Spatiotemporal Evolution and Key Factors of Coupling Coordination Between Water Ecological Carrying Capacity and Urbanization Quality: A Case Study of Hubei Province in the Yangtze River Economic Belt

Abstract

The coupling coordination between Urbanization Quality (UQ) and Water Ecological Carrying Capacity (WECC) represents a critical nexus for sustainable regional development within the Yangtze River Economic Belt (YREB). Focusing on 16 cities in Hubei Province over the period 2020–2024, this study constructed comprehensive indicator systems for UQ and WECC, Spatial Autocorrelation Analysis and Key Factor Analysis are then applied to analyze spatiotemporal evolution, identify key influencing factors. The results reveal that: (1) Both UQ and WECC demonstrated upward trajectories, with UQ increasing from 0.369 to 0.409, although WECC exhibited fluctuating patterns; (2) Spatial analysis identified pronounced “core–periphery” clustering effects with Wuhan as the dominant center, confirmed by the positive Global Moran’s I; (3) Hubei’s CCD advanced from 0.626 to 0.661, progressing toward initially coordinated stages, with Wuhan pioneering this transition, while 81.25% of cities remained at the moderately coordinated stage; (4) Grey relational analysis identified aquatic biological resources as the principal constraint, with piscivore biomass ratios and pension insurance participation rates (γ = 0.752) emerging as key biophysical and socioeconomic drivers, respectively. These findings provide empirical evidence for targeted interventions promoting balanced urban–water ecological development in the YREB, while contributing a novel analytical framework for examining UQ-WECC interactions in rapidly urbanizing regions globally.

1. Introduction

Urbanization is a global multi-dimensional process accompanied by increasing uncertainty due to climate change, human migration, and changes in the capacity to sustain ecosystem services [1]. China has experienced a rapid period of urbanization since the 1980s, which has not only reshaped the nation’s socioeconomic landscape but also provided crucial insights for global urbanization [2,3]. The Yangtze River Economic Belt (YREB), accounting for 43.1% of the national population and 45.2% of the GDP while holding 47% of the nation’s hydropower resources, has emerged as a key engine for this development [4]. However, rapid urbanization has imposed significant strain on the basin’s water resources and environment [5,6]. These challenges manifest as localized structural water shortages, exacerbated by spatiotemporally uneven distribution and extreme climate events [7], the eutrophication of lakes and reservoirs, excessive total phosphorus levels in the main channel, and frequent flooding incidents [8,9]. Water pollution disrupts aquatic food chains and diminishes ecosystem services, while the substantial discharge of wastewater further deteriorates aquatic ecological environments, imposing tremendous pressure on the basin’s sustainable development [6,10].

In response to water resource management challenges, the concept of the water resource carrying capacity (WRCC) was formulated. Its initial definition was the maximum scale of socioeconomic activities that a region’s water resources can sustain, subject to the condition that social and ecological systems remain uncompromised [11,12,13]. As research progressed, scholars gradually adopted a systemic perspective, recognizing water systems as complex entities involving intricate environmental and biological interactions [14]. However, the traditional WRCC framework—centered predominantly on the physicochemical assimilative capacity of water bodies—has yet to fully embrace this systemic vision. It reflects a fundamental dichotomy between hydrological engineering and ecological approaches that has long pervaded water resource management: the compartmentalized view of nature leads hydrologists to lose sight of opportunities for transdisciplinary engagement, while ecologists lack the hydrological tools to address problems related to ecosystem hydrology. This fragmentation brings about counterproductive risks—the management of one set of problems may exacerbate the other, with no functional understanding of potential trade-offs or synergistic pathways [12,15].

To overcome this limitation, the Water Ecological Carrying Capacity (WECC) has emerged as a comprehensive framework encompassing water quality, ecosystem health, resource availability, population dynamics, and economic scale [11,16]. Grounded in the ecohydrology paradigm, WECC treats water resources and biological communities as interconnected components. Aquatic organisms serve dual functions: maintaining ecosystem services through oxygen production, water purification, and trophic regulation [17,18], while acting as sensitive biological indicators of environmental quality [12,19]. Overexploitation and pollution—including eutrophication and heavy metal contamination—cause habitat destruction and biodiversity declines, underscoring the necessity of incorporating aquatic biodiversity into carrying capacity assessments.

A growing body of literature has investigated the relationship between UQ and WECC within the YREB, revealing two paradigms: one integrates UQ as a subsystem within the WECC framework (or vice versa) [18], while the other treats them as interdependent systems and analyzes their coupling coordination. Methodologically, existing WECC evaluation approaches fall into two categories. The first regards WECC as a system dynamics issue, employing concepts such as ecological footprints, cloud models, agent-based models, and fuzzy mathematics [10,20,21,22]. The second treats WECC as a component of regional sustainable development, utilizing integrated evaluation methods including principal component analysis, AHP, CRITIC, TOPSIS, and entropy weight methods [23]. To address methodological limitations in weight determination, combined approaches such as CRITIC, TOPSIS, and entropy weight–TOPSIS have been employed in recent years [18].

Despite the conceptual shift toward integrating aquatic biota into WECC, this paradigm has not yet been adequately reflected in empirical research [19]. Three critical gaps persist. First, existing studies predominantly employ static assessments at single time points or within individual administrative regions, and the lack of cross-regional comparative analysis further constrains the identification of regional heterogeneity patterns and the development of generalizable assessment frameworks [10]. Second, aquatic biodiversity indicators remain largely absent from operational WECC evaluation frameworks, which still rely on conventional metrics such as total water resources and water quality, resulting in the incomplete characterization of aquatic ecosystem health [12]. Third, the dynamic interaction patterns and spatial spillover effects between UQ and WECC under accelerating urbanization remain insufficiently explored, where the driving pathways and threshold effects of key factors under rigid water resource constraints are yet to be clarified [24,25]. Therefore, it is imperative to adopt dynamic analytical approaches such as panel data modeling and spatiotemporally weighted regression to construct a multi-dimensional WECC evaluation framework that incorporates aquatic biodiversity, thereby systematically revealing the spatiotemporal evolution and multi-factor driving mechanisms of UQ-WECC coupling coordination across multiple urban scales and providing a scientific basis for differentiated regional sustainability strategies.

The objectives of this study address these needs and are as follows. Firstly, this study will construct a comprehensive index system for evaluating WECC based on the DPSIR framework. Central to our framework’s novelty is the systematic incorporation of aquatic biodiversity indicators (D3, S3–S4, I2–I3), which redirects the assessment focus from resource availability toward ecosystem integrity—a fundamental extension of the WECC paradigm; it further enriches the UQ system by integrating a broader set of social equity indicators (SO2, CC1–CC4). Secondly, regarding the methodology, the improved TOPSIS method will be employed for weight determination, which involves initial weight allocation through three relatively independent models, followed by redistribution based on credibility coefficients. Building upon these optimizations, spatiotemporal evaluations and clustering analyses will be conducted for all cities within the study area to reveal evolutionary trends and inter-city spatial disparities. The findings are intended to provide theoretical support for understanding the UQ-WECC relationship and to offer a replicable methodological reference for WECC assessment in other regions.

2. Materials and Data Resources

2.1. Study Area

Hubei Province (29°01′ N–33°06′ N, 108°21′ E–116°07′ E) is situated in the middle part of the YREB, in the heart of Central China (Figure 1). It covers an area of about 185,900 km² and has a population of 58.3 million. It is bounded by Henan Province to the north, Anhui to the east, Jiangxi and Hunan to the south, Chongqing to the west, and Shaanxi to the northwest. Hubei is also known as the transportation center linking Eastern and Western China.

The Yangtze River flows west to east across the province for approximately 1061 km, giving Hubei the longest stretch of the river’s main channel among all provinces. This geographical advantage has positioned Hubei as a pivotal node in the YREB development strategy. The study area encompasses sixteen prefecture-level cities directly administered by Hubei Province.

The Han River, the largest tributary of the Yangtze River, flows southeast through Hubei province and eventually converges with the Yangtze at Wuhan. The Han River’s basin occupies about 174,300 km²; it is the largest tributary basin in the Yangtze River system. In addition to this major fluvial artery, Hubei possesses abundant lacustrine resources, with 755 lakes currently included in the provincial lake protection system, of which 231 have a surface area exceeding 1 km². This remarkable lake density has earned Hubei the epithet “Province of a Thousand Lakes”. Because these lakes respond rapidly to changes in water flow and chemical composition, they are well suited for monitoring the aquatic ecological health of the Yangtze basin—detecting early signs of degradation and providing a basis for developing effective conservation strategies across the region.

2.2. Data Resources

The data used in this study primarily include water resource data, aquatic biological resource data, socioeconomic statistics, and geographic information data from 2019 to 2024. Water resource data mainly originate from the water resource bulletins of various cities in Hubei Province and publicly available statistical information from the Water Resources and Lakes Bureau. Aquatic biological resource data are primarily derived from the Bulletin on Aquatic Biological Resources and Habitat Status in the Yangtze River Basin and publicly released statistical data from the Hubei Provincial Department of Agriculture and Rural Affairs. Socioeconomic data primarily originate from the Hubei Statistical Yearbook, municipal statistical yearbooks, and statistical bulletins.

3. Research Methods

The research framework of the study is illustrated in the figure below (Figure 2). This research focuses on three main aspects: (1) Steps 1 and 2 describe the UQ and WECC of cities in Hubei Province, along with their spatiotemporal variation analysis; (2) Step 3 provides the CCD of cities in Hubei Province and their visualization; (3) Step 4 presents critical factors influencing the CCD and their impact pathways.

3.1. Establishment of Index System

3.1.1. Index System for UQ

To enable a thorough analysis of the coupling coordination relationship between UQ and WECC, it is necessary to understand their respective evaluation frameworks. Scholars have traditionally assessed UQ by deconstructing urbanization into four subsystems: population urbanization, economic urbanization, social urbanization, and spatial urbanization [24,26,27]. Besides these four subsystems, environment, infrastructure, and demographic criteria are also frequently taken into account [28,29]. As urbanization research has deepened, scholars have expanded UQ into five subsystems, which has become an established convention [4,5,17]. Building upon these foundations, this study establishes an urbanization indicator system encompassing demographic urbanization (DE), economic urbanization (EC), infrastructure urbanization (IF), social urbanization (SO), and city–countryside coordination urbanization (CC) (

Table 1. Indicator system for Urbanization Quality (UQ).

System	Criterion	Indicator	Attribute	Weight	Reference
Urbanization Quality (UQ)	Demographic (DE)	Proportion of urban population (DE1)	+	0.033	[24]
		College Students (per 10,000 people) (DE2)	+	0.088	[26]
		Invention Patents (per 10,000 people) (DE3)	+	0.095	[24]
	Economic (EC)	Per capita GDP (CNY/person) (E1)	+	0.063	[27,28,29]
		Theil index in Industrial Structure (E2)	+	0.062	[24,30,31]
		Consumer Market Activity Index (E3)	+	0.068	[32]
		High-Tech Industry Proportion in GDP (E4)	+	0.05	[24,33]
	Infrastructure (IF)	Public Library Collection Size (per 10,000 people) (IF1)	+	0.069	[32,33]
		Urban Air Quality Composite Index (IF2)	−	0.045	[34,35,36]
		Per Capita Staple Food Production (IF3)	+	0.072	[29]
	Social (SO)	Health Technicians (per 10,000 people) (SO1)	+	0.065	[4,32]
		Proportion of Urban Residents in Elderly Care Insurance (SO2)	+	0.062	[21,37]
		Per Capita Government Expenditure (SO3)	+	0.051	[25,38]
	City–Countryside Coordination (CC)	Urban–Rural Resident Balance (CC1)	−	0.048	[31,39]
		Registered Resident Population (CC2)	−	0.023	[14]
		Primary Distribution of Income Ratio (CC3)	+	0.063	[14,39]
		Per Capita Rural Resident Medical Expenditure (CC4)	−	0.045	[14,37]

Note: A higher positive indicator (+) signifies better UQ, while a higher negative indicator (−) indicates poorer UQ.

).

(1)

Demographic Urbanization (DE). The degree of population and innovation capacity, which are fundamental driving forces for regional development [1]. Proportion of urban population (DE1) indicates the level of population agglomeration in urban areas. College students (per 10,000 people) (DE2) and invention patents (per 10,000 people) (DE3) represent the innovation potential and technological advancement capacity of a region, which are critical for sustainable development [24,26].

(2)

Economic Urbanization (EC). Economic development plays a crucial role in urbanization by creating job opportunities, improving infrastructure, and attracting investment. It drives migration from rural to urban areas, as people seek better their living standards and employment prospects. Per capita GDP (E1) is adopted to represent the overall economic development level [27,28,29]. The Theil index (E2) reflects the balance and rationality of the economy [24,30,31]. To evaluate market vitality and the upgrading process of the industrial structure—key determinants of urbanization quality—the consumer market activity index (E3) and high-tech industry proportion in GDP (E4) are selected [24,32,33].

(3)

Infrastructure Urbanization (IF). Some scholars argue that regional infrastructure development should be reflected from two dimensions: physical infrastructure construction and soft infrastructure development [33]. Therefore, for infrastructure urbanization, we consider dimensions such as cultural facility development, urban air quality levels, and food supply stability levels. The public library collection size (per 10,000 people) (IF1) represents the cultural infrastructure level [32,33]. The urban air quality composite index (IF2) reflects environmental infrastructure quality [34,35,36]. Per capita staple food production (IF3) is chosen to evaluate food security, which provides essential support for residents’ well-being [29].

(4)

Social Urbanization (SO). This dimension reflects the social security system and public health service level, which are important guarantees for residents’ quality of life [4,32]. Health technicians (per 10,000 people) (SO1) serves as a critical indicator for assessing a city’s public health level, as it directly determines the accessibility, quality, and equity of medical services. The proportion of urban residents in elderly care insurance (SO2) [21,37] and per capita government expenditure (SO3) represent social security coverage and government investment in public services [25,38]. These indicators reflect the levels of social welfare and protection.

(5)

City–Countryside Coordination Urbanization (CC). Empirical studies confirm that both income inequality and financial development have significant ecological effects [39]. Thus, it is necessary to incorporate equity indicators into research. City–countryside coordination is crucial for achieving comprehensive regional development. The urban–rural resident balance (C1), city–countryside registered resident population (C2), primary distribution of income ratio (C3), and per capita rural resident medical expenditure (C4) are adopted to evaluate the coordination level between urban and rural development, which helps to reduce regional disparities and promote equitable development [14,31,37,39].

3.1.2. Index System for WECC

The United Nations Economic Cooperation Organization (UNECO) proposed the Pressure–State–Response (PSR) model framework to assess the global environmental condition, which was soon adopted by the European Environment Agency (EEA). Subsequently, improved frameworks including the Pressure–State–Impact–Response (PSIR) model and Driver–Pressure–State–Impact–Response (DPSIR) model were successively introduced. To date, the PSR framework and its improved iterations have been widely applied in environmental impact assessments, sustainable development, and ecological risk assessments [40]. Compared to the PSR basic model, the DPSIR model incorporates additional “Driver” and “Impact” components, enabling a clearer depiction of environmental pressures, changes in system state, and their consequences for human society and ecosystems [16,18]. Moreover, analysis of the internal relationships among the water environment, water resources, and aquatic biological resources reveals that these three elements are not only influenced by multiple factors but also exhibit intricate connections with the quality of urbanization. Based on these two considerations, this study adopts the DPSIR framework to construct an indicator system for WECC (

Table 2. Indicator system for Water Ecological Carrying Capacity (WECC).

System	Criterion	Indicator	Attribute	Weight	Reference
Water Ecological Carrying Capacity (WECC)	Driver (D)	Wastewater Discharged Per Unit of GDP (D1)	−	0.047	[4,6,41]
		Electricity Consumption Per Unit of GDP (D2)	−	0.018	[41,42]
		Fishy Caught Per Unit Effort (D3)	−	0.068	[43]
	Pressure (P)	Per Capita Water Use (P1)	−	0.038	[10,16]
		Per Capita Aquatic Product Yield (P2)	−	0.049	[10,19]
		Per Capita Industrial Wastewater Discharge (P3)	+	0.033	[41]
		Pollution Monitoring Exceedance Rate (P4)	+	0.026	-
	State (S)	Per Capita Water Resources (S1)	−	0.075	[29,44]
		Water Yield Modulus (S2)	+	0.071	[43,45]
		Invasive Alien Species (S3)	−	0.027	-
		Fish Species Number (S4)	−	0.057	[43]
		Total Nitrogen (mg/L) (S5)	−	0.025	[24,46]
		Total Phosphorus (mg/L) (S6)	+	0.023	[17,19]
		pH in Water (S7)	+	0.044	[43,46]
	Impact (I)	Dissolved Oxygen Saturation Status in Water (I1)	+	0.060	[43,46]
		Proportion of Fish-Eating Fish (I2)	+	0.101	[43]
		Shannon–Wiener Index of Fish (I3)	−	0.050	[47]
		River Water Quality (I4)	−	0.071	[39]
	Response (R)	Water Consumption Per Industrial Value Added (R1)	−	0.030	[48,49]
		Energy Intensity Reduction Rate Per Unit GDP (R2)	−	0.027	[26]
		Government Environmental Protection Expenditure (R3)	+	0.063	[4,21]

Note: A higher positive indicator (+) signifies better WECC, while a higher negative indicator (−) indicates poorer WECC.

).

(1)

Driver (D). This subsystem reflects the stress on water resource consumption and aquatic ecosystem disturbances resulting from human activities [6]. Water consumption per unit of GDP (D1) and electricity consumption per unit of GDP (D2) represent the resource utilization efficiency and consumption pressure [4,40,41]. Caught Per Unit Effort (D3) is a core indicator in fishery assessments; more fish caught per unit of effort indicates a greater abundance of fish in the water [41,42,43].

(2)

Pressure (P). Pressure is the direct exertion of force to alter environmental conditions. Per capita water consumption (P1) and per capita aquatic product yield (P2) indicate the direct pressure on water resources [10,16,19]. Per capita industrial wastewater discharge (P3) and the pollution monitoring exceedance rate (P4) are selected to evaluate the pollution pressure on aquatic ecosystems [41].

(3)

State (S). State indicates changes in the physical, chemical, or biological conditions of water resources and aquatic ecosystem. Per capita water resources (S1) and the water yield modulus (S2) [29,43,44,45] represent the water resource availability. Invasive alien species (S3) and the fish species number (S4) indicate the biodiversity status. Total nitrogen (S5) [24,43,46], total phosphorus (S6) [17,19], and the pH in water (S7) are adopted to reflect the water quality status, as these are critical indicators for assessing aquatic ecosystem health [43,46].

(4)

Impact (I). This criterion reflects the ecological consequences and biological responses resulting from environmental pressures on aquatic ecosystems. The dissolved oxygen saturation status in water (I1) represents the fundamental condition for aquatic life [43,46]. The proportion of fish-eating fish (I2) and the Shannon–Wiener index of fish (I3) are selected to evaluate the fish community structure and ecosystem functioning, which indicate the overall impact on aquatic biodiversity and ecosystem stability [43,47]. River water quality (I4) indicators measure the health of a river, determining its safety for ecosystems and humans [39].

(5)

Response (R). This criterion reflects the management measures and policy interventions implemented to protect water resources and restore aquatic ecosystems [4]. Water consumption per industrial value added (R1) represents the improvement in water use efficiency [48,49]. The energy intensity reduction rate per unit GDP (R2) reflects the efforts to reduce environmental pressures through technological advancement [26]. Government environmental protection expenditure (R3) is adopted to evaluate the investment in environmental governance and restoration, which demonstrates a commitment to aquatic ecosystem protection and sustainable water resource management [4,21].

3.2. Improved TOPSIS

In the assessment of the coupling coordination between UQ and WECC, the limitations of single weighting methods have gradually become apparent: for subjective approaches such as the AHP, which are capable of incorporating expert experience, their effectiveness is susceptible to subjective cognition and mental rigidity, being particularly pronounced when confronted with complex indicator systems; conversely, objective methods including the entropy and CRITIC methods, albeit grounded in the data themselves, suffer from high sample sensitivity and tend to overlook ordinal distinctions or non-linear relationships. In response to the multi-dimensionality and heterogeneity inherent in urbanization–ecological environment data, combined weighting approaches have emerged as a critical direction in methodological evolution by systematically integrating the strengths of both subjective and objective methods to compensate for the structural deficiencies of individual approaches [10]. Among these, the entropy–AHP-CRITIC combined framework (improved TOPSIS) synthesizes objective data characteristics with subjective expert judgment, not only overcoming the one-sidedness of single methods but also enhancing the robustness and credibility of the evaluation results, thereby establishing itself as the predominant methodological paradigm within this research domain [10,11,23].

With established methodological frameworks, the extreme value normalization method was employed prior to entropy weight calculation. The entropy weight method is an objective weighting approach that determines indicator weights based on the degree of data dispersion, thereby mitigating subjective bias in weight assignment. The computational procedure is presented as follows.

(1)

Data Normalization

For positive indicators,

$$X^{\prime }={\displaystyle \frac{X-X_{min}}{X_{max}-X_{min}}}$$

(1)

For negative indicators,

$$X^{\prime }={\displaystyle \frac{X_{max}-X}{X_{max}-X_{min}}}$$

(2)

where X is the original indicator value, X′ is the normalized indicator value, and X_max and X_min are the maximum and minimum values of the indicators.

(2)

Entropy Calculation

The entropy weight is calculated through the following steps:

$$y_{\theta ij}={\displaystyle \frac{C_{ij}}{{\sum }_{i=1}^m{C}_{ij}}}$$

(3)

$$e_j=-{\displaystyle \frac{1}{\mathit{ln} m}}{{\displaystyle \sum }}_{i=1}^m{y}_{\theta ij}\mathit{ln}{y}_{\theta ij}$$

(4)

$$g_i=1-e_j$$

(5)

$$W_i={\displaystyle \frac{g_i}{{\sum }_{j=1}^n\left(1-e_j\right)}}$$

(6)

where Y_θij is the feature proportion, 0 ≤ i ≤ m, 0 ≤ j ≤ n; e_j is the entropy value of the indicator; g_j is the information utility value; and W_i is the entropy weight of the indicator.

(3)

AHP Method

The analytic hierarchy process (AHP) is a multi-criteria decision-making method that integrates qualitative and quantitative analysis; it is widely applied in comprehensive evaluations of carrying capacity and urban development:

$$Aw={\lambda }_{maxw}$$

(7)

where A is the pairwise comparison matrix; w is the priority weight vector; and λ_maxw is the maximum eigenvalue.

(4)

CRITIC method

Criteria Importance Through Intercriteria Correlation (CRITIC) is an objective weighting approach based on the correlations between criteria and their comparative intensities. It is suitable for determining criterion weights in multi-attribute decision-making problems. It effectively integrates the information content within criteria and the structural relationships among them, making it applicable to comprehensive evaluation scenarios that emphasize the inherent characteristics of the data. The calculation process is as follows:

$$r_{jk}={\displaystyle \frac{{\sum }_{i=1}^m\left(x_{ij}^{*}-\overline{x_j^{*}}\right)\left(x_{ik}^{*}-\overline{x_k^{*}}\right)}{\sqrt{{\sum }_{i=1}^m{\left(x_{ij}^{*}-\overline{x_j^{*}}\right)}^2{\sum }_{i=1}^m{\left(x_{ik}^{*}-\overline{x_k^{*}}\right)}^2}}}$$

(8)

$$C_j={{\displaystyle \sum }}_{k=1}^n\left(1-\left|r_{jk}\right|\right)$$

(9)

$$I_j={\mathsf{\sigma }}_j\times C_j$$

(10)

$$w_j^{CRITIC}={\displaystyle \frac{I_j}{{\sum }_{j=1}^n{I}_j}}$$

(11)

where r_jk is the Pearson correlation coefficient between criteria j and k; x_ij and x_ik are the standardized values of i alternative for criteria j and k; i is the alternative index; C_j is the conflict measure of criterion j; σ_j is the standard deviation of criterion j; I_j is the information content of criterion j; w_j^CRITIC is the objective weight of criterion j.

(5)

Integrated Weighting Approach

To overcome the inherent limitations of single weighting methodologies and achieve an optimal balance between subjective expert knowledge and objective data characteristics, this study adopts an integrated weighting approach grounded in credibility theory and information fusion principles:

$$C_i={\displaystyle \frac{R_i}{R_{AHP}+R_{CRITIC}+R_{ENTROPY}}}$$

(12)

$$w_j=C_{AHP}·w_j^{AHP}+C_{CRITIC}·w_j^{CRITIC}+C_{ENTROPY}·w_j^{ENTROPY}$$

(13)

where C_i specifically denotes the credibility coefficient of the method; I ∈ (AHP, CRITIC, ENTROPY) [11]; and C satisfies C_AHP + C_CRITIC + C_ENTROPY = 1.

3.3. Coupling Coordination Degree Model

The CCD model is employed to quantitatively characterize the interaction intensity and mutual influence among multiple systems. The urbanization process and ecological environment constitute an integrated system that necessitates balanced and coordinated development. The coupling mechanisms among subsystems govern the comprehensive development trajectory of the YREB and determine whether Hubei Province’s CCD evolves toward an ordered or disordered developmental state.

(1)

Coupling Degree Model

Based on the comprehensive evaluation indices derived from entropy weight calculations, the coupling degree between UQ and WECC for cities in Hubei Province is calculated using the following equation:

$$C={\displaystyle \frac{\sqrt{u_1\times u_2}}{\left(u_1+u_2\right)/2}}$$

(14)

where C quantifies the coupling degree between subsystems (here and hereafter, the notation C consistently denotes the coupling degree). U₁ and U₂ represent the scores of WECC and UQ.

However, this approach fails to adequately capture the comprehensive development level of the coupled systems. To address this methodological limitation, the comprehensive development index T and coupling coordination degree D are introduced.

(2)

Comprehensive Development Index

The comprehensive development index is calculated to assess the overall development level of the coupled systems:

$$T=\alpha U1+\beta U2$$

(15)

where α and β represent the contribution coefficients of UQ and WECC, respectively, with α + β = 1. In this study, α = β = 0.5, indicating the equal importance of both systems.

(3)

CCD Model

The CCD model integrates both the coupling degree and the comprehensive development level:

$$D=C\times T$$

(16)

where D represents the CCD between UQ and WECC, D ∈ [0, 1].

The CCD between UQ and WECC was classified using a hierarchical framework based on established thresholds in the literature. Following the distribution characteristics of the coordination index and drawing from previous studies [50,51], this research developed a ten-level classification system to characterize the coordination states between UQ and WECC (

Table 3. Classification standard for CCD between UQ and WECC.

D Classification	D Range	D Classification	D Range
Severely misaligned	0 ≤ D < 0.1	Barely coordinated	0.5 ≤ D < 0.6
Seriously misaligned	0.1 ≤ D < 0.2	Initially coordinated	0.6 ≤ D < 0.7
Moderately misaligned	0.2 ≤ D < 0.3	Moderately coordinated	0.7 ≤ D < 0.8
Mildly misaligned	0.3 ≤ D < 0.4	Well coordinated	0.8 ≤ D < 0.9
Nearly misaligned	0.4 ≤ D < 0.5	Highly coordinated	0.9 ≤ D ≤ 1

).

3.4. Spatial Autocorrelation Analysis

3.4.1. Global Moran’s I

To examine the spatial evolution characteristics of UQ and WECC, in this study, global spatial autocorrelation analysis was employed to detect spatial dependence patterns across the study area [51]. The Global Moran’s I statistic, a widely adopted measure in ecological and geographical research, is calculated as follows:

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

(17)

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

(18)

where x_i, x_j represent the attribute values of spatial units i and j; w_ij indicates elements of the spatial weight matrix defining neighborhood relationships; and S is the global mean of the attribute.

3.4.2. Local Moran’s I

Although the Global Moran’s I provides an overall measure of spatial association, it cannot identify specific locations exhibiting clustering or outlier behavior. Therefore, this study employed the Local Indicators of Spatial Association (LISA) [52] to decompose global patterns and detect local spatial clusters and outliers. The Local Moran’s I statistic was computed as follows:

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

(19)

$$S^2={\displaystyle \frac{1}{n}}{{\displaystyle \sum }}_{i=1}^n(x_i-\overline{x}{)}^2$$

(20)

3.5. Key Factor Analysis Method

3.5.1. Grey Relational Analysis

Multiple factors influence the CCD between UQ and WECC. Grey relational analysis was employed to identify key factors affecting the CCD within the UQ system and WECC system [53]. The calculation formula is as follows:

$${\gamma }_i={\displaystyle \frac{1}{n}}{{\displaystyle \sum }}_{k=1}^n{\displaystyle \frac{{\Delta }_{min}+\rho ·{\Delta }_{max}}{\left|x_0\left(k\right)-\frac{x_i\left(k\right)}{x_i\left(1\right)}\right|+\rho ·{\Delta }_{max}}}$$

(21)

where γ_i is the grey relational degree of indicator i; Δ_min and Δ_max represent the minimum and maximum absolute differences; ρ is the distinguishing coefficient; X₀(k) is the k-value of the reference sequence; and X_i(k) is the k-value of the comparative sequence for indicator i.

3.5.2. Spatial Durbin Model

The Spatial Durbin Model (SDM) effectively captures the dynamic interrelationships among system components under conditions of incomplete information and limited sample sizes, rendering it particularly suitable for analyzing complex urban–environmental interactions characterized by data limitations.

To further identify and assess the key factors influencing the CCD, and considering the characteristics of the available data, the Spatial Durbin Model (SDM) was employed to investigate the spatial influence mechanisms of critical factors:

$$y_{it}=\rho {{\displaystyle \sum }}_{j=1}^N{w}_{ij}{y}_{jt}+{{\displaystyle \sum }}_{k=1}^K{\beta }_k{x}_{k,it}+{{\displaystyle \sum }}_{k=1}^K{\theta }_k{{\displaystyle \sum }}_{j=1}^N{w}_{ij}{x}_{k,jt}+{\mu }_i+{\epsilon }_{it}$$

(22)

where y_it is the CCD of city i in year t; w_ij is an element of the spatial weight matrix W, representing the spatial relationship between city i and city j; ρ is the spatial autoregressive coefficient; x_k,it is the value of explanatory variable k for city i in year t; β_k is the direct effect coefficient of explanatory variable k; θ_k is the spatial lag coefficient of explanatory variable k; μ_i represents the random individual effect of city i; and ε_it is the random error term.

4. Results and Discussion

4.1. Temporal Evolution of UQ and WECC

In terms of the overall development level, the UQ of Hubei Province presented an overall upward trend from 2020 to 2024 (Figure 3). From 2020 to 2024, the UQ score of Hubei Province increased from 0.369 to 0.409, with an increase of 10.8%. Although there was a small decrease in 2021, the overall upward trend continued. Therefore, it can be seen that the UQ of Hubei is developing, and the overall CCD is constantly improving.

Based on the specific numerical data analysis, the spatial disparity of the UQ in Hubei Province from 2020 to 2024 presented a “decline followed by rise” pattern. The coefficient of variation decreased from 0.183 in 2020 to 0.177 in 2021 and then increased gradually to 0.195 in 2024. This means that the gap in regional development has been gradually amplified in recent years. In particular, the gap between Wuhan and other cities has continually increased. It should be noted that Wuhan’s score in 2024 was greater than the provincial average by 53.8%.

From the point of view of the overall level of WECC, the overall average value of Hubei Province showed an oscillating trend consisting of “first rising, then declining, and then recovery” from 2020 to 2024. In particular, the overall average value of WECC increased from 0.371 in 2020 to 0.421 in 2021 (the highest score in the whole period); it then decreased to 0.373 in 2022 and 0.375 in 2023 and then recovered to 0.393 in 2024. It can be seen that the overall average value of Hubei Province’s WECC is generally trending, and there are certain interannual fluctuations. This result is consistent with the conclusions of Lu (2022) [49]. Although the research methods, time scales, and spatial ranges of the two studies are different, our findings align with the evolutionary pattern of Hubei Province’s WECC and confirm the reliability of the research results in the mentioned paper.

Analyzing the detailed levels of UQ for each city (Figure 4), this research finds that there are still large differences in the levels of UQ in Hubei’s cities. Wuhan obtained the highest score in 2024 (0.629), and Xiaogan obtained the lowest score (0.347), which was only 55.2% of Wuhan’s score. Wuhan had a leading position throughout the entire research period. In each year, the score of Wuhan was more than 50% higher than the provincial average, showing that Wuhan was still a national central city with considerable strength in terms of economic development and innovation capacity. There was a clear “center–periphery” spatial differentiation pattern. Based on the temporal evolution, this research found that the UQ scores of the cities showed different changing trends from 2020 to 2024. Wuhan maintained a steady and consistent improvement, and it was always the first in the province. Yichang improved most obviously, and the improvement rate was 23.4%.

Assessing WECC implementation in different cities (Figure 5), it can be seen that there are still obvious differences in the level of WECC between cities in Hubei Province in 2024. The score of Xiangyang is the highest, reaching 0.526, while the score of Ezhou is the lowest, at only 0.292, accounting for 55.5% of Xiangyang’s score. The score of Wuhan is slightly higher than the provincial average, ranking at the upper-middle level among all cities. In terms of time, the changing trend of WECC in different cities from 2020 to 2024 showed a different pattern. Xiangyang, Xiantao, and Suizhou showed a rapid growth trend. Wuhan was overall stable, with its score increasing from 0.385 in 2020 to 0.409 in 2024—an increase of 6.2%. Although it reached its peak in 2021, it returned to a stable state in the following years. Differing from this, cities like Ezhou, Jingzhou, and Tianmen showed a decreasing trend.

4.2. Spatial Variation in UQ and WECC

According to the results of the temporal analysis of UQ and WECC, K = 5 was selected based on the optimal results of the algorithm to carry out spatial clustering analysis and spatial autocorrelation analysis (Figure 6) [53]. The clustering analysis showed that, as the provincial capital, Wuhan remained in an indisputable core position among all regions, while Xiangyang and Yichang formed secondary growth poles. The temporal evolution had two obvious phases: coordinated development in 2020–2021 and a state of imbalance and oscillation in 2022. From 2023 to 2024, a new “1 + 2 + N” pattern of differentiation and reconfiguration was gradually formed, and the gap between cities continued to widen. The spatial spillover effects were reflected in the following aspects: Wuhan’s influence on surrounding cities gradually strengthened every year—that is, the positive spillover of surrounding cities to the central city was gradually weakened. For example, Enshi Prefecture, in the far west of Hubei Province, was geographically isolated and could not benefit from the radiation from the central city.

Based on the spatial autocorrelation analysis (

Table 4. Global Moran’s I parameter for WECC and UQ.

System	Year	K-Value	Moran’s I	SD	Z	p-Value
WECC	2020	5	0.552 ***	0.027	0.164	3.760
WECC	2021	5	0.277 **	0.019	0.137	2.510
WECC	2022	5	0.595 ***	0.027	0.165	4.013
WECC	2023	5	0.179 *	0.020	0.142	1.729
WECC	2024	5	0.481 ***	0.023	0.151	3.623
UQ	2020	5	0.242 ***	0.005	0.068	4.545
UQ	2021	5	0.278 ***	0.010	0.100	3.440
UQ	2022	5	0.289 ***	0.008	0.092	3.858
UQ	2023	5	0.272 ***	0.006	0.079	4.297
UQ	2024	5	0.225 ***	0.007	0.082	3.545

Note: * p < 0.10, ** p < 0.05, *** p < 0.01.

), Hubei Province demonstrates significant positive spatial clustering in both UQ and WECC. The Global Moran’s I for WECC exhibits strong yet fluctuating spatial correlations during 2020–2024, increasing from 0.179 in 2020 to 0.595 in 2022 (p < 0.01). In contrast, UQ shows greater stability, maintaining values between 0.225 and 0.289, with p < 0.01, and revealing a “club convergence” phenomenon. Both indicators exhibit distinct H-H and L-L clustering patterns, suggesting robust spatial clustering effects and significant spillover dynamics among contiguous regions.

The Local Moran’s I analysis further reveals distinct spatial regimes. Wuhan and its seven surrounding prefecture-level cities constitute a typical “high–high (H-H)” agglomeration zone—a “hot spot” where high-performing areas are mutually reinforced by similarly high-performing neighbors, reflecting coordinated development driven by spatial spillover effects. Adjacent to Wuhan, Xiaogan presents an “low–high (L-H)” spatial mismatch—a pattern identifying a spatial outlier where a low-performing unit is embedded within a high-value neighborhood. This suggests that localized barriers—whether institutional or infrastructural—prevent Xiaogan from fully benefiting from the surrounding development momentum. Meanwhile, Xiantao, Qianjiang, and Tianmen in the Jianghan Plain hinterland form an “low–low (L-L)” agglomeration zone—a “cold spot” where persistently low values cluster together, indicating entrenched unfavorable conditions that constrain development across these contiguous regions.

The coexistence of “H-H” hot spots around the Wuhan metropolitan core, L-H outliers in transitional zones such as Xiaogan, and L-L cold spots in the Jianghan Plain provides a nuanced portrait of the spatial heterogeneity across Hubei Province. The “H-H” and “L-L” patterns reveal mutually reinforcing spatial processes, while the “L-H” pattern highlights spatial discontinuities where targeted policy interventions may yield the greatest marginal returns.

4.3. Spatial–Temporal Variations in CCD

The analysis of the overall coupling coordination degree showed that Hubei’s CCD level situated it in the primary coordinated stage throughout the whole period. Although the CCD presented a slight decreasing tendency in 2022, from 2020 to 2024, the overall CCD increased from 0.626 to 0.661, and it is constantly approaching the initially coordinated stage. This indicates that Hubei Province is constantly moving in the direction of an increasing CCD between UQ and WECC.

In terms of the characteristics of temporal evolution, from 2020 to 2024, the CCDs of different cities displayed different development trends (Figure 7). The CCD of Xiangyang increased the most remarkably, rising from 0.624 to 0.681. Xiantao rose from the barely coordinated to the initially coordinated stage. Wuhan, Yichang, Shiyan, and Suizhou had a steady upward trend. Meanwhile, Jingzhou’s position decreased slightly. Compared with other cities, the CCD of Tianmen fluctuated significantly. In 2023, that of Tianmen reached 0.633, while it decreased to 0.583 in 2024. Huangshi, Ezhou, and Xianning had a slight improvement in fluctuation.

From the perspective of spatial differentiation patterns, the CCDs across cities in 2024 ranged between 0.583 and 0.712, revealing significant spatial disparities. Wuhan achieved the highest CCD score (0.712), entering the moderately coordinated coordination stage (0.7 ≤ D < 0.8), fully demonstrating its dominant position in terms of resource allocation and social development. Meanwhile, Tianmen recorded the lowest (0.583), remaining in the barely coordinated stage. Regarding the coordination level distribution, 13 cities (81.25%) were in the initially coordinated stage (0.6 ≤ D < 0.7). Xiangyang serves as an illustrative example: its CCD value stood at 0.624 in 2020, situating the city within the initially coordinated stage. This metric exhibited a marginal increase to 0.629 in 2021, yet experienced a minor retreat to its 2020 baseline in 2022. A gradual recovery ensued, with the index returning to 0.629 in 2023 and advancing further to 0.681 in 2024. Despite intermittent fluctuations in certain years, the overarching momentum reflects a pronounced ascending tendency, with the region sustaining its position in the initially coordinated stage throughout the study period. While Tianmen and Ezhou remained in the barely coordinated stage, all cities had a considerable gap compared with Wuhan in terms of the CCD. This spatial pattern, characterized by “one core leading, multiple points supporting, and localized lagging”, was particularly pronounced.

4.4. Identification of Key Factors for CCD

4.4.1. Grey Relational Analysis of Key Factors Influencing CCD

Grey relational analysis is utilized in this study to quantitatively examine the influences exerted by the UQ and WECC system indicators on the CCD (

Table 5. Key factors of CCD.

No.	Indicator	Grey Relational Grade
1	Proportion of fish-eating fish (I2)	0.988 **
2	Water consumption per industrial value added (R1)	0.783 **
3	Per capita water resources (S1)	0.767 *
4	Proportion of urban residents in elderly care insurance (SO2)	0.752 ***
5	Total nitrogen (mg/L) (S5)	0.730 **
6	Water yield modulus (S2)	0.715 *
7	Pollution monitoring in exceedance rate (P4)	0.699 ***
8	Total phosphorus (mg/L) (S6)	0.785

Note: * p < 0.10, ** p < 0.05, *** p < 0.01.

). The highest relational indicator is the proportion of fish-eating fish (I4), whose relational degree is 0.988. The proportion of fish-eating fish is an overall indicator that reflects the health of water bodies. The appearance and quantity of fish not only depend on sufficient water resources and good water connectivity but also require the support of the whole food chain and plenty of prey fish resources. At the same time, this indicator is very sensitive to the water quality and habitat environmental quality. It is an important biological indicator to evaluate the health and sustainability of aquatic ecosystems. The rationality of its proportion reflects the balanced state and ecological integrity of the whole aquatic system.

The other primary influencing factors exhibit the following order from large to small: industrial value-added water consumption (R1), per capita water resources (S1), proportion of urban residents participating in elderly care insurance (SO2), total nitrogen content (S6). Among them, per capita water resources (S1) and the water production modulus (S2) also appear in previous research, which proves that the basic indicators of water resources are still the main limiting factors for coupling coordination. The strong correlation among the water resource quality indicators of total nitrogen (S5) and total phosphorus (S6) further proves that water environmental quality is a basic guarantee for coordinated ecological and social development. WECC indicators occupy five of the top six places in the significance test, as well as being the top three, proving that WECC is the main limiting factor currently restricting the CCD.

4.4.2. Mechanism Analysis of Key Factors Influencing CCD

To further investigate the mechanisms and spatial effects of various factors, this study employs the SDM for estimation (

Table 6. Panel estimation of Spatial Durbin Model.

Variable	Fixed Effects (FE)		Random Effects (RE)
	Coe	t	Coe	z
I2	−0.007 **	(−2.36)	−0.005 *	(−1.74)
R1	0.000	(0.03)	−0.009	(−1.42)
S1	0.072 ***	(3.10)	0.080 ***	(2.91)
SO2	0.066 **	(2.49)	0.071 ***	(2.79)
W × I2	−0.007	(−1.13)	0.005	(0.88)
W × R1	0.001	(0.06)	−0.007	(−0.69)
W × S1	−0.153 **	(−2.56)	−0.088 **	(−1.41)
W × SO2	−0.021	(−0.35)	−0.0478	(−0.88)
ρ (W × Y)	0.139	(1.01)	0.474
Log-likelihood	274.77		228.28
AIC	−489.54		−436.56
BIC	−418.08		−412.74
Hausman Test	χ2 = 2.26		p = 0.972

Note: * p < 0.10, ** p < 0.05, *** p < 0.01.

). The Hausman test statistic is 2.26, with a p-value of 0.972, indicating the acceptance of the null hypothesis and suggesting that the random effects model is more appropriate. The spatial autoregressive coefficient ρ of the random effects SDM model is 0.4743, indicating that a 1% increase in the CCD in a given region promotes a 0.47% increase in neighboring regions, demonstrating a significant positive spatial spillover effect. The spatial lag term W × S1 is significantly negative (−0.1527) in the fixed effects model, suggesting the potential existence of regional competition effects.

Further effect decomposition (

Table 7. SDM effect decomposition (RE).

Variable	Direct Effect	Indirect Effect	Total Effect	Indirect Effect (%)
I2	−0.005	0.004	−0.001	45.82
R1	−0.011	−0.020	−0.031	64.09
S1	0.072	−0.087	−0.015	54.79
SO2	0.069	−0.025	0.044	26.39

) reveals the dual role of per capita water resources (S1). Its direct effect (0.0721) significantly promotes the local CCD; however, its indirect effect (−0.0874) demonstrates a negative spatial spillover, indicative of inter-regional competition. The direct effect of the proportion of urban residents participating in elderly care insurance (SO2) is 0.0685, with an indirect effect of −0.0245 and a positive total effect (0.0439), making it a key factor in enhancing coordinated development. The indirect effect of water consumption per industrial value added (R1) accounts for 64.09% of its total effect, indicating that its influence is primarily transmitted through spatial spillover channels. The proportion of fish-eating fish (I2) exhibits a slight inhibitory effect on local regions but demonstrates a positive spillover effect on neighboring regions.

The effect decomposition reveals differences in the mechanisms of action for each factor. Although per capita water resources (S1) and the proportion of urban residents participating in elderly care insurance (SO2) have significant positive effects on local regions, both generate negative spatial spillovers to neighboring regions, suggesting potential regional competition. The influence of industrial water use efficiency (R1) is primarily transmitted through spatial channels.

Based on the estimation results of the Spatial Durbin Model (SDM), the proportion of urban residents participating in elderly care insurance (SO2) exhibits a significant positive local effect on the CCD, with a coefficient of 0.071 under the random effects specification. However, its spatial lag term (W × SO2) is statistically insignificant, indicating that its cross-regional influence is not straightforwardly captured by the model’s immediate spatial parameters. The effect decomposition further reveals a nuanced mechanism: while SO2 exerts a positive direct effect (0.069) on the local region, it generates a negative spatial spillover (−0.025) to neighboring areas, resulting in a net positive total effect (0.044). This suggests that improving elderly care insurance coverage enhances the local CCD by stabilizing residents’ consumption expectations and boosting the domestic demand, yet it may also trigger regional competition for resources, mildly dampening the CCD in surrounding regions.

This finding is reinforced by the grey correlation analysis, which identifies SO2 as a key institutional factor with a notably high correlation coefficient (0.752) within the UQ system. The integration of urban and rural pension insurance not only optimizes labor allocation and resource distribution but also establishes a stable social foundation for ecological conservation. Ultimately, by increasing governmental and societal investment in environmental protection, SO2 contributes to enhancing the CCD between UQ and WECC. Therefore, policy efforts should focus on expanding insurance coverage and promoting regional coordination to mitigate negative spillovers and harness the full potential of social security systems in fostering socioecological synergy.

5. Conclusions

(1)

Hubei Province exhibits a dual-track, divergent evolution in both UQ and WECC over time. UQ steadily increased from 0.369 in 2020 to 0.409 in 2024. Wuhan remained the absolute leader and further widened its gap with other cities, showing spatial disparities that first narrowed and then widened. WECC showed a fluctuating trend consisting of “first rising, then declining, and then rising again”. Other cities showed insufficient coordination between two systems. In other words, although the future development momentum will continue, the regional divergence trend will be enhanced.

(2)

The spatial distribution of UQ and WECC in Hubei Province exhibits a pronounced “core–periphery” clustering effect and a gradient differentiation pattern. The Global Moran’s I analysis shows that there is a significant positive spatial correlation between the two systems. The Local Moran’s I analysis results show that Wuhan has formed a network connecting surrounding prefecture-level cities, Xiangyang–Yichang forms the second growth pole, and the cities on Jianghan Plain form an “L-L” aggregation depression.

(3)

The CCD between UQ and WECC in Hubei Province remains at the initially coordinated stage, while steadily advancing toward the moderately coordinated stage. Hubei’s CCD rose from 0.626 in 2020 to 0.661 in 2024. Wuhan was the first city to enter the moderately coordinated stage, and the CCD for 81.25% of the cities indicate that they are in the initially coordinated stage. While the remain cities remain at barely coordinated stage.

(4)

The WECC system represents the primary constraint on CCD development in Hubei Province, with the piscivore biomass ratio (proportion of fish-eating fish) serving as a critical bioindicator of ecosystem health (p < 0.05). The proportion of urban residents participating in elderly care insurance (SO2) (γ = 0.752) is the dominant socioeconomic factor. The SDM analysis demonstrates significant positive spillover effects (ρ = 0.4743), while per capita water resources exhibits competitive dynamics (W × S1 = −0.1527). Achieving sustainable development requires integrated approaches addressing both aquatic ecosystem restoration and comprehensive social security system development.

6. Study Limitations and Future Prospects

This study covers 2020–2024, offering a focused perspective on recent developments while acknowledging that the findings may reflect period-specific conditions. Although the analysis incorporates major disruptions, such as public health emergencies and extreme climate events, their long-term consequences across different time cycles remain important avenues for future investigation. Methodologically, the existing models could be further enhanced in capturing complex interdependencies among indicators. As this study demonstrates, weight assignment and grey relational grades measure fundamentally different dimensions of “importance”—static structural relevance versus dynamic temporal similarity—suggesting that no single method alone can comprehensively characterize indicator contributions. Additionally, indicator selection inevitably involves a degree of subjectivity, and the conclusions drawn from individual indicators should be interpreted cautiously. Future research should employ sensitivity analyses and clarify intrinsic connections among the indicators to strengthen the analytical robustness.