Analysis of Spatiotemporal Evolution Characteristics and Influencing Factors of Vegetation Net Primary Productivity in Shendong Mining Area

Abstract

Coal plays a key role in the composition of energy production and consumption in China. The mineral is a crucial resource for human economic and social development. The surface subsidence caused by coal mining has attracted disturbance and damage to the surface ecological environments of mining areas. To quantitatively study vegetation carbon sequestration in coal mining areas, this article takes the Shendong mining area as its research object and analyzes the spatiotemporal distribution characteristics and influencing factors of vegetation net primary productivity (NPP). Based on meteorological data, remote sensing images, and land use data in the mining area, this research analyzed the vegetation NPP and its relationship with meteorological factors in the Shendong mining area from 2000 to 2023. The results show that (1) in the past 24 years, the range of NPP variation in the Shendong mining area has been 0–350 gC/m². The overall NPP shows a fluctuating upward trend. The annual average change in NPP in the mining area is bounded by the two central dividing lines, with a lower value in the central part and a gradually increasing trend in the western and eastern parts. (2) The average vegetation NPP of the Shendong mining area over 24 years was 0–250 gC/m²; after 2011, the annual average NPP of vegetation was higher than this average level, which shows that the ecological environment of the mining area is gradually improving. Coal mining in the mining area has not caused a significant impact on the overall environment of the mining area. (3) There is a significant relationship between NPP and annual cumulative rainfall, average annual temperature, and average annual Normalized Difference Vegetation Index (NDVI). Overall, it is not significantly related to mining factors. The mining factors affect local areas and have a certain correlation.

1. Introduction

Overall, 70–80% of China’s total carbon emissions come from coal production and consumption [1], which determines that coal will still occupy the leading position of energy in China’s future sustainable economic and social development. The concept of sustainable development was announced after the publication of the Brundtland Commission Report by the World Commission on Environment and Development [2]. Sustainable development means “development that meets the needs of the present without compromising the ability of future generations to meet their own needs”. The notion of economic sustainability, more specifically, focuses on market opportunities with regard to the welfare level that each generation passes on to the next one [3]. Sustainable development research should include a comprehensive historical background and a broader explanatory framework. More theories have shown a clear stable trend, gradually shifting their attention comprehensively to the context or developmental background. Smith, Mill, and Marx’s theory of developmental stages provides a powerful theoretical framework that can integrate multiple, diverse, and medium-sized contemporary clues in development research. Classical political economy provides an analytical foundation for certain elements, including the important role of history, the necessity of interdisciplinary methods, and the priority of analyzing social classes [4]. A major theory in development research, namely John Stuart Mill’s Grand Stage Theory, provides the necessary perspective for development research to fully understand the origins of contemporary events. Mueller’s theory regards the dialectical relationship between knowledge/innovation (human capital) and nature (natural capital) as an accelerator of economic change. In Mueller’s political economy, innovation and nature play a crucial role in achieving sustainable economic development. Contemporary society and development issues require the adoption of alternative research approaches in development studies. There is an urgent need to revive grand theories that can encompass multiple intermediate-range theories, provide greater historical depth, and offer necessary interdisciplinary perspectives to better understand and articulate complex and interconnected development dynamics. In this regard, we attempt to demonstrate that Mill’s theory of stages provides a dynamic and grand theory that can be expanded and enriched through contact with mid-term theories that address empirical challenges [5].

Coal enterprises have their own unique characteristics in development. Therefore, coal enterprises should pay more attention to implementing sustainable development strategies to prolong their development. After a long period of reckless and extensive development, coal enterprises have developed many problems that seriously affect their subsequent development. These are all challenges that coal enterprises need to face in implementing sustainable development strategies [6].

Coal resource mining not only creates and drives urban economic and social development, but also has varying degrees of impact on the ecological environments of mining areas. Many scholars have conducted research on coal environmental pollution, including wastewater pollution, air pollution, soil pollution, and so on [7]. The water pollution caused by coal mining activities is the main source of water pollution worldwide [8]. Untreated mine water discharged into surface rivers will cause serious water pollution [9]. High-intensity coal mine excavation can also cause serious drops in groundwater levels, resulting in a fragile ecological environment in some areas and severe damage [10]. In addition to the environmental impact factors mentioned above, the underground mining of coal can also lead to ground subsidence and the formation of subsidence areas, affecting the growth of surface vegetation and even causing a loss of land resources in mining areas, leading to an imbalance in the ecosystem [11].

To solve the problem of ecological restoration projects in mines left over from history, the Chinese government has made consolidating and improving carbon sequestration ability a key task [12]. Based on the special environment of mining areas, calculating vegetation NPP and analyzing its temporal and spatial evolution characteristics and influencing factors has a guiding role in understanding the impact of coal mining in western mining areas on the environment, as well as land reclamation and environmental governance in mining areas.

At present, most Chinese scholars are focusing on analyzing the ecological environments of surface mines and western mining areas. Most scholars use the improved CASA (Carnegie–Ames–Stanford Approach) model and RSEI (Remote Sensing Ecological Index) model as research methods to measure the ecological environments of mining areas. Qin Lu et al. used the RSEI to analyze the spatiotemporal evolution of the ecological environment in the Jiawang Mine in Xuzhou. They concluded that the overall environmental quality of the mine area steadily improved from 2009 to 2021. When Jiawang Mine was closed, the ecology of the Jiawang Mine area improved after long-term ecological restoration [13]. Li Zhixiang et al. used the MOD17A3H dataset to analyze the spatiotemporal changes in vegetation NPP in six major coalfields in Shanxi Province, and the impact of land cover types on vegetation NPP. The results showed that from 2005 to 2015, the vegetation NPP of the six major coalfields in Shanxi Province showed a fluctuating growth trend, and the increase in forest and cultivated land area led to a significant increase in vegetation NPP [14]. Fan Deqin et al. added vegetation NPP as a new indicator to the RSEI model and summarized the patterns of environmental changes in the Shenfu mining area from 2000 to 2016. Their conclusion shows that with coal mining, the ecological quality of the mining area has steadily improved and recovered continuously [15]. Sun Yinglong et al. calculated vegetation NPP by using the carbon flux TEC model. By analyzing the spatial variation in NPP in Lincang City from 2000 to 2018 and its relationship with meteorological factors, this study displayed the characteristics of sustained growth of NPP in Lincang City, and the most significant correlation with rainfall [16]. Zhang Wenkai et al. used an improved CASA model to calculate the distribution characteristics of vegetation NPP in the Shendong mining area from 2005 to 2015. This research has shown that the impact of mining levels at different scales and intensities on vegetation NPP in the Shendong mining area is not significant [17]. There are also studies emphasizing that inoculation with mycorrhizal fungi can increase plant biomass and effectively improve rhizosphere soil quality, targeting the surface environment of subsidence areas in the Shendong Coalfield and Mu Us desert mining subsidence areas [18,19,20,21]. The existing research has the following problems: (1) most literature focuses on the overall mining area as the research scale, which makes it difficult to analyze the impact of different coal mines; (2) regarding the specific impact of mining factors on vegetation NPP, the existing literature is not sufficiently combined with coal mining conditions; and (3) the methods used for analysis of the relationship between NPP and its influencing factors mainly consider the spatiotemporal distribution level, and further analysis is needed to study the spatial heterogeneity between variables.

Therefore, this article analyzes the spatiotemporal distribution and influencing fac-tors of vegetation NPP from three scales: mining area, coal mine, and working face. By combining remote sensing data and spatial econometric models, this article analyzed the spatiotemporal variation characteristics of vegetation NPP in 11 coal mines in the Shendong mining area over 24 years under high-intensity mining. This article further analyzed the relationship between meteorological factors and vegetation NPP in the study area, aiming to make up for the shortcomings of earlier research on ecological environment measurement in mining areas and supplement the time–space characteristics of NPP in mining areas based on long-term time series at a small scale.

In addition, this article combines spatial econometric methods to analyze the spatial distribution patterns of NPP in various mines, providing a reference basis for the evaluation of ecological environment quality in mining areas. The research results of this article contribute to providing more suitable methods for improving the ecosystems of mining areas, which also have reference significance for vegetation NPP research in mining areas with similar terrain and landforms.

2. Materials and Methods

2.1. General Situation of Shendong Mine

The Shendong mining area is located in Ordos City, Inner Mongolia Autonomous Region and Yulin City, Shaanxi Province. The mining area has abundant resources, stable coal seam occurrence, shallow burial, thick coal seams, excellent coal quality, and simple geological conditions. The climate in the mining area features strong sandstorms in spring, short hot summers, cool and rainy autumns, and long cold winters. The annual rainfall is low, the evaporation is high, and the frost period is long, belonging to a dry semi-desert plateau continental climate. The soil types in the mining area include chestnut soil, aeolian sand soil, and tidal soil. Due to the influence of the Mu Us desert, most of the surface of the mining area is covered by widely distributed aeolian sand. The mining area belongs to the desert grassland sub zone. Due to the terrain, topography, climate, and unique soil-forming conditions of the Ordos Plateau, the main plants are arid grasslands and decertified grasslands dominated by sandy and arid semi-shrubs, with sparse vegetation and a forest grass vegetation coverage of 23%. See Figure 1.

2.2. Datasets

The meteorological data of the research area include sunshine hours, precipitation, and average temperature, obtained from the National Earth System Science Data Center, National Science & Technology Infrastructure of China. Based on commonly used processing methods, we used the Kriging method to interpolate meteorological data; finally, ArcGIS was used to mask and resolve the study area, resulting in monthly meteorological grid data from 2000 to 2023.

The vegetation type map was sourced from the Geographic Science Data Network. Firstly, we performed grid projection on the vegetation type map to ensure that the spatial range and projection type of the output image were consistent with the mask file of the research area; secondly, mask extraction was performed on the data while eliminating the background values of the file; and finally, we converted the data into a file format consistent with meteorological data.

The NDVI of the study area was sourced from the geographic data cloud platform. This article selected Landsat TM products from each month of the study area from 2000 to 2023, with a spatial resolution of 30 m. Firstly, it was necessary to perform radiation calibration and atmospheric correction on the NDVI, calculate the NDVI using formulas, unify the scope with the research area, and finally unify the file format with meteorological data and vegetation type maps.

The remote sensing data consisted of TM image data from 2000 to 2023 and MOD17A3HV6 data products. This product provides information on annual NPP with a resolution of 500 m pixels and requires masking the data to eliminate background values in the file, finally converting it into a file format consistent with meteorological data.

2.3. Trend Analysis of NPP

To study the temporal trends in the net primary productivity of vegetation, the trend analysis method based on univariate linear regression was used to analyze the interannual NPP from 2000 to 2023 [22]. The calculation formula is as follows:

$${\mathrm{K}}_{\mathrm{s}\mathrm{l}\mathrm{o}\mathrm{p}\mathrm{e}}={\displaystyle \frac{\mathrm{n}\times \sum _{\mathrm{i}=1}^{\mathrm{n}}(\mathrm{i}\times {\mathrm{N}\mathrm{P}\mathrm{P}}_{\mathrm{i}})-\sum _{\mathrm{i}=1}^{\mathrm{n}}\mathrm{i}\times \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{N}\mathrm{P}\mathrm{P}}_{\mathrm{i}}}{\mathrm{n}\sum _{\mathrm{i}=1}^{\mathrm{n}}\mathrm{i}2-\left(\sum _{\mathrm{i}=1}^{\mathrm{n}}\mathrm{i}\right)2}}$$

(1)

In this formula, K_slpoe is the linear slope, NPP_i is the annual total net primary productivity in the ith year, and n is 24. When K_slpoe > 0, it means that the NPP is increasing over time; when K_slpoe < 0, the NPP is decreasing.

2.4. Correlation Analysis of NPP and Meteorological Factor Spatial Correlation Testing Model

Based on the correlation coefficient between the net primary productivity of vegetation and meteorological factors, the formula is [23]:

$${\mathsf{\rho }}_{\mathrm{x}\mathrm{y}}={\displaystyle \frac{\sum _{\mathrm{i}=1}^{\mathrm{n}}\left[\left({\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}}\right)\left({\mathrm{y}}_{\mathrm{i}}-\overline{\mathrm{y}}\right)\right]}{\sqrt{\sum _{\mathrm{i}=1}^{\mathrm{n}}\left({\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}}\right)2\sum _{\mathrm{i}=1}^{\mathrm{n}}\left({\mathrm{y}}_{\mathrm{i}}-\overline{\mathrm{y}}\right)2}}}$$

(2)

In the above formula, ${\mathsf{\rho }}_{\mathrm{x}\mathrm{y}}$ is the correlation coefficient of the two variables; n is 22; x_i and y_i are the NPP value and the meteorological value of the two variables in the ith year; and $\overline{\mathrm{x}}$ and $\overline{\mathrm{y}}$ are the mean values of the two, respectively.

2.5. Spatial Correlation Testing Model

Based on global spatial autocorrelation to reflect the distribution characteristics of in-terannual changes in vegetation NPP, to determine whether vegetation NPP changes in the study area are clustered at the spatial scale, a commonly used testing method is the global Moran’s index [24], and the calculation formula is as follows:

$$\mathrm{M}oran’sI={\displaystyle \frac{n{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{W}_{ij}(x_i-\stackrel{-}{x})(x_j-\stackrel{-}{x})}}}{({\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{W}_{ij}}}){\displaystyle \sum _{i=1}^n{(x_i-\stackrel{-}{x})}^2}}}$$

(3)

In this formula, n is the number of coal mines; x_i and x_j represent the vegetation NPP of coal mine i and coal mine j; W_ij is the spatial weight; and $\overline{\mathrm{x}}$ is the average value of x. When the global Moran’s index is greater than 0, it indicates a positive spatial correlation, which means a significant level of vegetation NPP aggregation in coal mining areas; when the global Moran’s index is less than 0, there are significant differences in the spatial distribution of vegetation NPP in the study area, with uniform or diffuse distribution characteristics; and when the index is 0, there is no obvious distribution pattern of vegetation NPP in each coal mine, and it has a random distribution characteristic.

2.6. Surface Movement Calculation Method

This article used the probability integration method to calculate surface movement and deformation, and the calculation formula is as follows [25]:

$$\begin{array}{c}\mathrm{S}\mathrm{u}\mathrm{b}\mathrm{s}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{e}\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}W(x,y)=W_{cm}\cdot {\iint }_D{\displaystyle \frac{1}{r^2}}\cdot \mathit{exp}(-\pi {\displaystyle \frac{(\eta -x{)}^2+(\xi -y{)}^2}{r^2}}d\eta d\xi )\hfill \end{array}$$

(4)

$$\begin{array}{c}\mathrm{I}\mathrm{n}\mathrm{c}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{d}\hfill \\ i_x(x,y)={\displaystyle \frac{\partial W(x,y)}{\partial x}}=W_{cm}\cdot {\iint }_D{\displaystyle \frac{2\pi (\eta -x)}{r^4}}\mathit{exp}(-\pi {\displaystyle \frac{(\eta -x{)}^2+(\xi -y{)}^2}{r^2}}d\eta d\xi )\hfill \\ i_y(x,y)={\displaystyle \frac{\partial W(x,y)}{\partial y}}=W_{cm}\cdot {\iint }_D{\displaystyle \frac{2\pi (\xi -y)}{r^4}}\mathit{exp}(-\pi {\displaystyle \frac{(\eta -x{)}^2+(\xi -y{)}^2}{r^2}}d\eta d\xi )\hfill \end{array}$$

(5)

$$\begin{array}{c}\mathrm{C}\mathrm{u}\mathrm{r}\mathrm{v}\mathrm{a}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e}\hfill \\ \hspace{1em}{k}_x(x,y)={\displaystyle \frac{\partial i(x,y)}{\partial x}}=W_{cm}{\iint }_D{\displaystyle \frac{2\pi }{r^4}}\left({\displaystyle \frac{2\pi (\eta -x{)}^2}{r^4}}-1\right)\mathit{exp}(-\pi {\displaystyle \frac{(\eta -x{)}^2+(\xi -y{)}^2}{r^2}}d\eta d\xi )\hfill \\ \hspace{1em}{k}_y(x,y)={\displaystyle \frac{\partial i(x,y)}{\partial y}}=W_{cm}{\iint }_D{\displaystyle \frac{2\pi }{r^4}}\left({\displaystyle \frac{2\pi (\xi -y{)}^2}{r^4}}-1\right)\mathit{exp}(-\pi {\displaystyle \frac{(\eta -x{)}^2+(\xi -y{)}^2}{r^2}}d\eta d\xi )\hfill \end{array}$$

(6)

In the formula: W_cm is the maximum surface subsidence value during full mining; U_cm is the maximum horizontal movement of the surface during full mining; r is the main impact radius; θ₀ mainly affects the propagation angle; D is the mining area (considering the inflection point offset); and x, y are the relative coordinates of the point.

2.7. Estimated Parameters and Calculation Schemes for Surface Movement

According to the analysis results of the measured data of surface movement in Shangwan Mine, the calculation parameters of surface movement are taken as shown in

Table 1. Probability integration method for predicting parameters.

Parameter	Subsidence Factor	Horizontal Movement Coefficient	Tangent of Major Influence Angle	Influence Propagation Angle Coefficient	Inflection Points Offset Coefficient
Symbols Value	Q 0.72	B 0.34	Tanβ 2.1	Kθ 0.6	Ks 0.21

[26].

The calculation plan was designed based on the coal mine production year and mining impact range. By calculating the surface movement of each working face in the fourth mining area of Shangwan Mine, the calculated results can be used for a correlation analysis between mining impact and surface vegetation NPP.

2.8. Residual Method

This article set up the linear regression model between NPP and other factors to calculate the NPP prediction, so that the difference between it and the actual NPP can be used to measure the increase or decrease in NPP caused by mining activities. The calculation formula is as follows [27]:

NPP_CM = NPP − NPP_M

(7)

The impact of natural factors on NPP is referred to as NPP_M. The vegetation NPP values in the fourth district were calculated with the CASA model. The impact of mining activities on NPP is defined as NPP_CM.

Based on the principle of least squares, this section used Origin software (OriginPro 8) to analyze the relationship between vegetation NPP and various influencing factors. When establishing the model, the limitations of different factor data in terms of units and amplitudes were eliminated, and the data were converted into dimensionless numbers (DNs). By fitting NPP with accumulated precipitation, annual average temperature, and NDVI, the calculation formula is as follows:

DN_NPP_M = β₀ + β₁ × P + β₂ × T + β₁ × NDVI

(8)

In the formula, β₀ is the constant term, and β₁, β₂, and β₃ are the coefficients of the precipitation, temperature, and NDVI.

2.9. Gaussian Fitting

The Gaussian fitting of the DN_NPP_CM and mining factors (FMs) is as follows:

$$DN\_NPPCM={\displaystyle \frac{A}{w\sqrt{\frac{\pi }{2}}}}{e}^{{\displaystyle \frac{-2{\left(FM-{FM}_c\right)}^2}{w^2}}}+FM0$$

(9)

In the above formula, FM₀ is the minimum value of the Gaussian function; A/w√(π/2) is the function peak; FM_C is its corresponding abscissa; and W is the standard deviation.

3. Results and Analysis

3.1. Analysis of Temporal Distribution Characteristics of NPP in Shendong Mining Area

By summarizing the average vegetation NPP data of the study area from 2000 to 2023, we selected NPP data from 11 coal mines in the Shendong mining area for 7 years to analyze. From Figure 2, it can be seen that since 2001, the trends in the vegetation NPP changes in the 11 coal mines have been consistent. The trends in vegetation NPP changes in each coal mine in the study area have shown a fluctuating upward trend, showing a “stepped” growth. The period from 2001 to 2008 was the first growth stage, followed by a significant downward trend. The period from 2011 to 2016 was the second growth stage, indicating a rapid increase in vegetation NPP levels in each coal mine. What is more, the NPPs of Buertai coal mine and Cuncaota Second coal mine reached their maximum values in 2016, reaching 263.2 gC/m² and 227.6 gC/m², respectively. The period from 2017 to 2019 was the third growth stage, and the NPPs of other coal mines reached their maximum values in 2019. The NPP level in Liuta coal mine has been the lowest over the years, while the vegetation NPP level in Shangwan coal mine is the highest. The lowest (87.2 gC/m²) and highest (280.7 gC/m²) NPP values appeared in Liuta coal mine in 2001 and Shangwan coal mine in 2019. Based on comprehensive data statistics and analysis, it can be seen that the vegetation NPP level in the Shendong mining area is gradually improving, the vegetation carbon sequestration ability is steadily increasing, and the ecological environment in the mining area and surrounding areas has been improved and continuously restored.

Based on a simple linear regression analysis of the changes in vegetation NPP pixels of 11 coal mines in the Shendong mining area over the past 24 years, the results were visualized using the natural break classification method. The NPP in the Shendong mining area showed an upward trend from 2000 to 2023 (Figure 3). The annual average change in NPP in the mining area is bounded by the two central dividing lines, with a lower value in the central part and a gradually increasing trend in the western and eastern parts.

3.2. Analysis of Spatial Distribution Characteristics of NPP in Shendong Mining Area

To correspond to the temporal variation characteristics of vegetation NPP, this section selects seven main change years, 2001, 2008, 2011, 2016, 2017, 2019, and 2021, as well as 24 years of annual average variation, for the spatial distribution analysis.

From the 24-year average annual distribution figure of vegetation in the research area, it can be seen that the NPP levels of vegetation in the 11 coal mines are concentrated at 150–200 gC/m² (see Figure 4). The maximum area accounts for 80% of the total area (Cuncaotaer coal mine), and the minimum is about 46% (Shangwan coal mine), mostly accounting for 50–60% of the total area. The vegetation NPP level in Liuta coal mine is concentrated at 100–150 gC/m², accounting for 87.5% of its area. The changing characteristics of NPP in each coal mine can be observed, which generally conform to the “three step” growth pattern. From the years with the lowest vegetation NPP levels in 2001, the NPP levels in Liuta coal mine, Ulanmulun coal mine, Shigetai coal mine, Halagou coal mine, and the eastern parts of Bulianta and Shangwan coal mines were the lowest. Combined with the spatial distribution in 2008 and 2016, it is shown that even though the overall NPP level in Shendong mining area is in the growth stage, the NPP level in Liuta coal mine is still in the low value area; although the overall level of vegetation NPP in the Ulanmulun and Shigetai coal mines significantly improved, there was still a small portion below the average level. Overall, the 2011 vegetation NPP level in the Shendong mining area was generally higher than the average NPP level of 24 years, indicating that the ecological environment of the mining area is gradually improving, and coal mining may not have had a significant impact on the overall environment of the mining area. The low NPP values of vegetation in the Shendong mining area are concentrated in the border area of each coal mine. The year 2008 can basically represent the average level of NPP in the study area from 2001 to 2021.

3.3. Analysis of Spatial Correlation Characteristics of Vegetation NPP in Shendong Mining Area

Table 2. Global Moran’s index statistics of NPP vegetation in Shendong mining area probability integration method for predicting parameters.

Year	Moran’s I	Z	p	Year	Moran’s I	Z	p
2000	−0.238	−1.615	0.053 *	2011	−0.283	−1.945	0.026 **
2001	−0.196	−1.006	0.157	2012	−0.301	−2.204	0.014 **
2002	−0.265	−2.042	0.021 **	2013	−0.249	−1.728	0.042 **
2003	−0.248	−1.751	0.040 **	2014	−0.263	−1.836	0.033 **
2004	−0.217	−1.487	0.068 *	2015	−0.264	−1.870	0.031 **
2005	−0.233	−1.488	0.068 *	2016	−0.237	−1.566	0.059 *
2006	−0.228	−1.476	0.070 *	2017	−0.264	−1.794	0.036 **
2007	−0.230	−1.510	0.066 *	2018	−0.245	−1.681	0.046 **
2008	−0.235	−1.548	0.061 *	2019	−0.288	−2.146	0.016 **
2009	−0.216	−1.750	0.040 **	2020	−0.185	−0.980	0.163
2010	−0.297	−2.171	0.015 *	2021	−0.227	−1.430	0.076 *

Explanatory note: *, **, represent significance levels of 10% and 5% respectively.

summarizes the Moran’s I, Z, and p values of 11 mines in the Shendong mining area from 2000 to 2021. The Moran’s I data from previous years passed the test, except for data from 2001 and 2020. The average value is −0.25 ∈ (−0.301~−0.185), which indicates a strong spatial negative correlation of vegetation NPP in each mine. The vegetation NPP in the Shendong mining area has diffusion distribution characteristics in space and is not randomly distributed. The high-vegetation-NPP coal mines are clustered with the surrounding low-vegetation-NPP coal mines, and the low-vegetation-NPP coal mines are surrounded by high-vegetation-NPP coal mines.

The local Moran scatter plot can be divided into four quadrants: the first quadrant (HH), the second quadrant (LH), the third quadrant (LL), and the fourth quadrant (HL). The first and third quadrants show that the vegetation NPP in the Shendong mining area is positively autocorrelated in space, while the second and fourth quadrants are negatively autocorrelated in space.

Figure 5 shows the regional distribution of vegetation NPP in the local Moran’s I of the Shendong mining area in 2001 and 2019. The research results show the following: (1) In 2001, the vegetation NPP in each coal mine mainly exhibited LH- and HL-type clustering, with Liuta coal mine, Halagou coal mine, Huojitu coal mine, Shigetai mine, and Ulanmulun coal mine falling into the LH-type area, mainly located in the eastern mining area of the boundary. Shangwan coal mine, Cuncaota coal mine, Cuncaotaer coal mine, and Burtai coal mine fall into the HL-type area, mainly in the western mining area of the boundary. (2) In 2019, Liuta coal mine, Shigetai coal mine, Ulanmulun coal mine, Cuncaota coal mine, and Cuncaotaer coal mine fell into LH-type areas, including areas centered around Liuta coal mine and Ulanmulun coal mine; all remaining coal mines fell into the HL-type area, mainly including the areas centered around the Bulianta and Shangwan coal mines. (3) By comparing the spatial agglomeration situations in 2001 and 2019, it can be found that the Cuncaota coal mine and Cuncaotaer coal mine have shifted from HL- to LH-type areas, while Halagou coal mine and Huojitu coal mine have shifted from LH- to HL-type areas. The high-concentration coal mines remain unchanged, while the number of low-value-concentration coal mines are increasing, showing an expanding trend.

Figure 5. Local Moran scatter map of vegetation NPP in Shendong mining area in 2001 and 2019. The numbers are explained in

Table 3. Corresponding serial numbers.

ID	Coal Mine Name	ID	Coal Mine Name
1	Shigetai	7	Cuncaota
2	Huojitu	8	Liuta
3	Halagou	9	Shangwan
4	Daliuta	10	Wulanmulun
5	Buertai	11	Bulianta
6	Cuncaotaer

.

Figure 5. Local Moran scatter map of vegetation NPP in Shendong mining area in 2001 and 2019. The numbers are explained in Table 3.

3.4. Analysis of Spatial Correlation Characteristics of Vegetation NPP in Shendong Mining Area

To delve into the impact of different categories of influencing factors on vegetation NPP, this section adopts a multi-scale analysis method of “region, mining area, coal mine, and working face” to sequentially analyze the correlation between meteorological, NDVI, and mining factors and NPP, effectively revealing the impact of various factors on vegetation NPP in different ways and levels.

3.4.1. Analysis of Vegetation NPP and Meteorological Factors

Climate change has a direct impact on the type, structure, distribution, and function of vegetation [27]. This article calculates the correlation coefficient between vegetation NPP, temperature, and rainfall in Ejin Horo Banner and Shenmu City for 24 years at the pixel scale (Figure 6). The results show that there is a significant correlation between annual vegetation NPP changes and cumulative annual precipitation changes, with a maximum of 0.95, and a relatively weak correlation with annual average temperature changes.

The correlation coefficient between vegetation NPP and annual cumulative precipitation is −0.03–0.95, with a positive correlation area of 99.99%. Among them, the number of pixels related to moderate intensity accounts for 8%, the number of pixels related to strong intensity accounts for 38%, and the number of pixels related to extreme intensity accounts for 54%. The correlation coefficient between vegetation NPP and annual average temperature is −0.65–0.56, with a positive correlation area of 99%. The number of pixels with extremely weak correlation accounts for 32%, the number of pixels with weak correlation accounts for 65%, and the number of pixels with strong correlation accounts for 1%.

3.4.2. Correlation Analysis between Vegetation NPP and NDVI in Shendong Mining Area

Through an analysis of the correlation between NPP and NDVI in the mining area, it can be concluded that the changes in vegetation NPP and NDVI in the Shendong mining area over the past 24 years are positively correlated, accounting for 97% of the study area. This indicates that as the vegetation index increases, the level of NPP in the mining area also increases. Among them, the correlation coefficient between vegetation NPP and NDVI in the mining area is −0.84–0.98, with the minimum and maximum correlation coefficients seen in Daliuta coal mine and Bulianta coal mine, respectively.

According to Figure 7, the area with the highest proportion of negative correlations between NPP changes and NDVI changes is Liuta coal mine, and the area with the lowest proportion is Boltai coal mine. Among them, the central zone of the mining area is an area with a concentrated proportion of negative correlations, and the central zone towards the east and west to the edge of the mining area is a highly correlated area.

As shown in the

Table 4. Statistical table of the correlation strength between vegetation NPP and NDVI in Shendong mining area.

Positive Correlation Ratio	Non-Correlation	Weak Correlation	Moderately Strong Correlation	Strongly Correlated	Very Strongly Correlated	Total
Buertai	1%	3%	5%	26%	63%	98%
Ulanmulun	2%	2%	3%	16%	75%	96%
Halagou	2%	4%	8%	33%	52%	96%
Bulianta	2%	3%	5%	22%	66%	95%
Shangwan	2%	4%	7%	27%	58%	95%
Daliuta	3%	5%	9%	29%	50%	93%
Shigetai	3%	5%	10%	35%	43%	93%
Huojitu	4%	8%	17%	44%	23%	92%
Cuncaotaer	4%	6%	9%	25%	51%	91%
Cuncaota	9%	16%	19%	43%	6%	84%
Liuta	8%	8%	8%	12%	54%	82%

, the coal mines with the lowest and highest proportions of positive correlation strength are Liuta coal mine and Burtai coal mine, with 82% and 98%, respectively. The coal mines with the smallest and largest proportions of non-correlation are Buertai coal mine and Cuncaota coal mine, at 1% and 9%, respectively. The coal mines with the smallest and largest proportions of weak-to-medium-strength correlations are Ulanmulun coal mine and Huojitu coal mine, which account for 5% and 25%, respectively. The coal mines with the smallest and largest proportions of strong correlations are Liuta coal mine and Huojitu coal mine, with 16% and 44%, respectively. The coal mines with the smallest and largest very strongly correlated proportions are Cuncaota coal mine and Ulanmulun coal mine, with 6% and 75%, respectively.

Overall, it can be concluded that the changes in NPP in the Shendong mining area increase with an increase in NDVI, indicating that the ecological restoration project in the mining area has a promoting effect on vegetation NPP, and that the effect is significant, enhancing the ecological environment of the mining area.

3.4.3. Correlation Analysis between NPP and Mining Factors in Shangwan Coal Mine

(1)

The situation in the fourth district of Shangwan coal mine

In order to further analyze the relationship between vegetation NPP and mining fac-tors, the fourth district of the Shangwan coal mine in this section was taken as the research area (Figure 8). Taking 30 m × 30 m pixels as the minimum unit, we extracted the coordinates of the center points of each pixel. This section used mining subsidence analysis software (MSAS 1.0) to calculate surface movement deformation and draw contour maps and movement deformation curves.

The four working faces in the fourth district area were mined from April 2018 to 2024. As of December 2023, the mining of working faces 12401, 12402, and 12403 has been completed. The 12404 working face started mining in November 2023 and is planned to continue until the end of 2024. The square measure of the four working faces is 1.5552 mil-lion m², and the estimated coal output is 94.342 million tons. The average dip angle of the coal seams in the fourth mining area is 0°~2.2°, the mining thickness is 8.4 m~8.8 m, the mining depth is 130 m~254 m, and the coal seam bulk density is taken as 1.7 t/m³.

Based on the expected parameters obtained from the observation data of Shangwan coal mine, the surface movement of each working face in the fourth district was calculated for each year. The maximum surface subsidence value and subsidence area data obtained are shown in

Table 5. Annual mining data statistics of the fourth mining area of Shangwan mine.

Year	NPP/gC/m2	Average Mining Depth/m	Area of Mining/ha	Annual Coal Output/10 kt	Maximum Subsidence/mm	Area of Subsidence /ha
						Subsidence >10 mm	Subsidence > 500 mm
2018	125.5	200	81.4	1149.3	6007	152.75	121.50
2019	114.3	193	114.3	1700.5	6367	224.00	169.75
2020	131.4	168	125.3	1832.1	6191	218.25	165.25
2021	119.1	214.5	8.4	122.9	6191	39.75	18.25
2022	107.5	183	120.9	1726.7	6075	224.75	170.00
2023	124.6	192	57.1	834.8	6192	119.75	85.50

. The maximum annual surface subsidence value in the fourth district of Shangwan coal mine is 6007 mm~6367 mm, and the annual subsidence area with a subsidence value greater than 500 mm is 182,500~205,500 m². The cumulative subsidence area in the fourth mining area is 9.3575 million square meters. Figure 8 shows the annual distribution of surface subsidence in the fourth mining area of Shangwan coal mine.

(2)

Correlation analysis between NPP and mining factors in the fourth district

From the annual NPP_CM data and the corresponding annual area of mining, the change in NPP_CM showed a law of increasing–decreasing trend, so the Gauss function was used for fitting. Figure 9 shows the gross fitting curve between the annual NPP_CM value and mining area (MA) for 6 years, with the fitting relationship DN-NPP_CM (MA); R² is 0.65.

$$\mathrm{DN}\_NPP\mathrm{CM}\left(MA\right)=-0.4+1.1\times e^{-(MA-0.4{)}^2/0.6}$$

(10)

From the annual NPP_CM data and the corresponding annual coal output, the change in NPP_CM showed a law of increasing–decreasing trend, so the Gauss function was used for fitting. Figure 10 shows the gross fitting curve between the annual NPP_CM value and coal output (CO) for 6 years, with the fitting relationship DN-NPP_CM (CO); R² is 0.68.

$${\mathrm{DN}}_{NPP\mathrm{CM}\left(CO\right)}=-0.43+1.58\times e^{{\displaystyle \frac{-(CO-0.4{)}^2}{0.07}}}$$

(11)

From the annual NPP_CM data and the corresponding annual area of subsidence, the change in NPP_CM showed a law of increasing–decreasing trend, so the GaussAmp function was used for fitting. Figure 11 shows the gross fitting curve between the annual NPP_CM value and area of subsidence (SA) for 6 years, with the fitting relationship DN-NPP_CM (SA); R² is 0.7.

$${\mathrm{DN}}_{NPP\mathrm{CM}\left(SA\right)}=-0.77+1.78\times e^{{\displaystyle \frac{-(SA-0.5{)}^2}{0.16}}}$$

(12)

From the annual NPP_CM data and the corresponding maximum subsidence/average mining depth, the change in NPP_CM showed a law of decreasing–increasing trend, so the GaussAmp function was used for fitting. Figure 12 shows the gross fitting curve between the annual NPP_CM value and maximum subsidence/average mining depth (S_M/H) for 6 years, with the fitting relationship NPP_CM (K); R² is 0.7.

$$NPP\mathrm{CM}\left(SM/\mathrm{H}\right)=177-7\times e^{{\displaystyle \frac{-(K-33{)}^2}{0.03}}}$$

(13)

(1)

NPP_CM and multifactor analysis of mining

Considering the interdependence between coal mining factors, this article selects the ratio of maximum subsidence value to average mining depth (K) and subsidence area as the factors used to measure the mining factors, in which:

DN_NPP_CM = −14.9 + 0.94 × DN_NPP(SA) + 15.1 × NPP(S_M/H)

(14)

By importing Formulas (11) and (12), the result is:

$${\mathrm{DN}}_{NP{P}_{CM}}=-0.6+1.7\times e^{{\displaystyle \frac{-(SA-0.5{)}^2}{0.16}}}-0.58\times e^{{\displaystyle \frac{-(SM-33{)}^2}{0.003}}}$$

(15)

In conclusion, the correction coefficient for the non-linear relationship between the mining factor and NPP_CM based on the ratio of the maximum subsidence value to the average mining depth (S_M/H) and subsidence area is 0.9, indicating a strong fit. The result supports that mining activities have a significant impact on the remaining NPP after excluding meteorological impacts.

4. Discussion

This study used the BLOME-BGC Model to calculate Shendong coal mine’s NPP from the year 2000 to 2010. The NPP value and change trend are similar to those seen in [28]. Niu Hongbo et al. used the MOD17A3 database and analyzed the NPP trend from 2000 to 2016 [29]. The results showed that the average NPP values were mainly concentrated at 150–200 gC/m², which is the same as in this article. Considering the NPP results in Inner Mongolia, some articles also support the changing results of the NPP, which are similar to those found in this article [30,31,32]. Thus, the NPP in Inner Mongolia showed a fluctuating rising trend from 2000 to 2023.

In terms of the correlation results between the NPP and meteorological elements, scholars concluded that the annual vegetation NPP in the Shendong mining area is significantly positively correlated with precipitation, but not significantly correlated with annual average temperature [28].

With regard to the effect of the coal mining factor, Niu Hongbo et al. concluded that the growth of NPP vegetation in mining areas is disturbed by mining activities, and that the degraded mines are mainly distributed in high-intensity mining areas [29], which can support the results of this article. Xiang Yang et al. showed that the growth status of vegetation NPP in different mines in the Shendong mining area was deteriorated [33]. Both of these results are similar to the conclusion of this article.

What is more, increasing numbers of articles are focused on the lag between nature factors and NPP [33,34,35]. Therefore, we will use the improved CASA model and nature data to analyze the long-term NPP in the mining area and study the time lag between the NPP and each factor.

There are still some shortcomings in this article that need to be further improved: (1) The changes in the values of different parameters when calculating NPP may affect the research results. (2) Because of the NDVI data, this article only analyzes the working face NPP and mining factors of high-precision remote sensing images from 2018 to 2023. The impact time and scope of mining factors on NPP still need to be further researched. (3) The influencing factors of vegetation NPP do not only include environmental factors, but also factors such as vegetation characteristics and human activities, which affect the changes in vegetation NPP.

5. Conclusions

Based on meteorological and remote sensing data in the mining area, this article calculates the vegetation NPP of each coal mine in the Shendong mining area from 2000 to 2023. The evolution and distribution characteristics of NPP are analyzed from the temporal and spatial perspectives. This article analyzed the relationship between NPP and meteorological, nature, and mining factors in the scale of city region—mining area—coal mine—working face.

(1)

This article analyzed and summarized the annual changes in vegetation NPP levels in the Shendong mining area, which showed that the vegetation NPP levels in 11 coal mines were mainly concentrated at 150–200 gC/m², revealing the vegetation NPP in the Shendong mining area and showing a stepwise growth pattern of vegetation NPP (step one from 2001 to 2008, step two from 2011 to 2016, and step three from 2017 to 2019).

(2)

The vegetation NPP level in the Shendong mining area has been generally higher than the average NPP level over the 24 years studied since 2011, indicating that the ecological environment of the mining area is increasingly improving, and that coal mining may not have had a significant impact on the overall environment of the mining area; over the past 24 years, the area of vegetation NPP growth trends has far exceeded the area of decrease trends, and the spatial distribution of significant increases in vegetation NPP is the most widespread.

(3)

The vegetation NPP of each coal mine shows a strong negative spatial correlation, and the vegetation NPP of Shendong mining area has diffusion distribution characteristics in space, not random distribution. The high-vegetation-NPP coal mining area is clustered with the surrounding low-vegetation-NPP coal mining area, which is surrounded by the high-vegetation-NPP coal mining area.

(4)

This article analyzes the correlation of different influencing factors on different scales in a progressive manner in the following order: cities—mining areas—coal mines—working faces. The results show that the vegetation NPPs in Ejin Horo Banner and Shenmu City are positively correlated with annual cumulative precipitation and annual average temperature; the correlation between annual vegetation NPP changes and annual cumulative precipitation changes is relatively strong, while the correlation with annual average temperature changes is relatively weak. There is a positive correlation between the vegetation NPP and NDVI in the Shendong mining area, while the correlation between annual vegetation NPP changes and annual NDVI changes is significant.

(5)

This article calculates the vegetation NPP_CM in the fourth district of Shangwan coal mine, which is not affected by meteorological factors. By fitting the non-linear functional relationship between each mining factor and NPP_CM, the mining factors that affect NPP_CM are proposed, which means the ratio of the maximum subsidence value to the average mining depth (S_M/H) and subsidence area. Through the analysis of NPP_CM and multiple mining factors, the fitting relationship between NPP_CM and multiple mining factors was obtained, and the conclusion was drawn that mining activities have a significant impact on NPP_CM after excluding meteorological effects.