Assessing Rural Development Vulnerability Index: A Spatio-Temporal Analysis of Post-Poverty Alleviation Areas in Hunan, China

Abstract

Rural post-poverty alleviation areas are not on a solid developmental footing and therefore remain at risk of returning to poverty in the midst of rapid urbanization. Vulnerability assessment of socio-ecological systems is critical for identifying risks and enhancing resilience in rural areas transitioning out of poverty. Based on research data from 2012, 2017, and 2022 in the post-poverty alleviation areas of Hunan Province, this research establishes a Vulnerability-Scoping-Diagram (VSD) assessment framework for rural development vulnerability and Spatially-Explicit-Resilience-Vulnerability (SERV) analysis model from a socio-ecological system perspective. It comprehensively analyzes the spatial and temporal variations of the Rural Development Vulnerability Index (RDVI) in the study area. Geodetector is used to explore the main factors influencing the spatial and temporal variability of RDVI, and vulnerability type zones are classified by combining the dominant elements method. The findings indicate that: (1) The rural development vulnerability index of post-poverty alleviation areas in Hunan Province has obvious characteristics of spatial and temporal differentiation. The RDVI in western Hunan and southern Hunan is always high, while the RDVI in ChangZhuTan and Dongting Lake regions decreases year by year. (2) The RDVI of post-poverty alleviation areas in Hunan Province is determined by the three dimensions of exposure, sensitivity, and adaptability, exhibiting significant spatial and temporal variations. (3) Spatial autocorrelation analysis showed that areas with similar rural socio-ecological vulnerability in post-poverty alleviation areas of Hunan Province were significantly clustered spatially. (4) The core influencing factors of RDVI in Hunan’s post-poverty alleviation areas have shifted from natural disaster risk to multiple risk dimensions encompassing social resource load and ecological environment risk superimposition, resulting in more complex and diversified influencing factors. (5) By combining results from the RDVI assessment with the dominant elements method, the regions can be classified into multiple vulnerability type districts dominated by multiple elements or single-element dominance, leading to corresponding development suggestions. The study aims to examine the process of changes in vulnerability within rural development in post-poverty alleviation areas and its causal factors from a socio-ecological system perspective. This will provide a foundation for policy formulation to consolidate the results of post-poverty alleviation and promote the sustainable development of rural areas.

1. Introduction

Poverty and environmental issues persist as challenges in human society, and addressing poverty is a crucial step in achieving sustainable human development [1]. Ending poverty in all its forms worldwide is the primary objective among the 17 Sustainable Development Goals (SDGs) established by the United Nations for 2030 [2]. China has undergone four main stages in its poverty alleviation efforts (

Table 1. Four main stages of poverty alleviation in China after 1949.

Period	Characteristics	Key Measures
1949–1977	Relief-based poverty alleviation	Provided fiscal subsidies and material relief to ensure minimum living standards for impoverished populations.
1978–2011	Promote poverty alleviation through institutional reform and solve the problem of food and clothing for the poor.	Implementing the household contract responsibility system in rural areas and promoting the reform of township enterprises. Carrying out regional poverty alleviation initiatives, such as the Western Development Strategy.
2012–2020	Targeted governance with clear objectives	Accurately identify people living in poverty. Based on local conditions and individual circumstances, develop distinctive industries, implement relocation poverty alleviation plans, strengthen education poverty alleviation, etc.
2021–Present	In the post-poverty alleviation period, consolidating poverty alleviation achievements and effectively linking them with rural revitalization	Strengthen dynamic monitoring of low-income populations. Implement consumption assistance. Implement employment promotion measures to prevent return to poverty.

). In the early stages after the founding of the People’s Republic of China in 1949, the main approach was relief-based poverty alleviation. Since 1978, China has established specialized agencies to implement systematic and extensive poverty alleviation efforts in rural areas [3]. The early 21st century witnessed significant progress in global poverty reduction, prompting a shift in research focus from absolute to relative poverty [4,5]. During the 2020s, China successfully eradicated region-based absolute poverty and transitioned into a new development phase characterized by secondary poverty emergence [6]. However, previously impoverished areas remain at risk of regression, particularly in rural regions lacking firmly established foundations for sustainable development [7,8,9]. Urgent enhancement of risk recognition and understanding is required to implement effective measures that prevent regression of poverty alleviation efforts and consolidate gains.

Vulnerability studies represent a prominent focus in global sustainability science [10,11] and are increasingly applied in socio-ecological system sustainability research [12,13,14]. Poverty originates in socio-ecological system vulnerability, establishing an inextricable linkage [15]. The majority of villages in post-poverty alleviation areas are characterized by unique natural conditions that directly and profoundly influence the area’s land use, ecological characteristics, socio-economic patterns, and vulnerability issues, which tend to be more diverse and complex than in other areas [8,16]. Although the existing socio-ecological perspective provides valuable insights for vulnerability research, there is little research on the spatial characteristics and impact mechanisms of vulnerability under the combined influence of the complex natural ecological environment and socio-economic conditions in post-poverty alleviation areas, which requires further exploration. Therefore, by approaching ecological, social, and economic issues as the starting point, conducting vulnerability research on post-poverty alleviation areas from a socio-ecological systems (SES) perspective, and re-evaluating the process of changes in rural development vulnerability in post-poverty alleviation areas as well as its underlying causes can establish a theoretical basis for enhancing the region’s resilience and self-development capabilities. This is of significant practical importance for consolidating post-poverty alleviation achievements and further advancing comprehensive rural revitalization.

This study investigated the post-poverty alleviation areas of Hunan province in China from 2012 to 2022 with the county as the basic research unit. The main research objectives are: (1) Constructing a rural development vulnerability index (RDVI) based on a socio-ecological perspective to characterize spatio-temporal evolution patterns of rural development vulnerability; (2) Identifying spatial agglomeration characteristics of RDVI using spatial autocorrelation analysis; (3) Determining core influencing factors and their temporal dynamics via Principal Component Analysis (PCA) and geodetector; (4) Classifying rural vulnerability into different types. Accordingly, this research addresses three core questions: (1) What spatial distribution and evolutionary trends characterize RDVI in the post-poverty alleviation areas of Hunan Province? (2) What spatial aggregation patterns emerge for RDVI? (3) What key factors drive RDVI, and how do their influences evolve temporally? This study aims to address the above questions, reveal the spatial changes in rural development vulnerability in poverty-stricken counties, identify the core factors influencing vulnerability, and classify different vulnerability zones. The findings are intended to provide reference for local governments in advancing rural revitalization and governance efforts during the post-poverty alleviation period.

2. Literature Review and Framework Presentation

2.1. SES Perspective

Since the beginning of the 21st century, rural development in China has faced significant challenges, encompassing demographic issues [17], infrastructure development [18], ecosystem degradation [19], and environmental pollution [20]. These challenges involve economic, social, ecological, and cultural dimensions [21]. Poverty-stricken areas, as “natural-economic-social” complex systems [22], exhibit heightened vulnerability due to the direct impact of unique natural conditions on land use, ecological patterns, and socio-economic structures [8]. While China’s precision poverty alleviation strategy eliminated absolute rural poverty, post-poverty areas still struggle with sustainable development [23], as existing research predominantly focuses on livelihood risks [24] rather than the diverse, complex interplay of economic, social, and ecological factors [25]. Therefore, there is a need for a theoretical perspective that can comprehensively consider a variety of factors to analyze the sustainable development of rural post-poverty alleviation areas.

SES theory provides a critical framework to address this limitation. Driven by globalization and technological advancement, human activities increasingly interconnect social and ecological subsystems, amplifying their coupled effects [26,27]. SES theory emphasizes the interdependence between human societies and natural resources: socio-economic activities rely on ecosystem services but also impact ecological health, rendering single-subsystem analyses inadequate [28]. Conceptualized as complex adaptive systems by Holland [29], SES integrate social, economic, and ecological components as tightly intertwined yet semi-autonomous entities. Scholars like Cumming [30] and Walker [31] further define SES as human-earth symbiosis systems, highlighting the bidirectional interactions between socio-cultural practices and environmental processes [32,33].

For post-poverty alleviation areas, SES theory offers a holistic lens to analyze vulnerability. By treating social and ecological elements as an integrated system, it captures their inseparable wholeness and mutual influence, aligning with the multi-dimensional complexity of post-poverty challenges. This perspective enables more accurate identification of potential risks and mechanisms, providing a theoretical foundation for enhancing rural resilience and guiding sustainable development policies.

2.2. Poverty-Returning Vulnerability

Global poverty has garnered growing international attention, prompting scholars to investigate its economic causes across disciplines to explain persistent cycles of poverty [9,34,35,36]. Since 2000, “poverty-returning”—a phenomenon where individuals, households, regions, or even communities that have escaped poverty relapse into poverty due to various shocks—has emerged as a critical research focus [37,38,39]. This dynamic process is characterized by uncertainty and recurrence, rooted in incomplete or unsustainable poverty exit [40,41,42]. Existing studies address causes, early-warning systems, and governance [43,44,45,46,47,48] but inadequately capture the complex socio-ecological risks and mechanisms driving this phenomenon.

Vulnerability, a cross-disciplinary concept, refers to a system’s inherent propensity to suffer damage from exposure to risks, shaped by its sensitivity and adaptive capacity [49,50]. This framework aligns closely with poverty studies [51,52,53,54,55], as the poor are both more exposed to risks [52] and lack resources to mitigate them [55]. The World Bank defines poverty vulnerability as the poor’s resilience to shocks [56], while scholars like Azeem and Chiwaula characterize it as the probability of falling into poverty due to risk exposure [52,57]. On this basis, analytical frameworks such as the exposure-sensitivity-adaptability analytical framework [58], the vulnerability-sustainable livelihoods analytical framework [59,60], and the sensitivity-adaptability framework [61] are widely applied in poverty vulnerability studies.

Following China’s eradication of absolute poverty, research focuses on “vulnerability to poverty return”—a critical concept defined as the risk of regression to poverty thresholds in post-poverty alleviation areas due to adverse environmental shocks, social transformations, and insufficient resilience of SES to mitigate such risks [62]. While recent studies have explored this in China [51,63,64], research on the spatio-temporal differentiation and driving mechanisms of socio-ecological vulnerability in post-poverty rural areas remains underdeveloped, hindering accurate predictions of poverty dynamics, and still needs to be further improved.

2.3. Methodological Frameworks in Vulnerability Studies

Vulnerability research methods are diverse and interdisciplinary, and mathematical models and statistical methods are commonly used to conduct research. Currently, the commonly used models in vulnerability research are the R-H (Risk-Hazard) model [65,66], HOP (Hazard-of-Place) model [67,68], PAR (Pressure-and-Release) model [69,70], PSR (Pressure-State-Response) model [71,72], VSD (Vulnerability-Scoping-Diagram) model [58,73,74], SERV (Spatially-Explicit-Resilience-Vulnerability) model [75], and so on. Different vulnerability models have different research focuses and characteristics (Table 2). Among them, the VSD model explicitly defines systemic vulnerability as three dimensions: exposure, sensitivity and adaptability [58], and this framework has been widely used in vulnerability assessment where multiple elements are intertwined and multiple risks are intertwined in their impacts [73], which provides a basis for the study of complex and multifaceted vulnerability of SES in post-poverty alleviation areas.

Based on previous research, the VSD model provides a more explanatory framework for studying the vulnerability of rural development in poverty-stricken areas from a SES perspective compared to other vulnerability models. Firstly, the vulnerability of rural development in the post-poverty alleviation area is the result of the combined effects of exposure, sensitivity, and adaptive capacity [6]. Additionally, the VSD model’s advantage in studying the nonlinear interactions of vulnerability aligns well with the emphasis on system integrity in the SES perspective, which is more helpful for understanding the mechanisms influencing the vulnerability of rural development in the post-poverty alleviation area [73]. Existing research on the VSD model is primarily conducted at the regional scale, which lays the foundation for this study’s analysis of county-level units at the provincial scale [74,76]. Furthermore, this study draws on Chen et al.’s research [77] to incorporate the SERV model into the VSD vulnerability research framework. By integrating geospatial data with statistical techniques, the study achieves the visualization and mapping of the Rural Development Vulnerability Index (RDVI) (Figure 1).

Table 2. Comparison of methodological models and their characteristics related to vulnerability research.

Model Name	Characteristics	References
R-H (Risk-Hazard)	Emphasizes multi-causal interactions between hazard drivers and exposure elements, focusing on hazard-consequence linkages and systemic complexity.	Burton, 1993 [66]; Costa, 2013 [65]
PAR (Pressure-and-Release)	Examines vulnerability dynamics under imbalanced societal pressures and institutional responsiveness, decoding systemic fragility formation.	Fadigas, 2017 [70]; Huelssiep, 2021 [69]
PSR (Pressure-State-Response)	A sustainable development assessment framework for ecological vulnerability based on the Pressure-State-Response model, with emphasis on studying linear pressure and its corresponding response mechanisms.	Talukdar, 2020 [71]; Zhang, 2023 [72]
HOP (Hazard-of-Place)	Assesses coupled natural-socioenvironmental impacts on regional vulnerability through risk-quantified spatial analytics.	Frigerio, 2016 [67]; Guo, 2021 [68]
VSD (Vulnerability-Scoping-Diagram)	The model breaks vulnerability down into three parts: “exposure, sensitivity, and adaptability,” integrates and constructs vulnerability assessment indicators, and considers nonlinear interactions.	Polsky, 2007 [58]; Nicholas, 2012 [73]; Cao, 2022 [74]
SERV (Spatially-Explicit-Resilience-Vulnerability)	Conducting vulnerability assessment and characterization of spatial systems affected by natural and anthropogenic factors within defined geographical areas. This methodology integrates geospatial data with statistical techniques to enable vulnerability pattern mapping.	Frazier, 2014 [75]; Chen, 2018 [77]

Table 2. Comparison of methodological models and their characteristics related to vulnerability research.

Model Name	Characteristics	References
R-H (Risk-Hazard)	Emphasizes multi-causal interactions between hazard drivers and exposure elements, focusing on hazard-consequence linkages and systemic complexity.	Burton, 1993 [66]; Costa, 2013 [65]
PAR (Pressure-and-Release)	Examines vulnerability dynamics under imbalanced societal pressures and institutional responsiveness, decoding systemic fragility formation.	Fadigas, 2017 [70]; Huelssiep, 2021 [69]
PSR (Pressure-State-Response)	A sustainable development assessment framework for ecological vulnerability based on the Pressure-State-Response model, with emphasis on studying linear pressure and its corresponding response mechanisms.	Talukdar, 2020 [71]; Zhang, 2023 [72]
HOP (Hazard-of-Place)	Assesses coupled natural-socioenvironmental impacts on regional vulnerability through risk-quantified spatial analytics.	Frigerio, 2016 [67]; Guo, 2021 [68]
VSD (Vulnerability-Scoping-Diagram)	The model breaks vulnerability down into three parts: “exposure, sensitivity, and adaptability,” integrates and constructs vulnerability assessment indicators, and considers nonlinear interactions.	Polsky, 2007 [58]; Nicholas, 2012 [73]; Cao, 2022 [74]
SERV (Spatially-Explicit-Resilience-Vulnerability)	Conducting vulnerability assessment and characterization of spatial systems affected by natural and anthropogenic factors within defined geographical areas. This methodology integrates geospatial data with statistical techniques to enable vulnerability pattern mapping.	Frazier, 2014 [75]; Chen, 2018 [77]

Therefore, this study applies a SES perspective, integrated with the VSD vulnerability framework and SERV model, to analyze the spatio-temporal vulnerability characteristics and influence mechanisms governing rural development in post-poverty alleviation areas. This re-examination of socio-economic-ecological transitions and their drivers enriches theoretical foundations for sustainable development in post-poverty alleviation areas and advances practical implementation of rural revitalization strategies.

3. Materials and Methods

3.1. Study Area

The study area of this study is the post-poverty alleviation area of Hunan Province, China (Figure 2). Located in central and western China, the post-poverty alleviation area of Hunan Province contains 51 poverty-alleviating counties in 11 prefectural-level cities and one autonomous prefecture, of which 40 are national-level poverty-alleviating counties and 11 are provincial-level poverty-alleviating counties. This region is a typical composite area of “ethnic minority areas, border areas, and poverty-stricken areas” in central and western China.

The terrain in the region is complex, with various terrain types such as semi-high mountains, low mountains, hills, basins, and plains. The area of mountains accounts for 68% of the total area, and karst landforms are widely distributed. The climate varies greatly from year to year, with colder winters, hotter and more humid summers, rainy springs and summers, and dry autumns and winters, among which the vertical changes in the climate of the mountainous areas of western Hunan and southern Hunan are the most obvious, with the average annual temperatures mostly ranging from 16 to 19 degrees Celsius.

The latest data from the Seventh Population Census Report of Hunan Province shows that as of 2020, the resident population of the post-poverty alleviation areas in Hunan Province was 23.06 million, accounting for 34.71% of the total population of Hunan Province. The rural population is predominantly rural, with 12.58 million people, accounting for 54.55% of the total regional population. The urbanization rate of the study region is 45.45%, lower than the provincial average (60.3%). The aging level of the region is high, with people aged 65 and above accounting for 15.46% of the total resident population. The study area is characterized by multi-ethnic culture and ecology, with one ethnic minority autonomous prefecture and seven ethnic minority autonomous counties, and 24 counties in which the proportion of ethnic minority population exceeds half of the total population of the administrative division.

In terms of economy and industry, the post-poverty alleviation areas in Hunan Province are dominated by agriculture, with arable land accounting for 18.7% of the total land area in the region, and the leading industries include rice, tea, citrus, etc. The industrial base of the study area is relatively weak. The industrial base of the study area is relatively weak, with the number of industrial enterprises above large scale only 45% of the provincial average, and primary industries such as the processing of agricultural products are dominant.

3.2. Data Sources

The data and information in this study mainly include remote sensing images, DEM, land use type data, meteorological and hydrological data, natural disaster data, and socio-economic data of post-poverty alleviation areas in Hunan Province in 2012, 2017, and 2022. Among them, remote sensing images and DEM (Digital elevation models) data come from Geospatial Data Cloud Landsat8 OLI and Landsat7ETM SLC satellite digital products; land use data are obtained from the official website of the Resource and Environment Science Data Center of the Chinese Academy of Sciences; Data on rainfall, population, and economic development data come from the Hunan Provincial Statistical Yearbook.

3.3. Methods

3.3.1. Assessment of the Rural Development Vulnerability Index (RDVI)

This study constructs a framework for analysing the vulnerability of rural development in post-poverty alleviation areas based on the socio-ecological system perspective (Figure 1). The VSD assessment framework and SERV model are integrated to comprehensively analyse the Rural Development Vulnerability Index (RDVI) of post-poverty alleviation areas in Hunan Province.

(1)

VSD Assessment Framework

This study adopts the VSD assessment framework to construct the vulnerability assessment index system of rural development in post-poverty alleviation areas.

The construction of an evaluation indicator system must adhere to the following three principles: theoretical relevance, data availability, and regional representativeness. Based on the theoretical research framework proposed earlier, this study draws on the evaluation indicator systems used by scholars in their vulnerability studies [21,78,79,80,81,82,83,84]. First, from a socio-ecological perspective, three typological dimensions are identified: exposure, sensitivity, and adaptability. In each dimension, we identify the ecological, social, and economic factors that contribute to the vulnerability of rural development in post-poverty alleviation areas and select relevant indicators. In addition, the author’s research team has been committed to the study of rural development and traditional villages in Hunan Province for nearly a decade and has a deep research foundation on the characteristics of rural areas in Hunan Province [85]. Therefore, based on the overview of the research area, we can select relevant characteristic evaluation indicators that are representative of the region.

In summary, based on legal norms, literature references, and relevant research materials, combined with the geographical characteristics of the research area, we compiled a set of indicators to measure the vulnerability of rural development in the post-poverty alleviation areas of Hunan Province. On this basis, we used the expert consultation method to consult 11 scholars, including professors, associate professors, and doctoral students who have long been concerned with underdeveloped areas and rural areas, and asked them to screen the content of the indicators. Since this study considers the characteristics of vulnerability evolution in time series, it has high requirements for the completeness and availability of data in long series. Therefore, based on the availability and completeness of the data, the indicators compiled after consulting with experts were screened a second time, and 30 indicators for evaluating the vulnerability of rural development in the post-poverty alleviation areas of Hunan Province were finally determined, including 10 exposure evaluation indicators, 10 sensitivity evaluation indicators, and 10 adaptability evaluation indicators (

Table 3. Indicator system for assessing Rural Development Vulnerability Index in post-poverty alleviation areas.

Type	Subtype	Indicator	Direction
A1. Exposure	B1. Ecological exposure	C1. Number of days with heavy rainfall	+ *
		C2. Relative rate of change in precipitation	+
		C3. Area of rocky desertification	+
		C4. Number of important geological hazard sites	+
		C5. Industrial sulfur dioxide, wastewater, and soot emissions	+
		C6. Surface PM2.5 concentration	+
	B2. Social exposure	C7. Density of resident population	+
		C8. Urbanization rate	+
	B3. Economic exposure	C9. Share of primary sector output	+
		C10. Per capita disposable income of rural residents	−
A2. Sensitivity	B4. Ecological sensitivity	C11. Average slope	+
		C12. Average elevation	+
	B5. Social sensitivity	C13. Per capita food possession	−
		C14. Proportion of population over 65 years old	+
		C15. Proportion of population aged 15–64	−
		C16. Ratio of male to female population	+
		C17. Number of traditional villages	+
	B6. Economic sensitivity	C18. Share of output value of secondary and tertiary industries	−
		C19. Per capita arable land area	−
		C20. Disposable income gap ratio between urban and rural residents	+
A3. Adaptability	B7. Ecological adaptability	C21. Average NDVI	+
		C22. Number of environmental protection penalty cases	+
		C23. Area of state-level nature reserves	+
	B8. Social adaptability	C24. Number of educated population with tertiary education and above	+
		C25. Number of beds in medical and healthcare institutions per 1000 people	+
		C26. Highway density	+
	B9. Economic adaptability	C27. Fiscal expenditures on healthcare, education and social security	+
		C28. Local fiscal revenue	+
		C29. GDP per capita	+
		C30. Total retail sales of consumer goods	+

* Indicators marked with “+” means the higher the value, the higher the corresponding dimension value; indicators marked with “−” means the lower the value, the higher the corresponding dimension value.

).

(1) Exposure: Exposure refers to the degree to which the system is subject to external risk disturbances and pressures, and the size of exposure depends not only on the intensity of risk disturbances but also has a strong relationship with the system’s own characteristics [78]. In the selection of exposure indicators, in addition to ecological exposure factors consistent with regional characteristics, such as climate risk (C1, C2) and geological disaster disturbance (C4), as well as soil erosion (C3) and environmental pollution (C5, C6) [80], it also includes indicators reflecting population pressure (C7), regional development (C8), agricultural development (C9), and rural income (C10), among other social and economic exposure indicators [79], to assess the long-term exposure to poverty.

(2) Sensitivity: Sensitivity refers to the degree to which a system is internally affected, positively or negatively, by external risk perturbations and pressures, and is determined by the properties of the system itself, which is relatively less sensitive when it is stable. The importance of economic and demographic indicators is of particular interest, as the differences in sensitivity within regions facing similar risk challenges are more indicative of inequalities in the social development dimension [81]. The selected social sensitivity indicators include five indicators reflecting food security (C13), population structure (C14, C15, C16), and cultural diversity (C17) [86]. Among these, cultural diversity is measured by the number of traditional villages selected based on regional representativeness, reflecting regional culture and rural social capital conditions. Economic sensitivity includes three indicators reflecting industrial transformation (C18), arable land resources (C19), and income disparities (C20) [87]. Additionally, two indicators were selected from the terrain and topography (C11, C12) to characterize ecological sensitivity, as terrain and topography characteristics significantly impact agricultural production and are key factors influencing the occurrence of disasters such as landslides and mudslides [81].

(3) Adaptability: Adaptability refers to the ability of a system to adjust and recover in the face of risky perturbations and pressures. Given that adaptability places special emphasis on the flexible adjustment and active response of human society to changes in the external environment, the adaptability indicators established in this study are closely related to a series of proactive interventions and processes implemented by human beings to cope with risks [82,83]. Ecological adaptability includes three indicators reflecting ecological governance (C21, C22) and the scale of regional nature reserves (C26) [88]. Social adaptability is characterized by three indicators reflecting educational attainment (C24), healthcare coverage (C25), and road accessibility (C26). Economic adaptability includes four indicators reflecting government fiscal security (C27, C28) and basic socio-economic levels (C29, C30) [84]. Increased values for most indicators tend to reflect a significant increase in the adaptive capacity of the system, a higher degree of resilience to withstand and cope with potential risk challenges.

In this study, the polar standardisation method was applied to process the raw data of the indicators, and the entropy method was used to determine the weights of the indicators.

(2)

SERV model and visualisation

The SERV (Spatially-explicit-resilience-vulnerability) model proposed by Frazier [75] combines location, space, and scale-specific indicators in vulnerability studies and is more suitable for analyses at the regional level. Therefore, this study adopts the SERV model for RDVI calculation. The SERV model identifies the different distributions of vulnerability in the study area using a number of socio-economic, spatial and location-specific indicators. Spatial statistics (spatial autocorrelation technique) and multivariate techniques (factor analysis) are used to determine the differential impacts of the various vulnerability and adaptability indicators [89]. The formula of the analytical model is as follows:

$${EI}_i={\sum }_{j=1}^n{w}_j\times x_{ij}$$

(1)

$${SI}_i={\sum }_{j=1}^n{w}_j\times x_{ij}$$

(2)

$${AI}_i={\sum }_{j=1}^n{w}_j\times x_{ij}$$

(3)

$${RDVI}_i={EI}_i+{SI}_i-{AI}_i$$

(4)

where RDVI_i is the rural development vulnerability index of the ith county, EI_i is the exposure index, SI_i is the sensitivity index, AI_i is the adaptability index, w_j is the weight of the jth indicator, x_ij is the standardized value of the jth indicator of the ith county.

The calculation results were spatially visualized using ArcGIS 10.6 spatial analysis tools. The results of Exposure, Sensitivity, Adaptability, and RDVI calculations were imported into the attribute tables corresponding to the administrative divisions of each county in the study area.

3.3.2. Spatial Autocorrelation Analysis

Spatial autocorrelation analysis was used to investigate the spatial clustering characteristics of regional RDVI. In this study, spatial autocorrelation is analyzed using the spatial statistical analysis tool of ArcGIS 10.6. The spatial autocorrelation analysis includes global spatial autocorrelation and local spatial autocorrelation. Global spatial autocorrelation is mainly used to describe the overall spatial distribution of accessibility levels throughout the study area, using Global Moran’s I to determine whether there is spatial agglomeration of facility accessibility levels. The formulas are as follows:

$$I={\displaystyle \frac{n}{S_0}}\times {\displaystyle \frac{{\sum }_{i=1}^n{\sum }_{j=1}^n{w}_{ij}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}{{\sum }_{i=1}^n{\left(x_i-\overline{x}\right)}^2}}$$

(5)

$$S_0={\sum }_{i=1}^n{\sum }_{j=1}^n{w}_{ij}$$

(6)

where: n is the number of all spatial units in the study area; x_i and x_j are the observed values of the ith and jth spatial units; $\overline{x}$ is the mean of all the observed values; w_ij is the spatial weight matrix; S₀ is the sum of all the w_ij values. This study applies a binary adjacency matrix, with w_ij = 1 for spatial adjacency and w_ij = 0 for spatial dispersion. The value of Moran’s I is in the range of −1~1. At a given significance level, when Moran’s I is significantly positive, it indicates that spatial units with similar observations (e.g., high-high, low-low) are significantly clustered. When Moran’s I is significantly negative, it indicates that there are significant differences in the spatial patterns of high-low clustering or low-high clustering among observations. When the value of Moran’s I tends to converge, it indicates that there are significant differences between observations. Moran’s I tends to zero, it indicates that the observations are randomly distributed in space and there is no spatial autocorrelation.

Local spatial autocorrelation uses Local Moran’s I to identify spatial agglomeration patterns at different locations, which can more accurately capture the pattern of spatial elements’ divergence in the whole region, and correct the problem of potential smoothing at different locations in the region masked by global autocorrelation. The formulas are as follows:

$$I_i={\displaystyle \frac{x_i-\overline{x}}{S^2}}\sum _{j=1,j\ne i}^n{w}_{ij}\left(x_j-\overline{x}\right)$$

(7)

where: the meaning of each parameter is basically the same as the Global Moran I.

3.3.3. Principal Component Analysis

In order to further explore the intensity of the specific impacts of multifaceted disturbance factors on the vulnerability of social-ecological systems, this study drew on a wide range of domestic and international academic research results and experiences, and utilized the SPSS 27.0 statistical software platform to carry out a principal component analysis. This aimed to reveal the potential complex associations and mechanisms of action between each disturbance factor and vulnerability.

The KMO test and Bartlett’s sphericity test were conducted to determine whether the dataset was suitable for principal component analysis. The KMO test is designed to assess the bias correlation between variables, with a range of values from 0 to 1. When the KMO value tends to be close to 1, it indicates that there are more cofactors between variables and the PCA will be more effective, in particular, it is ideal to be greater than 0.9, whereas values lower than 0.5 are usually considered unsuitable for principal component analysis. The Bartlett’s sphericity test was also used to determine whether the dataset was suitable for principal component analysis. Bartlett’s test of sphericity is designed to test whether the correlation matrix is a unit matrix, the original hypothesis is that the correlation matrix is a unit matrix, the test results in a significance level of less than 0.05 (i.e., p < 0.05), then the original hypothesis is rejected, indicating that there is sufficient correlation between the data to be suitable for principal component analysis.

Next, the eigenvalues of its correlation matrix and the contribution of each principal component are required. When the eigenvalue of its principal component exceeds 1, the principal component is determined to have utility. Meanwhile, when the cumulative principal component contribution rate is higher than 70%, it indicates that the selected principal component can better represent the characteristics of the overall data and has high explanatory power.

3.3.4. Identification and Analysis of Factors Influencing Vulnerability

Geodetectors are a set of mathematical techniques for identifying spatial differences and resolving their underlying driving forces [90,91]. In this study, we mainly use the factor detector in the Geodetector to identify the differences in vulnerability influences, find the independent variable factor with the strongest explanatory power for the dependent variable, and then use the interaction detector to identify the interaction between the two factors.

(1)

Factor detector

The spatial variation in the dependent variable Y was examined, and the extent to which the factor X was able to elucidate differences in the characteristics of Y was determined, adopting the q-value as an assessment criterion. The formulas are as follows:

$$q=1-{\displaystyle \frac{{\sum }_{h=1}^L{N}_h{\delta }_h^2}{N{\delta }^2}}=1-{\displaystyle \frac{SSW}{SST}}$$

(8)

$$SSW={\sum }_{h=1}^L{N}_h{\delta }_h^2, SST=N{\delta }^2$$

(9)

where L represents the hierarchical stratification of the dependent variable Y or the independent variable X, N_h and N are the number of cells in stratum h and the whole region, δ_h² and δ² are the variance of the Y values in stratum h and the whole region, SSW is the total sum of the variances of the strata, SST is the total variance in the whole region, q is the explanatory power of the independent variable for the dependent variable, with the value range of [0, 1]. The closer the q value is to 1, the more prominent X itself is in the explanation of Y, and the more prominent Y’s spatial variability will also be.

(2)

Interaction detector

The main feature of the interaction detector is the ability to examine the interaction between bivariate variables and whether they enhance or reduce the explanatory power of the Y variable when X1 and X2 work together. The first step requires measuring the q-value of each element individually with respect to Y, and in the case where they interact, a further step of comparison is carried out. There are five forms of interaction between the two factors: Weaken-nonlinear; Weaken-univariate-nonlinear; Enhance-bivariate; Independent; and Enhance-nonlinear [90].

The “Weaken-nonlinear” effect means that the two driving factors weaken each other, and the interaction effect of driving factors X1 and X2 is weaker than each of their separate effects. The “Weaken-univariate-nonlinear” effect means that the less active factor reduces the effect of the other one, and the interaction effect of driving factors X1 and X2 lies between each of their separate effects. The “Enhance-bivariate” effect means a greater interactive effect of the driving factors X1 and X2 than each of their separate effects. The “Independent” effect means that the two driving factors are independent of each other. And the “Enhance-nonlinear” effect means the strongest interactive effect of the driving factors X1 and X2 over the sum of their separate effects, a strong enhancement that does not show a simple (linear) proportional relationship [92].

3.3.5. Type Classification of Vulnerability

In order to explain the vulnerability differentiation and its multidimensional complexity more clearly and intuitively, and at the same time provide a basis for effective governance. This study refers to Wang’s study [93], based on the results of exposure, sensitivity and adaptability measurement in each county and district, adopts the dominant elements method to analyse the main contributing dimensions of rural development vulnerability in different geographical areas of the study area, and integrates the results of RDVI to classify the vulnerability type zones.

The dominant elements method is based on the quantitative results of the core evaluation indicators and comprehensively analyses the influence mechanism of other related index items, so as to precisely identify the key elements that drive the formation of differences in each research unit and occupy a dominant position [84]. From the perspective of socio-ecological systems, post-poverty alleviation areas are composed of multiple elements, such as natural, ecological, social, economic, and historical elements, and there are complex and variable interactions between these elements. This interaction leads to the three dimensions of rural development vulnerability, namely, exposure, sensitivity and adaptability, showing obvious differences in the intensity of their roles, and the use of dominant elements analysis to analyse the main contributing dimensions that play a decisive role in determining the degree of rural development vulnerability in post-poverty alleviation areas is of vital significance in effectively classifying the types of vulnerability and promoting the subsequent sustainable development of the region.

At present, the dominant elements method has been widely used in environmental function types [94] and county spatial function area delineation [84], which provides an important theoretical reference for the delineation of rural development vulnerability types from the perspective of the socio-ecological system. Therefore, this study innovatively introduces the dominant elements classification method in vulnerability type area delineation, aiming to precisely define and quantify the type and number of dominant elements in the vulnerability performance of each county. The formulas are as follows:

$$CE=E-\overline{E}$$

(10)

$$CI=I-\overline{I}$$

(11)

$$CA=A-\overline{A}$$

(12)

where CE, CI, and CA denote the off-mean deviation of exposure, sensitivity, and adaptability, respectively; $\overline{E}$,$\overline{I}$, and $\overline{A}$ denote the mean value of exposure, sensitivity, and adaptability, respectively. If the calculation results in CE > 0 or CI > 0, it indicates that the main contributing dimension of vulnerability of the county or district is exposure or sensitivity; if CE ≤ 0 or CI ≤ 0, exposure or sensitivity is the auxiliary contributing dimension of the vulnerability construct of the county or district. If CA < 0, then adaptability is the main contributing dimension of vulnerability for that county and district; if CA ≥ 0, then adaptability is the secondary contributing dimension of vulnerability for that county and district.

Based on the identification of the main contributing dimensions of vulnerability, the results of the RDVI assessment are combined to classify vulnerability types (Figure 3).

4. Results

4.1. Spatial Characteristics of the Rural Development Vulnerability Index

In this study, the RDVI of post-poverty alleviation areas in Hunan Province in 2012, 2017, and 2022 were calculated and visualised on the ArcGIS 10.6 platform. The spatial differentiation status of the three dimensions of exposure, sensitivity, and adaptability was also analysed. The specific analysis results are as follows.

4.1.1. The Spatial Pattern of the Rural Development Vulnerability Index

The RDVI of post-poverty alleviation areas in Hunan Province showed a general trend of decreasing and then increasing from 2012 to 2022. The RDVI decreased from 0.5660 in 2012 to 0.5346 in 2017, and then increased to 0.5382 in 2022. In terms of spatial pattern, RDVI showed the characteristic of ‘low in the east and high in the south and north-west’. The RDVI in western and southern Hunan is always high, while the RDVI in ChangZhuTan and Dongting Lake areas decreases year by year (Figure 4).

In 2012, there were 15 counties with high RDVI or above, including 9 counties in western Hunan, 4 counties in southern Hunan, and 2 counties in the Dongting Lake region. In 2017, there were 17 counties with high RDVI or above, including 13 counties in western Hunan, 3 counties in southern Hunan, and 1 county in the Dongting Lake region. In 2022, there were 13 counties with high RDVI or above, including 9 counties in western Hunan and 4 counties in southern Hunan. The results show that western Hunan and southern Hunan are the areas where the vulnerability of rural development in post-poverty alleviation areas in Hunan Province is more concentrated. They need to be focused on in the future.

4.1.2. The Spatial Patterns of Exposure, Sensitivity, and Adaptability

The exposure of rural development in post-poverty alleviation areas in Hunan Province is generally above the medium level, and the higher exposure counties are mainly concentrated in the southern Hunan and the Dongting Lake region. From 2012 to 2022, the regional exposure shows a trend of ‘decreasing and then increasing’. Among them, the Dongting Lake region is always located in the high exposure area, the exposure of the southern Hunan region increases year by year, and the exposure of the western Hunan region decreases (Figure 5).

The sensitivity of rural development in post-poverty alleviation areas in Hunan Province is generally above the medium level, higher in western Hunan, and lower in the Dongting Lake region. From 2012 to 2022, the regional sensitivity shows a trend of ‘decreasing and then increasing’. The sensitivity of the southern Hunan region and the Dongting Lake region decreases year by year, but the sensitivity of the western Hunan region increases. Therefore, it is necessary to focus on the stability of the socio-ecological system after the disturbance of the rural development process in this region (Figure 6).

The adaptability of rural development in post-poverty alleviation areas in Hunan Province generally shows a below-medium level. The adaptability of the western Hunan region is high, while that of the southern Hunan and Dongting Lake region is relatively low. From 2012 to 2022, the adaptability shows a trend of ‘strengthening first and then weakening’. The adaptability of the western Hunan region generally increases, but the adaptability of the southern Hunan and Dongting Lake region decreases (Figure 7).

4.2. Spatial Clustering Characteristics of the Rural Development Vulnerability Index

Based on the spatial characterization of the vulnerability of rural development in poverty-alleviation areas of Hunan Province, the spatial clustering characteristics of RDVI were further analyzed through spatial autocorrelation analysis. The Moran I of RDVI in post-poverty alleviation areas in Hunan Province was 0.133, 0.084, and 0.075 in 2012, 2017, and 2022, respectively. Moran, I was greater than 0 and passed the significance test at the 0.001 level, which indicated that the areas with similar levels tended to be spatially clustered. However, it showed a decreasing trend with the development of time, indicating that the spatial clustering characteristics gradually weakened with the development of time. In terms of spatial distribution, the spatial agglomeration of RDVI in Hunan Province is mainly dominated by high-high agglomeration and low-low agglomeration. In 2012, the spatial agglomeration characteristics of RDVI in Hunan Province were more obvious, mainly forming a high-high agglomeration area centered on the northwestern region of Xiangxi and a low-low agglomeration area centered on the southwestern region of Xiangxi. From 2017 to 2022, the agglomeration characteristics show a weakening trend, with high-high agglomeration areas mainly concentrated in Xiangxi City and some counties in Huaihua City, and the low-low agglomeration area is mainly concentrated in Loudi City (Figure 8).

4.3. Factors Influencing Rural Development Vulnerability in Post-Poverty Alleviation Areas

Based on the previously constructed vulnerability evaluation index system for rural development in post-poverty alleviation areas in Hunan Province and the results of vulnerability index assessment, the main influencing factors of vulnerability are explored. Firstly, the original indicators were simplified into 10 comprehensive dimensions using PCA (

Table 4. Results of PCA of the original indicators.

Principal Component	Name	Original Indicators Included
X1	Level of rural social service security	C8. Urbanization rate, C13. Per capita food possession, C19. Per capita arable land area, C27. Fiscal expenditures on healthcare, education and social security, C29. GDP per capita, C26. Highway density
X2	Rural population structure	C14. Proportion of population over 65 years old, C15. Proportion of population aged 15–64
X3	Resource and environmental loads	C5. Industrial sulfur dioxide, wastewater, and soot emissions, C7. Density of resident population
X4	Geographical and environmental characteristics	C11. Average slope, C12. Average elevation
X5	Level of development of non-agricultural industries	C18. Share of output value of secondary and tertiary industries, C20. Disposable income gap ratio between urban and rural residents
X6	Urban-rural socio-economic balance	C10. Per capita disposable income of rural residents
X7	Educational level	C24. Number of educated population with tertiary education and above
X8	Level of local finance	C28. Local fiscal revenue
X9	Geological disaster risk	C4. Number of important geological hazard sites
X10	Climate risk	C1. Number of days with heavy rainfall

), thus eliminating the correlation of the 30 original indicators and retaining the original data information to the maximum extent. The main drivers of vulnerability are further investigated using a geodetector.

4.3.1. Factor Detection Results

The explanatory power of each principal component on the vulnerability of rural development in 2012, 2017, and 2022 was probed using a factor detector. The magnitude of the q-value directly reflects the significance of the impact of the principal component on the differences in the spatial distribution of vulnerability. Specific probing results are shown in

Table 5. Ranking results for factor detection.

Principal Component Dimension	2012	Rank	2017	Rank	2022	Rank
X1 Level of rural social service security	0.2387	2	0.1719	1	0.2469	1
X2 Rural population structure	0.0098	10	0.0370	10	0.0401	9
X3 Resource and environ-mental loads	0.1305	5	0.1186	3	0.1646	3
X4 Geographical and environmental characteristics	0.0531	9	0.0675	7	0.0550	8
X5 Level of non-agricultural industries	0.1154	6	0.1075	4	0.1220	4
X6 Urban-rural socio-economic balance	0.0727	8	0.0409	9	0.0559	7
X7 Educational level	0.1516	4	0.0601	8	0.0076	10
X8 Level of local finance	0.0870	7	0.1284	2	0.0695	6
X9 Geological disaster risk	0.2863	3	0.0793	6	0.2451	2
X10 Climate risk	0.3543	1	0.1061	5	0.1042	5

. In 2012, climate risk, the level of rural social service security, and geological disaster risk were the top three principal component dimensions with the strongest explanatory power on the vulnerability of rural development in post-poverty alleviation areas in Hunan. In 2017, the level of rural social service security, the level of local finance, and the resource and environmental loads were the three main dimensions influencing the differences in the vulnerability distribution. In 2022, the greater impacts of the three main dimensions are rural social service security level, geological disaster risk, and rural population structure. The results of the analysis show that the main component dimension of the level of security of rural social services is always ranked in the top three, reflecting that social development and infrastructure development are always the main dimensions affecting the vulnerability of regional rural development. Ecological and environmental risks, such as climatic and geological disasters, are also an important influence on the region’s vulnerability, and the impact of resource and environmental loads and demographic pressures on the vulnerability of the region’s rural development is also increasing and needs to be focused on in the future.

4.3.2. Interaction Detection Results

By detecting the interactions among the principal components in 2012, 2017, and 2022, combined with the degree of individual explanatory power of each principal component in Table 4, it is found that the two-factor interaction demonstrates a more significant explanatory power compared to the independent effect of a single factor (Figure 9). The results of the analyses show that the principal component interactions of rural development vulnerability in post-poverty alleviation areas are dominated by nonlinear enhancement. It indicates that the factors have multiple combined effects on vulnerability.

Table 6. Principal component dimensions of main interactions from 2012 to 2022.

Year	Main Interaction Principal Components with Top 3
	1	2	3
2012	X9∩X10 (0.7235 *)	X1∩X10 (0.7175)	X3∩X7 (0.6006)
2017	X1∩X10 (0.7610)	X1∩X5 (0.7254)	X5∩X7 (0.6368)
2022	X2∩X9 (0.7689)	X1∩X4 (0.6776)	X3∩X10 (0.6765)

* Numbers in parentheses indicate q-values obtained with interaction detection.

lists the principal component dimensions that have major interaction effects on vulnerability in different years. The results of factor interaction detection in 2012 show that geological disaster risk and climate risk (X9∩X10) have the strongest interaction effect with a q-value of 0.7235, followed by the level of rural social service security and climate risk (X1∩X10), with an interaction effect of 0.7175, and thirdly, resource environmental loads and educational level (X3∩X7), with an interaction effect of 0.6006. The results of the factor interaction probes in 2017 show that the level of rural social service security and climate risk (X1∩X10) have the strongest interaction effect, with a q-value of 0.7610; followed by the level of rural social service security and the level of non-farming industry (X1∩X5), with an interaction effect of 0.7254; the third is the level of non-farming industry and educational level (X5∩X7), with an interaction influence of 0.6368. The results of the factor interaction probes in 2022 show that the strongest interaction influence is found in the rural population structure and the geological disaster risk (X2∩X9), with a q-value of 0.7689; the second is the level of rural social service security and the geographical and environmental characteristics (X1∩X4), with a q-value of 0.6776. The third is resource environmental loads and climate risk (X3∩X10), with a q-value of 0.6765. The above findings indicate that, in general, the spatial and temporal differences in the vulnerability of rural development in post-poverty alleviation areas in Hunan Province are caused by a combination of factors. The main interaction influence gradually shifted from natural disaster risk to the influence of multiple risk dimensions superimposed on social resource pressure and ecological and environmental risks, and its vulnerability influencing factors are more complex and diversified.

4.4. Type Classification of Vulnerability in Post-Poverty Alleviation Areas

The dominant elements method was first applied to determine the type of dominant elements of the RDVI, and then combined with the results of the RDVI assessment to comprehensively classify the vulnerability type.

According to the dominant elements method, if the number of dominant elements in a county or district is zero, it is classified as No leading elements types. If the number of dominant elements is 1, it is subdivided into E (Exposure) types, S (Sensitivity) types, or A (Adaptability) types according to the specific type. When the number of dominant elements is 2, it will be further subdivided into E-S types, E-A types, or S-A types. If the number of dominant elements is 3, it is defined as E-S-A types. Accordingly, the types of dominant elements of vulnerability in counties and districts of post-poverty alleviation areas in Hunan Province in 2012, 2017, and 2022 are determined (Figure 10). The results show that the dominant element type in the study area from 2012 to 2022 is always dominated by the S-A type, and this type is mainly distributed in western Hunan.

Precise identification of highly vulnerable areas helps to allocate limited resources efficiently and increase the potential of vulnerable areas to achieve sustainable development. Therefore, the results of the type classification of dominant elements were superimposed on the results of the vulnerability assessment to obtain the results of the vulnerability type classification (Figure 11). The results show that from 2012 to 2022, the high-vulnerability areas of rural development in post-poverty alleviation areas in Hunan Province are mainly dominated by multi-elements dominance, which presents a more complex typological characteristic. And the high vulnerability type areas dominated by multiple elements are mainly concentrated in western Hunan and southern Hunan.

Multi-elements-dominated high vulnerability zones have complex vulnerability characteristics, and should firmly grasp the new development opportunities given by major national strategies such as ‘Rural Revitalisation’ and ‘Rise of the Central Region’, put high-quality development at the top of the priority list, uphold the core concept of green development, do well in disaster prevention and mitigation, ensure food security, and strive to build an agricultural sector deeply rooted in agriculture, with a view to improving food security. The core problems of rural development in single- element vulnerability zones are more prominent and obvious, and it is necessary to focus on the dimensions of agricultural development, pollution control, arable land protection and improvement of social services, so as to make up for the shortcomings and promote the comprehensive and sustainable development of rural areas in poverty-stricken zones. Given the evolving nature of the internal and external environments of the socio-ecological system, it is also necessary to make dynamic adjustments to the environmental changes and the adaptability strategies required during the system’s operating cycle. Successful experiences and lessons learnt from past practices should be continuously refined to enhance the adaptability of the system, so that it can effectively respond to the dual pressures and challenges posed by external uncertainty and risk factors to both the natural ecosystem and the human-social system, and ultimately ensure that the post-poverty alleviation rural areas move forward steadily on the track of sustainable development.

5. Discussion

5.1. External Evidence and Theoretical Thinking

5.1.1. Spatial Characteristics of Vulnerability: Convergences with Global Mountainous-Ethnic Vulnerability Studies

Spatial analysis reveals high vulnerability in southwestern mountainous areas and low vulnerability in northeastern plains of post-poverty alleviation areas in Hunan Province, aligning with global mountainous vulnerability research. Nguyen et al. found that the livelihood vulnerability of ethnic minority communities in northwestern Vietnam is significantly higher for the H’mong community in mountainous areas than for the Thai community. This is mainly due to their higher sensitivity to natural disasters [95]. Fetoui et al. and Poudel et al. found that households in mountainous regions of Tunisia and Nepal are more vulnerable in terms of livelihood strategies [96,97]. These studies reveal how mountainous terrain and cultural marginalization interact to form hotspots of dual ecological and social vulnerability [98]. This characteristic is highly similar to the vulnerability of rural development in the post-poverty alleviation areas of southwestern Hunan Province.

Most post-poverty alleviation areas in Hunan Province are located in hilly and mountainous areas, where geological and climatic disasters are frequent [99]. The post-poverty alleviation rural areas are basically in the key prevention and control zones of geological disasters, and the conditions and development characteristics of geological disasters are difficult to change in the short term [100], which means that the post-poverty alleviation areas will be in a long-term state of vulnerability to geological disasters, which not only causes casualties and property losses, but also damages the local infrastructures and agricultural production conditions. Hunan’s post-poverty alleviation areas are also inhabited by Tujia, Miao, and other ethnic minority people, and the traditional culture of ethnic minorities is also declining in the process of urbanisation due to population loss and environmental changes. The ability of culturally marginalized areas with complex terrain to respond to risks and challenges, as well as their governance capabilities, is relatively weak. This also adds to the challenges of sustainable development in poverty alleviation areas.

For post-poverty rural areas with complex terrain, ethnic minority settlements, and traditional villages, sustainable development paths should integrate ecological conservation with traditional ecological knowledge systems to mitigate vulnerability to climate risks while preserving cultural heritage through community-led traditional village revitalization programs. Additionally, culturally sensitive infrastructure development and targeted policy interventions are critical to foster economic diversification and social cohesion, ensuring equitable growth that honors local cultural identities.

5.1.2. Evolving Influences on Vulnerability: From Natural Disasters to Pressures on Social Resources

The research results show that during the decade 2012–2022, the influencing factors of rural development vulnerability in post-poverty alleviation areas of Hunan Province gradually shifted from natural disaster risks to the influence of multiple risk dimensions superimposed on social resource loads and ecological and environmental risks, and the influencing factors of the vulnerability of social and ecological systems were more complex and diversified. A finding that is consistent with a number of existing studies. Li, Arouri, and Ahmad et al. found that the risk of natural disasters, such as extreme weather and climate change, increases the likelihood of poverty among rural residents through empirical studies in different countries [101,102,103]. At the level of social risk research, McCulloch, Yang et al. argue that the establishment of transport infrastructure and the universalisation of compulsory basic education in rural areas can effectively reduce the likelihood of structural poverty and vulnerability to poverty [104,105]. Hernández shows that the lack of public utilities, education, and healthcare facilities is an important influence on the vulnerability to multidimensional poverty and impoverishment in rural areas of Colombia [106]. Sun, Chen, Li et al. analysed the association between fiscal and healthcare-related social policies such as bank credit, digital inclusive finance, and public transfers and the vulnerability of rural residents, and found that for rural residents with higher post-poverty alleviation area, active social policies such as fiscal and healthcare policies have a greater impact on the vulnerability to return to poverty in poverty-returning areas [107,108,109]. The findings of the above scholars further corroborate the basic conclusions of this study, demonstrating that in post-poverty alleviation areas, social policies and natural ecological risk factors exert a substantial influence on rural development vulnerability. It also enriches the approach to exploring rural development vulnerability in post-poverty alleviation areas from a socio-ecological system perspective.

The findings of this study are closely intertwined with the evolutionary trajectory of China’s urbanization process. China’s urbanization has brought about multiple economic and social development opportunities to post-poverty alleviation areas, while simultaneously presenting challenges in social resource allocation and environmental load, thus exacerbating the urban-rural development imbalance [110]. With the acceleration of urbanization, a significant influx of high-quality skilled laborers and working-age adults from rural areas is flocking to the cities, leading to the problems of ‘hollowing out’ and ‘ageing’ in post-poverty alleviation rural areas [85]. Rural residents’ demand for public services such as education, medical care, and pensions is also increasing, but due to financial pressures and limited resources, it is often difficult to meet the diversified needs of the public for public services [111]. These represent typical “after-effects” of rapid urbanization in the rural development of post-poverty alleviation areas, necessitating urgent intervention from the government and relevant authorities. Targeted measures should prioritize mitigating demographic hollowing, enhancing public service resilience, and fostering endogenous social capital.

5.2. Academic Contributions and Future Research Directions

5.2.1. Theoretical Contributions: Advancing Coupled Socio-Ecological Vulnerability Analysis

In the current global studies on rural development or vulnerability of poverty areas, most scholars focus on a particular dimension of development issues or factors to assess and measure vulnerability. Bui, Lohmann, McCulloch et al. explored the vulnerability of rural development in Southeast Asian regions such as Vietnam and Thailand through agricultural research and development, off-farm employment opportunities, food security, and location factors [105,112,113]. Eshetu showed that participation in urban-rural migration reduced multidimensional poverty and vulnerability by 20.63% and 11.42%, respectively, in rural areas of southern Ethiopia [114]. Fernandez assessed the vulnerability of rural sustainable carding in 10 districts in the province of Huesca, Spain, based on ecosystem service indicators [115]. Cai assessed the vulnerability of rural sustainable carding in the province of Wuhan, China, by evaluating spatial patterns of rural ecological landscapes and water ecosystem vulnerability to understand the current status of ecosystems in China [116]. The above studies have explored the mechanisms that characterise the vulnerability of rural development and poverty areas around the world from one dimension or another, such as socio-economic, rural-urban relations, and ecological environment, but lacked exploration and analysis from a coupled socio-ecological perspective. This study explores the socio-ecological system perspective of rural development vulnerability in post-poverty alleviation areas, which helps to analyse the connotation and mechanism of rural development vulnerability in post-poverty alleviation areas in a more comprehensive and integrated way. This contributes to global efforts to operationalize the SDGs’ cross-sectoral targets by demonstrating how integrated vulnerability analysis can inform holistic development strategies.

5.2.2. Future Research Prospects

Future research needs to further explore the risk factors affecting poverty rebound in rural development in post-poverty alleviation areas and establish a risk resilience correlation mechanism to promote the comprehensive revitalization and development of rural areas after poverty alleviation.

Regional research is an important means of controlling the development of vulnerability at the macro level, and in the future, the scale of research and analysis can be further extended to the micro level (county/village/household). The integration of field research, qualitative research, and other social science methods into the research system is conducive to a more in-depth exploration of the micro-mechanisms of rural development vulnerability in post-poverty alleviation areas. In addition, machine learning methods such as random forests and neural networks can be used to simulate the complex interactions between rapidly changing variables and address limitations such as the lack of nonlinear research in this study. It is also possible to explore policy design based on cultural foundations, taking into account regional characteristics. Conduct participatory research with ethnic minority communities to jointly design “socio-ecological-cultural” three-dimensional resilience interventions. Conduct comparative studies with multi-ethnic regions around the world to further transform these into scalable policy tools.

5.2.3. Practical Implications for Policy and Practice

This study emphasizes vulnerability research based on a coupled socio-ecological perspective, which requires policymakers not to limit themselves to sector-specific solutions, but to take into account the intricate interactions between social structures, ecological processes, and cultural systems to design programs that ultimately promote more equitable and resilient post-poverty transitions. In China, the rural revitalization strategy should be effectively combined with the characteristics of various elements of the region to explore locally adapted development paths and models.

By combining theoretical innovation with rigorous empirical research, this study opens up new frontiers for understanding vulnerability in an era of accelerated urbanization and environmental change. Successful experiences and lessons learned from past practices should be continuously summarized and improved in order to enhance the adaptability of rural territorial systems in poverty-stricken areas, so that they can effectively respond to the dual pressures and challenges posed by external uncertainties and risk factors to both natural ecosystems and human social systems, and ultimately to ensure that rural areas are steadily put on a track of sustainable development after poverty eradication.

5.3. Limitations

This study still has some limitations.

(1) The vulnerability assessment indicator system does not comprehensively cover all the elements of social-ecological systems involved in the rural development process. The vulnerability of rural development encompasses a wide range of interacting natural, ecological, social, economic, and historical factors. The social-ecological system is also a complex system influenced by multiple factors and with complex mechanisms. In the course of the study, some indicators were not included in the assessment system due to the lack of sample data in some counties and years (e.g., energy and infrastructure data, population migration data, biodiversity data). This may lead to unavoidable errors in the assessment results and make it difficult to fully and accurately reflect the actual situation of regional vulnerability. In the future, the study can expand the ways of obtaining indicators and data, and at the same time explore the connotative characteristics of vulnerability from different scales, so as to make the results of the study more accurate and reasonable.

(2) At the level of influencing factors, the study only focuses on the influence of the principal component dimensions on the vulnerability index, but does not study in depth the influencing mechanisms of exposure, sensitivity, and adaptability. In the future, we can explore the influencing factors of exposure, sensitivity, and adaptability separately, so as to explore the influencing mechanism of vulnerability in depth.

(3) Although county-level data supports regional comparisons, it may obscure differences within counties. Future research could combine macro-scale modeling with micro-level case studies to address this gap.

(4) The RDVI is no comparison with the average level of China or the region; the classification of vulnerability types and the proposed strategies may lack certain rationality. In the future, the study area and time period can be expanded to explore the spatial and temporal characteristics of rural vulnerability before and after China achieves full poverty alleviation through comparative analyses. In view of these shortcomings, the author will continue to make efforts to deepen the study in the future, with a view to providing more comprehensive and in-depth analyses and laying a more solid theoretical and practical foundation for the formulation of rural revitalisation and sustainable development strategies in post-poverty alleviation areas.

6. Conclusions

This study assessed the rural development vulnerability index of post-poverty alleviation areas in Hunan Province from a socio-ecological system perspective in 2012, 2017, and 2022, and explored its main influencing factors and type classification. The main research conclusions are as follows:

(1) The rural development vulnerability index of post-poverty alleviation areas in Hunan Province has obvious characteristics of spatial and temporal differentiation. It presents the characteristic of ‘low in the east and high in the south and north-west’ in the spatial pattern. From 2012 to 2022, the rural development vulnerability index of post-poverty alleviation areas in Hunan Province generally shows a downward and then upward trend. The RDVI in western Hunan and southern Hunan is always high, while the RDVI in ChangZhuTan and Dongting Lake regions decreases year by year.

(2) The degree of socio-ecological vulnerability of post-poverty alleviation rural areas in Hunan Province is determined by the three dimensions of exposure, sensitivity and adaptability, and the spatial differentiation is obvious. Counties with higher exposure are mainly concentrated in the southern Hunan and Dongting Lake areas. Sensitivity was higher in western Hunan and lowest in the Dongting Lake region. The adaptability of the western Hunan region is higher, while the adaptability of the southern Hunan and Dongting Lake regions is relatively lower.

(3) Spatial autocorrelation analysis showed that areas with similar rural socio-ecological vulnerability in post-poverty alleviation areas of Hunan Province were significantly clustered spatially. The spatial clustering characteristics gradually weakened with the development of time.

(4) The spatial and temporal differences in the vulnerability of rural development in post-poverty alleviation areas in Hunan Province are caused by a variety of factors, and the core influencing factors have gradually shifted from natural disaster risks to the influence of multiple risk dimensions superimposed on social resource loads and ecological and environmental risks, and the dimensions of the influencing factors have become more complex and diversified. 2012–2022, the level of security of rural social services has always been the main factor influencing the vulnerability of the regional rural development At the same time, climate, geological disasters, and other risk issues are also important influencing factors of the region’s vulnerability, and the impact of resource and environmental loads and demographic pressures on the vulnerability of the region’s rural development is also increasing.

(5) The dominant elements method was used to clearly define the types of vulnerability dominant factors in each county and district, and the results of the vulnerability assessment were combined with the superposition analysis, so as to classify multiple vulnerability type classifications of multi-elements dominant and single-element dominant, and put forward the corresponding development suggestions.