Evolution of Vegetation Coverage in the Jinan Section of the Basin of the Yellow River (China), 2008–2022: Spatial Dynamics and Drivers

Abstract

The Yellow River Basin serves as a critical ecological barrier in China. However, it has increasingly faced severe ecological and environmental challenges, with soil erosion and overgrazing being particularly prominent issues. As an important region in the middle and lower reaches of the Yellow River, the Jinan section of the Yellow River Basin is similarly affected by these problems, posing significant threats to the stability and sustainability of its ecosystems. To scientifically identify areas severely impacted by soil erosion and systematically quantify the effects of climate change on vegetation coverage within the Yellow River Basin, this study focuses on the Jinan section. By analyzing the spatio-temporal evolution patterns of the Normalized Difference Vegetation Index (NDVI), this research aims to explore the driving mechanisms behind these changes and further predict the future spatial distribution of NDVI, providing theoretical support and practical guidance for regional ecological conservation and sustainable development. This study employed the slope trend analysis method to examine the spatio-temporal variation characteristics of NDVI in the Jinan section of the Yellow River Basin from 2008 to 2022 and utilized the FLUS model to predict the spatial distribution of NDVI in 2025. The Optimal Parameters-based Geographical Detector (OPGD) model was applied to systematically analyze the impacts of four key driving factors—precipitation (PRE), temperature (TEM), population density (POP), and gross domestic product (GDP) on vegetation coverage. Finally, correlation and lag effect analyses were conducted to investigate the relationships between NDVI and TEM as well as NDVI and PRE. The research results indicate the following: (1) from 2008 to 2022, the NDVI values during the growing season in the Jinan section of the Yellow River Basin exhibited a significant increasing trend. This growth suggests a continuous improvement in regional vegetation coverage, likely influenced by the combined effects of natural and anthropogenic factors. (2) The FLUS model predicts that, by 2025, the proportion of high-density NDVI areas will rise to 55.35%, reflecting the potential for further optimization of vegetation coverage under appropriate management. (3) POP had a particularly significant impact on vegetation coverage, and its interaction with TEM, PRE, and GDP generated an amplified combined effect, indicating the dominant role of the synergy between socioeconomic and climatic factors in regional vegetation dynamics. (4) NDVI exhibited a significant positive correlation with both temperature and precipitation, further demonstrating that climatic conditions were key drivers of vegetation coverage changes. (5) In urban areas, NDVI showed a certain time lag in response to changes in precipitation and temperature, whereas this lag effect was not significant in suburban and mountainous areas, highlighting the regulatory role of human activities and land use patterns on vegetation dynamics in different regions. These findings not only reveal the driving mechanisms and influencing factors behind vegetation coverage changes but also provide critical data support for ecological protection and economic development planning in the Yellow River Basin, contributing to the coordinated advancement of ecological environment construction and economic growth.

1. Introduction

Currently, under the combined influence of intensified human activities and climate change, the global environment is facing severe challenges such as accelerating land desertification and extensive deforestation. As a critical ecological barrier and a vital water resource security zone in China, the Yellow River Basin plays a core role in national and regional sustainable development. However, in recent years, the Yellow River Basin has faced a series of ecological and environmental challenges, including wetland degradation, land desertification, soil erosion, and fragmentation of ecological corridors. These issues not only threaten the stability of the basin’s ecosystems but also impose significant pressures on regional economic development and national food and water security [1]. To address these challenges, the Chinese government has identified ecological security protection and high-quality development of the Yellow River Basin as a strategic priority critical to the great rejuvenation of the Chinese nation. It has also proposed the ecological protection framework of “one belt, five zones, and multiple points” to curb ecological degradation and accelerate the restoration of key ecosystems [2]. The Jinan section, located in the lower reaches of the Yellow River, is closely connected to major economic hubs and represents an integral part of the Yellow River Basin with significant ecological importance [3]. However, due to the dual impacts of human activities and climate change, this region has long faced severe ecological problems such as soil erosion and land desertification [4,5,6,7,8,9,10,11,12]. Given the increasingly severe vegetation degradation [13,14], systematically exploring the trends in vegetation coverage, the spatial patterns of regional vegetation cover, and the underlying driving mechanisms are essential for mitigating these ecological problems and providing scientific evidence for ecological restoration and sustainable development of the basin. The period from 2008 to 2022 encompasses a critical phase of rapid economic development and ecological environmental changes in China, reflecting the comprehensive impacts of human activities and climate change [15]. Additionally, NDVI data from this period are relatively abundant, providing a solid foundation for systematic analysis. Studying vegetation dynamics during this timeframe can reveal trends in ecological changes and provide scientific evidence for regional ecological protection and sustainable management. Therefore, based on NDVI data from 2008 to 2022, this study comprehensively analyzes the vegetation dynamics and driving factors in the Jinan section of the Yellow River Basin. This research will contribute to advancing the ecological protection and high-quality development strategies for the Yellow River Basin and provides valuable scientific references for studying vegetation dynamics in similar basins.

NDVI provides continuous spatio-temporal data, enabling the quantitative monitoring of dynamic changes in vegetation coverage on the Earth’s surface and serving as an optimal indicator for characterizing vegetation growth conditions [16,17,18,19,20,21,22,23,24]. In recent years, researchers worldwide have extensively utilized NDVI data to analyze the spatio-temporal evolution patterns of vegetation coverage and its response to climatic factors. In China, Wang et al. [25] investigated the spatiotemporal characteristics of NDVI in the Xin’an River Basin, while Lu et al. [26] analyzed the spatiotemporal changes in green land density in Gansu Province from 2000 to 2020. Internationally, scholars have examined the spatiotemporal variation in NDVI in regions such as the Baltic Sea, the wetlands of the Andes Mountains in Argentina, and the Great Plains of the United States [27,28]. Additionally, to explore the future trends of NDVI, some researchers have conducted predictive studies. For instance, Wu et al. utilized the Hurst exponent and coefficient of variation to investigate the spatio-temporal fluctuations of green density and its potential future pathways [29]. However, traditional NDVI forecasting methods, such as the Hurst index analysis, have certain limitations in capturing the complexity of NDVI dynamics, and cannot fully account for its variations. Recently, the FLUS model, which integrates machine learning with multi-source data, has demonstrated great potential in improving the accuracy of vegetation coverage predictions [30]. These studies contribute to understanding the trends in vegetation changes from a spatio-temporal perspective, thereby providing scientific support for vegetation conservation [31,32,33,34]. Regarding the study of factors influencing NDVI, domestic and international scholars have focused on the intrinsic relationships and response mechanisms between NDVI changes and precipitation (PRE) and temperature (TEM) [35,36,37]. For example, Li et al. used variable separation methods to analyze the internal relationships and response mechanisms between NDVI variations and PRE and TEM [38]. Wang’s investigation of the central Great Plains revealed that annual fluctuations in NDVI are primarily driven by the variance in PRE, and the overall geographical distribution of NDVI is closely related to the spatial patterns of mean annual PRE [39]. Past researchers have primarily utilized techniques like partial analysis of correlation, regression modeling, and residual analysis in multiple regression to investigate the factors influencing changes in vegetation cover. These approaches operate under the presumption of a linear association between vegetation cover and environmental factors. The Geodetector framework, pioneered by Wang et al., adeptly tackles the challenge of multicol-linearity stemming from interactions among predictor variables [40]. However, the Geodetector usually relies on preset parameters, and the selection of these parameters is often based on experience or trial and error, lacking a systematic approach and scientific rigor. To tackle the challenge of the Geodetector framework not being optimized for identifying levels of driving factors, Song et al. introduced a Geographical Detector model based on Optimal Parameters (OPGD) [41]. This model incorporates a parameter optimization module that can select the optimal discretization method and the optimal number of breakpoints, thereby enhancing the significant differences among various categories and providing a more accurate study for driver analysis.

Based on these findings, several shortcomings in current NDVI research were identified. (1) Most studies focus on provincial or larger-scale regions, with limited detailed research at the urban scale (such as the Jinan section). (2) The classification of driving factors often relies on empirical methods, lacking optimization strategies. (3) There is limited research on the nonlinear response of NDVI to driving factors and lag effects. (4) NDVI predictions have not fully accounted for the dynamic interactions between climate change and human activities.

Therefore, this study focuses on the Jinan section of the Yellow River Basin, investigating the spatiotemporal dynamics and driving mechanisms of vegetation cover in this region. Specifically, the aims of this study are as follows. (1) Apply slope trend analysis to examine the spatiotemporal variation characteristics of NDVI. (2) Combine climate factors and human activities to predict the spatial distribution of NDVI in 2025 using the FLUS model, improving prediction accuracy. (3) Quantitatively assess the contribution of driving factors to vegetation coverage using the Optimal Parameter-based Geodetector model. (4) Explore the response relationship between NDVI, temperature (TEM), and precipitation (PRE) through correlation analysis and lag effect testing. Based on this, the following hypotheses are proposed:

(1) vegetation cover in the Jinan section is jointly driven by climate factors (TEM, PRE) and human activities, with significant differences in trends and spatial distribution; (2) the high-density spatial distribution of NDVI will significantly increase in 2025; (3) TEM and PRE have significant lag effects on NDVI, and the lag relationship can reveal the long-term impact of climate change on vegetation cover. This study can provide a scientific basis for regional ecological restoration and vegetation conservation, offering practical insights for addressing the impacts of climate change and human activities.

2. Materials and Methods

2.1. Study Area

The Jinan segment of the Yellow River Basin is located between 36°21′4″ N and 36°57′4″ N latitude and 116°49′12″ E to 117°25′12″ E longitude, covering a total length of 183 km. This segment starts from Dong’e Town in Pingyin County and ends at Renfeng Town in Jiyang District, passing through nine districts and counties, including Pingyin County, Changqing District, Huaiyin District, Tianqiao District, Licheng District, Jinan High-tech Zone, Zhangqiu District, Jiyang District, and Gangcheng District (Figure 1). The Jinan section of the Yellow River Basin falls within the warm temperate continental monsoon climatic zone, characterized by distinct seasonal variations, abundant sunlight, and concentrated precipitation during summer. It exhibits typical ecological features of northern plain river systems. The region supports various types of vegetation, including farmland, plantations, and natural vegetation, which play a crucial role in maintaining ecosystem services. However, prolonged human activities and urban expansion have disrupted the local ecosystems to varying degrees. From a socio-economic perspective, Jinan, as the capital city of Shandong Province, is a key city within the Bohai Economic Rim and the Yellow River middle and lower reaches economic region. It has a high population density and a rapidly developing urban economy. The Jinan segment of the Yellow River Basin is not only vital for water supply and flood control but also contributes significantly to ecological, cultural, and tourism functions of the region. However, with the acceleration of urbanization, conflicts between resource development and ecological conservation have become increasingly prominent, particularly issues such as water resource shortages and degradation of aquatic ecosystems.

In response, the Jinan Municipal Government has prioritized ecological conservation in recent years, formulating and implementing a series of policies and regulations focusing on ecological protection. For instance, soil and water conservation projects and vegetation restoration efforts have been carried out to enhance ecosystem services and promote the sustainable development of the Yellow River Basin. In 2020, the Shandong Provincial Government launched the Jinan Pilot Zone for Transitioning to New Growth Engines (2020–2035) initiative, which explicitly emphasized prioritizing ecological protection, strengthening ecological restoration in the Yellow River Basin, increasing vegetation coverage, and improving the quality of the basin’s ecological environment. These measures provide critical support and context for the ecological conservation and socio-economic development of the study area.

2.2. Sources of Data and Preprocessing

NDVI data can effectively reflect vegetation growth conditions and the dynamic changes in ecosystems. Furthermore, NDVI data have been extensively validated in numerous studies, demonstrating high reliability and stability, making them particularly suitable for research in specific regions. The NDVI data are processed and extracted using MODIS tools, including clipping and data extraction operations, and are ultimately summarized for the growing season data from March to November. TEM and PRE datasets were stored in NETCDF (NC) format and compressed into int16 format to enhance storage efficiency. ArcMap software can be used to open NC files for visualization and to separate monthly TEM and PRE data by their respective bands. Although these meteorological data have certain limitations in spatial resolution, they feature high temporal resolution and wide coverage, allowing for precise capture of climatic variability in the Jinan section of the Yellow River Basin. These data provide crucial climatic background for analyzing ecological dynamics. They are used to investigate the driving forces and correlations of TEM and PRE with NDVI and to predict the future spatial distribution of NDVI. GDP and POP datasets were cropped to extract data specific to the Jinan section of the Yellow River Basin. The selection of these data is based on their strong representativeness of economic and social activities. GDP and POP data provide a foundation for analyzing regional economic conditions and population density. These data are utilized to analyze their driving effects on NDVI and to predict the future spatial distribution of NDVI. Nighttime light data compensate for the limitations of traditional statistical data by reflecting the spatial distribution of socioeconomic activities. This dataset complements other data types and is used to predict the future spatial distribution of NDVI. The origins of the aforementioned data are described in

Table 1. Remote sensing data information and sources.

Data	Spatial Resolution	Time Resolution	Year	Sources
NDVI	1 km	1 month	2008–2022	NASA
TEM	1 km	1 month	2008–2022	ESSDNC
PRE	1 km	1 month	2008–2022	ESSDNC
GDP	1 km	1 year	2008, 2013, 2018, 2022	ESSDNC
POP	1 km	1 year	2008, 2013, 2018, 2022	ESSDNC
Nighttime Light	1 km	1 year	2022	NTPDC

. All the data used in this study were obtained from authoritative institutions, ensuring the accuracy of the data.

2.3. Methods

The study aims to explore the spatio-temporal variations and driving forces of NDVI, and to predict its future spatial distribution. Firstly, slope trend analysis was used to investigate the spatio-temporal change patterns of the NDVI in the Jinan region of the basin of the Yellow River from 2008 to 2022. Furthermore, a FLUS model was employed to predict the distribution of NDVI in 2025. Subsequently, the OPGD was utilized to examine the impact of four key driving forces: PRE, TEM, POP, and GDP, on vegetation coverage. Finally, correlation analysis and lag analysis were conducted to further investigate the correlation between NDVI and TEM, and NDVI and PRE. The specific process is detailed in Figure 2.

2.3.1. Maximum Value Compositing

The Maximum Value Compositing (MVC) method is an internationally recognized method for the statistical analysis of NDVI data. It involves overlaying multiple identical raster images and selecting the maximum value for each grid cell across all images, resulting in a synthesized image [26]. The interference caused by factors such as air pollution, clouds, and solar altitude angle can be eliminated by MVC method. In this study, we utilized MVC method to synthesize NDVI values during the vegetation period (March to November) by choosing NDVI values annually from 2008 to 2022 within the Yellow River region of Jinan Watershed. The calculation formula is as follows:

$$M_{NDVI}=\max \left({\sum }_{i=1}^{15}{NDVI}_{ik}\right),$$

(1)

where i denotes the $M_{NDVI}$ value in the $i$th year; k denotes the NDVI value in the $k$th ten-day period of the $i$th year (i = 1, 2, …, 15).

2.3.2. Slope Trend Analysis

The trend analysis approach describes the linear relationship between the two variables and is capable of simulating each grid’s changing trend. To reflect the intra-annual characteristics of vegetation changes, we use the composite NDVI values during the growing season (March to November) to represent vegetation density [42].

We employ the slope of the univariate linear regression, where an NDVI trend is indicated by a slope > 0 when it is increasing and by a slope < 0 when it is declining. Combined with the significance level (p-value) of the regression coefficient, the inter-annual changes in NDVI in our study region are divided into four types: significant decrease (0.01< p < 0.05, slope < 0), insignificant decrease (p > 0.05, slope < 0), insignificant increase (p > 0.05, slope > 0), and significant increase (0.01 < p < 0.05, slope > 0) [42].

The following is the calculating formula:

$$Slope={\displaystyle \frac{n{\sum }_{k=1}^n\left(k{NDVI}_k\right)-{\sum }_{k=1}^{n}k{\sum }_{k=1}^n{NDVI}_k}{n{\sum }_{k=1}^n{k}^2-{\left({\sum }_{k=1}^{n}k\right)}^2}},$$

(2)

where $k$ indicates the chronological order of years, where $k$ = 1, 2, …, $n$; $n$ represents the cumulative number of years; and ${NDVI}_k$is the NDVI value of the $k$th year.

2.3.3. FLUS Model

The FLUS model is a land type change simulation model built using Cellular Automata (CA) and Markov Chain. It is used to predict future land type scenarios and analyze the effects of land type changes on the environment and socioeconomics. In this study, the FLUS model is utilized in conjunction with natural environmental factors and human activities to predict the NDVI’s geographical dispersion in 2025 [30]. The model setup and parameters are as follows.

1. Calculation of suitability probability: the Artificial Neural Network (ANN) algorithm is employed to generate suitability probabilities for different NDVI classes (

Table 2. NDVI Grade.

Grade	NDVI Value
Very Low Vegetation Density	NDVI ≤ 0.1
Low Vegetation Density	0.1 < NDVI ≤ 0.3
Medium Vegetation Density	0.3 < NDVI ≤ 0.5
High Vegetation Density	0.5 < NDVI ≤ 0.7
Very High Vegetation Density	0.7 < NDVI

) based on natural factors (such as PRE and TEM) and social factors (such as human activities). These probability data are used to represent the suitability levels of NDVI under various vegetation cover types.

2. Setting neighborhood weight factors and transition cost matrix: suitability probability data are input into the model, along with the relevant transition matrix and weight factors. The NDVI transition matrix describes the potential conversions between different NDVI values, where a matrix value of 1 indicates a possible conversion and 0 indicates an impossible conversion. Taking into account the real NDVI conditions in the research region, the transition matrix values for all NDVI classes are set to 1, allowing for mutual conversion. The neighborhood weight factors range from 0 to 1, where values near 0 indicate an easier conversion and values near 1 indicate a more difficult conversion.

3. Verification of simulation accuracy: NDVI data from 2019 and the relevant driving factors are selected for simulation to predict the NDVI values for 2022. The simulated and real NDVI values for 2022 are compared using the Kappa coefficient, assessing the accuracy of the model’s predictions. Furthermore, the NDVI data and driving factors for 2022 are used to predict the spatial changes in NDVI for 2025. The following is the formula used to calculate the $Kappa$ coefficient:

$$Kappa={\displaystyle \frac{P_0-P_e}{1-P_e}}.$$

(3)

In the formula, $P_0$ represents the observed agreement proportion, which is the proportion of the quantity of predictions that match the actual data. $P_e$ denotes the expected agreement proportion, calculated based on random probabilities, reflecting the expected level of agreement in the data between the expected and actual.

The $Kappa$ coefficient varies between 0 and 1, where ideally a $Kappa$ value of 1 indicates complete consistency. A $Kappa$ coefficient nearer 1 denotes increased precision in the simulation outcomes. When the $Kappa$ coefficient exceeds 0.6, it suggests that the model’s overall performance is satisfactory, making it suitable for predicting future trends in vegetation changes.

2.3.4. Correlation Analysis

The correlation analysis primarily focuses on an in-depth study of two or more variables with relevant characteristics, aiming to quantify the degree of association between these variables. There must be a particular connection or probability between the associated elements for correlation analysis to be performed [43]. It is commonly quantified by R or R², where a higher value suggests a more robust correlation between the two factors, and a diminished magnitude indicates a poorer correlation [42]. The equation is as follows:

$$R={\displaystyle \frac{{\sum }_{i=1}^n\left[\left(X_i-\overline{X}\right)\left(Y_i-\overline{Y}\right)\right]}{\sqrt{{{\sum }_{i=1}^n\left(X_i-\overline{X}\right)}^2}\sqrt{{{\sum }_{i=1}^n\left(Y_i-\overline{Y}\right)}^2}}},$$

(4)

where $R$ represents the linear correlation coefficient between the variables $X$ and $Y$; $X_i$ and $Y_i$ represent the values of $X$ and $Y$ in the i-th year, respectively; and $\overline{X}$ and $\overline{Y}$ represent the two variables’ average values over a period of n years, respectively. When $R$ > 0, it signifies a direct relationship existing between the two variables; when $R$ < 0, it denotes an inverse correlation between the two variables.

In this study, we explore the connection between climate change and NDVI by calculating the relationship between NDVI, PRE, and TEM, to understand the corresponding trends between climate variation and vegetation alteration.

2.3.5. Improved OPGD Model

As a spatial analysis tool, the Geodetector aims to reveal the Variability in geographic distribution and further explore the driving elements that underlie it. This method is widely used in various driving force and factor analyses. Its core idea is based on the assumption that if a certain independent variable has an important influence on a dependent variable, then the spatial distributions of the independent variable and the dependent variable should be similar. Through such analysis, the geographic sensor helps us better understand spatial phenomena and the mechanisms behind them. The geospatial sensor comprises interaction detection, environmental detection, parameter detection, and hazard detection [40]. Traditional Geodetector often relies on pre-set parameters, the selection of which is often based on experience or trial and error, lacking a systematic approach and scientific rigor. Therefore, we resorted to the improved Optimal Parameter-based Geographic Detector (OPGD) model, which incorporates an additional parameter optimization module in addition to the four components included in the Geodetector. This module can discretize the independent variables in various ways (such as natural breaks, quantiles, equal interval, and geometric interval) and with multiple numbers of breakpoints based on their characteristics. It then computes the q-score for each pairing of discretization approach and breakpoint count, and identifies the optimal discretization technique and the optimal quantity of breakpoints determined from the maximum q-value. This module can compensate for the defects of poor analysis caused by relying on experience or unified discretization method to deal with multiple independent variables, and precisely evaluate the independent variables quantitatively. It can reduce the fluctuation of dependent variables within the optimal segmentation groups, increase the significant difference among groups, and attain the optimal detection of the influence of influential elements on the geographical distribution of dependent variables [41].

We employ the factor identification and interaction identification within the OPGD to explore the impact of driving factors on NDVI. Factor detection primarily serves to quantify the degree to which independent variables account for the spatial diversity of reliant variables, while identifying interactions focuses on studying whether the interplay between two autonomous variables amplifies or diminishes the influence on reliant variables. The factor identification framework utilized in this study is delineated below:

$$q_{x_i}=1-{\displaystyle \frac{{\sum }_{h=1}^{L_{x_i}}{N}_h{\sigma }_h^2}{N{\sigma }^2}}=1-{\displaystyle \frac{SSW}{SST}},$$

(5)

where $q_{x_i}$ (∈[0, 1]; $i$ = 1, 2, …, $n$) represents the elucidating capability of a certain influencing variable $X_i$, and the greater the q value, the more robust the elucidatory capability; $h$ is the number of stratifications of the predictor variable $X_{i }$or the response variable $Y$; $L_{xi}$ ($x_i$ = 1, 2, …, $n$) is the overall quantity of stratifications; $N_h$ and $N$ depict the quantity of specimens in the $h$ stratum and the area of research, respectively; ${\mathrm{\sigma }}_{\mathrm{h}}^2$ and ${\mathrm{\sigma }}^2$ are the deviations of NDVI in the $h$ layer and the study region, respectively; $SSW$ and $SST$ depict the accumulation of variation in the strata and the aggregate of overall variance within the research region, respectively.

2.3.6. Lagged Correlation Analysis

Lagging refers to the delayed relationship between data and lagged analysis, and it can determine the lag time between data, i.e., whether one data point is influenced by the previous period or several periods. Drawing from the lagged correlation analysis of Fu et al. [42], we investigate the characteristics of the association between NDVI and climate variables, as well as the reaction to changes in climate variables over time. The time-delayed effect of NDVI on changes in climatic conditions is represented by the lagged correlation coefficient $R*$ such as TEM or PRE. Lagged correlation can reveal whether the occurrence and evolution of two processes have a certain correlation in terms of time sequence [44]. The calculation formula is as follows:

$$R*=\max \left\{R_0,R_1,R_2,\dots ,R_n\right\},$$

(6)

where $R_1$, $R_2$, and $R_3$ represent the association coefficients between NDVI and a certain meteorological variable for the current month, the previous month, and two months prior, respectively. If $R*$ equals $R$, it indicates that the reaction lag time of NDVI to fluctuations in that climatic factor is $n$ months.

3. Results and Analysis

3.1. Inter-Annual Dynamic Changes in NDVI

The values of NDVI in the Jinan section of the Yellow River Basin significantly increased from 0.476 in 2008 to 0.5241 in 2022, with an average increase of 0.095 (see

Table 3. NDVI values for the growing season.

Year	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021	2022
NDVI Value	0.476	0.466	0.446	0.473	0.473	0.469	0.456	0.469	0.479	0.488	0.490	0.469	0.506	0.512	0.524

and Figure 3). This change reflects the gradual improvement in vegetation cover. Environmental protection measures implemented by the government, such as afforestation, returning farmland to forests and grasslands, wetland protection, and bans on grazing and hunting, have effectively reduced land degradation and desertification, promoting vegetation recovery. These measures have played a positive role in vegetation growth.

3.2. Spatio-Temporal Variation in Patterns of Vegetation Cover

In order to clearly portray the changes in vegetation covering over the last 15 years in terms of both space and time, the data from this period were divided into three time periods (2008–2012, 2013–2017, and 2018–2022). A slope trend analysis was conducted on the NDVI data for these three groups of data to derive the trends in variation in NDVI (Figure 4 and Figure 5).

During the period of 2008–2012, 53% of the Jinan section of the Yellow River Basin showed an increasing vegetation trend, with 2% exhibiting a significant increase and 51% showing an insignificant increase. Conversely, 47% of the region displayed a decreasing trend, with 3% showing a significant decrease and 44% showing an insignificant decrease.

By 2013–2017, the area with an increasing vegetation trend rose significantly to 84%, with 11% experiencing a significant increase and 73% showing an insignificant increase. Areas with a decreasing trend declined to 16%, with no significant decreases observed and 16% showing an insignificant decrease. In 2018–2022, the area with an increasing vegetation trend further expanded to 92%, with a significant increase observed in 16% of the region and an insignificant increase in 76%. Areas with a decreasing trend dropped to 7%, with no significant decreases and 7% showing an insignificant decrease. These findings indicate that the trend of increasing vegetation cover has strengthened significantly in recent years, reflecting the effectiveness of national ecological governance and environmental protection policies, particularly in promoting ecological restoration. From a spatial perspective, the areas showing an increasing vegetation trend were primarily located in the mountainous regions of Changqing District, Pingyin County, and Laiwu District. These areas have benefited from ecological restoration projects and soil and water conservation efforts. In contrast, areas with a decreasing vegetation trend were mainly concentrated in urban areas, where the expansion of urban land use has resulted in the conversion of natural or semi-natural land, leading to reduced vegetation cover. Overall, between 2008 and 2022, 81% of the region exhibited an increasing vegetation trend, while 19% showed a decreasing trend. Suburban and mountainous areas demonstrated significant improvements in vegetation cover, indicating the success of ecological protection policies in these regions. Urban areas, however, showed a slight decline in vegetation cover, highlighting the localized environmental impact of urban development. The study reveals that ecological protection measures implemented by national and local governments in recent years, such as reforestation and soil and water conservation initiatives, have significantly improved vegetation cover in the Jinan section of the Yellow River Basin, particularly in mountainous regions. However, the observed reduction in vegetation cover in urban areas underscores the need to prioritize ecological principles during urbanization to achieve a balance between economic development and environmental protection.

3.3. NDVI Prediction for the Jinan Section of the Basin of the Yellow River in 2025

This study used the FLUS model based on 2019 NDVI data and related climate factors and human activity data to predict the distribution of NDVI in 2022. Compared with the actual distribution in 2022, the obtained Kappa coefficient is 0.655, which satisfies the standards for accuracy and can be applied to forecast the NDVI’s geographical distribution in the research region in 2025.

The predicted spatial distribution of NDVI in 2025 (Figure 6 and Figure 7) shows that, compared with 2022, the proportion of high-density coverage areas has increased, rising from 53.91% in 2022 to 55.35%. At the same time, the proportion of low coverage density areas decreased from 0.74% in 2022 to 0.54%, the percentage of areas with medium coverage density dropped from 44.26% to 44.03% and the change in extremely high coverage density areas was relatively small, from 0.07% to 0.06%. These changes are mainly reflected in the transformation of medium coverage density areas to high coverage density areas, as well as the reduction in low coverage density areas (Figure 8). This trend can be attributed to several factors: firstly, the improvement of climate conditions in recent years (such as increased PRE and TEM changes) may have promoted vegetation growth, thereby increasing NDVI values. Second, some facets of human endeavors, such as those aimed at restoring vegetation and land management measures, may also have a positive impact on vegetation cover. The results suggest that vegetation cover in the study area will continue to improve, with a spatial distribution pattern tending towards higher coverage density. This provides a scientific basis for regional ecological protection and sustainable land management, while also highlighting the need to continue monitoring the impact of human activities on vegetation cover to ensure ecosystem stability and healthy development.

3.4. Assessment of Influencing Factors

3.4.1. Selection of Indicators

TEM and PRE are important factors affecting vegetation growth. Consequent to the phenomenon of global alterations in climatic conditions, resulting in deleterious effects on the vegetation’s growth environment and a marked reduction in vegetation coverage. GDP is a significant gauge to assess the economic advancement status of a locality, mirroring the level of productivity and economic activity within that region. An increase in GDP usually means that the region has more resources for environmental protection and vegetation construction, which may have a positive impact on NDVI. POP acts as a marker of the degree of congestion and the intensity of human activities within a given region. High POP areas may imply greater land use pressure and resource consumption, which could influence the development of plants and NDVI. At the same time, POP may also affect land utilization patterns and the execution of ecological and environmental protection measures, thereby indirectly affecting NDVI. Studying the impact of these elements on plant NDVI has important practical significance. Therefore, the above four types of data are selected as driving factors for our research.

3.4.2. Analysis of Factor Detection

Utilizing the factor detection outcomes of the Optimal Parameter-based Geographic Detector (

Table 4. Factor detection results.

Year	2008			2013			2018			2022
	q	p		q	p		q	p		q	p
X1	0.015	0		0.073	0		0.031	0		0.032	0
X2	0.019	0		0.014	0		0.008	0		0.005	0
X3	0.040	0		0.046	0		0.024	0		0.040	0
X4	0.021	0		0.048	0		0.053	0		0.018	0

X1: TEM; X2: PRE; X3: GDP; X4: POP, the same below.

), we delved into the extent of influence exerted by various drivers on vegetation coverage. It can be observed that, given p ≤ 0.05, the foremost three factors affecting NDVI in 2008 were POP (0.021), GDP (0.04), and PRE (0.019). In 2013, the three main variables influencing NDVI were POP (0.048), TEM (0.073), and GDP (0.046). In 2018, the three main variables influencing NDVI were POP (0.053), TEM (0.031), and GDP (0.024). In 2022, the three main variables influencing NDVI were POP (0.018), PRE (0.032), and GDP (0.04). From 2008 to 2022, POP consistently emerged as the most significant driver of NDVI, likely due to its close association with the intensity of human activities and changes in land use. The impact of TEM and PRE exhibited temporal fluctuations, indicating that climatic conditions exhibit spatial and temporal variability in their effects on vegetation coverage. The role of GDP reflects the potential of economic development to drive ecological system construction. These findings suggest that both natural and socio-economic factors should be considered comprehensively, with their varying influences over time, to provide a scientific basis for formulating regional ecological protection policies.

3.4.3. Examination of Interaction Identification

Detection of individual factors solely assesses the impact of individual drivers on vegetation coverage, as well as interaction detection offers a more nuanced analysis, examining how the interplay between these factors influences the spatiotemporal variations in vegetation coverage. Utilizing the Optimal Parameter-based Geospatial Detector (OPGD), we analyzed the interaction among driving factors. The results show that, in 2008, the factors with significant impacts on NDVI were TEM ∩ GDP (0.124), PRE ∩ GDP (0.168), and TEM ∩ POP (0.124). The factors that had a greater impact on NDVI in 2013 were PRE ∩ GDP (0.266), TEM ∩ GDP (0.266), and GDP ∩ POP (0.269). In 2018, the factors that had a greater impact on NDVI were PRE ∩ GDP (0.082), PRE ∩ POP (0.107), and TEM ∩ POP (0.152). In 2022, the major factors affecting NDVI were GDP ∩ TEM (0.175), PRE ∩ GDP (0.165), and POP ∩ GDP (0.079) (Figure 9). These results indicate that the combined effects of driving factors have a significantly greater impact on vegetation coverage than the independent effects of individual factors, suggesting that vegetation coverage changes are the result of the interplay of multiple factors. In particular, interactions involving GDP repeatedly emerged as key factors, highlighting the regulatory role of economic development in influencing vegetation coverage when combined with natural factors. Additionally, the interactions between temperature, precipitation, and population demonstrated dynamic changes across different years. This suggests that when formulating ecological protection policies, it is essential to consider the synergistic effects of multiple factors, especially the mutually reinforcing relationship between economic development and the natural environment, in order to develop more comprehensive and science-based management strategies.

3.5. Correlation Exploration

Heat and moisture are essential conditions for plant growth, and appropriate PRE and TEM can promote rapid plant development. To reveal the association between NDVI, TEM, and PRE, we processed the NDVI information to ascertain the correlation analysis results among NDVI, TEM, and PRE, to understand the relationship between PRE, TEM, and NDVI in the Jinan section of the Yellow River Watershed throughout the previous fifteen years.

3.5.1. Association Between NDVI and PRE

To fully reflect the spatio-temporal evolution of the correlation between NDVI and PRE, the past fifteen years were divided into three time periods (2008–2012, 2013–2017, and 2018–2022). As can be seen from Figure 10 and Figure 11, during these three time periods, the proportion of regions where there is a positive correlation between NDVI and PRE in the Jinan section of the basin of the Yellow River represented approximately 37%, 76%, and 89% of the entire region, respectively. These areas were mainly distributed in mountainous regions such as Changqing, Pingyin, Laiwu, Gangcheng. With the progress of urbanization, some originally non-vegetated areas may have been transformed into green spaces or agricultural land, and these changes have also contributed to increased vegetation coverage and have established a positive correlation with PRE. During these three time periods, the percentage of areas with a negative relationship between PRE and NDVI in the Jinan section of the basin of the Yellow River represented approximately 63%, 24%, and 11% of the entire region, respectively. The majority of these areas were found in urban agglomeration areas. Overall, during the period of 2008–2022, 90% of the areas showed positive correlation between NDVI and PRE over the past fifteen years, while 10% had showed negative correlation. With the development of urbanization, vegetation in urban agglomeration areas may have been damaged to some extent. The study indicates that PRE is a key factor influencing vegetation cover change in the Jinan section of the Yellow River Basin, particularly in mountainous areas, where its positive correlation with NDVI has significantly strengthened. This reflects the crucial role of natural factors in vegetation restoration. At the same time, the impact of urbanization on vegetation cover shows dynamic changes. Initially, urbanization may lead to vegetation degradation, but in the later stages, urban greening and ecological restoration efforts can effectively promote vegetation recovery. Future research should further optimize urban ecological management and strengthen the integrated study of the interaction between natural factors and human activities, providing a scientific basis for regional ecological protection and sustainable development.

3.5.2. Relationship Between NDVI and TEM

To fully reflect the spatio-temporal evolution of the correlation between NDVI and TEM, the past fifteen years were divided into three periods (2008–2012, 2013–2017, and 2018–2022). As can be seen from Figure 12 and Figure 13, during these three time periods, the regions where NDVI and TEM had a positive association in the Jinan section of the basin of the Yellow River represented approximately 76%, 85%, and 18% of the entire region, respectively. These areas were mainly distributed in the mountainous regions of Changqing, Pingyin, Laiwu, Gangcheng. During these three time periods, the areas with a negative correlation between NDVI and TEM in the Jinan section of the basin of the Yellow River represented approximately 25%, 15% and 81% of the entire region, respectively. These areas were mainly distributed in urban agglomeration areas. From 2018 to 2022, the percentage of regions where NDVI and TEM have a positive association dropped sharply to 18%, while the proportion of areas with a negative correlation rose to 81%. The acceleration of urbanization may have led to a significant increase in TEM in urban agglomeration areas, resulting in the “urban heat island effect”. This effect not only raises the TEM in urban areas but may also have a negative impact on vegetation growth, resulting in a negative relationship between TEM and vegetation covering. Overall, between 2008 and 2022, the proportion of regions in the Jinan section of the Yellow River Basin showing a positive correlation between vegetation cover and temperature was 78%, while the negative correlation regions accounted for 22%. Despite the significant impact of the “urban heat island effect” in urban areas, the relationship between NDVI and temperature showed phased changes. However, the positive correlation remained dominated, suggesting that the natural environmental conditions in the region have a positive effect on vegetation growth. Future management should particularly focus on the dual impact of urbanization on temperature and vegetation, developing targeted ecological regulation measures to mitigate the negative effects of urbanization while preserving and enhancing the stability of the regional ecosystem.

3.5.3. Temporal Delay of NDVI in Summer on TEM

Due to the significant characteristics of TEM in summer, we analyzed the correlation coefficients between NDVI in summer and TEM, TEM one month earlier, and TEM two months earlier in the Jinan section of the basin of the Yellow River from 2008 to 2022 based on NDVI data in summer, and analyzed the time-lagged response of NDVI to TEM. Figure 14 showed the distribution of correlation coefficients between NDVI and TEM at different lag times: June–August, May, and April.

As can be seen from Figure 14, the lag effect of NDVI in summer on TEM was not strong in the Jinan section of the basin of the Yellow River. The areas with a positive correlation between NDVI in summer and TEM represented 63% of the entire region. The areas where NDVI is positively correlated with TEM in May and April account for 76% and 49% of the entire region, respectively. It can be observed that the proportions of the areas with different lag times between NDVI and TEM showing positive correlation coefficients were roughly similar. Among them, the area with a positive correlation between NDVI in summer and TEM in May was the largest, and the area with a significant positive correlation was also relatively large. Therefore, the lag time of NDVI’s response to TEM in summer was approximately one month.

In addition, spatially speaking, as shown in the analysis of the lag effect between NDVI in summer and TEM (Figure 14a), the proportion of areas with a positive correlation between vegetation coverage and TEM in mountainous regions of the Jinan section of the basin of the Yellow River was relatively high. According to the analysis of the lag effect between summer NDVI and TEM lagged by one month (Figure 14b), the proportion of areas with a positive correlation between vegetation coverage and TEM in mountainous regions of the Jinan section of the basin of the Yellow River was relatively high. The analysis of the lag effect between NDVI in summer and TEM lagged by two months (Figure 14c) revealed that the proportion of areas with a positive correlation between vegetation coverage and TEM in urban areas of the Jinan section of the basin of the Yellow River was relatively high. The study indicates that the response of summer NDVI in the Jinan section of the Yellow River Basin to temperature changes exhibits a certain time-lag effect, with the lag time being approximately one month. Spatially, vegetation in urban areas shows a stronger time-lag response to temperature, while vegetation in mountainous regions is more sensitive to temperature changes with longer time lags. This result suggests that future studies should consider the temperature variation characteristics at different lag times to more accurately assess the dynamic impact of temperature on vegetation growth, providing scientific basis for urban greening and ecological protection in mountainous areas.

3.5.4. Time Delay of NDVI in Summer Regarding PRE

As PRE in summer is relatively abundant, it can better reflect the lag effect of NDVI on PRE. Therefore, based on the NDVI data in summer from 2008 to 2022 in the Jinan section of the basin of the Yellow River, we calculated the correlation coefficients between NDVI in summer (June–August) and PRE in summer, as well as PRE one month and two months earlier, respectively, to analyze the time-lagged response of NDVI to PRE. Figure 15 shows the distribution of correlation coefficients between NDVI and PRE at different lag times: June–August, May, and April.

As can be seen from Figure 15, the lag effect of NDVI in summer on PRE was not strong in the Jinan section of the basin of the Yellow River. The areas with a positive correlation between NDVI in summer and PRE represented 54% of the entire region. The areas where NDVI was positively correlated with PRE in May and April represented 55% and 20% of the entire region, respectively. From the analysis of the correlation between NDVI in summer and PRE at different lag times, it can be seen that there was no significant lag effect of summer NDVI on PRE.

Spatially speaking, as shown in the analysis of the lag effect between NDVI in summer and PRE (Figure 15a), the proportion of areas with a positive correlation between vegetation coverage and PRE in the mountainous regions of the Jinan section of the basin of the Yellow River was relatively high. According to the analysis of the lag effect between NDVI in summer and PRE lagged by one month (Figure 15b), the proportion of areas with a positive correlation between vegetation coverage and PRE in urban areas of the Jinan section of the basin of the Yellow River was relatively high. The analysis of the lag effect between NDVI in summer and PRE lagged by two months (Figure 15c) revealed that the correlation between vegetation coverage and PRE in various regions of the Jinan section of the basin of the Yellow River was generally low. The spatial distribution of the correlation between NDVI in summer and PRE lagged by one month and two months was roughly the same. The areas with positive correlation with PRE lagged by two months were smaller. Additionally, there was a small lag effect of PRE on suburbs, mountainous areas, and urban areas. The study indicates that the lag effect of precipitation on summer NDVI in the Jinan section of the Yellow River Basin is relatively short, primarily concentrated in the same month or within one month of lag, while the effect with a two-month lag is not significant. Vegetation in mountainous areas shows a more direct dependence on precipitation, whereas urban green spaces exhibit higher sensitivity to lagged precipitation. This suggests that ecological management and greening measures should pay closer attention to current precipitation and climatic conditions within a one-month lag to more precisely optimize the positive impact of precipitation on vegetation coverage, thereby promoting the sustainable development of the regional ecological environment.

4. Discussion

4.1. Main Research Methods

In terms of research methods, this study employed slope trend analysis to explore the spatio-temporal evolution characteristics of vegetation cover in the Jinan section of the Yellow River Basin [27,28]. Then, based on the FLUS model, the spatial distribution of NDVI in 2025 was predicted [30], and the spatio-temporal changes and future trends of NDVI in the Jinan section of the Yellow River Basin were analyzed. Compared with previous spatial analysis methods (standard deviational ellipse and Hurst index analysis) [29], these methods provide a more detailed observation of subtle regional changes and higher predictive accuracy. Subsequently, the study used an OPGD to analyze the impact of driving factors on NDVI [41]. Unlike traditional geographic detectors [40], this method can automatically select the optimal parameter combinations, thereby reducing the arbitrariness and subjectivity of manual parameter selection. By analyzing the driving forces of influencing factors, the study offers deeper insights into the effects of climate factors and human activities on NDVI. Finally, correlation analysis and lag effect tests were applied to determine the relationship between meteorological factors, such as TEM and PRE, and vegetation cover [35,36,37]. This approach enables a more accurate identification of the spatial distribution of the correlation between meteorological factors and NDVI, providing scientific evidence for ecological protection and climate change adaptation.

4.2. Main Research Findings and Significance

From 2008 to 2022, the NDVI values in the Jinan section of the Yellow River Basin showed a significant upward trend, with an average annual growth rate of 3.6% per decade. This is broadly consistent with the findings of Han et al. regarding the NDVI growth trend in the upper reaches of the Yellow River [44]. This trend can be attributed to the implementation of a series of ecological protection and restoration projects, such as the construction of the Yellow River Ecological Landscape Belt and the development of model cities along the river. The study found that vegetation coverage in mountainous and sparsely populated suburban areas continued to increase, while urban areas experienced degradation. Urban expansion and the growing demand for construction land significantly contributed to the reduction in urban vegetation, which is consistent with existing research on the impacts of urbanization on the ecological environment [45]. The 2025 NDVI forecast indicates that high-density vegetation areas will continue to expand, suggesting that the long-term effects of ecological protection policies are gradually becoming evident. The changes in the spatial pattern of vegetation coverage in the Jinan section may have profound impacts on regional climate regulation as well as soil and water conservation [46]. The increase in vegetation in mountainous and suburban areas contributes to enhancing the region’s soil retention capacity and water source conservation functions. However, the reduction in urban vegetation may weaken the ecological functions of urban areas, such as exacerbating the urban heat island effect [47]. Therefore, future urban planning should further optimize land use layout to mitigate the adverse effects of urbanization on ecosystems. The findings of this study validate the initial hypothesis that changes in vegetation cover in the Jinan section of the Yellow River Basin are driven by both climate factors (TEM, PRE) and human activities, with significant differences in trends and spatial distributions across different regions. This is consistent with the close relationship between climate and vegetation identified by Pang et al. in their NDVI survey of Asian vegetation [48]. Additionally, lag effect analysis shows that temperature and precipitation have time-lag effects on NDVI, further revealing the long-term impacts of climate change on vegetation cover, which is also consistent with findings in the related literature [49,50].

This study systematically analyzed the spatiotemporal characteristics, driving factors, and future trends of vegetation cover change in the Jinan section of the Yellow River Basin, providing scientific evidence for ecological protection and high-quality development in the region [51]. The FLUS model was used to predict future NDVI changes and offer guidance for regional ecological planning, the OPGD quantified the contribution of driving factors to provide a basis for formulating more targeted ecological protection policies and the lag effect study revealed the long-term impact of climate change on vegetation cover to support climate adaptation management.

4.3. Methodological Limitations and Future Research Directions

Despite the achievements of this study, several limitations remain. Firstly, the resolution of NDVI data may be insufficient to capture small-scale vegetation changes, such as urban micro-greenspaces and riparian vegetation, and lower-resolution data may hinder the understanding of spatial heterogeneity in complex terrain areas. Secondly, the study focused solely on the influences of climatic data and human activities on vegetation cover without considering other potential factors such as soil type [52]. Future research can focus on improving data accuracy by integrating high-resolution remote sensing data and multi-source data to better investigate localized vegetation coverage changes. Meanwhile, a cross-regional comparative study can be carried out within the Yellow River Basin to explore the similarities and differences in vegetation coverage changes driven by ecological protection policies and natural conditions, so as to formulate more universally applicable protection strategies [53].

5. Conclusions and Suggestions

In this paper, the slope trend analysis method was used to explore the spatio-temporal evolution of vegetation coverage in the Jinan section of the basin of the Yellow River. The future trend of NDVI was then simulated and predicted based on the FLUS model. Then, combined with the improved optimal parameters of the geographic detector, the impact of driving factors on vegetation coverage was analyzed. Finally, correlation analysis and lag test methods were used to obtain the correlation between TEM, PRE meteorological factors, and vegetation coverage. The following conclusion can be drawn.

(1) Vegetation cover trends and hypothesis validation.

Over the past 15 years, the NDVI in the Jinan section of the Yellow River Basin has shown an overall increasing trend, rising from 0.476 in 2008 to 0.5241 in 2022, validating the hypothesis of continuous improvement in vegetation cover. However, this trend varies across regions, with vegetation cover in suburban and mountainous areas significantly increasing, while urban areas experience some degree of degradation due to expansion and land development. Projections indicate that by 2025, high-coverage areas will increase from 44.26% in 2022 to 55.35%, suggesting that the positive effects of ecological protection policies will continue.

(2) Dynamic changes in driving factors and hypothesis comparison.

The driving factors of NDVI exhibited dynamic changes across different years, confirming the hypothesis that climate factors and human activities jointly drive changes in vegetation cover. In the early years (2008), population and GDP were the primary driving factors. After 2013, temperature (TEM) became the key factor, and from 2018 onwards, the influence of precipitation (PRE) significantly increased, reflecting the gradually stronger dominance of climate factors in influencing vegetation cover.

(3) Correlation and lag effects of meteorological factors.

NDVI showed different correlations with PRE and TEM in different periods. From 2008 to 2017, there was a significant positive correlation with temperature and precipitation, particularly in mountainous areas. However, from 2018 onward, NDVI maintained a positive correlation with precipitation but shifted to a negative correlation with temperature, exhibiting a one-month lag effect, especially in urban areas. This suggests that the long-term impact of climate change on vegetation cover has regional differences.

This study provides a scientific basis for understanding the dynamics and driving mechanisms of vegetation cover in the Jinan section of the Yellow River Basin, offering support for regional ecological conservation and policymaking. Future research should focus on the following: (1) integrating remote sensing big data and higher-precision models to optimize NDVI prediction accuracy; and (2) assessing the long-term impacts of climate change on regional ecosystems to guide adaptive management strategies.

The above conclusion can provide several policy implications for environmental protection and sustainable development in the Jinan section of the basin of the Yellow River. We propose the following suggestions for ecological protection in the Jinan section of the basin of the Yellow River. (1) Climate adaptation strategy: conduct regional climate change risk assessment to understand the impact of climate change such as PRE, TEM changes, and extreme weather on the Jinan section of the basin of the Yellow River. Develop climate adaptation strategies to mitigate the negative impacts of climate change, such as improving water resource management, flood control measures, forest and grass vegetation restoration, and drought management plans. (2) Urban development: develop urban development plans to ensure coordination between urban expansion and farmland, wetlands, and nature reserves, promote sustainable development within the city, and improve the quality of urban life.