Remote Sensing Evaluation and Monitoring of Spatial and Temporal Changes in Ecological Environmental Quality in Coal Mining-Intensive Cities

Featured Application

This study proposes a new method for assessing the spatial and temporal changes in the ecological quality in resource cities. The rapid and effective assessment of ecological quality changes in resource cities was realized. This study provides valuable data support and a policy reference for the ecological protection and sustainable development of resource cities.

Abstract

In coal-dependent urban economies, the dichotomy between resource exploitation and ecological conservation presents a pronounced challenge. Traditional remote sensing ecological assessments often overlook the interplay between mining activities and urban environmental dynamics. To address this gap, researchers developed an innovative Resource-Based City Ecological Index (RCEI), anchored in a Pressure–State–Response (PSR) framework and synthesized from six discrete ecological indicators. Utilizing geodetic remote sensing data, the RCEI facilitated a comprehensive spatiotemporal analysis of Jincheng City’s ecological quality from 1990 to 2022. The findings corroborated the RCEI’s efficacy in providing a nuanced portrayal of the ecological state within mining regions. (1) Jincheng City’s ecological quality predominantly sustained a mudhopper-tier status, exhibiting an overarching trend of amelioration throughout the study period. (2) Disparities in ecological landscape quality were pronounced at the county level, with Moran’s Index exceeding 0.9, signifying a clustered ecological quality pattern. High–high (H–H) zones were prevalent in areas of elevated altitude and dense vegetation, whereas low–low (L–L) zones were prevalent in urban and mining sectors. (3) Further, a buffer zone analysis of two coal mines, differing in their mining chronology, geographical positioning, and operational status, elucidated the ecological impact exerted over a 32-year trajectory. These insights furnish a robust scientific and technical foundation for resource-centric cities to fortify ecological safeguarding and to advance sustainable development stratagems.

1. Introduction

Resource-based cities (RBCs) refer to the type of cities that make the exploitation and processing of minerals, forests, and other natural resources their leading industries [1,2].

RBCs are vital for national energy and resource security. With their extensive numbers and widespread distribution, these cities have significantly contributed to China’s economic development. However, rapid urbanization has intensified the conflict between resource exploitation and ecological protection, leading to severe environmental degradation, including land subsidence [3], solid waste accumulation [4], water resource degradation [5], and air pollution [6]. This environmental decline directly impacts residents’ health and well-being, posing threats to sustainable development [7,8]. Therefore, accurately assessing the ecological quality of resource-based cities is crucial for their sustainable development. Currently, ecological quality evaluation at various spatial scales is a prominent topic in geographical and ecological research [9,10], both domestically and internationally. With the increasing global awareness of environmental issues and the demand for sustainable development, more and more scholars and government officials are realizing that it is urgent to address the environmental challenges in RBCs. Consequently, ecological vulnerability [11,12], ecological security [13,14,15], and ecological resilience [16,17,18] have emerged as key topics in the ecological research of these cities. Furthermore, the evaluation and monitoring of the ecological environment play a crucial role in ecological research, and it has become a hot topic in geographical and ecological research [19,20]. According to the China Sustainable Development Plan for Resource-Based Cities (2013–2020), 262 cities are defined as being resource-dependent and classified into four types: growing cities, mature cities, declining cities, and regenerating cities [21]. Notably, prefecture-level resource-based cities are categorized into five categories depending on their resource types, coal, oil and gas, ferrous metal, nonferrous metal, and forestry, with their respective proportions being 44.4%, 9.5%, 33.3%, 9.5%, and 3.3% [22,23].

Remote sensing technology has emerged as a principal tool for ecological environmental monitoring, attributed to its characteristics of rapid data acquisition and periodic observation [24,25], especially in mining areas. As early as 1973, NASA began to use Earth resources satellites to monitor land use, vegetation coverage, and water accumulation in the Appalachian mining area [26]. Significant advancements in multispectral and hyperspectral remote sensing technologies [27], using satellite [28], airborne [29,30], and ground-based platforms [31], have greatly enhanced ecological evaluation in mining regions. Due to the severe surface damage from resource extraction, ground surveys have become crucial for monitoring and assessing these environments [32]. Leveraging quantitative remote sensing techniques, we can precisely monitor ecological components, including vegetation, water, and soil, enabling a comprehensive analysis and assessment of an ecological environment’s quality.

Vegetation serves as the clearest indicator of ecological and environmental changes in mining areas [33]. Among various indices, the Normalized Difference Vegetation Index (NDVI) is widely recognized as one of the most effective for characterizing vegetation changes [34,35,36,37]. With the rapid development of remote sensing satellite sensors, a series of innovative red-edge vegetation indices (REVIs), represented by the Sentinel-2 satellite, have been proposed to monitor and evaluate the growth status and health condition of vegetation [38,39,40]. Ma et al. [41] analyzed the spatiotemporal variation in the vegetation indices in subsidence with the help of the NDVI and five typical REVIs.

In mining areas, relying on a single environmental indicator is insufficient to reflect ecological trends due to the complexity and diversity of ecosystems. A single indicator, such as the NDVI, only captures one aspect, potentially leading to misjudgments. It also overlooks the interactions between water, soil, and air quality and lacks the capacity for comprehensive assessment. Therefore, a multi-indicator integrated assessment is necessary for a more accurate and holistic evaluation of an ecological environment. To address this, the Ministry of Environmental Protection of China introduced the Ecological Environment Index (EI), a comprehensive metric that includes six weighted indicators: biological richness, vegetation coverage, water network density, land stress, pollution load, and environmental restrictions [42]. Despite its significant role in ecological assessments, the EI’s indicator selection and weight determination processes are subjective and complex. Firstly, the selection of EI indicators often relies on expert judgment, which can lead to inconsistent results across different studies. Secondly, the determination of indicator weights lacks uniform standards and is typically conducted through methods like expert scoring or the Delphi method, both of which are inherently subjective. Additionally, the EI model’s applicability across different regions and time scales is debated, as varying ecological characteristics and stressors may necessitate different indicators and weights.

To address these limitations, researchers have developed more objective methodologies for weighting individual ecological indicators. For instance, principal component analysis (PCA) can automatically assign indicator weights, reducing the influence of human subjectivity. Furthermore, optimization models that incorporate regional ecological features can better reflect regional ecological changes. Future research should focus on developing more objective and adaptable assessment models to enhance the accuracy and reliability of ecological environmental evaluations. With the help of PCA, Xu et al. proposed the Remote Sensing Ecological Index (RSEI), which encompasses four key indicators: greenness, wetness, dryness, and heat [43]. Many scholars have since enhanced the RSEI model, successfully adapting it for ecological quality assessments and monitoring in diverse regions, including mining areas, cities, river basins, and nature reserves [44]. For example, considering the characteristics of mining regions, a specialized Iron Mine Remote Sensing-Based Ecological Index (IM-RSEI) has been proposed for dense mining areas with the iron ore industries [45].

Current research on ecological quality evaluation using remote sensing data is extensive, but studies on resource-based cities often rely on panel data and statistical yearbooks, lacking objective quantitative assessments. This study addressed this gap by focusing on Jincheng City, a mature resource-based city, to analyze the dynamic impacts of mining cycles and urban construction on ecological changes.

Urban development in resource-based cities typically centers around mining areas. Existing remote sensing assessments often overlook the combined effects of mining and urban expansion on the environment. The large-scale exploitation of resources and rapid urban growth significantly impact local air quality, with coal dust posing serious health risks to residents and vegetation. To provide a comprehensive evaluation, this study integrated the Indexed Coal Dust Index (ICDI) and Enhanced Vegetation Index (EVI) into the Remote Sensing Ecological Index (RSEI), creating the Resource Cities Ecological Index (RCEI).

By examining the temporal and spatial gradients of these impacts, this study offers strategic recommendations and a feasible framework for sustainable mining operations and ecological restoration. It proposes an ecological evaluation model tailored to coal resource cities, guiding future environmental studies.

The objectives of this study are the following:

(1)

Utilizing Landsat data retrieved from the GEE platform, we aim to develop an ecological index, namely the Resource-Based City Ecological Index (RCEI), specifically designed to assess the ecological quality of cities relying on coal resources. This index takes into account the characteristics of coal exploitation as well as the synergistic effects between mining areas and urban environments.

(2)

Employing the RCEI index, we intend to monitor the spatial and temporal changes in the ecological environmental quality of Jincheng City from 1990 to 2022. Additionally, we will conduct a thorough analysis of the spatial distribution patterns of the ecological environmental quality in the city, offering insights into its evolving ecological environment and guiding effective management strategies.

(3)

Utilizing the Resource City Ecological Index (RCEI), this investigation will delve into the ecological dynamics encircling two emblematic mining districts within designated buffer zones. This study will contrast the ecological variances at divergent junctures of mining activity—pre-extraction, during active extraction, and post-closure—alongside spatial disparities proximal and distal to urban centers.

(4)

Furthermore, this study will scrutinize the intricate repercussions of coal extraction on ambient ecosystems, dissecting temporal and spatial gradients of influence. Synthesizing these multifaceted analyses, this research will proffer strategic counsel and actionable frameworks for mining operations and ecological rehabilitation, charting a course for the sustainable evolution of resource-dependent cities.

2. Study Area and Data Sources

2.1. Study Area

Jincheng City, a prefecture-level city, is located in the southeastern region of Shanxi Province. Its geographical coordinates span from 112°45′10″ E to 112°55′08″ E in longitude and from 35°24′55″ N to 35°35′45″ N in latitude. The city is surrounded by mountainous landscapes and features a relatively flat terrain in its central area (refer to Figure 1). Jincheng stands as a pivotal anthracite coal base in China, with its coal-bearing area encompassing 49% of the total territorial area. Its annual anthracite coal production comprises nearly 50% of the national aggregate output. Furthermore, Jincheng boasts over 110 active coal mines, predominantly located in Yangcheng, Zezhou, Gaoping, and the urban center, collectively contributing a substantial 93.3% to the city’s total coal production capacity.

Jincheng City, with a history of over 50 years of large-scale coal mining, is currently confronted with a series of severe environmental issues resulting from high-intensity mining activities. These issues primarily encompass ground subsidence and collapse, reduction in and pollution of water resources, as well as the accelerated degradation of vegetation and soil erosion [46,47]. In this context, the efficient employment of remote sensing technology in assessing and monitoring the ecological environmental quality of Jincheng City is crucial in fostering the sustainable development of this resource-intensive urban landscape [48].

2.2. Data Source

A series of Landsat 5 TM and Landsat 8 OLI images were utilized to construct the ecological index. Specifically, this study employed the surface reflectance data products of Landsat 5 TM images from 1990, 1996, 2002, 2006, and 2008, as well as Landsat 8 OLI images from 2013, 2018, and 2022. The details of the Landsat data are presented in

Table 1. Landsat image series data of Jincheng City from 1990 to 2022.

1990 (Landsat TM)		1996 (Landsat TM)		2002 (Landsat TM)		2006 (Landsat TM)
Path/Row	Data Times	Path/Row	Data Times	Path/Row	Data Times	Path/Row	Data Times
124,036	7 July 1990	124,035	13 June 1996	124,035	8 July 2002	124,035	1 June 2006
125,035	31 August 1990	124,036	16 August 1996	124,036	24 July 2002	124,036	20 August 2006
124,036	24 August 1990	125,036	7 August 1996	125,036	13 June 2002	125,035	10 July 2006
125,035	28 June 1990	124,035	29 June 1996	125,035	16 August 2002	125,036	24 June 2006
125,036	14 July 1990	125,035	4 June 1996	125,036	29 June 2002	125,036	11 August 2006
2008 (Landsat TM)		2013 (OLI/TIRS)		2018 (OLI/TIRS)		2022 (OLI/TIRS)
Path/Row	Data Times	Path/Row	Data Times	Path/Row	Data Times	Path/Row	Data Times
125,036	29 June 2008	124,035	4 June 2013	124,035	12 June 2018	125,036	23 June 2022
125,036	31 July 2008	124,035	7 August 2013	124,035	30 July 2018	125,035	16 June 2022
124,035	25 August 2008	124,035	6 July 2013	124,035	31 August 2018	124,035	19 August 2022
124,036	24 July 2008	124,036	23 August 2013	124,036	28 June 2018	125,035	25 July 2022
124,035	22 June 2008	125,036	27 June 2013	125,035	22 August 2018	125,036	9 July 2022

. To mitigate the uncertainties associated with seasonal variability, this study focused on the timeframe from June to September, coinciding with the peak of vegetative growth [49]. These Landsat data were accessed from the public data archive of the Google Earth Engine (GEE) platform (https://code.earthengine.google.com/, accessed on 3 October 2023).

In this study, the Landsat Ecosystem Perturbation Adaptive Processing System (LEDAPS) [50] and the Landsat Surface Reflectance Code (LaSRC) (https://earthengine.google.com accessed on 3 November 2023) [51] were used to calibrate the at-sensor radiance of Landsat 5 and Landsat 8 images for surface reflectance. We also utilized the C implementation of the mask function (CFMASK) [52] to determine useful pixel data quality information such as clouds, shadows, water, and snow. The thermal infrared (TIR) band (band 6) of the Landsat 5 TM Collection 1 Tier 1 raw scene was initially acquired at a resolution of 120 m/pixel and resampled to 30 m using three convolutions. The two TIR bands (bands 10 and 11) of the Landsat 8 Collection 1 Tier 1 raw scene were initially acquired at a resolution of 100 m/pixel resolution and resampled to 30 m using three convolutions [53].

In addition, we utilized the China Land Cover Dataset (CLCD), which has a spatial resolution of 30 m, to investigate the relationship between land use change and ecological quality transformation. This dataset, compiled from 335,709 Landsat data on the GEE platform and published by Dr. Xin Huang of Wuhan University, provides yearly land cover information for China (1985,1990–2022) [54] (https://zenodo.org/records/4417810, accessed on 3 October 2023).

2.3. Methodology

2.3.1. Research Process

The overall workflow of this study is shown in Figure 2. Firstly, the ecological evaluation indicators were selected by the Pressure–State–Response framework based on the Landsat data. Then, the RCEI was constructed based on the GEE platform to invert the past ecological and environmental quality changes in Jincheng City from 1990 to 2022. Finally, spatial autocorrelation analysis was conducted to explore the spatial distribution patterns of ecological environmental quality. The relationship between land cover and ecological environmental quality was further examined and discussed.

A novel Resource-Based City Ecological Index (RCEI) was constructed based on a Pressure–State–Response framework. This framework, conceptualized by the Organization for Economic Cooperation and Development and the United Nations Environmental Programme, incorporates the three dimensions of anthropogenic pressure, environmental state, and climate response.

In the early development of coal resource cities, rapid urban expansion centered around coal mining areas. As mining intensified and the workforce grew, population activities related to coal mining increased, leading to population concentration around mining areas. This resulted in the rapid development of production and residential zones, with significant ground hardening, replacing vegetation and causing ecological damage. The Normalized Differential Buildup and Bare Soil Index (NDBSI) is a remote sensing index used to describe surface dryness. It integrates multiple spectral bands to distinguish between bare soil and vegetation cover, reflecting changes in surface dryness and bare soil.

Large-scale mining leads to substantial coal piles on the surface, and coal dust from preprocessing and transport disperses into the air, spreading widely. Early mining areas in resource cities are typically near urban centers, causing coal dust to disperse in populated areas, leading to respiratory diseases and stressing surrounding vegetation, rapidly degrading the ecological environment. Therefore, coal dust pollution’s impact on the ecological environment cannot be ignored. This paper used the Coal Dust Index (ICDI) to describe coal dust’s impact on the ecological environment of resource cities. The NDBSI describes surface aridity caused by urbanization, built-up land expansion, and increased rock or soil exposure, while the ICDI quantifies coal dust pollution from mining. Both indices assess ecological pressure from human activities [55]. Large-scale urban expansion, the establishment of over 100 coal mines, and extensive road construction have encroached on vegetated areas, leading to reduced biodiversity, altered water cycles, soil erosion, and decreased carbon storage. The Normalized Difference Vegetation Index (NDVI) is an effective vegetation index known for its sensitivity, simplicity, wide application, and ease of interpretation, making it widely used in ecological monitoring and assessment. However, the NDVI is limited in quantifying vegetation health. In coal resource cities, large-scale mining and severe coal dust pollution have damaged vegetation health. Therefore, we propose the Enhanced Vegetation Index (EVI) to quantitatively assess vegetation health. This study used the NDVI and EVI to reflect the environmental conditions of resource-based cities [56,57], As urban expansion accelerates and mining areas increase, soil degradation and surface temperature around cities and mines rise, causing ecological damage. The Moisture Index (WET) accurately reflects soil moisture content and assesses uncovered vegetation and bare soil areas. Land surface temperature (LST) accurately reflects urban surface temperature changes, crucial for assessing the urban heat island effect and climate change impact. Resource-based cities are prone to heat island effects due to industrial activities and urbanization. The LST helps identify high-temperature areas, which exert pressure on ecosystems, affecting plant and animal survival. The LST helps identify these areas and formulate mitigation measures, such as increasing green spaces and water bodies to reduce surface temperature. The WET and LST can be integrated with the NDVI and EVI for a comprehensive ecological environmental quality assessment [58,59]. These six components were carefully integrated to formulate the RCEI, as shown in Equation (1):

$$RCEI=f(WET,NDVI,EVI,NDBSI,LST,ICDI)$$

(1)

$$RSEI=f(NDVI,WET,LST,NDBSI)$$

(2)

Before the calculation of the RSEI and RCEI, the water body was masked using the Modified Normalized Water Body Index (mNDWI) to ensure that the WET index accurately represented the humidity of the study area [60]. The mathematical formulations for the aforementioned remote sensing indicators are detailed in

Table 2. The mathematical formulations of each component index of the RCEI.

Index	Equation
WET	${\mathit{W}\mathit{E}\mathit{T}}_{\mathit{T}\mathit{M}}\mathbf{=}\mathbf{0.0315}{\mathit{\rho }}_{\mathit{B}\mathit{l}\mathit{u}\mathit{e}}\mathbf{+}\mathbf{0.2012}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{+}\mathbf{0.3102}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{+}\mathbf{0.1594}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{-}\mathbf{0.6806}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{0.6109}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{2}}$	(5)
	${\mathit{W}\mathit{E}\mathit{T}}_{\mathit{O}\mathit{L}\mathit{I}}\mathbf{=}\mathbf{0.1511}{\mathit{\rho }}_{\mathit{B}\mathit{l}\mathit{u}\mathit{e}}\mathbf{+}\mathbf{0.1973}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{+}\mathbf{0.3283}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{+}\mathbf{0.3407}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{-}\mathbf{0.7117}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{0.4559}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{2}}$	(6)
NDVI	$\mathit{N}\mathit{D}\mathit{V}\mathit{I}\mathbf{=}\left({\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{-}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\right)\mathbf{/}\left({\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{+}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\right)$	(7)
EVI	$\mathbf{N}\mathbf{D}\mathbf{S}\mathbf{V}\mathbf{I}\mathbf{=}\mathbf{(}{\mathsf{\rho }}_{\mathbf{S}\mathbf{W}\mathbf{I}\mathbf{R}\mathbf{1}}\mathbf{-}{\mathsf{\rho }}_{\mathbf{R}\mathbf{e}\mathbf{d}}\mathbf{)}\mathbf{/}\mathbf{(}{\mathsf{\rho }}_{\mathbf{S}\mathbf{W}\mathbf{I}\mathbf{R}\mathbf{1}}\mathbf{+}{\mathsf{\rho }}_{\mathbf{R}\mathbf{e}\mathbf{d}}\mathbf{)}$	(8)
	$\mathit{N}\mathit{R}\mathit{I}\mathbf{=}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{/}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}$	(9)
	${\mathit{N}\mathit{I}}_{\mathit{i}}\mathbf{=}\mathbf{(}{\mathit{I}}_{\mathit{i}}\mathbf{-}{\mathit{I}}_{\mathit{m}\mathit{i}\mathit{n}}\mathbf{)}\mathbf{/}\mathbf{(}{\mathit{I}}_{\mathit{m}\mathit{a}\mathit{x}}\mathbf{-}{\mathit{I}}_{\mathit{m}\mathit{i}\mathit{n}}\mathbf{)}$	(10)
	$\mathit{E}\mathit{V}\mathit{I}\mathbf{=}\mathit{P}\mathit{C}\mathbf{1}\mathbf{(}\mathit{N}\mathit{D}\mathit{V}\mathit{I}\mathbf{,}\mathit{N}\mathit{R}\mathit{I}\mathbf{,}\mathit{N}\mathit{D}\mathit{S}\mathit{V}\mathit{I}\mathbf{)}$	(11)
	$\mathit{E}\mathit{V}\mathit{I}\mathbf{=}\mathbf{(}\mathit{P}\mathit{C}\mathbf{1}\mathbf{-}\mathit{P}\mathit{C}{\mathbf{1}}_{\mathit{m}\mathit{i}\mathit{n}}\mathbf{)}\mathbf{/}\mathbf{(}\mathit{P}\mathit{C}{\mathbf{1}}_{\mathit{m}\mathit{a}\mathit{x}}\mathbf{-}\mathit{P}\mathit{C}{\mathbf{1}}_{\mathit{m}\mathit{i}\mathit{n}}\mathbf{)}$	(12)
NDBSI	$\mathit{I}\mathit{B}\mathit{I}\mathbf{=}{\displaystyle \frac{\mathbf{2}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{+}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{)}\mathbf{-}\left[{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{+}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{)}\mathbf{+}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{+}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{)}\right]}{\mathbf{2}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{+}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{)}\mathbf{+}\left[{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{+}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{)}\mathbf{+}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{G}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{+}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{)}\right]}}$	(13)
	$\mathit{S}\mathit{I}\mathbf{=}\left[\begin{array}{c}\mathbf{(}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{+}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{)}\mathbf{-}\mathbf{(}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{+}{\mathit{\rho }}_{\mathit{B}\mathit{l}\mathit{u}\mathit{e}}\mathbf{)}\end{array}\right]\mathbf{/}\left[\begin{array}{c}\mathbf{(}{\mathit{\rho }}_{\mathit{S}\mathit{W}\mathit{I}\mathit{R}\mathbf{1}}\mathbf{+}{\mathit{\rho }}_{\mathit{R}\mathit{e}\mathit{d}}\mathbf{)}\mathbf{+}\mathbf{(}{\mathit{\rho }}_{\mathit{N}\mathit{I}\mathit{R}}\mathbf{+}{\mathit{\rho }}_{\mathit{B}\mathit{l}\mathit{u}\mathit{e}}\mathbf{)}\end{array}\right]$	(14)
	$\mathit{N}\mathit{D}\mathit{B}\mathit{S}\mathit{I}\mathbf{=}\mathbf{(}\mathit{I}\mathit{B}\mathit{I}\mathbf{+}\mathit{S}\mathit{I}\mathbf{)}\mathbf{/}\mathbf{2}$	(15)
LST	$\mathit{L}\mathit{S}\mathit{T}\mathbf{=}\mathit{T}\mathbf{/}\left[\mathbf{1}\mathbf{+}\mathbf{\left(}{\displaystyle \frac{\mathit{\lambda }\mathit{T}}{\mathbf{1.438}\mathbf{\times }{\mathbf{10}}^{\mathbf{-}\mathbf{2}}}}\mathbf{\right)}\mathit{l}\mathit{n}\mathbf{}\mathit{\epsilon }\right]\mathbf{-}\mathbf{273.15}$	(16)
ICDI	$ICDI={\displaystyle \frac{\{3{\rho }_{Blue}/({\rho }_{Blue}+{\rho }_{Green})-[{\rho }_{Green}/({\rho }_{Blue}+{\rho }_{Green})+{\rho }_{NlR}/({\rho }_{NlR}+{\rho }_{Red})+{\rho }_{Green}/({\rho }_{Green}+{\rho }_{SlvlR1})\left]\right\}}{\{3{\rho }_{SlveR1}/({\rho }_{SlveR1}+{\rho }_{MR})+[{\rho }_{Green}/({\rho }_{Blue}+{\rho }_{Green})+{\rho }_{NlR}/({\rho }_{NlR}+{\rho }_{Red})+{\rho }_{Green}/({\rho }_{Green}+{\rho }_{SlvlR1})\left]\right\}}}$	(17)
mNDWI	$\mathit{m}\mathit{N}\mathit{D}\mathit{W}\mathit{I}\mathbf{=}\mathbf{(}{\mathit{\rho }}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{-}{\mathit{\rho }}_{\mathit{s}\mathit{w}\mathit{i}\mathit{r}\mathbf{1}}\mathbf{)}\mathbf{/}\mathbf{(}{\mathit{\rho }}_{\mathit{g}\mathit{r}\mathit{e}\mathit{e}\mathit{n}}\mathbf{+}{\mathit{\rho }}_{\mathit{s}\mathit{w}\mathit{i}\mathit{r}\mathbf{1}}\mathbf{)}$	(18)

Notes: Red, Blue, Green, NIR, SWIR1, and SWIR2 are the reflectance values of Landsat images in the blue, green, red, near-infrared, short-wave-infrared 1, and short-wave-infrared 2 bands, respectively [62]. In the LST equation, T is the sensor’s thermal value; λ is the central wavelength of the thermal infrared band; and ε is the surface reflectance [63]. In the EVI equation, NI_i is the standardized index value; I_i is the index value at the ith pixel; and I_max and I_min are the maximum and minimum values of this index value, respectively. In order to facilitate the subsequent calculation, after obtaining PC1, it was also standardized to obtain the final EVI value [64].

. These six-component indicators were then integrated through principal component analysis (PCA), and the first principal component (PC1) was selected to formulate the RCEI and RSEI. Given the diverse units and value ranges of each indicator, normalization of the six indicators’ values within the range of [0, 1] was essential prior to PCA analysis to ensure consistency and comparability [61]. This normalization was achieved using GEE feature analysis to compute PC1. Subsequently, based on PC1, the initial RCEI (RCEI₀) and RSEI (RSEI₀) was derived and expressed as outlined in Equations (3) and (4). Finally, using the normalization method to derive the RCEI from the RCEI₀ and the RSEI from the RSEI₀ yielded values in the range [0, 1]. The resulting RCEI values were then graded in accordance with the “Specification for Ecological Condition Evaluation Indices”, with the grading criteria detailed in

Table 3. The standards for classification of ecological environmental conditions based on the RCEI.

Level	Excellent	Good	Moderate	Fair	Poor
Index	0.75–1	0.55–0.75	0.35–0.55	0.2–0.35	0–0.2

.

$$RCE{I}_0=1-PC1\left[f\right(WET,NDVI,EVI,NDBSI,LST,ICDI\left)\right]$$

(3)

$$RSE{I}_0=1-PC1\left[f\right(NDVI,WET,LST,NDBSI\left)\right]$$

(4)

2.3.2. Spatial Autocorrelation Analysis

Spatial autocorrelation is an important indicator for testing whether the ecological environmental quality of an element is related to the ecological environmental quality of its neighboring spaces. Spatial autocorrelation analysis of ecological environmental quality can describe the spatial homogeneity distribution of the ecological environmental quality in a study area. The global spatial autocorrelation index (Global Moran’s I) [65] and the local spatial correlation index (Local Moran’s I) were used to analyze the spatial correlation of the RCEI [66].

The Global Moran’s I (GMI) measures the correlation of the attribute values between neighboring spatial units, with its value ranging from −1 to 1. When the absolute value of Moran’s I approaches 1, it indicates a strong spatial autocorrelation, suggesting that similar values tend to cluster together in space [67,68]. The mathematical formula for the GMI is shown as Equation (19):

$$\mathit{G}\mathit{M}\mathit{I}={\displaystyle \frac{\mathit{n}{\displaystyle {\sum }_{\mathit{i}=1}^{\mathit{n}}}\hspace{1em}{\displaystyle {\sum }_{\mathit{j}=1}^{\mathit{m}}}\hspace{1em}{\mathit{W}}_{\mathit{i}\mathit{j}}({\mathit{x}}_{\mathit{i}}-\stackrel{\mathit{‾}}{\mathit{x}})({\mathit{x}}_{\mathit{j}}-\stackrel{\mathit{‾}}{\mathit{x}})}{\left({\displaystyle {\sum }_{\mathit{i}=1}^{\mathit{n}}}\hspace{1em}{\displaystyle {\sum }_{\mathit{j}=1}^{\mathit{m}}}\hspace{1em}{\mathit{W}}_{\mathit{i}\mathit{j}}\right){\displaystyle {\sum }_{\mathit{i}=1}^{\mathit{n}}}\hspace{1em}({\mathit{x}}_{\mathit{i}}-\stackrel{\mathit{‾}}{\mathit{x}}{)}^2}}$$

(19)

where n is the total number of elements, x_i denotes the ecological quality value of the i position, respectively, $\overline{x}$ denotes the average value of the ecological quality of all elements in the study area, and W_ij is the spatial weight.

The Local Moran’s I (LMI) can effectively capture the correlation between the ecological quality of the grid cells within a study area. Therefore, local spatial autocorrelation analysis is important. In the absence of global spatial autocorrelation, the LMI enables the identification of potentially masked locations exhibiting local spatial autocorrelation. Conversely, when global spatial autocorrelation is present, the LMI facilitates the analysis of spatial heterogeneity. Specifically, the LMI computes the local spatial aggregation characteristics of the Regional Comprehensive Ecological Index (RCEI) and visualizes these characteristics in the local spatial context through the Local Spatial Association Indicator (LISA). The mathematical formula for calculating the LMI is shown as Equation (20):

$$\mathit{L}\mathit{M}\mathit{I}={\displaystyle \frac{\mathit{n}({\mathit{x}}_{\mathit{i}}-\stackrel{\mathit{‾}}{\mathit{x}}){\displaystyle {\sum }_{\mathit{j}=1}^{\mathit{m}}}\hspace{1em}{\mathit{W}}_{\mathit{i}\mathit{j}}({\mathit{x}}_{\mathit{j}}-\stackrel{\mathit{‾}}{\mathit{x}})}{{\displaystyle {\sum }_{\mathit{i}=1}^{\mathit{m}}}\hspace{1em}({\mathit{x}}_{\mathit{i}}-\stackrel{\mathit{‾}}{\mathit{x}}{)}^2}}$$

(20)

In this study, the GMI and LMI indices were calculated using Geo Da software (v.1.22) to assess spatial autocorrelation. Five distinct spatial aggregation patterns were identified as the following: high–high clustering (H–H), low–low clustering (L–L), low–high outlier (L–H), high–low outlier (H–L), and insignificant. Specifically, H–H indicates that both the selected area and its neighboring areas exhibit high ecological quality values. Conversely, L–L indicates low ecological quality values for both the selected area and its neighboring spaces. L–H denotes that the ecological quality of the selected area is low while its neighboring areas possess high ecological quality. In contrast, H–L implies that the selected area has high ecological quality while its neighboring spaces exhibit low ecological quality values. These spatial patterns provide insights into the spatial heterogeneity and dependencies of ecological quality within the study area.

3. Results

3.1. Comparative Analysis of the Remote Sensing Ecological Index RSEI and Resource-Based City Ecological Index RCEI

This study constructed the Remote Sensing Ecological Index (RSEI) and the Resource Cities Ecological Index (RCEI) using principal component analysis (PCA) with the assistance of the Remote Sensing Data Processing Platform. To demonstrate that the RCEI was superior to the RSEI in assessing the ecological quality of resource cities, this study conducted ecological calculations and assessments for both indices in Jincheng City using the Remote Sensing Data Processing Platform. A comparison of the PCA results indicated that the RCEI provided a more comprehensive representation of ecological quality than the RSEI (

Table 4. Comparison of the RCEI and RSEI principal component analysis results.

Date	Results of PCA
	Eigenvalue		Contribution Rate (%)
	RSEI	RCEI	RSEI	RCEI
1990	0.0837	0.1317	70.92	88.62
1996	0.0786	0.1459	70.05	81.07
2002	0.0354	0.0876	60.45	85.35
2006	0.0491	0.0924	79.42	84.31
2008	0.0783	0.1045	68.23	73.13
2013	0.0786	0.1407	80.32	89.36
2018	0.1127	0.1815	73.81	79.51
2022	0.0882	0.1367	79.69	87.22

).

Upon careful comparison of the PCA results of the RSEI and RCEI models, several key differences emerged, as illustrated in Figure 3a and Table 4. Firstly, the first principal component eigenvalues of the RCEI model ranged from 0.0876 to 0.1815, significantly higher than the range of 0.0354 to 0.1127 observed in the RSEI model. This indicates that the RCEI model was more effective in capturing the primary variability in the data. Notably, the smallest first principal component eigenvalues of the RCEI model were just below those of the RSEI model for the years 2018 and 2022.

Further analysis of the contribution rate of the first principal component revealed significant differences between the RSEI and RCEI models. The contribution rate of the RSEI model fluctuated considerably, ranging from 60.45% to 80.32%, with an extreme deviation of 19.87%, as shown in Figure 3b. In contrast, the RCEI model demonstrated greater stability, with contribution rates varying from 73.13% to 89.36% and an extreme variance of only 16.23%. This smaller variance indicates that the RCEI model performed more consistently across different years. Moreover, when comparing the first principal component contributions of the two models for corresponding years, the RCEI model consistently showed higher contributions than the RSEI model. This finding further underscores the superiority of the RCEI model in characterizing ecological quality.

3.2. Comprehensive Evaluation of Ecological Environmental Quality in Jincheng City

As shown in

Table 5. Eigenvalues of six discrete ecological indices in the principal component analysis.

Date	WET	NDVI	EVI	NDBSI	LST	DI
1990	0.3651	0.5251	0.3560	−0.5081	−0.3017	−0.3388
1996	0.3573	0.4635	0.0166	−0.5468	−0.4640	−0.3781
2002	0.2704	0.5437	0.2834	−0.5907	−0.3519	−0.3891
2006	0.3860	0.5092	0.0159	−0.5498	−0.4185	−0.3377
2008	0.4080	0.5462	0.0181	−0.6164	−0.2300	−0.3197
2013	0.3461	0.4953	0.3898	−0.5347	−0.3210	−0.3062
2018	0.2951	0.4992	0.3670	−0.5438	−0.3872	−0.2883
2022	0.4631	0.4847	0.3706	−0.4958	−0.2840	−0.2943

, the eigenvalues of the six individual ecological indicators in PC1 reflect the variance contribution of each ecological index in the dataset through principal component analysis (PCA). Specifically, the larger the eigenvalue, the more variance is explained by that principal component. Therefore, these eigenvalues can be used to assess the importance of each principal component in a composite index. In contrast, the eigenvalues of PC2 to PC6 exhibited anomalies in both magnitude and sign. Notably, in PC1, the eigenvalues of the WET, NDVI, and EVI were positive, while those of the NDBSI, LST, and IDCI were negative. As illustrated in Figure 4 and Table 5, this distribution aligns with the observed environmental conditions.

The principal component analysis from 1990 to 2022 revealed a positive correlation between the first principal component and key ecological indicators: the NDVI, WET and EVI. The NDVI’s substantial contribution to the PC1 loading values underscores the critical role of vegetation cover in the RCEI. Concurrently, increases in the WET reflect heightened soil and vegetation moisture content, while enhancements in the EVI denote improved vegetation health. These trends collectively affirm the positive influence of these indicators on the study area’s ecology. In contrast, the NDBSI, LST, and ICDI demonstrated a negative correlation with PC1 load values. Rising NDBSI values suggest escalating desiccation of surface soils and vegetation, elevated LST levels imply an intensifying greenhouse effect, and increased ICDI values signal detrimental coal dust impacts on air quality. These indicators’ trajectories highlight their adverse effects on the ecological environment. In essence, the NDVI, WET, and EVI are pivotal in monitoring and evaluating regional ecosystem health, while the NDBSI, LST, and ICDI offer vital insights into ecological degradation.

Figure 5 depicts the distribution of the RCEI values across eight temporal snapshots spanning the period from 1990 to 2022 in Jincheng City. Remarkably, during the years 2006, 2008, 2013, and 2022, a subset of RCEI values exceeded the 0.8 threshold, indicating robust ecological conditions. Furthermore, the lower quartile consistently remained above 0.4, while the upper quartile exceeded the 0.6 mark, indicating a generally healthy ecological state. Conversely, the years 1990, 1996, 2002, and 2018 did not exhibit any RCEI values above 0.8, reflecting periods of ecological stress. A longitudinal analysis revealed a discernible degradation in Jincheng City’s ecological quality from 1990 to 2002. However, a subsequent rebound was observed from 2006 to 2022, with the exception of a brief dip in 2018, indicating a progressive improvement in the overall ecological environment.

Figure 6 illustrates the spatiotemporal progression of ecological environmental quality in Jincheng City from 1990 to 2022. A color gradient ranging from red to dark green demarcates five tiers of ecological quality, ascending from low to high. The analysis indicated that, throughout this period, Jincheng City’s ecological status predominantly sustained an upper–middle level. At the county and district levels, areas of diminished ecological quality were principally clustered in regions experiencing rapid urban development, such as the central urban locales of Jincheng City, Gaoping City, and the county hubs of Zezhou, Yangcheng, Lingchuan, and Qinshui. These sectors are distinguished not only by extensive construction and increased human activity but also by a large number of coal mines, exceeding a hundred in total and varying in scale. Notably, the ecological condition of major coal production sites, including the Sihe, Chengzhuang, and Fenghuangshan mines, was evaluated as low. Conversely, areas of elevated ecological quality were primarily located in the city’s peripheral wetland reserves and forested areas, characterized by their annular layout, higher altitudes, and robust forest coverage.

Figure 7 illustrates the dynamics of the mean value of the regional ecological environmental quality index (RCEI) in each district and county of Jincheng City (including urban areas, Zizhou County, Yangcheng County, Qinshui County, Gaoping County, and Lingchuan County) from 1990 to 2022. Through comparative analysis, we found that the average values of the RCEI in Gaoping County and urban areas were significantly lower than those of the other four counties, reflecting the relatively poor ecological quality of these two regions. Specifically, the mean value of the RCEI in urban areas fluctuated between 0.4612 and 0.5884, indicating that their ecological environmental quality was at a medium level, while the mean value of the RCEI in Gaoping County fluctuated between 0.5244 and 0.6454, indicating that its ecological environmental quality was a little better at an intermediate–upper level. The analysis of the coal mine distribution map revealed that the urban areas were not only the core area for urban development, but also the key area for large-scale mining operations, especially the five major million-ton coal mines mainly concentrated in these two jurisdictions. This phenomenon highlights the far-reaching impact of urban expansion and concentrated mining activities on the ecological structure of Jincheng City.

The average RCEI values for Yangcheng County and Zezhou County fluctuated between 0.6266 and 0.7429 and 0.5566 and 0.7029, respectively, showing a continuously increasing trend. Over the past 32 years, the average RCEI values of these two counties have increased by 0.1163 and 0.1463, respectively. The map of mine distribution shows that the distribution of coal mines in the Zezhou and Yangcheng counties showed obvious agglomeration, which was mainly concentrated in the flat terrain in the center of Jincheng City. At the same time, these two counties were surrounded by a large number of forested areas, which reflects the fact that although the average RCEI values of the Zezhou and Yangcheng counties were classed as good, mine development and urban construction still had significant impacts on the ecological environment. Zezhou County, which borders urban areas, had a higher urban area, population size, and degree of industrialization than Yangcheng County. This further indicates that the ecological quality of Zezhou County was lower than that of Yangcheng County, and its ecological improvement was smaller, which is consistent with the actual observation.

The mean RCEI values of Qinshui County and Lingchuan County fluctuated between 0.6684 and 0.7958 and 0.6947 and 0.8083, respectively, showing high ecological quality. The mining distribution maps showed that these two counties and districts had fewer coal mines and relatively lower levels of urban construction, so their mean RCEI values reached good levels, especially for Lingchuan County, where the highest mean RCEI value reached the excellent level. This finding provides a valuable comparative perspective for exploring the ecological quality of resource cities and demonstrates the effective applicability of RCEI indicators in resource cities.

This study examined the ecological trajectory of Jincheng City across a span of 32 years, tracing the spatial dynamics of ecological quality through eight pivotal years from 1990 to 2022, as depicted in

Table 6. Proportions of areas graded on their ecological quality from 1990 to 2022.

Year	Form	Poor	Fair	Moderate	Good	Excellent	Total
1990	Area (km2)	667.42	4126.9	2354.36	1772.38	568.94	9490
	Pct. (%)	7.03	43.48	24.81	18.66	6.02	100
1996	Area (km2)	297.98	2269.12	3440.82	2477.06	1005.02	9490
	Pct. (%)	3.14	23.91	36.25	26.11	10.59	100
2002	Area (km2)	595.35	3433.23	2367.28	2183.18	910.96	9490
	Pct. (%)	6.27	36.17	24.94	23.01	9.61	100
2006	Area (km2)	195.09	2156.05	3063.2	2334.11	1741.55	9490
	Pct. (%)	2.05	22.71	32.27	24.59	18.38	100
2008	Area (km2)	312.61	1103.77	2210.31	3323.15	2540.86	9490
	Pct. (%)	3.28	11.63	23.29	35.02	26.78	100
2013	Area (km2)	422.19	1214.65	2986.5	3537.3	1329.36	9490
	Pct. (%)	4.44	12.80	31.46	37.27	14.03	100
2018	Area (km2)	782.39	1606.02	3147.34	3239.36	714.29	9490
	Pct. (%)	8.24	16.92	33.17	34.13	7.54	100
2022	Area (km2)	341.88	839.9	1767.02	4067.83	2473.37	9490
	Pct. (%)	3.60	8.85	18.61	42.86	26.08	100

and Figure 8. In 1990, a substantial portion of the city, comprising approximately 6481.26 km² or 68.29% of its total area, was classified under poor and average ecological conditions. These regions were primarily located in the mining and exploitation zones central to Jincheng City. In stark contrast, areas of favorable ecological quality were confined to the peripheral mountainous forests and grasslands. The data from 1990 unequivocally indicate a prevalent state of ecological degradation across the majority of the cityscape.

This study examined the ecological trajectory of Jincheng City over a period of 32 years, tracing the spatial dynamics of its ecological quality through eight key years from 1990 to 2022, as shown in Table 6 and Figure 8. Figure 8 clearly illustrates the flow of ecological change within the study area during these selected years. The right section of Figure 8 displays the total area of the study region, which was 9490 km² for the years 1990, 1996, 2002, 2006, 2008, 2013, 2018, and 2022. The left section categorizes the RCEI index into five classes: poor, fair, moderate, good, and excellent. Notably, the area classified as “good” was the largest among the five categories, marked as 22,934 km² in Figure 8, while the area classified as “poor” was the smallest, at 3615 km². This indicates a positive trend in the spatial and temporal changes in the ecological quality of the study area.

By 2008, the area of regions with good ecological quality exceeded that of those of moderate ecological quality for the first time, while the area of regions with poor and poorer ecological quality decreased annually. In 2018, the area of good ecological quality regions declined after reaching a peak in 2008, but it rapidly recovered and exceeded 2008 levels to become the dominant ecological class by 2022. Additionally, the area of good ecological quality regions peaked between 1990 and 2022. This shift was primarily attributed to the urban expansion in Jincheng City between 2008 and 2018, as well as the commissioning of numerous small- and medium-sized coal mines in Gaoping County and Qinshui County, which significantly reduced the ecological level of Jincheng City.

By 2022, the areas with good and excellent ecological quality underwent a rapid transformation compared to 2018. This was primarily due to the cessation of production in three large coal mines, the Fenghuangshan Coal Mine [69], the Wangtai Coal Mine, and the Gu Mine Coal Mine, all with an annual output of 10 million tons, as well as several small coal mines in the surrounding urban area of Jincheng City. This led to the improvement in the ecological environment in the urban area of Jincheng City and Zezhou County. Furthermore, as a resource-oriented city, Jincheng City continued to promote its transformation and development, increasing its investment in and protection of the ecological environment [70]. Concurrently, areas with originally good ecological quality were transformed into having excellent grades, particularly through the restoration of forests, grasslands in the suburbs, and the ecological vegetation in the Baima Temple mountainous area. The quality of the ecological environment was significantly improved. This transformation not only reflects Jincheng’s efforts in ecological environmental protection, but also the effectiveness of its ecological restoration and sustainable development strategies.

3.3. Dynamic Change in Ecological Environmental Quality in Jincheng City

Remote sensing change monitoring serves as an important tool for analyzing the spatial and temporal variations in ecological environmental quality over different periods.

Table 7. Monitoring changes in the level of the remote sensing ecological index in resource-based cities in 1990–2022.

Year	Form	−3	−2	−1	0	1	2	3	Total
1990 to 1996	Area (km2)	77.98	72.47	123.22	1466.88	5898.76	1608.46	242.23	9490
	Pct. (%)	0.8	0.8	1.3	15.5	62.2	16.8	2.6	100
1996 to 2002	Area (km2)	111.25	455.95	1258.65	3871.34	3177.99	484.97	129.85	9490
	Pct. (%)	1.2	4.8	13.3	40.7	33.5	5.1	1.4	100
2002 to 2006	Area (km2)	50.34	45.71	55.74	1059.37	7120.34	1077.85	80.65	9490
	Pct. (%)	0.5	0.5	0.6	11.2	75	11.4	0.8	100
2006 to 2008	Area (km2)	73.08	85.14	109.15	2469.43	5319.74	1177.79	255.67	9490
	Pct. (%)	0.8	0.9	1.2	26	56.1	12.3	2.7	100
2008 to 2013	Area (km2)	84.98	136.45	1002.46	4969.59	3037.56	151.78	107.18	9490
	Pct. (%)	0.9	1.4	10.6	52.4	32	1.6	1.1	100
2013 to 2018	Area (km2)	81.89	102.54	736.13	6543.48	1799.06	156.55	70.35	9490
	Pct. (%)	0.9	1.1	7.7	69	19	1.6	0.7	100
2018 to 2022	Area (km2)	75.31	129.28	112.01	798.59	5746.48	2367.69	260.64	9490
	Pct. (%)	0.8	1.4	1.2	8.4	60.6	24.9	2.7	100
1990 to 2022	Area (km2)	82.73	111.83	245.75	849.63	3032.64	3657.09	1510.33	9490
	Pct. (%)	0.9	1.2	2.6	9	32	38.5	15.8	100

and Figure 9 depict the dynamic changes in the RCEI in Jincheng City. The results indicate that from 1990 to 2022, the ecological environment of Jincheng City generally exhibited a trend of degradation followed by improvement, with 1996 and 2022 being the years with the most significant ecological changes.

During 1990–1996, 81.6% of the regional ecological environmental quality in Jincheng City improved, primarily concentrated in the urban built-up area, the built-up area of Zezhou County, and the coal mining regions of Fenghuangshan, Ancient Mining, and Chengzhuang, which are located in the southern center of Jincheng City and displayed a concentrated area of improvement. From 1996 to 2002, the area of improved ecological quality decreased by 40%, while the total area of ecological degradation increased by 16.4%. From 2002 to 2006, the ecological quality improved significantly, and the percentage of ecologically restored area recovered to the level of 1990–1996, while the degraded area accounted for only 1.6%. From 2008 to 2013 and 2013 to 2018, the area of ecological improvement consecutively decreased, and the degraded area continued to rise, primarily concentrated in the northeastern part of Jincheng City. Feature classification showed that most of the ecologically degraded area comprised cropland and fallow land, as well as small- and medium-sized coal mine areas in and around the urban built-up area of Gaoping County. During 2013–2018, the area with no ecological change (level 0) accounted for the highest proportion of ecological change over the five-year period, reaching 69%.

Between 2018 and 2022, Jincheng City witnessed a significant increase in ecological environmental quality, with 60.6% of the study area exhibiting mild improvement (level 1) and 2.7% achieving significant enhancement. The implementation of comprehensive ecological policies catalyzed substantial improvements in the ecological quality of urban built-up areas, coal mine production zones, and residential sectors across Jincheng City and its five counties and districts, attaining a significant level of improvement (+2 level). Conversely, the remaining 3.4% of areas experiencing ecological decline were primarily situated in Zezhou County, Yangcheng County, and Qinshui County, coinciding with the operational commencement of new coal mines.

Over the seven analyzed periods, the proportion of areas undergoing ecological improvement exhibited a W-shaped pattern of fluctuation, while the proportion of degraded areas presented a contrasting M-shaped trend. The share of areas remaining unchanged was more volatile, displaying an inverted V shape, and experienced a sharp decline post-peak during the 2013–2022 interval, with variations ranging from 8.4% to 69%. Specifically, these unchanged areas were predominantly situated between degradation zones (depicted as red zones in Figure 9) and buffer zones surrounding improved green patches. They were centered around urban built-up locales, agricultural lands associated with coal mine production and habitation, fallow lands, and grassland regions. Throughout the entire study period (1990–2022), the proportions of improved, unchanged, and degraded areas were 86.3%, 4.7%, and 9%, respectively. Notably, areas that underwent significant degradation were mainly coal mining and built-up areas prior to 2018. Post-2018, the overall ecological quality showed marked improvement following the decommissioning of certain built-up and coal mining areas proximal to the city. In summary, the ecological quality of Jincheng City has displayed a pronounced upward trend.

3.4. Analysis of Ecological Environmental Quality and Land Use Change in Jincheng City

The land use dynamics of Jincheng City from 1990 to 2022, as depicted in Figure 10, revealed substantial transformations. Data presented in Figure 10 indicate a decrease in cultivated land from 40.83% to 32.36% of the total area, while fallow forest areas expanded to 804.2 km². Concurrently, grassland areas decreased from 20.11% to 12.22%. A discernible trend is evident in Figure 10: grasslands at higher elevations around Jincheng City progressively transformed into forested lands, which is a change that significantly elevated the RCEI value in the Qingshui, Lingchuan, Yangcheng, and Zezhou counties compared to urban areas and Gaoping County. The forested land area notably increased from 36.76% in 1990 to 49.89% in 2022, now constituting nearly half of Jincheng City’s total area, with a net gain of 1246.1 km², primarily attributed to the conversion from grassland and cropland. Among the land use types of cropland, grassland, and woodland, only woodland experienced positive growth, indicative of extensive land use conversion. Urban built-up areas expanded from 1.45% to 4.46%, reflecting rapid urbanization and indicating an increase in coal production activities. Water-filled areas more than doubled since 1990, signifying Jincheng City’s significant strides in water resource conservation and environmental enhancement, thereby markedly improving the ecological environment’s quality. The analysis of Figure 5, Figure 6 and Figure 9 showed an improvement in the mean RCEI value within the urban area in 2022, despite urban sprawl. This positive shift, coupled with the distribution of coal mines and the decommissioning of major coal mines—Wangtaipu, Fenghuangshan, and Guming—underscores the impact of urban development and coal mining on the ecological environment of Jincheng City. These observations affirm the efficacy of the RCEI as a robust indicator for intuitively capturing the ecological quality of resource-based cities.

4. Discussion

4.1. Evaluation Based on RCEI Sample Images

In Figure 11, the ecological environmental quality of Jincheng City in 2022 is presented through six carefully selected sample points covering the urban areas, coal mine production area, mine living area, national forest park, and national ecological wetland area. The mean RCEI values of these sample points (0.9089, 0.3874, 0.3863, 0.1363, 0.6711, and 0.8688, respectively) were calibrated in Figure 11a,b based on the RCEI images and the RGB satellite images. Parts c through h of Figure 11 show in detail the RGB imagery of these areas, corresponding to (c) Lishan National Forest Area—Qinshui County; (d) Yicheng Coal Mine—Yangcheng County; (e) Sihe Coal Mine—Qinshui County; (f) Gu Mine—Urban Area; (g) Fenghuangling Forest Wetland—Zezhou County; and (h) Chessing Mountain National Forest Wetland—Lingchuan County. These images cover the details of the five counties of Jincheng City, ensuring a broad and diverse sample, thus eliminating the possibility of interactions. The analytical results in Figure 11 confirmed the close connection between the Resource-Based City Ecological Environment Index (RCEI) and the distribution of mining areas, urban development, and forest ecosystems, and they validated the RCEI as a reliable indicator of ecological change in resource-based cities.

4.2. Analysis of the Correlation between Component Indicators and RCEI

To ensure the integrity of information and the precision of quantitative assessments within the specified scale, this study integrated the intrinsic attributes of the study area with its classification patterns and ecosystem characteristics. Consequently, image data were resampled to a spatial resolution of 1 km × 1 km. A total of 3043 sample points were extracted from RCEI imagery spanning from 1990 to 2022 to ascertain the spatial correlation of variables and their magnitudes. Moreover, this research examined the interplay between positive vegetation indices (NDVI, EVI), hydric and xeric indicators (WET, NDBSI), and adverse indicators (surface LST and coal dust index ICDI) in relation to the RCEI through three-dimensional spatial projections, as depicted in Figure 12.

Figure 12a revealed a direct correlation between the NDVI, EVI, and RCEI, where higher index values corresponded to elevated RCEI scores, indicating the effectiveness of these vegetation indices in reflecting ecological environmental quality. Conversely, Figure 12b demonstrates the contrasting impacts of the hydric indicator WET and the xeric indicator NDBSI on the RCEI, with increased RCEI values aligning with higher WET and lower NDBSI values, thus indicating a positive association with ecological quality for the WET and a negative one for the NDBSI. Figure 12c illustrates the inverse relationship of the LST and ICDI with the RCEI, where higher values of these indicators were associated with reduced ecological quality.

As shown in Figure 13, the RCEI index showed a significant correlation with several ecological indicators, indicating its effectiveness in synthesizing these indicators. As revealed in Figure 13, the RCEI showed positive correlation with the NDVI, EVI, and WET indicators and negative correlation with the NDBSI, ICDI, and LST indicators. By comparing the six-year average correlation coefficients, we found that the NDBSI had the highest correlation coefficient among the six single indicators, which reached 0.7. However, the six-year average correlation coefficient of the RCEI was 0.765, which exceeded that of the NDBSI by 0.065. The average correlation coefficients of the RCEI were higher than those of any single ecological environmental indicator during the period from 1990 to 2022, which further confirmed the importance of the RCEI in assessing the ecological quality of resource-based cities.

4.3. Spatial Autocorrelation Analysis of Ecosystem Quality

To guarantee the robustness and comprehensiveness of the analytical outcomes, this study integrated the unique attributes of the study area and resampled the imagery to a spatial resolution of 300 m × 300 m. This study collected 129,075 sample points, which supported the spatial autocorrelation analysis of Jincheng City’s RCEI utilizing Moran’s I and the local indicators of spatial association. Figure 14 delineates the overall spatial autocorrelation of the RCEI within Jincheng City, with a preponderance of data points clustering in the first and third quadrants. This distribution indicated a positive spatial autocorrelation, predominantly characterized by high–high (H–H) and low–low (L–L) clustering patterns. The trajectory of Moran’s I generally followed a pattern of initial fluctuation, subsequent consolidation, and eventual fluctuation, with notable perturbations in 2002 and 2008. In 2006, Moran’s I peaked at 0.95796, reflecting a pronounced spatial agglomeration of the RCEI. Conversely, by 2022, the index had declined to its lowest point at 0.90062, suggesting a decline in the RCEI’s spatial clustering and an emergent fragmentation of ecological patches. Over the period from 1990 to 2022, Moran’s I values (0.9349, 0.9459, 0.9574, 0.9579, 0.9202, 0.9249, 0.9168, and 0.9006) reinforced the inference that the spatial distribution of the ecological environmental quality in Jincheng City was clustered, not random. The evolution of Moran’s I aligns with the trajectory of ecological environmental quality levels.

In the present study, local indicators of spatial association clustering diagrams were utilized to delineate the spatial patterns of Jincheng City’s RCEI from 1990 to 2022 (Figure 15). In 1990, low–low (L–L) clusters primarily occupied the central region of Jincheng City, including the adjacent coal mining areas of Fenghuangshan, Wangtai, and Gushuyuan; the urbanized sectors of Yangcheng and Qingshui Counties, along with the mining zones of Sihe and Duanshi; and the developed areas and mining clusters of Gaoping City. By 2002, the expanse of the L–L clusters had significantly diminished, primarily retreating to the eastern precincts of Jincheng City. Concurrently, the L–L clusters in Qingshui County expanded in an east–west trajectory toward Gaoping County, with the L–L clusters near the Sihe coal mine in southern Yangcheng County transitioning into high–high (H–H) clusters.

Figure 15 illustrates that by 2006, the L–L clusters had intensified in concentration and spatial extent, surpassing those observed in 1990, 1996, and 2002. However, by 2008, the domain of the L–L clusters commenced a discernible contraction, centralizing around the urban core of Jincheng City and the Gaoping District, while the H–H clusters witnessed a progressive annual expansion. The enactment of pivotal governmental policies and regulations, such as the revised Environmental Protection Law (2014) and the Overall Program for Reform of the Ecological Civilization System (2015), catalyzed the momentum toward sustainable development, reinforcing environmental stewardship and ecological conservation. A comparative analysis of the LISA clustering diagrams and the RCEI change detection maps from 1990 to 2022 revealed a general alignment between the H–H clusters and ecological amelioration, and conversely, the L–L clusters with environmental degradation. These findings corroborate prior research, affirming the nonrandom, clustered nature of the RCEI’s spatial distribution. This evidence further elucidates the impact of coal mining and urban sprawl in fostering L–L clusters, while ecological rehabilitation initiatives, such as the targeted development, preservation, and augmentation of the Baima Temple Ecological Park within urban precincts, along with the sequential decommissioning of coal mines like Fenghuangshan, Wangtaipu, and Guming within the Jincheng metropolitan area, facilitated the proliferation of H–H clusters, thereby contributing to the enhancement of Jincheng City’s ecological milieu.

4.4. Buffer Analysis of a Typical Mine Sites

As shown in Figure 16 the ecological impacts of coal mining and urban development were assessed in the Sihe Mine Area in Qingshui County, Jincheng City, and the Fenghuangshan Mine Area in the urban area. By analyzing Landsat imagery from 1990 and 2002, buffer analysis was used to explore changes in the mean RCEI value per 0.5 km within a range of 0 to 8 km.

The results revealed significant negative impacts of coal mining activities on the ecosystem. In the Sihe mine, a notable decrease in the mean RCEI within 0.5 to 1.5 km of the mine center was observed, dropping from 0.5791 in 1990 to 0.3476 in 2002. This decline is attributed to construction and mining activities, which increased coal dust pollution, expanded built-up areas, hardened land surfaces, reduced vegetation cover, and deteriorated vegetation health. Since its inception in 1996, the Sihe mine has produced 15 million tons of coal annually. The surrounding small-scale villages had relatively little ecological impact, making it easier to visualize the effects of mining activities on the surrounding environment. Although the Sihe mine had not yet been mined in 1990, the area’s ecological environment was already moderately low due to neglect and local farming activities. The average RCEI value fluctuated within the 0.5 to 8 km range. Since 2002, as the mine entered the production phase, the mean RCEI value in the center dropped sharply. However, beyond 2 km, the mean RCEI value increased due to the government’s 1999 policy of returning farmland to forest.

In the Fenghuangshan Mine Area, the mean RCEI values within 0.5 to 1 km were relatively low but still higher than those in the Sihe Mine Area by more than 0.1 percentage points. This difference is due to the lower production at the Fenghuangshan mine (about 4 million tons) compared to the Sihe mine (15 million tons), reflecting the differential ecological impacts of coal production levels. The change in mean RCEI values for the Fenghuangshan mine buffer zone indicated significant ecological quality changes from 1990 to 2022. This change was directly linked to the cessation of mining in 2018 and the mine’s closure between 2019 and 2020. With the cessation of production and closure, population activity decreased significantly, and urban sprawl was curbed. Government-supported ecological restoration efforts significantly improved the mine site’s ecological quality, as evidenced by the increased average RCEI value. Beyond 1.5 km, the mean RCEI value showed a rapidly increasing trend, attributed to the improved ecological environment of the Baima Temple forest wetland on the west side of the mine area. However, extending the distance to 3 to 7 km into the urban area, the destructive impacts of urban activities on the ecological environment caused the mean RCEI value to decline stepwise until it reached its lowest point at the center of the 8 km urban area. Within the 7 to 8 km range, the mean RCEI value increased slightly as some areas entered the forested zone.

The results of the buffer zone analyses of the two typical mines in the study area showed that the mining cycle of coal mines had different impacts on the surrounding ecosystem. These impacts varied with the distance from the buffer zone. Exploring the impacts of mines on the surrounding ecological environment at different ranges and with different mining cycles can provide guidance and a theoretical basis for coal mine siting, coal mining, and environmental protection in resource-based cities in the future.

Specifically, coal mining activities can lead to problems such as land hardening, reduced vegetation cover, and increased coal dust pollution, thus negatively affecting an ecosystem. The spatial distribution and temporal changes in these impacts can be quantified through buffer zone analysis. For example, in the Sihe Mine Area, the ecological quality of the central area declined significantly with mining activities, while in the area away from the mine area, the ecological quality improved as a result of the government-implemented policy of returning farmland to forest. In addition, there were differences in the ecological impacts of different mine areas. For example, the ecological impact of the Fenghuangshan mine was relatively small due to its low production, but ecological restoration measures significantly improved the ecological quality of the area as the mine ceased production and was closed. This suggests that the scale of production and management measures at the mines played an important role in their ecological impact.

Through these analyses, certain beneficial steps were determined as the following: (1) When selecting a new mine area, its potential impact on the surrounding ecological environment should be fully considered, and priority should be given to selecting areas of lower ecological sensitivity. (2) In the process of the planning and construction of mine areas, measures should be taken to reduce land hardening and vegetation destruction and to minimize negative impacts on the ecological environment. (3) After the suspension or closure of mine areas, ecological restoration work, such as planting trees and restoring wetlands, should be carried out in a timely manner in order to improve the quality of the ecological environment. (4) Long-term ecological real-time monitoring and restoration efforts play a crucial role in protecting the ecology of mine areas.

The results of this study emphasized the efficiency and practical value of the RCEI in assessing the quality of an ecological environment. The construction of the RCEI was based on selected ecological indicators that not only accurately reflected the physical characteristics of the surfaces in resource-based cities, but could also be efficiently computed by remote sensing technology. Therefore, the RCEI provides a convenient and rapid means for evaluating the ecological environmental status of resource cities. In addition, the RCEI supports the visualization and analysis of spatial and temporal dynamics of ecological environmental quality. Compared with traditional EI indicators, the RCEI is easier to popularize and apply. It is especially worth mentioning that the RCEI, as a comprehensive evaluation tool based on the PSR model, is more comprehensive than most existing indicators in considering the impacts of production pressure in mine areas and urban construction and expansion. Therefore, the RCEI not only guides the standardization of coal mining activities and urban construction, but also reduces irrational development in the development of resource-based cities, which provides important decision-making support for improving and upgrading the quality of the ecological environment.

4.5. Limitations and Future Work

In this study, the spatial and temporal changes in the ecological environmental quality of Jincheng City from 1990 to 2022 were evaluated using the Resource City Ecological Index (RCEI). The results provided a scientific basis and guidance for the planning and development of coal mine areas, urban expansion, and ecological restoration in resource-based cities, particularly those reliant on coal. However, this study had some limitations:

• Limitations of Remote Sensing Imagery—The establishment of the RCEI was entirely based on remote sensing imagery, which is inevitably affected by cloud cover, resulting in missing images for some months. Although interpolation and multi-source data fusion techniques can supplement these images, they are not original data and may introduce small errors in subsequent operations. While these errors were within acceptable limits, they still imposed certain limitations on the research results.
• Complexity of the Ecological Environment—The ecological environment of resource-based cities is complex and variable, with frequent human activities and intricate surface biophysical features. The RCEI consists of six ecological indicators, including the Indexed Coal Dust Index (ICDI) and the Enhanced Vegetation Index (EVI), which were added to represent the impact of coal mining on the atmosphere and vegetation establishment, respectively. However, the lack of socioeconomic factors (e.g., population density, economic activities, and policy changes) in these indicators limits their comprehensive description of resource-based cities.
• Limitations of Model Validation—Although this study validated the reliability of the RCEI through the spatiotemporal correlation, coefficient correlation, and buffer zone analysis of two typical coal mines, complex large-scale coal mining and rapid urban expansion raised questions about whether the RCEI could adequately describe the temporal and spatial changes in the ecological environment of resource-based cities. Further research and validation are needed.
• Deficiencies in the Analysis of Driving Factors—While we obtained data on the spatial and temporal changes in the ecological environmental quality of Jincheng City from 1990 to 2022, we did not deeply analyze the driving factors, such as climatic, anthropogenic, and topographical influences.

Future work should focus on the following:

• In future studies, we plan to incorporate more socioeconomic factors (e.g., population density, economic activities, and policy changes) and coal mining data that are accurate in terms of working surfaces into the Resource City Ecological Index (RCEI) to more comprehensively assess spatial and temporal changes in ecological quality in resource cities. In addition, we plan to place ground-based field monitoring equipment, such as ecosticks, in different ecological change areas to obtain field-based ecological change data. These data will be processed and analyzed in parallel with remote sensing satellite data to effectively verify the accuracy and reliability of the ecological index of resource cities.
• The RCEI could be applied to other coal resource cities for more detailed and reliable validation analyses.
• The application potential of the RCEI on satellite sensors such as MODIS, ASTER, AVHRR, IRS, Sentinel, and Gaofen-5 should be explored. This could lead to enhancing the temporal, spatial, and spectral resolutions of remote sensing images through multi-source data fusion to establish a long-term, high-precision RCEI dataset for typical resource-based cities.
• The analysis of the driving forces behind changes in the ecological environmental quality of resource-based cities should be strengthened and the impact of climate change and human activities on their ecological environmental quality should be quantified.
• In subsequent studies, we will verify the reliability and validity of the PCA results using regression analyses, support vector machines (SVMs), decision trees, and neural networks. Future research will focus more on these nonlinear relationships and interactions to further enhance the accuracy of our analyses.

5. Conclusions

Utilizing the GEE platform in conjunction with Landsat TM/OLI remote sensing data, this study embarked on a comprehensive spatiotemporal dynamic monitoring and analysis of Jincheng City’s ecological environmental quality over the past 32 years, thereby employing the RCEI as a metric. This study explored the spatial differentiation characteristics of the city’s ecological environmental quality. The findings derived from the remote sensing assessment and the monitoring of spatiotemporal variations within the 1990–2022 timeframe revealed the following:

(1)

The Resource Cities Ecological Index (RCEI) is a comprehensive evaluation index based on remote sensing information. It integrates the majority of the information from six ecological environmental indicators, the NDVI, EVI, WET, NDBSI, LST, and ICDI, through principal component analysis. This index enables an objective and quantitative assessment and monitoring of the spatial and temporal distributions and change characteristics of the ecological environmental quality in mining areas.

(2)

Based on an analysis spanning from 1990 to 2022, the RCEI of Jincheng City ranged between 0.6 and 0.8, indicating that the city maintained an overall ecological quality at the upper–middle level. During this period, the city’s ecological environmental quality exhibited a general trend of improvement.

(3)

At the county-level scale, significant disparities in ecological quality were observed among various districts within Jincheng City. In particular, the central urban area and Gaoping County exhibited relatively lower ecological levels compared to other regions, with average RCEI values ranging from 0.4 to 0.5. This phenomenon was directly attributed to the dense distribution of coal mines and the high degree of urbanization in these areas.

(4)

The Moran’s I values were all over 0.9, suggesting a positive spatial autocorrelation in the distribution of ecological quality within the study area. This implied that the spatial distribution of ecological quality exhibited a clustered pattern, deviating from a random distribution. Specifically, the high–high (H–H) clusters were primarily located in the high-altitude regions surrounding the study area, characterized by dense ecological vegetation and fewer mining areas. Conversely, the low–low (L–L) clusters were primarily situated in urban construction areas and regions with large-scale mining distributions, where human production and living activities were frequent.

(5)

This study revealed the close link between coal production and ecological quality by analyzing the buffer zones of the Sihe and Fenghuangshan mining areas. The high production in the Sihe Mine Area caused a significant decrease in the mean RCEI value within the core mining area (0–1 km), a phenomenon that highlights the profound impact of mining activities on the surrounding ecological environment. Meanwhile, the Fenghuangshan Mine Area ceased production and underwent ecological restoration in 2018, and by 2022, the mean RCEI value of the area significantly increased from 0.4–0.45 to 0.65. Although urban development led to a decline in the ecological quality of the area 5 km away, the ecological quality of the environment within the city still showed a trend of improvement year by year.

This study provides a feasible method for evaluating the spatial and temporal changes in ecological environmental quality in resource cities. Selecting suitable remote sensing indicators, establishing a resource-based ecological environmental quality evaluation system, and carrying out the evaluation and monitoring of spatiotemporal changes hold great significance for ecological protection management and sustainable development in resource-based cities. Our assessment study entirely relied on Landsat images to construct the RCEI in the GEE platform, which provides a feasible method for evaluating spatial and temporal changes in ecological quality. Further research is needed to integrate our method with socioeconomic data to evaluate the ecological quality of resource-based cities in a more comprehensive and integrated manner.