Novel Planning Strategies for Ecological Restoration of Abandoned Mines: A Case of Toli County, China

Abstract

With the development of mineral resources, the inevitable creation of numerous abandoned mines has impacted environmental resources. Numerous studies have been conducted on the restoration and management of individual abandoned mines. However, there has been no systematic study on the overall ecological restoration planning of abandoned mine clusters. Hence, there is an urgent need to research restoration planning strategies, focusing on the characteristics of abandoned mines and their environmental impacts. In this study, abandoned mines in Toli County, Xinjiang, were selected as the case. The K-means clustering analysis method was employed to study the spatial distribution of abandoned mines, selecting longitude, latitude, and road access as analytical factors. Based on spatial location attributes, three groups of abandoned mines were identified. The Analytic Hierarchy Process (AHP) was used to study the ecological importance evaluation model in Toli County, selecting eight evaluation factors including vegetation, precipitation, and population density, and dividing the ecological importance levels of various sectors to establish a three-stage restoration project. The Principal Component Analysis (PCA) method was used to assess the hazards of abandoned mines, selecting distance, land type, area, and ecological impact as influencing factors and determining the management sequence of abandoned mines within each project. The results show that (1) longitude, latitude, and road indices help to mitigate geographical obstacles such as mountains and rivers, ensuring a high degree of continuity in abandoned mine areas; (2) the AHP reveals that the combined weight of population density, gross domestic product, and vegetation index exceeds 80%, which are key factors affecting the priority of ecological restoration; and (3) the application of PCA provides a scientific basis for the hazard assessment and management of abandoned mines, prioritizing those close to densely populated areas and with larger areas. The significance of this study lies in providing a systematic method for ecological restoration planning of abandoned mines, as well as offering important references for future research and practice in related fields.

1. Introduction

China’s extensive mining history has resulted in a substantial number of abandoned mines. In the context of advancing ecological civilization, there is an imperative to enhance the management of these abandoned sites [1]. Projections indicate that by 2030, China will have approximately 15,000 abandoned mines, primarily non-metallic [2]. Despite this, the reclamation rate remains below 10%, leading to the degradation of over 3.61 million hectares of land resources. Furthermore, abandoned mines contribute to various geological and environmental challenges, including mine-related geological disasters and soil and water pollution [3,4]. Currently, China is intensifying its efforts to manage abandoned mines. However, ecological restoration efforts encounter significant obstacles due to the uneven spatial distribution of abandoned mines, intricate ecological issues [5], prolonged remediation timelines, and high associated costs [6]. These challenges are not unique to China, as international studies have similarly highlighted the difficulties in restoring ecosystems affected by mining activities [7,8]. Global research has focused on integrating both surface and subsurface restoration strategies, emphasizing the need for tailored approaches that account for regional environmental, socio-economic, and policy factors [9,10]. Consequently, there is an urgent need for research focused on developing effective ecological management and planning strategies [11] that can draw from both international best practices and the unique challenges faced by China’s mining regions.

To address the issue of ecological management planning for abandoned mines, various methods of governance and restoration of abandoned mines have been explored worldwide [12,13]. In response to the necessity of ecological restoration in mining areas, we studied the concept and current status of ecological restoration in mining areas, proposing a strategy and comprehensive technical system for ecological and environmental restoration in China’s mining areas [14]. Based on the necessity of restoration, we analyzed the effects of natural recovery and artificial intervention in land restoration in mining areas, identifying the main methods of land restoration in mining areas as pre-mining prevention and post-mining rehabilitation [15]. Regarding the ecological restoration in mining areas, we conducted a study on the ecological restoration strategies for abandoned mines, clarifying the overall approach and basic principles of ecological restoration of abandoned mines in China [16]. Meanwhile, foreign scholars have proposed the concept of using cemented paste to backfill and repair severely disturbed surfaces caused by complex iron ore mining [17]. Given the varying environmental conditions in the restoration process of abandoned mines [18], we proposed territorial spatial planning for mine ecological restoration, adopting a tailored approach to ecological restoration in different regions [19]. International research was also focused on the ecological restoration of individual or local abandoned mines. In particular, extensively studied the effect of surface vegetation on restoring mining areas to their pre-mining ecological state [20]. A comprehensive assessment framework for entire abandoned mine regions has been developed, exploring sustainable remediation strategies within the contexts of ecotourism and the circular economy [21]. Overall, restoration technologies for abandoned mines, whether on individual or small-scale regions, have demonstrated increasing maturity both domestically and internationally. In summary, the available technologies for restoring abandoned mines become increasingly mature [22]. Still, there is a lack of research on overall ecological restoration planning strategies for abandoned mines within particular regions and specific local conditions.

This study integrates the overall approach and requirements of ecological management planning for abandoned mines in China [23,24]. Utilizing such statistical analysis methods as cluster [25,26], hierarchical [27], and principal component analyses [28], An ecological restoration sequencing model for abandoned mines was developed through the collection of field data and the analysis of remote sensing data, it identifies feasible strategies for abandoned mine reclamation planning, facilitating the arrangement of a three-phase restoration project for abandoned mines in Toli County. Through the collection of field data and analysis of remote sensing data, an ecological restoration sequence planning model for abandoned mines was developed. This model identifies viable strategies for the reclamation planning of abandoned mines, thereby facilitating the implementation of a three-phase restoration project in Toli County. The study encompasses four primary objectives: (1) Dividing the abandoned mines into three contiguous and spatially concentrated regions based on spatial continuity and road accessibility, thereby reducing transportation and remediation costs in management projects; (2) Assessing the ecological significance of Toli County and categorizing the three regions into phased restoration projects to sequentially implement remediation efforts, prioritizing areas of high ecological importance; (3) Evaluating the hazards posed by abandoned mines during each project phase and determining the sequence of remediation activities to prioritize the treatment of mines that present significant threats to human safety; and (4) Completing the zoning, classification, and graded reclamation planning for abandoned mines in Toli County, establishing a restoration sequence for all abandoned mines, and providing a reference framework for planning in other similar locations and regions.

2. Methodology

2.1. Study Area

Toli County is located in the northwest of Xinjiang Uygur Autonomous Region, on the west side of the Junar Basin, in the southeast of the Tarim Basin, and in the geographical heart of the Eurasian continent and lies between 44°58′ and 46°24′ north latitude and 82°28′ and 85°20′ east longitude, with a border line of approximately 58 km. The county stretches 221.6 km from east to west and 159.3 km from north to south, covering a total area of 21,300 square kilometers km² (Figure 1). The County has favorable conditions for mineralization and serves as an important mineral resource reserve base in Xinjiang. Long-term mining activities have resulted in numerous abandoned mines, with non-metallic, metal, and metalloid mines accounting for 84.1%, 13.5%, and 2.4%, respectively. The main types of land damaged by these activities are grasslands, croplands, and construction land. Consequently, the development of mineral resources in Toli County has triggered regional environmental problems, including vegetation die-off, land degradation, soil erosion, and air pollution. Therefore, there is an urgent need to carry out ecological restoration in mining areas, particularly for historically abandoned mines.

2.2. Data Source

To investigate the current situation of abandoned mines in Toli County, China, we conducted field surveys and analyzed remote sensing data. The study primarily involved the collection of data pertaining to abandoned mines in Toli County, including the number and types of mines, their spatial coordinates (latitude and longitude), distances from human settlements and transportation infrastructure, the area occupied by each abandoned mine, and the types of land degradation associated with them, The “General Plan for Territorial Space of Toli County (2023)” was selected for analysis to determine the socio-economic planning of Toli County. Relying on the “Mine geological environment protection and land reclamation plan” of typical mines in the county, a survey was conducted on the mines and surrounding environment in Toli County. To clarify the sources and generation methods of the subsequent indicator data, such as population density, slope, and aspect, population density data were sourced from national census databases, while slope and aspect were derived from DEM (Digital Elevation Model) data using GIS software, specifically ArcGIS 10.8.

Landsat 8 OLI images for 2024 were downloaded from NASA Earth Science Data and subjected to radiometric and atmospheric correction, mosaicking, and clipping to the boundary of Toli County. On this basis, land use information and the vegetation index spatial distribution were interpreted, and the Normalized Difference Vegetation Index (NDVI) was calculated using the red and near-infrared bands. The 30 m DEM data were used to derive elevation and related topographic variables. gross domestic product (GDP) and precipitation data were obtained from ready-made gridded spatial datasets provided by the Chinese Academy of Sciences and the National Earth System Science Data Center, respectively. These datasets were reprojected, resampled to ensure compatibility with the Landsat-based layers, and clipped to the study area, and were then integrated into the GIS for subsequent ecological evaluation. These datasets were utilized to evaluate and analyze the ecological environment of the county (

Table 1. Research data collection.

Values	Data Types	Survey Time	Sources
Land utilization rate	Landsat 8 imagery (30 m)	17 March 2024	NASA Earth Science Data [29]
Vegetation index spatial distribution	Landsat 8 imagery (30 m)	17 March 2024	NASA Earth Science Data
GDP	Economic Factor Data	10 March 2024	Chinese Academy of Sciences Center for Resource Environment Sciences and Data [30]
Precipitation	Environmental Data	15 March 2024	National Earth System Science Data Center [31]
DEM elevation	Shuttle Radar Topography Mission (30 m)	2000	NASA Earth Science Data

).

2.3. Methods

The planning sequence for abandoned mines in Toli County is formulated by the requirements of ecological civilization construction, considering such factors as contiguity, ecological safety [32], and ecological functions for scheduling work phases [33]. A sequential restoration planning and design for abandoned mines in Toli County was undertaken.

Initially, a zoning plan for the abandoned mines in Toli County was formulated based on the degree of contiguity, dividing the county into three abandoned mine zones to facilitate the arrangement of subsequent work phases. Furthermore, to establish the prioritization sequence for the remediation of three zones of abandoned mines, the ecological importance of Toli County was evaluated based on the extent of human impact. A classification scheme was developed for the three primary regions, prioritizing ecological restoration in areas of greater human significance, thereby delineating the first, second, and third phases of the abandoned mine management project. Subsequently, for each phase, the hazards posed by abandoned mines to human populations were assessed, and remediation efforts were prioritized for mines presenting the highest threats to human safety. This methodology ultimately determined the restoration sequence of abandoned mines within each phase. Ultimately, this achieved the research goal of studying the management planning of abandoned mines in Toli County (Figure 2).

2.3.1. Abandoned Mine Zoning Management Planning Methodology

The zoning criteria for abandoned mines aimed to delineate regions where the mines are contiguous and spatially concentrated. To facilitate this zoning planning, cluster analysis was employed [34]. Specifically, the K-means clustering method was selected due to its efficacy in classifying samples or indices and its ability to distinguish similar data points effectively. Consequently, a cluster analysis with K = 3 was implemented to define the zoning framework.

Three factors were selected for the clustering and zoning of abandoned mines: longitude, latitude, and road index [35]. Since the K-means cluster analysis method uses the distances between data points, large differences in the dimensions of different factors can lead to the clustering results being directly controlled by a few high-dimension variables. Therefore, to ensure the same data dimensions, it was necessary to standardize the factor data [36]. Standardization used the Z-Score function, calculating the difference between each data value and the average value of the dataset AVERAGE (range) and dividing it by the standard deviation STDEV (range). The calculation formula Equation (1) is as follows:

O = (AVERAGE (range) − value)/STDEV (range).

(1)

After standardizing the basic data of abandoned mines in Toli County, the K-means cluster analysis with K = 3 was performed. Initially, three cluster centers, M₁, M₂, and M₃, were arbitrarily selected. The distance between each abandoned mine and the three cluster centers was calculated, and each abandoned mine was assigned to the group of the nearest cluster center, completing the first iteration of the analysis and resulting in three zones [37]. Then, the average values of the data in each zone were taken as the cluster centers for the second iteration, and the distances from each abandoned mine to these new cluster centers were calculated to obtain three zones, completing the second iteration. After five iterations of this cluster analysis, the calculations converged, the abandoned mines within each zone remained unchanged, and the clustering zoning was completed.

2.3.2. Management Planning Strategy for Abandoned Mine Classification

The above zoning plan was based on the contiguity degree and achieved zoning planning of abandoned mines but did not specify the management construction phases of each zone. This study used the analytic hierarchy process (AHP) [38], taking the degree of human activity impact as the analytical standard to evaluate the overall environment of Toli County [39,40]. Based on the zoning results, the first, second, and third phases of management projects for abandoned mines in Toli County were determined [41].

The key evaluation factors were used to study the extent of human impact [42] based on the significance of Toli County’s environment on human impact [43]. The hierarchical model of ecological importance in Toli County was categorized into three levels: the decision-making goal, criteria layer, and benchmark layer (Figure 3) [44]. The scaling method of the AHP was used to compare evaluation factors. The data were collected by distributing questionnaires to experts. Data were collected by distributing questionnaires to a total of 45 experts in the field of mine ecological restoration, resulting in a judgment matrix [45]. In the matrix, b_ij represents the relative importance of B_i to B_j, the level of importance is usually represented by numbers.

First, the consistency index (CI) was calculated (to check whether the hierarchical single ordering was confirmed) via the following formula Equation (2):

$$CI={\displaystyle \frac{{\lambda }_{max}-n}{n-1}},$$

(2)

where λ_max is the maximum eigenvalue of the matrix, and n is the matrix order. The closer the CI is to zero, the better the consistency, and vice versa [46]. The random consistency index (RI) was introduced to measure the magnitude of CI.

The consistency ratio (CR) is a parameter used to test the consistency of a matrix, which is derived as follows Equation (3):

CR = CI/RI.

(3)

The results of single rankings within the same level were used to test the overall ranking and consistency [47]. The composite weights of each element in each level were calculated using the Eigenvector method, which determined the weights of evaluation factors at both the criterion and reference levels. The following formula Equation (4) was used to calculate the weights of ecological importance evaluation factors in Toli County:

A × W = λ_max × W,

(4)

where A represents the matrix of weight ratios; W is the weight vector [48]. By calculating each weight vector W via Equation (4) and normalizing it, the factor weights within each judgment matrix were obtained.

2.3.3. Strategy for the Grading Planning of Abandoned Mines

The classification planning conducted in this study established three-phase restoration projects for abandoned mines in Toli County. However, the sequencing of restoration for multiple abandoned mines within each phase still required determination [49]. To address this, the degree of harm caused by each abandoned mine was utilized as a grading criterion, and principal component analysis (PCA) was employed to establish the restoration sequence within each project phase [50].

The PCA is a linear dimensionality reduction technique that uses variance to measure the differences in data [51]. Based on this characteristic, it was used to formulate the grading planning strategy. Abandoned mine characteristics were selected as influencing factors to calculate their comprehensive scores [52]. Based on the latter values, the ranking of the danger levels of abandoned mines within the affected area was analyzed, determining the order for the restoration of abandoned mines [53].

The selection of grading influencing factors primarily considered the harmfulness of abandoned mines and chose the respective evaluation indices [54]. Data preprocessing was performed to prepare the indices for analysis. Initially, the data size related to influencing factors should be positively correlated with the analysis objectives [55]. Therefore, to appropriately classify and grade abandoned mines, we applied a Z-Score method to normalize the data and assess the relative position of each mine within the dataset. The Z-Score formula is given as

$$Z={\displaystyle \frac{\left(X-\mu \right)}{\sigma }},$$

(5)

where Z is the Z-Score, X is the value of the indicator (e.g., area, distance, etc.) for each mine, $\mu$ is the mean of the indicator values, $\sigma$ is the standard deviation of the indicator values.

This method allows us to standardize the indicators, enabling comparison of different variables with varying units and scales. By applying this Z-Score normalization, the mines are classified into various categories, making it easier to assign them to appropriate zoning areas for restoration.

The grading management of abandoned mines primarily considers hazards posed by abandoned mines to humans and the ecological environment [56]. PCA is a method of linear dimensionality reduction that uses variance to measure the variability in data. Based on this characteristic, PCA was used in this study to determine grading planning strategies. The comprehensive score of abandoned mines was evaluated by selecting characteristics of abandoned mines as influencing factors. According to the comprehensive score, the level of danger of abandoned mines within the impact area was analyzed, and a ranking was made for the order of restoration of abandoned mines [57]. Therefore, the hazards of abandoned mines were also comprehensively classified into those concerning (i) safety for humans and nature, and (ii) economic impact. In the PCA, four factors were ultimately selected: distance, land category, area, and ecology indices (

Table 2. Influencing factors of the hierarchical analysis method.

Influencing Factor	Impact Principle
Distance indicator	Based on the safety level, the closer an abandoned mine is to densely populated areas, the greater its hazard.
Land category indicator	Based on the economic level, the higher the value of the land type destroyed, the greater the hazard of the abandoned mine.
Area indicator	Based on the safety level, the larger the impacted area, the greater the hazard of the abandoned mine.
Ecological indicator	Based on the ecological safety level, the more severe the disruption to ecological functions, the greater the hazard of the abandoned mine.

) [58].

3. Results

3.1. Abandoned Mine Zoning Planning

To facilitate the implementation of future remediation projects, the zoning requirements for abandoned mines necessitated the delineation of areas exhibiting high contiguity and concentrated spatial distributions of abandoned mines earmarked for remediation. Spatial coordinates, specifically longitude and latitude, were selected as evaluation factors due to their direct representation of the spatial distribution of abandoned mines. However, geographically proximate abandoned mines may be separated by natural barriers such as mountains and rivers, potentially impeding the practical execution of remediation efforts. Recognizing the critical role of transportation routes in ensuring mine accessibility, a road index was incorporated as an additional evaluation factor [59]. The assignment of road index values followed a systematic approach: first, the primary transportation routes within the region were identified and assigned specific values; second, the distances between each abandoned mine and these main roads were measured to determine the nearest road for each mine; finally, the road index value for each mine was defined based on its proximity to the nearest road. This methodology aimed to digitize and differentiate transportation routes, thereby enhancing the meaningfulness and utility of the road indices in the zoning process.

The purpose of assigning values to roadways is to digitize and differentiate roads, while also ensuring that road indicators contribute meaningfully to the clustering and zoning process [60]. The road value assignment follows a specific rule: roads oriented east–west are assigned odd numbers, with values increasing sequentially from east to west, while roads oriented north–south are assigned even numbers, with values increasing sequentially from north to south. Under this assignment scheme, adjacent or intersecting roads will have minimal value differences, whereas roads with greater spatial separation will exhibit larger value differences, thus optimizing the clustering and zoning outcomes. As illustrated in Figure 4, taking the roads and abandoned mines shown in the figure as examples: Highway X101 is assigned a value of 1, Highway X102 is assigned a value of 3, Highway X103 is assigned a value of 2, and Highway X104 is assigned a value of 4. Abandoned mine 1, located near Highway X102, is assigned a road indicator value of 3, while abandoned mine 2, located near Highway X101, is assigned a road indicator value of 1.

The standardized results on the longitude [61], latitude, and road index data for the abandoned mines in Toli County were designated from M001 to Mn (

Table 3. Standardized basic data for abandoned mines in Toli County, China.

Abandoned Mine Number	Longitude	Latitude	Road Index
M001	0.89235	−0.87623	0.75899
M002	0.89519	−0.87518	0.75899
…	…	…	…
Mn−1	0.89109	0.6214	0.75899
Mn	0.8852	0.55729	−0.70478

).

The final clustering results for standardized basic data of abandoned mines in Toli County with K = 3 in K-means clustering analysis [62]. Three abandoned mine zones were identified in Toli County. Zone 1 had abandoned mines, numbered from Q₁ to Q_a. Zone 2 had B abandoned mines, numbered from M₁ to M_b. Finally, Zone 3 had C abandoned mines, numbered from N₁ to N_c.

3.2. Abandoned Mine Classification Planning

While the zoning plan successfully categorized abandoned mines in Toli County based on contiguity, it did not specify the management timelines for each zone. To address this limitation, a comprehensive environmental assessment of Toli County was conducted using the hierarchical analysis method, with the degree of impact on human activities serving as the primary evaluation criterion. Based on the results of this zoning analysis, the remediation projects for abandoned mines in Toli County were subsequently delineated into first, second, and third phases.

An evaluation was conducted in Toli County based on the importance of the environmental impact on local inhabitants. Three ecological factors were primarily chosen: topography and geomorphology, vegetation and precipitation, and socio-cultural aspects. The degree of human impact was analyzed through eight secondary factors: elevation, aspect, slope, vegetation index, land use type, precipitation index, population density, and GDP [63,64]. The ecological importance hierarchical model of Toli County was subdivided into the following three layers: the decision goal criterion layer, the criterion layer, and the benchmark layer. The decision goal was the final evaluation result; the criterion layer comprised the three ecological factors, and the benchmark layer comprised eight secondary factors under these three ecological factors.

Pairwise comparisons of evaluation factors were conducted based on the degree of impact on humans, assigning values according to expert scoring. The judgment matrices for ecological importance, topography, geomorphology, vegetation, precipitation, and socio-cultural aspects of Toli County were I₁, I₂, I₃, and I₄, respectively, and had the following forms Equations (6)–(9):

$$I_1=\left[\begin{array}{ccc}1& {\displaystyle \frac{1}{2}}& {\displaystyle \frac{1}{6}}\\ 4& 1& {\displaystyle \frac{1}{4}}\\ 6& 4& 1\end{array}\right],$$

(6)

$$I_2=\left[\begin{array}{ccc}1& {\displaystyle \frac{1}{2}}& 2\\ 2& 1& 3\\ {\displaystyle \frac{1}{2}}& {\displaystyle \frac{1}{3}}& 1\end{array}\right],$$

(7)

$$I_3=\left[\begin{array}{ccc}1& 2& {\displaystyle \frac{1}{2}}\\ {\displaystyle \frac{1}{2}}& 1& {\displaystyle \frac{1}{3}}\\ 2& 3& 1\end{array}\right],$$

(8)

$$I_4=\left[\begin{array}{cc}1& 1\\ 1& 1\end{array}\right].$$

(9)

The consistency ratios of the four judgment matrices I₁, I₂, I₃, and I₄ were 0.0088, 0.0088, 0.0088, and 0, respectively, all passing the consistency test [65]. The analysis’s overall hierarchical sorting CR value was 0.0088, i.e., it passed the consistency test, allowing for calculating weights for the evaluation factors.

By integrating Equations (2)–(4) with Equation (6), the weights of the criteria layer factors were determined to be 0.1061, 0.1929, and 0.7010, respectively. Subsequently, the evaluation indicators within each criterion were calculated for the benchmark layer. By combining Equations (2)–(4) with Equations (7)–(9), the weights of all evaluation factors in the benchmark layer were obtained. Finally, the weights of ecological environmental importance evaluation factors for Toli County were obtained (

Table 4. Weights of ecological environmental importance evaluation factors for Toli County, China.

Factor of Criterion Layer	Weight of Ecological Factor	Factor of Benchmark Layer	Weight of Secondary Factor
Topography and geomorphology	0.1061	Slope	0.0573
		Elevation	0.0315
		Aspect	0.0173
Vegetation and precipitation	0.1929	Vegetation index	0.1041
		Land use type	0.0305
		Precipitation	0.0573
Socio-cultural factors	0.7010	Population density	0.3505
		GDP	0.3505

).

After determining the weights of each evaluation factor, the importance of each factor’s impact on humans in Toli County was assessed. Given that different factors vary in their degree of significance to human populations, each factor’s importance was evaluated individually. For example, regarding slope, human activities typically occur in areas with gentler gradients; therefore, regions with lower slopes hold greater significance for human habitation and production. Consequently, the impact of slope on humans is inversely proportional to its numerical value. The zoning principles for each evaluation factor are outlined in (

Table 5. Evaluation factor of zoning principles.

Ecological Factor	Secondary Factor	Grading Principle
Topography and geomorphology	Slope	The impact of slope on humans is inversely proportional to its numerical value.
	Elevation	The impact of elevation on humans is inversely proportional to its numerical value.
	Aspect	Aspect mainly affects factors such as sunlight duration and water accumulation capacity. Therefore, due north, northeast, and northwest have the greatest impact on humans, followed by due east and due west, with the least impact from east, southwest, due south, and flat land.
Vegetation and precipitation	Vegetation index	The magnitude of the vegetation index is inversely proportional to the degree of human impact.
	Land use type	Urban construction land and arable land have the greatest impact on humans, followed by forest land, grassland, water bodies, and wetlands, with the least impact from unused land and wasteland.
	Precipitation	The impact of precipitation on humans is directly proportional to its numerical value.
Socio-cultural factors	Population density	The impact of population density on humans is directly proportional to its numerical value.
	GDP	The impact of GDP on humans is directly proportional to its numerical value.

).

The remote sensing data for each ecological factor in Toli County were reclassified, and the impact of each category was evaluated and assigned numerical values. To eliminate the influence of dimensionality and enhance the assessment of importance, the data for the eight evaluation factors were categorized into Primary Impact Zones, Secondary Impact Zones, and Tertiary Impact Zones based on a hierarchical evaluation principle. Assigned values of 3, 2, and 1 were allocated to these zones, respectively, with the magnitude of the assigned values being proportional to their degree of importance (

Table 6. List of ecological importance evaluation factors in Toli County, China.

Ecological Factor	Secondary Factor	Category	Impact Level	Assigned Value
Topography and geomorphology	Slope	0–5	Primary impact zone	3
		5–10	Secondary impact zone	2
		≥10	Tertiary impact zone	1
	Elevation	188–1000	Primary impact zone	3
		1000–2000	Secondary impact zone	2
		2000–3067	Tertiary impact zone	1
	Aspect	Due north, northeast, northwest	Primary impact zone	3
		Due east, due west	Secondary impact zone	2
		East, southwest, due south, flat land	Tertiary impact zone	1
Vegetation	Vegetation index		Primary impact zone	3
		0.3–0.6	Secondary impact zone	2
		0–0.3	Tertiary impact zone	1
	Land use type	Urban construction land, arable land	Primary impact zone	3
		Forest land, grassland, water bodies, wetlands	Secondary impact zone	2
		Unused land, wasteland	Tertiary impact zone	1
	Precipitation	235–270	Primary impact zone	3
		217–235	Secondary impact zone	2
		190–217	Tertiary impact zone	1
Socio-cultural	Population density	6–9.8148	Primary impact zone	3
		3–6	Secondary impact zone	2
		0–3	Tertiary impact zone	1
	GDP	≥60	Primary impact zone	3
		18–60	Secondary impact zone	2
		0–18	Tertiary impact zone	1

).

Based on the aforementioned analysis of the importance and weight assignment of the evaluation factors, spatial overlay processing and analytical calculations were conducted on the remote sensing data for the eight assessment factors in Toli County. This process categorized the factors into three distinct classes: Primary Impact Zone, Secondary Impact Zone, and Tertiary Impact Zone. Taking slope as an example, as illustrated in Table 4, the remote sensing data were reclassified by assigning a value of 3 to slopes between 0 and 5°, a value of 2 to slopes between 5 and 10°, and a value of 1 to slopes exceeding 10°. This reclassification resulted in the zonation map for the slope factor in Toli County (Figure 5a). Similarly, the remaining factors were reclassified according to their respective levels, thereby producing zoning maps for all eight evaluation factors in Toli County (Figure 5a–h).

By combining the latter map with the impact weights of each factor and performing grid calculations and overlaying, the comprehensive evaluation map of the ecological importance of Toli County was constructed (Figure 6). The most significant factors affecting the importance levels of abandoned mine areas in Toli County were the population density, GDP, and vegetation index. The primary Impact Zone in Toli County, influenced by anthropogenic factors like population density and GDP, was mainly located near Toli County City and Tiechanggou Town. Influenced by natural factors such as precipitation, vegetation, and land use, the Secondary Impact Zone was concentrated in the northwestern part of Toli County. Other areas in Toli County with less significant impacts formed the Tertiary Impact Zone, mainly covering the eastern and southern parts of Toli County.

Using mathematical quantitative modeling, the ecological impact importance evaluation system was established for Toli County. The numbers and ratios of importance levels for abandoned mine areas were determined by integrating zone planning, ultimately subdividing the project into three phases. The base map’s remote sensing data is sourced from Maxar Technologies (Figure 7).

The red area in the figure represents Phase I of the project, the purple area denotes Phase II, and the green area corresponds to Phase III. The rationale for this delineation is as follows:

(1)

Abandoned Mine Zone 1: This zone is predominantly characterized by primary and secondary ecological impacts in Toli County, significantly affecting both human society and the ecological environment. Consequently, it has been designated as the first phase of the abandoned mine restoration project in Toli County. Priority remediation efforts will focus on the abandoned mines within this area, which is termed the Human Factors-Dominated Ecological Priority Restoration Zone.

(2)

Abandoned Mine Zone 2: This zone primarily exhibits primary and tertiary impacts in Toli County. It is therefore designated as the second phase of the abandoned mine restoration project. Although this area experiences lower levels of human activity, it benefits from favorable natural conditions such as adequate precipitation and temperature, indicating substantial ecological potential. As a result, abandoned mines within this zone will undergo secondary remediation. This area is named the Climate Factors Dominated Ecological Key Restoration Zone.

(3)

This zone is mainly characterized by secondary and tertiary impacts in Toli County, posing minimal threats to human society and the natural environment. Additionally, it has limited potential for ecological development. Consequently, it is classified as the third phase of the abandoned mine restoration project in Toli County, with remediation activities scheduled subsequently. This zone is designated as the Topography Factors-Dominated Ecological Routine Restoration Zone.

3.3. Abandoned Mine Grading Planning

The above classification plan outlined the three phases of restoration projects for the abandoned mines in Toli County. However, within each phase, the sequence for restoring multiple abandoned mines still had to be planned. The restoration sequence for abandoned mines within each restoration project was established based on the degree of hazard posed by the mines, using principal component analysis as the grading standard.

In the first phase of the Toli County project, data regarding the distance, land category, area, and ecological indices of abandoned mines were systematically analyzed. Due to policy requirements, the abandoned mines were anonymized and assigned numerical identifiers ranging from Q₀₀₁ to Q_a (

Table 7. Basic data for grading abandoned mines in the severely impacted area of Toli County, China.

Abandoned Mine Number	Area	Land Category Indicator	Ecological Indicator	Distance Indicator
Q001	0.90661	−0.19807	−0.12121	1.64870
Q002	0.09268	4.95169	3.58763	1.86197
…	…	…	…	…
Qa−1	2.90446	−0.19807	0.03149	−0.41150
Qa	−0.25628	−0.19807	−0.21008	0.41677

). To mitigate the influence of differing units and ensure comparability, the raw data collected were standardized using the Z-Score normalization method (Table 6).

PCA was used for data analysis to rank abandoned mines within the project’s first phase. PCA applicability is known to require the Kaiser-Meyer-Olkin (KMO) value of no less than 0.6 and the significance (Sig) of Bartlett’s test of sphericity not exceeding 0.05. The KMO value of this model is 0.613, and Bartlett’s test of sphericity value is 0.0001, both meet the requirements for analysis. The indices of the performed PCA had a good degree of association, allowing for the PCA to be conducted [66].

Further extract principal components from the total variance with initial eigenvalues greater than 1, calculated as principal components 1 (Y₁) and 2 (Y₂). Together, they accounted for 87.11% of the total variance. (

Table 8. Abandoned mine comprehensive score total variance explanation table based on the AHP.

Principal Component	Initial Eigenvalue (%)			Extraction Sums of Squared Loadings (%)
	Total	Variance	Cumulative	Total	Variance	Cumulative
1	2.445	61.115	61.115	2.445	61.115	61.115
2	1.040	25.998	87.113	1.040	25.998	87.113
3	0.488	12.193	99.306	\	\	\
4	0.028	0.694	100.000	\	\	\

) shows the total variance explanation for the PCA [67].

Consequently, the hazard assessment model for abandoned mines was characterized by Principal Components 1 and 2. To elucidate the coupling relationships between the four evaluation indicators and the hazards associated with abandoned mines, the interactions between these four rating indicators and the principal components were subsequently calculated (

Table 9. Component matrix table for the comprehensive score of abandoned mines based on the AHP.

Evaluation Index	Component
	Principal Component 1	Principal Component 2
Ecological index (X1)	0.945	−0.244
Land category index (X2)	0.920	−0.332
Distance index (X3)	0.763	0.296
Area index (X4)	0.350	0.885

). A comprehensive score calculation for the principal components was conducted to determine the restoration order of abandoned mines in the severe impact zone, where X₁ represents the ecological index, X₂ is the land category index, X₃ is the distance index, and X₄ is the area index. The weights of the evaluation indices in each principal component were calculated

The corresponding coefficients of each evaluation index were calculated using the data in the component matrix table and the eigenvalues of the corresponding principal components. The following formulas Equations (10) and (11) were used to determine the composite scores for principal components 1 and 2:

Y₁= 0.6014 × X₁ + 0.5884 × X₂ + 0.4880 × X₃ + 0.2238 × X₄,

(10)

Y₂ = −0.2393 × X₁ − 0.3256 × X₂ + 0.2903 × X₃ + 0.8678 × X₄.

(11)

Combining various ratios of principal components Y₁ and Y₂, the final comprehensive score Y for each abandoned mine was derived using Equation (12):

Y = 0.61115 × Y₁ + 0.25998 × Y₂.

(12)

Using the above formulas for each principal component and the data of each index (after preprocessing), the comprehensive scores of principal components 1 (Y₁) and principal component 2 (Y₂) for each abandoned mine, as well as the overall mine score Y, were obtained. The abandoned mines within the project phase were graded and ranked based on the comprehensive score Y. The comprehensive scores and grading order of the abandoned mines in phase one of the project (

Table 10. Grading for abandoned mines in the first phase of the Toli County, China.

Mine Number	Area	Land Type	Ecology	Distance	Y1	Y2	Y	Grading
Q1	0.724	4.952	6.000	3.147	8.220	−1.506	4.632	1
Q2	0.093	4.952	3.588	1.862	6.001	−1.850	3.186	2
…	…	…	…	…	…	…	…	…
Qa−1	−0.832	−0.198	−0.254	−0.601	−0.749	−0.771	−0.658	A−1
Qa	−0.890	−0.198	−0.259	−0.594	−0.761	−0.818	−0.678	A

).

The grading and ranking of abandoned mines were directly proportional to their threat to human health and the ecological environment, meeting the requirements of grading planning. This approach achieved graded management planning for restoring abandoned mines from ecological safety and functional perspectives. The arrangement of the first phase of the abandoned mine project in Toli County (Figure 8).

By applying the same methodology, the restoration sequences for abandoned mines exhibiting secondary (Figure 9) and tertiary (Figure 10) impacts in Toli County were subsequently determined.

4. Discussion

4.1. Advantages of Zoning, Classification, and Grading Planning for Abandoned Mines

4.1.1. Advantages of Zoning Planning for Abandoned Mines

As can be seen from Figure 7, the abandoned mines in Toli County were primarily distributed in the central, northwest, and northeast parts of the county, with only a few scattered in other areas. zones 1 and 2 show contiguous effects. However, the continuity of Zone 3 is significantly poorer (Figure 11a), as the complex terrain such as mountains and rivers, the road does not directly connect the north and south sections, which increases the actual driving distance, affects the management cost, and reduces the zoning benefit. The distribution of abandoned mines in Zone 2 and Zone 3 is concentrated (Figure 11b), which are connected by county road and S201 expressway, and have a good contiguous effect.

By incorporating longitude, latitude, and the road index into the K-means clustering, the delineated zones are not only spatially coherent but also accessible along the main transport corridors of Toli County [68]. Compared with the control scheme that uses only geographic coordinates, the zoning scheme with the road index produces clusters that correspond to groups of mines that can be reached via the same highway or county-road network, rather than mines that are close in Euclidean distance but separated by mountains or river valleys. This improves the practicality of the planning results, because remediation projects can be organized as contiguous engineering packages, reducing mobilization costs, travel time, and difficulties in on-site supervision. Similar accessibility-based zoning ideas have been adopted in regional ecological restoration and watershed management planning, where transport or hydrological connectivity is used to group management units; our results demonstrate that this concept is also effective for abandoned mine clusters in arid mountainous regions.

4.1.2. Advantages of Classification Planning for Abandoned Mines

A work schedule can be arranged for each zone by combining the comprehensive ecological importance evaluation map with the three abandoned mine zones (

Table 11. Table of the proportion of impact zones of different importance in each phase of abandoned mine remediation pro-jects in Toli County, China.

Construction Phase	Proportion of Primary Impact Area (%)	Proportion of Secondary Impact Area (%)	Proportion of Tertiary Impact Area (%)
Phase I project	19.6	35.3	45.1
Phase II project	8.7	21.7	69.6
Phase III project	0	5.8	94.2

).

The proportions of abandoned mines situated within various impact zones for the Phase III restoration project in Toli County are presented in Table 11. The conclusions drawn are as follows: (1) In Phase I, abandoned mines comprised 19.6% of the primary impact zone, 35.3% of the secondary impact zone, and 45.1% of the tertiary impact zone. (2) In Phase II, abandoned mines accounted for 8.7% of the primary impact zone, 21.7% of the secondary impact zone, and 69.6% of the tertiary impact zone. (3) In Phase III, no abandoned mines were located within the primary impact zone, 5.8% were situated within the secondary impact zone, and 94.2% were found within the tertiary impact zone.

The AHP-based ecological importance evaluation provides an explicit link between the spatial zones and the underlying ecological–socioeconomic context [69]. By weighting vegetation index, precipitation, population density, GDP, and other indicators, the method highlights areas where abandoned mines intersect with dense human activities and high-value ecosystems. The fact that Phase I restoration projects contain the largest share of mines located in primary and secondary impact zones indicates that the planning scheme successfully concentrates early remediation efforts in places where ecological functions and human safety are most vulnerable. In contrast, Phases II and III are dominated by secondary and tertiary impact zones, reflecting a gradual shift from urgently needed interventions to more routine ecological improvement. This priority pattern agrees with regional ecological sensitivity and restoration studies in other parts of China, where population density, economic output, and vegetation condition are also identified as dominant drivers of ecological importance and high-impact zones are preferentially targeted for protection and restoration. Thus, the classification planning in this study is not only internally consistent but also aligned with broader regional experiences.

4.1.3. Advantages of Grading Planning for Abandoned Mines

The first principal component (Y₁) accounts for 61% of the cumulative score contribution. According to Equation (6), both ecological indicators and land category indicators significantly influence Y₁, indicating that Y₁ primarily represents the hazard impacts related to the ecological functionality of abandoned mines. The second principal component (Y₂) accounts for 26% of the cumulative score contribution. As demonstrated by Equation (7), area indicators and distance indicators exert positive influences on Y₂, whereas ecological indicators and land category indicators have negative influences. This suggests that Y₂ mainly reflects the hazard impacts associated with the ecological safety of abandoned mines.

Furthermore, by integrating Equations (10) and (11) into Equation (12), the following equation for the comprehensive score Y of the abandoned mines and the four indices, namely ecological index X₁, land category index X₂, distance index X₃, and area index X₄, was obtained Equation (13):

Y₁ = 0.3053 × X₁ + 0.2749 × X₂ + 0.3737 × X₃ + 0.3623 × X₄

(13)

The standardized coefficients for the ecological index (X₁), land category index (X₂), distance index (X₃), and area index (X₄) are 0.3052, 0.2749, 0.3737, and 0.3623, respectively, in relation to the comprehensive hazard scores of abandoned mines. All four indicators exert a positive influence on the hazards associated with abandoned mines, with the distance and area indices having relatively greater impacts. This underscores the significant role of ecological safety levels in the planning of abandoned mine remediation. Consequently, project prioritization should focus on remediating abandoned mines that are located near densely populated areas and those that encompass large, disturbed areas.

The PCA-based grading further refines restoration priorities within each construction phase by revealing which combinations of mine characteristics drive hazard levels [70]. In our analysis, the first principal component, dominated by ecological and land use indicators, explains around 60% of the total variance and can be interpreted as an “ecological function” dimension, while the second component, controlled mainly by area and distance indicators, represents an “exposure and safety” dimension. This two-axis structure is comparable to PCA-based evaluations of water-quality risk, regional drought vulnerability, and other environmental hazards in China and elsewhere where the first component typically captures natural or land-cover conditions and the second reflects human disturbance or exposure. The similarity suggests that the relative influence of ecological status versus exposure is robust across different regions and environmental problems.

Transforming the principal components into a single comprehensive score yields a continuous index that can be used to rank individual mines within the same classification zone. Mines with higher scores are simultaneously located in ecologically important areas, closer to settlements, and with larger disturbed areas, meaning that remediation there is likely to deliver the greatest reduction in ecological and safety risk per unit investment. Compared with approaches that only consider administrative priority or single indicators, the PCA-based grading offers a more transparent and quantitative basis for sequencing projects and can be readily combined with other decision-making tools in future work.

While this study focuses on the specific conditions of Toli County, the framework developed for zoning, classification, and grading of abandoned mines is flexible and can be adjusted to suit different environmental contexts. The integration of AHP and PCA allows for the customization of factors depending on the local context, making the methodology applicable to a wide range of mining areas with varying ecological and socio-economic conditions. In addition, future studies could explore the adaptation of this approach to different mining types (e.g., coal, metals, non-metallic minerals) and the corresponding changes in the factor weights and hazard assessments.

By refining these elements, this model can be applied globally, offering a robust tool for mine restoration planning in diverse geographic and environmental settings.

4.2. Policy Implications

The Chinese authorities promoted the ecological restoration of regional historical legacy mines, issuing such regulatory documents as “Notification on supporting the demonstration project for ecological restoration of historical legacy abandoned mines” and “Opinions on exploring the use of market-oriented approaches to promote mine ecological restoration” [71]. However, the ongoing restoration process is still susceptible to subjective factors and lacks scientific and standardized scheduling, resulting in unreasonable restoration plans, low quality of restoration, slow progress, and untimely management of key abandoned mines for restoration.

Based on the above analysis of the case study area, the following suggestions have been proposed: (1) Prioritize and categorize management policies for abandoned mines, classify the mines to ensure effective allocation and use of resources, and prioritize the management of high-hazard abandoned mines. (2) Strengthen scientific and standardized scheduling to reduce restoration costs and improve management efficiency and quality. (3) Establish demonstration applications for management planning and formulate and improve relevant laws and policies. To ensure effective and orderly management of mine restoration.

4.3. Limitations and Future Work

Despite the above useful findings, this study has the following two main limitations. Firstly, it formulated restoration plans only for one case study region, namely Toli County, while differences in the environments of various sites and regions imply respective differences in the main influencing factors for the grading, classification, and zoning planning of abandoned mines. Secondly, the proposed planning covered overall abandoned mines in the region under study without considering the specific management plans for individual abandoned mines [72].

In the follow-up study, more attention will be dedicated to the following aspects: (1) In the planning strategy stage, adjust the evaluation factors and their weights in the plan according to the environmental characteristics of different regions, forming planning strategies for arid areas, cold areas, alpine areas, monsoon areas, and other different environments. (2) Based on the overall planning of abandoned mines and the analysis of the mines’ hazards, guide the arrangement of individual management projects for abandoned mines. (3) Since the restoration cycle of abandoned mines is long, it is necessary to establish an evaluation system for restoration effects and strengthen the maintenance, monitoring, and management of the restored legacy mines [73].

Furthermore, future work could combine the AHP and PCA derived indices with multi-objective optimization techniques to refine restoration prioritization under budgetary and temporal constraints [74,75]. For example, integer programming or evolutionary algorithms could be used to identify portfolios of projects that simultaneously maximize ecological improvement (such as reductions in comprehensive hazard scores or increases in restored areas within high-importance zones) and minimize implementation costs or total travel distance along the road network. Such an approach would transform the current stepwise planning strategy into a decision-support tool that explicitly quantifies trade-offs among competing management objectives and provides more flexible scheduling options for local governments.

This planning approach sequentially employs three methods, advancing the research on planning strategies. A comprehensive summary of the different methods used, evaluation factors, and results is provided, along with an outline of the study’s objectives and key findings. The summary is presented in the following

Table 12. Overview of Abandoned Mine Planning.

TEST Aim	Abandoned Mine Remediation Planning
Zoning planning	Method	Cluster analysis (CA)
	Evaluation factors	Longitude, latitude, road index
	Results	Based on spatial connectivity, three relatively concentrated zones of abandoned mines were delineated.
Classification planning	Method	Analytic hierarchy process (AHP)
	Evaluation factors	Slope, elevation, aspect, vegetation index
	Results	Land use type, precipitation, population, density, GDP
Grading planning	Method	The three zones were designated as three phases of the remediation project, based on their ecological significance.
	Evaluation factors	Principal component analysis (PCA)
	Results	Area

.

In addition to surface ecological factors, it is essential to consider subsurface conditions, particularly the geological environment, when planning ecological restoration in abandoned mine areas [76]. Mineral resource development causes significant disturbances not only to the surface but also to the subsurface, including soil composition, groundwater quality, and geological stability. These underground factors can influence the success of restoration efforts, as changes in the subsurface environment may affect water retention, soil fertility, and overall ecosystem stability.

Geological conditions such as soil structure, permeability, and the presence of contaminants in the groundwater should be incorporated into the restoration strategy [77]. For instance, areas with unstable geological conditions may require more intensive interventions to stabilize the soil and prevent further degradation. Additionally, the quality of groundwater, which can be affected by previous mining activities, should be monitored and addressed to ensure the health of ecosystems in restored areas.

To further validate the model results, future research could incorporate ground-truth verification by collecting field data from abandoned mine sites in Toli County or similar regions. This would allow for a direct comparison between the predicted restoration priorities and actual ecological conditions. In addition, we plan to compare the model’s outcomes with existing regional restoration plans from similar mining areas, such as region, to assess the accuracy and applicability of the proposed methodology. The validation through both ground-truthing and comparative analysis with regional plans would provide further support for the reliability of the restoration prioritization framework developed in this study.

5. Conclusions

This study introduces a novel framework for the zoning, classification, and grading management of abandoned mines, which mitigates the influence of subjective factors in the decision-making process. By integrating Cluster Analysis (CA), Analytic Hierarchy Process (AHP), and Principal Component Analysis (PCA), this research establishes a spatially and ecologically prioritized approach for restoration planning. The findings of this study contribute significantly to the field of ecological restoration by providing a systematic methodology for evaluating and addressing the hazards posed by abandoned mines.

(1)

Spatial Concentration and Connectivity: The use of longitude, latitude, and road indices facilitated the identification of spatially contiguous mine zones, effectively overcoming geographical barriers such as mountains and rivers. The cluster analysis methodology allowed for the delineation of three contiguous abandoned mine zones, significantly reducing the logistical challenges and costs associated with subsequent remediation efforts.

(2)

Zoning Classification Based on Ecological Importance: The study identified population density, economic indicators, vegetation, and precipitation as the primary factors influencing the ecological importance of mine zones. A quantitative ecological impact evaluation system was developed for Toli County, leading to the classification of the study area into three distinct zones based on ecological priority: Human Factors-Dominated Ecological Priority Restoration Zone; Climate Factors-Dominated Ecological Key Restoration Zone; Topography Factors-Dominated Ecological Routine Restoration Zone. This classification provides a clear basis for prioritizing restoration efforts according to the ecological significance of each zone.

(3)

Hazard Assessment and Grading: PCA was employed to assess the ecological hazards of abandoned mines, focusing on ecological function and safety. By selecting four key indicators—distance, area, ecology, and land category—the study quantified the environmental and safety risks posed by abandoned mines. The analysis demonstrated that distance and area are the primary indicators influencing the severity of these hazards, thereby aiding in the prioritization of remediation efforts.

Although this study focuses on abandoned coal mines located in an arid region, where land degradation is the dominant ecological issue and water pollution is relatively limited, the evaluation framework can be readily extended to metal mining areas in which off-site water quality is the primary concern. In such cases, water-quality indicators (e.g., concentrations of heavy metals in surface and groundwater, pH, and major ions) can be incorporated as additional evaluation factors in the AHP system, and their spatial distributions can be used to further refine the priority zoning for ecological restoration. This demonstrates that the proposed model is adaptable and capable of supporting region-specific planning for different types of mines and associated environmental stressors.