Analyzing Coupling Coordination and Driving Factors of Social–Ecological Resilience: A Case Study of the Lower Yellow River

Abstract

Flat terrain and economically prosperous downstream regions face significant challenges in achieving a balance between socio-economic development and ecological preservation. The Social–Ecological System (S-ES) serves as a vital framework for quantifying the interactions between human activities and the natural environment, providing insights into the development status of regional social and ecological systems. This study utilizes the Coupling Coordination Degree (CCD) method to construct an S-ES coupling model, integrating correlation analysis, geographic detectors, and grey relational analysis to explore the driving factors influencing Social–Ecological Resilience (S-ER) coordination. The research focuses on 25 counties in Shandong Province, situated in the lower Yellow River Basin, over the period from 2000 to 2022. Key findings include (1) significant spatial clustering, with identifiable hotspots and cold spots in S-ER distribution; (2) substantial changes in S-ER CCD around 2010 and 2020; and (3) persistent mismatches between socio-economic development and ecological improvement, presenting a major challenge for enhancing coordination. These insights provide valuable guidance for sustainable development strategies in the lower Yellow River Basin.

1. Introduction

The lower reaches of river basins, with their flat terrain and fertile soil, have historically been areas of economic prosperity, population concentration, and agricultural development [1,2]. However, behind these natural advantages lie significant inconsistencies and challenges. While the flat terrain is conducive to agricultural irrigation, it also brings risks such as soil erosion and flood disasters; the overdevelopment and singular use of fertile soil have led to ecological degradation, soil depletion, and loss of biodiversity [3,4]. The concentration of population further exacerbates pressures on water, land, and energy resources, with environmental pollution and overconsumption of ecological resources becoming increasingly prominent [5]. The imbalance in society and the contradiction between economic development and environmental protection have led these areas to face unsustainable growth patterns, with declining ecological carrying capacity being a key issue limiting long-term stable development [6]. In the face of these challenges, the UN Sustainable Development Goals (SDGs) provide a strategic guiding framework for regions worldwide, particularly those such as lower river basins, which face challenges of imbalanced development and sustainability [7]. The SDGs, which include goals like ‘building inclusive, safe, resilient, and sustainable cities and communities’, guide efforts to address increasingly severe socio-economic imbalances and ecological degradation by enhancing regional social and ecological resilience. The lower river basin areas, which play a crucial role in food production, are confronted with challenges of resource-carrying capacity limits. SDG 2 (‘Zero Hunger’), SDG 6 (‘Clean Water and Sanitation’), and SDG 11 (‘Sustainable Cities and Communities’) highlight the importance of food security and water resource management, which are key for sustainable urban development. At the same time, against the backdrop of climate change threatening the ecological environment and economic development, SDG 13 (‘Climate Action’) has become increasingly important in the lower river basin.

Since Holling introduced the concept of resilience into ecology in 1973 [8], the idea of Social–Ecological Resilience (S-ER) has become crucial for assessing the diversity and adaptability of Social–Ecological Systems (S-ESs), integrating both ecological and social science perspectives. S-ER is essential for addressing challenges and promoting sustainable futures, focusing on the capacity of the entire system to adapt to changes and disturbances through the interaction of social and ecological components.

Social Resilience (SR) refers to the capacity of communities to withstand risks, maintain economic stability during shocks, and adapt and recover post-disturbance. SR encompasses economic, social, and environmental dimensions, aiming to benefit all social groups inclusively. Current research primarily measures socioeconomic resilience from economic and social perspectives [9,10]. Economically, improvements in indicators like GDP growth and increased investment are crucial for achieving SDGs related to economic growth and employment (SDGs 8 and 9) [11,12]. Socially, enhancing welfare and strengthening social stability and governance are vital for realizing SDGs related to social inclusion, equality, and justice (SDGs 10 and 11) [13,14].

Ecological Resilience (ER) focuses on the ability of natural ecosystems to recover and maintain stability after disturbances, emphasizing inherent stability and biodiversity [15,16]. ER is also shaped by policy and economic development levels. Compared to social systems [17,18], the consequences of ecological disturbances are more direct and geographically diverse, necessitating varied quantitative indicators and assessment methods without a unified standard [19].

The CCD model emphasizes the importance of dynamic interactions between systems [20,21], providing a scientific foundation for formulating targeted management strategies. Current research on S-ER CCD in river basins primarily focuses on the interactions between key resources within S-ES and other indicators, such as socio-economic factors and water quality [22], water environmental protection [23], carbon emission reduction [24], habitat quality, and resource and environmental carrying capacity [25]. These studies have deepened the understanding of the complexity of S-ER in river basins. However, water resources, a key indicator in these systems, are often treated in isolation or simplistically categorized, overlooking their complex interactions with other system components.

Addressing existing research gaps, this study investigates the fundamental principles of the S-ES by explicitly dividing it into social and ecological components. The coupling mechanisms of top-level indicators in downstream basins were reconstructed, and key indicators reflecting the interactive dynamics of these systems were identified and validated.

The CCD model, aligned with SDG principles, is applied in this study to analyze the CCD and dynamic evolution of S-ER in 25 counties of Shandong Province in the lower reaches of the Yellow River from 2000 to 2022. Correlation analysis (CA), geographical detector (GD), and grey relational analysis (GRA) are utilized to investigate the driving factors influencing the evolution of S-ER CCD in these regions. This comprehensive approach establishes a scientific basis for sustainable development in downstream basins, deepens understanding of the interrelationship between social behaviors and ecological changes, and supports the formulation of sustainable strategies for the future.

2. Research Design and Study Area

2.1. Research Design

This study aims to investigate the coupling and coordination relationship between society and the ecosystem in economically dense downstream regions of rivers, specifically focusing on the lower reaches of the Yellow River from 2000 to 2022. It consists of two main components: the CCD assessment and the analysis of driving factors.

This study aims to explore the coupling and coordination relationship between society and the ecosystem in the downstream regions of economically intensive rivers, with a focus on the lower Yellow River Basin from 2000 to 2022. The research design, as shown in Figure 1, consists of two main components: the assessment of CCD and the analysis of driving factors. The CCD assessment focuses on six key time points: 2000, 2005, 2010, 2015, 2020, and 2022. The year 2000 was selected as the starting point because it marked the year following the official impoundment of the Xiaolangdi Reservoir, which significantly reduced the threat of flooding and marked a new phase in the water–sediment relationship of the Yellow River. Subsequent time points were chosen at five-year intervals to maintain temporal consistency and avoid redundant analysis. The final time point, 2022, was selected based on the completeness of available data post-2020.

Building upon this, a S-ER CCD evaluation system for the lower Yellow River Basin was established to quantify the spatial and temporal distribution of S-ER. This system accurately measures the coupling coordination between the socio-economic and ecological systems during specific periods. The evaluation framework comprehensively covers both the social and ecological dimensions and uses a combination of the entropy method and AHP to determine the weight of evaluation indicators. The second part of the study focuses on identifying the driving factors influencing system coordination. Based on existing research, the study employs a combination of CA, GD, and GRA to explore the constraints and promoting factors at both the basin and county levels. The findings will provide scientific support for policy development aimed at promoting coordinated development between socio-economic and ecological systems in economically intensive downstream regions.

This comprehensive research approach aims to reveal the complex interactions between socio-ecological dynamics in the lower river basin and offers valuable scientific insights and practical guidance for sustainable management and policymaking in similar regions.

First, an evaluation system for the CCD of the S-ER in the lower Yellow River is established to quantify the spatiotemporal distribution characteristics of S-ER. This system accurately measures the degree of coupling coordination between the social and ecological systems within the study area over specific time periods. The constructed evaluation framework comprehensively encompasses both social and ecological dimensions. In determining the weights of the evaluation criteria, a combination of the entropy method and the Analytic Hierarchy Process (AHP) is employed. The second part of the study focuses on identifying the driving factors that influence the coordinated development of these systems. Building on existing research, three methods—CA, GD, and GRA—are utilized to explore both constraining and facilitating factors at the basin and county levels. The results of this analysis will provide a basis for policy recommendations aimed at promoting the coordinated development of social and ecological systems in these economically intensive downstream regions.

This comprehensive approach seeks to provide a nuanced understanding of the complex interactions between social and ecological dynamics in river downstream areas, contributing valuable insights for sustainable management and policy formulation.

2.2. Study Area

The study focuses on the Shandong segment of the lower Yellow River, spanning 628 km and covering 18,300 square kilometers. This region exhibits a diverse landscape, transitioning from wide to narrow, and steep to gentle terrains [26]. It includes 9 municipalities and 25 counties, with emphasis on those bordering the river and their floodplain areas (Figure 2).

Balancing economic growth and ecological sustainability in this area is challenging due to its complex terrain and historically unstable river course, resulting in unique supra-river features and shifting deltas. The coastline, shaped by sediment deposition, poses constraints on port development and water resource management.

Despite its economic and agricultural productivity, the region faces persistent issues such as water scarcity, ecological degradation, and extensive flood-prone areas caused by the “hanging river” phenomenon. These natural constraints distinguish the lower Yellow River Basin from other basins [27].

This study aims to analyze the coupling and coordination of social and ecological resilience across 25 counties, providing insights to harmonize ecological protection and economic development in the Shandong segment of the lower Yellow River.

2.3. Indicator System and Data Sources

The lower Yellow River Basin, characterized by concentrated agriculture, dense population, and limited water resources, faces significant imbalances between social and ecological development. To evaluate the comprehensive resilience of this region, S-ER, an evaluation system comprising 18 indicators, was constructed. Six indicators related to SR encompass social and economic aspects, reflecting development conditions, population quality, and living standards (

Table 1. Evaluation Indicators for Social Resilience.

No	Subsystem	Criterion Layer	Unit	Calculation Method/Data Source	Attribute	Weight	SDG
S1	Social	Urbanization Rate [28,29]	%	Obtained from statistical yearbook	+	22.29	SDG 11: Sustainable Cities and Communities
S2		Road Network Density [30]	km/km2	Ratio of total road length to regional area	+	14.47
S3		Per Capita Cultivated Land Area [31]	hectares/10,000 people	Cultivated land area/year-end permanent population	+	9.23	SDG 2: Zero Hunger
S4	Economic	Per Capita GDP [32,33]	yuan/person	GDP/year-end permanent population	+	16.35	SDG 10: Reduced Inequalities
S5		Fiscal Revenue as a Proportion of GDP [34]	%	Fiscal revenue/GDP	+	13.03	SDG 16: Peace, Justice, and Strong Institutions
S6		Per capita total industrial output [35]	yuan/person	Obtained from statistical yearbook	+	24.63	SDG 9: Industry, Innovation, and Infrastructure

). Twelve indicators related to ER are divided into four dimensions: resources, S-ER factors, ecosystem service provision, and ecological integrity. Particular attention was given to water scarcity issues, with key indicators such as the Proportion of Water Supply from the Yellow River (E4), Agricultural Water Use Rate (E6), and Water Consumption Rate (E7) included (

Table 2. Evaluation Indicators for Ecological Resilience.

No	Subsystem	Criterion Layer	Unit	Calculation Method/Data Source	Attribute	Weight	SDG Goal
E1	Resources	Surface Water Resources [36]	10,000 m3	Obtained from water resources bulletin	+	7.34	SDG 6: Clean Water and Sanitation
E2		Annual Average Precipitation [36,37]	mm	Obtained from water resources bulletin	+	5.71
E3	Socio-economic factors	Per Capita Water Resources	m3/10,000 people	Total water resources/permanent population	+	6.45
E4		Proportion of Water Supply from the Yellow River	%	Proportion of water diverted from the Yellow River in regional water supply	+	5.41	SDG 7: Affordable and Clean Energy
E5		Area Proportion of the Beach Area within the Yellow River Embankment	km2	Beach area within the county/corresponding county area	+	14.7	SDG 13: Climate Action
E6		Agricultural Water Use Rate	%	Agricultural water use/total water use	+	4.12	SDG 2: Zero Hunger
E7		Water Consumption Rate [38]	%	Obtained from water resources bulletin	-	5.00	SDG 12: Responsible Consumption and Production
E8		Industrial Wastewater Discharge per Unit of GDP [39,40]	10,000 tons/10,000 yuan	Industrial wastewater discharge/GDP	-	2.35	SDG 6: Clean Water and Sanitation
E9	Ecological Integrity	Interspersion and Juxtaposition Index (IJI)	%	Fragstats	+	9.51	SDG 15: Life on Land
E10	Ecological Service Supply	Habitat Quality	-	InVEST (Appendix B)	+	5.91
E11		Soil Conservation	tons/hectare	InVEST (Appendix C)	+	26.09	SDG 14 and 15: Life Below Water and Life on Land
E12		Water Production	mm	InVEST (Appendix D)	+	7.41

). The indicator selection process was guided by regional literature and practical conditions, considering human–environment relationships, food systems, and water resource constraints, to accurately capture the region’s social system dynamics and the level of economic and social development.

The data used in this study span the period from 2000 to 2022, providing comprehensive temporal support for the analysis. For missing regional rainfall data, linear interpolation was applied to estimate the gaps, ensuring the completeness of the dataset (

Table 3. Data Sources and Pre-processing ways.

Data Name	Data Source	Data Parameters	Data Preprocessing
Population and Economic Data	China City Statistical Yearbook, Shandong Provincial and Local Statistical Yearbooks, Annual Bulletins, Official Policy Documents	Data for 25 counties	Data initialization, rasterization
Pollutant Emissions
Water Consumption Rate, Agricultural Water Use, Total Water Resources, Surface Water Resources	Shandong Provincial, City, and County Water Resources Bulletins
Yellow River Water Allocation	Yellow River Water Resources Bulletin, Shandong Provincial and Local Statistical Yearbooks
Open Map Data	Open Street Map website	Vector data	Supplement missing road data using ArcGIS 10.4 software to obtain street density
Land-Use Data	China Land-Cover Dataset (CLCD, https://zenodo.org/records/8176941) accessed on 10 September 2023	30 m	Analyze land use properties using ArcGIS 10.4 software
Digital Elevation Model (DEM) Data	Digital Elevation Model (DEM) Data	30 m	Projected onto the WGS_1984_Albers projection coordinate system, missing data supplemented through linear interpolation, clipped using the administrative boundaries of the 25 counties along the Yellow River in Shandong Province
Normalized Difference Vegetation Index (NDVI)	Resource and Environmental Science Data Center, Chinese Academy of Sciences (https://www.resdc.cn)	250 m
Annual Precipitation		1 km

). All data were standardized to the WGS-1984 geographic coordinate system and projected to the WGS_1984_Albers coordinate system to ensure spatial consistency. Furthermore, the data were clipped to align with the administrative boundaries of the 25 counties in Shandong Province along the Yellow River, enhancing the specificity and accuracy of the dataset in relation to the study area.

3. Methodology

3.1. Data Pre-Processing

The selected indicators exhibit differences in dimensions, scale, and qualitative attributes, necessitating standardization to harmonize these variations. To address this, the Min–Max normalization method was applied to both positive and negative indicators [41].

For positive indicators:

$$x_{ijk\prime }=(x_{ijk}-{\lambda jk}_{min})/({\lambda jk}_{max}-{\lambda jk}_{min}) (i=1, 2, \dots , m)$$

(1)

For negative indicators:

$$x_{ijk\prime }=({\mathsf{\lambda }}_{jkmin}-x_{ijk})/({\lambda jk}_{max}-{\lambda jk}_{min}) (i=1, 2, \dots , m)$$

(2)

Here, x_ijk′ represents the original data value, λ_jkmin is the minimum value of the original data for the indicator, λ_jkmax is the maximum value, and x_ijk′ is the standardized indicator value. This process eliminates dimensional inconsistencies and ensures comparability across all indicators.

3.2. Combined Weights

To enhance the reliability and accuracy of weight determination, this study employs a combined weighting approach that integrates the Analytic Hierarchy Process (AHP) for subjective weights and the entropy method for objective weights.

(1)

Subjective Weight Calculation

The AHP method establishes a hierarchical structure, constructs a judgment matrix, and performs consistency checks to ensure the rationality of the ranking results [42,43]. The consistency index (CI) and consistency ratio (CR) are calculated as follows:

$$CI={\displaystyle \frac{{\lambda }_x-n}{n-1}}$$

(3)

$$\mathrm{C}\mathrm{R}={\displaystyle \frac{CI}{RI}}$$

(4)

A CR value below 0.10 indicates acceptable consistency; otherwise, adjustments are required.

(2)

Objective Weight Calculation

The entropy method calculates weights based on the variability of indicator data. After standardizing the data, normalized indicator values X_ijk are obtained. The indicator weights are then computed using the entropy formula:

$$yj=x^{\prime }j/\sum _i\sum _{j}x\prime j$$

(5)

$$W_{jo}={\displaystyle \frac{1-[-\frac{1}{\theta }{\sum }_i{\sum }_j{y}_{ijk}\mathit{ln}(y_{ijk}\left)\right]}{{\sum }_k{g}_k}}$$

(6)

Here, y_ijk represents the proportion of each indicator, W_jo is the calculated weight, and k corresponds to the specific indicator [44].

(3)

Weight Combination

The combined weight is derived using the following formula:

$$wj={\displaystyle \frac{W_{js}\times W_{jo}}{\sum _{I=1}^p{W}_{js}\times W_{jo}}}$$

(7)

Here, W_js and W_jo represent the subjective and objective weights, respectively. This approach integrates expert knowledge with data-driven insights to achieve a balanced and rational weight allocation [45].

3.3. Coupling Coordination Degree

The CCD model is employed to analyze the interaction mechanisms and dynamic relationships among various components within a system by quantifying both the coupling degree and coordination level of each element [46]. This model helps in understanding the extent to which different subsystems interact and develop synergistically. The coupling degree (C) measures the level of interdependence and mutual restraint between or within system elements, ranging from 0 to 1, where higher values indicate stronger interactions between systems. The coordination degree (D), on the other hand, reflects the extent of positive interaction and synergistic development among the elements during the coupling process [20].

The CCD model is mathematically represented as:

$$C={\displaystyle \frac{2\sqrt{S_1\times E_1}}{S_1+E_1}};T=\alpha S_1+\beta J_2;D=\sqrt{\mathrm{C}\times \mathrm{T}},$$

(8)

where C is the coupling degree, with a value range of [0, 1], indicating the strength of interactions between systems. S1 and E1 represent the resilience of the social and ecological systems, including water-related subsystems. T is the comprehensive development index of the resilience of the social and ecological systems. D is the coupling coordination degree, which quantifies the level of coordination between the systems.

The coupling degree is then scored from 1 to 10, reflecting the degree of coupling and coordination among the elements, as shown in

Table 4. Classification Standards for Coupling Coordination Degrees.

Coordination Level	Degree of Coordination	Coordination Level	Degree of Coordination
1	Extreme Imbalance	6	Barely Coordinated
2	Severe Imbalance	7	Primary Coordination
3	Moderate Imbalance	8	Intermediate Coordination
4	Mild Imbalance	9	Good Coordination
5	Near Imbalance	10	Excellent Coordination

.

3.4. Exploratory Spatial Data Analysis

The Exploratory Spatial Data Analysis (ESDA) model is employed to investigate spatial patterns and relationships within the data [47,48], particularly focusing on the coupling between SR and ER. In this study, spatial autocorrelation analysis and hot/cold spot analysis were utilized to uncover the underlying spatial characteristics and patterns of coupling coordination. Both analyses were performed using ArcGIS 10.4.

3.4.1. Spatial Autocorrelation Analysis

Spatial autocorrelation analysis assesses the correlation and significance of spatial data across a region [30]. This method is essential for understanding the spatial distribution of variables and detecting patterns of spatial clustering. In this study, Global Moran’s I was used to evaluate the spatial characteristics and agglomeration of the coupling coordination degree between social and ecological systems across the entire study area. Global Moran’s I quantifies the overall spatial correlation by measuring whether the spatial distribution of the coupling coordination degree is clustered, dispersed, or random.

$$Global Mora{n}^{\prime }sI={\displaystyle \frac{{\sum }_{I=1}^n{\sum }_{j=1}^n{W}_{ij}(X_i-\overline{X})(X_j-\overline{X})}{s^2{\sum }_{I=1}^n{\sum }_{j=1}^n{W}_{ij}}}$$

(9)

where n is the number of spatial units indexed by i and j. x_i and x_j are the values of the variable of interest at locations i and j. w_ij is the spatial weight matrix.

3.4.2. Hot/Cold Spot Analysis

Hot/cold spot analysis is used to identify clusters of high (hot spots) and low (cold spots) values of the coupling degree between social and ecological systems [43]. This method goes beyond global spatial autocorrelation by pinpointing specific locations where values are statistically significantly high or low, thus highlighting areas of significant spatial clustering. The analysis tests the statistical significance of clustering using standardized z-values, which indicate the extent to which observed clustering deviates from a random distribution. Hot spots correspond to areas of high coupling degree, while cold spots represent areas of low coupling degree, indicating significant spatial dependence among these regions.

$$G{i}^{*}={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n{w}_{i,j}{\mathrm{x}}_j}-\overline{X}{\displaystyle {\sum }_{j=1}^n{w}_{i,j}}}{S\sqrt{\frac{n{\displaystyle {\sum }_{j=1}^n{w}_{i,j}-({\displaystyle {\sum }_{j=1}^n{w}_{i,j}{)}^2}}}{\mathrm{n}-1}}}};\overline{X}={\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n{X}_j}}{n}};S=\sqrt{{\displaystyle \frac{{\displaystyle {\sum }_{j=1}^n{X}_j^2}}{n}}-{(\overline{X})}^2},$$

(10)

where Gi^∗ is the Getis–Ord statistic. x_j is the attribute value for feature j. w_ij is the spatial weight between feature i and j. S is the standard deviation.

By applying ESDA, this study provides a comprehensive analysis of the spatial patterns of coupling coordination between social and ecological resilience, offering insights into regional disparities and potential areas for targeted intervention in the study area.

3.5. Analysis of Driving Factors

To thoroughly explore the driving factors influencing the CCD between SR and ER in the lower reaches of the Yellow River basin, this study integrates CA, GD, and GRA. These methods allow for a comprehensive evaluation from both quantitative and spatial perspectives.

Firstly, CA and the GD model were employed to investigate the overall driving factors affecting the lower reaches of the Yellow River. The CA provided a quantitative assessment [49], while the GD model addressed spatial variations in these factors [50]. Secondly, GRA was applied to delve deeper into the driving factors of typical counties, focusing on areas with varying coupling degrees [51].

3.5.1. Multicollinearity Test

To prevent spurious regression results between the S-ER CCD and its driving factors, it is essential to address multicollinearity in the regression model. Multicollinearity refers to a high degree of linear correlation between two or more independent variables in the model, which can distort the estimation of regression coefficients [50,52]. This issue was addressed by conducting a Variance Inflation Factor (VIF) test using STATA15.0 software on panel data.

VIF = 1/(1 − R²),

(11)

In the formula, R² represents the coefficient of determination for the auxiliary regression with multiple explanatory variables. The collinearity results are detected based on the VIF and tolerance (TOL) among the independent variables, and factors with significant VIF > 10 and TOL < 0.1 are eliminated.

3.5.2. Correlation Analysis

Correlation analysis was conducted to examine the relationships between variables, using the correlation coefficient r to measure the strength and direction of the linear relationship between two variables [53].

$$\mathrm{r}={\displaystyle \frac{{\displaystyle \sum (X_i-\overline{X})(Y_i-\overline{X})}}{\sqrt{{\displaystyle \sum {(X_i-\overline{X})}^2}(}\sqrt{{(Y_i-\overline{Y})}^2}}},$$

(12)

where r is the correlation coefficient, X_i and Y_i respectively represent the i th item of two different variable sequences, and $\overline{X}$ and $\overline{Y}$ respectively represent the average value of two different variable sequences. The coefficient is between −1 and 1. The closer the absolute correlation coefficient is to 1, the more closely the data points fall on a straight line. Among them, 0 < |r| ≤ 0.3 is weak relevant, 0.3 < |r| ≤ 0.5 is low-alcohol related, 0.5 < |r| ≤ 0.8 is significant correlation, 0.8 < |r| ≤ 1 is highly relevant.

3.5.3. Geographical Detector Model

The GD model is a statistical tool used to detect the driving factors of spatial variability and their impact on the dependent variable, highlighting spatial differentiation [54].

$$q=1-{\displaystyle \frac{1}{n{a}^2}}{\sum }_{i=1}^L{n}_i{{\sigma }_i}^2$$

(13)

where q represents the detection value of the influencing factor’s spatial heterogeneity, ranging from 0 to 1. A larger q value indicates stronger explanatory power of the factor for the spatial heterogeneity of the dependent variable. In this formula, i represents different strata of the influencing factor, ranging from 1 to L. n denotes the total sample size of the study area, while n_i represents the sample size in the ith stratum. σ² is the variance of the dependent variable across the entire study area, and σi² represents the variance of the dependent variable in the ith stratum. The GD model categorizes interactions between two factors as nonlinear weakening, single-factor nonlinear weakening, two-factor enhancement, independence, or nonlinear enhancement, allowing for a nuanced understanding of spatial influences [55,56,57].

3.5.4. Grey Relation Analysis

The GRA model examines the similarity or dissimilarity of development trends between variables using time series or data sequences [58]. This method evaluates the degree of correlation based on the geometric similarity of curves.

$$\gamma \left[\left[X_O\left(k\right),X_i\left(k\right)\right]\right]={\displaystyle \frac{m{in}_i{min}_k⌊X_O\left(k\right)-X_i\left(k\right)⌋+\in {max}_i{max}_k⌊X_O\left(k\right)-X_i\left(k\right)⌋}{⌊X_O\left(k\right)-X_i\left(k\right)⌋+\in {max}_i{max}_k⌊X_O\left(k\right)-X_i\left(k\right)⌋}}$$

(14)

$$\gamma \left(X_o,X_I\right)={\displaystyle \frac{1}{n}}\sum _{k=1}^n\gamma \left[\left[X_O\left(k\right),X_i\left(k\right)\right]\right]$$

(15)

where γ(X0,X_i) represents the grey relational degree between X₀ and X_i, where ε is the resolution coefficient used to enhance the significance of differences among grey relational coefficients, with ε∈(0,1), and the empirical value is typically set at 0.5.

4. Results

4.1. Spatio-Temporal Distribution Pattern of Social Resilience

From 2000 to 2022, the SR in the lower Yellow River Basin exhibited significant changes in both temporal and spatial dimensions. At the city level, Jinan and Dongying exhibited relatively high social resilience, while counties such as Pingyin, Licheng, and Hekou showed higher resilience, whereas counties like Mudan and Liangshan showed room for improvement. Overall, SR displayed an upward trend during this period, with the SR total value increasing by 259.43 points, a growth rate of 42% (Figure 3). However, this growth was not stable, with negative growth observed between 2005–2015 and 2020–2022. In terms of key indicators, social-related indicators such as urbanization rate, road network density, and per capita arable land showed an upward trend, with urbanization primarily occurring in the central and eastern regions, while road network density grew more in the western regions. Economic indicators, such as per capita GDP and industrial output in the eastern and central regions, displayed significant regional disparities (Appendix E, Figure A1). In general, the changing trends of social resilience in the lower Yellow River Basin reflect both the region’s economic development and infrastructure improvements, and also reveal its capacity to cope with natural disasters and socio-economic transformations.

4.2. Spatio-Temporal Distribution Pattern of Ecological Resilience

In terms of ER, the lower Yellow River Basin showed a relatively stable trend. At the city level, Jinan and Dongying were still more prominent, while counties such as Dongping, Licheng, and Hekou exhibited higher ecological resilience, while Mudan and Binzhou showed relatively lower resilience. Overall, ER also showed an upward trend (Figure 4). However, it is worth noting that the increase in ecological resilience was relatively slow, with a total increase of only 0.9%, and negative growth observed during the periods 2005–2010 and 2015–2020. Regarding natural resource-related indicators, surface water resources and precipitation fluctuated over time, but there was a stable upward trend in the eastern region. Water-related socio-economic indicators such as agricultural water usage rate and industrial wastewater discharge showed a general downward trend. Ecological service indicators, including habitat quality and soil protection services, fluctuated, but fishery output continued to rise (Appendix E, Figure A2). The trends in ecological resilience reflect the achievements in ecological protection and restoration in the region, but also highlight the resilience of the ecosystem under multiple pressures.

4.3. Coupling Coordination Degree and Correlation Analysis

4.3.1. Spatio-Temporal Distribution of CCD

The CCD showed fluctuating growth, with low points in 2020 and recovery by 2022. Spatially, higher CCD values were concentrated in the eastern and central regions, while the western regions remained vulnerable to disorder. CCD inflection points occurred in 2010 and 2020, marking significant shifts in regional coordination levels (Figure 5).

Based on the CCD relationships among the 25 counties in the Shandong section, four categories were identified: (1) CCD I: Counties with a CCD above the eighth level, indicating highly coordinated. (2) CCD II: Counties with a CCD between the seventh and eighth levels, meeting criteria for well coordinated. (3) CCD III: Counties with a CCD between the fourth and sixth levels, representing near-disorder to barely coordinated. (4) CCD IV: Counties with a CCD at the third level or lower, indicating a moderate-to-serious imbalance.

In 2010, the highest number of counties fell into Category III, with some counties exhibiting a severe imbalance with a CCD score of two for the first time. Conversely, 2020 marked the lowest number of counties in Category I and the highest number in Category IV, highlighting a critical period of imbalance. Post these inflection points, from 2010 to 2015 and 2020 to 2022, slight improvements were observed in coupling and coordination between the systems compared to earlier years. However, enhancements in CCD were predominantly concentrated in economically developed eastern coastal regions, while counties in the western regions continued to experience severe imbalances in CCD (Figure 6).

4.3.2. Spatial Autocorrelation and Hotspot Analysis

Significant spatial autocorrelation in CCD during the study period was revealed by Moran’s I index, with spatial aggregation effects fluctuating over time. The spatial autocorrelation status of annual S-ER CCD is recorded in

Table 5. Global Spatial Autocorrelation of S-ER CCD.

Year	Moran’s I	Z-Score	p-Value
2000	0.385161	2.519651	0.011747
2005	0.347995	2.317354	0.020484
2010	0.313162	2.123586	0.033705
2015	0.393359	2.594356	0.009477
2020	0.487724	3.149162	0.001637
2022	0.360683	2.427154	0.015218

, showing that CCD spatial distribution became increasingly clustered. Moran’s I increased from 0.385161 in 2000 to a peak of 0.487724 in 2020, before decreasing to 0.360683 in 2022. The lowest Moran’s I was observed in 2010, when CCD across counties was most balanced (CCD = 5), achieving a basic state of coordination. Despite the highest Moran’s I in 2020, regional differences were pronounced, with the lowest average CCD recorded that year. A significant number of counties fell into the IV category (CCD < 3), primarily in the western region, with the fewest counties in the I category. The high concentration of low-coordination regions contributed to a sharp increase in Moran’s I.

Geographic correlation features were further analyzed through cold and hotspot spatial clustering. Figure 7 illustrates the temporal and spatial distribution of cold and hotspot areas in the lower Yellow River Basin from 2000 to 2022. The analysis shows that CCD in the eastern and central regions remained stable or improved during this period, whereas the western region faced persistent challenges. From 2010 onward, cold spots were primarily concentrated in counties in the western region, including Liangshan, Mudan, Donge, and Juancheng.

4.4. Analysis of Driving Mechanisms of CCD

4.4.1. Validation of Factor Independence

The selected indicators for both SR and ER were tested for multicollinearity, and the results confirmed the absence of significant linear relationships between indicators. This ensures that the explanatory power of the model is not adversely affected by collinearity issues.

Table 6. Multicollinearity Diagnostic Results.

System	Indicator	TOL	VIF	Indicator	TOL	VIF
Social system	Urbanization Rate	0.787	1.271	Per Capita GDP	0.880	1.136
	Road Network Density	0.859	1.165	Fiscal Revenue as % of GDP	0.751	1.331
	Per Capita Cultivated Land Area	0.926	1.080	per capita total industrial output	0.851	1.175
Ecological system	Surface Water Resources	0.593	1.685	Area proportion of the beach area (within the Yellow River embankment)	0.831	1.204
	Average Annual Precipitation	0.218	4.582	Agricultural Water Use Rate	0.334	2.994
	Per Capita Total Water Resources	0.502	1.994	Interspersion and Juxtaposition Index (IJI)	0.311	3.217
	Industrial Wastewater Discharge per Unit of GDP	0.755	1.325	Habitat Quality	0.663	1.508
	Water Consumption Rate	0.620	1.612	Soil Conservation	0.263	3.806
	Proportion of Water Supply from the Yellow River	0.304	3.292	Water Yield	0.176	5.693

VIF value greater than 1, with values closer to 1 indicating weaker multicollinearity. TOL value ranges from 0 to 1, with smaller values indicating stronger collinearity between the independent variables.

summarizes the multicollinearity diagnostic results, showing acceptable tolerance (TOL) and variance inflation factor (VIF) values for all indicators across the social and ecological systems (Table 6).

4.4.2. Correlation Analysis of Driving Factors and CCD

CA revealed fluctuations in the relationships among driving factors over time, with notable turning points in 2010 and 2020. Before 2010, negative correlations between social and ecological system factors were predominant, but a gradual shift toward positive correlations was observed post-2010. By 2020, all high-value relationships had transitioned to positive correlations, with intensified linkages among factors. The weakest correlations were observed in 2000 and 2020, where many weak positive and negative correlations persisted, especially between social and ecological factors. The year 2022 exhibited the highest number of positive correlations within the ecological system, marking an evolution toward improved system integration (Figure 8).

4.4.3. Key Drivers and Interaction Effects

Although the CA approach quantified the relationships, it failed to comprehensively capture spatial interactions. The GD further identified the key driving factors and their spatial interactions within the SR, ER, and integrated S-ER (Figure 9).

The primary drivers of SR included road network density (S2), GDP per capita (S4), and per capita total industrial output (S6). Among these, S4 consistently held a significant proportion across six stages from 2000 to 2022. During the early phase of rapid urbanization (2000–2015), S6 showed relatively higher importance, while the role of S2 became increasingly prominent with the progression of urbanization, eventually emerging as the major driver. Compared to the consistent drivers of SR, ER exhibited more dynamic changes, with an increasing number of high-impact drivers. Factors such as surface water resources (E1), the proportion of water supplied by the Yellow River (E4), water utilization rate (E7), IJI (E9), and habitat quality (E10) showed variable rankings and influence across different years, acting as dominant forces driving regional changes. The top-ranking drivers of integrated S-ER included GDP per capita (S4), annual precipitation (E2), agricultural water utilization rate (E6), IJI (E9), and water output (E12). Notably, S4, as the core driver of SR, consistently played a crucial role in S-ER rankings. Meanwhile, ecological factors such as E2, E6, and E12 gained prominence with the implementation of agricultural and ecological protection policies.

The GD model further explored the explanatory power of individual and interactive drivers in the evolution of the CCD between social and ecosystem systems. The results revealed nonlinear and dual-factor enhancement effects among interactive drivers, with significant changes observed in 2010 and 2020. For SR, road network density (S2), per capita arable land area (S3), and the proportion of fiscal revenue to GDP (S5) exhibited strong interactive effects characterized by nonlinear enhancement (Figure 10). These factors demonstrated high interaction at different stages, with strong interactions involving S5 appearing in 2000 and 2010, S2 in 2005, and S3 in 2015 and 2022.

In ER, factors such as annual precipitation (E2), surface water resources (E3), agricultural water utilization rate (E6), industrial wastewater discharge per unit of GDP (E8), and water output (E12) showed strong interactions with other factors across different years (Figure 11). E2 demonstrated high interaction effects only in 2005, while E12 was prominent during the early stages of evolution (2000 and 2005). From 2005 onwards, E8 emerged as an important driver, exhibiting strong interactions with other factors. Furthermore, from 2015 onward, the interactive relationships involving E3 and E6 significantly increased (Figure 11).

Further interaction analysis revealed the relationships between S-ER drivers and CCD. Key factors, including S2, S3, S5, E2, E3, E8, E10, and E12, exhibited strong interactions with other drivers in various years. These interactions became particularly prominent in 2010 and 2020, driven by adjustments in economic structure and advancements in ecological protection policies. Notably, 2022 marked the year with the highest number of significant interactive relationships among drivers (Figure 12).

4.4.4. Verification in Typical Counties

To validate the findings, GRA was conducted on four typical counties representing different CCD categories. Results indicated that the most prevalent coupling level from 2000 to 2022 was CCD III, with scores ranging from 4 to 6, observed across western, central, and eastern counties. Consequently, three counties were selected to represent this dominant range, while one county was chosen to represent each of the other three CCD categories.

Further analysis revealed that CCD I and CCD II couplings were primarily influenced by ecological indicators, while CCD III exhibited a shift toward social system indicators. Conversely, CCD IV couplings were predominantly driven by social system factors, with E5 (area proportion of the beach area within the Yellow River embankment) being the only significant ecological factor identified.

Figure 13 shows the ranking of driving factors in these typical counties across different CCD levels, verifying the significant role of water-related ecosystem indicators in the coupling and coordination of S-ER in the downstream Shandong segment of the Yellow River basin.

5. Discussion

5.1. Temporal–Spatial Distribution Patterns of Social and Ecological Resilience

Between 2000 and 2022, SR and ER showed significant spatial and temporal variation in the lower Yellow River Basin. The eastern region displayed higher resilience than the western region, with cities such as Jinan and Dongying performing well. The CCD between social and ecological systems displayed marked regional differences, with coastal developed counties and provincial capital cities exhibiting higher coordination, while the western region showed more fluctuation and was mostly in an imbalanced state. The improvement in SR was mainly due to economic development, industrial restructuring, policy support, and enhanced disaster prevention capabilities. However, during 2005–2015 and 2020–2022, SR declined due to natural disasters, socio-economic transformation pressures, resource and environmental constraints, and policy uncertainty. ER fluctuations were mainly influenced by social development and natural disasters. During 2005–2010, natural disasters and urbanization/industrialization issues weakened ER, while during 2015–2020, although ecological protection policies were strengthened, the delayed implementation of policies, global economic fluctuations, high-pollution industries, and climate change still posed challenges to ER.

5.2. Water-Related Indicators as Key Determinants of S-ER CCD Dynamics

Human societal development and economic prosperity are closely linked, with changes in individual demands serving as a core driving force [59]. In examining the interaction between humans (SR) and the environment (ER), ensuring regional sustainable well-being is critical, particularly in downstream areas. In analyzing the changes in the ranking of driving factors for the CCD between the social and ecological systems in the Shandong section, several key aspects emerge: First, the coordinated advancement of socioeconomic development and ecological protection is a key factor. As societal attention to ecological environmental protection increases, the role of ecological factors in driving the system gradually strengthens, making them crucial elements influencing the CCD. Second, the sustainable use of water resources and the impacts of climate change cannot be ignored. The growing importance of water resource factors such as annual precipitation and agricultural water use efficiency reflects the critical role of water resource management in maintaining ecological balance and supporting socioeconomic development. Furthermore, the promotion of ecological policies is a key driver in enhancing ecosystem service capacity. Through the implementation of various ecological policies, the driving effects of factors such as the interspersion and juxtaposition index (IJI) and aquatic product yield (E12) have been significantly amplified.

Notably, all three patterns of CCD changes are closely linked to water-related indicators. Existing research suggests that water ecosystems are complex, multi-factor systems involving the coupling of natural and social water cycles. A healthy water ecosystem drives regional water system optimization and development by improving key indicators such as per capita water resources and wastewater recycling rates [60]. This conclusion aligns with the findings of this study, emphasizing the critical role of water systems in the overall water ecosystem. Properly coordinated water systems not only provide essential services to human society but also maintain ecological stability, alleviate water scarcity, and enhance resource utilization efficiency [61,62,63]. When misalignment occurs, ecological vulnerability increases, further restricting sustainable development in river basins [64].

Compared to other studies in the Yellow River Basin, strong connections were observed between water-related ecological indicators (e.g., freshwater resources, water-saving practices, and water production) and ecosystem services such as soil protection and carbon sequestration. These indicators and services jointly drive trade-offs and synergies within the region [65]. Similarly, large-scale studies have demonstrated a high degree of coupling between water resources, energy, and food systems, with water resources identified as the dominant factor influencing the other two [26]. Reduced performance in water-related systems was confirmed to weaken interactions between urbanization and ecosystems, limiting the scale of regional economic and social development. This outcome has also been observed in the Yangtze River Basin [66].

Previous studies on the CCD of S-ES often relied on single indicators or insufficient metrics, limiting their ability to capture the complexities of regional dynamics. To address these limitations, this study incorporated 18 indicators reflecting regional characteristics and applied methods such as CA, GD, and GRA to investigate the high driving influence of water-related indicators, including annual precipitation and water consumption rate, on the S-ER CCD. The results underscore the importance of integrating water-related indicators into CCD evaluations, which is equally crucial for other challenged river systems, such as the Colorado River and Niger River.

5.3. Policy Implications: Balancing Socioeconomic Growth and Ecological Sustainability

Existing policies exert a multidimensional impact on S-ER CCD, encompassing both positive effects, such as economic development, and negative effects, including fragmented county-level policies and a lack of specificity in basin-wide policies. An analysis of 55 policy documents from the lower Yellow River Basin, covering nine cities and 25 counties in Shandong Province, reveals that these policies are generally formulated based on regional characteristics and local conditions. The diversity and number of policies indicate key future directions for regional planning and development.

Economic development currently dominates policy goals in Shandong Province, especially given the province’s role as a pilot for new and old energy transitions. The long-term strategy of promoting a cleaner and more ecologically friendly production model remains a central focus. Adjustments in the economic structure have also brought significant changes to land use patterns, water consumption models, and regional economic frameworks, reflecting increasing attention to sustainable development and laying a foundation for future progress. However, balancing economic development and environmental protection remains a critical challenge for policymakers [67,68]. Fragmentation in county-level ecological governance policies is a significant concern. The lack of systemic coherence often undermines the effectiveness of governance efforts. In areas with higher CCD (CCD I), ecological policies prioritize biodiversity restoration and ecological protection, while in areas with lower CCD (CCD IV), policies emphasize infrastructure construction to stimulate economic growth. Figure 13 and Figure 14 illustrate how the fragmented policy environment weakens governance implementation and exacerbates the misalignment between ecosystems and socio-economic systems. Greater systematization and coherence in county-level policies are essential, alongside the development of more actionable ecological governance measures to improve S-ER CCD. Additionally, the vagueness and lack of specificity in national and basin-level policies often hinder their effectiveness at the county level. Basin-wide policies, while strategic, frequently lack the precision needed to address local issues effectively. Improved coordination and alignment between basin-level and county-level policies are critical to tailoring policies to local needs and enhancing their enforceability.

This study highlights that even minor ecosystem changes can significantly affect S-ER CCD, underscoring the strong coupling between socio-economic and ecological systems. Policy formulation must fully account for these mutual influences to ensure that economic development is promoted while effectively protecting the ecological environment. In conclusion, existing policies impact S-ER CCD both positively and negatively. A holistic approach to policymaking is necessary to develop feasible, specific policies that foster effective governance and promote coordinated development of ecological and socio-economic systems.

5.4. Adaptive Strategies for Enhancing Coupled Coordination

High interaction in coupled coordination arises not from a simple combination of social and ecological systems but from the complex mutual influences among internal driving factors. The lower Yellow River Basin retains significant potential for improving its coupling status and advancing sustainable development. Existing research has examined various aspects of the basin, such as land use and temperature correlations [69] and the interplay between flow dynamics and pollution [70]. Although studies on social and ecological system coupling in the lower basin remain limited, the literature indicates relatively strong mutual coordination in the region [60]. Consistent findings highlight better coupling in economically developed coastal and capital city areas compared to less developed regions, mirroring the spatial patterns observed in this study.

Analysis of high CCD in typical counties suggests that social system indicators provide a foundation for achieving a basic CCD score. Advancing to higher CCD levels increasingly relies on ecological system indicators. As CCD improves, water-related ecological indicators, such as water consumption rates and pollution control, become dominant drivers. In economically developed regions, further enhancement of S-ER CCD requires leveraging ecological systems by minimizing water consumption and controlling pollution.

Based on these findings, comprehensive adaptive strategies are proposed to improve CCD in the lower Yellow River Basin. At the basin scale, priority ecological zones should be designated for protection, restoring natural connections between mountains and rivers. Strengthening green space networks, optimizing natural water replenishment, and improving rainwater and sewage collection systems are critical actions. In western and central regions with low coupling, efforts should prioritize economic transformation and reductions in agricultural water consumption. Meanwhile, provincial capitals and coastal regions should intensify ecological protection to safeguard biodiversity.

Similar strategies can be applied to other downstream river basins, such as the Colorado and Niger Rivers, where water scarcity and high population density present shared challenges. Variations in CCD conditions and administrative structures necessitate the identification of water-related driving factors with direct impacts on CCD. Targeted governance and improvement measures tailored to each region’s specific conditions are essential to promote sustainable and balanced development.

6. Conclusions

In the face of growing ecological pressures, such as climate change and water scarcity, the S-ER CCD in the lower Yellow River Basin is marked by long-term, complex systemic characteristics. To better understand the trends and driving forces of S-ER CCD in this region, an S-ER CCD assessment framework was applied, utilizing panel data from 25 counties in the lower reaches of the Yellow River. The findings offer valuable insights into the intricate interactions between regional economic development and ecological protection in these downstream areas.

The analysis reveals that, between 2000 and 2022, several counties in the eastern and central parts of the lower Yellow River Basin exhibited high CCD. However, a comprehensive, region-wide coordinated development pattern has yet to be established. By identifying key driving factors, the study confirms that strengthening regional cooperation and cross-boundary governance—particularly through coordinated management of water-related indicators—is essential for improving the coordination of regions with low coupling. Regional cooperative governance can more effectively optimize water resource allocation, mitigate water pollution risks, and enhance agricultural water-use efficiency, providing a replicable governance model for other regions facing similar socio-ecological challenges.

Promoting the harmonious coexistence of socio-economic and ecological systems remains a shared challenge for water-scarce regions globally, including the lower Yellow River Basin. Establishing a water-related S-ER coupling evaluation system tailored to regional characteristics, identifying misalignments, and proposing targeted governance strategies are crucial for achieving sustainable development. The CCD verification system developed in this study offers new insights and practical examples for addressing ecological challenges in other downstream river basins. Furthermore, this research provides a solid scientific foundation and theoretical support for the development of sustainable policies and the enhancement of regional resilience, offering valuable guidance for sustainable development practices in similar regions around the world.