Analysis of the Coupling Coordination and Spatial Difference Between Economic and Ecological Environment: A Case Study of China

Abstract

This study adopts a sustainable development perspective to examine the economic and ecological coordinated progression and spatial disparities across 30 regions in China from 2011 to 2022. Firstly, detailed analysis of CCD reveals that coordination between the ES (economic subsystem) and the EES (ecological environment subsystem) in 30 regions of China has been rising annually. However, the overall level of coordination remains relatively limited. Second, the analysis of kernel density estimation (KDE) shows that the coordination degree between ES and EES across various regions exhibits considerable variability, with the disparity becoming increasingly pronounced. Third, trend surface (TS) analysis indicates that there exist regional variations in the degree of coordination between ES and EES. Specifically, the east experiences an upward trend, while the west exhibits a downward trend. Similarly, the south shows an increase, whereas the north demonstrates a decrease. With ongoing development, it has been observed that the coordination degree remains relatively stable in the east–west direction; however, the disparity between the north and south is increasing. Fourth, an analysis of global Moran’s I reveals a pronounced positive spatial autocorrelation. Finally, the analysis of local Moran’s I reveals that Jiangsu, Fujian, Anhui, and Jiangxi provinces exhibit significant high–high clusters, while the three regions of Xinjiang, Gansu, and Ningxia have always been low–low clusters.

1. Introduction

The ongoing population growth and swift economic development, to a large extent, may place some strain on the ecological environment. Nevertheless, the primary contributor to environmental pollution and ecological degradation lies in human socio-economic activities [1]. A developing economy will, to some extent, cause ecological pollution. There is a contradiction between ES and EES. Therefore, we have to minimize the damage to the environment while developing the economy as fast as possible and balance the co-ordinated development of ES and EES aspects, which is a topic we have to pay attention to. Consequently, what is the degree of coordination between ES and EES in China? What are the disparities? And what differences exist in the intricate spatial distribution of ES and EES at various developmental stages? Environmental pollution and ecological degradation have emerged as significant issues of societal concern. Therefore, examining sustainable development in ES and EES holds significant practical value. This endeavor aspires to promote an enduring harmonious coexistence of humanity and nature while alleviating the ecological pressures resulting from economic growth.

China’s economy has shifted from a stage of rapid growth to one characterized by high-quality development. Consequently, the economy’s progress is increasingly centered around quality and efficiency [2]. For instance, data published by the National Bureau of Statistics (NBS) of China in 2024 indicated that the real GDP growth rate for the initial quarter of 2024 reached 5.3%. The United States occupies the status of being the world’s largest economy, and according to website of the US government, its real GDP growth in the first quarter of 2024 was 1.6%, which is still higher than China’s total economic output, despite the relatively slow pace of economic growth. Meanwhile, in developed countries like Germany and Japan, economic recovery remains sluggish, with Germany’s GDP growth rate declining by 0.8% in the first quarter of 2024, according to gross domestic product data in the Federal Statistical Office of Germany. According to Quarterly Estimates of GDP in the Statistics Bureau of Japan, Japan’s GDP growth rate declined by 0.5% in the first quarter of 2024. China, however, is confronted with more severe issues of air pollution and water shortage as a result of its rapid industrialization and urbanization. Nevertheless, there has been an improvement in environmental quality in recent years. In contrast, Europe has a long-standing tradition of environmental protection and generally enjoys better environmental quality, despite also facing some emerging environmental challenges, such as climate change. In Japan, after years of treatment, industrial pollution has been effectively controlled and environmental quality has improved, but there are new problems in areas such as nuclear pollution.

Consequently, attaining a harmonious equilibrium between the sustainable and co-ordinated development of ES and EES has emerged as a crucial task. BenDor et al. (2015) discovered increasing evidence suggesting that the restoration industry safeguards the environment and significantly fosters economic growth and creating jobs [3]. Raihan (2024) found that financial growth, energy use, and economics in Singapore significantly influence ecological health [4]. When there is rapid economic growth, there is a tendency to focus on the economy itself and ignore the consequences associated with the environment. Yasin et al. (2024) provide examples of the consequences associated with economic growth. We may take the BRICS economies as an example. These nations have witnessed high economic growth rates, yet their industrial expansion has led to environmental pollution [5]. Many countries are grappling with the challenge of ES and EES [6]. This research centers on sustainable and ES and EES coordinated development in 30 regions (excluding Hong Kong, Macao, and Taiwan regions and the Xizang Autonomous Region, the same below) of China.

There has been a lot of research on economics and ecological environment-related topics by many researchers. In the 1950s, Kuznets put forward the Kuznets curve [7]. Based on the Kuznets curve, researchers have conducted extensive research on the economy and environment [8,9,10]. The CCD model can effectively uncover interactions among various systems [11]; as a result, research on the coordinated development of the ES and EES is increasingly gaining attention. For instance, Ma et al. (2013) undertook a study examining the coordination degree between ES and EES in 350 regions across China. The findings suggested that the coordination degree was comparatively low [12]. Fan et al. (2019) explored economic and environmental coordination degree among different provincial capital cities in China, focusing on the perspective of large-scale urban centers. The results demonstrated that there is a relatively large gap in their coupling coordination levels [13]. Liu et al. (2021) [14] and Deng (2022) [15] conducted a study focusing on the perspective of urban clusters. Liu et al. (2021) identified a significant disparity in coordination levels between the ES and EES among cities along the Yellow River Basin [14]. Deng (2022) demonstrated that the level of coordination between ES and EES is on the rise [15]. Other studies have analyzed coupling coordination among three systems, among which economic and environmental systems are included. For instance, research has been undertaken to investigate the synergistic coordination among ecology, economy, and tourism [16,17]; economy, ecological environment, and society [18]; economy, ecological environment, and climate [19]; and environment, economy, and energy [20]. Spatial difference in coupling coordination relationships has been researched from different perspectives, for example, air pollution and carbon emissions [21], ecosystem service value and urbanization [22], and agricultural development and rural revitalization [23]. However, comprehensive and systematic research on the spatial differences in coordination development of ES and EES is lacking. Therefore, this research concentrates on analyzing the coordinated development of ES and EES and their spatial variation in 30 regions of China.

Multi-Criteria Decision Analysis (MCDA) is widely employed across various aspects of the environmental, educational, and economic domains. With the development of technology, new research methods are constantly emerging, such as TOPSIS-COMET, ESP-COMET, and SESP-SPOTIS. These new methods have different characteristics. For instance, the TOPSIS-COMET method is applicable to both linear and nonlinear problems [24]. The introduction of the ESP concept in the ESP-COMET methodology significantly streamlines the process of discerning expert preferences [25]. The SESP-SPOTIS methodology broadens knowledge and comprehension of the issue [26], considering the complexity of the system of indicators of ecological environment, as well as the ease of use of the methodology. There are relatively few applications of new methods in the field related to ecological environment. There are numerous methods employed in MCDA. Examples of methodologies include the Analytic Hierarchy Process (AHP), Fuzzy Analytic Hierarchy Process (FAHP), entropy-weight method (EWM), and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) method, among others. The AHP assesses the comparative significance of various indicators by comparison, but over-reliance on subjective evaluation methods leads to the problem of insufficient persuasive results [25]. Although the FAHP method introduces fuzziness for subjective judgment, it still provides subjective evaluation. In MCDA, the weights of indicators are of great significance. Traditional subjective weighting methods, such as the expert scoring method, are highly susceptible to the subjective judgment of experts. As a result, the weights assigned by different experts can vary significantly [27]. Sun (2021) employed EM to assess indicators of economic performance [28]. Lee et al. (2024) employed the EWM framework to meticulously assess ecological resilience [29]. The EWM is a technique for determining weights by calculating the entropy values of indicators. As a method of weight assignment grounded in objectivity, entropy weighting ascertains the significance of various objectives by leveraging the intrinsic information entropy embedded in the data. TOPSIS can reflect the relative strengths and weaknesses of evaluation objectives. It does so by assessing the proximity of the evaluation objectives to their idealized counterparts, enabling the evaluation of the comprehensive levels of different objects. Although the TOPSIS method typically depends on subjective techniques for weight determination, the weights can be combined with objective methods to determine the weights. For example, EWM represents a method of objective weighting that reduces the subjectivity of the TOPSIS method. Therefore, this research integrates the EWM with the TOPSIS evaluation method. The calculated weights are then employed as the entropy-based weights. The EWM not only reflects the importance of each indicator within ES and EES but also captures the temporal variation in each indicator’s weight. As a consequence, it is widely employed to ascertain the weights of ES and EES indicators [30]. Then, we employ the TOPSIS method to perform a comprehensive evaluation of both ES and EES. By combining EWM with the TOPSIS method, this approach fully leverages the information provided by data attributes [31].

Expert scoring methods may rely too much on expert judgment. Therefore, when there exists a marginal likelihood that the expert’s judgment will be wrong, the RANCOM emerges as an exceptionally suitable solution [32]. The RANCOM method is a robust and reliable subjective weighting method [33]. Because this research mainly considers the information of the fully used data to evaluate the ecological and economic system, subjective evaluation methods were not used.

Building on previous research, indicator systems of ES and EES in most studies have only taken into account certain regions. Moreover, the construction of the indicator system has not been comprehensive enough. We thus present a thorough evaluation of both indicator systems. Firstly, the EWM-TOPSIS methods were employed to provide an integrated assessment of both the economic and ecological subsystems. Then, the CCD model was subsequently utilized to integrate ES and EES, so as to measure the coordination level between ES and EES [34]. After the CCD analyses, it is equally important to investigate their coordination relationship from the perspective of spatial differences. Therefore, this research employs kernel density estimation (KDE), trend surface (TS) analysis, and a spatial autocorrelation (SA) model to analyze the characteristics of spatial differences. Finally, through the study of the coordinated development and spatial differences in ES and EES, this research can enrich the relevant theoretical foundation and offer some specific suggestions on ES and EES. This research is meticulously organized as follows:, Section 2 delineates the selection of indicators and the construction of the model. Section 3 presents an analysis of the model’s results. Section 4 engages in a discussion regarding the implications of these findings. Section 5 summarizes and plans for future research. The following Figure 1 is a theoretical analysis framework for ES and EES.

2. Materials and Methods

2.1. Selection of Variables

Drawing on the existing literature regarding the economic and ecological environment, this research utilizes data collected from 30 regions in China spanning from 2011 to 2022. The system establishes an evaluation framework for indicators from the perspectives of ES and EES. So, this research analyzes two subsystems, which are the economic subsystem (ES) and the ecological environment subsystem (EES). In

Table 1. Indicator evaluation system for ES and EES.

Subsystem	Variables	Units	Direction	Weight
Economic Subsystem (ES)	Regional Per Capita GDP (RPCGDP)	CNY	+	0.125
	Growth Rate of GDP (GDPGR)	%	+	0.014
	Total Retail Sales of Consumer Goods (CGTR)	CNY 100 Million	+	0.208
	Growth Rate of Total Investment in Fixed Assets (FATIGR)	%	+	0.009
	Total of International Trade in Goods (ITG)	USD Thousand	+	0.512
	Per Capita Disposable Income of Urban Households (PCDUHI)	CNY	+	0.132
Ecological Environment Subsystem (EES)	Forest Coverage Rate (FCR)	%	+	0.132
	Per Capita Water Resources (PCW)	m3/Person	+	0.333
	Public Recreational Green Space Per Capita (PCGS)	m2/Person	+	0.066
	Green Covered Area as of Completed Area (GCA)	%	+	0.031
	Rate of Domestic Garbage Harmless Treatment (DGHTR)	%	+	0.020
	Integrated Reuse Rate of Industrial Solid Wastes (ISWIR)	%	+	0.085
	Investment Completed the Treatment of Industrial Pollution (IPTI)	CNY 10,000	+	0.253
	SO2 Emission	10 000 Tons	−	0.026
	Common Industrial Solid Wastes Generated (ISWG)	10,000 Tons	−	0.035
	Particulate Matter Emission (PM)	10,000 Tons	−	0.020

, “Units” denotes the units of each variable, “Direction” indicates the influence of each variable on the subsystem, “+” indicates a positive influence on the subsystem, and “−” indicates a negative influence on the subsystem. The weights are calculated using the EWM. All the data have been meticulously drawn from the National Statistical Yearbook (NSY). However, data for Xizang were excluded because some of the relevant data were missing. According to the 2011 National Bulletin on the Status of Environmental Protection-Related Industries released by the Ministry of Environmental Protection, a national survey of environmental protection-related industries was carried out with 2011 set as the base year. However, the statistics in the China Statistical Yearbook changed in 2011. For example, the statistical approach to particulate matter emissions has changed. Before 2011, the particulate matter indicator consisted of three components—industrial soot emissions, domestic soot emissions, and industrial dust emissions—whereas after 2011, the particulate matter indicator became composed of two components: soot emissions and dust emissions. To ensure the availability, uniformity, and completeness of data, this research commenced its study period in 2011. Due to the fact that some of the data for 2023 were not updated, the research period for this study is 2011–2022.

Referring to the existing research, six variables, including RPCGDP, GDPGR, CGTR, FATIGR, ITG, and PCDUHI, are selected as the economic development indicators. RPCGDP reflects the development of the regional economy. The higher the RPCGDP, the more robust the economic development of the region [35]. GRGDP refers to the annual growth rate of GDP. It serves as an indicator that reflects the pace, whether rapid or slow, at which the GDP is expanding [36]. CGTR reflects the purchasing power of social goods and the size of the retail market [37]. The higher the CGTR, the higher the purchasing power of consumers and the better the standard of living of people in the region. Referring to the explanation of the indicator in the NBS, FATIGR serves as a comprehensive indicator that reflects the magnitude, composition, and growth rate of investment in fixed assets. A higher FATIGR bodes well for a region’s development [38]. The total imported and exported trade in goods from the NSY is used to express the Total of International Trade in Goods, which reflects the size of a region’s trade in the importation and exportation of goods. The higher the aggregate volume of trade, the larger the scale of a region’s import and export trade activities. This also indicates a higher degree of external openness for the region [39]. PCDUHI can visually reflect the income and expenditure situation and living standard of the residents [37].

Ten variables, including FCR [40], PCW [41], PCGS [42], GCA [42], DGHTR [43], ISWIR [40], IPTI [42], SO₂ emission [44], ISWG [44], and PM [45], are selected as the indicator evaluation system of the ecological environment. Among these, FCR, PCW, PCGS, and GCA can show how good or bad the ecological environment is. DGHTR, ISWIR, and IPTI are all indicators reflecting ecological and environmental treatment. SO₂ emission, ISWG, and PM are all indicators reflecting environmental pollution. The influence of this indicator on the ecological environment is detrimental; thus, lower emissions are preferable.

2.2. Research Methods

2.2.1. Entropy-Weight TOPSIS Method Construction

The CCD model is utilized to evaluate the interactions among two or more systems. The CCD reflects the process of their coordinated development [46]. Before the CCD analysis, the EWM-TOPSIS method was employed to perform a comprehensive evaluation of the ES and EES. Therefore, the EWM-TOPSIS model is constructed initially.

The TOPSIS method adeptly illustrates the relative merits and demerits of evaluative criteria by assessing their proximity to idealized objectives [47]. The TOPSIS method is capable of identifying both the Positive Ideal Solution (PIS) and the Negative Ideal Solution (NIS) across multiple objectives. Subsequently, one can determine the degree of proximity of each objective to the optimal solution [48]. Finally, through proximity degree, relative advantages or disadvantages of the evaluation objectives are judged. The proximity degree is in the range of $[0, 1]$. A value near 1 suggests that the corresponding evaluation objective approaches an optimal level. On the contrary, values near 0 indicate that the evaluation objective is close to the worst level [49]. The method of operation is as follows:

First, we construct an evaluation matrix; usually there are $n$ evaluation objectives $D_1,D_2,\dots D_m$ ($i=1,2,\cdots m)$, and each evaluation objective has $m$ evaluation indicators $X_1,X_2,\dots n$ ($j=1,2,\cdots n$) [48]. In this study, the evaluation objectives pertain to various regions within China. Each region has $n$ evaluation indicators to obtain the evaluation matrix as Equation (1).

$$D=\left[X_1\left(x_1\right),\cdots ,X_j\left(x_i\right),\cdots ,X_n\left(x_m\right)\right]$$

(1)

Indicators were normalized in order to compare indicators with different attributes [50]. There are various normalization techniques available, including min–max normalization and Z-score normalization, among others. The min-max normalization involves rescaling the data to a fixed range of [0, 1] through the min-max scaling method [51]. The Z-score normalization method assumes that data should follow a normal distribution, which may lead to biased results if data do not follow a normal distribution. When analyzing multivariate data with different values, the min–max normalization method usually produces better results compared to the Z-score method [52]. Therefore, we selected the min–max normalization method.

$$Z_{ij}=\left\{\begin{array}{c}{\displaystyle \frac{x_{ij}-x_j^{min}}{x_j^{max}-x_j^{min}}}, Positive indicator\\ {\displaystyle \frac{x_j^{max}-x_{ij}}{x_j^{max}-x_j^{min}}}, Negative Indicator\end{array}\right.$$

(2)

In Formula (2), $Z_{ij}$ indicates the value after normalization, and $x_{ij}$ denotes the original value. Then, normalization produces relative weights $P_{ij}$ [53].

$$\begin{gathered} P_{ij}=Z_{ij}/\sqrt{{\sum }_{i=1}^n{Z}_{ij}} \\ i=1,2,\cdots m and j=1,2,\cdots n \end{gathered}$$

(3)

The entropy value $E_j$ and the weights $w_j$ are determined through Equation (3) [54].

$$E_j=-k{\sum }_{i=1}^n{P}_{ij}\ln P_{ij}\mathrm{},\mathrm{}k={\displaystyle \frac{1}{\ln m}}$$

(4)

$$w_j=(1-E_j)/{\sum }_{j=1}^m\left(1-E_j\right),w_j\in \left[\mathrm{0,\; 1}\right], and {\sum }_{j=1}^m{w}_j=1$$

(5)

The computed normalization values are multiplied by their weights to derive the normalization matrix $v_{ij}$ [55].

$$\begin{gathered} v_{ij}=P_{ij}\times w_j \\ i=1,2,\cdots m and j=1,2,\cdots n \end{gathered}$$

(6)

We then determine the PIS $V^{+}$ and the NIS $V^{-}$ [56].

$$\begin{array}{c}{V}^{+}=\left\{{\mathit{max}v}_{ij}\right.\left.\left|i=1,2,\cdots n\right.\right\}\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.5em}=\left\{v_1^{+},v_2^{+},\cdots ,\right.\left.v_j^{+},\cdots ,v_m^{+}\right\}\end{array}$$

(7)

$$\begin{array}{c}{V}^{-}=\left\{{\mathit{min}v}_{ij}\right.\left.\left|i=1,2,\cdots n \right.\right\}\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.5em}=\left\{v_1^{-},v_2^{-},\cdots ,\right.\left.v_j^{-},\cdots ,v_m^{-}\right\}\end{array}$$

(8)

In calculating the weighted Euclidean distance from evaluation indicator to both PIS and NIS, the distance from the evaluation indicator to PIS, denoted as $V^{+}$, is represented by $D^{+}$, while the distance from the evaluation indicator to NIS, denoted as $V^{-}$, is represented by $D^{-}$ [57].

$$D_j^{+}=\sqrt{{\sum }_{j=1}^n{\left(v_{ij}-v_j^{+}\right)}^2}$$

(9)

$$D_j^{-}=\sqrt{{\sum }_{j=1}^n{\left(v_{ij}-v_j^{-}\right)}^2}$$

(10)

In Formula (9), $v_j^{+}$ represents the maximum value of the evaluation indicator across the entire region. In Equation (10), $v_j^{-}$ denotes the minimum value. Finally, the relative proximity degree $C_j$ of multiple evaluation objectives to the ideal solution in the whole region is calculated.

$$C_j={\displaystyle \frac{D_j^{+}}{D_j^{+}+D_j^{-}}} (1\le j\le n)$$

(11)

In Formula (11), the value of $C_j$ ranges from 0 to 1. A value of $C_j$ closer to 1 indicates a relatively high level of integration in the region. On the contrary, a value of $C_j$ closer to 0 indicates that the integration level of the region is relatively weak [57].

2.2.2. CCD Model Construction

This research has two subsystems, the ES and EES. The CCD is modeled as Formula (12) [14,58].

$$C=2\sqrt{U_1·U_2/{(U_1+U_2)}^2}$$

(12)

$$D=\sqrt{C\times T}, T=\alpha U_1+\beta U_2$$

(13)

In Formula (12), $C$ represents the coupling degree, which reflects the intensity of interaction between ES and EES. $U_1$ denotes integrated economic development, while $U_2$ signifies the integrated ecological environment. The values for $U_1$ and $U_2$ are derived using the EWM-TOPSIS modeling method described above. The proximity degree $C_j$ calculated by Formula (11) indicates $U_1$ and $U_2$ here. In Formula (13), $T$ is the coordination index and $\alpha$ and $\beta$ are the degree of contribution, and in this research, ES and EES are equally important, so $\alpha =\beta =0.5$ [21]. In Formula (13), $D$ represents the coordination degree, which illustrates the extent of interaction among ES and EES, as well as the level of coordinated development of their respective strengths and weaknesses. $D\in \left[0,1\right]$, where $D$ closer to 0 indicates worse ES and EES coordination development [21]. The classification of CCDs is shown in

Table 2. CCD classification standards.

D Value	Coordination Level
0~0.099	Extremely uncoordinated
0.1~0.199	Seriously uncoordinated
0.2~0.299	Moderately uncoordinated
0.3~0.399	Slightly uncoordinated
0.4~0.499	On the verge of uncoordinated
0.5~0.599	Barely coordinated
0.6~0.699	Slightly coordinated
0.7~0.799	Moderately coordinated
0.8~0.899	Well coordinated
0.9~1	Quality coordination

[59].

2.2.3. Kernel Density Estimation Model Construction

Through examination, it has been observed that there exists an imbalance in coordination level between ES and EES in China considering spatio-temporal differences between ES and EES. This research was subsequently analyzed utilizing kernel density estimation (KDE).

KDE represents a nonparametric approach employed to evaluate distributional characteristics of random variables [60]. Kernel density estimation has become a common methodology for researching spatio-temporal differences in regional economic and ecological environments [61]. The formula for kernel density estimation is shown in Formula (14).

$$f\left(x\right)={\displaystyle \frac{1}{nh}}{\sum }_{i=1}^{n}k\left({\displaystyle \frac{{x-X}_i}{h}}\right)$$

(14)

In Formula (14), $X_i$ indicates the coordination degree between ES and EES. $x$ indicates the average value of the coordination degree. $n$ represents how many observations are in the sample. $h$ represents the bandwidth; as the $h$ value increases, the KDE curve becomes smoother, but it also increases the deviation of the fit.

$$K(\eta )={\displaystyle \frac{1}{\sqrt{2\pi }}}{e}^{-{\displaystyle \frac{1}{2}}{\eta }^2}$$

(15)

$$\eta ={\displaystyle \frac{{x-X}_i}{h}}$$

(16)

$k(·)$ is a weighting function, including a Gaussian kernel, Epanechnikov kernel, etc. This research utilizes the Gaussian kernel function [62]. The formula is shown in Formula (15).

2.2.4. Trend Surface Analysis Model Construction

According to the aforementioned kernel density estimate analysis of the coordination degree of ES and EES in China, the gap between the two groups has grown in each location. Taking into account the regional variations in ES and EES, in this research, the coordination degree of both ES and EES is utilized as the observation value. Trend surface analysis offers insights into the spatial distribution and changing trends of the coordination degree between ES and EES across north–south and east–west directions.

A trend surface serves as an approximation of a real-world surface, effectively illustrating the spatial distribution of observed values [63]. In this research, the coordination degree is placed in geospatial coordinates, and through the fitted curves formed, the trends of the values in different directions are observed. The formula for trend surface analysis is shown in Formula (17) [64].

$$Z_i\left(x_i,y_i\right)={\beta }_0+{\beta }_1{x}_i+{\beta }_2{y}_i+{\beta }_3{x}_i^2+{\beta }_4{y}_i^2+{\beta }_5{x_{i}y}_i+{\epsilon }_i$$

(17)

In Formula (17), $Z_i\left(x_i,y_i\right)$ is the value of the object of research, and in this research, it is the degree of coordination. The coordinates $\left(x_i,y_i\right)$ represent the geographic location of the $i\mathrm{t}\mathrm{h}$ object. The term ${\epsilon }_i$ represents the random error.${\beta }_0$ is a constant term. ${\beta }_1$–${\beta }_5$ are variable coefficients.

2.2.5. Spatial Autocorrelation Model Construction

This research investigates the potential spatial correlation between the coordination levels of ES and EES, taking into account the spatial disparities inherent in these levels [65]. Commonly used methods for spatial correlation tests are global spatial (GS) and local spatial (LS) autocorrelation analysis.

This research used two methods, GS autocorrelation and LS autocorrelation, to analyze the ES and EES coordination degree [22]. The formula for Global Moran’s I is as follows (18):

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

(18)

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

(19)

In Formula (18), $x_i$ and $x_j$ indicate the coordination degree of regions $i$ and $j$. The symbol $\overline{x}$ represents the mean coordination degree across all regions. The variable $n$ represents how many regions there are. Moran’s I value falls within the range $[-1,1]$ [66]. $Moran\mathrm{’}s I>0$ suggests a positive spatial correlation. Conversely, when $Moran\mathrm{’}s I<0$, it signifies a negative spatial correlation. A value of $Moran\mathrm{’}s I=0$ suggests that the distribution across space is random and there exists no spatial correlation. Higher values of Moran’s I reflect a stronger spatial correlation. The term $w_{ij}$ denotes the spatial weight matrix, which takes on a value of 1 when $x_i$ and $x_j$ are adjacent. When $x_i$ and $x_j$ are not adjacent, it is 0. Significance was determined using the Z-value test [67]. The formula is as follows (20):

$$Z={\displaystyle \frac{1-E\left(I\right)}{\sqrt{V\left(I\right)}}}$$

(20)

In Formula (20), $E\left(I\right)$ indicates the mean Moran’s I value, and $\sqrt{V\left(I\right)}$ indicates the variance of Moran’s I.

Global spatial autocorrelation shows overall spatial correlation, while local spatial autocorrelation shows the spatial correlation of adjacent regions. Spatial correlations include high–high values, low–low values, high–low values, and low–high values [68]. The local Moran’s I formula is (21) as follows:

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

(21)

In Formula (21), $S^2$, $x_i$, $x_j$, $\overline{x}$ and $w_{ij}$ indicate the same meaning as in Formula (17). When $Moran\mathrm{’}s I>0$, it signifies that high values (or low values) are clustered among other high values (or low values). Conversely, when $Moran\mathrm{’}s I>0,$ it indicates the presence of a high value (or low value) surrounded by lower values (or higher values) [69].

3. Results

3.1. CCD Model Result Analysis

3.1.1. Time Perspective

In

Table 3. CCD ES and EES (time perspective).

Year	C	T	D	Coordination Level
2011	0.823	0.169	0.366	Slightly uncoordinated
2012	0.830	0.181	0.380	Slightly uncoordinated
2013	0.837	0.198	0.401	On the verge of uncoordinated
2014	0.854	0.208	0.415	On the verge of uncoordinated
2015	0.878	0.199	0.411	On the verge of uncoordinated
2016	0.889	0.208	0.422	On the verge of uncoordinated
2017	0.912	0.211	0.431	On the verge of uncoordinated
2018	0.921	0.218	0.439	On the verge of uncoordinated
2019	0.932	0.221	0.445	On the verge of uncoordinated
2020	0.933	0.221	0.445	On the verge of uncoordinated
2021	0.941	0.235	0.459	On the verge of uncoordinated
2022	0.945	0.237	0.462	On the verge of uncoordinated

, C represents the coupling value. T denotes the comprehensive evaluation for both ES and EES. D signifies the coordination degree (D-level). The degree of coordination is based on Table 2.

From Table 3, it can be seen that the ES and EES D-level in 30 regions of China increased from 0.366 in 2011 to 0.462 in 2022. The D-level changed from “slightly uncoordinated” to “on the verge of uncoordinated”. It can be found that the coordination degree between ES and EES increases year by year. Notably, there was a slight decrease in coordination in 2015, which may be due to the decrease in government investment in environmental pollution treatment. According to the China Statistical Yearbook, the total investment in environmental pollution treatment for the year 2015 amounted to CNY 880.630 billion. In comparison, the total investment in this sector for 2014 was CNY 957.550 billion, reflecting a decrease of CNY 76.920 billion from the previous year. Due to the severe environmental pollution and the ongoing deterioration of environmental quality, addressing environmental treatment has become a long-term challenge. There is a significant demand for effective environmental treatment but insufficient investment. In general, the coordination degree is low, indicating significant potential for improvement.

3.1.2. Regional Perspective

Table 4. CCD ES and EES (regional perspective).

Regions	C	T	D	Coordination Level
Beijing	0.973	0.283	0.524	Barely coordinated
Tianjin	0.990	0.175	0.417	On the verge of uncoordinated
Hebei	0.950	0.162	0.390	Slightly uncoordinated
Shanxi	0.893	0.124	0.332	Slightly uncoordinated
Inner Mongolia	0.908	0.163	0.383	Slightly uncoordinated
Liaoning	0.983	0.161	0.397	Slightly uncoordinated
Jilin	0.860	0.143	0.350	Slightly uncoordinated
Heilongjiang	0.832	0.156	0.360	Slightly uncoordinated
Shanghai	0.943	0.301	0.531	Barely coordinated
Jiangsu	0.942	0.380	0.597	Barely coordinated
Zhejiang	0.989	0.349	0.586	Barely coordinated
Anhui	0.922	0.175	0.401	On the verge of uncoordinated
Fujian	0.954	0.265	0.502	Barely coordinated
Jiangxi	0.817	0.197	0.401	On the verge of uncoordinated
Shandong	0.982	0.310	0.551	Barely coordinated
Henan	0.972	0.181	0.418	On the verge of uncoordinated
Hubei	0.956	0.186	0.422	On the verge of uncoordinated
Hunan	0.916	0.194	0.422	On the verge of uncoordinated
Guangdong	0.903	0.518	0.683	Slightly coordinated
Guangxi	0.802	0.196	0.397	Slightly uncoordinated
Hainan	0.743	0.180	0.365	Slightly uncoordinated
Chongqing	0.908	0.179	0.403	On the verge of uncoordinated
Sichuan	0.958	0.188	0.424	On the verge of uncoordinated
Guizhou	0.787	0.154	0.347	Slightly uncoordinated
Yunnan	0.799	0.182	0.381	Slightly uncoordinated
Shaanxi	0.886	0.159	0.375	Slightly uncoordinated
Gansu	0.864	0.094	0.285	Moderately uncoordinated
Qinghai	0.604	0.262	0.397	Slightly uncoordinated
Ningxia	0.866	0.113	0.312	Slightly uncoordinated
Xinjiang	0.837	0.134	0.334	Slightly uncoordinated

examines ES and EES from a regional perspective. From Table 4, it can be seen that the D-level of Gansu is moderately uncoordinated. The coordination level among the 14 regions (including Hebei, Shanxi, Inner Mongolia, Liaoning, Jilin, Heilongjiang, Guangxi, Hainan, Guizhou, Yunnan, Shaanxi, Qinghai, Ningxia, and Xinjiang) is found to be “slightly uncoordinated”. The coordination level among the eight regions of Tianjin, Anhui, Jiangxi, Henan, Hubei, Hunan, Chongqing, Sichuan is found to be “on the verge of uncoordinated”. The coordination level among the six regions of Shanghai, Fujian, Beijing, Zhejiang, Jiangsu, and Shandong is found to be “barely coordinated”. The coordination degree of Guangdong is “slightly coordinated”.

From Table 4, Gansu exhibits the lowest coordination degree with a value of 0.285, indicating that it is moderately uncoordinated. Due to the geographic location of Gansu in the northwest of China and the Loess Plateau region, the region’s economic development primarily relies on traditional agriculture and resource-based industries, which can lead to detrimental effects on the environment [70]. Gansu’s economic development is relatively backward and its resources are relatively insufficient, so the quality of its ecological environment is worse compared to other regions. Hebei, Shanxi, Inner Mongolia and 14 other regions are slightly uncoordinated. The D-level of Tianjin, Anhui, Jiangxi and 14 other regions is “on the verge of uncoordinated”. The D-level of these 23 regions shows that they are not coordinated. Most of these 23 regions are situated in the central and western parts of China. The level of economic development in these central and western regions is relatively less advanced compared to that of the eastern regions. This phenomenon may be attributed to the prioritization of the economy in these regions, which inevitably leads to a certain degree of pollution, affecting the ecological environment. Consequently, this results in a low coordination level between ES and EES.

Beijing, Shanghai, Jiangsu and six other regions are barely coordinated. The coordination degree of Guangdong is “slightly coordinated”, with the highest value of 0.683. According to annual data from NBS, Guangdong is located in the Pearl River Delta and has a well-developed economy, with the highest GDP of CNY 13,567.320 billion in 2023. Its natural resources are abundant, so compared with other regions, the coordination degree between ES and EES is higher. These seven regions are coordinated. These seven regions are all located in the eastern of China, which as a whole has a more developed economy. This shows that these seven regions have reasonably managed the ecological environment while developing their economies.

3.2. Kernel Density Estimation Result Analysis

Due to limited space, only some years are analyzed for spatio-temporal differences. In this research, four years, 2011, 2015, 2019, and 2022, are selected to represent the trend of spatio-temporal differences from 2011 to 2022. The change trend is visualized in Figure 2.

As illustrated in Figure 2, the KDE curve exhibits a rightward shift from a positional perspective. This trend indicates an overall increase. From the distribution of curves, the right tail of the KDE curves in these four years lengths, indicating that the spatial gap is gradually expanding. The kernel density curves for each of the four years exhibit twin peaks in terms of peak count, indicating a multipolar polarization in the coordination degree of ES and EES. In terms of the width of the shape of the curve, the 2011–2022 kernel density curve becomes wider and wider, which indicates increasing differences between regions. So, what kind of difference? This research uses trend surface analysis to further analyze the differences in the D-level of ES and EES.

3.3. Trend Surface Result Analysis

As with the spatio-temporal difference analysis, four years, 2011, 2015, 2019, and 2022, were selected for analysis. The trend surfaces are depicted in Figure 3. In Figure 3, the X-axis is oriented toward the east, and the Y-axis is directed toward the north. The green curve illustrates the spatial trend in the east–west direction. Conversely, the blue curve represents the spatial trend of ES and EES in the north–south direction.

As shown in Figure 3, in the east–west direction, the eastern region exhibits a higher D-level compared to the western region. From a north–south perspective, the southern region exhibits a greater degree of harmonization compared to the northern region. From the perspective of time, the variations in D-level have remained relatively stable in the east–west direction; however, they have been on the rise in the north–south direction.

The eastern region has a more developed economy and richer resources, with a higher coordination degree. However, the economic development of the western region is comparatively less advanced than that of the eastern region. Consequently, there is a greater emphasis on rapid economic growth, often at the expense of ecological considerations. This oversight has resulted in a low degree of coordination among ES and EES. The southern economy is more developed than the northern one, and the southern economy is given more assistance by the national economic growth policy. The southern region is better than the north in terms of economic openness to the outside world [71]. For example, the pilot free trade zones in Shanghai and Guangdong already belong to the forefront of China.

3.4. Spatial Autocorrelation Result Analysis

3.4.1. Global Spatial Autocorrelation Result Analysis

In this research, the global Moran’s I statistic was employed to examine the spatial correlation of D-levels between ES and EES across 30 regions in China from 2011 to 2022. The findings regarding global spatial autocorrelation are presented in

Table 5. GS Moran’s I autocorrelation.

Year	Moran’s I	Z	p-Value
2011	0.356	3.297	0.001
2012	0.372	3.432	0.001
2013	0.350	3.250	0.001
2014	0.343	3.163	0.002
2015	0.437	3.972	0.000
2016	0.465	4.165	0.000
2017	0.407	3.720	0.000
2018	0.400	3.661	0.000
2019	0.464	4.164	0.000
2020	0.468	4.212	0.000
2021	0.426	3.872	0.000
2022	0.504	4.495	0.000

.

In Table 5, Moran’s I values for 2011–2022 are all more than 0, and the p-values are all significant at 1% and 5% levels. The findings indicate a positive spatial correlation. From a temporal perspective, Moran’s I demonstrates an overall increasing trend, rising from 0.356 in 2011 to 0.504 in 2022. The findings indicate a progressive enhancement in spatial correlation between ES and EES in terms of coordination.

3.4.2. Local Spatial Autocorrelation Result Analysis

In this research, the local Moran’s I statistic is employed to examine the spatial correlation. As with the kernel density estimates, the four years of 2011, 2015, 2019, and 2022 were selected to analyze spatial correlations. The results of the LS autocorrelation are shown in Figure 4. The regions depicted in white within Figure 4 represent Hong Kong, Macao and Taiwan regions and the Xizang Autonomous Region.

As shown in Figure 4, the number of high–high cluster regions is increasing, with two regions, Jiangsu and Fujian, in 2011 and four regions, Jiangsu, Fujian, Anhui, and Jiangxi, in 2022. This shows a better coordination of ES and EES with the four regions and the surrounding regions. There is a slight change in high–low outliers, with two regions, Heilongjiang and Sichuan, being high–low outliers for all four analyzed years. Inner Mongolia, Qinghai, and Shaanxi had slight changes during these four years. The analysis reveals instability in the coordinated development of ES and EES across these three regions (Inner Mongolia, Qinghai and Shaanxi) and their surrounding regions. The number of low–low cluster regions is on the rise; specifically, there were four such regions (Xinjiang, Gansu, Ningxia, and Shaanxi) identified in 2011. By 2022, this figure increased to six regions (Xinjiang, Gansu, Ningxia, Shaanxi, Inner Mongolia, and Liaoning). This trend indicates that the coordinated development of ES and EES within these areas and their surroundings remains low. Consequently, this situation has led to a proliferation of low–low clusters in adjacent regions.

4. Discussion

4.1. Discussion on the CCD

The coordination degree between ES and EES has been analyzed across 30 regions in China. From a temporal perspective, the findings indicate that the coordination degree among these regions has gradually increased from 2011 to 2022; however, the overall CCD level remains relatively low. From a regional standpoint, it is observed that Guangdong exhibits the highest degree of coordination compared to other regions. This suggests that the economic and ecological environments in Guangdong are more effectively coordinated. Guangdong has a widespread industrial chain and a large market size, and it attracts a lot of foreign investment due to the local government’s emphasis on an open economic policy and its active participation in foreign trade. Guangdong is not only economically developed but also rich in natural resources due to its geographic location, so Guangdong’s economic and ecological environment can develop in a sustainable and coordinated manner. And Gansu has the lowest D-level. This may be due to Gansu’s geographical location in the northwestern region as well as its lack of natural resources and relatively backward economic development. Therefore, when the local government develops economically, due to the lack of natural resources, the economic and ecological environment cannot be developed in a well-coordinated manner.

4.2. Discussion on the KDE

This study utilizes KDE to examine the spatial disparities. The findings indicate that the overall D-level of ES and EES factors is on an upward trend. However, there is a growing spatial disparity in the D-level of ES and EES among these 30 regions. This suggests that the relationship between ES and EES sustainability remains unbalanced, with an increasing gap evident across different areas in China.

4.3. Discussion on the TS Analysis

As the disparities between regions continue to widen, this research further examines the coordination degree of ES and EES through trend surface analysis. The findings indicate that the coordination degree is significantly higher in the east than in the west and is also greater in the south than in the north. As they continue to develop, the east–west disparity is relatively stable, but it is noteworthy that the north–south disparity is gradually widening. This phenomenon may be attributed to the geographical advantages of the southern region, which predominantly lies along coastal areas. This positioning has facilitated a more rapid integration with global markets compared to its northern counterpart. Additionally, the south possesses richer natural resources than the north, contributing to a more mature, advanced, and accelerated coordinated development of both ES and EES in that area.

4.4. Discussion on the SA Analysis

The SA method was used to analyze the spatial correlation. The results from the GS Moran’s I autocorrelation analysis indicate a positive spatial autocorrelation in the coordination degree of ES and EES from 2011 to 2022. The four regions of Jiangsu, Fujian, Anhui, and Jiangxi eventually show high–high clustering. Although Anhui and Jiangxi presented as low–high outliers in 2011, they became high–high clusters in 2015, 2019, and 2022. This finding suggests that these four regions, along with their surrounding areas, exhibit relatively well-coordinated development between ES and EES. The trend surface analysis corroborates this observation by highlighting the geographic advantages and abundant natural resources present in the southern regions, which encompass all four studied areas. Economic development is more advanced in the four regions of Jiangsu, Fujian, Anhui, and Jiangxi. Consequently, it can be inferred that the degree of coordination between ES and EES is high. The three regions of Xinjiang, Gansu, and Ningxia have always been low–low clusters, which indicates that the level of coordination between ES and EES in these three regions and their surrounding regions is relatively poor. In alignment with the findings from the trend surface analysis, three additional regions located in the northwest demonstrate a lower level of coordination between ES and EES due to limited natural resources and comparatively underdeveloped economies.

Similar studies have been conducted on other countries and regions; for example, Liu and Suk (2021) conducted a study on the level of coordination between the tourism economic system and the ecosystem in Nagasaki Prefecture, Japan. Their findings indicated that this degree is relatively low. And the coordination degree generally went through a stage of decreasing and then increasing [72]. The research of Yan and Chen (2019) showed that the ecological environment, economy, and energy systems in Australia had a low coordination degree in most years from 2007 to 2016, but the coordination degree showed an increasing trend [73]. These studies are more consistent with the findings of this research, which shows that the overall D-level of ES and EES is not high, which has become a problem faced by some countries and regions. But the good aspect is that the overall coordination degree of ES and EES in some countries and regions demonstrates a general upward trend.

Due to changes in the statistics of the China Statistical Yearbook, the data are not consistent, and data prior to 2011 cannot be accurately obtained. This resulted in this research only analyzing the data from 2011 to 2022. If the period of this research can be extended, it can reflect the coordination degree and spatial difference in ES and EES more clearly. In order to assess ES and EES, this study solely used an objective entropy weighting approach in conjunction with the TOPSIS method, ignoring subjective techniques like the expert score method. More information would be revealed by contrasting subjective and objective assessment techniques. Because the economy and ecological environment are in constant flux, the coordination degree relationship between ES and EES will also change. Therefore, we will focus on dynamic modeling in our future research.

5. Conclusions

This research examines the sustainable coordinated development of ES and EES across 30 regions in China. The analysis of coordination degree reveals significant variations in the D-level of ES and EES among these regions. Therefore, on this basis, we performed further spatial difference analysis of the D-level of ES and EES. The spatial difference analysis can clearly determine the characteristics of the spatial distribution of the D-level of ES and EES. What variables contribute to the observed regional variations in the degree of coordination between the ES and EES? In future research, the focus will be on what factors influence the coordination degree of ES and EES.

The practicality of this paper is that by researching the coordination degree of ES and EES, it can provide policy recommendations for local governments in different regions, such as areas with faster economic development but greater damage to the ES and EES, where local governments can set strict environmental protection standards, such as restricting enterprises that cause high pollution to the environment, and at the same time support the green development of the industry, such as the new energy industry, green environmental protection industry, etc. Through this research, it has been observed that the ES and EES coordination level in Gansu is low, and the ecology in Gansu is weaker, so the government can obtain funds from economically developed areas through the ecological compensation mechanism based on the “polluter pays” principle for ecological restoration and ecological management, so that resources can be reasonably distributed [74]. A sound ecological environment serves as the cornerstone of a healthy life for individuals; through this research of the ES and EES coordination degree, it can improve people’s quality of life, improve people’s awareness of environmental protection, reduce the emission of pollutants, and give people a healthy and comfortable living environment.

The theoretical significance of this paper is that research on the development of the coordination degree of ES and EES can promote the cross-integration of different disciplines, and this research enriches the theoretical research of economy and ecology by combining both economic and ecological disciplines. It enriches the interdisciplinary research methodology. When constructing models of the coordination degree of ES and EES, a variety of econometric models should be utilized, and the indicators of the economic system and ecosystem should be comprehensively assessed. Geographic models are used to examine spatial differences across various regions. Such an interdisciplinary research method can provide reference for other disciplines and promote communication and cooperation between different disciplines.