Assessment of the Spatiotemporal Evolution Characteristics and Driving Factors of Ecological Vulnerability in the Hubei Section of the Yangtze River Economic Belt

Abstract

The Hubei section of the Yangtze River Economic Belt (YREB) has an important strategic position as the core zone of the central part of the YREB, and the advantages and disadvantages of its ecological environment are closely related to the development quality of the whole YREB. Moreover, the systematic assessment of ecological vulnerability is of great significance to regional ecological environmental protection, the rational exploitation and utilization of resources, and sustainable development. Based on the pressure–state–response–management model, this study analyzes the spatial and temporal evolution characteristics of the ecological vulnerability of the Hubei section of the YREB and its influencing factors using G1–CRITIC–game theory combination weighting, the Theil index, and the Ridge regression model. The results show that from 2010 to 2023, the area was characterized by medium ecological vulnerability, with an average area share of 58.2%; the degree of vulnerability rose and then fell; the ecological environment gradually improved; and there was an overall spatial distribution pattern of high in the central part and low in the east and west. On the trend of vulnerability transformation, 62.2% of the area remained unchanged, 21% of the area shifted to low vulnerability, and 16.8% of the area increased in vulnerability level. The Theil index decreased and then rose, the degree of spatial agglomeration was floating in a “V” shape, and the spatial pattern of vulnerability was essentially the same in the hot- and cold-spot areas. Among the six ecological functional protection zones, the soil preservation function zone exhibited the lowest average ecological vulnerability index (EVI) at 0.371. From 2010 to 2023, the water source conservation function zone demonstrated a significant decline in EVI, while the remaining zones showed a gradual upward trend in EVI. The human disturbance index was the main driver affecting the change in ecological vulnerability, and the pressure layer was the key influence criterion layer. This study can provide a reasonable evaluation model and analytical framework for the scientific and objective assessment of ecological vulnerability.

1. Introduction

As the material foundation of human existence and the cornerstone of socio-economic development, the ecological environment plays a crucial role in realizing sustainable development [1]. Against the backdrop of global climate change, human social development, and ecosystem deterioration, ecological environmental protection has increasingly become one of the key issues of concern for scholarly organizations around the world [2,3,4,5]. Ecological vulnerability refers to the response and resilience of ecosystems to external disturbances at specific spatial and temporal scales and is the result of the joint action of natural factors and socio-economic activities [3,6,7]. Ecological vulnerability assessment can objectively analyze the condition of ecosystems and explain the change patterns and causes of vulnerability [8]. Therefore, the scientific evaluation of ecological vulnerability is a key measure for preventing ecological degradation, improving the ecological environment, and promoting green and sustainable development, and is of great significance to the realization of the balanced and coordinated development of the region in terms of “economy, society, and ecology” [9,10].

Currently, the research focus of ecological vulnerability evaluation mainly concentrates on the indicator system, weighting method, and analysis model used [11]. For example, in [12], 24 indicators were selected to construct the evaluation index system based on the exposure–sensitivity–adaptation (Vulnerability Scoping Diagram, VSD) model. The authors of [13] evaluated the ecological vulnerability of the Wuliangsu Sea Basin using the sensitivity–pressure–recovery (SRP) model. The authors of [14] used the pressure–state–response (PSR) model combined with fuzzy analysis to analyze the spatial and temporal characteristics of ecological vulnerability in the Tarim River Basin. The above evaluation model lacks horizontal comparison, and each criterion layer is relatively fragmented, ignoring the interaction between indicators. The pressure–state–response–management (PSRM) model takes into account the interactions between the criterion layers and is more systematic and comprehensive than the previous evaluation system [15]. At the level of weighting methods, the weights of indicators are crucial to the evaluation results, and the current common weight calculation methods include hierarchical analysis (AHP) [16], entropy weighting [17], principal component analysis [18], and fuzzy assessment [19]. Single-level calculation methods inevitably have certain limitations; for example, the hierarchical analysis method is overly dependent on expert scoring, which is too subjective; the entropy weighting method can easily ignore information on differentiation between indicators, and a horizontal comparison of indicators is lacking. Therefore, a scientific and reasonable method of combining the results of multi-level weighting should take into account the advantages of both subjective and objective weighting methods, thus preventing the disadvantage of using a single method from affecting the analysis results to the greatest extent possible [20].

The Hubei section of the YREB, located in the center of the Yangtze River Basin, has a prominent position as a transportation hub, an area rich in natural resources, and a good industrial base, and is a natural link between the eastern and western regions of the YREB. However, in the process of rapid economic development, the region faces problems, such as insufficient resource integration, water pollution, and unstable ecosystems [21]. With the introduction of the “Provincial People’s Government’s Opinions on the Implementation of the National Strategy for the Development of the YREB” and other relevant policies, Hubei Province plans to develop the Hubei section of the YREB into a modern industry-intensive and ecological civilization demonstration belt that will lead Hubei’s economic and social development, enhance the function of the golden waterway in the middle reaches of the Yangtze River, promote the rise of the central region, and ensure the high-quality development of the YREB. The regional strategy has been transformed into joint economic–environmental development. Therefore, a comprehensive and systematic ecological vulnerability assessment is the key to understanding the current ecological environment of the Hubei section of the YREB, maintaining the stability of the ecosystem, and realizing the harmonious development of the YREB in terms of the economy, society, and ecology [22]. Since the concept of the YREB was proposed [23], many scholars have conducted a series of ecosystem-related studies around the region. The authors of [24] analyzed the spatial and temporal changes in ecological quality in the Yangtze River Basin and its relationship with environmental and topographical factors from 2001 to 2019 based on the GEE platform using the remotely sensed ecological index. Their results showed that the remotely sensed ecological index of the Yangtze River Basin exhibits an upward trend, and the ecological quality generally improves. The authors of [25] proposed a “habitat–structure–function” framework and combined it with a geographically weighted regression model to analyze and evaluate the pattern of ecological vulnerability in the Yangtze River Basin and the spatial heterogeneity of various factors on the change in vulnerability from 1990 to 2018. Their results showed that the ecological vulnerability of the Yangtze River Basin has generally decreased, and human activities are the main factor contributing to the change in ecological vulnerability. The authors of [26] used the SRP model, combined with AHP-SPCA and the geodetector system, to analyze the spatial and temporal characteristics of ecological vulnerability and influencing factors in the Jiangsu section of the YREB. According to their results, the ecological vulnerability in the study area showed an increasing trend from Nanjing to Nantong, and the proportion of arable land was the dominant factor in the spatial pattern of vulnerability. The above studies mainly focus on the analysis of the whole and parts of the ecosystem of the YREB, but there is a lack of studies on the ecological vulnerability of the Hubei section of the YREB, which makes it difficult to effectively reflect on its ecological environment quality and ecological vulnerability. Therefore, based on the current ecological status of the Hubei section of the Yangtze River Economic Belt, the following research objectives are proposed: (1) to reveal the spatiotemporal evolution characteristics of ecological vulnerability in the Hubei section by constructing a multidimensional ecological vulnerability assessment model; (2) to analyze the temporal and spatial differences in the spatial agglomeration patterns of ecological vulnerability; and (3) to identify and quantify the driving factors influencing regional ecological vulnerability dynamics. Furthermore, targeted restoration recommendations are proposed by integrating evaluation results with ecological functional protection zone types, thereby providing a scientific basis for ecological vulnerability assessment, as well as the formulation and implementation of ecological conservation policies in the YREB.

The principal contributions of this study are as follows: (1) A comprehensive multidimensional ecological vulnerability assessment framework was developed by integrating the G1–CRITIC–game theory combination weighting with the Pressure–State–Response–Management (PSRM) model. This approach overcomes the limitations of conventional methods in the coupling analysis of complex human–environment systems, achieving dynamic equilibrium between subjective and objective weights and enabling multidimensional vulnerability evaluation. (2) The spatiotemporal evolution patterns of ecological vulnerability in the Hubei section of YREB from 2010 to 2023 were systematically quantified, addressing gaps in traditional methods for analyzing coupled complex systems. (3) The spatial heterogeneity driving mechanisms of ecological vulnerability were elucidated through the combined application of the Theil index, hot-spot analysis, and ridge regression modeling, filling the research void in the high-precision dynamic monitoring and attribution analysis of ecological vulnerability in this region. This methodological framework effectively addresses the issues of fixed weighting and mechanistic rigidity inherent in conventional models for assessing vulnerability within complex human–environment coupled systems, thereby achieving a methodological breakthrough in ecological vulnerability evaluation at the watershed scale. These findings provide a scientific foundation for advancing the construction of an ecological civilization demonstration zone and formulating targeted environmental protection policies in the Hubei section, thereby promoting the balanced and coordinated development of the economy–society–ecology nexus. The technical workflow is illustrated in Figure 1.

2. Materials and Methods

2.1. Overview of the Study Area

The Hubei section of the YREB is located in the middle reaches of the Yangtze River Basin, at the “waist” of the Yangtze River, which is the golden zone connecting the east and west of the YREB and has the key significance of starting and ending in the whole YREB [27]. The study area lies between longitude 110°29′~115°93′ E and latitude 29°74′~30°98′ N, and spans about 590 km from east to west. There are many water systems in the area, including the Yangtze River, which flows through eight cities, namely, Wuhan, Yichang, Huanggang, Ezhou, Xianning, Huangshi, Jingzhou, and Enshi Tujia and the Miao Autonomous Prefecture [28]. The total area is about 5 × 10⁴ km², and the elevation gradually decreases from east to west to the center, with a large undulating topography, whereas a large area in the center is a relatively flat plain. The region has a subtropical monsoon climate, with four distinct seasons, cold winters and hot summers, and simultaneous rain and heat. The average annual temperature is 16.7°, and the average annual precipitation is 1200 mm, with a decreasing trend from west to east. A modern industrial pattern with primary industry as the supporting industry, secondary industry as the dominant industry, and tertiary industry as the leading industry characterizes the study area. An overview of the study area is shown in Figure 2.

2.2. Data Sources and Processing

In this study, NDVI, topography, land use, water system, meteorology, soil type, population, GDP, and socio-economic data were selected as basic data.

The NDVI data were the 250 m Normalized Vegetation Index dataset for the Chinese region [29], and the like-element dichotomous model was used to calculate the vegetation cover in 2010, 2015, and 2023. Topographic data were ASTER GDEM data at 30 m resolution. Land use data were provided by [30]. The density of the water system was calculated by nuclear density analysis. Meteorological data including mean annual precipitation and mean annual temperature were obtained from [31,32,33]. The human disturbance index (HDI) was calculated according to the formula proposed in [34]. Socio-economic data were obtained by reviewing the Hubei Provincial and Municipal Statistical Yearbooks from 2010 to 2023, and vector data were rasterized using the inverse distance interpolation method in order to ensure the consistency of the different data study scales. The habitat quality index, biological abundance, and landscape diversity index (SHDI) were calculated based on land use type as the base data using the InVEST model, the biological abundance formula [35], and Fragstatsv4.2 software, respectively, and the values of the relevant parameters of the InVEST model were referenced to previous studies and user manuals [36,37]. Soil erosion intensity was calculated using the Modified Soil Erosion Model (RUSLE) equation [38,39]. Due to the absence of 2023 data in the statistical yearbook, data from the most recent available year were substituted. The data on ecological functional protection zones were obtained from the China Ecosystem Assessment and Ecological Security Database (https://www.ecosystem.csdb.cn/, accessed on 10 January 2025). Based on the ecosystem functional characteristics of the study area and integrated with regional geographical attributes, the zones were classified into six distinct types of ecological functional protection zones. Details are shown in

Table 1. Data sources and content.

Data Type	Data Content	Data Sources	Data Use
NDVI	MOD13Q1	Institute of Tibetan Plateau Research Chinese Academy of Sciences (https://data.tpdc.ac.cn, accessed on 1 October 2024).	Calculation of vegetation cover, cover and management factors
Terrain	ASTER GDEM	Geospatial Data Cloud (https://www.gscloud.cn/, accessed on 1 October 2024).	Calculate DEM, slope, terrain undulation, slope length and gradient factor.
Land use	30 m CLCD Land Use	Zenodo (https://zenodo.org/records/12779975, accessed on 1 October 2024).	Calculation of HDI, habitat quality index, biological abundance, landscape diversity index, soil and water conservation factors
Water system	Hubei Provincial Water System	OpenStreetMap (https://www.openstreetmap.org/, accessed on 1 October 2024).	Calculation of water system density
Hydrological	1000 m precipitation, temperature	Institute of Tibetan Plateau Research Chinese Academy of Sciences (https://data.tpdc.ac.cn, accessed on 2 October 2024).	Calculation of mean annual precipitation, mean annual temperature, rainfall erosivity factor
Soil type	1000 m grid	Harmonized World Soil Database (HWSD, accessed on 2 October 2024).	Calculation of soil erodibility factor
Population	1000 m grid	LandScan Global (https://landscan.ornl.gov/, accessed on 2 October 2024).	Calculation of population density
GDP per capita	1000 m grid	Resource and Environmental Science Data Platform (https://www.resdc.cn, accessed on 2 October 2024).	Calculation of GDP per capita
Socio-economic	Socio-economic indicators from 2010 to 2023	Hubei Province, Municipal Statistical Yearbook	Calculation of sewage treatment rate, green area ratio in built-up areas

.

In this study, multi-source data with different data spatial resolutions and coordinate systems are used, and to ensure the accuracy and reasonableness of the results, the raster resolution of each indicator is converted to 300 m×300 m, and the projection is uniformly defined as CGCS2000 3 Degree GK Zone 37.

2.3. Research Methodology

2.3.1. Evaluation Indicators

The PSRM model is “pressure (P)–state (S)–response (R)–management (M)”, and the interaction of the four guideline levels constitutes the logical relationship of “what happens, why it happens, how to deal with it, and based on what” [15]. The PSRM model was chosen to construct an ecological vulnerability index evaluation system for the Hubei section of the YREB, aiming to analyze the mutual influence of different criterion layers on the ecological environment of the study area.

The pressure layer refers to the pressure on regional ecosystems that are disturbed by human activities. The greater the pressure on the ecosystem, the more serious the damage to the local natural environment caused by human activities and economic development. The rapid economic and social development of the Hubei section of the YREB, industrial restructuring, and human production and life have an important impact on the local ecological environment. Therefore, GDP per capita was chosen to characterize the development status of the region; population density and HDI characterize the impact of production and living activities on the ecological environment; and water system density characterizes the carrying capacity of regional water resources, higher values indicating greater densities of rivers, lakes, and wetlands, which enhance the regulation of the hydrological cycle, mitigate drought stress, and strengthen the region’s ability to sustain water resources, thereby contributing to ecosystem stability.

The state layer refers to the state of the regional ecosystem when it is disturbed by the outside world. The higher the vegetation cover, the stronger the stability of the ecosystem. Precipitation and temperature jointly determine the hydrothermal conditions essential for vegetation growth. Increased precipitation enhances vegetation productivity by elevating plant water content and thereby promoting ecosystem carbon sequestration. The Hubei section of the Yangtze River Economic Belt, situated in plain regions with relatively high mean annual temperatures, faces intensified drought stress as rising temperatures cause plant transpiration rates to exceed precipitation recharge capacity. Elevation indirectly regulates vegetation growth conditions by modulating local precipitation patterns and thermal regimes. Slope gradient and topographic relief critically shape geomorphological characteristics: Higher values accelerate surface runoff velocities and significantly elevate the soil erosion modulus, exacerbating water and soil loss. Concurrently, fragmented landscapes caused by steep terrain obstruct biological migration corridors, leading to decline in biodiversity indices and ultimately heightening ecological vulnerability. The intensity of soil erosion affects the soil environment in which vegetation grows. Therefore, average annual precipitation, average annual air temperature, slope, DEM, topographic relief, and soil erosion intensity were chosen to characterize the local ecosystem.

The response layer refers to the ability of regional ecosystems to withstand external pressure, mainly reflecting the strain capacity of regional ecosystems under environmental pressure. The habitat quality index reflects the structural and functional integrity of ecosystems, with higher values indicating lower levels of habitat degradation. Bioabundance, a measure of biodiversity, enhances ecosystem stability by increasing the complexity of food webs. Vegetation cover quantifies the density of plant biomass in a region, where higher values promote soil and water conservation and regulate microclimates. Landscape diversity, often characterized by the Shannon Diversity Index (SHDI), is positively correlated with the degree of land use fragmentation, where elevated values typically signify heterogeneous but ecologically disconnected patches that compromise habitat connectivity. Therefore, the habitat quality index, biological abundance, vegetation cover, and SHDI were chosen to reflect the responsiveness to ecological changes.

Management refers to the measures taken to reduce and prevent the negative impacts of human activities on the environment, including the reduction in the negative impacts of governmental and individual activities on the environment, as well as the measures taken by human beings to restore the ecological environment and reduce ecological vulnerability. Increasing the sewage treatment rate can significantly reduce the discharge of key pollutants, such as chemical oxygen demand (COD) and ammonia nitrogen into water bodies. The resultant decline in pollutant fluxes mitigates eutrophication risks, thereby enhancing the stability of aquatic ecosystems. As the study area is located within the mainstream basin of the Yangtze River, improving sewage treatment rates effectively suppresses the deterioration in regional ecological vulnerability. Green spaces in built-up areas (e.g., parks, shelterbelts, and wetland parks) reduce surface runoff, alleviate the urban heat island effect, and enhance carbon sequestration capacity through vegetation coverage. Notably, green spaces within industrial zones effectively adsorb pollutant gases such as particulate matter with a diameter of 2.5 μm or less emitted from industrial areas. These mechanisms directly diminish the ecosystem’s exposure and sensitivity, reinforcing ecosystem stability and reducing the severity of ecological vulnerability. Therefore, the sewage treatment rate and the green area ratio in built-up areas were chosen to indicate the changes in environmental awareness and the behavior of the government and people. The specific evaluation indicators are shown in

Table 2. Ecological vulnerability evaluation indicator system.

Criterion Layer	Indicator	Resolution	Direction	Ecological Significance
Pressure (P)	Population density (A1)	1000 m	+	Population pressure on land and environment
	GDP per capita (A2)	1000 m	−	Level of regional economic development
	Water system density (A3)	1:1 million	+	Waterway connectivity
	HDI (A4)	30 m	+	Level of human activity
State (S)	Slope (B1)	30 m	+	Topographic and geomorphologic conditions
	DEM (B2)	30 m	+	Topographic and geomorphologic conditions
	Terrain undulation (B3)	30 m	+	Topographic and geomorphologic conditions
	Soil erosion intensity (B4)	300 m	+	Soil erosion conditions
	Average annual precipitation (B5)	1000 m	−	Hydrothermal conditions
	Average annual temperature (B6)	1000 m	+	Hydrothermal conditions
Response (R)	Habitat quality index (C1)	30 m	−	Habitat quality status
	Bioabundance (C2)	30 m	−	Ecosystem diversity
	Vegetation cover (C3)	250 m	−	Degree of surface vegetation cover
	SHDI (C4)	30 m	−	Landscape heterogeneity
Management (M)	Sewage treatment rate (D1)	300 m	−	Disposal of pollutants
	Green area ratio in built-up areas (D2)	300 m	−	Urban greening status

Note: “’+’ indicates a positive driving force, while ‘−’ indicates a negative driving force”.

.

Different indicators have different quantitative outlines, and in this study, the indicators were normalized using the method of change in extreme deviation with the following formula:

Positive indicators:

$${\omega }_i=(j_i-j_{\min })/(j_{\max }-j_{\min })$$

(1)

Negative indicators:

$${\omega }_i=(j_{\max }-j_i)/(j_{\max }-j_{\min })$$

(2)

where ω_i is the normalized value, j_i are the raw data of the indicator, j_max is the maximum value of the indicator, and j_min is the minimum value of the indicator.

2.3.2. G1 Method

The G1 method is a subjective weighting method which can improve the shortcomings of hierarchical analysis; compared with the previous subjective weighting method, the process is more concise and clear without the need for a consistency test [40]. The calculation process is as follows:

1.

Determine the indicator order relationship and the importance of the indicator relative to the evaluation criteria as x₁ > x₂ > … > x_m.

2.

Rational judgment based on expert experience in determining the ratio of relative importance levels between neighboring indicators r_k.

$$r_k=Y_{k-1}/Y_k$$

3.

3.

Calculate the weighting factor wk.

$$w_k={(1+{\displaystyle \sum _{k=2}^m{\displaystyle \prod _{i=k}^m{r}_k}})}^{-1}$$

(4)

$$w_{k-1}=r_k{w}_k$$

(5)

2.3.3. CRITIC Method

As an objective empowerment method, the CRITIC method can consider the amount of information contained in the indicators themselves, fully reflect the conflict and difference between different programs, and pay more attention to the horizontal comparison between the indicators than the entropy weight method; additionally, the calculation results are more accurate and reasonable [41]. The formulas are as follows:

1.

For data standardization, see Equations (1) and (2).

2.

The standard deviation of each indicator and the correlation coefficient between indicators of the standard matrix X are as follows:

$${\omega }_i=\sqrt{{\displaystyle \frac{1}{m}}{{\displaystyle \sum _{i=1}^m(x_{ij}-\stackrel{\_}{x_j})}}^2}\begin{array}{cc} & i=1,2,\dots ,n\end{array}$$

(6)

$${\rho }_{ij}=\mathrm{cov}(X_i,X_j)/({\omega }_i,{\omega }_j)\begin{array}{cc} & i,j=1,2,\dots ,n\end{array}$$

(7)

3.

Indicator informativeness:

$$W_j={\omega }_j{\displaystyle \sum _{i=1}^{n}1-{\rho }_{ij}}$$

(8)

4.

Calculate objective weights for indicators:

$$E_j={\displaystyle \frac{W_j}{{\displaystyle \sum _{j=1}^n{W}_j}}}$$

(9)

2.3.4. Game Theory Combination Weighting

The game theory combination weighting method achieves equilibrium among different weighting approaches by minimizing the deviations between potential weights and fundamental weight sets. This equilibrium is not a simple arithmetic mean but rather seeks a Pareto-optimal solution for subjective and objective weights through the Nash equilibrium, thereby enabling the synergistic integration of advantages from distinct weighting methodologies and enhancing the scientific rigor and reliability of comprehensive evaluations. To ensure the objectivity and rationality of assessment outcomes, in this study, the game theory combination weighting method was employed to scientifically optimize the weight coefficients derived from the G1 method and CRITIC method, ultimately determining the finalized indicator weights [42,43].

N ecological vulnerability evaluation index weight vectors are obtained by using N methods, constituting a weight vector set $w=\sum _{k=1}^N{\alpha }_k{w}_k^T$. Based on the idea of the game ensemble model, in order to find the optimal weight vectors, the above M linear combination coefficients α_k are optimized to achieve the minimization of the deviation, based on which the conceptual model is determined:

$$\min ‖{\displaystyle \sum _{k=1}^N{\alpha }_j{w}_j^T-w_i^T}‖(i=1,2,\dots ,N)$$

(10)

The system of linear equations based on the matrix differentiation property converting the above equation to the optimized first-order derivative condition is:

$$\left[\begin{array}{cccc}{w}_1{w}_1^T& w_1{w}_2^T& \dots & w_1{w}_{\delta }^T\\ w_2{w}_1^T& w_2{w}_2^T& \dots & w_1{w}_{\delta }^T\\ ⋮& ⋮& ⋮& ⋮\\ w_{\delta }{w}_1^T& w_{\delta }{w}_2^T& \dots & w_1{w}_{\delta }^T\end{array}\right]\left[\begin{array}{c}{\alpha }_1\\ {\alpha }_2\\ ⋮\\ {\alpha }_{\delta }\end{array}\right]=\left[\begin{array}{c}{w}_1{w}_1^T\\ w_2{w}_2^T\\ ⋮\\ w_{\delta }{w}_{\delta }^T\end{array}\right]$$

(11)

For (α₁, α₂, …, α_m) normalization, ${\alpha }_k^{\mathrm{*}}={\alpha }_k^{\mathrm{*}}/{\sum }_{k=1}^M{\alpha }_k^{\mathrm{*}}{w}_k^T$; the final combined weights are:

$$W={\displaystyle \sum _{k=1}^M{\alpha }_k^{*}}{w}_k^T$$

(12)

The results of the weighting operation are shown in

Table 3. Indicator weights.

Indicator	G1 Method	CRITIC Method	Game Theory Combination Weights
Population density	0.0814	0.0077	0.0441
GDP per capita	0.0471	0.1160	0.0820
Water system density	0.0565	0.0807	0.0688
HDI	0.0678	0.0429	0.0552
Slope	0.0288	0.0505	0.0397
DEM	0.0171	0.0528	0.0351
Terrain undulation	0.0240	0.0439	0.0340
Soil erosion intensity	0.0345	0.0073	0.0208
Average annual precipitation	0.0580	0.0701	0.0641
Average annual temperature	0.0483	0.0310	0.0396
Habitat quality index	0.0913	0.0931	0.0922
Bioabundance	0.0913	0.0874	0.0893
Vegetation cover	0.1460	0.0474	0.0962
SHDI	0.0761	0.0734	0.0747
Sewage treatment rate	0.0549	0.0915	0.0734
Green area ratio in built-up areas	0.0768	0.1045	0.0908

after the optimized combination of the two linear coefficients of the G1 method and the CRITIC method based on game theory combination weighting.

2.3.5. Ecological Vulnerability Evaluation Model

In this study, the ecological vulnerability index (EVI) was used to analyze and evaluate the ecological condition of the Hubei section of the YREB, with reference to the existing EVI model [44], which was calculated as follows:

$$EVI={\displaystyle \sum _{i=1}^n{a}_j}\times r_{ij}$$

(13)

where EVI is the ecological vulnerability index, ɑ_j is the indicator weight, r_ij is the normalized value, and n is the number of indicators. Based on the field condition of the Hubei section of the YREB and the existing classification standards [45,46,47,48], the ecological vulnerability results were classified into the five grades of slight vulnerability [0~0.2], mild vulnerability (0.2~0.4], moderate vulnerability (0.4~0.6], severe vulnerability (0.6~0.8], and extreme vulnerability (0.8~1.0), using the equidistant method.

2.3.6. Theil Index

The Theil index was first proposed by Henri Theil and used as an important statistical index to measure the difference in regional income levels [49]. In this paper, the Theil index was introduced into ecological vulnerability assessment to reflect the degree of spatial clustering of ecological vulnerability in the Hubei section of the YREB by measuring the differences in EVI values between regions, with the following formula [1]:

$$T={\displaystyle \sum _{a=1}^n\left[\frac{EV{I}_a}{EVI}\times \log \left(\frac{EV{I}_a/EVI}{A_a/A}\right)\right]}$$

(14)

where T is the Theil index; n is the total number of districts in each county; EVI_a and EVI are the ecological vulnerability indexes for the a-th district and overall, respectively; and A_a and A are the area of the a-th district and overall, respectively.

2.3.7. Getis–Ord Gi*

The Getis–Ord Gi* method is able to identify statistically significant spatial clustering distribution patterns of high and low values in the region, with high G* values indicating the clustering of high values of ecological vulnerability in the region and low G* values indicating the clustering of low values of ecological vulnerability, with the following formula [50]:

$$G_i^{*}={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n{\alpha }_{i,j}{y}_j-\overline{X}{\displaystyle {\sum }_{j=1}^n{\alpha }_{i,j}}}}{S\sqrt{\frac{\left[n{\displaystyle {\sum }_{j=1}^n{\alpha }_{i,j}^2-{\left({\displaystyle {\sum }_{j=1}^n{\alpha }_{i,j}}\right)}^2}\right]}{n-1}}}}$$

(15)

where n is the total number of delineation units, y_j is the value of the ecological vulnerability index at unit j, and α_ij is the spatial weight matrix of regional units i and j.

2.3.8. Ridge Regression Model

The ridge regression model addresses limitations in the study of the multicollinearity of the independent variables and corrects the bias that may be induced by the ordinary least squares method when estimating the parameters [51,52]. The basic idea lies in introducing a normal number matrix kI to the singular matrix X′X to alleviate its singularity, so as to model more reasonable estimated coefficients. In addition, this model efficiently integrates the information of independent variables, which makes the model construction more scientific and reasonable [53]. In this study, a ridge regression model was introduced with the aim of dissecting the intrinsic influence mechanism of ecological vulnerability with the following formula:

$$\stackrel{\wedge }{\beta }(k)={(X^{\prime }X+kI)}^{-1}{X}^{\prime }Y$$

(16)

where k is the ridge estimation parameter, I is the unit matrix of the same order as X′X, Y is the matrix of the explanatory variables, and X is the row matrix consisting of the explanatory variables. k = 0, the ridge regression estimation parameter, is the least squares regression model (OLS) estimation parameter [54]. The absolute value of the regression coefficient indicates the degree of influence of each indicator on ecological vulnerability; a positive value indicates a positive driver, while a negative value indicates a negative driver.

3. Results

3.1. Ecological Vulnerability Spatiotemporal Evolution Characteristics

From 2010 to 2023, the ecological vulnerability of the Hubei section of the YREB shows a trend of increasing and then decreasing, with an overall decreasing trend, as shown in Figure 3. The region as a whole is dominated by moderate vulnerability, with the proportion of moderately vulnerable areas being 60.7%, 59.9%, and 54.1% in 2010, 2015, and 2023, respectively. From 2010 to 2015, the areas of slight, severe, and extreme vulnerability increased by 223.82 km², 2019.6 km², and 145.26 km², with the proportion of the areas increasing by 0.48%, 4.12% and 0.30%, respectively; the mildly and moderately fragile area decreased by 2026.17 km² and 353.88 km², respectively, accounting for 4.15% and 0.75%, respectively. The severe vulnerability area had the largest growth, the mild vulnerability area was obviously reduced, and ecological vulnerability increased. From 2015 to 2023, the areas of slight and mild vulnerability increased by 1280.16 km² and 2802.06 km², respectively, and their proportion increased by 2.61% and 5.7%, respectively. The moderately, severely, and extremely vulnerable areas decreased by 2860.2 km², 1064.07 km², and 123.21 km², respectively, and their proportion decreased by 5.88%, 2.18%, and 0.25%, respectively. Among them, the minimally and mildly vulnerable areas increased substantially, the moderately, severely, and extremely vulnerable areas decreased greatly, and the moderately, severely, and extremely vulnerable types shifted to the minimally and mildly vulnerable types, resulting in a significant reduction in ecological vulnerability and a sustained improvement in the ecological environment.

Spatially, the EVI of the Hubei section of the YREB during 2010 and 2015 exhibited a distinct east–central high and west low distribution pattern, as illustrated in Figure 4. Cities along the Yangtze River mainstream—including Jingzhou, Wuhan, Ezhou, Huanggang, and northern Xianning—exhibited moderate to severe vulnerability grades due to rapid industrialization and urbanization. In contrast, regions such as Enshi Tujia and the Miao Autonomous Prefecture, western Yichang, and Huangshi, characterized by high forest coverage rates and low human disturbance intensity, maintained slight to mild vulnerability grades over this period. By 2023, the spatial pattern shifted to a central-to-peripheral gradient decline, reflecting overall improvements in ecological quality compared to 2015. The mild vulnerability area in western Enshi expanded substantially, while moderate, severe, and extreme vulnerability areas in the east largely transitioned to slight and mild grades. For instance, Wuhan and Ezhou shifted from severe to mild vulnerability, and Huangshi and Huanggang transitioned from moderate to mild vulnerability.

3.2. Spatial Clustering Patterns of Ecological Vulnerability

From 2010 to 2023, the spatial pattern of ecological vulnerability in the Hubei section of the YREB showed a centralized distribution, with the degree of spatial agglomeration decreasing and then increasing. The Theil index decreased from 0.371 in 2010 to 0.365 in 2015, indicating that the degree of spatial agglomeration was slightly weakened, and the spatial pattern tended to be randomly distributed. The Theil index increased to 0.398 in 2023, indicating an overall enhancement of spatial agglomeration and demonstrating strong interactive effects on ecological vulnerability within the study area. In order to verify the reasonableness of the calculation results of the Theil index, the Moran index was used for testing, and the results are shown in Figure 5. From 2010 to 2023, the Moran index values were 0.883, 0.875, and 0.88, respectively. Both the Theil index and Moran’s I exhibited a “V”-shaped fluctuation trend. Furthermore, the Moran index results showed a p-value of less than 0.01 and a Z-score greater than 1.96, confirming that the Theil index calculations adhere to statistical principles and demonstrate significant clustering characteristics.

The spatial distribution pattern of hot- and cold-spots in the Hubei section of the YREB is shown in Figure 6. The overall ecological vulnerability hot-spots are mainly concentrated in the east and part of the central region. The hot-spots in the central and eastern parts are mainly affected by factors such as high population density and frequent production activities because the area is mostly a plain. Cold-spots are mainly concentrated in the western part of the region, where the vulnerability index is low and the area of cold-spots is large due to the sparse population, the low impact of human activities on the area, vast forest and grassland areas, and relatively stable ecosystems. The hot-spots in the Hubei section of the YREB from 2010 to 2023 account for 28.1%, 23.8%, and 32.2% of the total area, while the cold-spots are 21.3%, 24.1%, and 26.5% of the total area, respectively. In 2010~2015, the hot-spot area share decreased, and the cold-spot area share increased; specifically, the central hot-spot area decreased, and the western cold-spot area expanded. In 2015~2023, the cold- and hot-spot area share both increased; specifically, the central hot-spot and the eastern cold-spot areas increased.

3.3. Trends in Ecological Vulnerability

3.3.1. Trends in Overall Vulnerability

The transfer matrix of ecological vulnerability level is shown in Figure 7. The values in the figure represent the percentage of areas (km²) of transitions into and out of different ecological vulnerability types. From 2010 to 2015, the ecological vulnerability level of the Hubei section of the YREB maintained the current vulnerability level as a whole and part of it shifted from a lower vulnerability to a higher vulnerability level, which was mainly manifested in the conversion of mild vulnerability to moderate vulnerability, with a conversion area of 4103.3 km². Moderate vulnerability was converted to severe vulnerability, with a conversion area of 2961.8 km². From 2015 to 2023, the vulnerability level decreased, and most of the vulnerability level was converted to a lower level of vulnerability, with 1699.8 km² of mild vulnerability converted to slight vulnerability, and 55.8 km² of medium vulnerability converted to slight vulnerability. In mild vulnerability, a 1699.8 km² area was converted to slight vulnerability. In moderate vulnerability, a 5573.8 km² area was converted to mild vulnerability. In severe vulnerability, a 2692.8 km² area shifted to moderate vulnerability. The area transferred from each vulnerability level to a higher level was small: the area transferred from mild to moderate vulnerability was 3100.7 km², and the area transferred from moderate to severe vulnerability was 2791.1 km², which had little influence on the overall trend of ecological vulnerability conversion in the study area. From 2010 to 2023, the overall trend of vulnerability was decreasing, and the sum of the area transferred from mild, moderate, severe, and extreme vulnerability to a low vulnerability level was 1818.81 km², 6169.95 km², 2229.57 km², and 50.4 km², respectively.

Between 2010 and 2023, 62.2% of the study area maintained the current vulnerability level, 21% of the area shifted to a low vulnerability level, and 16.8% of the area increased in vulnerability level, as shown in Figure 8. In terms of the spatial distribution pattern, the central plains area maintained a moderate vulnerability grade in general, and some towns and cities shifted from moderate and severe to severe and extreme vulnerability grades. The western region had a large area of declining ecological vulnerability, which was mainly characterized by the trend of “moderate–mild” and “mild–slight”, with some areas remaining slightly or mildly vulnerable. In the eastern region, the type of vulnerability transformation was more complex, with an overall shift from vulnerability to low vulnerability, which was reflected in the shift from severe, moderate, and mild vulnerability to moderate, mild, and slight vulnerability, respectively. Due to the large number and decentralized distribution of land use types in the region, there were various types of ecological vulnerability level transformations.

3.3.2. Trends in Vulnerability of Ecological Functional Protection Zones

From 2010 to 2023, among the six types of ecological functional reserves, the soil conservation functional area had the lowest average ecological vulnerability index, with an EVI of 0.371, as shown in Figure 9. This was followed by the forest product provision functional area, the water conservation functional area, the agricultural product provision functional area, the flood storage functional area, and the key town functional area, with average ecological vulnerability index values of 0.394, 0.417, 0.421, 0.477, and 0.519, respectively.

According to the area share of ecological vulnerability level in different ecological function protected areas, it can be seen (Figure 10) that from 2010 to 2015, the share of slightly, moderately, severely, and extremely vulnerable areas in each protected area showed an overall increasing trend, and the share of mildly vulnerable areas showed a decreasing trend. Except for the key town functional area where the degree of vulnerability decreased, the degree of vulnerability of each protected area increased. Among them, the total ratio of moderately and severely vulnerable areas in the functional area of agricultural product provision increased by 11.8% compared with 2010, and the degree of vulnerability significantly increased compared with 2010. From 2015 to 2023, the slightly and mildly vulnerable areas in the functional area of water conservation increased substantially, and the total ratio of slightly and mildly vulnerable areas increased by 15.21%. The area of moderate, severe, and extreme vulnerability decreased; the percentage of severe vulnerability area decreased by 7.78%; the degree of ecological vulnerability decreased significantly; and the ecological environment improved significantly. The remaining protected areas as a whole showed a trend of decreasing the proportion of areas with low vulnerability and increasing the proportion of areas with high vulnerability.

3.4. Analysis of Ecological Vulnerability Drivers

3.4.1. Indicator Correlation Analysis and Covariance Diagnosis

In order to take into account the selection criteria for the interaction of different indicators in the PSRM model and to eliminate the bias of the calculation results caused by the possible influence of indicator covariance, Pearson correlation coefficient analysis and multivariate covariance tests were conducted for each indicator. The degree of correlation between indicators is usually measured by the correlation coefficient, and if there is a significant correlation between the indicators, this indicates the existence of analytical significance. The absolute value of the correlation coefficient ranges from 0 to 1. The larger the absolute value, the stronger the correlation, and vice versa. If the VIF value of the indicators is greater than 7.5, this indicates that there is multicollinearity between the indicators.

The results of the correlation coefficient calculation are shown in Figure 11. Except for B4 and B6, the p-value of each indicator is less than 0.001, which indicates that it is significantly correlated at the 99.9% confidence level, and the correlation between the indicators is strong. The selection of indicators meets the modeling criteria for the interaction between indicators and has regression significance. According to the covariance diagnostic (

Table 4. Covariance results for each indicator.

Indicator	Tolerance	VIF
A1	0.740	1.35
A2	0.645	1.55
A3	0.648	1.54
A4	0.085	11.72
B1	0.115	7.51
B2	0.028	35.55
B3	0.088	11.43
B4	0.951	1.05
B5	0.287	3.48
B6	0.036	27.60
C1	0.224	4.46
C2	0.058	17.18
C3	0.548	1.82
C4	0.822	1.22
D1	0.280	3.57
D2	0.273	3.67

), the VIF values of indicators A4, B1, B2, B3, B6, and C2 are all greater than 7.5, indicating that there is a more obvious multiple covariance between the above indicators. If the conventional method was chosen to analyze the contribution of the factors at this time, it would lead to a decrease in the accuracy of the calculation, affecting the results of the analysis. Therefore, ridge regression was chosen for the calculation of the indicator driving force in this study.

3.4.2. Factors Influencing Ecological Vulnerability

The ridge regression model was used to analyze the factors affecting the ecological vulnerability of the Hubei section of the YREB, and the results are shown in (

Table 5. Results of ridge regression analysis.

K = 0.196	Non-Standardized Coefficient		Standardized Coefficient	T Value	p	R2
	Regression Coefficient B	Standard Error	Standard Coefficients Beta
A1	0.207	0.001	0.032	150.865	0.000 ***	0.968
A2	0.105	0.001	0.027	123.254	0.000 ***
A3	0.116	0	0.178	813.538	0.000 ***
A4	0.209	0	0.197	873.237	0.000 ***
B1	0.005	0	0.005	22.583	0.000 ***
B2	-0.019	0	−0.02	−112.39	0.000 ***
B3	-0.025	0	−0.023	−101.469	0.000 ***
B4	0.1	0.001	0.021	101.587	0.000 ***
B5	0.065	0	0.109	465.336	0.000 ***
B6	0.031	0	0.027	134.009	0.000 ***
C1	0.157	0	0.344	1373.421	0.000 ***
C2	0.114	0	0.246	1185.908	0.000 ***
C3	0.161	0	0.149	647.541	0.000 ***
C4	0.133	0	0.162	777.082	0.000 ***
D1	0.106	0	0.175	740.633	0.000 ***
D2	0.157	0	0.197	856.449	0.000 ***

Note: ***: p ≤ 0.001.

). Combined with the ridge trace plot, the ridge traces of the indicators tend to be stable when k is 0.196, and the R² value of the model is 0.968 (p <= 0.001), indicating that the fitting equations of the 16 indicators are able to explain 96.8% of the changes in the ecological vulnerability index, and the model fits well.

From the perspective of model coefficients, the p-values of the 16 indicators were all less than 0.001, indicating that the model regression results passed at least 99.9% of the significance test. In terms of single indicators, except for elevation and terrain undulation, all the other indicators were positively driven. Among them, the indicator with the greatest influence on ecological vulnerability was the human interference index, with a regression coefficient of 0.209, followed by population density, vegetation cover, the habitat quality index, and the greening area rate of built-up areas, with regression coefficients of 0.207, 0.161, 0.157, and 0.157, respectively. Slope, elevation, and topographic relief were relatively weak drivers of ecological vulnerability. Among the four criterion layers, the pressure layer had the strongest influence on ecological vulnerability, with an average regression coefficient of 0.159, followed by the response layer, management layer, and state layer, with average regression coefficients of 0.141, 0.132, and 0.041, respectively. Combining the dimensions of the single indicators and criterion layers, human activities were the strongest driving force on the change in ecological vulnerability index in the region.

4. Discussion

4.1. Spatiotemporal Evolution Pattern of Ecological Vulnerability in the Hubei Section of the YREB

The overall ecological quality of the Hubei section of the YREB is medium, and the vulnerability level is moderately fragile, which is consistent with the results of the existing ecological environment evaluation in related areas [24,55]. In terms of spatial distribution, the ecological vulnerability in the study area generally exhibits a pattern of higher values in the central region and lower values in the eastern and western parts. This pattern is influenced by the fact that much of the central region consists of flat terrain with dense population and frequent human activities and industrial production, leading to severe air pollution. Additionally, high temperatures and rapid urbanization in the central area contribute to limited land availability and low vegetation coverage, resulting in relatively poor ecosystem stability. Furthermore, human activities, such as irrational cultivation, extensive farming, and excessive grazing, exacerbate environmental degradation, leading to higher ecological vulnerability in this region. Areas of slight and mild vulnerability are primarily concentrated in the western region, where most areas are located within water conservation zones. These areas benefit from favorable hydrothermal conditions, high vegetation coverage, extensive forest areas, and a smaller population, resulting in weaker human impact on the ecosystem and lower environmental pressure. As a result, the western section of the YREB in Hubei Province exhibits lower ecological vulnerability and better environmental conditions. From 2010 to 2023, ecological vulnerability in the study area showed a decreasing trend. This is attributed to the implementation of policies such as the “YREB Development Outline,” the overall plan for the development of the YREB in Hubei Province (2009–2020), and the “14th Five-Year Plan for Green Development in Hubei Province.” Additionally, projects like “Greening the Jingchu Region” and the integrated protection of mountains, rivers, forests, farmlands, lakes, grasslands, and sands have significantly contributed to ecosystem restoration and protection. In terms of the degree of spatial agglomeration, the Theil index first declined and then increased from 2010 to 2023, with an overall “V”-shaped fluctuation trend, indicating that the high and low values of ecological vulnerability in the study area tended to be centrally distributed, with stronger interaction force of vulnerability indexes, and the distribution of hot- and cold-spots was essentially consistent with the spatial pattern of vulnerability.

4.2. Trends in Ecological Vulnerability in Six Types of Ecological Functional Protection Zones

The average EVI of each ecological function reserve was B > C > D > A > E > F. The EVI of A showed a decreasing trend, and the rest of the districts showed an increasing trend. The land use type of the A area is mostly forest, the vegetation cover is higher than the rest of the districts, and it shows a trend of increasing year by year; moreover, the mountainous and hilly area is wider, farther away from the center of the urban area, with a small population density, and the intensity of the disturbance caused by the production and living activities of mankind is low; therefore, the EVI is the lowest. With the implementation of measures such as “Promote development with water” and the construction of ecological civilization demonstration zones in the study area, the ecological environment in the water conservation function area was restored, and the ecological vulnerability index decreased significantly. B is urban land, and the rapid population growth and urbanization pressure are the reasons for the increase in the ecological vulnerability index in this region.

4.3. Drivers of Ecological Vulnerability

According to the driving force analysis, compared with natural factors, anthropogenic factors have a stronger influence on the spatial and temporal changes in ecological vulnerability in the Hubei section of the YREB, where HDI becomes a key factor regulating the dynamic evolution of ecological vulnerability in the region. This result is consistent with the findings of [25] that human activities are the dominant factor leading to changes in ecological vulnerability in the Yangtze River Basin. Human activities as a whole have an important impact on the changes in ecological vulnerability in the Hubei section of the YREB from multiple aspects and dimensions, such as production and living activities. As the region is the center of the YREB, there are frequent socio-economic activities along the route, and most of the area is located in the plains, which is suitable for human habitation in terms of topography, so the region is densely populated, and any human activities will have a direct impact on the regional ecological environment. In addition, the region’s strategic positioning and future development policies cause the region’s economic attributes to be weighted higher.

4.4. Recommendations for Ecologically Sustainable Development

As the central core of the YREB, the Hubei section has a key role to play in the sustainable development of the YREB. For key urban functional areas (Wuhan, Ezhou, and Huanggang cities), which exhibit the highest EVI (0.519), future strategies should prioritize ecological restoration coupled with industrial green transition. This includes mandatory low-emission retrofitting for high-polluting industries (e.g., steel and chemical sectors) and the establishment of ecological corridors linking critical nodes such as East Lake and Yangtze River wetlands. In water conservation function areas (Enshi Tujia and Miao Autonomous Prefecture, western Yichang), characterized by the lowest EVI (0.371), conservation efforts must focus on delineating core protected zones to maintain soil and water integrity. The implementation of an ecological compensation mechanism and agricultural intensification—such as piloting vertical farming in mountainous areas—are essential to balance ecological and economic goals. Soil conservation functional areas (western Jianghan Plain and parts of Huangshi), with a moderate EVI (0.417), require integrated soil protection and ecological rehabilitation. Key measures include adopting crop rotation and fallow systems, enhancing carbon sequestration through organic amendments, and prioritizing natural restoration in forested areas to minimize anthropogenic disturbances, thereby gradually reducing vulnerability while stabilizing current ecological conditions. Flood storage functional areas (e.g., Honghu Lake), display a relatively high EVI (0.477), necessitating rigorous water quality management. Strategies include enforcing strict controls on industrial effluent discharge, mitigating agricultural non-point source pollution, and eliminating untreated sewage inflows. For agricultural supply areas (eastern Jianghan Plain and northern Xianning), with a moderately high EVI (0.421), transitioning to eco-agricultural practices is critical. Examples include remediating soil contamination through phytoremediation and implementing straw return programs to curb erosion. Adaptive management frameworks should be established to foster synergies among “production–ecology–economy” objectives. In forest product supply areas (Shennongjia Forestry District and remote Yichang), where the EVI is relatively low (0.394), sustainable forest management and ecological value realization are paramount. Priorities include protecting primary forests, developing non-timber forest product economies, and advancing carbon sink trading initiatives. Degraded forests should undergo targeted rehabilitation to enhance ecosystem stability, while ecological networks must be established to improve habitat connectivity and biodiversity conservation.

4.5. Limitations and Future Directions

The authors of this study transcend conventional unidimensional evaluation systems by innovatively constructing an integrated analytical framework that couples the PSRM model with the G1–CRITIC–game theory combination weighting method. By further incorporating the Theil index, hot-spot analysis, and ridge regression modeling, the multi-scale spatiotemporal heterogeneity patterns of ecological vulnerability in the Hubei section of the YREB are systematically unraveled. However, this study has several limitations. Since the data of the statistical yearbook are only recorded until 2022, the data of the indicators of the sewage treatment rate and the rate of greening area in built-up areas are replaced by the most recent available years, which may result in subtle changes between different years not being captured. Therefore, in future studies, the timeliness of the data should be ensured as much as possible so as to improve the accuracy of the evaluation results. In addition, the time span of this study is from 2010 to 2023, which lacks the prediction of the future development of ecological vulnerability. Future studies can apply scientific prediction models to broaden the span of analysis and predict future trends in ecological vulnerability development to address the governance needs of ecologically fragile areas in the context of global change.

5. Conclusions

(1)

Spatially, there was significant variation in the distribution of ecological vulnerability within the study area, with the eastern and central regions showing higher vulnerability compared to the western region. Temporally, from 2010 to 2023, the overall ecological vulnerability in the Hubei section of the Yangtze River Economic Belt remained at a moderate level, with a trend of initially increasing and then decreasing vulnerability. The area of severe vulnerability increased from 2010 to 2015, but from 2015 to 2023, there was a noticeable decrease in vulnerability, particularly in the moderate, severe, and extreme categories.

(2)

The Theil index values for the YREB in Hubei from 2010 to 2023 were 0.371, 0.365, and 0.398, showing an overall “V”-shaped upward trend. This indicates that the spatial pattern of ecological vulnerability in the study area was concentrated and exhibited a decreasing-then-increasing degree of spatial aggregation. Hot-spots were mainly concentrated in the eastern and central areas, while cold-spots were primarily observed in the western region. From 2010 to 2015, the central hot-spot area decreased, while the western cold-spot area expanded. From 2015 to 2023, both the central hot-spots and eastern cold-spots increased in size.

(3)

Spatially, the central plain area of the study region generally maintained a moderate level of vulnerability, but some urban areas transitioned from moderate or severe to severe or extreme vulnerability. The western region saw a significant decrease in vulnerability levels. Temporally, from 2010 to 2023, approximately 62.2% of the vulnerable area remained at the same level, with a trend of transitioning from higher to lower vulnerability levels overall.

(4)

Among the six types of ecological function conservation areas, soil conservation areas had the highest proportion of low and moderate vulnerability, while key urban areas had a high proportion of severe and extreme vulnerability (up to 90%). From 2010 to 2023, except for water conservation areas, overall ecological vulnerability increased in other protected areas.

(5)

Based on ridge regression modeling, HDI was the main driving factor affecting ecological vulnerability in the study area. Among the four criterion layers, the pressure layer had the highest average regression coefficient, indicating its significant impact. When considering individual indicators and criterion dimensions, human activities were the strongest driver of changes in ecological vulnerability index in this research area.