Study on the spatiotemporal evolution of vegetation and the spatial heterogeneity of its influencing factors based on land use policy zoning in Anhui Province: providing new ideas for vegetation protection

Identifying the primary factors influencing changes in vegetation cover within different land use policy zones, along with their spatial heterogeneity and the spatiotemporal evolution characteristics of past and future vegetation cover, is essential for effective Fractional Vegetation Cover (FVC) management. This study utilizes a Self-Organizing Feature Map (SOFM) neural network to identify land use policy zones in Anhui Province. By coupling the SEN+MK trend test with the Hurst index, the study predicts future vegetation change trends. A multiscale geographically weighted regression model is employed to reveal the spatial heterogeneity of the influencing factors of FVC, and a geographical detector is used to analyze the main influencing factors in different land use policy zones. The results indicate the following: From 2000 to 2023, the FVC in Anhui Province predominantly ranged between 0.4 and 0.6. Over these 24 years, the highest average FVC was found in the core ecological protection zone. The areas with the most significant improvement and fluctuation were located in the core agricultural production zone. Future improvements and degradations are primarily concentrated in the core agricultural production zone. The main driving factors differ across various land use policy zones. Climate factors dominate in Anhui Province as a whole, the core agricultural production zone, and the ecological protection transition zone. Topographical factors are predominant in the core ecological protection zone, whereas socioeconomic factors are the primary drivers in the core economic development zone. This study provides theoretical guidance for optimizing vegetation protection policies, coordinating human-environment relationships, and achieving sustainable development across different policy zones.

1 Introduction

In terrestrial ecosystems, vegetation serves as a pivotal constituent, profoundly influencing global environmental dynamics, carbon cycling mechanisms, and climate regulation functions (Shen et al., 2022). The FVC, a crucial indicator measuring vegetation growth status and density, has increasingly become a fundamental basis for monitoring ecosystem transitions and assessing the sensitivity of global environmental changes (Xu et al., 2018; Chu et al., 2019; Bai, 2021). Given the frequent occurrence of extreme climate events in recent years, the role of FVC in revealing ecosystem responses and driving mechanisms has become increasingly prominent. Research on FVC is significant for comprehending regional and even global ecosystem dynamics, unraveling the underlying mechanisms of ecological element cycling, and guiding the formulation of ecological environmental protection and socio-economic sustainable development strategies (He et al., 2021; Mao et al., 2022).

Land, the cornerstone of human society’s existence and development, exerts direct or indirect impacts on ecological processes, which further manifest in fluctuations in the quantity and quality of FVC, thereby influencing food security, economic development trends, and ecological security patterns. Beyond being a spatial carrier for socio-economic development, land morphology evolution mirrors the shifts in policy orientations and development strategies within specific geographical regions, objectively reflecting the dynamic adjustments in socio-economic structures (Liu et al., 2020). Critically, the optimal allocation of land use structures, a key means to regulate regional ecosystem integrity and stability, often accompanies corresponding changes in FVC, with these variations displaying notable differences under varying land use policy frameworks. Therefore, a profound analysis of the evolution trends and driving mechanisms of FVC within different land use policy zones holds immense value for promoting vegetation conservation, optimizing agricultural production layouts, and fostering coordinated regional economic development.

With the rapid advancements in remote sensing technology and the profound integration of multi-source remote sensing data with Geographic Information Systems (GIS), remarkable progress has been achieved in FVC research across global, national, and watershed spatial scales (Ouyang et al., 2010; Pei et al., 2021; Shi et al., 2022). These studies have not only revealed the distribution characteristics of FVC in different regions but also delved into the multifaceted factors driving its changes, emphasizing the central role of FVC in ecological security and economic development (Omer et al., 2020; Wei et al., 2021). However, current research primarily focuses on the long-term spatio-temporal evolution of FVC within a single scale (Shi et al., 2019; Chen et al., 2023; Pandey et al., 2023), with insufficient attention paid to cross-scale dynamics, particularly FVC variations in the context of different land use policy zones. Additionally, in the practical delineation of land use policy zones, conventional methods often focus on the static attributes within individual patches (the vertical dimension), such as those evaluated in traditional land suitability assessments, at the expense of the dynamic transitions of land-use types between patches (horizontal transfers). This oversight, to some extent, limits the comprehensiveness and depth of research.

To overcome these limitations, this study innovatively introduces the SOFM neural network, aiming to scientifically delineate different land use policy zones by recognizing attribute similarities and pattern characteristics during land use transitions. Furthermore, by integrating advanced analytical tools such as the Seasonal and Trend decomposition using Loess (STL), Mann-Kendall (MK) trend test, and Hurst index, this study aims to establish a comprehensive predictive framework to accurately forecast the future trends of FVC within different land use policy zones. This methodological system not only enriches the theoretical perspectives of FVC research but also provides robust technical support and data-driven insights for deeply understanding the spatio-temporal evolution patterns of FVC and scientifically formulating vegetation management policies and regional development strategies.

Investigating the intricate spatial non-stationarity between FVC and external environmental factors, current research has increasingly acknowledged the limitations of traditional methods such as the Globally Weighted Regression (GWR) model in handling multi-scale effects. The GWR model, grounded on the assumption of uniform driver scales, fails to fully capture the heterogeneity of relationships across different scales, thus limiting its effectiveness in explaining the mechanisms underpinning FVC variations (Wang et al., 2022). In response, this study introduces the Multi-scale Geographically Weighted Regression (MGWR) model, which, through the concept of variable bandwidths, offers an innovative pathway for unraveling the multi-scale characteristics of spatial interactions within FVC (Fotheringham et al., 2017; Niu et al., 2021). Coupled with the Geodetector technique, this research endeavors to delve into the spatial heterogeneity and interaction mechanisms among FVC’s influencing factors, ultimately providing a scientific basis for the formulation of precise vegetation management strategies.

As a pivotal component of the Yangtze River Delta economic zone, Anhui Province not only holds a strategic position in China’s economic development but is also intimately connected to multiple economically developed regions due to its unique geographical location (Huang and Gui, 2018; Chen et al., 2020). Boasting 59 significant wetlands, this region is rich in wetland resources, serving as a treasure trove of biodiversity and indispensable habitats for rare migratory birds, underscoring its remarkable ecological service functions (Liu, 2016). Although ecological restoration projects like “returning farmland to forests” have significantly improved Anhui’s ecological environment, issues such as excessive land development intensity, water scarcity, and habitat degradation remain pressing concerns. Consequently, continuous monitoring of vegetation cover changes in Anhui Province is crucial not only for regional ecological protection and enhancing the quality of economic development but also for safeguarding national and even global ecological balance (Yu et al., 2021; Zhang and Jin, 2021).

Based on a case study of Anhui Province, this research is designed to achieve the following three core objectives: (1) Delineation of Land Use Policy Zones: By systematically investigating the dynamic transition relationships of land use types within administrative units from 2000 to 2023, a SOFM neural network will be employed to conduct cluster analysis and delineate distinct land use policy zones across Anhui Province. (2) Analysis of FVC Temporal Evolution Characteristics: Integrating Sen’s slope estimator, Mann-Kendall trend analysis, the coefficient of variation (CV), and the Hurst index, this study aims to reveal the spatiotemporal evolution characteristics of FVC over the past 24 years, and to identify areas at potential risk of future FVC degradation within each land use policy zone. (3) Exploration of Spatial Heterogeneity in Influencing Factors: Utilizing Geographical Detector and MGWR techniques, this research will thoroughly dissect the dominant factors driving FVC changes and their spatial heterogeneity across different land use policy zones. By quantifying the contribution rates and interaction effects of various factors at multiple scales, this study will provide a theoretical foundation for formulating targeted vegetation conservation strategies. Collectively, this work offers a novel perspective for understanding the spatiotemporal evolution and underlying mechanisms of FVC in Anhui Province, while also supplying scientific evidence and technical support for promoting coordinated and sustainable development of the regional economy, agriculture, and ecological systems.

2 Materials and methods

2.1 Study area

Located in the eastern geographical region of China, Anhui Province encompasses a broad range of longitudes, spanning from 114°54′E to 119°37′E, and latitudes, extending vertically from 29°41′N to 34°38′N, thereby demonstrating a remarkable geographical span and diversity. The province features a complex terrain with various landforms including plateaus, hills, plains, and mountains, making it representative for research purposes (Wang et al., 2023). Anhui has an extensive network of rivers and lakes, including the Yangtze River, Huai River, and Xin’an River basins, contributing to diverse wetlands and habitat types (Teng et al., 2022). Situated in a transitional zone between subtropical and warm temperate climates, as well as humid and semi-humid areas, Anhui’s climate and vegetation are distinctly characterized. The annual mean temperature is demarcated within a range of 14 °C–17 °C, while the annual precipitation varies between 800 mm and 1600 mm, displaying a prominent north-to-south climatic gradient, as documented by Jiang and He (2023). The province is divided into three main geographical regions: Northern Anhui, Central Anhui, and Southern Anhui. As of December 2023, Anhui has 59 significant wetlands covering a total area of 1.6037 million hectares, comprising 16 prefecture-level cities and 104 counties. These wetlands account for 2.85% of the total wetland area in China, making Anhui one of the provinces with the richest wetland resources in the country. Notably, Shengjin Lake Wetland in Chizhou City is an internationally important wetland. Hydrological analysis reveals that the Huai River Basin boasts 30 wetlands, the Yangtze River Basin possesses 28, and the Xin’an River Basin holds only 1 wetland. Regionally, the number of wetlands in Northern Anhui (Wanbei) amounts to 26, 16 are situated in Central Anhui (Wanzhong), and Southern Anhui (Wannan) boasts 17 wetlands (refer to Figure 1). Anhui Province plays a crucial role in ecological protection within the Yangtze River Delta and serves as an essential wintering and habitat site for migratory waterbirds along the East Asian-Australasian Flyway.

Figure 1

Figure 1. Study area location diagram.

2.2 Data sources

This study employs a multi-source dataset encompassing land use, vegetation coverage, and various influencing factors. The specific data sources and processing methodologies are detailed as follows: (1) The provenance of land use data can be traced back to the Resource and Environment Science and Data Center of the Chinese Academy of Sciences (http://www.esdc.cn), spanning a temporal range from 2000 to 2023. These data have undergone rigorous classification, categorizing them into six primary types: arable land, forest land, grassland, water bodies, built-up land,and unutilized land. Each category of data possesses a high degree of spatial accuracy, with a resolution precision of 30 m, ensuring precision and reliability in geospatial analysis (Qu et al., 2025).(2) The acquisition of FVC data relies on the advanced remote sensing cloud computing platform, Google Earth Engine (GEE), which maintains a spatial resolution of equally high precision at 30 m (3) Factors influencing FVC include topography, climate, and socio-economic factors. The data acquisition methods are listed in Table 1, and the spatial distribution is shown in Figure 2. To unify the data format and ensure the consistency of spatial scope, all the data in this paper are projected into the Albers_Conic_Equal_Area coordinate system, and the spatial resolution is uniformly set to 30 m × 30 m, which is consistent with the spatial resolution of the FVC.

Table 1

Table 1. Impact factor data source table.

Figure 2

Figure 2. Spatial distribution map of influencing factors in 2023.

2.3 Methods

Figure 3 presents a meticulously crafted technical roadmap for this study, centered on three core dimensions: data source integration, analysis of the spatiotemporal dynamics of FVC across diverse land use policy zones, and an in-depth exploration of the dominant factors and spatial differentiation characteristics influencing FVC within these specific regions. Regarding data sources, a multifaceted dataset has been compiled, consisting of meticulously processed land use data, FVC data, and ten driving factor datasets that encompass a wide range of influencing factors. In investigating the spatiotemporal evolution of FVC, this study innovatively integrates the advanced interpretive capabilities of the GEE platform, the robustness of the Sen+MK trend analysis method, the precise quantification of data dispersion offered by the CV, and the predictive power of the Hurst index for future trends, jointly constructing a comprehensive analytical framework for assessing the stability and trend changes in long-term FVC series under varying land use policies in Anhui Province. Furthermore, in dissecting the factors influencing FVC, this study employs cutting-edge techniques such as the MGWR model and geographical detector analysis, aiming to profoundly reveal the heterogeneity in FVC spatial distribution and pinpoint the dominant factors that impact FVC within various land use policy zones in Anhui Province, thereby providing a scientific basis for policy formulation and regional management.

Figure 3

Figure 3. Research technology roadmap.

2.3.1 Land use policy zoning identification

The SOFM neural network, a competitive neural network architecture under an unsupervised learning paradigm, is characterized by its high adaptability, spontaneous organization properties, and exceptional self-learning capabilities. The SOFM network architecture consists of topological structured input and output layers, with its core operational mechanism relying on “predefined learning rules”—an iterative algorithm based on competition and cooperation. Specifically, the network first calculates the similarity between input vectors and the weight vectors of output layer neurons to select the Best Matching Unit (BMU) as the winning neuron. Subsequently, a neighborhood function is applied to adjust the connection weights of the winning neuron and its adjacent nodes, enabling their topological configuration to better approximate the feature distribution of the input patterns. Through this mechanism, the SOFM autonomously extracts the intrinsic structure of input data, achieving unsupervised clustering and visual representation of complex patterns. In this study, we leveraged the SOFM implementation embedded within the R programming platform to conduct an in-depth analysis of the similarities in land use pattern changes among the 104 counties in Anhui Province. Based on the similarity assessment, we employed clustering strategies to group counties with similar characteristics into the same type clusters, while assigning those with significant dissimilarities to distinct type clusters. Furthermore, through a thorough examination of land use type transitions in Anhui Province spanning from 2000 to 2023, we identified six predominant land use conversion patterns: A - conversion of arable land to forest land, B - conversion of arable land to water bodies, C - conversion of arable land to built-up land, D - conversion of forest land to arable land, E - conversion of forest land to built-up land, and F - conversion of built-up land to arable land. Drawing upon similarity analyses and pattern recognition techniques of these conversion patterns, we successfully delineated land use policy zones within Anhui Province, a process efficiently facilitated by the “kohonen” package in R (Table 2). In this study, the input layer of the SOFM model employs 104 county-level administrative units in Anhui Province as analytical samples. The input variables comprise six principal categories of land-use conversion and their respective areas during the period 2000–2023, serving to characterize the spatiotemporal dynamics of land-use transformation across regions. The output layer adopts a two-dimensional topological structure, facilitating the clustering and visualization of land-use transition features for each administrative unit. Ultimately, four categories of land-use policy zones are delineated based on the clustering outcomes.

Table 2

Table 2. Transfer area of main land use types in Anhui Province from 2000 to 2023.

2.3.2 Vegetation coverage inversion

Within the GEE platform’s data processing pipeline, Landsat satellite imagery spanning from 2000 to 2023 underwent systematic preprocessing measures, encompassing the elimination of cloud interference, the correction of radiometric effects, and the compensation for atmospheric factors, all aimed at ensuring data quality. Subsequently, algorithmic models were employed to extract the minimum, maximum, and average values of the Normalized Difference Vegetation Index (NDVI) for each year within the study area, providing a solid foundation for an in-depth analysis of the dynamic changes in vegetation cover (Peng et al., 2019). Using the annual maximum value method to eliminate cloud and atmospheric interference, NDVI values for Anhui from 2000 to 2023 were obtained. The FVC was then estimated using the pixel dichotomy model (Jiang et al., 2021; Nie et al., 2021). The specific formula for this estimation is presented as follows Formula 1:

$FVC=\left(NDVI-NDV{I}_{\mathit{min}}\right)/\left(NDV{I}_{\mathit{max}}-NDV{I}_{\mathit{min}}\right)(1)$

In this computational model, FVC is designated as the pivotal metric for quantifying vegetation cover extent, while NDVI serves as a quantitative representation of vegetation vigor and density. Specifically, NDVI_min is demarcated as the lower threshold of NDVI values observed in bare soil regions, and NDVI_max represents the upper threshold of NDVI values found in densely vegetated areas. This study rigorously selects the 5th and 95th percentiles of the NDVI value distribution as the scientific basis for these two thresholds, respectively.

2.3.3 Sen+MK trend analysis

This study delves into the variability and statistical significance of FVC at the pixel level, achieving precise characterization and evaluation of FVC dynamics through the integration of Sen’s trend analysis and the MK non-parametric test (Gu et al., 2018; Li et al., 2019). Furthermore, to elucidate the overall trend of FVC changes at the regional scale, a linear regression model is introduced for quantitative analysis. The formulas are expressed as follows Formulas 2–5:

$S_{FVC}=median\frac{FV{C}_j-FV{C}_i}{j-i}(2)$

$Z=\left\{\begin{array}{cc}\frac{S-1}{\sqrt{s\left(S\right)}},& S>0\\ 0,& S=0\\ \frac{S+1}{\sqrt{s\left(S\right)}}& S<0\end{array}\right.(3)$

$S={\displaystyle \sum _{j=1}^{n-1}}{\displaystyle \sum _{i=j+1}^n}sgn\left(FV{C}_j-FV{C}_i\right)(4)$

$s\left(S\right)=\frac{n\times \left(n-1\right)\times \left(2n+5\right)}{18}(5)$

Within the framework of the aforementioned formulas, S_FVC stands for the trend of vegetation change, FVC_i and FVC_j distinctly represent the quantitative values of FVC for a specific pixel point in the time series at years i and j, respectively. The parameter n, serving as the measure of the time series length, constitutes a fundamental basis for the analysis. The function sgn functions as a discriminator of the directionality of differences, while the statistic Z encompasses a range from negative to positive infinity, comprehensively covering all possible trends of change.

Theil Sen-MK trend analysis was realized by Matlab programming, and the result of vegetation change trend was obtained. Table 3 presents a five-level classification system.

Table 3

Table 3. Vegetation trend change classification table.

2.3.4 Coefficient of variation

The CV quantifies the relative variability of observed data, serving as an indicator of data stability (Vicente-Serrano et al., 2016; Zheng et al., 2019). A high CV denotes significant data fluctuations, indicative of notable changes in vegetation. In contrast, a low CV points to minor data variations and stable vegetation patterns. The calculation formula is expressed as follows Formulas 6, 7:

$\sigma =\sqrt{\frac{{\displaystyle {\sum }_{i=1}^n}\left(FV{C}_i-\overline{FVC}\right)}{n-1}}(6)$

$C_v=\frac{\sigma }{\overline{FVC}}(7)$

In the formula, n represents the observation year, σ represents the standard deviation, FVC_i denotes the vegetation coverage in year i, $\overline{FVC}$ represents the average coverage, and $C_v$ represents the CV.

2.3.5 Sen+MK trend test coupled with hurst index

The Hurst index is a robust method for quantitatively characterizing the long-range dependence of time series (Qu et al., 2020). In recent years, it has been increasingly applied in studies investigating the future sustainability of changes in vegetation cover. This study utilized MATLAB for the calculation of the Hurst index, abbreviated as the H-value, to determine the persistence of time series in vegetation cover. Values of H range from 0 to 1 and are divided into three categories: 0 < H < 0.5 (indicating anti-sustainability), H = 0.5 (signifying no significant change), and 0.5 < H < 1 (implying persistence). An elevated H-value suggests a higher degree of persistence, whereas a lower H-value indicates less persistence.

The coupling analysis of SEN+MK trend test and Hurst index is carried out to obtain the sustainable trend of FVC change in the future. It is divided into five levels, as shown in Table 4.

Table 4

Table 4. Classification of future vegetation change trends.

2.3.6 Multi-scale geographical weighted regression

The GWR model holds a prominent position in geographical research due to its capability of modeling spatial non-stationarity using data from proximate observation points. However, a notable limitation of this model in handling multivariate data lies in its adoption of a uniform bandwidth setting, which overlooks the potential distinct spatial scales associated with different independent variables. To address this issue, the MGWR model has been introduced, allowing for the concurrent optimization of multiple bandwidths to precisely capture the unique spatial heterogeneity and scale effects of individual variables. This study aims to employ the MGWR model framework to delve into the intricate spatial non-stationarity and scale dependence between the spatial distribution patterns of vegetation cover and its influencing factors in Anhui Province, with the mathematical model representation detailed subsequently:

$y_i={\displaystyle \sum _{j=1}^k}{\beta }_{bwj}\left(u_i,v_i\right)x_{ij}+{\epsilon }_i(8)$

In the Formula 8, $y_i$, $x_{ij}$, ${\epsilon }_i$ represent the dependent variable and independent variable at location i in space, respectively, with $\left(u_i,v_i\right)$ being the random error at location i; j denotes the number of independent variables, β_j is the regression coefficient at location i for the j variable, and β₀ is the intercept. ${\beta }_{bwj}$ represents the bandwidth used for the regression coefficient of the j variable.

2.3.7 GeoDetector

The GeoDetector model, a sophisticated statistical analysis tool, is primarily designed to delve into the spatial distribution patterns of a specific object and the underlying driving mechanisms. In its examination of influencing factors, this model employs the q-value as a quantitative metric to precisely evaluate the contribution and driving force of individual factors towards the spatial variability of FVC within the study area. Furthermore, GeoDetector boasts advanced capabilities in detecting interaction effects among factors, systematically identifying and analyzing the interplay and synergistic effects between different factors in driving FVC spatial variability. This comprehensive approach provides robust support for gaining a profound understanding of the spatial complexity inherent in vegetation distribution. The model expression is:

$q=1-\frac{{\displaystyle {\sum }_{h=1}^L}{N}_h{\sigma }_h^2}{N{\sigma }^2}(9)$

In the Formula 9 framework, q is defined as a pivotal metric assessing the explanatory power of influencing factors on the spatial variability of FVC, with its range strictly confined to values between 0 and 1. A higher q value signifies a more pronounced potential influence of the independent variable X on the dependent variable Y (i.e., FVC), whereas a lower q value implies a relatively weaker influence. N_h represents the number of sampling units within the h layer; h represents the stratification or classification of the independent variable X; L represents the total number of independent variables X, and h stands for one of the types among them. ${\sigma }_h^2$ and ${\sigma }^2$ are the variances of FVC in layer h and the study area, respectively.

3 Results and analysis

3.1 Identification of different land use policy zoning

Based on land use changes reflecting human socio-economic activities and the current development status of Anhui Province, SOFM was applied to identify six clusters using the 104 counties of Anhui Province as administrative units. Based on a comprehensive analysis of the current situation in Anhui Province, this study has definitively delineated four strategic land use zones: the Core Agricultural Production Area, the Ecological Protection Buffer Zone, the Core Ecological Protection Area, and the Core Economic Development Zone. The detailed classification criteria are comprehensively documented in Table 5 and visually illustrated in Figure 4, aiming to provide a rigorous and comprehensive framework for land use policy zoning.

Table 5

Table 5. Land use function zoning.

Figure 4

Figure 4. Self-organizing map of land use policy zoning.

As shown in Figure 5, the land use policy zones in Anhui Province exhibit certain clustering characteristics. The Core Agricultural Production Zone is mainly concentrated in parts of Northern and Central Anhui, characterized by the conversion of built-up land to arable land, and includes 15 counties. The Ecological Protection buffer Zone is primarily located in Central and Southern Anhui, with a few areas in Northern Anhui. This zone emphasizes the coordinated development of agricultural production, economic development, and ecological protection, covering a large area and comprising 61 counties. The Core Ecological Protection Zone is mainly situated in the Dabie Mountains and Southern Anhui, where the ecological value is high and requires focused environmental protection, encompassing 15 counties. The Core Economic Development Zone is predominantly found in the Central Anhui cities of Hefei, Wuhu, and Chuzhou, which are part of the Yangtze River Delta urban agglomeration and are experiencing rapid economic growth, including 13 counties.

Figure 5

Figure 5. Land use policy zoning map of Anhui Province.

3.2 Analysis of temporal and spatial evolution characteristics of FVC

3.2.1 Spatio-temporal evolution of average FVC

This study, leveraging the GEE cloud computing platform, conducted a thorough analysis of the spatial distribution characteristics of the annual mean FVC in Anhui Province from 2000 to 2023, and subsequently generated a detailed geospatial distribution map (see Figure 6). The map clearly illustrates a prominent pattern in the spatial distribution of FVC in Anhui Province: the regions at the southern and northern extremities exhibit higher levels, while the central belt comparatively demonstrates lower coverage. Specifically, the FVC in the cities of Anqing, Tongling, Wuhu, Ma’anshan, Hefei, Huainan, and Bengbu, which span the Yangtze River, Huai River, and Chaohu Lake basins, ranges mainly between 0–0.2 (Figures 6c,d). According to Figures 6a,b, the FVC in the southern part of Liuan, northern Anqing, Chizhou, Xuancheng, and Huangshan is mainly between 0.8 and 1, which is the highest, covering about 19.90% of the total area. The FVC in Fuyang, Bozhou, Huaibei, and Suzhou is relatively high (0.6–0.8), accounting for approximately 19.54% of the total area. The FVC in the regions north of Liuan, north of Hefei, and Chuzhou is lower, ranging between 0.2–0.4.

Figure 6

Figure 6. Spatial distribution of average FVC in Anhui Province from 2000 to 2023. (a-d) High average FVC values indicate an area.

By analyzing the FVC proportion in various land use policy zones (Figure 7), it was found that high FVC areas (0.8–1) are mainly concentrated in the Core Ecological Protection Zone, accounting for about 60.95%, followed by the Ecological Protection buffer Zone, Core Economic Development Zone, and Core Agricultural Production Zone. This is primarily because the Core Ecological Protection Zone is located in the Dabie Mountains and southern Anhui, where there is extensive forest cover and high vegetation coverage. Higher FVC areas (0.6–0.8) are mainly concentrated in the Core Agricultural Production Zone, covering about 28.41% of the area. This is mainly because the Core Agricultural Production Zone is located in northern Anhui, where arable land is abundant and vegetation coverage is relatively high.

Figure 7

Figure 7. FVC proportion of different land use policy areas in Anhui Province.

3.2.2 Spatio-temporal evolution trend of FVC

The analysis of the spatial change trend of FVC in Anhui Province from 2000 to 2023 (Figure 8) shows that the overall trend of vegetation change is positive, with the total area of regions showing FVC improvement accounts for approximately 44.64%. Specifically, Figure 8a indicates that the intersection of Wuhu, Ma’anshan, and Xuancheng cities experienced significant FVC degradation, marking it as a severely degraded area, covering 0.38% of the total area. There is a slight degradation trend in areas north of Wuhu, at the border of Ma’anshan and Chuzhou, and the border of Anqing and Chizhou, accounting for about 4.28% of the total area. Figures 8b–d show that cities such as Liuan, Bozhou, Huaibei, Bengbu, Suzhou, and Chuzhou primarily exhibit significant vegetation improvement trends, covering 13.14% of the total area.

Figure 8

Figure 8. Spatial distribution of FVC trend change in Anhui Province. (a-d) High FVC trend change values indicate an area.

By analyzing the proportion of FVC change trends in various land use policy zones (Figure 9), it was found that the area with the most significant improvement is concentrated in the Core Agricultural Production Zone, accounting for 63.38% of the area, followed by the Ecological Protection buffer Zone, Core Economic Development Zone, and Core Ecological Protection Zone. The area with the most significant degradation is concentrated in the Core Ecological Protection Zone, accounting for 6.71% of the area, indicating the need for focused protection in the future.

Figure 9

Figure 9. FVC change trend ratio of different land use policy areas in Anhui Province.

3.2.3 FVC change stability

This Study analyzed the stability of FVC in Anhui Province from 2000 to 2023 via Matlab programming, calculating the CV for each pixel. The stability is meticulously categorized into five tiers: the low fluctuation range (<0.05), the relatively low fluctuation interval (0.05–0.10), the moderate fluctuation threshold (0.10–0.15), the relatively high fluctuation stage (0.15–0.20), and the high fluctuation domain (>0.20). As shown in Figure 10, the vegetation changes in Anhui Province during this period exhibited considerable fluctuation. High fluctuation areas were mainly distributed in northern and central Anhui, accounting for the largest proportion at 36.40%. As depicted in Figures 10a–d, the northern regions, including Fuyang, Bozhou, Liuan, Huainan, Bengbu, Chuzhou, Ma’anshan, and Wuhu, showed significant fluctuation and require special attention. Regions exhibiting relatively high fluctuations account for 17.71% of the total area, spatially concentrated in the urban clusters of Fuyang, Bozhou, Huaibei, and Suzhou, located in the northern part of Anhui Province. In summary, the vegetation in Anhui Province over the past 24 years has experienced significant fluctuation, with areas having a stability coefficient greater than 0.15 accounting for 54.11% of the total area.

Figure 10

Figure 10. Spatial distribution of FVC change stability in Anhui Province. (a-d) High FVC change stability values indicate an area.

Through a rigorous analysis of the stability of FVC change proportions within various land use policy zones (as depicted in Figure 11), it is evident that areas characterized by high fluctuation are predominantly concentrated in the Core Agricultural Production Zone, accounting for 57.50% of its total area. Following this, are the Core Economic Development Zone, the Ecological Protection Buffer Zone, and the Core Ecological Protection Zone, each exhibiting distinct fluctuation patterns. Areas with low fluctuation are mainly concentrated in the Core Ecological Protection Zone, accounting for 31.97% of the area, indicating relatively minor fluctuation.

Figure 11

Figure 11. FVC change stability ratio of different land use policy areas in Anhui Province.

3.2.4 FVC future change trend analysis

By coupling the analysis of FVC change trends with the Hurst index, the spatial distribution of future FVC trends in Anhui Province was determined (Figure 12). As shown in Figure 12, the dominant trend for future FVC in Anhui Province is indetermination, covering 51.22% of the area and spanning all cities and counties in the province. The second largest area, accounting for 27.19% of Anhui’s total area, is where the trend shifts from improvement to degradation. Figures 12a–d indicate that these regions are mainly concentrated in the northern and southern parts of Anhui, particularly in Bozhou, Huaibei, Suzhou, and Bengbu in the north, and Liuan, Anqing, Tongling, Xuancheng, and Huangshan in the south. Areas with a persistent improvement trend occupy 17.35% of the province, primarily in Fuyang, Bengbu, and Huainan, with fewer regions in other areas.

Figure 12

Figure 12. Spatial distribution of future FVC change trend in Anhui Province. (a-d) High future FVC change trend values indicate an area.

By analyzing the proportions of future FVC change trends across different land use policy zones (Figure 13), it was found that the areas likely to shift from improvement to degradation are mainly concentrated in the Core Agricultural Production Zone, accounting for 35.86% of the area, followed by the Core Ecological Protection Zone, Ecological Protection buffer Zone, and Core Economic Development Zone. In the future, it is essential to strengthen the protection of arable land in the Core Agricultural Production Zone, enhance the quality of arable land, and prevent vegetation degradation.

Figure 13

Figure 13. FVC future change trend ratio of different land use policy areas in Anhui Province.

3.2.5 Spatial autocorrelation characteristics of vegetation coverage

Spatial autocorrelation techniques provide a powerful method for analyzing the correlation between similar characteristics across different spatial units. This study utilizes Moran’s I statistic to conduct a global spatial autocorrelation analysis, aiming to delve into the spatial distribution characteristics of FVC in Anhui Province, with a particular focus on identifying its underlying clustering patterns. It was found that the Moran’s I index for FVC in Anhui Province is 0.27, with a z-value of 104, exceeding the statistically significant threshold (P < 0.05). This finding emphasizes that the distribution of FVC has significant positive spatial autocorrelation, revealing a spatial clustering pattern among regions with similar values.

The local Moran’s I analysis of the average FVC from 2000 to 2023 identified major patterns of high-high clusters, high-low clusters, and non-significant clusters, with clear boundaries between these classifications (Figure 14). Figure 14 provides a clear illustration of the “high-high” clustering phenomenon of FVC in Anhui Province, which is primarily concentrated in the northwestern and southern regions of the province. Notably, these clusters are situated within the core areas of agricultural production and ecological protection, encompassing cities such as Fuyang, Bozhou, Anqing, Chizhou, Huangshan, and Xuancheng, forming a prominent high-FVC belt. These zones not only individually exhibit remarkable vegetative abundance but are also spatially contiguous, surrounded by geographical units that possess similarly elevated FVC values. In contrast, the “high-low” FVC clusters emerge distinctly in the central region of Anhui Province, specifically within the ecological protection buffer zone, demonstrating a significant divergence in vegetation cover from their immediate surroundings. These areas have high FVC levels, while the neighboring areas have lower FVC levels. The “low-low” FVC clusters are scattered, with a higher concentration in the Core Economic Development Zone among the land use policy zones.

Figure 14

Figure 14. Cluster distribution of FVC spatial autocorrelation analysis in Anhui province.

3.3 Analysis of factors affecting FVC

3.3.1 Spatial differentiation of FVC influencing factors

By applying the MGWR model, a comprehensive spatial heterogeneity analysis was conducted on the factors influencing FVC across 104 counties in Anhui Province. This model precisely estimated the regression coefficients for each driving factor and ingeniously utilized visualization techniques to intuitively illustrate the intricate spatial interdependencies between each factor and FVC (as detailed in Figure 15), thereby providing a solid data foundation for a deeper understanding of regional ecological dynamics. As shown in Figure 15, these factors exhibit significant spatial heterogeneity in Anhui Province, affecting FVC in a non-stationary manner across different regions. The impacts of topographic, climatic, and socio-economic factors on FVC vary distinctly.

Figure 15

Figure 15. Distribution of regression coefficients of vegetation impact factors in Anhui Province. (a) Elevation. (b) Slope. (c) average annual precipitation. (d) average annual temperature. (e) Relative humidity. (f) Evapotranspiration. (g) Sunshine duration. (h) GDP. (i) Population density. (j) Night light index.

In terms of topographic factors, differences in climate, hydrology, soil conditions, and human activities at different altitudes lead to regional variations in FVC. Figures 15a,b show a positive correlation between elevation, slope, and FVC in Anhui Province. Expansion of built-up land in low-altitude areas tends to reduce FVC, whereas afforestation and ecological restoration activities increase FVC in high-altitude areas. The regression coefficients for elevation and slope show minimal variation among the counties. The influence of elevation on FVC decreases gradually from northwest to southeast, with the least impact in the eastern regions (Figure 15a). The effect of slope on FVC decreases towards the northern and southern regions but is more pronounced in central Anhui (Figure 15b). Overall, the regression coefficients for topographic factors highlight a significant positive correlation with FVC across counties.

Climate plays a crucial role in affecting FVC changes, with favorable climatic conditions promoting vegetation growth and restoration, thereby significantly improving ecological quality. Figure 15c shows that the average annual precipitation in northern and southern Anhui has a greater impact on FVC than in central regions. A noticeable negative correlation is observed in the northern regions, while both positive and negative correlations are found in central regions, and a significant positive correlation is observed in the southern regions. The average annual temperature exhibits a complex relationship with FVC, which cannot be simply categorized as purely positive or negative. While a general decreasing trend is observed from west to east—particularly in Fuyang, Liuan, and Anqing—high temperatures may in some cases exacerbate soil erosion and flooding, thereby hindering vegetation recovery. Under appropriate hydrothermal conditions, elevated temperatures can also enhance plant photosynthesis and productivity. This dual effect requires careful consideration (Figure 15d). Similarly, relative humidity and evapotranspiration demonstrate predominantly positive but non-linear associations with FVC. Although regression coefficients show little variation across counties, the strength of the relationship between relative humidity and FVC increases from north to south, with the most significant effects observed in southern Anhui (Figure 15e). Dense vegetation, through its role in humidity regulation, contributes positively to regional climate stability and biodiversity. Nevertheless, excessively high humidity may also promote fungal diseases or inhibit transpiration, suggesting a context-dependent influence. Evapotranspiration generally supports vegetation growth, with its positive impact strengthening from east to west, most notably in Fuyang, Bozhou, Huaibei, and Suzhou (Figure 15f). Yet, under water-limited conditions, high evapotranspiration may intensify soil moisture deficit, potentially offsetting its benefits. Sunshine duration also shows a spatially varying relationship with FVC. While a significant negative correlation is observed overall—with a stronger adverse effect from west to east, particularly in Chuzhou, Ma’anshan, and Xuancheng (Figure 15g)—adequate solar radiation remains essential for photosynthesis. Excessive duration, however, may couple with high temperature and low moisture, lead to water stress and photoinhibition, thereby suppressing vegetation growth.

At the socio-economic level, the Gross Domestic Product (GDP) in Anhui Province, as one of the key factors, exhibits a negative influence trend on FVC, with this trend gradually weakening from north to south, specifically manifested in the regression coefficients fluctuating within a narrow range of −0.072 to −0.071 (as shown in Figure 15h). Notably, despite the relatively limited extent of this negative impact, regions with higher GDP typically coincide with more intense demands for various resources, such as energy, raw materials, and water. This heightened pattern of resource demand readily prompts overexploitation and consumption, thereby exerting pressure on natural vegetation and leading to a decline in its coverage. Conversely, population density positively affects FVC, increasing from north to south (Figure 15i), with a relatively small impact (0.029–0.030). The nighttime light index negatively impacts FVC, with the influence increasing from south to north. The most affected areas include Bozhou, Huaibei, Suzhou, and the northern regions of Fuyang and Bengbu (−0.545 to −0.448) (Figure 15j).

3.3.2 Analysis of influencing factors of FVC in different regions

Compared to the MGWR model, the geographical detector can rank the importance of factors using the q value. This model incorporates geographic spatial factors, avoids the impact of multicollinearity, and considers the effects of dual factors on the dependent variable. Building on the quantitative analysis of the spatiotemporal characteristics and spatial heterogeneity of vegetation evolution in Anhui Province, this study further explores the factors influencing FVC changes in Anhui and its various subregions. Elevation, slope, average annual precipitation, average annual temperature, relative humidity, evapotranspiration, sunshine duration, GDP, population density, and nighttime light index are denoted as X1 to X10, respectively, with FVC denoted as Y. Using the geographical detector’s factor detection module, the driving factors of FVC in Anhui Province were analyzed, revealing the extent of each factor’s impact on FVC changes.

As shown in Figure 16, FVC in Anhui Province is influenced by various factors, with topographic and climatic factors having a more significant impact on vegetation than socio-economic factors. The top five factors explaining FVC across the province are X5, X3, X2, X1, and X10.

Figure 16

Figure 16. q values of impact factors in Anhui Province and each region.

In the Core Agricultural Production Zone, climatic factors are significantly more influential than other factors, with the top five being X5, X3, X6, X2, and X1. In the Ecological Protection Buffer Zone, climatic factors are also notably more influential, with the top five being X7, X5, X1, X3, and X2. In the Core Ecological Protection Zone, topographic factors are more prominent, with the top five being X1, X2, X7, X3, and X4. In the Core Economic Development Zone, socio-economic factors are more significant, with the top five being X8, X9, X10, X1, and X2.

Using the interaction detection module, the interactive effects of pairs of 10 factors on FVC in Anhui Province and its various subregions were analyzed, as shown in Table 6. The interaction between factors significantly enhanced the influence of individual factors on vegetation, demonstrating a nonlinear amplification effect. No single factor could independently influence vegetation, highlighting the critical importance of the complex interactions among factors in the process of vegetation growth and change.

Table 6

Table 6. The ranking of the top 5 influencing factors and interaction relationships between Anhui Province and various regions.

In Anhui Province and its land use policy zones, vegetation distribution and changes are jointly influenced by topographic, climatic, and socio-economic factors. Climatic factors dominate in Anhui Province, the Core Agricultural Production Zone, and the Ecological Protection buffer Zone. Topographic factors are predominant in the Core Ecological Protection Zone, while socio-economic factors are primarily significant in the Core Economic Development Zone.

4 Discussion

4.1 Comparison of GWR and MGWR models

To delve into the intricacies and spatial variability of factors influencing FVC in Anhui Province, this study embraces the MGWR model, aiming to conduct a nuanced analysis of the spatial heterogeneity of FVC drivers. Both the MGWR model and its predecessor, the GWR model, acknowledge the importance of spatial non-stationarity and multi-scale characteristics (Yu et al., 2019; Zhu et al., 2020), yet the MGWR model marks a significant methodological advancement. Specifically, the MGWR model transcends the limitations of the GWR model by assigning unique spatial scales and variable bandwidths to individual influencing factors, enabling a more intricate regression analysis framework for FVC in Anhui Province (Fotheringham et al., 2015). This refinement not only maintains the capacity to capture spatial heterogeneity but also uncovers the differential intensities and patterns of influence exerted by distinct drivers on FVC across geographical regions.

In contrast to the GWR model’s generalized average impact assessment for each variable, the MGWR model portrays a richer and more dynamic picture of how different factors uniquely interact with FVC at specific spatial locations, with varying intensities and modalities (Dai et al., 2019; Ariken et al., 2021; Xiong et al., 2021). This revelation underpins the formulation of region-specific and nuanced ecological conservation strategies, offering a solid theoretical foundation and empirical evidence for the sustainable management and optimal utilization of ecological resources. A comparative analysis of the bandwidth differences under all variables between the GWR and MGWR models (Figure 17) reveals that the scale coefficients of the MGWR model help distinguish specific factors, whereas the GWR model applies a uniform average scale of influence for all factors on vegetation (Guo et al., 2017).

Figure 17

Figure 17. Bandwidth of geographical weighted regression and multi-scale geographical weighted regression models in Anhui Province.

As shown in Figure 17, the classic GWR model has a bandwidth of 64, meaning it utilizes data from approximately 64 counties to estimate the coefficients for each local regression, which covers 61.5% of the total 104 counties (the sample units in this study). However, the MGWR model demonstrates significant differences in the effect scales of various vegetation influencing factors across Anhui Province. Specifically, the effect scales of average annual precipitation and nighttime light index are relatively low (with a bandwidth covering only ∼43 counties), explaining their spatial variation based on approximately 41.3% of the counties. This indicates significant spatial heterogeneity in their influence. In contrast, factors such as elevation, slope, relative humidity, evapotranspiration, sunshine duration, GDP, and population density have a much larger effect scale, with a bandwidth of 103. This implies that the processes of these factors are spatially stable, as their coefficients are calibrated using data from nearly all (99%) of the 104 counties. The effect scale of average annual temperature is larger than that of precipitation and the nighttime light index, with a bandwidth of 51, meaning its coefficients are estimated using data from about 49% of the county samples.

The MGWR model is superior to the GWR model in accurately capturing the different scales of influence of various variables, providing crucial insights for developing localized vegetation restoration strategies.

4.2 The theme and significance of the article

This study aims to analyze the long-term evolutionary trends of FVC across different land-use policy zones and to elucidate the underlying driving mechanisms. The research holds significant implications for advancing vegetation conservation, optimizing agricultural land-use planning, and fostering coordinated regional economic development. Anhui Province exhibits pronounced transitional and complex geographical characteristics, encompassing diverse landforms such as plains, hills, and mountains, rendering it a microcosm of China’s topographic diversity. This unique and representative geographical setting provides an ideal natural laboratory for the present investigation. Owing to its substantial internal environmental heterogeneity, the findings and methodologies derived from this study on land-use policy zoning, vegetation dynamics, and driving mechanisms possess considerable demonstrative value and broad applicability, offering robust theoretical and technical references for similar research in other complex geographical regions worldwide.

Conventional land-use zoning studies often emphasize static assessments of land patches, while overlooking the dynamic transitions between different land-use types. In traditional approaches, establishing suitability evaluation indicator systems, determining weights, and assigning suitability levels constitute the core and most commonly adopted procedures for delineating functional zones (Zong et al., 2023; Zhang et al., 2024). This study addresses the aforementioned gap by explicitly incorporating the dynamic transition relationships between land patches, thereby overcoming the limitations of previous static evaluations that neglect inter-categorical flow characteristics. By leveraging the characteristics of land-use transitions between patches, we innovatively introduce the SOFM—a machine learning technique—to perform feature learning and pattern recognition across the region. This is further integrated with hierarchical clustering analysis to develop a comprehensive land-use policy zoning methodology that synthesizes dynamic transition information with spatial clustering techniques. The proposed framework enables a scientifically grounded and dynamic partitioning of land-use patterns within the topographically complex context of Anhui Province.

Furthermore, based on the established policy zones, this study systematically examines the long-term evolutionary trends of FVC within each zone by employing theoretical constructs such as the SEN + MK, CV, Hurst exponent, and Moran’s index. Methodologically, we integrate R language, MATLAB, coupled MGWR, and geographical detector models to identify areas at risk of future vegetation degradation and to conduct an in-depth investigation of the driving mechanisms. Beyond its practical relevance for promoting regional ecological conservation, optimizing agricultural spatial arrangements, and harmonizing regional economic development, this research highlights the interconnectedness of ecological processes such as land-use transitions and offers effective solutions for vegetation preservation. In comparison with existing studies, the proposed methodological framework allows for more accurate and rational identification of multi-objective functional zones, thereby facilitating more effective vegetation conservation strategies.

4.3 Implications of vegetation management in different land use policy areas

Based on land use transition trends, this study used the SOFM model to delineate four land use policy zones in Anhui Province: the Core Agricultural Production Zone, Ecological Protection buffer Zone, Core Ecological Protection Zone, and Core Economic Development Zone. These zones were used to explore the spatiotemporal distribution characteristics of past, present, and future FVC in Anhui Province. The results are fairly reliable and were compared with previous studies. We found that from 2000 to 2023, the average spatial distribution pattern of FVC in Anhui Province exhibited a trend of high values in the north and south, with lower values in the central region. Similar results were obtained in previous studies (Liu, 2016). Elevation, slope, and other topographic factors significantly influence FVC dynamics in Anhui Province, consistent with our findings.

This study reveals that over the past 24 years, the FVC in core agricultural production zones has exhibited an overall increasing trend, albeit with periodic fluctuations. Notably, the Hurst exponent further indicates a potential future shift in vegetation dynamics from improvement to degradation in these areas (Figures 12, 13). This trend aligns with the ‘vegetation growth saturation’ phenomenon observed in several globally distributed intensive agricultural regions (Liu et al., 2022). Consequently, we recommend the systematic promotion of agricultural technological innovation in these zones, focusing on water-saving irrigation, precision fertilization, and smart farming systems. This approach not only aims to enhance resource-use efficiency in grain production but also serves as a proactive strategy to mitigate potential vegetation degradation risks. Simultaneously, it is imperative to uphold stringent farmland protection policies, ensuring regional food security while exploring sustainable development pathways that synergize economic growth with the enhancement of ecosystem services. Conversely, vegetation conditions in core ecological protection areas also warrant attention, as these areas exhibit a prospective trend of vegetation degradation (Figures 12, 13). Given their relatively higher elevations, implementing hydrologically-oriented forest structure optimization-by adjusting tree species composition, regulating planting density, and optimizing spatial configuration-is recommended to bolster ecosystem stability. As multifunctional transitional belts, ecological protection buffer zones require spatial planning that holistically balances ecological conservation, controlled urban expansion, and agricultural activities. In key vegetation protection areas, particularly in cities such as Lu’an, Huangshan, and Xuancheng, promoting efficient water-saving technologies like drip and sprinkler irrigation could significantly improve vegetation restoration outcomes and water-use efficiency. The effectiveness of such measures has been validated in ecological restoration projects conducted in arid and semi-arid regions. Among the four land-use policy zones, the core economic development area exhibits the lowest average FVC, reflecting the persistent pressure exerted by rapid urbanization on vegetation cover. To alleviate the degradation of urban ecological functions, we propose the systematic integration of green infrastructure concepts into urban development. This can be achieved by optimizing green space layout, increasing the proportion of native vegetation, and enhancing the connectivity of ecological corridors-measures that collectively strengthen the regulatory functions and climatic resilience of urban green space systems, as supported by multiple urban ecological studies (Chen and Wang, 2022).

5 Conclusion

Based on a thorough analysis of the transformation characteristics of land use types in Anhui Province, this study innovatively integrates land use transition theory with SOFM technology to precisely delineate four land use policy zones: Core Agricultural Production Zone, Ecological Protection Buffer Zone, Core Ecological Conservation Zone, and Core Economic Development Zone. Employing the SEN + MK trend analysis framework and the CV as a quantitative tool, the study systematically dissects the dynamic evolution trends and fluctuation characteristics of FVC within different land use policy zones in Anhui Province from 2000 to 2023. Furthermore, by seamlessly integrating the SEN + MK methodology with the Hurst exponent prediction model, the study achieves a scientific forecast of future FVC trends in these zones. Additionally, this research introduces the MGWR model, which profoundly reveals the intricate spatial heterogeneity relationships among various influencing factors during the FVC changes in Anhui Province. Simultaneously, utilizing the advanced analytical tool of Geodetector, the study delves into the comprehensive influence mechanisms of topographical, climatic, and socio-economic factors on FVC variations within different land use policy zones, ultimately yielding a series of academically valuable research conclusions:

1. Clustering Characteristics of Land Use Policy Zones. The Core Agricultural Production Zone is mainly concentrated in the northern and central parts of Anhui Province, encompassing 15 counties. The Ecological Protection Buffer Zone is widely distributed across the central and southern regions, with scattered presence in the northern area, encompassing a total of 61 counties. The Core Ecological Protection Zone, on the other hand, is prominently centered on the Dabie Mountains and southern Anhui, encompassing 15 county-level administrative units within its boundaries. The Core Economic Development Zone is primarily distributed in the central regions of Hefei, Wuhu, and Chuzhou, comprising 13 counties.

2. The distribution of FVC in Anhui Province exhibits notable geographical characteristics, with the southern and northern regions displaying higher average coverage; whereas the central region registers relatively lower levels, creating a distinct contrast in the geographical pattern. The vegetation condition in Anhui Province is relatively good, with FVC primarily between 0.4–0.6, accounting for about 39.75% of the area. Over the past 24 years, the region with the highest average FVC has been the Core Ecological Protection Zone, covering about 60.95% of the area. From 2000 to 2023, FVC in Anhui Province has been mainly stable or improving, with the most improvement observed in the Core Agricultural Production Zone and the most degradation in the Core Ecological Protection Zone. The future trend of FVC changes, excluding indetermination, also indicates a significant proportion of areas shifting from improvement to degradation, accounting for 27.19%, primarily in the Core Agricultural Production Zone.

3. Spatial Heterogeneity of FVC Drivers: The driving factors of FVC in Anhui Province exhibit significant spatial heterogeneity at different scales. Topographic factors such as elevation and slope play a dominant role in the Core Ecological Protection Zone. Climatic factors like relative humidity and average annual precipitation are crucial in Anhui Province, the Core Agricultural Production Zone, and the Ecological Protection Transition Zone. Socio-economic factors like GDP and population density are the main influencing factors in the Core Economic Development Zone.

4. There is a significant clustering effect among high FVC areas in Anhui Province. The “high-high” FVC clusters are prominently concentrated in the northwestern and southern regions, particularly at the core of the Core Agricultural Production Belt and the Core Ecological Protection Zone. In contrast, the “high-low” FVC clusters are specifically distributed in the central part of Anhui Province, predominantly within the scope of the Ecological Protection Buffer Zone. The “low-low” FVC clusters are scattered, with a higher concentration in the Core Economic Development Zone among the land use policy zones.

These research findings provide valuable references for vegetation protection and ecosystem restoration in other regions. Future research should further explore the specific mechanisms and strategies for vegetation changes at different scales to achieve more effective ecological management and sustainable development.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

XZ: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. BD: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing.

Funding

The authors declare that financial support was received for the research and/or publication of this article. Funding for this research was provided by National Natural Science Foundation of China (32071600; 41571101), provided by the Natural Resources Science and Technology Project of Anhui Province (2022-k-1). Remote sensing science and technology cross peak cultivation discipline (23103107). We appreciate the above projects for their assistance in this study.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Ariken, M., Zhang, F., Chan, N., and Kung, H. t. (2021). Coupling coordination analysis and spatio temporal heterogencity between urbanization and cco-environment along the silk road fconomic belt in China. Ecol. Ind. 121, 107014. doi:10.1016/j.ecolind.2020.107014

CrossRef Full Text | Google Scholar

Bai, Y. (2021). Analysis ofvegetation dynamics in the qinling-daba mountains region from modistime series data. Eeol. Ind. 129, 108029. doi:10.1016/j.ecolind.2021.108029

CrossRef Full Text | Google Scholar

Chen, J., Dong, B., Li, H. R., Zhang, S., Peng, L., Fang, L., et al. (2020). Study on landscape ecological risk assessment of hooded crane breeding and overwintering habitat. Environ. Res. 187, 190649. doi:10.1016/j.envres.2020.109649

PubMed Abstract | CrossRef Full Text | Google Scholar

Chen, Y., Ma, Y., Pan, J., Zhang, S., Zhang, X., Wu, R., et al. (2023). Integrating ecosystem health diagnosis into the construction of ecological security network-a case study in Qujing City, China. Ecol. Ind. 146, 109780. doi:10.1016/j.ecolind.2022.109780

CrossRef Full Text | Google Scholar

Chu, H., Venevsky, S., Wu, C., and Wang, M. (2019). NDVI-based vegetation dynamics and its response to climate changes at Amur-Heilongjiang River Basin from 1982 to 2015. Sci. Total Environ. 650, 2051–2062. doi:10.1016/j.scitotenv.2018.09.115

PubMed Abstract | CrossRef Full Text | Google Scholar

Dai, L., Li, S., Lewis, B., Wu, J., Yu, D., Zhou, W., et al. (2019). The influence of land use change on the spatial temporal variability of habitat quality between 1990 and 2010 in northeast China. J. For. Res. 30, 2227–2236. doi:10.1007/s11676-018-0771-x

CrossRef Full Text | Google Scholar

Fotheringham, A. S., Crespo, R., and Yao, J. (2015). Geographical and temporal weighted regression (GTWR). Eogr. Anal. 47, 431–452. doi:10.1111/gean.12071

CrossRef Full Text | Google Scholar

Fotheringham, A. S., Yang, W., and Kang, W. (2017). Multiscale gicographically weighted Regression(MGWR). Ann. Am Assoc Gcogr 107 (6), 1247–1265. doi:10.1080/24694452.2017.1352480

CrossRef Full Text | Google Scholar

Gu, Z., Duan, X., Shi, Y., Li, Y., and Pan, X. (2018). Spatiotemporal variation in vegetation cover-age and its response to climate factors in the Red River Basin, China. Ecol. Ind. 93, 54–64. doi:10.1016/j.ecolind.2018.04.033

CrossRef Full Text | Google Scholar

Guo, Y., Tang, Q., Gong, D. Y., and Zhang, Z. (2017). Estimating ground-level PM2.5 concentrations in Beijing using a satellite-based geographically and temporally weighted regression model. Remote Sens.Environ. 198, 140–149. doi:10.1016/j.rse.2017.06.001

CrossRef Full Text | Google Scholar

He, J., Shi, X., and Fu, Y. (2021). Identifying vegetation restoration effectiveness and driving factors on different micro-topographic types of hilly Loess Plateau: from the perspective of ecological resilience. J. Environ. Manage. 289, 112562. doi:10.1016/j.jenvman.2021.112562

PubMed Abstract | CrossRef Full Text | Google Scholar

Huang, D. W., and Gui, H. R. (2018). Distribution features and internal relations of heavy metals in soil–maize system of mining area, Anhui Province, Eastern China. Hum. Ecol. Risk Assess. 25 (4), 863–881. doi:10.1080/10807039.2018.1453298

CrossRef Full Text | Google Scholar

Jiang, H., and He, G. (2023). Analysis of spatial and temporal evolution of regional water resources carrying capacity and influencing factors Anhui province as an example. Sustain 15, 11255. doi:10.3390/su151411255

CrossRef Full Text | Google Scholar

Jiang, M., He, Y., Song, C., Pan, Y., Qiu, T., and Tian, S. (2021). Disaggregating climatic and anthropogenic influences on vegetation changes in Beijing-Tianjin-Hebei region of China. Sci. Total Environ. 786, 147574. doi:10.1016/j.scitotenv.2021.147574

CrossRef Full Text | Google Scholar

Li, G., Sun, S. B., Han, J. C., Yan, J., Liu, W., Wei, Y., et al. (2019). Impacts of Chinese grain for green program and climate change on vegetation in the loess plateau during 1982-2015. Sci. Total Environ. 660, 177–187. doi:10.1016/j.scitotenv.2019.01.028

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, H., Shi, C., Zhang, Y., Ning, T., Sun, P., Liu, X., et al. (2020). Using the budyko hypothesis for detecting and attributing changes in runolf to climate and vegetation change in the soft sandstone area of the middle yellow river basin. China. Sci. Total Environ. 703, 135588. doi:10.1016/j.scitotenv.2019.135588

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, X. T. (2016). Study on spatiotemporal change of vegetation cover and its driving factors. China University of Mining and Technology.

Google Scholar

Liu, Y., Li, Y., and Song, X. (2022). Global evidence of vegetation growth saturation in intensive agricultural landscapes. Nat. Geosci. 15 (4), 282–288.

Google Scholar

Mao, P., Zhang, J., Li, M., Liu, Y., Wang, X., Yan, R., et al. (2022). Spatial and temporal variations in fractional vegetation cover and its driving factors in the hulun lake region. Ecol. Ind. 135, 108490. doi:10.1016/j.ecolind.2021.108490

CrossRef Full Text | Google Scholar

Nie, T., Dong, G., Jiang, X., and Lei, Y. (2021). Spatio-temporal changes and driving forces of vegetation coverage on the Loess Plateau of Northern Shaanxi. Remote Sens. 13 (4), 613. doi:10.3390/rs13040613

CrossRef Full Text | Google Scholar

Niu, L., Zhang, Z., Peng, Z., Liang, Y., Liu, M., Jiang, Y., et al. (2021). Identifying surface urban heat Island drivers and their spatial heterogeneity in China’s 281 cities: an empirical study based on multiscale geographically weighted regression. Remote Sens. 13 (21), 4428. doi:10.3390/rs13214428

CrossRef Full Text | Google Scholar

Omer, A., Zhuguo, M., Zheng, Z., and Saleem, F. (2020). Natural and anthropogenic influences on the recent droughts in yellow river basin, China. Sci. Total Environ. 704, 135428. doi:10.1016/j.scitotenv.2019.135428

PubMed Abstract | CrossRef Full Text | Google Scholar

Ouyang, W., Hao, F., Skidmore, A. K., and Toxopeus, A. (2010). Soil erosion and sediment yield and their relationships with vegetation cover in upper stream of the yellow river. Sci. Total Environ. 409, 396–403. doi:10.1016/j.scitotenv.2010.10.020

PubMed Abstract | CrossRef Full Text | Google Scholar

Pandey, H. P., Aryal, K., Aryal, S., and Maraseni, T. N. (2023). Understanding local ecosystem dynamics in three provinces of the lowlands of Nepal. Sci. Total Environ. 867, 161501. doi:10.1016/j.scitotenv.2023.161501

PubMed Abstract | CrossRef Full Text | Google Scholar

Pei, H., Liu, M., Jia, Y., Zhang, H., Li, Y., and Xiao, Y. (2021). The trend of vegetation greening and its drivers in the Agro-pastoral ecotone of northern China, 2000–2020. Ecol. Ind. 129, 108004. doi:10.1016/j.ecolind.2021.108004

CrossRef Full Text | Google Scholar

Peng, W., Kuang, T., and Tao, S. (2019). Quantifying influences of natural factors on vegetation NDVI changes based on geographical detector in Sichuan, western China. J. Clean. Prod. 233, 353–367. doi:10.1016/j.jclepro.2019.05.355

CrossRef Full Text | Google Scholar

Qu, S., Wang, L. C., Lin, A. W., Yu, D., and Yuan, M. (2020). Distinguishing the impacts of climate change and anthropogenic factors on vegetation dynamics in the Yangtze River Basin, China. Ecol. Ind. 108, 105724. doi:10.1016/j.ecolind.2019.105724

CrossRef Full Text | Google Scholar

Qu, J., Xu, Z., Dong, B., Wang, H., and Han, Y. (2025). Analyzing trade-offs, synergies, and driving factors of ecosystem services in Anhui Province using spatial analysis and XG-boost modeling. Ecol. Ind. 171, 113098. doi:10.1016/j.ecolind.2025.113098

CrossRef Full Text | Google Scholar

Shen, R. N., Zhang, Y. J., Zhu, J. T., Chen, N., Chen, Y., Zhao, G., et al. (2022). The interactive effects of nitrogen addition and increased precipitation on gross ecosystem productivity in an alpine meadow. J. Plant Ecol. 15 (1), 168–179. doi:10.1093/jpe/rtab081

CrossRef Full Text | Google Scholar

Shi, P., Zhang, Y., Ren, Z., Yu, Y., Li, P., and Gong, J. (2019). Land-use changes and check dams reducing runoff and sediment yield on the Loess plateau of China. Sci. Total Environ. 664, 984–994. doi:10.1016/j.scitotenv.2019.01.430

PubMed Abstract | CrossRef Full Text | Google Scholar

Shi, P., Li, P., Li, Z., Sun, J., Wang, D., and Min, Z. (2022). Effects of grass vegetation coverage and position on runoff and sediment yields on the slope of Loess plateau, China. Agric. Water Manag. 259, 107231. doi:10.1016/j.agwat.2021.107231

CrossRef Full Text | Google Scholar

Teng, Y., Ke, Y. Y., Zhou, X., Tao, R., and Wang, Y. (2022). Derived regional soil-environmental quality criteria of metals based on Anhui soil-crop systems at the regulated level. Sci. Total Environ. 825, 154060. doi:10.1016/j.scitotenv.2022.154060

PubMed Abstract | CrossRef Full Text | Google Scholar

Vicente-Serrano, S. M., Camarero, J. J., Olano, J. M., Martín-Hernández, N., Peña-Gallardo, M., Tomás-Burguera, M., et al. (2016). Diverse relationships between forest growth and the normalized difference vegetation index at a global scale. Remote Sens. Environ. 187, 14–29. doi:10.1016/j.rse.2016.10.001

CrossRef Full Text | Google Scholar

Wang, X., Shi, S., Zhao, X., Hu, Z., Hou, M., and Xu, L. (2022). Detecting spatially non-stationary between vegetation and related factors in the Yellow River Basin from 1986 to 2021 using multiscale geographically weighted regression based on Landsat. Remote Sens. 14 (24), 6276. doi:10.3390/rs14246276

CrossRef Full Text | Google Scholar

Wang, C., Li, Z., Chen, H., and Wang, M. (2023). Comprehensive evaluation of agricultural water resources’ carrying capacity in Anhui Province based on an improved TOPSIS model. Sustain 15, 13297. doi:10.3390/su151813297

CrossRef Full Text | Google Scholar

Wei, J., Lei, Y., Yao, H., Ge, J., Wu, S., and Liu, L. (2021). Estimation and influencing factors of agricultural water efficiency in the yellow river basin, China. J. Clean. Prod. 308, 127249. doi:10.1016/j.jclepro.2021.127249

CrossRef Full Text | Google Scholar

Xiong, Y., Li, Y., Xiong, S., Wu, G., and Deng, O. (2021). Multi-scale spatial correlation between vegetation index and terrain attributes in a small watershed of the upper Minjiang River. Ecol. Ind. 126, 107610. doi:10.1016/j.ecolind.2021.107610

CrossRef Full Text | Google Scholar

Xu, G. C., Zhang, J. X., Li, P., Li, Z., Lu, K., Wang, X., et al. (2018). Vegetation restoration projects and their influence on runoff and sediment in China. Ecol. Ind. 95, 233–241. doi:10.1016/j.ecolind.2018.07.047

CrossRef Full Text | Google Scholar

Yu, T., Bao, A., Xu, W., Guo, H., Jiang, L., Zheng, G., et al. (2019). Exploring variability in landscape ecological risk and quantifying its driving factors in the Amu Darya Delta. ERPH 17, 79. doi:10.3390/ijerph17010079

PubMed Abstract | CrossRef Full Text | Google Scholar

Yu, X., Xie, J., Jiang, R., Zhao, Y., Li, F., Liang, J., et al. (2021). Spatio-temporal variation and predictability of vegetation coverage in the beijing tianjin hebei metropolitan region, China. Theor. Appl. Climatol. 145, 47–62. doi:10.1007/s00704-021-03616-x

CrossRef Full Text | Google Scholar

Zhang, X., and Jin, X. (2021). Vegetation dynamics and responses to climate change and anthropogenic activities in the three-river headwaters region, China. Ecol. Ind. 131, 108223. doi:10.1016/j.ecolind.2021.108223

CrossRef Full Text | Google Scholar

Zhang, Y. P., Lin, J. P., Huang, Y. M., Chen, Z., Zhu, C., and Yuan, H. (2024). Delineation of urban growth boundary based on FLUS model under the perspective of land use evaluation in hilly mountainous areas. J. Mt. Sci-engl. 21 (5), 1647–1662. doi:10.1007/s11629-023-8499-2

CrossRef Full Text | Google Scholar

Zheng, K., Wei, J. Z., Pei, J. Y., Cheng, H., Zhang, X. L., Huang, F. Q., et al. (2019). Impacts of climate change and human activities on grassland vegetation variation in the Chinese Loess Plateau. Sci. Total Environ. 660, 236–244. doi:10.1016/j.scitotenv.2019.01.022

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhu, C., Zhang, X., Zhou, M., He, S., Gan, M., Yang, L., et al. (2020). Impacts of urbanization and landscape pattern on habitat quality using OLS and GWR models in Hangzhou, China. Ecol. Ind. 117, 106654. doi:10.1016/j.ecolind.2020.106654

CrossRef Full Text | Google Scholar

Zong, S., Hu, Y., and Bai, Y. (2023). Spatio-temporal pattern and driving mechanisms of land use conflicts changes (2010–2018) in the Bohai Rim transition zone. Land Degrad. Dev. 34 (12), 3451–3466. doi:10.1002/ldr.4694

CrossRef Full Text | Google Scholar