Spatiotemporal Dynamics and Scenario Simulation of Regional Green Spaces in a Rapidly Urbanizing Type I Large City: A Case Study of Changzhou, China

Abstract

The rapid urbanization observed in major Chinese cities has resulted in the degradation of both urban and rural environments. In response to this challenge, the concept of regional green spaces has emerged as an innovative approach to coordinate and manage green space resources across urban and rural areas. This study focuses on conducting a comprehensive analysis of the evolution, driving factors, and future scenarios of regional green spaces in Changzhou, which serves as a representative Type I large city in China. To accomplish this analysis, Landsat satellite images from 1992, 2002, 2012, and 2022 were utilized. Various methodologies, including landscape pattern indices for quantitative evaluation, the CLUE-S model, logistic regression for qualitative evaluation, and the Markov–FLUS model, were employed. The findings indicate a continuous decline in the area of regional green spaces in Changzhou, decreasing from 248.23 km² in 1992 to 204.46 km² in 2022. Landscape pattern analysis reveals an increase in fragmentation, complexity, irregularity, and human interference within these green spaces. Logistic regression analysis identifies key driving factors influencing regional green spaces, including elevation, urban population, and proximity to water bodies and transportation. The scenario simulations provide valuable insights into potential future trends of regional green spaces. According to the economic priority scenario, a modest increase in regional green spaces is anticipated, while the ecological priority scenario indicates substantial growth. Conversely, the inertial development scenario predicts a continued decline in regional green spaces. This research emphasizes the significance of achieving a harmonious coexistence between economic progress and environmental preservation. It emphasizes the necessity of optimizing the arrangement of green areas within a region while fostering public engagement in the conservation of these spaces. The findings contribute to the protection and sustainable development of the urban environment in the Yangtze River Delta region.

1. Introduction

China’s unique urban–rural divide has resulted in fragmented environments in both urban and rural areas, leading to a deterioration of overall environmental conditions. To address this issue, China has prioritized harmonizing urban–rural relations and promoting integrated urban–rural development as key objectives in its new urbanization initiatives and rural revitalization strategies [1,2,3]. The theory of integrated urban–rural development emphasizes the implementation of system theory to facilitate the equitable exchange of ecological resources between urban and rural areas, as well as the balanced allocation of public ecological resources. This approach aims to achieve a development mode that maximizes benefits by integrating all elements and multiple domains of the living environment [4,5,6,7].

The Ministry of Housing and Urban–Rural Development in China adopted the Urban Green Spaces Classification Standard (CJJ/T 85-2017) in 2018 to address the need for well-structured urban–rural ecological resources planning [8]. This standard introduced the concept of “regional green spaces”, replacing the previous classification of “other green spaces” under code G5 in the 2002 edition of the green space classification standard. Regional green spaces encompass green areas located outside built-up areas and serve diverse functions, including the protection of urban–rural environmental and cultural resources, recreational and fitness activities, safety measures and isolation, species conservation, and garden and nursery plant production [9,10,11]. Examples of regional green spaces include scenic recreation green spaces, scenic spots, forest parks, wetland parks, suburban parks, ecological conservation green areas, regional facility protective green spaces, and productive green lands [12]. The presence of regional green spaces is crucial for ensuring regional ecological safety and promoting unified and coordinated management of urban–rural ecological spatial resources [13].

Previous research on regional green spaces has primarily concentrated on conceptual definitions, functional analyses, and planning strategies [14]. Ding and Zhang [15] conducted a comprehensive review of the research status and trends in the ecological functions of regional green spaces. Tang et al. [16] utilized remote sensing (RS) technology to study the ecological networking of green space landscape patterns in the central urban area of Xuzhou City, comparing ecological green spaces in the main urban area and peripheral suburbs. Xu et al. [14] utilized Landsat remote sensing imagery to analyze the transformations occurring in the green spaces of the Nanjing urban region. They employed dynamic measurements and various indices to evaluate the underlying factors influencing these transformations. Jiang et al. [17] investigated the spatial structural attributes of green spaces in the riverside region of Shanghai. They employed diverse viewpoints and Fragstats landscape indices to analyze these characteristics. Kucsicsa et al. [18] evaluated Romania’s significant structural changes in the land cover system caused by political and socioeconomic transformation, utilizing the CLUE-S model and CORINE land cover database. Kamal et al. [19] utilized satellite images and a multilayer perceptron Markov model to predict land-use changes in 2030 and investigate the issue of insufficient green spaces resulting from extreme urbanization in Bangkok. Nadoushan conducted an analysis of dynamic changes in land use and landscape pattern changes in Khomeyni Shahr County, Iran, using an artificial neural network classification method to generate land-use maps and compute landscape-level metrics with Fragstats software version 4.2. Hashemi Aslani et al. [20] employed proxy models, multilayer perceptron neural network technology, and Landsat images to analyze the North Awaz basin in Iran, assessing the impacts of human decision-induced land-use changes. The research conducted in these studies offers valuable perspectives on how the arrangement of urban green spaces impacts the urban environment and enhances the overall quality of life. These findings contribute significantly to the fields of urban planning and sustainable development.

The existing body of literature predominantly focuses on conceptual definitions, functional analyses, and planning strategies pertaining to regional green spaces, as evidenced by the works of Ding and Zhang [15], Tang et al. [16], and Xu et al. [14]. These studies have provided valuable insights into the ecological functions and landscape patterns associated with green spaces. However, a noticeable research gap exists in the exploration of spatiotemporal dynamics and future scenarios specific to regional green spaces, particularly within rapidly urbanizing Type I large cities in China. The literature predominantly concentrates on super and mega cities with well-established urban–rural development, thus overlooking the unique challenges faced by Type I large cities undergoing rapid urbanization and incremental updates to their urban–rural integration [14,17]. The introduction of the Urban Green Spaces Classification Standard in China, as noted by Ji et al. [13] and Wang et al. [12], has introduced the concept of “regional green spaces”, necessitating a comprehensive understanding of the evolving nature of these spaces and the factors driving their transformation. While recent studies by Kucsicsa et al. [18] and Kamal et al. [19] have utilized simulation-based approaches to investigate land-use changes and their impacts on green spaces in other countries, the application of these methodologies within the specific context of Chinese Type I large cities remains limited. Furthermore, although qualitative analyses conducted by Nadoushan [21] and Hashemi Aslani et al. [20] have laid a valuable foundation by examining driving factors that influence regional green spaces, further expansion is necessary to account for the unique circumstances present in rapidly urbanizing urban centers in China. This study aims to address these aforementioned research gaps and contribute to the growing body of knowledge on regional green spaces and their role in promoting sustainable urban–rural development. By conducting a comprehensive analysis of the evolution, driving factors, and future scenarios of regional green spaces in Changzhou, a representative Type I large city, crucial insights will be provided to inform policymaking and planning strategies in addressing the pressing environmental challenges faced by major Chinese cities amidst rapid urbanization. Moreover, the integration of the theory of urban–rural integration with the protection and expansion of regional green spaces, as explored in our research, presents a novel framework for addressing the complex interplay between economic progress and ecological preservation [22,23]. This approach aligns with the broader national priorities outlined in China’s new urbanization initiatives and rural revitalization strategies [24,25,26,27,28], thereby emphasizing the significance of our work in supporting the sustainable development of urban–rural communities.

Previous studies have predominantly focused on regional green spaces in super and mega cities characterized by well-established urban–rural development, often neglecting Type I large cities undergoing incremental updates. According to the 2014 notification issued by the People’s Republic of China, Type I large cities are defined as urban areas with a permanent resident population exceeding 3 million but below 5 million. Presently, Type I large cities in the Yangtze River Delta are experiencing rapid urbanization, which presents various challenges to the integrated development of urban–rural ecology. In this study, our specific focus is on Changzhou as a representative Type I large city, aiming to analyze the spatiotemporal evolution patterns of regional green spaces. To comprehensively analyze the spatiotemporal dynamics and future scenarios of regional green spaces in Changzhou, a diverse array of research methods was employed. These encompassed the application of landscape pattern indices to assess the structural composition and spatial distribution of green spaces, the implementation of the CLUE-S model and logistic regression analysis to investigate the driving factors influencing their transformation, and the integration of the Markov–FLUS model to simulate future land-use scenarios under varying development priorities [15,16,18]. The findings derived from this analysis contribute to the construction of a macro-scale ecological safety pattern for urban–rural development. This study aims to address the aforementioned research gaps and contribute to the existing body of knowledge concerning regional green spaces and their role in facilitating sustainable urban–rural development. Consequently, a comprehensive analysis was conducted to examine the evolution, driving factors, and future scenarios of regional green spaces in Changzhou, a Type I large city selected as a representative case. The findings from this analysis aimed to provide crucial insights that can inform policymaking and planning strategies. Furthermore, this research explored the integration of the theory of urban–rural integration with the protection and expansion of regional green spaces, presenting a novel framework that addresses the intricate interplay between economic progress and ecological preservation.

2. Materials and Methodology

2.1. Survey Area Overview

Changzhou holds strategic significance in the preservation and joint development of the Yangtze River Delta, located between 119°08′ E–120°12′ E and 31°09′ N–32°04′ N. The urban area covers approximately 4385 km², characterized by a subtropical monsoon climate with an average annual precipitation of approximately 1200 mm. The landscape of Changzhou is predominantly characterized by a combination of plains and hills, which provide abundant ecological resources, including evenly distributed forests and grasslands. Notable mountainous areas in the region include Mao Mountain in Jintan City and the residual ranges of Tianmu Mountain in Liyang City, located in the western and southern regions respectively (Figure 1). As of the end of 2022, the urbanization rate had reached 78.01%, leading to continuous updates in the urban–rural built-up areas. To promote sustainable development, Changzhou has implemented an “Ecological Green City Construction” strategy, which encompasses the concept of “one city, two lakes, three mountains, and six corridors”. The term “one city” signifies the development of an ecological green city, while “two lakes” refers to the ecological construction centered around Changdang Lake and Ge Lake. The “three mountains” represent the establishment of forest parks around Mao Mountain, Wawu Mountain, and the southern hills near Tianmu Lake. The “six corridors” consist of ecological green corridors along the Yangtze River, the Beijing-Hangzhou Grand Canal, the Taihu protection zone, the Changzhou–Taicang high-speed railway-based ecological corridor, the Xinmeng River clean water corridor, and the Xingou River clean water corridor. With rapid urban–rural development, Changzhou has become a crucial component in shaping the green space network within the Yangtze River Delta urban agglomerations, laying a foundation for regional green space construction [29,30,31,32,33].

2.2. Data Sources

For this research, remote sensing image data were acquired from the Geographic Spatial Data Cloud Platform and collected through the utilization of the United States Land Resources Satellite. The image data consisted of Landsat4-5TM satellite digital products, Landsat8-9OLI/TIRS C2 L2 satellite digital products, and Landsat7TM C2L2 products, with a spatial resolution of 30 m. The selected image data covered the years 1992, 2002, 2012, and 2022, and had specific orbital numbers (119/038, 120/38, 120/38, and 119/38) with cloud coverage of less than 10%. The coordinate system used was WGS-1984-UTM-Zone-50N [34]. The vector data for Changzhou’s administrative boundary were acquired from the National Geomatics Center of China and utilized the WGS-84 coordinate system. The DMSP/OLS nighttime light data, acquired from the Harvard Dataverse, were accessible for the years 1992, 2002, 2012, and 2022. These data were used to extract information on built-up areas. Various driving factor data were used in the study. Data on annual precipitation and average temperature were obtained from the Jiangsu Provincial Bureau of Statistics. Population-related data, such as total population, urban population, urbanization rate, and gross regional product, were obtained from the Changzhou Statistical Yearbook published by the Changzhou Municipal Bureau of Statistics.

2.3. Methodology

2.3.1. Processing of Remote Sensing Images and Extraction of Regional Green Spaces

This study utilized data from 1992, 2002, 2012, and 2022 to analyze the changes in regional green spaces in Changzhou. The identification of regional green spaces involved data extraction from Landsat satellite imagery and Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light data. Various metrics, such as dynamic degree, landscape pattern indices, and change rates, were employed to evaluate the alterations in the scale and pattern of these green spaces. Additionally, the study employed the CLUE-S model and logistic regression model to analyze the factors influencing the changes in the pattern of regional green spaces [35]. Furthermore, the study simulated three different scenarios using the Markov–FLUS (Future Land Use Simulation) model. These scenarios encompassed inertial development, economic prioritization, and ecological prioritization of regional green spaces in Changzhou. The primary objective was to explore a pathway that integrates the theory of urban–rural integration with sustainable regional ecological development.

The remote sensing images from 1992, 2002, 2012, and 2022 underwent various processing steps using ENVI 5.6 software. The process involved various steps, such as radiometric calibration, atmospheric correction, geometric correction, image clipping, and image splicing. The L2 series of image products were corrected using GROUP = L1 METADATA FILE and merged with satellite images having similar imaging conditions in the survey area from both east and west directions. The survey area’s landscapes were categorized into six groups using the Current Land Use Classification (GB/T 21010-2017) and the Standard for Classification of Urban Green Spaces (CJJ/T 85-2017). These categories encompassed construction land, forest land, cultivated land, grassland, water bodies, and unused land, wherein forest land and grassland were deemed as urban green spaces. The remote sensing images underwent supervised classification employing the maximum likelihood method. Post-processing and the confusion matrix were utilized to refine and evaluate the classification results, generating interpretation maps of land use in Changzhou at a 30 m resolution [21]. Furthermore, the classification results were processed using ArcMap 10.8 software to convert them into vector format and extract the segments corresponding to urban green spaces. To ensure accuracy up to 1 m, the remote sensing images were visually compared and adjusted, taking into account references from Google Earth, field investigations, and high-resolution satellite images accessed through Google Earth. To validate the accuracy of the results, 50 random points were generated as validation samples using ArcMap, and accuracy reports were generated based on the confusion matrix. The Kappa coefficient was verified to be higher than 0.8, and the classification accuracy exceeded 80%, meeting the requirements of the study.

The nighttime light image data captured by DMSP/OLS is a valuable indicator of human activity intensity and is commonly utilized for extracting segmented urban built-up areas. According to Xie et al. [36], a light threshold of 50 or higher closely corresponds to the range of built-up areas. In this study, ArcGIS 10.8 software was employed to convert the DMSP/OLS nighttime light images into vector format, allowing for the delineation of built-up area boundaries. Through the integration of these data with the interpretation maps obtained from the remote sensing land use images, the built-up areas were refined to exclude green spaces. Consequently, this refinement facilitated the extraction of regional green spaces situated beyond the built-up areas.

In conclusion, the remote sensing image data used in this study were acquired from the Geographic Spatial Data Cloud Platform and collected through the utilization of the United States Land Resources Satellite. The image data consisted of Landsat4-5TM satellite digital products, Landsat8-9OLI/TIRS C2 L2 satellite digital products, and Landsat7TM C2L2 products, all with a spatial resolution of 30 m. The selected image data covered the years 1992, 2002, 2012, and 2022, with specific orbital numbers (119/038, 120/38, 120/38, and 119/38) and cloud coverage of less than 10%. The coordinate system used was WGS-1984-UTM-Zone-50N. The vector data for Changzhou’s administrative boundary were acquired from the National Geomatics Center of China and utilized the WGS-84 coordinate system. The DMSP/OLS nighttime light data, acquired from the Harvard Data verse, were accessible for the years 1992, 2002, 2012, and 2022, and were used to extract information on built-up areas.

The spatial and temporal resolutions of the data used in this study were crucial for capturing the spatiotemporal dynamics of regional green spaces in Changzhou. The 30 m spatial resolution of the remote sensing imagery and the annual temporal resolution of the climatic and socioeconomic data provided a robust dataset to analyze the evolution and driving factors of these green spaces over the study period. The consistent geographic coordinate system and the use of high-quality, cloud-free satellite images further ensured the reliability and comparability of the data across the different time points. By providing detailed information about the data sources and their characteristics, we sought to demonstrate the appropriateness of the dataset in addressing the research objectives and facilitating the replication of this study in other rapidly urbanizing cities.

2.3.2. Calculation of the Dynamic Degree

The dynamic degree (K) model is an analytical approach utilized to evaluate alterations in the count of land-use categories within a designated timeframe. It functions as an indicator of the magnitude and pace of changes occurring in various land-use types within a specific region. In this study, the dynamic degree was employed to quantify the annual rate of change for regional green areas within the surveyed area. The formula (Equation (1)) used to express the dynamic degree is as follows:

$$\mathrm{K}={\displaystyle \frac{{\mathrm{U}}_{\mathrm{b}}-{\mathrm{U}}_{\mathrm{a}}}{{\mathrm{U}}_{\mathrm{a}}}}\times {\displaystyle \frac{1}{\mathrm{T}}}\times 100\%.$$

(1)

In this equation, K symbolizes the dynamic degree of regional green areas within the specified timeframe, while U_a and U_b represent the green space areas at the commencement and conclusion of the survey period, respectively. The time interval, denoted as T and measured in years, is also incorporated [37].

2.3.3. Analysis of the Change Degree

To perform the analysis, ArcGIS 10.8 software and overlay analysis tools were utilized. The regional raster data were converted into vector format, enabling further calculations and evaluations. Summarization techniques were employed to determine the regional area, increased area, and reduced area. Joint operations were then used to calculate the rates of increase and decrease in green spaces during different time periods. The survey area was divided into 300 m × 300 m rasters, resulting in a total of 4972 basic evaluation units. For each raster, the increase and decrease rates of regional green spaces were computed, providing the change degree of green spaces within each unit. The following formulas (Equations (2) and (3)) were used for these calculations:

$$\mathrm{S}={\displaystyle \frac{{\mathrm{Q}}_{\mathrm{b}}}{\mathrm{Qa}}}\ast 100\%,$$

(2)

$$\mathrm{P}={\displaystyle \frac{{\mathrm{Q}}_{\mathrm{c}}}{{\mathrm{Q}}_{\mathrm{a}}}}\ast 100\%.$$

(3)

Within these equations, Q_a signifies the regional green space area within the survey area during the previous year. Q_b represents the reduction in regional green spaces, while Q_c represents the increment in regional green spaces. The variables P and S represent the rates of increase and decrease in regional green spaces, respectively [38].

2.3.4. Selection and Computation of Landscape Pattern Indices

Landscape pattern indices serve as quantitative measures that capture the structural composition, spatial distribution, and other attributes of landscapes. These indices are employed to establish connections between patterns observed and underlying landscape processes [35]. In this study, our focus centered on examining the properties of patches and the overall pattern of regional green spaces using five distinct landscape pattern indices. These indices consist of patch density (PD), shape index (LSI), aggregation index (AI), perimeter–area fractal dimension (PAFRAC), and connectivity (CONNECT).

Table 1. Calculation formulas and ecological implications of landscape pattern indices.

Landscape Pattern Indices	Name in Short	Ecological Implication	Calculation Formulas
Patch density	PD	PD represents the number of patches per unit area, and measures the degree of landscape fragmentation.	$\mathrm{PD}={\displaystyle \frac{{\mathrm{n}}_{\mathrm{i}}}{\mathrm{A}}}$(1000)(100)
Landscape shape index	LSI	LSI represents the complexity of landscape patch shapes.	$\mathrm{LSI}={\displaystyle \frac{25{{\displaystyle \sum }}_{\mathrm{k}=1}^{\mathrm{m}}{\mathrm{e}}_{\mathrm{ik}}}{\sqrt{\mathrm{A}}}}$
Aggregation index	AI	AI measures the spatial aggregation degree of similar patches in the landscapes.	$\mathrm{AI}=\left[{\displaystyle \sum }_{\mathrm{j}=1}^{\mathrm{n}}\left({\displaystyle \frac{{\mathrm{g}}_{\mathrm{ij}}}{\max \to {\mathrm{g}}_{\mathrm{ij}}}}\right){\mathrm{P}}_{\mathrm{i}}\right]$
Connectivity	CONNECT	CONNECT reflects the spatial connectivity between patches among landscapes.	$\mathrm{CONNECT}={\displaystyle \frac{{{\displaystyle \sum }}_{\mathrm{j}=\mathrm{k}}^{\mathrm{n}}{\mathrm{c}}_{\mathrm{ijk}}}{\frac{{\mathrm{n}}_{\mathrm{j}}\left({\mathrm{n}}_{\mathrm{i}}-1\right)}{2}}}\times 100$
Perimeter–area fractal dimension	PAFRAC	PAFRAC used to measure the degree of anthropogenic disturbance to patch shapes.	$PAFAC={\displaystyle \frac{2}{\frac{N{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\left(\ln p_{ij}\times \ln a_{ij}\right)-{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\ln p_{ij}}\times {\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\ln a_{ij}}}}}}}{N{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\ln {P^2}_{ij}}-{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\ln P_{ij}}}}}}}\left(1\le PAFAC\le 2\right)$

provides an overview of the calculation formulas for these landscape pattern indices, accompanied by their ecological implications.

2.3.5. Qualitative Analysis of Driving Factors Based on the CLUE-S Model and Logistic Regression Analysis

The transformation of land use is influenced by a multifaceted interaction of natural and socioeconomic factors, occurring in both static and dynamic contexts. Therefore, it is essential to consider both these elements when selecting driving factors to accurately simulate land-use change. In this research, the process of selecting driving factors for land-use change adhered to four primary principles:

-

Accessibility: The factors chosen should be readily available for studying driving factors and ensuring the accuracy of related models.

-

Consistency: The factor data should be consistently measured across both time and space.

-

Quantifiability: The factors should be quantifiable and suitable for input into models for simulation and prediction.

-

Comprehensive selection: The factors should be selected in a comprehensive manner, considering various aspects.

Based on these principles and the relevant literature, the study identified 11 driving factors for land-use change. These factors encompassed natural elements such as temperature, slope, precipitation, digital elevation model (DEM) elevation, and aspect. Additionally, human-related factors, including urban population and gross domestic product (GDP), as well as transportation factors, such as distances to water bodies, roads, railways, and residential areas, were considered [39]. Figure 2 indicates the driving factors considered in the study to analyze the changes in regional green spaces in Changzhou. The figure encompasses various components. Firstly, the aspect representing the terrain’s orientation can influence vegetation distribution and growth. The DEM provides elevation data that is crucial in determining regional green spaces, as topography affects factors such as soil moisture, temperature, and accessibility. Additionally, the slope of the terrain plays a role in shaping the type and distribution of regional green spaces. Climatic factors, including average annual temperature and annual precipitation, significantly impact the suitability and growth of regional green spaces by influencing factors such as water availability and vegetation patterns. The size of the urban population exerts pressure on regional green spaces through land use changes and development activities. Moreover, the level of economic development, as represented by GDP, drives the conversion of regional green spaces for industrial, commercial, or residential purposes. Proximity to built-up residential areas affects the accessibility and usage of regional green spaces by the local population. Distance to railway infrastructure influences the connectivity and integration of regional green spaces with the broader transportation network. Similarly, proximity to road networks impacts accessibility and susceptibility to encroachment or fragmentation. Furthermore, the distance from regional green spaces to water bodies, such as rivers, lakes, and wetlands, is a significant factor in determining their ecological functions and overall sustainability. The figure presents these driving factors in a contour-based visualization, highlighting the range of low and high values for each factor across the study area.

The CLUE-S model is a valuable instrument employed to simulate land-use changes and evaluate their environmental consequences at a localized scale. Analyzing driving factors is facilitated by the file convert function within the CLUE-S model, coupled with logistic regression analysis in SPSS version 29. By conducting regression analysis, it becomes feasible to ascertain the likelihood of each land type occurring in specific pixels within the region. This analysis also establishes the correlation between the spatial distribution of various land-use types and their corresponding driving factors [40]. This study primarily focused on investigating the regional green spaces in Changzhou, which were considered the dependent variable, while the driving factors were treated as independent variables. The land-use data for the year 2022 in Changzhou were classified into six distinct classes, and the regional green spaces were subsequently extracted. The raster data associated with the 11 selected driving factors, along with their autocorrelation factors, were processed and converted into ASCII format. Normalization was applied to these factors, followed by text processing using the CLUE-S model. The processed factors were then imported into SPSS for logistic regression analysis.

To examine the association between different influencing factors and the probability of changes in regional green space patterns, a logistic regression model was employed. The probability was calculated using the following formula (Equation (4)):

$$\log \left({\displaystyle \frac{P_i}{1-p_i}}\right)=\alpha +{\beta }_1{X}_{1,i}+{\beta }_2{X}_{2,i}+\cdots {\beta }_n{X}_{n,i}.$$

(4)

In this equation, P_i denotes the probability of a specific land type i occurring in a raster. X _{(n, i)} represents the driving factors influencing the distribution pattern of that land type, such as population, GDP, distance from a raster to roads, and terrain conditions. The coefficient β corresponds to the regression coefficient in the logistic regression equation, quantifying the relationship between the land type and the driving factors. Furthermore, α signifies the intercept term [41].

The accuracy of the regression outcomes was assessed using the receiver operating characteristic (ROC) curve. The computation of the area under the ROC curve (AUC) followed the formula provided in Equation (5).

$$\mathrm{AUC}={\displaystyle \frac{1}{\mathrm{n}}}{\displaystyle \sum }_{\mathrm{j}=1}^{{\mathrm{n}}_{\mathrm{b}}}{\displaystyle \sum }_{\mathrm{i}=1}^{{\mathrm{n}}_{\mathrm{a}}}\mathsf{\phi }\left({\mathrm{X}}_{\mathrm{ai}},{\mathrm{X}}_{\mathrm{bj}}\right),\mathsf{\phi }\left({\mathrm{X}}_{\mathrm{ai}},{\mathrm{X}}_{\mathrm{bj}}\right)=\left\{\begin{array}{c}1,{\mathrm{X}}_{\mathrm{ai}}>{\mathrm{X}}_{\mathrm{bj}}\\ 0.5,{\mathrm{X}}_{\mathrm{ai}}={\mathrm{X}}_{\mathrm{bj}}\\ 0,{\mathrm{X}}_{\mathrm{ai}}<{\mathrm{X}}_{\mathrm{bj}}\end{array}\right..$$

(5)

In the equation, X_ai (i = 1, 2, 3, …, n_a) represents the observed values in the abnormal data set, and X_bj (j = 1, 2, 3, …, n_b) represents the observed values in the normal data set. Typically, the AUC value falls between 0.5 and 1.0. A higher AUC value indicates a superior fit, implying that the driving factors can more aptly elucidate the particular land type [42].

2.3.6. Simulation of Changes in Land-Use Types Based on the FLUS Model and Markov Model

The FLUS model combines the artificial neural network (ANN) algorithm with conventional cellular automata (CA) to evaluate land-use changes. By taking into account the initial land-use types and driving factors, this model assesses the appropriateness of different land transformations. It effectively handles the complexity and uncertainty associated with natural and human-induced influences on land-use development and change. This is accomplished through an adaptive inertial competition mechanism based on roulette selection, which produces simulated land-use outcomes [43]. In order to forecast forthcoming alterations in land-use types, the FLUS model employs a Markov multi-year transition probability matrix of land-use types. By integrating the intricacies of land use influenced by both natural and human factors, the CA-based FLUS model can simulate future land-use scenarios [44].

In this study, the utilization of ANN was employed to predict the land use and cover in Changzhou, specifically for the year 2022. This process resulted in the generation of prediction and simulation maps. An overlay analysis was then conducted using a raster computer in ArcGIS 10.8, comparing the simulated maps with the actual raster maps of 2022. For rasters coded as 0, it indicated that the simulation was accurate. The Kappa coefficient was calculated using the following formula (Equation (6)):

$$\mathrm{Kappa}={\displaystyle \frac{{\mathrm{p}}_0-{\mathrm{p}}_{\mathrm{c}}}{{\mathrm{p}}_{\mathrm{p}}-{\mathrm{p}}_{\mathrm{c}}}}$$

(6)

In Equation (6), p₀ signifies the observed proportion of accurate simulation outcomes, p_c represents the anticipated proportion of accurate simulation outcomes under random conditions (set at 1/5 in this study due to the consolidation of grassland and forest land into a single category among the five land types), and pp denotes the ideal proportion of accurate simulation outcomes (set as 1 in this study). If the Kappa coefficient exceeded 80%, it indicated a high degree of similarity between the simulated and actual results. Meeting the numerical requirement of the Kappa coefficient suggested the potential use of the driving factors for simulating future land cover/use. To obtain simulated prediction data for 2022, ANN was utilized to simulate prediction maps based on the driving factors observed in 2012. Subsequently, a CA analysis and a comparison of land-use matrices were conducted. The weights for the neighborhood factors, which refer to the elements in the surrounding or adjacent areas influencing land-use changes, are presented in

Table 2. Weights of neighborhood factors.

	Cultivated Land	Green Space	Water Body	Construction Land	Unused Land
Weights of neighborhood factors	0.49	0.50	0	1	0.5

[45].

By applying the Kappa coefficient calculation formula, the land use simulation based on the 2012 data yielded a Kappa coefficient of 81% for the year 2022. This indicates a strong similarity between the simulated and actual conditions under the considered driving factors. Therefore, these driving factors can be effectively utilized to simulate land use and cover in future years. To project future land conditions in 2032, an external Markov model was integrated, which leverages the land use status of 2022. The Markov model is a reliable quantitative prediction model that estimates changes in land-use types using a multi-year transition probability matrix. It can be effectively combined with the FLUS model. The formula (Equation (7)) for predicting land use with the Markov model is as follows:

$${\mathrm{S}}_{\left(\mathrm{t}+1\right)}={\mathrm{P}}_{\mathrm{ij}}{\mathrm{S}}_{\mathrm{t}}.$$

(7)

In Equation (7), S_(t+1) symbolizes the land use condition at time t + 1, P_ij represents the transition probability matrix of land-use types, and St denotes the land use condition at the initial time t.

This study employed the Markov model to predict and simulate multiple land-use scenarios within the surveyed area, with the aim of providing decision-makers insights into regional green space development. Three primary scenarios were designed for analysis. The first scenario, referred to as the “Inertial Development Scenario”, simulated future land-use types and scales based on historical change rates and the natural and anthropogenic driving factors from 1992 to 2022. The Markov model was utilized in the coupling model to forecast the scales of various land-use types, which were then treated as parameters for scale demand in the FLUS model. The second scenario, referred to as the “Ecological Priority Scenario”, incorporated ecological redline restriction areas into the inertial development scenario. In this scenario, adjustments were made to the transition probability matrix of the Markov model. The probability of converting arable land, water bodies, construction land, and unused land into green space was increased by 50%, while the probability of converting green spaces into other land-use types was regulated and reduced by 50%. The third scenario, known as the “Economic Priority Scenario”, focused on examining the impacts of economic development on urban expansion. In this scenario, the expansion of human settlement areas was considered, and the green space system was designated as a transformable land type with an increased probability of transformation into a human settlement ecological system. The transition probability matrix of the Markov model was adjusted to reflect the trend of natural development. More specifically, the likelihood of converting arable land, green spaces, water bodies, and unused land into construction land was enhanced by 50%. Conversely, the probability of transforming construction land into other land-use types was regulated and reduced by 50%. Additionally, the ecological space within the areas designated by red lines for ecological protection was considered a restricting factor, aligning with the development policies of Changzhou. These scenarios provide different perspectives on regional green space development, considering ecological and economic priorities as well as natural and anthropogenic factors [44].

When analyzing the conversion of land-use types, it is crucial to consider the potential interconversion between different land categories. To facilitate this, it is necessary to establish a cost matrix and define restricted areas. A value of 0 indicates that a particular land-use type cannot undergo transformation, while a value of 1 signifies its transformability. In the context of urbanization, it is improbable for construction land to convert into other land-use types. Thus, restrictions are placed on the transformation of construction land into other categories. However, the transformability of other land-use types cannot be determined directly and must be defined based on the specific scenario being examined. The detailed transformation settings for each land-use type are outlined in

Table 3. Optimal land-type transformation rules.

Scenario Setting	Inertial Development Scenario					Ecological Priority Scenario					Economic Priority Scenario
	CL	GS	WB	CT	UN	CL	GS	WB	CT	UN	CL	GS	WB	CT	UN
CL	1	1	1	1	0	1	1	1	1	0	1	1	1	1	0
GS	1	1	0	0	0	0	1	0	0	0	1	1	1	1	0
WB	1	0	1	1	0	1	1	1	1	0	1	1	1	1	0
CT	1	0	0	1	0	1	1	1	1	0	0	0	0	1	0
UN	0	0	0	0	1	0	1	0	0	1	0	0	0	1	1

CL: cultivated land; GS: green space; WB: water body; CT: construction land; UN: unused land.

[46].

To comprehensively examine the spatiotemporal dynamics and future prospects of regional green spaces in Changzhou, a diverse range of research methods was employed. These methods included the application of landscape pattern indices to evaluate the structural composition and spatial distribution of green spaces, the utilization of the CLUE-S model and logistic regression analysis to investigate the driving factors that influence their transformation, and the integration of the Markov–FLUS model to simulate future land-use scenarios considering different development priorities. The insights derived from this analysis significantly contribute to the establishment of a macro-scale ecological safety pattern for urban–rural development.

3. Results and Analysis

3.1. Change in the Area of Regional Green Spaces

The examination of land use dynamics from 1992 to 2022 yielded noteworthy observations, visually depicted using a Sankey diagram (Figure 3). The key findings are summarized as follows:

-

From 1992 to 2022, a noteworthy and escalating interchange of land-use types occurred, with varying magnitudes observed on an annual basis. Among all land-use types, cultivated land displayed the most substantial degree of transformation, consistently declining each year and undergoing modifications in conjunction with other land-use categories, ultimately transitioning into construction land.

-

The overall modification of regional green spaces was relatively constrained. However, regional green spaces predominantly underwent a conversion into cultivated land and construction land, indicating a shift in their land-use categorization.

These findings provide insights into the land use dynamics within the examined region and emphasize notable shifts observed in particular land-use categories, specifically the decline of cultivated land and the conversion of regional green spaces into alternative classifications.

Figure 4 illustrates the distribution of regional green spaces within Changzhou City, indicating a prominent concentration in the southwestern region, while being relatively scarce in the northeastern area. Notably, significant clusters of green spaces were situated along the boundaries of the administrative divisions within Changzhou. More specifically, the districts of Liyang and Jintan encompassed the majority of regional green spaces, whereas the Wujin district exhibited a comparatively lower prevalence of such areas. Moreover, a considerable number of regional green spaces were located adjacent to expansive water bodies or mountainous terrains. Nevertheless, it is crucial to acknowledge that the overall coverage of regional green spaces was restricted and exhibited a gradual decrease throughout the examined timeframe. In order to evaluate the changes in these areas, ArcGIS 10.8 was utilized to identify the precise extents of regional green spaces within the various districts of Changzhou City during the periods of 1992–2002, 2002–2012, and 2012–2022.

The comprehensive changes in regional green spaces are summarized in

Table 4. Area and dynamics of green spaces in Changzhou from 1992 to 2022.

Administrative District	Area, km2				Dynamic Degree during 1992–2002, %	Dynamic Degree during 2002–2012, %	Dynamic Degree during 2012–2022, %
	1992	2002	2012	2022
Jintan district	38.37	36.81	21.62	23.04	0.00	−0.04	0.01
Liyang district	195.01	190.80	181.06	172.14	0.00	−0.01	0.00
Tianning district	1.37	0.83	0.40	0.44	−0.04	−0.05	0.01
Wujin district	12.48	11.28	9.09	8.42	−0.01	−0.02	−0.01
Xinbei district	0.98	0.81	0.38	0.43	−0.02	−0.05	0.01
Zhonglou district	0.02	0.01	0.01	0.00	−0.05	−0.04	−0.07
Total area	248.23	240.53	212.56	204.46	0.00	−0.01	0.00

. Between 1992 and 2022, all administrative districts in Changzhou City experienced a decrease in the area of regional green spaces, resulting in a total reduction of 7.7 km² (approximately 3.1%). Although the overall reduction is not substantial, specific areas, such as the Tianmu Lake Scenic Area, have undergone significant decreases in green space. In the Tianmu Lake region alone, green space has diminished by approximately 44.3% of its original area, with other major green patches also witnessing notable reductions. This decline has resulted in a shift from comprising 5.66% of the total area in 1992 to 4.66% in 2022. Among the areas with decreased green spaces, Liyang and Jintan districts experienced significant reductions of 4.21 km² and 1.56 km², respectively. The Tianning and Zhonglou districts, which are major urban areas in Changzhou, displayed noticeable decreases in green space, with dynamic changes of −0.04% and −0.05%, respectively, indicating larger changes compared to other regions. However, both districts fell below the overall dynamic degree of the city. Between 2002 and 2012, there was a notable decrease in the area of regional green spaces within the city, amounting to approximately 27.97 km² or approximately 11.63% compared to the preceding period. The overall dynamic degree experienced a shift from 0 to −0.01%. The changes in green spaces across the administrative districts increased compared to the previous period, with a continued declining trend observed in regional green spaces. Notably, Tianning and Xinbei districts exhibited the most significant reduction in green spaces, with a decrease of −0.05%, followed by Jintan and Zhonglou districts with a decrease of −0.04%. From 2012 to 2022, the overall area of regional green spaces continued to decrease by approximately 8.1 km², reducing the rate of reduction to 3.81%. Compared to the previous period, the change amplitude became more gradual, and the degree of area reduction decreased by 7.82%. Jintan, Tianning, and Xinbei districts showed slight growth in the area of regional green spaces, with increases of 6.57%, 10%, and 13.16%, respectively. However, Wujin and Zhonglou districts exhibited a continuing decreasing trend with a dynamic degree of −0.01%. While Jintan, Tianning, and Xinbei districts exhibited growth, the Liyang district witnessed a decline in green space area, amounting to 8.92 km², equating to a reduction of approximately 4.93%. As a result, the total area of regional green spaces in Changzhou City continued to decrease.

3.2. Change Degrees of Regional Green Spaces

Rates of Increase and Decrease in the Area of Regional Green Spaces in Changzhou

The spatial distribution of decreased regional green spaces in Changzhou from 1992 to 2022 is illustrated in Figure 5a–c, revealing sporadic patches of reductions, particularly in the northwest and southwest regions. These reductions were more pronounced along the peripheries of the administrative districts. Figure 6a–c depicts the overall change in the rate of increase, which was relatively small, mostly ranging between 0% and 30%. The areas of significant change were concentrated near key regional green space patches. During the period from 1992 to 2002, there were numerous areas exhibiting a 100% reduction rate in regional green spaces, primarily situated at the edges of the green space patches and surrounding the densely built-up central urban areas. In total, 1493 rasters indicated decreases in regional green spaces ranging from 70% to 100%. These areas accounted for 20.43% of the total regional green spaces. This reduction can be attributed to Changzhou’s planning strategy during that time, which focused on “controlling the east and west, developing the north and south, with an emphasis on the north” (Sourced from the work report of the Changzhou People’s Government). Consequently, a significant portion of green areas located in the northern part of the main urban area experienced a complete reduction of 100%. The Mao Mountain Scenic Area, located in the western region of Jintan district, also experienced significant declines in regional green spaces. The rapid urbanization in Changzhou resulted in the conversion of forest land and grassland into built-up areas, leading to the reduction of green spaces. However, there were only a few instances where regional green spaces exhibited an increase in area, and these notable increases were primarily observed within large green space patches.

From 2002 to 2012, there was a decrease in patches with a 100% reduction in regional green spaces compared to the previous period. However, significant reductions were still observed in the majority of regional green spaces located at the boundaries of the Mao Mountain Scenic Area in Jintan and the Tianmu Lake Scenic Area in Liyang. The overall increase rate of regional green spaces remained limited and concentrated in small sections of the patches. During this time, efforts were focused on the integrated development and conservation of tourism resources in certain areas, including the Tianmu Lake area, Changdang Lake area, and Qianzi Lake area. As a result, there was a 100% increase in green spaces within the ecological zone of Tianmu Lake and at the intersection of Changdang Lake and Qianzi Lake.

Between 2012 and 2022, Changzhou enacted an ecological protection redline, encompassing a total area of 345.46 km², as a measure to restrict the excessive expansion of regional green spaces. The proportion of green spaces experiencing a reduction rate of 70–100% decreased, while the proportion of areas with the most common reduction rate (0–30%) increased, primarily due to the slower reduction of regional green spaces in the Changdang Lake Scenic Area. During this period, Changzhou pursued an urban–rural integration strategy known as the “Two Lakes (Changdang Lake and Ge Lake)-Oriented Innovation Zone”, which accelerated the development of urban construction land between Changdang Lake and Ge Lake. However, the implementation of the ecological redline led to a gradual decrease in the reduction rate of regional green spaces within the delimited area, shifting from 70–100% before 2012 to 0–30%. Despite this, the main urban areas of Changzhou continued to experience incremental growth. The integration between Changzhou and Jintan, as well as the expansion to the east and south, resulted in a reduction of regional green spaces between Qianzi Lake and Changdang Lake. The central urban areas expanded continuously, leading to a significant decrease in regional green spaces along the urban–rural borders. Currently, Changzhou’s urban development is still in a phase of incremental expansion. The concept of urban–rural integration has influenced a focus on the development of regional green spaces, leading to a noticeable increase in green spaces in certain regions in recent times. However, the overall reduction rate of regional green spaces within the city still exceeds the increase rate. Therefore, it remains crucial to strengthen ecological protection for regional green spaces.

3.3. Changes in the Overall Landscape Pattern of Regional Green Spaces

The city’s land-use types were reclassified using ArcGIS 10.0 in order to identify and extract regional green spaces. Fragstats 4.0 software was then employed to obtain landscape pattern indices specifically pertaining to the identified regional green spaces within the survey area. The graphical information summary provided insights into the changes in landscape pattern characteristics within the city’s regional green spaces between the years 1992 and 2022.

Patch density (PD) serves as an indicator of landscape heterogeneity and is positively associated with landscape fragmentation. Based on the data presented in

Table 5. Landscape indices of regional green spaces in Changzhou from 1992 to 2022.

Year	PD	LSI	AI	PAFRAC	CONNECT
1992	0.63	51.23	89.16	1.33	0.5562
2002	0.74	52.41	89.13	1.36	0.5054
2012	0.79	54.80	88.94	1.37	0.469
2022	0.88	57.92	89.56	1.38	0.4483

PD: patch density; LSI: landscape shape index; AI: aggregation index; PAFRAC: perimeter–area fractal dimension; CONNECT: connectivity.

, the PD values in the surveyed area exhibited an annual increase, ranging from 0.63 to 0.88, indicating a greater degree of fragmentation in the pattern of regional green spaces. The aggregation index (AI) measures the connectivity between patches within a specific landscape type, where smaller AI values suggest a more dispersed landscape. Table 5 demonstrates relatively consistent AI values for the years 1992, 2002, 2012, and 2022, all hovering around 89, indicating a concentrated distribution of patches. However, between 1992 and 2012, the AI of regional green spaces in the surveyed area displayed a gradual, albeit weak, decreasing trend, totaling 0.22. This indicates a shift towards a more dispersed pattern. Following 2012, the city implemented a controlled and optimized approach to urban and rural development, resulting in an increase in the AI of regional green spaces from 88.94 in 2012 to 89.56 in 2022. This indicates a more concentrated distribution of patches compared to 1992. The connectivity index (CONNECT) describes the clustering or dispersion trend of patches. A higher CONNECT value indicates well-connected patches, while a lower value signifies fragmented landscapes with multiple elements. As depicted in Table 5, the CONNECT value decreased from 0.5562 in 1992 to 0.4483 in 2022, indicating a declining trend in patch connectivity during this period. Consequently, the patches became less connected over time. The perimeter–area fractal dimension (PAFRAC) reflects the impact of human activities on the landscape pattern. A PAFRAC value approaching 2 indicates significant human interference. The data presented in Table 5 shows an increase in PAFRAC from 1.2 in 1992 to 1.2135 in 2022, indicating a yearly rise in human-induced interference on regional green spaces due to urbanization. The shape index (LSI) represents the complexity of patch shapes. A higher fluctuation in LSI suggests unstable patch shapes. From 1992 to 2011, the regional green spaces in Changzhou experienced a continuous increase in LSI, ranging from 51.23 to 57.92, indicating increasingly complex patch shapes and an overall more irregular pattern. In summary, the regional green spaces in the surveyed area exhibited fragmented, complex, and irregular patterns, with a higher degree of human interference due to urbanization.

3.4. Qualitative Analysis of the Driving Factors for Regional Green Spaces

The regression results, along with the corresponding regression constants and coefficients, as well as the ROC values of green spaces, are presented in

Table 6. Beta values of logistic regression for each land-use type.

Regression Coefficient	Cultivated Land	Greenbelt	Water Body	Construction Land	Unused Land
Waterways	2.014	1.907	−1.097	−2.288	−4.488
Temp	−4.385	−2.087	10.672	2.141	−29.394
Slope	−5.064	15.688	5.81	−1.541	−36.663
Roads	−15.28	2.971	29.7	−31.99	19.986
Railways	−0.547	0.255	0.822	0.251	−1.74
Pre	−0.457	4.535	−1.08	−0.752	3.123
Tpop	0.232	−7.025	1.757	1.24	5.981
Dsr	3.261	−4.974	2.964	−6.546	−7.646
Gdp	−4.091	3.988	−4.321	2.1	−3.884
DEM	−18.063	42.021	−6.918	−0.232	−98.198
Aspect	−	0.252	−0.034	0.073	0.181
ROC	0.755	0.983	0.868	0.849	0.784

Temp: annual average temperature; Pre: precipitation; Tpop: total population; dsr: distance to residential areas; Gdp: gross domestic product; DEM: digital elevation model.

.

The analysis revealed that several factors had an impact on the distribution of regional green spaces. Specifically, variables such as slope, elevation, distance to waterbody, distance to road, distance to railway, precipitation, and GDP demonstrated positive influences on regional green spaces. Conversely, temperature, population, and distance to residential areas had negative effects on the presence of regional green spaces. Notably, the elevation derived from the digital elevation model (DEM) and the urban population emerged as significant driving factors with higher explanatory power for regional green spaces, characterized by absolute β values of 42.021 and 7.025, respectively. These findings are consistent with prior research. Changzhou is predominantly composed of low-lying plains and hilly areas. During the early stages of urbanization, there was an accelerated exploitation of mountainous regions to meet the construction demands. The Mao Mountain mining area, in particular, experienced extensive exploitation, leading to the fragmentation and barrenness of certain mountainous areas, which resulted in significant fluctuations in the DEM elevation. Subsequently, in line with the concept of urban–rural integration and the objective of regreening regional green spaces, Changzhou initiated a comprehensive geological environment control project. This project focused on the closure of abandoned mine openings and the restoration of the Maolu Mining Treatment Zone. These efforts played a crucial role in rehabilitating the forested mountain areas, thereby contributing to the restoration and enhancement of green spaces.

During the process of urbanization, a negative correlation can be observed between the size of regional green spaces and the urban population. This relationship can be examined by considering the proximity of green spaces to residential areas. As urbanization progressed in Changzhou, both urban and peri-urban regions experienced an increase in population density, resulting in a growing demand for land. Regional green spaces, with their comparatively lower development and construction costs compared to other land uses, became attractive targets for developers. As a consequence, with the expansion of the urban population and the spread of residential areas, a significant portion of the original regional green spaces was encroached upon. This encroachment led to notable ecological and environmental changes, as well as a reduction in the extent of regional green spaces.

In general, the selected driving factors demonstrated ROC values exceeding 75% for various land types. This suggests that these driving factors can effectively explain the variations in regional green spaces and can be utilized for simulating future land use patterns.

3.5. Analysis of the Land Simulation Results under Different Scenarios

By 2022, the extent of regional green spaces in Changzhou had reached 204.46 km². Figure 7 showcases three scenarios depicting the land-use patterns of the city in 2032. It is apparent that the land-use composition in the upcoming decade will bear a strong resemblance to that of 2022, with cultivated land, construction land, and green spaces being the predominant categories. Nevertheless, concentrated transformations are expected, especially in relation to the expansion of construction land. The regions labeled as ①, ②, and ③ in Figure 7 represent the primary patches of regional green spaces in Changzhou. It is evident that under different simulated scenarios, the regional green spaces will exhibit distinct coupling structures compared to the predominant land-use types into which regional green spaces are primarily transformed, namely cultivated land and construction land.

In the scenario of inertial development, the projected area of construction land in Changzhou is estimated to reach 948.5757 km², while the regional green spaces will cover approximately 198.6 km². Figure 8 illustrates that the area of regional green spaces will be slightly larger by about 0.09 km² (approximately 2.87%) compared to 2022, whereas the area of water bodies will decrease by approximately 3.73 km². In the absence of policy constraints, the expansion of construction land will occur rapidly to meet the demands of urban development. Regional green spaces will be converted into cultivated land to fulfill agricultural needs and will also be appropriated for construction purposes, serving as the primary source of land for other land-use types influenced by urbanization. Examining the outcomes of land-use conversion, it is evident that the changes in regional green spaces will primarily manifest in the Liyang Tianmu Lake Ecological Area (marked as ②) and the Mao Mountain Ecological Area (marked as ③). These areas will experience a transition from previous external reductions to a combined reduction of both internal and external regions. If land-use conversion remains unrestricted, the regional environment will inevitably suffer significant damage, and the fragmentation of regional green spaces will intensify, undermining their inherent ecological functions.

In the economic priority scenario, the expansion of construction land will undergo significant changes. According to simulations, the projected area of construction land will reach 1178.1 km², reflecting a considerable increase of approximately 26.82 km². To account for different policies concerning Changzhou’s economic development, ecological system constraints were applied in this study. Consequently, in this scenario, the area of regional green spaces is expected to expand by about 211.3 km² (3.2%) compared to 2022. However, the closely associated water bodies will experience a substantial decrease of approximately 11.12 km², which will also exert a notable impact on the regional ecological system. Figure 7 provides a visual representation of the consequences of the economic priority scenario. Regional green spaces located on the outskirts of built-up areas within the designated area marked as ③ will be extensively utilized for other purposes, leading to significant reductions in their size. These reductions will particularly affect the areas surrounding Tianning district, Xinbei district, Zhonglou district, and other regions. Furthermore, the interior of core regional green spaces will also be impacted by urban development, resulting in the intersection of built-up areas with the Mao Mountain Scenic Area and Ge Lake Scenic Area.

The ecological priority development scenario was simulated in accordance with Changzhou’s environmental protection plan for the “Two Lakes” Innovation Zone. The simulation results indicate that this scenario successfully safeguards regional green spaces and other ecological land-use types. Figure 8 visually presents that, under the ecological priority scenario, the area of regional green spaces is projected to expand to 265.54 km², signifying a substantial increase of 29.87% compared to the year 2022. Conversely, the areas of other land types are expected to undergo minor fluctuations in comparison to the other two scenarios. The area of cultivated land will decrease by approximately 7.67 km², indicating a gradual trend of converting cultivated land back to forest land. The reduction in the area of water bodies, in comparison to the economic priority scenario (11.12 km² reduction) and the inertial development scenario (3.73 km² reduction), will better meet the requirements for regional ecological protection. Although cultivated land and construction land will undergo significant changes, their expansion rates will be effectively controlled. The spatial patterns of land types reveal that the area of green spaces surrounding core regional green spaces such as Mao Mountain, Ge Lake, Changdang Lake, and Tai Lake (marked as ② and ③) will increase, displaying enhanced internal connectivity. Additionally, there will be scattered expansions of regional green spaces near the Two Lakes and other ecological wetland areas. In summary, the ecological priority scenario promotes an increase in the area of regional green spaces and reduces the conversion of ecological land for construction purposes, thereby enhancing regional ecological safety.

In order to provide a more comprehensive analysis of the potential future developments of regional green spaces in Changzhou, we have expanded the descriptions of the three simulated scenarios and their underlying assumptions. The “Inertial Development Scenario” assumes a continuation of historical trends and development patterns, without significant policy interventions or prioritization of ecological protection. Under this scenario, the projected results indicate a moderate increase in the area of regional green spaces by 2032. However, it is accompanied by a substantial expansion of construction land and a decrease in water bodies. On the other hand, the “Ecological Priority Scenario” places a strong emphasis on the implementation of strict ecological protection measures. It involves adjustments to the Markov model’s transition probability matrix to favor the conservation and expansion of green spaces. The simulation results of this scenario suggest a notable 29.87% increase in the area of regional green spaces by 2032, with relatively minor fluctuations in other land-use types. Lastly, the “Economic Priority Scenario” concentrates on the impacts of economic development on urban expansion. This scenario allows for a higher probability of converting various land-use types, including green spaces, into construction land. While the results of this scenario project a considerable increase in the area of construction land, it also anticipates a 3.2% expansion in regional green spaces. However, it is important to note that there is a significant reduction in water body areas. By providing these detailed descriptions of the scenario assumptions and their respective projections, our aim is to offer a comprehensive understanding of the complex interplay between urban development, ecological protection, and the potential future trajectories of regional green spaces in Changzhou.

4. Results and Discussion

Between 1992 and 2022, there has been a consistent decrease in the extent of regional green spaces, declining from 248.23 km² in 1992 to 204.46 km² in 2022. Particularly noteworthy is the period from 2002 to 2012, which exhibited the highest reduction in the rate of change for the area of regional green spaces, with a decrease of −0.01% compared to other years. The process of urban–rural integration has played a significant role, prompting a greater focus on integrating the concept of an ecological civilization into urban and rural planning. This integration aims to ensure that the layout, construction standards, and development intensity align with ecological conservation requirements. Consequently, during the period from 2012 to 2022, there has been a more stabilized fluctuation in the dynamic degrees of regional green spaces. Although the overall trend in the study area indicates a decline, it is worth noting that certain regions have witnessed an expansion in their regional green spaces [47].

The analysis of changes in regional green spaces reveals an increasing trend in the number of units experiencing a reduction of 70% to 100% in their area from 1992 to 2012. Particularly, within the timeframe of 1992 to 2002, a substantial decline was observed in approximately 20.43% of the total regional green space rasters. The peripheries of regional green spaces in the studied region experienced extensive development, resulting in an overall decrease in the stability of regional green space areas. The majority of the increase rates in regional green spaces fell within the range of 0% to 30%, indicating a relatively modest expansion. On the other hand, units with increase rates ranging from 70% to 100% were sporadic, demonstrating an intermittent pattern of regional green space growth. Between 2012 and 2022, Changzhou implemented a series of policies, including the 14th Five-Year Plan for Ecological Environment Protection and the Ecological Environment Protection Plan for the “Two Lakes” Innovative Zone, which emphasized the theory of urban–rural integration. As a result, the increase rate of regional green spaces witnessed a slight rise, mainly concentrated in larger regional green space areas, such as the mountainous region within the Mao Mountain ecological area and the surrounding zones of major ecological wetlands.

The alterations in landscape pattern indices provide evidence that the regional green space patches underwent fragmentation, complexity, irregularity, scattering, and heightened interference from human factors between 1992 and 2022. Despite the implementation of ecological policies during 2012–2022, the regional green spaces, as a whole, displayed a prevailing trend of degradation.

The qualitative analysis of driving factors for green spaces in Changzhou highlights several key considerations. First, it is crucial to address abandoned mine openings by implementing measures such as repair and ecological landscape design. Similarly, quarries should be closed, and ecological restoration efforts should be undertaken on mountains and water bodies. In cases where there are originally abandoned mines, preservation should be prioritized while maintaining regional green spaces. Furthermore, restoration and management measures, including slope trimming, anchoring, enhancing green vegetation coverage on slopes, and soil improvement, should be implemented to ensure the rejuvenation of regional green spaces. The positive correlation with water systems emphasizes the need to promote afforestation along water bodies and strengthen the protection of landscape forest systems during the establishment of regional green spaces. Special attention should also be given to the ecological construction of water-coupled green spaces, such as the governance of green spaces along the Danjin Licao River. Moreover, ecological compensation can be utilized in the economic development zone of Changzhou to create regional green spaces. By treating Changdang Lake wetland, Ge Lake wetland, and Tai Lake wetland as three interconnected networks, ecological protection can be reinforced. Additionally, the periphery of built-up areas can be targeted for regreening efforts, and the conversion of some cultivated land to forest land can be pursued, resulting in a multi-point regreening ecological pattern.

Drawing from a comprehensive survey on the evolution trends and driving forces of regional green spaces in Changzhou, this study conducted simulations of three scenarios to examine the future states of land-cover/use types and their impact on regional green spaces in 2032. The scenarios presented diverse outcomes for regional green spaces. In the ecology-oriented scenario, significant growth in regional green spaces was observed, while the mode emphasizing economic development but within ecological redlines showed a slight increase. Conversely, the inertia-driven development scenario predicted a decrease in the area of regional green spaces. Analyzing the landscape pattern perspective, the encroachment on regional green spaces for construction purposes generally began at their edges. However, in the natural development mode, such encroachment was more evident from both the interior and the exterior of regional green spaces. As a Type I large city, Changzhou is experiencing progressive urbanization, rendering the expansion of construction land irreversible. Nevertheless, this continuous expansion poses a substantial threat to regional ecological safety. Therefore, it is essential to achieve a harmonious equilibrium between economic progress and ecological preservation. The planning of ecological redlines for regional green spaces, particularly along the three major green space networks (Changdang Lake, Ge Lake, and Tai Lake), should be prioritized. This planning should restrict the development of construction land within and adjacent to the ecological redlines. Within the Mao Mountain Scenic Area, specific actions are necessary, including defining the boundaries of eco-sensitive zones, rehabilitating abandoned mine openings, and promoting regreening of mining areas to increase regional green spaces. The simulation outcomes also demonstrated that the green spaces surrounding key urban regions (Tianning district, Zhonglou district, Xinbei district, and Wujin district) underwent significant reutilization and were predominantly converted into developed areas. Therefore, it is imperative to develop and enforce strict land use planning for green spaces around urban built-up areas to safeguard them against illegal occupation or excessive development. This planning should integrate regional green spaces with urban green spaces within built-up areas, establish ecological corridors, and enhance the connectivity of isolated green spaces to foster ecological connectivity. Ongoing monitoring of green spaces is essential to ensure the effectiveness of these measures. Lastly, promoting public participation in green space protection is critical, fostering public awareness of the significance of green spaces [48].

The methodological approaches employed in this study, such as the utilization of landscape pattern indices, the CLUE-S model, logistic regression, and the Markov–FLUS model, align with the simulation-based techniques used in prior research on the dynamics of green spaces in rapidly urbanizing areas. For instance, simulation modeling was effectively applied by Zhang et al. [49] to analyze the impact of infrastructural changes on environmental and logistical outcomes, providing a useful methodological reference for studying the dynamics of green spaces in rapidly urbanizing areas. Similarly, the spatiotemporal distribution and driving factors of regional green spaces during rapid urbanization in the Nanjing metropolitan area were examined by Liu et al. [50], offering valuable insights into the factors influencing green space dynamics in rapidly developing cities. Furthermore, urban growth patterns and the loss of urban green space in Kolkata, India, were assessed by Dinda et al. [51] using an integrated simulation approach and GIS-based analysis, providing a methodological framework that can be applied to similar studies in other rapidly urbanizing cities. By situating this study within this broader body of research, the findings presented here contribute to the growing understanding of the complex interplay between urbanization, environmental preservation, and the dynamics of regional green spaces in the context of China’s rapidly developing urban centers.

The findings of this study on the spatiotemporal dynamics and future scenarios of regional green spaces in Changzhou align with and build upon the existing body of research on this topic. A comprehensive review of the ecological functions of regional green spaces was conducted by Ding and Zhang [15], highlighting their importance for ensuring regional ecological safety and promoting integrated urban–rural development. The patterns identified in their review are corroborated by the observations made in this study regarding the fragmentation, complexity, and human interference impacting the regional green spaces in Changzhou. Similarly, the analysis of landscape pattern indices presented in this study echoes the work of Tang et al. [16], who utilized remote sensing technology to study the ecological networking and landscape patterns of green spaces in Xuzhou City. The trends observed in Changzhou, such as increasing fragmentation and decreasing connectivity of green spaces, align with their findings, underscoring the common challenges facing rapidly urbanizing cities in China. The scenario simulations conducted in this research provide new insights that expand upon the work of Kucsicsa et al. [18] and Kamal et al. [19], where similar modeling approaches were employed to investigate land-use changes and their impacts on green spaces in Romania and Bangkok, respectively. By considering the varying priorities of economic development and ecological preservation, a more nuanced understanding of the potential future trajectories of regional green spaces in Changzhou is offered in this study, highlighting the delicate balance required to ensure sustainable urban–rural integration. Furthermore, the qualitative analysis of the driving factors influencing regional green spaces, as presented in this manuscript, complements the research conducted by Nadoushan [21] and Hashemi Aslani et al. [20], who explored the relationships between land-use dynamics and underlying natural and socioeconomic drivers in Iran. The identification of elevation, urban population, and proximity to water bodies and transportation as key determinants for regional green spaces in Changzhou provides additional empirical evidence to support the understanding of these complex, multifaceted processes. By situating the findings of this study within the broader context of existing research, the valuable contributions this work makes to the field of urban planning and sustainable development are underscored. The comprehensive analysis of spatiotemporal trends, landscape pattern changes, driving factors, and future scenarios offers a robust framework for addressing the pressing environmental challenges faced by rapidly urbanizing Type I large cities in China and beyond.

Therefore, this study aimed to investigate the extent of change and pattern characteristics of regional green spaces in Changzhou, as well as to simulate the future regional green space patterns in the city over the next decade. The findings of this research provide valuable insights into the protection and development of Changzhou’s environment. However, it is important to acknowledge that the final results were influenced by challenges in accurately extracting the area of small urban green spaces in certain regions, primarily due to limitations in data acquisition and the resolution of remote sensing images. Future studies conducted by this research group will strive to address these limitations by obtaining more precise data and conducting comprehensive classification analyses of regional green spaces. Additionally, efforts will be made to evaluate the ecological suitability of the regional landscape and establish a safety pattern for regional ecology, contributing to a more comprehensive understanding of the region’s environmental dynamics

5. Conclusions

This study focuses on investigating the spatiotemporal dynamics and conducting scenario simulations of regional green spaces in Changzhou, a Type I large city in China experiencing rapid urbanization. The research holds significant relevance in addressing the pressing environmental challenges faced by major Chinese cities. The accelerated urbanization has led to the fragmentation and degradation of both urban and rural environments, necessitating the adoption of innovative strategies for managing green space resources. The concept of regional green spaces has emerged as a strategic solution to coordinate the preservation and development of ecological resources across the urban–rural continuum.

To analyze the evolution, driving factors, and future scenarios of regional green spaces in Changzhou, this study employed Landsat satellite imagery and various research methods, including landscape pattern indices, the CLUE-S model, logistic regression, and the Markov–FLUS model. The analysis reveals a consistent decline in the area of regional green spaces in Changzhou, decreasing from 248.23 km² in 1992 to 204.46 km² in 2022, with the most significant reduction observed between 2002 and 2012. Landscape pattern analysis indicates an increase in fragmentation, complexity, irregularity, and human interference within these green spaces. Noteworthy driving factors influencing changes in regional green spaces include elevation, urban population, and proximity to water bodies and transportation infrastructure. Scenario simulations present different perspectives on the future of regional green spaces in Changzhou; the ecological priority scenario projects a substantial increase, the economic priority scenario suggests a slight expansion, while the inertial development scenario predicts a continued decline.

Future research should investigate additional factors that may influence the dynamics of regional green spaces in Changzhou and other rapidly urbanizing cities in China. While the current study has identified significant drivers, such as elevation, urban population, and proximity to water bodies and transportation, it is important to consider other variables, including socioeconomic factors, policy-related influences, and ecosystem services. Expanding this investigation to encompass other Chinese cities with diverse characteristics would yield a more comprehensive understanding of the intricate interactions that shape regional green space landscapes across various urban contexts. Furthermore, assessing the ecological suitability and safety of the regional landscape, encompassing evaluations of habitat quality, biodiversity, and critical ecological corridors, would contribute to the formulation of well-informed conservation strategies. Notably, fostering increased public participation and awareness in green space protection is essential for the long-term sustainability of these invaluable resources. Strategies such as community-based stewardship programs, educational campaigns, citizen science initiatives, and the integration of green space planning into local decision-making processes can empower residents and ensure that their needs and concerns are considered in the management of regional green spaces. By pursuing these future research directions and implementing innovative public engagement approaches, the findings of this study can be further strengthened and applied to support the sustainable development of Changzhou and other rapidly urbanizing cities in China.

While this study provides comprehensive insights into the dynamics of regional green spaces in Changzhou, it acknowledges limitations related to the accuracy of extracting small urban green spaces and the resolution of remote sensing data. Future research should aim to address these limitations by acquiring more precise data and conducting comprehensive classification analyses of regional green spaces. Furthermore, evaluating the ecological suitability of the regional landscape and establishing a safety pattern for regional ecology would contribute to a more comprehensive understanding of environmental dynamics in the region.

The results of this research may hold practical significance for urban planners, policymakers, and environmental managers engaged in formulating strategies for promoting sustainable development in rapidly urbanizing metropolitan areas. By integrating the theory of urban–rural integration with the protection and expansion of regional green spaces, this research offers a framework for addressing the pressing environmental challenges faced by Type I large cities in China and beyond. Ultimately, it contributes to the creation of livable, resilient, and ecologically balanced urban–rural communities.