Spatial–Temporal Multivariate Correlation Analysis of Ecosystem Services and Ecological Risk in Areas of Overlapped Cropland and Coal Resources in the Eastern Plains, China

Abstract

The overlapped areas of cropland and coal resources play a fundamental role in promoting economic and social progress. However, intensive mining operations in high water-level areas have brought significant spatial–temporal heterogeneity and ecological problems. From the dual dimensions of the ecosystem service value (ESV) and ecological risk (ER), it is of great significance to explore the influence characteristics of underground mining on the landscape, such as above-ground cultivated land, which is valuable to achieving regional governance and coordinated development. In this study, taking Peixian as the research area, a multiple-dimensional correlation framework was constructed based on the revised ESV and ER, integrating the grey relational degree, spatial–temporal heterogeneity, disequilibrium, and inconsistency index to explore the ESV and ER assessment and correlation characteristics from 2010 to 2020. The results show that (1) the ESV showed a high agglomerated distribution pattern in the east, with a net decrease of 13.61%. (2) The ER decreased by 78.18 and was concentrated in the western and southern regions, with overall contiguous and local agglomeration characteristics. This indicates that the ecological security of the region has improved. (3) The comprehensive grey correlation between the cultural service value and the ecological risk index was the highest. Furthermore, the spatial–temporal heterogeneity of the ESV and ER weakened, and the disequilibrium rose and then fell, indicating that the ecosystem gradually tended to be stable. The study is crucial for overlapped cropland and coal resource areas to maintain stability and sustainable development. The multivariate correlation framework provides practical value for ecosystem management and risk control.

1. Introduction

Rapid economic growth inevitably puts pressure on ecosystems. Moreover, the achievements obtained at the expense of resource consumption have exacerbated the unsustainable use of land [1,2]. The scientific evaluation of ecosystem service value and ecological risk is the basis for solving the dilemma of ecological conflict. Overlapped cropland and coal resources areas are administrative regions composed of major grain- and coal-producing areas [3]. Coal mining activities continue to impact the original landscape, change the functions and processes of ecosystems [4,5], and easily cause ecological damage, such as surface subsidence, vegetation degradation, and land desertification. Therefore, applying the ecosystem service theory and ecological risk assessment to the ecological protection and restoration of the overlapped areas of cropland and coal resources at the landscape level is an effective measure to solve the problem and a meaningful way to achieve regional sustainable development.

Ecosystem service value (ESV) and ecological risk (ER) are essential indicators in ecological environment assessment. ESV is the benefit people obtain from the ecosystem [6], directly or indirectly affected by artificial interference. ER is defined as the likelihood of adverse ecological effects due to exposure to multiple risk stressors [7]. The assessment of ESV is an economic process that assigns economic value to an ecosystem and its ecosystem services. The research methods of ESV are mainly based on the equivalent factor of unit area value [8,9] or the price of unit service function [10]. Currently, many studies have shown that changes in ESV are often driven by land use/land cover (LUCC) change [11]. ESVs have been determined by combining remote sensing and other multi-source data [12,13] to analyze the impact of different regions [14,15] or multiple factors [16,17] on the regional ESV. Ecological risk assessment is mainly based on landscape pattern changes [18], with attention paid to the spatial–temporal heterogeneity and scale effect of risk [19,20]. Shifts in the landscape dominated by human activities have regional and cumulative impacts on the ecological environment and have become a hot spot in ecological risk assessment. Ecosystem services are the bridges that link nature and society [21,22], while the degradation of ecosystem services can cause a series of adverse effects or ecological problems [23]. Therefore, research on the integration dimension of ESV and ER could provide important insights into strengthening the utility of decision support for regional ecological environment protection [24,25,26].

The overlapped areas of cropland and coal resources with a high water-level in the eastern plains of China have distinct characteristics [27]. Coal mining leads to different degrees of influence on the structure and function of the original ecosystem [28]. The differences in the mining area and depth directly affect the severity of local soil pollution and aquifer damage, affecting regional ecological security [29]. The land destruction caused by mining will directly affect the ecosystem value of the native landscape [30,31]. Generally, land subsidence will lead to a decrease in the ecological value. However, because the study area is located at a high water-level plain, underground coal mining is likely to cause seasonal or perennial waterlogging [32]. When ecological restoration and governance measures are carried out, such as reusing subsidence water as fish ponds or wetlands, a stable landscape will be formed, increasing the ESV to a certain extent and improving the ER caused by landscape fragmentation. Ultimately, the ESV and ER exhibit more significant spatial heterogeneity [33,34]. However, previous studies on the spatial–temporal equilibrium relationship and diversity of ESV and ER in particular regions are rare [35], and related studies need to be discussed further.

As mentioned above, this study took Peixian, a typically overlapped cropland and coal resource area with a high water-level in the eastern plain of China, as the study area. The revised calculation and correlation analysis of the ESV and ER revealed the spatial–temporal evolution of the ESV and ER in Peixian from 2010 to 2020. The study makes three significant contributions to the literature: (i) we improved the ESV and ER models by analyzing the particular area’s characteristics; (ii) we analyzed the spatial–temporal correlation changes in the ESV and ER; (iii) we discuss some significant implications for ecological risk control and ecological restoration of high water-level overlapped areas of cropland and coal resources in China.

2. Research Methods and Data Sources

2.1. Study Area

Peixian (34°28′–34°59′ N, 116°41′–117°09′ E) is located on the border of Jiangsu and Shandong on the Huang Huaihai Plain and is adjacent to Weishan Lake in the east (Figure 1), with a total area of 1576 km². By the end of 2019, the total population of the study area was about 1.3 million, with a GDP of CNY 76.262 billion (1 CNY = 0.1404 USD). The territory has flat terrain, abundant water sources, concentrated precipitation, and fertile soil, and the productivity index is 0.918. It is an important agricultural area and the main grain-producing area of wheat and corn with flood and drought rotation, and the annual grain output reaches 657,800 tons.

Peixian is the only central coal-producing area in Jiangsu, with over 40 years of coal mining history. It is also an essential part of the Lianghuai coal base and one part of the East China coal accumulation area [36]. The domestic coal field has many thick coal seams, large reserves, good coal quality, and a stable distribution law. It covers 160 km² and has a proven 2.37 billion tons of reserves consisting of eight coal mines, including Zhangshuanglou, Longdong, Yaoqiao, Sanhejian, Longgu, Kongzhuang, and Peicheng (Figure 1). As a typical high diving mining area in the eastern plain, the underwater level is mostly 1–3 m. Under this circumstance, mining would cause varying surface subsidence, typically affecting each area’s agricultural land and water bodies (

Table 1. Status of mining area in Peixian.

Mining Area	Surface Subsidence Area of Mining Area (hm2)			Affected Area of Agricultural Land (km2)	Affected Area of Water Bodies (km2)
	Subsidence Depth < 1.5 m	1.5 m ≦ Subsidence Depth ≦ 3.0 m	Subsidence Depth > 3.0 m
Longdong Mine	204.05	138.98	396.3	11.99	7.71
Yaoqiao Mine	1720.8	590.18	506.09	23.94	19.71
Xuzhuang Mine	355.62	282.28	428.73	13.72	8.92
Kongzhuang Mine	353.29	0.04	0	10.68	13.78
Sanhejian Mine	1658.79	416.39	366.76	15.73	4.69
Zhangshuanglou Mine	1465.02	537.04	328.59	18.77	6.05
Longgu Mine	542.58	19.55	0.48	4.49	5.93
Peicheng Mine	382.92	6.15	0	29.69	1.83

). Compared to conventional mining areas, the study area has unique characteristics, such as a high phreatic level, deep burial, and a complex geological structure.

2.2. Methods

2.2.1. Research Framework

The theoretical framework of the multivariate correlation analysis is presented in Figure 2. The overlapped area of cropland and coal resources is a compound functional region. The change in landscape patterns caused by underground resource exploitation will objectively affect the surrounding natural ecology, social economy, human culture, etc. In this study, the socio-economic data and land use/cover data were used as the primary data for calculation, and the relevant data of the mining area were used as the basis for correction. Then, the ESV and ER in 2010, 2015, and 2020 were used as the primary evaluation indicators. The ESV represents a positive ecological function status of the study area, and the ER reflects the possibility of environmental problems. The analysis of the ESV and ER evolution characteristics intuitively showed the study area's ecological quality. Finally, this study integrated the grey relational degree, spatial–temporal heterogeneity, disequilibrium, and an inconsistency analysis model to explore the correlation between the ESV and ER under the action of multiple factors from multiple perspectives.

2.2.2. Method of Evaluating the Ecosystem Service Value

Based on the ESV method proposed by [37], the net profit of food production in the 1hm2 farmland ecosystem is defined as the standard equivalent. Based on the grain production and socio-economic data in the study area, the coefficients of the standard match were revised to make it suitable for the relevant research in Peixian. The calculation formula is as follows:

$$\left(S{w}_i\times P{w}_i\times y{w}_i+S{c}_i\times P{c}_i\times y{c}_i+S{r}_i\times P{r}_i\times y{r}_i\right)\times e_i\times f_i$$

(1)

$$\overline{\mathrm{V}}={\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{V}}_{\mathrm{i}}/\mathrm{n}$$

(2)

where $\overline{\mathrm{V}}$ represents the standard equivalent (CNY/hm²); ${\mathrm{V}}_{\mathrm{i}}$ represents the standard equivalent in the year i (CNY/hm²); n represents the research period; $S{w}_i$, $S{c}_i,$ and $S{r}_i$ represent the percentages of the sown area of wheat, corn, and rice, respectively, in Peixian in the year i to the total planted area of the three crops; $P{w}_i$, $P{c}_i,$ and $P{r}_i$ represent the national average net profits of the three crops in the year I; $y{w}_i$, $y{c}_i,$ and $y{r}_i$ represent the yield correction coefficients of the year I; $e_i$ and $f_i$ represent the economic correction and fertility correction coefficients, respectively, for the year i. The summary statistics of the data are shown in the

Table 2. Coefficient correction data table.

Year	Proportion of Sown Area (%)			Net Profit (CNY/hm2)			Yield Correction Factor			Economic Correction Factor	Fertility Correction Factor
	Rice	Wheat	Corn	Rice	Wheat	Corn	Rice	Wheat	Corn
2010	0.46	0.44	0.10	270.20	351.79	159.01	0.82	0.94	0.90	0.87	0.76
2015	0.34	0.51	0.14	701.30	207.11	436.21	0.68	0.86	0.87	1.19	1.12
2020	0.34	0.48	0.18	1319.21	1565.20	602.03	0.93	0.99	0.93	1.04	1.03

.

Based on the standard equivalent, the service value coefficients of each ecosystem were corrected to obtain the $\overline{E}$ of different land types per unit area. Different coal mining intensities have discrepant effects on the ESV [38,39,40,41]. Therefore, a revision of the ESV with R for the inner area of the mine boundary was carried out. A standard equivalent correction calculation was only carried out for the area affected by non-coal mining. The revised formula is as follows:

$$ESV=\left\{\begin{array}{c}{\displaystyle \sum _{r=1}^n}\left(S_r·{\overline{E}}_r\right) \mathrm{Outside\; mine\; boundaries}\\ {\displaystyle \sum _{r=1}^n}\left(S_r·{\overline{E}}_r\right)R \mathrm{Within\; mine\; boundaries}\end{array}\right.$$

(3)

In the formula, $S_r$ represents the area of land use type r; n is the number of land use types; ${\overline{E}}_r$ represents the ESV coefficient of land use type r (CNY/${\mathrm{hm}}^2$); R is a correction factor. The mining intensity of the underground resources was divided according to the mining height and mining ratio of each mine (

Table 3. Assessment of mining intensity.

Mining Heights	Plane Mining Ratio
	>60~100%	30~60%	10~30%	0~<10%
≥3.50 m	Extremely High	High	Medium	Low
1.30~3.50 m	High	Medium	Medium	Low
≤1.30 m	Medium	Low	Low	Low

). The expert scoring method was used to quantify the impact of different mining intensities on the ESV. The underground resource mining intensities ranged from low to high, and the quantitative values of the correction factors R were 0.95, 0.74, 0.79, and 0.82, respectively.

2.2.3. Method of Evaluating the Ecological Risk

The destruction of the landscape structure caused by the exploitation of underground resources in the coal mining area will change the corresponding ecological processes, leading to changes or even a loss of ecological functions [42,43]. The ER assessment was constructed based on the risk characterization and evaluation paradigm of landscape patterns [44,45] to quantitatively describe the ecological risk index of the evaluation community.

$$ER={\sum }_{i=1}^n\frac{W_i}{W}\left(a{\phi }_i+b{\gamma }_i+c{\mu }_i\right)F_i$$

(4)

where ER is the ecological risk index; $W$ is the plot area; $W_i$ is the area of the landscape type $i$ in the plot; ${\phi }_i$ is the fragmentation index demonstrating the disturbance degree of the ecosystem; ${\gamma }_i$ is the separation index reflecting the tolerance; ${\mu }_i$ is the dominance index reflecting the stability; a, b, and c are the corresponding index weights (assigned as 0.5, 0.3, and 0.2, respectively); $F_i$ is the landscape vulnerability index, which was modified based on the expert consultation method (

Table 4. Consultation results of experts on the vulnerability of different land use types.

Land Use Type	Regional Status	${\mathit{F}}_{\mathit{i}}\left(\mathbf{Uncorrected}\right)$	Regional Status	${\mathit{F}}_{\mathit{i}}(\mathbf{After} \mathbf{Correction})$
Cultivated land	No mining disturbance area	0.27	Mining disturbance area	0.29
Forest		0.07		0.10
Waters area		0.13		0.14
Built-up land		0.20		0.19
Bare land		0.33		0.28

).

2.2.4. Grey Relational Degree Model

The grey relational degree model represents the geometric proximities among different discrete sequences, reference sequences, and at least one comparison sequence in the system [46]. By adjusting the correlation benchmark value ($\mathsf{\theta }$), the study focused on the absolute quantity and rate of change characteristics between the ER and ESV in the mining area to provide a reference for the decision-making of ecological protection or ecological restoration. When the benchmark value is smaller, it is biased toward the trend correlation of the rate of change; otherwise, it is biased toward the absolute quantity correlation. Assuming that sequences ${\mathrm{X}}_{\mathrm{o}}$ and ${\mathrm{X}}_{\mathrm{j}}$ have the same length and the initial value is not 0, the grey comprehensive correlation degree of the two sequences is as follows:

$${\mathsf{\rho }}_{0\mathrm{i}}={\mathsf{\theta }\mathsf{\epsilon }}_{0\mathrm{i}}+\left(1-\mathsf{\theta }\right){\mathrm{r}}_{0\mathrm{i}} \mathsf{\theta }\in \left[0,1\right]$$

(5)

In the formula, $\mathsf{\theta }$ is the reference value; ${\mathsf{\rho }}_{0\mathrm{i}}$ is the grey comprehensive correlation degree of the two sequences; ${\mathsf{\epsilon }}_{0\mathrm{i}}$ and ${\mathrm{r}}_{0\mathrm{i}}$ are the grey absolute correlation degree and the grey relative correlation degree, respectively, of the two sequences. The relevant indicators can be calculated using Grey System Theory Modeling Software (GSTA V7.0).

2.2.5. Spatial–Temporal Heterogeneity Index

The spatial heterogeneity index and the Theil coefficient reflect the spatial and temporal differences. The spatial heterogeneity assessment model was constructed based on the ESV and ER unit evaluations, exploring the fundamental spatial–temporal differences in the study area. This study used the Theil coefficient based on the ESV and ER to analyze the degree of regional imbalance in different regions. The calculated result is between 0 and 1. The larger the value, the more significant the difference between the ESV and ER in the study area, and the more unbalanced it is. The formulas are as follows:

$${\mathrm{Q}}_{\mathrm{i}}=\frac{1}{2}(\frac{\ln {\mathrm{ESV}}_{\mathrm{i}}{}^{\prime }}{\ln \overline{\mathrm{ESV}{}^{\prime }}}+\frac{\ln {\mathrm{ER}}_{\mathrm{i}}{}^{\prime }}{\ln \overline{\mathrm{ER}{}^{\prime }}})$$

(6)

$$\mathrm{P}=\frac{{10}^4}{\mathrm{N}}\sum _{\mathrm{i}}^{\mathrm{N}}\log \frac{\overline{\mathrm{Q}}}{{\mathrm{Q}}_{\mathrm{i}}}$$

(7)

where ${\mathrm{Q}}_{\mathrm{i}}$ is the spatial heterogeneity coefficient of an evaluation area; ${\mathrm{ESV}}_{\mathrm{i}}{}^{\prime }$ and ${\mathrm{ER}}_{\mathrm{i}}{}^{\prime }$ are the normalized ESV and ER of the evaluation area; $\overline{\mathrm{ESV}{}^{\prime }}$ and $\overline{\mathrm{ER}{}^{\prime }}$ are the normalized mean values; the $\overline{\mathrm{Q}}$ is a global average spatial heterogeneity coefficient; N is the number of evaluation units; and P is the difference in the spatial and temporal distribution of the entire study area.

2.2.6. Disequilibrium Index and Inconsistency Index

The disequilibrium index was used to analyze the evolution of the ESV and ER patterns in the study area, reflecting the concentration and distribution of value and risk in the region [47].

$$U_{ESV-S}=\sqrt{\frac{{\sum }_{\mathrm{i}}^{\mathrm{n}}{[\frac{\sqrt{2}\mathrm{(}\frac{{\mathrm{ESV}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{ESV}}_{\mathrm{i}}}-\frac{{\mathrm{S}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{S}}_{\mathrm{i}}}\mathrm{)}}{2}]}^2}{\mathrm{n}}}$$

(8)

$$U_{ER-S}=\sqrt{\frac{{\sum }_{\mathrm{i}}^{\mathrm{n}}{[\frac{\sqrt{2}\mathrm{(}\frac{{\mathrm{ER}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{ER}}_{\mathrm{i}}}-\frac{{\mathrm{S}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{S}}_{\mathrm{i}}}\mathrm{)}}{2}]}^2}{\mathrm{n}}}$$

(9)

$$U_{ESV-ER}=\sqrt{\frac{{\sum }_{\mathrm{i}}^{\mathrm{n}}{[\frac{\sqrt{2}\mathrm{(}\frac{{\mathrm{ESV}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{ESV}}_{\mathrm{i}}}-\frac{{\mathrm{ER}}_{\mathrm{i}}}{{\sum }_{\mathrm{i}}^{\mathrm{n}}{\mathrm{ER}}_{\mathrm{i}}}\mathrm{)}}{2}]}^2}{\mathrm{n}}}$$

(10)

In the formula, $U_{ESV-S}$, $U_{ER-S}$, and $U_{ESV-ER}$ are the disequilibrium indices of the ESV, ER, and both, respectively; $ES{V}_i$ and $E{R}_i$ represent the ESV and ER, respectively, of the grid cell I; ${\mathrm{S}}_{\mathrm{i}}$ represents the area of the grid cell I; n is the number of grid cells. The larger the value $U$, the higher the imbalance between the ESV and the ER in the study area and vice versa.

As an essential part of the disequilibrium index, the inconsistency index was used to measure the coordination degree of ESV agglomeration and ER agglomeration in the spatial distribution of the entire study area.

$$CPE=\frac{ES{V}_i/{\sum }_{i=1}^{N}ES{V}_i}{E{R}_i/{\sum }_{i=1}^{N}E{R}_i}$$

(11)

where CPE is the inconsistency index, and the other variables have the same meaning as above. If the CPE is less than 1, it means that the agglomeration of the ESV is lower than that of the ER, and conversely, the agglomeration of the ESV is higher than that of the ER. Furthermore, the closer the CPE value is to 1, the more consistent the spatial evolution trend of the ESV and ER.

2.3. Data Sources and Processing

The study selected statistical and biological data from the Statistical Yearbook and National Compilation of Agricultural Product Cost and Benefit Information to calculate the ESV and ER of Peixian. In addition, mine boundaries and conditions from the Fund projects were used. Geographic data came from the geospatial data cloud platform (http://www.gscloud.cn accessed on 23 November 2022). Three periods of remote sensing images of Peixian in 2010, 2015, and 2020 spanned from April to September.

The images were processed using ENVI 5.1 software, mainly including atmospheric radiation correction, geometric correction, and other operations [48]. The natural environment and land use characteristics of Peixian were divided into five types: cultivated land, forest, built-up land, water area, and bare land. To facilitate subsequent calculations, we unified the spatial resolution of all raw data to 900 m × 900 m. For accuracy verification, we randomly selected one thousand sample points and the actual survey land types for each classification result. The overall accuracy rates were 71 %, 79 %, and 83 %, and the Kappa coefficients were 0.79, 0.81, and 0.82, which met the research requirements.

3. Results

3.1. Spatial–Temporal Change of ESV

According to Formulas (1) and (2), the average value of the equivalent factor of the criterion ESV in 2010, 2015, and 2020 was 636.54 CNY/hm², which was used as the standard equivalent value during the study period. The changes in the ESV of different land types in Peixian from 2010 to 2020 are shown in

Table 5. Changes in landscape type and ESV in Peixian.

Years	Statistic Type	Landscape Type					Total
		Cultivated Land	Forestland	Water Area	Built-Up Land	Bare Land
2010	Area (km2)	904.51	25.81	423.11	331.82	120.53	1805.78
	ESV (108)	2.31	0.38	23.92	0	0.96	27.57
2015	Area (km2)	904.66	28.98	428.49	302.80	140.85	1805.78
	ESV (108)	2.31	0.42	24.22	0	1.12	28.07
2020	Area (km2)	908.13	28.51	407.82	322.49	138.83	1805.78
	ESV (108)	2.32	0.42	23.06	0	1.02	26.82
2010-2020	Area Change (%)	0.40	10.49	-3.61	-2.81	6.34	10.81
	ESV Change (%)	0.39	10.61	-3.59	0.00	6.37	13.78

. During the last ten years, the total ESV of Peixian was CNY 27.49 × 10⁸ on average, from CNY 27.56 × 10⁸ to CNY 26.81 × 10⁸, with a change rate of −13.61%. There were apparent differences in the total value contribution of system services. In terms of the ESV of each landscape type, the contribution rate of water area to the ESV was the highest, followed by cultivated land and bare land.

During the study period, coal mining resulted in the mutual conversion of various land use types in the study area. Cultivated land and water area were the most sensitive components of the ESV in the mining area. In addition, the total ESV of cultivated land increased by CNY 90 × 10⁴, which is related to the continuous land reclamation and ecological rehabilitation of subsided land during the past ten years in Peixian. The total value of the water area decreased by CNY 0.87 × 10⁸ due to land remediation, ecological restoration, and other projects, and the reduction in the water area has led to a 3.61% drop in the total value of water ecosystem services.

Figure 3 shows the spatial distribution of the ESV in Peixian during the three periods. The high value of ecosystem services in the study area was concentrated in the waters of Weishan Lake and some mining areas in the eastern part of Peixian. In contrast, the western and southern regions had a lower ESV and tended to be scattered. On the one hand, coal mining activities in the northern part of Peixian have formed subsided land and waterlogged areas. With land reclamation and ecological restoration in Peixian, the total ESV has increased to a certain extent. On the other hand, under the dual influence of coal mining subsidence and land consolidation, the ESVs of mining areas such as Zhang Shuangqiao, and Kongzhuang changed dramatically. The ESV of the surrounding waters presents a relatively unstable situation.

3.2. Spatial–Temporal Change of ER

Using Formula (4), the risk index and proportion of different landscapes in 2010, 2015, and 2020 were calculated (

Table 6. Changes and contributions of the ER in Peixian.

Years	Statistic Type	Land Use Type
		Cultivated Land	Forestland	Water Area	Built-Up Land	Bare Land
2010	ER	527.15	2.22	920.69	2067.66	171.91
	Proportion (%)	14.29	0.06	24.95	56.04	4.66
2015	ER	527.23	2.49	932.39	1886.85	200.85
	Proportion (%)	14.85	0.07	26.27	53.15	5.66
2020	ER	529.26	2.45	887.41	2009.51	182.81
	Proportion (%)	14.66	0.07	24.57	55.64	5.06
2010-2020	ER Variation	2.11	0.23	−33.28	−58.15	10.9
	Change Ratio (%)	0.40	10.36	−3.61	−2.81	6.34

). The total ER values were 3689.62, 3549.82, and 3611.44 in 2010, 2015, and 2020, respectively; the risk index decreased by 78.18. The ER of built-up land was the highest, with an average proportion of 54.94% in the total risk index, while the ER of forestland was the lowest, with a small comprehensive fluctuation range and an average change rate of 8.13%. The water and cultivated land risk levels fluctuated, with the former slightly decreasing and the latter slightly increasing from 2010 to 2020.

Figure 4 presents the spatial distribution of the ER in Peixian. The regions with high ecological risks were mainly concentrated in the southwest of the study area, showing the distribution characteristics of overall contiguous and local agglomeration. The low ER was located primarily on the narrow and long zone in the eastern study area, which is consistent with the spatial distribution of Weishan Lake, and the variation fluctuation was slight. High ER areas were mainly distributed in the central and southern townships with a high proportion of construction land and higher population density. It is worth noting that the levels of ER in the mining areas and surrounding areas were not high, indicating that the restoration of subsidence and polluted land effectively improved the negative impact of coal mining and relevant production activities. Weishan Lake has become a stable and improved ER zone. With the acceleration of northern Peixian integration and new urbanization construction, built-up land characterized by low ER has increased, and high ER areas have gradually decreased, becoming fragmented.

3.3. Multivariate Correlation Analysis of ESV and ER

3.3.1. Grey Relational Analysis of ESV and ER

Based on grid analysis, the study selected provisioning services, supporting services, regulating services, cultural services, and comprehensive ecosystem services [49] to evaluate the ESV with ER for grey relational analysis. The calculation results of the grey comprehensive correlation degree revealed the relationship between the absolute quantity and the rate of change characteristics of the ESV and ER (

Table 7. Grey relational degree between ESV and ER.

Years	θ Value	Value of Services Category (Unit Area Values)
		Total ESV	Provisioning Services	Regulating Services	Support Services	Cultural Services
2010	0.3	0.655	0.724	0.517	0.714	0.769
	0.5	0.735	0.752	0.584	0.746	0.781
	0.7	0.782	0.765	0.623	0.796	0.786
2015	0.3	0.645	0.722	0.508	0.712	0.767
	0.5	0.727	0.751	0.577	0.744	0.798
	0.7	0.775	0.764	0.617	0.759	0.785
2020	0.3	0.656	0.724	0.518	0.714	0.768
	0.5	0.735	0.752	0.583	0.745	0.780
	0.7	0.781	0.765	0.621	0.761	0.786

). (1) Cultural services had the highest comprehensive correlation with the ER among the four functional values, indicating that cultural services were the closest to the ER regarding the geometric characteristics and rate of change. (2) In 2015, the comprehensive correlation between the regulating services and the ER was the lowest, indicating that the ecological system regulation function had little correlation with the ER in the study area. (3) When increased by order, the increase rate of its comprehensive correlation decreased, indicating that the change rates of the ESV and ER were more synchronous, and the response sensitivity was more similar when the ESV and ER were disturbed.

3.3.2. Analysis of Spatial–Temporal Heterogeneity

The spatial–temporal heterogeneity indices of the study area in 2010, 2015, and 2020 respectively, were 0.020, 0.017, and 0.010, showing a downward trend as a whole. Figure 5 shows the spatial pattern based on the spatial interpolation method. The high heterogeneity area was mainly located at Weishan Lake and its surrounding zone in eastern Peixian, with typical characteristics of a high ESV and low ER in the contiguous and extended regions, which require attention for ecological protection. The proportion of areas with a low heterogeneity index shows a decreasing trend, mainly distributed in the northern mining area and the surrounding south areas in Peixian. Due to the frequent disturbance of human activities and the apparent impact of underground resource exploitation, a wide-ranging, small-piece distribution feature is needed to implement ecological restoration. In addition, the northern mining area in Peixian County, as a mining area governance pilot project in Jiangsu, shows a distinct ESV-ER coordination zone because of the governance and ecological restoration measures of coal mining subsidence in recent years.

3.3.3. Disequilibrium and Inconsistency Analysis

According to Equations (8)–(10), the ESV disequilibrium index ($U_{ESV-S}$), ER disequilibrium index ($U_{ER-S}$), and ESV-ER disequilibrium index ($U_{ESV-ER}$) in 2010, 2015, and 2020 are shown in

Table 8. Disequilibrium index between ESV and ER.

Type	2010	2015	2020	2010-2015	2015-2020	2010-2020
				Change Rate (%)	Change Rate (%)	Change Rate (%)
ESV disequilibrium index	2.90	2.98	3.19	2.67	6.85	9.70
ER disequilibrium index	1.34	1.32	1.30	−1.05	−1.84	−2.88
ESV-ER disequilibrium index	4.27	4.35	4.15	1.77	−4.50	−2.80

. From 2010 to 2020, the $U_{ESV-S}$ values in the study area were 2.90, 2.98, and 3.19, respectively, showing a continuously rising trend and indicating that the agglomeration of ecosystem service functional areas gradually strengthened and the uniformity weakened. The total $U_{ER-S}$ value decreased slightly by 2.88%, and the concentrated distribution of risks weakened, indicating that the integrated ecosystem stability improved. Due to this, Peixian has formed high-ESV areas, such as Anguo Lake Wetland and Hongfu Wetland, through the comprehensive management of coal mining subsidence, promoting ecological environment protection along Weishan Lake. At the same time, the overlapped areas of cropland and coal resources with landscape fragmentation have gradually formed a compound functional area characterized by stable “production–living–ecological” functions and a complete landscape structure.

The inconsistent spatial distribution of Peixian in 2010, 2015, and 2020 is shown in Figure 6. From 2010 to 2020, the areas with high ESV agglomeration were concentrated along Weishan Lake. Due to adherence to the ecological concept of “ecological, healthy, and high-quality”, the rectification actions for arbitrary occupation, construction, and arrangement reduced artificial destruction to the environment, finally forming a high and stable ESV area. The high ER agglomeration areas were distributed in the southwest and hinterland in the study area. This phenomenon occurred since the landscapes in the central and southern parts of Peixian are primarily agricultural land and built-up land, where the ER was easily increased and concentrated because of human intervention. Eventually, this promoted a significantly higher degree of agglomeration of the ESV than the ER. The subsidence pits formed by mining in the high water-level mining areas were prone to developing wetlands and other waters with high ESV, such as Yaoqiao Mine and Xuzhuang Mine in the north. In contrast, the ESV agglomeration was minor around the mines with low annual production in Pei and Long.

4. Discussion

4.1. Reasonableness of Revision of ESV and ER

Many studies have been carried out on the concepts [50], service types [51], and method systems [52] of ESV and ER. These studies have initiated a research hotspot regarding the theory of ecological services. Subsequently, Chinese scholars [9] developed a unit area value-equivalent factor suitable for evaluating China’s terrestrial ecosystem service value based on Constanze’s research, which promoted the study of ecosystem services. However, this method is more suitable for assessing the service value of various ecosystems at regional and large scales [53]. The value-equivalent factor method of per unit area is unsuitable for specific study areas, such as mining areas that are strongly disturbed by humans. Due to the spatial heterogeneity of services within each ecosystem, there may be a certain deviation in the conclusion of the assessed value. The differential impact of mining areas with different mining intensities on the ecological functions of the surrounding landscape [32].

Peixian is a typical high-water-level overlapped cropland and coal production area. Compared to other regions, the natural background characteristics and industrial structure in Peixian form a complex and comprehensive ecosystem. Long-term, large-scale, and high-intensity coal mining inevitably occupies and destroys a large amount of land and further damages and pollutes the original ecological landscapes, such as farmland and water areas [54].

For the uniqueness of the study area, the study set a correction factor to revise the results of the ESV based on the equivalent value method within the mining industry. It aimed to reduce the evaluation deviation of the ESV caused by mining activities in the mining area, which, to a certain extent, is a rich expansion of the existing ESV studies. Similarly, mining also objectively affects the ecological vulnerability of the landscape. Therefore, the study adopted the expert consultation method to correct the vulnerability index of the study area and then obtained the ER. The results show that at the early stage, due to ground subsidence, some mining areas of Peixian, such as the Zhang Shuangqiao and Kongzhuang coalfields, had a large number of stagnant water areas. The ecosystem service value was high (Figure 3). After the land reclamation and ecological rehabilitation, the mining area was mainly transformed into wetlands and construction land, forming relatively stable landscape types. The ER shows a trend of fluctuation and decline (Table 6), revealing that the overall ecological status in Peixian has tended to be stable and indicating that Peixian has played an active role in ecological management and restoration projects in mining areas in recent years.

4.2. Multivariate Correlation between ESV and ER

ESV and ER assessments are often used to assess the ecological environment of regional socio-economic–ecological systems. A single ER or ESV cannot fully or profoundly explain the changes in an ecological environment. The effective combination of both is of great significance for comprehensively considering ecological and environmental effects and improving human well-being. Related studies have mostly used the global autocorrelation model and grey correlation model [55,56] and have mainly focused on the degree of spatial correlation and the levels and composition ratio of the ESV and ER [57].

This study selected the grey relational degree model, spatial–temporal heterogeneity model, and disequilibrium and inconsistency indices to construct a multivariate correlation analysis framework to evaluate the ecological quality of the study area based on the order of magnitude, spatial configuration, and trend surface. The grey relational degree model focuses on the correlation between the absolute quantity and the rates of change of the ESV and ER for quantitative analysis. The spatial–temporal heterogeneity model aims to present the spatial heterogeneity characteristics of the ecosystem structure, process, and function after disturbance. The disequilibrium and inconsistency models focus on analyzing the coordination characteristics and agglomeration trends of the ESV and ER from the index evaluation and spatial representation, respectively, to guide the critical areas of environmental protection or ecological restoration.

The study shows that the waterlogging area formed by the land subsidence had a high degree of value agglomeration, which has improved the site’s ecological function to a certain extent. This is related to the local government’s concentrated development of aquaculture and other industrial policies. Only by clarifying the formation mechanism of the spatial identification of the ESV and ER and scientifically assessing changes in the ecological environment caused by human activities can we provide decision-making management references for environmental protection and high-quality, sustainable development.

4.3. Deficiencies and Prospects

Although necessary revisions have been made, they may only partially reflect the complex ecosystem’s actual situation since the analysis was based on land use types in the study. Because Peixian is located on the eastern plain and has a typical natural climate and economic and social environment, many factors affected the final ESV. Therefore, a more scientific and reasonable ESV evaluation model close to the region’s actual characteristics needs to be developed in the follow-up study. At the same time, the study is a case study evaluating the ESV of Peixian based on the equivalent factor method. Whether the revised index system is universal in other regions needs to be further verified. The essence of the study on the temporal and spatial correlation between ESV and ER was to analyze the interaction, but the internal mechanism has not been well explained and needs to be further investigated.

5. Conclusions

This study focused on the high water-level overlapped areas of cropland and coal resources in the eastern plains of China. Based on the correction of the ESV coefficient and ER assessment, a multivariate correlation analysis was carried out on the ESV and ER in the study area. From 2010 to 2020, the ESV showed a general feature of high ESV in the east and low ESV in the west, with a trend of increasing and then decreasing, indicating that the ecological environment in Peixian has been dynamic. The overall ER improved, meaning that the contradiction between human activities and the ecological environment was eased. According to the calculation of the grey relational degree, the ecosystem cultural services had the most significant impact on the regional ER. The spatial and temporal differences in the ESV and ER show a continuous slow decline trend from 2010 to 2020, revealing that the overall ecological status of Peixian has tended to be stable. According to the results of the disequilibrium and inconsistency indices, it was found that the ESV agglomeration was also distributed in the northern mining area. Although the location of the saturated site caused by coal mining subsidence was relatively small, the ESV agglomeration was high. ER agglomeration existed in the central and western regions with somewhat scattered landscapes. It should be noted that in the process of protection and restoration, the ER and ESV of Peixian and the control of natural ecological space use should be combined. When focusing on economic development, according to the spatial pattern of mining and farming, land use should be reasonably arranged by the type of area. Furthermore, the protection of ecological land with high ESV per unit area, such as Weishan Lake, should be strengthened in Peixian. Finally, the reclamation and ecological reconstruction of coal mining subsidence should be promoted to effectively increase the area of ecological land and realize the coordinated development of social economy and ecology.

From the perspective of the ecological environment, this study deeply analyzed the extraordinary impact of mineral resource development on ESV and ER. Evaluating ecological status from the perspective of multiple correlations provides a theoretical reference for alleviating land use conflicts and realizing the comprehensive and coordinated development of society, economy, and ecology in the overlapped areas of cropland and coal resources.