the urban–rural divide. Therefore, it is imperative to develop a consistent and comprehensive evaluation framework to analyze urban and rural resilience across both temporal and spatial dimensions. Such an approach would not only enable accurate assessment of adaptive capacities to uncertainties and reveal vulnerabilities in resilience planning but also help bridge the urban–rural institutional divide and integrate cohesive development principles throughout the entire process of resilience building.

Moreover, existing studies have predominantly been conducted at the scale of specific provinces or urban agglomerations, resulting in constrained spatial and temporal perspectives. The county level serves as a critical intermediary between urban and rural societies, playing a pivotal role in socio-political and economic systems [20]. In the absence of comprehensive township-level statistical data, a county-based perspective offers a more granular understanding of the evolution of regional resilience and enables systematic identification of vulnerabilities in resilience development. Thus, this study adopts the concept of “county-level resilience” [21] to delineate its spatial focus, defining it—based on the notion of community resilience—as the capacity of a county to utilize sufficient reserve resources and rapid adaptive capabilities in responding to crises and disruptions [18,22]. In light of urban–rural relationship theory, the drivers and evolutionary pathways of county-level resilience are expected to vary across different urban–rural contexts. It is essential to elucidate the heterogeneity in urban and rural resilience and its underlying mechanisms to inform targeted and effective resilience-building strategies, thereby promoting balanced and sustainable enhancement of resilience in both settings.

Therefore, county-level statistical and geospatial data are employed to systematically evaluate the development of resilience across Chinese counties from 2010 to 2021. The urban–rural disparities and spatiotemporal evolution of county-level resilience are further elucidated through multidimensional analysis encompassing resilience subsystems and geographical regions, thereby providing a scientific basis for the design of context-specific resilience strategies.

2. Materials and Methods

2.1. Framework Construction

A methodological flowchart depicting the complete analytical procedure—from data acquisition and processing to resilience evaluation and spatial analysis—is provided in Figure 1.

2.2. Data Sources

Initial socio-economic data were obtained from the China County Statistical Yearbook and various provincial and municipal statistical yearbooks spanning 2010 to 2021. PM2.5 emission data were sourced from the Atmospheric Composition Analysis Group at Dalhousie University. Road data were acquired from OpenStreetMap (OSM). CO₂ emission data were derived from the Emission Database for Global Atmospheric Research (EDGAR) developed by the European Commission. Gridded electricity consumption data at a 1 km × 1 km resolution were obtained from a globally calibrated dataset [23].

Cropland, green space, and wetland areas were calculated based on the 30 m resolution Annual China Land Cover Dataset [24]. Net Primary Productivity (NPP) data (MOD17A3) and Normalized Difference Vegetation Index (NDVI) data (MOD13Q1) were provided by NASA.

Due to partial missing values in the original datasets, a manual verification process was conducted. Counties with extensive data gaps were excluded and remaining missing values were addressed using linear interpolation. The final study area encompassed 2004 counties across China.

2.3. Resilience Index System

Existing composite indicator systems for measuring urban–rural resilience generally incorporate three to seven subsystems. In addition to economic, social, infrastructural, and ecological resilience, some studies have proposed the inclusion of institutional [25], community capital [11], and traditional knowledge resilience [26], among others. This study adopts the widely applied four-dimensional analytical framework—economic, social, infrastructural, and ecological resilience [5,27,28,29,30]—to structure the criterion layer for assessing county-level resilience. Specific evaluation indicators were designed to comprehensively reflect the conceptual depth of resilience by accounting for performance across pressure-state-response stages and characteristics such as robustness, redundancy, diversity, and rapidity. Considering the need for nationwide applicability, 18 indicators were selected (

Table 1. Community resilience indicator system.

Dimension	Metric (Refs.)	Definition	Effect	ID	Unit
Economic Resilience	GDP [31]	Aggregate Economic Output	+	EcR1	CNY
	GDP Growth Rate [5]	Economic Growth Stability	+	EcR2	%
	Share of Secondary and Tertiary Industries in GDP [1]	Industrial Diversity	+	EcR3	%
	Retail Sales per Capita [32]	Market Resilience	+	EcR4	CNY/person
Social Resilience	Population Density [33]	Human Capital	+	SoR1	persons/km2
	Savings Balance per Capita [31]	Household Risk Coping Capacity	+	SoR2	CNY/person
	General Public Budget Revenue per Capita [30]	Emergency Relief Affordability	+	SoR3	CNY/person
	Minimum Wage Standard [1]	Minimum Income Security	+	SoR4	CNY
	Number of Higher Education Institutions [1]	Talent Pool	+	SoR5	institutions
Facility Resilience	Road Network Density [34]	Transport Accessibility	+	InR1	km/km2
	Hospital Beds per Capita [32]	Healthcare System Capacity	+	InR2	units/person
	Social work agencies per capita [11]	Unemployment Mitigation	+	InR3	units/person
	Electricity Consumption per Capita [1]	Electricity Supply-Demand Balance	-	InR4	kWh/person
Ecological Resilience	Annual Average Net Primary Productivity (NPP) [17]	Vegetation Productivity	+	EnR1	-
	Annual mean FVC [33]	Vegetation Coverage	+	EnR2	%
	Annual mean PM2.5 [33]	Atmospheric Pollution	-	EnR3	μg/m3
	Annual CO2 emissions [35]	Low-Carbon Level	-	EnR4	t
	Green Space per Capita [1]	Vegetation Redundancy	+	EnR5	km2/person

Notes: “+” in the “Effect” column indicates that the indicator exerts a positive contribution to community resilience. The selection of all indicators is grounded in their established theoretical relevance.

).

2.4. Indicator Weighting

To ensure comparability among indicators and unify their directions (so that higher values consistently indicate greater resilience), this study applied a simplified min-max normalization method that simultaneously performs both standardization and positive/negative direction conversion prior to weighting. The final empowerment results are shown in

Table 2. Indicator weight.

Economic Resilience	EcR1	EcR2	EcR3	EcR4
0.2596	0.6208 (0.1612)	0.0101 (0.0026)	0.0261 (0.0068)	0.3430 (0.0891)
Social Resilience	SoR1	SoR2	SoR3	SoR4	SoR5
0.2552	0.1381 (0.0352)	0.0788 (0.0201)	0.1477 (0.0377)	0.0364 (0.0093)	0.5990 (0.1528)
Facility Resilience	InR1	InR2	InR3	InR4
0.1987	0.5752 (0.1143)	0.1026(0.0204)	0.3214 (0.0639)	0.0008 (0.0002)
Ecological Resilience	EnR1	EnR2	EnR3	EnR4	EnR5
0.2865	0.0643 (0.0184)	0.0146 (0.0042)	0.0090 (0.0026)	0.0005 (0.0002)	0.9116 (0.2612)

Note: Values in parentheses indicate the final weights derived from the integrated subjective and objective weighting approach.

.

Let the original data matrix be $X=\left[X_{ij}\right]$, where i = 1, …, n, denotes counties and j = 1, …, m, denotes indicators. Each indicator j is first classified by its direction $d_j$: $d_j$ = +1 for benefit indicators and $d_j$ = −1 for cost indicators. The normalization process is expressed as follows:

$$r_{ij}=\left\{\begin{array}{ll}{\displaystyle \frac{x_{ij}-{min}_i{x}_{ij}}{{max}_i{x}_{ij}-{min}_i{x}_{ij}}},& d_j=+1\\ {\displaystyle \frac{{max}_i{x}_{ij}-x_{ij}}{{max}_i{x}_{ij}-{min}_i{x}_{ij}}},& d_j=-1\end{array}\right.$$

(1)

where $r_{ij}\in \left[\mathrm{0,1}\right]$.

After normalization, the entropy weight method is applied to derive objective weights. The proportion of each indicator is calculated as:

$$p_{ij}={\displaystyle \frac{r_{ij}}{{\sum }_{i=1}^j{r}_{ij}}}$$

(2)

$$e_j={\displaystyle \frac{1}{lnn}}{\sum }_{i=1}^n{p}_{ij}ln{p}_{ij}$$

(3)

$$d_j=1-e_j$$

(4)

$$w_j^{\mathit{Entropy}}={\displaystyle \frac{d_j}{{\sum }_{j=1}^m{d}_j}}$$

(5)

To integrate the advantages of both subjective and objective weighting approaches, a combined weighting strategy was adopted, following established methodologies [36]. Specifically, the entropy weight method was applied to determine objective weights for individual indicators, while the analytic hierarchy process (AHP) was used to assign subjective weights to the criterion dimensions.

The AHP weights were derived from pairwise evaluations conducted by a panel of seven experts from academic, governmental, and industrial sectors specializing in urban–rural development. The consistency of expert judgments was tested using the Consistency Ratio (CR), calculated as $CR=CI/RI$, where CI is the consistency index and RI is the random index; all matrices satisfied CR < 0.1, confirming acceptable logical consistency. To integrate subjective and objective components, the final comprehensive weight $W_j$ for each indicator was obtained as:

$$W_j=\alpha W_j^{AHP}+(1-\alpha )W_j^{ENTROPY}$$

(6)

where $W_j^{AHP}$ and $W_j^{ENTROPY}$ represent the subjective and objective weights of indicator $j$, respectively, and $\alpha$ denotes the balance coefficient.

2.5. TOPSIS Method

The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) was employed to rank county-level resilience by measuring the relative closeness of each county to an ideal resilience state. The specific computational procedure was implemented as follows [37]:

$$R={\left(r_{ij}\right)}_{n*m}=z_j^{*}{ \times \left(x_{ij}^{\prime }\right)}_{n*m}=\left[\begin{array}{c}\begin{array}{cccc}{r}_{11}& r_{11}& \cdots & r_{1j}\\ r_{21}& r_{21}& \cdots & r_{2j}\\ ⋮& ⋮& ⋮& ⋮\\ r_{i1}& r_{i1}& \cdots & r_{ij}\end{array}\end{array}\right]$$

(7)

$$\begin{array}{c}{R}^{+}=\{{maxr}_{ij}\mid j=\mathrm{1,2}\cdots ,m\}=\{r_1^{+},r_2^{+},\cdots ,r_m^{+}\}\end{array}$$

(8)

$$R^{-}=\{{minr}_{ij}\mid j=\mathrm{1,2}\cdots ,m\}=\{r_1^{-},r_2^{-},\cdots ,r_m^{-}\}$$

(9)

$$D_i^{+}=\sqrt{{\sum }_{j=1}^m(R_j^{+}-r_{ij}{)}^2}$$

(10)

$$D_i^{-}=\sqrt{{\sum }_{j=1}^m(R_j^{-}-r_{ij}{)}^2}$$

(11)

$$T_i={\displaystyle \frac{D_i^{-}}{D_i^{+}+D_i^{-}}}$$

(12)

where R denotes the weighted normalized decision matrix; r_ij represents the weighted normalized value of indicator j in county i; n and m correspond to the total number of counties and indicators, respectively; $R^{+}$ and $R^{-}$ refer to the positive- and negative-ideal solutions of the county-level resilience system, respectively; $D_i^{+}$ and $D_i^{-}$ indicate the Euclidean distances from each county’s resilience level to the positive- and negative-ideal solutions, respectively; and $T_i$ denotes the comprehensive evaluation score derived using the TOPSIS method.

2.6. Spatial Cluster Analysis

The Hot Spot Analysis tool in ArcGIS (version 10.2) was employed to identify spatial clusters of high and low resilience values at the county level. A higher Z-score indicates a statistically significant clustering of high resilience values, whereas a lower Z-score suggests significant spatial aggregation of low resilience values. The Getis-Ord statistic is defined as follows [38]:

$$G_i^{*}={\displaystyle \frac{{\sum }_{i=1}^n{w}_{i,r}{T}_i-\overline{T}{\sum }_{i=1}^n{w}_{i,r}}{S\sqrt{\frac{[n{\sum }_{i=1}^n{w}_{i,r}^2-{\left({\sum }_{i=1}^n{w}_{i,r}\right)}^2]}{n-1}}}}$$

(13)

$$\overline{T}={\displaystyle \frac{{\sum }_{i=1}^n{T}_i}{n}}$$

(14)

$$S=\sqrt{{\displaystyle \frac{{\sum }_{i=1}^n{T}_i^2}{n}}-(\overline{T}{)}^2}$$

(15)

where $w_{i,r}$ denotes the spatial weight between counties i and r, $\overline{T}$ represents the mean value of county-level resilience, and S indicates the standard deviation of county-level resilience.

2.7. Binary Logistic Regression

The binary logistic regression model is widely used to examine urban–rural differences in resilience [11,26]. A critical prerequisite for applying this model is the clear and reasonable classification of urban and rural areas. Zheng et al. [39] suggested that municipal districts typically contain limited rural space, with minimal rural population and agricultural activity; hence, such districts can be approximated as urban areas when analyzing socioeconomic data. Accordingly, at the county level, this study classified county-level cities and municipal districts as urban areas, while counties and autonomous banners were designated as rural areas, reflecting differences in economic scale and structure. This binary classification served as the dependent variable in the model:

$$\mathit{ln}\left({\displaystyle \frac{P_i}{1-P_i}}\right)=\alpha +{\sum }_{k=1}^k{\beta }_k{X}_{ki}$$

(16)

where $X_{ki}$ denotes the resilience level of a specific subsystem in county i, and k represents the total number of subsystems; $P_i$ indicates the probability that a high-resilience area is classified as urban, while $1-P_i$ denotes the probability that a high-resilience area is classified as rural.

2.8. Bivariate Spatial Autocorrelation Analysis

The bivariate spatial autocorrelation model was applied to examine the spatial dependence and association between resilience levels and urban–rural types. The bivariate Moran’s I index was computed using GeoDa software (version 1.16.0.12) and visualized in ArcGIS. The formula is given as follows [40]:

$$L={\displaystyle \frac{{\sum }_{h=1}^n{\sum }_{k=1}^n{w}_{hk}(a_h-\overline{a})(b_k-\overline{b})}{S^2{\sum }_{h=1}^n{\sum }_{k=1}^n{w}_{hk}}}$$

(17)

$$L_h=z_h{\sum }_{k=1}^n{w}_{hk}{z}_k$$

(18)

where L denotes the bivariate global Moran’I index, and $L_h$ represents the bivariate local Moran’I index; $w_{hk}$ indicates the inverse-distance spatial weight matrix between county h and county k; $\overline{a}$ and $\overline{b}$ correspond to the mean values of the variables; $S^2$ signifies the variance of the total sample; and $z_h$ and $z_k$ refer to the variance-normalized values of county h and county k, respectively.

3. Results

3.1. Spatio-Temporal Patterns of County-Level Resilience

Figure 2 presents the descriptive statistics of the TOPSIS-based evaluation results for overall county-level resilience and its subsystem components across China from 2010 to 2021. Overall, county-level resilience, along with economic, social, and facility resilience, showed steady improvement, with facility resilience exhibiting the highest growth rate. As illustrated in Figure 3, since 2010, China has introduced a series of policy measures aimed at enhancing resilience, gradually forming a new pattern of urban–rural resilience development driven by multi-dimensional synergies.

Specifically, social resilience displayed the largest standard deviation, which increased year by year. Although the standard deviation of facility resilience was relatively small in 2010, it grew rapidly thereafter. These patterns reflect significant disparities in social resilience among counties and a widening gap in facility resilience over time. In contrast, ecological resilience remained largely unchanged throughout the study period, showing the slowest growth. This stagnation may be explained by the encroachment on ecological spaces, which likely hindered improvements in ecological resilience [41].

3.2. Spatio-Temporal Evolution of County-Level Resilience

Figure 4 shows the quantile classification map of county-level resilience and its subsystem resilience. Regions such as the Greater Khingan Mountains, the eastern fringe of the Tarim Basin, and western Sichuan consistently demonstrated the highest resilience levels, likely benefiting from abundant natural resources. The most pronounced improvement was observed in the southeastern coastal areas, where higher socioeconomic development and an early transition toward intensive growth modes have facilitated more complete social infrastructure and environmental protection systems. In contrast, the Taklamakan Desert region remained consistently at the lowest resilience tier. Additionally, a contiguous zone spanning the border of northeastern China and eastern Inner Mongolia was predominantly classified within the second-lowest resilience category, highlighting these as critical areas for future policy intervention.

Figure 4. Spatiotemporal hierarchy plots of county resilience and subsystems in China. Note: This map is based on the standard map (Review Map No. GS (2019) 1822) issued by the Ministry of Natural Resources of China. The same source applies to Figure 5 and Figure 6.

Figure 4. Spatiotemporal hierarchy plots of county resilience and subsystems in China. Note: This map is based on the standard map (Review Map No. GS (2019) 1822) issued by the Ministry of Natural Resources of China. The same source applies to Figure 5 and Figure 6.

Figure 5. Cold and hot spot mapping of county resilience in China.

Figure 5. Cold and hot spot mapping of county resilience in China.

Figure 6. Bivariate LISA cluster diagram of county resilience and county types in China.

Figure 6. Bivariate LISA cluster diagram of county resilience and county types in China.

Analysis of changes in resilience subsystems reveals that economic resilience improved markedly across most counties. The proportion of counties achieving the highest economic resilience tier increased from 7.78% to 34.43%, indicating that economic advancement served as a major endogenous driver of overall resilience improvement.

Regarding social resilience, the disparity between high-performing and low-performing regions generally narrowed over the decade. The share of counties in the lowest social resilience category decreased significantly from 69.81% to nearly 0%. Notably, among counties that consistently maintained the highest level of social resilience, 84.39% were county-level cities or municipal districts, which are spatially dispersed. Regions with relatively lower social resilience were concentrated in parts of Yunnan, Guangxi, and Hunan, as well as areas within the Taklamakan Desert.

Facility resilience showed pronounced spatial clustering. Counties with the most substantial improvements were predominantly located in a contiguous zone stretching from the southeastern coast to central China. In contrast, counties with slower development were largely found in remote regions such as Xinjiang, the Greater Khingan Mountains, and Heilongjiang.

Ecological resilience exhibited minimal change in spatial pattern throughout the study period. High-performing counties were primarily distributed in a belt-like formation along the southern coast, western Yunnan, and western Sichuan, with additional clusters in the Greater Khingan Mountains and the eastern Tarim Basin. A large contiguous area extending from the Taklamakan Desert through the Turpan Basin to the Hexi region consistently demonstrated the lowest level of ecological resilience.

3.3. Spatial Aggregation Effects of County-Level Resilience

As shown in the cold and hot spot distribution in Figure 5, the spatial pattern of high-resilience clusters—extending from the eastern Tarim Basin and the Turpan Basin to the Hexi Corridor in Gansu and further to western Sichuan—exhibited an eastward shift and a gradual reduction in areal extent. Similarly, the Greater Khingan Mountains, which consistently exhibited high resilience values, also showed a contracting trend. In contrast, the Yangtze River Delta region demonstrated a longitudinally expanding and coordinated development pattern of county-level resilience. By 2021, the eastern coastal area had become the largest agglomeration of high-resilience counties.

Conversely, low-resilience clusters were primarily distributed across southern North China, eastern Northwest China, and parts of Central and South China, with a discernible westward expansion trend. Persistent low-resilience agglomerations were observed in Northeast China and the Taklamakan Desert region, which have been gradually expanding eastward. Harsh natural conditions and resource depletion have hindered improvements in resilience, sustaining the characteristics of a development “depression” in these regions [42].

3.4. Urban–Rural Differentiation in County-Level Resilience

All 18 indicators of county-level resilience exhibited variance inflation factors (VIF) below 10, indicating no significant multicollinearity. Overall, as shown in

Table 3. Binary logistic regression results at the criterion layer.

Variable	B	Categories of High-Resilience Counties
	(S.E.)
County-Level Resilience	0.94 ***	Urban
	(0.02)
Constant	Yes
Sample Size	24,048
Nagelkerke R2	0.17
Model sig	0.00
Variable	B	Categories of High-Resilience Counties
	(S.E.)
Economic Resilience	1.64 ***	Urban
	(0.04)
Social Resilience	0.59 ***	Urban
	(0.03)
Facility Resilience	−0.21 ***	Rural
	(0.02)
Ecological Resilience	−0.23 ***	Rural
	(0.03)
Constant	Yes
Sample Size	24,048
Nagelkerke R2	0.36
Model sig	0.00

Note: *** p < 0.001.

, the regression coefficient for county-level resilience was significantly positive, suggesting that urban areas possess markedly higher overall resilience than rural regions. At the subsystem level, economic and social resilience were significantly higher in urban settings, whereas infrastructural and ecological resilience were greater in rural areas. These results reflect substantial disparities in resilience strengths between urban and rural China and highlight a consistent pattern of imbalanced development in regional resilience.

To identify the key drivers of urban–rural resilience disparities, a binary logistic regression model was constructed at the indicator level. The results indicated that 17 out of the 18 selected indicators were statistically significant at the 1% level, with the exception of “annual average NPP”.

Table 4. Binary logistic regression results at the indicator layer.

Variable	B	Categories of High-Resilience Counties
	(S.E.)
GDP	1.48 ***	Urban
	(0.05)
Savings per Capita	0.73 ***	Urban
	(0.04)
Minimum Wage Standard	−0.59 ***	Rural
	(0.03)
Number of Higher Education Institutions	0.44 ***	Urban
	(0.03)
Share of Secondary and Tertiary Industries in GDP	0.40 ***	Urban
	(0.03)
Other Indicators	Yes
Constant	Yes
Sample Size	24,048
Nagelkerke R2	0.42
Model sig	0.00

Note: *** p < 0.001. Note: Following established methodologies [11], all independent variables were standardized using the z-score method prior to regression analysis to ensure comparability of regression coefficients.

reports the five indicators with the largest regression coefficients. GDP exhibited the highest coefficient, followed by savings per capita, suggesting that economic disparity is the most prominent factor differentiating urban and rural resilience and that enhancing rural economic development is crucial for achieving balanced resilience. This conclusion is further supported by the significant positive influence of the minimum wage standard on rural resilience.

Table 5. Differences in urban–rural resilience on both sides of the Hu’s Line.

	Northwestern Part of China		Southeastern Region
Variable	B	Categories of High-Resilience Counties	B	Categories of High-Resilience Counties
	(S.E.)		(S.E.)
County-Level Resilience	0.48 ***	Urban	1.26 ***	Urban
	(0.03)		(0.03)
Constant	Yes		Yes
Sample Size	6840		17,208
Nagelkerke R2	0.07		0.22
Model sig	0.00		0.00
Variable	B		B
	(S.E.)		(S.E.)
Economic Resilience	1.67 ***	Urban	1.60 ***	Urban
	(0.09)		(0.04)
Social Resilience	1.15 ***	Urban	0.43 ***	Urban
	(0.08)		(0.04)
Facility Resilience	−0.06	n.s.	−0.23 ***	Rural
	(0.05)		(0.03)
Ecological Resilience	−0.17 ***	Rural	−0.35 ***	Rural
	(0.03)		(0.02)
Constant	Yes		Yes
Sample Size	6840		17,208
Nagelkerke R2	0.37		0.35
Model sig	0.00		0.00

Note: *** p < 0.001.

presents the urban–rural disparities in county-level resilience between the southeastern and northwestern segments of China, demarcated by the Hu Line. The results indicate that urban resilience was significantly higher than rural resilience on both sides of the line. A larger regression coefficient in the eastern segment suggests a more pronounced urban–rural disparity there compared to the western region.

Regarding resilience subsystems, the patterns of urban–rural in economic, social, and ecological resilience were generally consistent with the national trend. However, a distinct east–west divergence was observed in facility resilience: rural areas in the east exhibited significantly higher facility resilience than their urban counterparts, whereas no significant difference was found in the west.

Table 6. Differences in urban–rural resilience among the seven major geographical regions in China.

	Northwest China	Northeast China	North China	Central China	East China	South China	Southwest China
Variable	B	B	B	B	B	B	B
	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)
County-Level Resilience	0.29 ***	0.85 ***	0.86 ***	0.72 ***	1.13 ***	2.21 ***	1.59 ***
	(0.04)	(0.11)	(0.05)	(0.07)	(0.05)	(0.14)	(0.06)
Constant	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample Size	3456	1812	3936	3156	4920	2040	4728
Nagelkerke R2	0.03	0.07	0.17	0.07	0.26	0.28	0.38
Model sig	0.00	0.00	0.00	0.00	0.00	0.00	0.00
Variable	B	B	B	B	B	B	B
	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)	(S.E.)
Economic Resilience	2.83 ***	1.20 ***	1.51 ***	2.81 ***	1.55 ***	2.24 ***	2.30 ***
	(0.19)	(0.14)	(0.11)	(0.13)	(0.06)	(0.18)	(0.15)
Social Resilience	1.23 ***	0.35 ***	1.32 ***	−0.54 ***	0.1 ***	1.98 ***	1.13 ***
	(0.14)	(0.12)	(0.12)	(0.07)	(0.04)	(0.20)	(0.09)
Facility Resilience	−0.20 **	0.41 ***	0.004	−0.59 ***	−0.24 ***	−0.61 ***	−0.52 ***
	(0.09)	(0.11)	(0.05)	(0.08)	(0.05)	(0.10)	(0.08)
Ecological Resilience	−0.52 ***	−0.06	−0.13 ***	−0.42 ***	−0.71 ***	−2.79 ***	−0.56 ***
	(0.10)	(0.05)	(0.04)	(0.25)	(0.15)	(0.34)	(0.11)
Constant	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample Size	3456	1812	3936	3156	4920	2040	4728
Nagelkerke R2	0.44	0.26	0.41	0.37	0.41	0.50	0.56
Model sig	0.00	0.00	0.00	0.00	0.00	0.00	0.00

Note: *** p < 0.001, ** p < 0.01.

displays urban–rural disparities in county-level resilience across seven geographic regions. Urban resilience was significantly higher than rural resilience in all regions, with the disparity increasing in the following order: Northwest, Central, Northeast, North, East, Southwest, and South China.

The most pronounced urban–rural resilience gap was observed in South China. Regression results at the subsystem level revealed that economic divergence is the primary driver of this disparity. Furthermore, South China also exhibited the largest urban–rural difference in social resilience.

In contrast, the disparity in facility resilience was relatively modest. Notably, and unlike the national trend, facility resilience in Northeast China was significantly higher in urban than in rural areas. This pattern can be attributed to the region’s high urbanization level coupled with substantial outmigration [5].

3.5. Spatial Coupling Between County-Level Resilience and Urban–Rural Types

Results from the bivariate global spatial autocorrelation analysis indicate that the global Moran’s I values between county-level resilience and urban–rural type were 0.04 and 0.06 in 2010 and 2021, respectively, suggesting not only that urban areas generally exhibit higher resilience than rural areas, but also that this disparity widened over time.

As shown in the bivariate LISA cluster map in Figure 6, four distinct types of spatial coupling between resilience and urban–rural type were identified:

High Resilience–Urban type: Areas with high resilience surrounded predominantly by urban regions. These are mainly concentrated in the Chengdu-Chongqing and Yangtze River Delta economic zones, with scattered distributions in Northeast China. Characterized by flat terrain and significant socioeconomic advantages, these regions benefit from early development, well-established public infrastructure, and environmental protection systems, contributing to their high resilience.

Low Resilience–Rural type: Areas with low resilience predominantly embedded in rural surroundings. These are extensively distributed across contiguous regions in Shanxi, Shanxi, Ningxia, and Gansu, as well as parts of Yunnan, Guizhou, and Guangxi. The clustering of rural settlements and relatively underdeveloped industrial structures in these areas constrain resilience growth.

Low Resilience–Urban type: Low-resilience areas located mainly within urban agglomerations. These are primarily observed in the Chengdu-Chongqing economic circle, with additional occurrences in Northeast and Eastern Coastal China. Despite high urban density, economic, social, infrastructural, and natural resources in these regions are insufficient relative to their large populations, hindering resilience improvement.

High Resilience–Rural type: High-resilience areas situated within predominantly rural contexts. This pattern is identified in the eastern Tarim Basin and Western Sichuan. Low population density and abundant resource redundancy in these regions support notably high rural resilience.

4. Discussion

4.1. Integrated Drivers and Multidimensional Structure of Urban–Rural Resilience Disparities

Analysis indicates that although county-level resilience in China showed overall improvement from 2010 to 2021, pronounced urban–rural disparities persisted, closely linked to economic concentration and population siphon effects [43]. Urban areas demonstrated clear advantages in economic and social resilience, consistent with existing literature. For instance, Liu [44] highlighted that economic agglomeration and dense social capital are key factors enhancing urban resilience. However, most previous studies have focused on economic and demographic clustering at provincial [43,45] or city levels [46], paying insufficient attention to intra-county urban–rural differences—particularly the comparative advantages of rural areas within the infrastructural and ecological subsystems.

By directly comparing urban and rural subsystem performance, this study revealed that certain rural regions, especially those with favourable ecological conditions or strong policy support, exhibited relative strengths in infrastructural and ecological resilience. This finding not only complements the perspective on urban–rural divergence but also underscores the complex and multidimensional nature of resilience within different regional contexts. Temporally, although the urban–rural gap persists overall, the sustained enhancement of ecological and infrastructural resilience in some rural areas provides new evidence that policy interventions and imbalanced resource allocation [47,48] are critical drivers, not only reinforcing socioeconomic disparities but also shaping the distinctive resilience profiles of rural subsystems. In summary, the drivers of urban–rural disparities are shown to operate through a multidimensional structure: sustained by fundamental socioeconomic forces while being modulated by policy-driven improvements in specific rural subsystems.

The observed divergence in resilience patterns—specifically, the coexistence of “high-resilience rural” and “low-resilience urban” areas—reflects underlying socioeconomic and institutional factors. This pattern demonstrates variations in policy prioritization and governance effectiveness across different spatial and administrative contexts. In certain rural counties, adaptive capacity has been strengthened through livelihood diversification, strong community cohesion, and targeted policy interventions such as the Rural Revitalization Strategy and Ecological Conservation Redlines, which collectively enhance environmental quality and public service provision [49]. Conversely, in rapidly urbanizing zones, industrial expansion and population growth have strained governance capacity and accelerated ecological degradation, leading to relative resilience decline [50].

4.2. Spatial Heterogeneity, Typologies, and Underlying Mechanisms

Spatially, urban–rural resilience disparities in China exhibit pronounced geographical heterogeneity, with the most marked contrasts observed in the southeastern region relative to the northwestern half of the Hu Line, particularly in South China. While this finding partially aligns with previous studies emphasizing macro-scale regional divides such as “east-central-west” [51] or “coastal-inland” [52] gradients, it further provides a more refined spatial delineation leveraging county-level data.

The underlying mechanisms for these spatial patterns are multifaceted. The distinct east–west divergence in infrastructural resilience underscores the role of demographic and policy drivers. This discrepancy may be attributed to China’s persistently uneven population distribution [47]. In the eastern region, the siphon effect of large central cities has led to highly concentrated urban populations, thereby reducing infrastructural redundancy [48]. In contrast, policy initiatives such as the “Western Development Campaign” have prioritized comprehensive improvements in urban infrastructure in the western region, contributing to more balanced urban–rural facility resilience.

Furthermore, the pronounced multi-dimensional urban–rural gap in South China can be interpreted through interconnected socioeconomic and environmental mechanisms. As the most active region for population mobility in China, South China experiences substantially higher urban immigration than rural areas [53], resulting in pronounced demographic and economic imbalances that hinder coordinated social development. Simultaneously, continuous population growth and urban expansion accelerate land urbanization and diminish ecological space [39], primarily accounting for the marked ecological resilience gap.

By integrating coupling analysis with urban–rural typologies, this study identifies specific spatial types—including “high-resilience rural” and “low-resilience urban” areas—that deviate from conventional urban–rural dichotomies. These patterns highlight the complexity of resilience landscapes and offer an empirical basis for precisely targeting policy interventions in critical zones.

The distinct southeast-northwest resilience divide reflects the enduring influence of national development strategies and structural economic disparities. Southeastern China, as the primary beneficiary of the Reform and Opening-up Policy and Coastal Development Strategy, has accumulated stronger fiscal capacity, diversified industrial bases, and advanced infrastructure networks [54]. Institutional adaptability and technological diffusion have been reinforced through sustained investment in public services, environmental governance, and innovation-driven growth. Conversely, the northwestern region—despite support from the Western Development Campaign and ecological restoration programs—remains constrained by fragile ecosystems, limited market connectivity, and high dependence on resource-based industries [55]. Weaker fiscal autonomy and lower policy implementation capacity at the local level have further hindered resilience building. These spatial patterns demonstrate how China’s regionally differentiated policy system and uneven economic development have collectively shaped resilience outcomes, reflecting both historical policy legacies and varying regional capacities to transform policy inputs into adaptive gains.

4.3. Subsystem Variations and Theoretical Implications

The analysis at the subsystem level reveals critical variations that enrich the theoretical understanding of regional resilience. The general consistency of economic, social, and ecological resilience disparities with the national trend underscores the pervasive influence of macro-level drivers. However, the exception of infrastructural resilience in the east, where rural areas outperformed urban ones, indicates that the mechanisms of resilience formation are subsystem-dependent.

This finding necessitates a refinement of resilience theory. It suggests that the attributes of resilience (e.g., robustness, redundancy, efficiency) may be configured differently across subsystems and spatial contexts. The observed rural advantage in eastern infrastructural resilience could be attributed to newer, more decentralized infrastructure networks, as opposed to the potentially overburdened and centralized systems in dense urban cores [56]. This perspective shifts the focus from a monolithic view of resilience to a networked, context-specific one, aligning with emerging literature on the interdependencies of infrastructure systems.

5. Summary

The main findings of this study indicate that county-level resilience in China exhibited a steady increasing trend from 2010 to 2021, with notable improvements in economic, social, and infrastructural dimensions. Nevertheless, significant regional and urban–rural disparities persist. Overall, urban areas demonstrated stronger economic and social resilience, while rural regions showed comparative advantages in infrastructural and ecological resilience. These disparities were particularly pronounced in the southeastern part of the Hu Line, with South China representing the most evident case. Furthermore, coupling analysis between urban–rural typology and resilience level revealed characteristic patterns such as “high-resilience rural” and “low-resilience urban” types, thereby enriching the perspective of existing research on regional resilience.

Based on the results, first, priority should be given to upgrading critical rural infrastructure through targeted funding for digital connectivity, clean energy networks, and decentralized water-supply systems to close persistent public service gaps. Economic resilience should be strengthened by fiscal-guided funds that support the development of locally distinctive industrial chains and foster endogenous growth capacity within rural communities. Second, in urban areas, enhancing ecological redundancy through sponge-city retrofitting and the construction of green infrastructure networks can improve climate adaptability in high-density built environments. In rural regions, market-oriented ecological compensation mechanisms should be established to transform natural capital advantages—such as forests, grasslands, and watersheds—into sustainable development assets, thus linking ecological restoration with livelihood security. Third, urban–rural coordination policies to promote factor mobility. Functional complementarity should be strengthened by developing shared infrastructure platforms and extending high-quality public services—such as education and healthcare—from cities to surrounding rural areas. Furthermore, resilience network construction should be incorporated into territorial spatial planning by coordinating the layout of emergency supply depots and evacuation corridors, thereby fostering regional resilience communities capable of joint crisis response. Collectively, these measures provide a practical roadmap for enhancing resilience in alignment with China’s long-term goals of urban–rural integration and sustainable regional development.

Several limitations should be acknowledged. First, the reliance on county-level statistical data, though suitable for systematic nationwide comparison, inevitably introduces the Modifiable Areal Unit Problem (MAUP) and may obscure micro-level variations in governance, infrastructure, and community adaptive capacity—particularly between peri-urban fringes and remote rural settlements. Incorporating multi-source and higher-resolution data in future research could help capture these finer spatial differences and reveal more localized resilience mechanisms. Second, this study primarily provides a descriptive diagnosis rather than a causal explanation of resilience dynamics. The complex interactions between resilience evolution and external shocks—such as major natural disasters, policy interventions, and urban–rural linkages—require deeper investigation. Future studies should employ explicit causal modeling approaches to examine these feedback mechanisms, thereby strengthening the explanatory power of resilience research and enhancing its policy relevance in guiding resilience building across diverse urban–rural contexts.