Spatial and Temporal Characteristics of Land Use Changes in the Yellow River Basin from 1990 to 2021 and Future Predictions

Abstract

Studying spatial and temporal characteristics of land use changes and the driving factors in the Yellow River Basin as well as simulating and predicting future land use is crucial for resource management, ecological protection, and regional sustainable development in the Yellow River Basin. Based on the China Land Cover Dataset (CLCD) of the Yellow River Basin from 1990 to 2021, this study employs various methods such as the Mann–Kendall test and sliding t-test, land use dynamics, the land use transfer matrix, the standard deviation ellipse, the center of gravity migration model, and a geographic detector to explore the spatial and temporal characteristics of land use changes and driving forces in the Yellow River Basin over the past 30 years. Additionally, the study predicts land use types in the study area for the year of 2030 by using the Future Land Use Simulation (FLUS) model. The results show the following: (1) From 1990 to 2021, the area of forest, grassland, water, and impervious surfaces increased significantly, while the area of cropland, shrub, barren land, and wetlands decreased significantly. The most actively changing land use types are cropland, grassland, barren land, and impervious surfaces. (2) The center of gravity for shrub and impervious surfaces shifted westward, while wetlands showed a trend of obvious concentrated distribution, and the remaining land use types exhibited stable directional distributions. (3) Economic factors had a stronger driving effect on land use changes than topographic and climatic factors. The land use changes in the Yellow River Basin are influenced by the coordinated driving forces of multiple factors. (4) In 2030, the main land use types in the Yellow River Basin are still expected to be cropland, grassland, and forest. However, there will be a significant expansion of impervious surfaces and forest land, with substantial encroachment on cropland and grassland.

1. Introduction

Land cover (LC) is a fundamental parameter for environmental and climate change research [1]. Land use/cover change (LUCC) refers to the transformation of land use types and structures within specific temporal and spatial ranges, encompassing changes in the quantity, spatial pattern, and function of land use. LUCC directly reflects the interaction between human activities and the natural environment [2] and impacts regional economic and ecological environmental quality [3]. The “International Geosphere and Biosphere Program” (IGBP) and the “International Human Dimensions Programme on Global Environmental Change” (IHDP) listed LUCC as a core content of global change research in 1993 [4,5]. The study of the driving forces of land use change involves analyzing the factors that lead to changes in land use patterns and purposes, as well as the underling mechanisms driving these changes [6]. In recent decades, rapid economic development and population growth have led to significant land use changes, exacerbating the conflict between human needs and land resources [7]. Therefore, studying land use change and its driving forces is crucial for understanding the current status of land use and achieving coordinated regional development.

In recent years, the rapid development of remote sensing technology has provided a wealth of satellite image data products for the study of land use change [8], including FROM_GLC [9], Globeland30 [10], MCD12Q1 [11], ESACCI_LC [12], and CLCD [1]. Among these, the China Land Cover Dataset (CLCD) stands out, as it is based on all available Landsat data on Google Earth Engine (GEE). Professor Huang Xin constructed spatiotemporal features and combined them with a random forest classifier to obtain classification results. He also proposed a post-processing method that includes spatiotemporal filtering and logical reasoning to further improve the spatiotemporal consistency of the CLCD. Wu and Shao demonstrated that the CLCD dataset has higher recognition accuracy [13,14,15]. Therefore, this paper utilizes the CLCD dataset to analyze LUCC.

Many scholars have employed various methods to study land use change and its driving forces, as well as to simulate and predict land use. They have used methods such as land use dynamics [16,17,18], the land use transfer matrix [19], and geo-information maps [20,21] to analyze the characteristics of land use quantity and structural changes. Chord diagrams and Sankey diagrams have been utilized to visualize transfer matrices. The gravity center migration model [22,23] and the landscape pattern index [24,25] have been applied to study changes in the spatial patterns of land use. Additionally, methods like principal component analysis [26,27,28] and geographic detectors [29,30,31] have been used to explore the driving forces of land use change. Compared with traditional driving force analysis methods, geographic detectors can not only identify complex non-linear relationships but also explore the interactions between different driving factors. The FLUS model is often used to simulate and predict future land use changes. By integrating artificial neural networks (ANNs) and various geographic information system (GIS) technologies, it can dynamically adapt to different driving factors of land use change, thereby improving the accuracy of future land use change simulations [32].

The Yellow River Basin is an important ecological barrier area in China, shouldering the dual responsibilities of ecological protection and economic development [33]. Most existing studies focus on specific parts of the Yellow River Basin, such as the Ten Tributaries Basin [34], the Henan section [35], the Ningxia–Inner Mongolia section [36], and the water conservation zone [37]. There is a lack of research on the entire region, and comprehensive analyses of the spatial and temporal changes in land use, their driving forces, and future simulation predictions for the Yellow River Basin are scarce. Building on previous research, this study takes the entire Yellow River Basin as the study area, based on the CLCD dataset from 1990 to 2021. The Yellow River Basin is divided into upper, middle, and lower reaches, and the study period is divided into three phases using mutation analysis. This study combines land use dynamics and a transfer matrix, visualizing the land use transfer matrix with a Sankey diagram, to comprehensively analyze the quantity and structural changes in land use in the Yellow River Basin. Furthermore, this study uses the center of gravity migration model and standard deviation ellipse to analyze changes in spatial patterns, and employs a geographic detector to analyze the driving effects of economic, climate, and topographic factors on LUCC in the Yellow River Basin. Based on these analyses, the FLUS model is used to predict land use changes in the study area for 2030. This research contributes to a comprehensive understanding of the dynamic evolution of land use in the basin and provides references for reconciling human–environment conflicts and promoting sustainable development in the Yellow River Basin.

2. Materials and Methods

2.1. Study Area

The Yellow River is the second largest river in China, originating from the Yugur Basin at the northern foot of the Bayankala Mountains on the Qinghai–Tibet Plateau. It flows west to east through nine provinces and regions, including Qinghai, Sichuan, Gansu, Ningxia, Shaanxi, Henan, and Shandong, finally emptying into the Bohai Sea. The Yellow River Basin is illustrated in Figure 1. The total basin area is approximately 79.5 × 10⁴ km², accounting for around 8% of the country’s total area. The Yellow River Conservancy Commission of the Ministry of Water Resources divides the Yellow River into three sections: from the source of the river to Hekou Town in Tokto County, Inner Mongolia, as the upper reaches, with a basin area of 42.8 × 10⁴ km², from Hekou Town to the Huayuankou area of Taohuayu in Henan as the middle reaches, with a basin area of 34.4 × 10⁴ km², and from Taohuayu to the estuary as the lower reaches, with a basin area of 2.3 × 10⁴ km². The terrain of the basin is high in the west and low in the east, spanning China’s first, second, and third terraces. The upper reaches are primarily mountainous, while the lower reaches consist mainly of plains and hills. The basin mainly belongs to the southern temperate, mid-temperate, and plateau climate zones. Annual precipitation and average annual temperature increase from northwest to southeast, and the level of economic development, also show an increasing trend from the upper reaches to the middle and lower reaches.

2.2. Data

The land-use-type data from 1990 to 2021 in this paper adopt the CLCD data published by Professor Huang Xin of Wuhan University, with a spatial resolution of 30 m (available at https://zenodo.org/records/5816591#.ZAWM3BVBy5c accessed on 15 September 2024); China’s administrative division data come from CnOpenData (available at https://www.cnopendata.com/); the river dataset is provided by the Yellow River Data Center (available at http://huanghe.ncdc.ac.cn); the DEM data use the SRTM data from the Geospatial Data Cloud platform (available at https://www.gscloud.cn/), with a spatial resolution of 90 m, and the slope data are from the DEM data processing; population and GDP data are obtained from the Resource and Environmental Science Data Platform (available at https://www.resdc.cn/), with a spatial resolution of 1 km; temperature and precipitation data are from National Earth System Science Data Center, with a spatial resolution of 1 km (available at http://www.geodata.cn).

2.3. Methodologies

2.3.1. Mann–Kendall and Sliding t-Test

The Mann–Kendall (M-K) test is a quantitative non-parametric method that is particularly useful for analyzing time series data. It does not require the sample to follow a specific distribution, can handle missing data without affecting the results, and is robust against a few outliers. The M-K test is highly practical and is used to detect trends and identify whether mutations occur in the time series data [38]. In this study, the M-K test is conducted at a significance level of 0.05. If the absolute value of the statistical variable Z is greater than 1.96, the trend is considered significant; otherwise, it is insignificant. The sliding t-test is a statistical method used to detect significant changes or “mutations” in data over time by comparing the average values of different time periods within a dataset [39]. This paper combines the M-K test and sliding t-test to analyze the area change trends of land use types in the Yellow River Basin and to identify the mutation point years.

2.3.2. Land Use Dynamics and Transfer Matrix

Single land use dynamics can directly reflect the average change rate of each land use type area in the Yellow River Basin over a specific period. Its expression is as follows [20]:

(1)$K={\displaystyle \frac{U_m-U_n}{U_n}}\times {\displaystyle \frac{1}{t}}\times 100\mathrm{\%}$

where $K$ is the single dynamic degree; $U_n$ is the area (km²) in the first year of the study period; $U_m$ is the area in the last year of the study period; and $t$ is the length of the study period.

Comprehensive land use dynamics can reflect the average change rate of the overall land use type in the Yellow River Basin within a certain period. Its expression is represented as follows [40]:

(2)$LC=\left[{\displaystyle \frac{{\sum }_{\mathrm{a}=1}^n∆L{U}_{\mathrm{a}-\mathrm{b}}}{{\sum }_{\mathrm{a}=1}^{n}L{U}_a}}\right]\times {\displaystyle \frac{1}{t}}\times 100\mathrm{\%}$

where $LC$ is the comprehensive land use dynamics during the study period; ${LU}_a$ is the area (km²) in the first year of the $a$-th land use type during the study period; $∆L{U}_{a-b}$ is the absolute value of the net change area from the $a$-th land use type to the $b$-th land use type during the study period; $t$ is the length of the study period; and $n$ is the number of land use types.

The land use transfer matrix is a method used to analyze land use changes. It can clearly show the increase or decrease in various types of land use in the form of a matrix, and can quantitatively describe the transfer amount and transfer direction between different land use types, which is helpful for accurate analysis of land use changes. Its expression is typically represented as follows [41]:

(3)$S_{ij}=\left[\begin{array}{cc}\begin{array}{cc}{S}_{11}& S_{12}\\ S_{21}& S_{22}\end{array}& \begin{array}{cc}\cdots & S_{1n}\\ \cdots & S_{2n}\end{array}\\ \begin{array}{cc}⋮& ⋮\\ S_{n1}& S_{n2}\end{array}& \begin{array}{cc}\ddots & ⋮\\ \cdots & S_{nn}\end{array}\end{array}\right]$

where $S_{ij}$ is the area of the $i$-th land use type in the initial period of the study period converted to the $j$-th land use type at the end, and n is the number of land use types.

2.3.3. Standard Deviational Ellipse and Center of Gravity Migration Model

The size of the standard deviation ellipse can be used to characterize the degree of concentration of data distribution. The distribution direction of the major axis reflects the main trend of spatial distribution, and the length of the minor axis reflects the degree of concentration. This paper uses the flattening of the ellipse, that is, the ratio of the difference between the major and minor axes to the major axis, to judge whether the directionality of data distribution is obvious. The larger the flattening of the ellipse, the more obvious the directionality of data distribution, which can be used to reflect the distribution characteristics of the spatial pattern of land use types [42,43].

The center of gravity migration model can analyze the movement characteristics of the center of gravity for each land use type within a study area. This model is used to illustrate how the center of gravity of a specific land use type changes over time in spatial evolution processes. The formula for calculating the center of gravity migration rate is as follows [44]:

(4)$V_{\left(i,m-n\right)}={\displaystyle \frac{H\times \sqrt{{(X_m-X_n)}^2+{(Y_m-Y_n)}^2}}{t_m-t_n}}$

where $X_m$, $Y_m$, $X_n$, and $Y_n$ are the longitude and latitude coordinates of the center of gravity of the land use type in the $m$-th and $n$-th years respectively; $H$ is a constant with a value of 111.111 km; $V_{(i,m-n)}$ is the migration rate of the center of gravity of the ith land use type from $n$ to $m$ years; and $t_m$ and $t_n$ are the last and first years of the study.

2.3.4. Geographical Detector

The geographical detector method is widely recognized for its ability to detect spatial heterogeneity, making it highly applicable in studying the influencing factors of natural, economic, and social phenomena. Unlike many other methods, geographical detector analysis is less constrained by specific assumptions and demonstrates notable advantages in handling diverse types of data. In this paper, the factor detection module of the geographical detector is utilized to assess the explanatory power of each influencing factor on land use change within the Yellow River Basin. Additionally, the interactive detection module is employed to analyze how these factors interact and collectively influence land use change. Its expression is as follows [45]:

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

where $q$ represents the explanatory power of the factor on land use change, and its value range is 0 to 1. The larger the $q$ value, the stronger the explanatory power of the factor on the results of land use change; $L$ represents the number of types of impact factors; $h$ represents the type quantity; $N$ represents the number of samples in the study area; $N_h$ represents the number of samples of type $h$; and ${\sigma }_h^2$ and ${\sigma }^2$ represent the discrete variance of the type quantity $h$ and the variance of land use in the study area, respectively.

In this paper, two core factors from each of the three aspects—economy, climate, and topography—are selected for analysis using the geographic detector method. These factors include population and GDP (economic aspect), temperature and precipitation (climate aspect), and slope and DEM (topography aspect). These factors are processed to match the resolution of the land use data. Subsequently, 2000 random sampling points are selected, and both the land use data and corresponding factor data are extracted at these points. Since the independent variables in the geographic detector must be type quantities, it is necessary to discretize the six factors using the equal spacing method, quantile method, geometric interval method, etc. Different discrete methods and number of categories will have a significant impact on the results of the geographic detector model. Therefore, the optimal discrete method and number of categories in this paper are automatically calculated according to the “optidisc()” function provided in the GD package of R language [46]. This function automatically calculates the optimal discretization method and the appropriate number of categories based on the data, ensuring robust and objective results in the analysis.

2.3.5. FLUS Model

The FLUS model enhances the traditional cellular automata model by incorporating an adaptive inertia competition mechanism based on roulette wheel selection and using neural networks to obtain suitability probabilities [47]. This approach effectively addresses the competitive relationships among various land use types, improves simulation accuracy, and offers high computational efficiency and a wide simulation scope. The FLUS model primarily consists of the following three modules [48]:

(1) ANN-based Suitability Probability Estimation: Based on 6 driving factors and 2010 land use data, we used an artificial neural network (ANN) algorithm. The sampling pattern was set to random sampling of 10/1000, with the number of hidden layers in the neural network set to 12. This approach calculated the suitability probabilities of 9 different land use types for each pixel within the study area. These probabilities determine the potential distribution of each land use type across the region.

(2) Self-Adaptive Inertia and Competition-Mechanism-Based Cellular Automata: The conversion cost matrix indicates the possibility of conversion between different land use types. A value of 1 is set if conversion is possible, and 0 if not. In this study, all matrix values were set to 1 to allow mutual conversion between all land use types, reflecting the actual land use situation in the Yellow River Basin. The neighborhood weight factor ranges from 0 to 1, with values closer to 1 indicating a stronger expansion capability of the land use type. Based on previous research and multiple experiments [49,50], the neighborhood weight factors were adjusted and ultimately set as shown in Table 1.

(3) To validate the simulation accuracy, we used the 2010 land use data and driving factors to simulate the 2020 land use data, and then performed a Kappa test on the simulated 2020 data and the actual 2020 data to verify the model’s predictive performance. A Kappa coefficient greater than 0.6 indicates that the model has good overall performance and can be used to predict future land use data [51].

3. Results and Discussion

3.1. Trend and Mutation Test Analysis

The trend test and mutation test were performed at a given significance level of α = 0.05, and the results are shown in Table 2. The area of forest, grassland, water, and impervious surfaces showed a clear increasing trend. The area of snow and ice increased, but the trend was not obvious. The area of cropland, shrub, barren land, and wetlands exhibited a significant decrease. After conducting both the Mann–Kendall (M-K) and sliding t mutation tests, it was determined that 2002 and 2013 were the mutation point years. Consequently, the research period from 1990 to 2021 is divided into three distinct stages: 1990–2002, 2002–2013, and 2013–2021. This segmentation allows for a more detailed analysis of land use dynamics and changes in the Yellow River Basin over time.

3.2. Analysis of the Change in Land Use Quantity Structure

The analysis focused on key years in the Yellow River Basin: the initial study year of 1990, the final year of 2021, and the mutation years of 2002 and 2013. Figure 2 illustrates the distribution of land use types across these pivotal periods. The land use types in the Yellow River Basin are mainly grassland, cropland, and forest, accounting for around 57%, 25%, and 10%, respectively. Shrubs are mainly concentrated in the southwestern forest area and its periphery. Snow, ice, and wetlands are mainly located in the western part of the basin. Barren land is primarily located in the northern part of the upper reaches. Wetlands are mainly distributed in the southern part of the upper reaches. Impervious surfaces are prominent in the southeastern part of the basin.

The proportion of land use in each basin section is shown in Figure 3. The upper reaches are predominantly grassland, covering approximately 77% of the area. The middle reaches are marked by a mix of grassland (around 41%) and impervious surfaces (around 35%), and the lower reaches are dominated covered by cropland (around 73%) and impervious surfaces (around 18%). This shows that the upper reaches focus on ecological protection, the lower reaches emphasize agricultural production and economic activities, and the middle reaches are a transition area between the two.

The dynamics of land use in the Yellow River Basin from 1990 to 2021 are illustrated in Table 3. Over this period, comprehensive land use in different sub-basins showed an upward trend, indicating increasingly drastic land use changes. The middle reaches exhibit the highest dynamics, with an average rate of 1.24%, highlighting the most active land use changes in this area. For the entire Yellow River Basin, the dynamics of wetlands in 1990–2002, 2002–2013, and 2013–2021 were −6.8%, 7.5%, and 9.06%, respectively. This represents the largest growth rate, suggesting that the “National Wetland Protection Project Plan (2002–2030)” enacted in 2003 has significantly contributed to wetland restoration and protection. Between 2002 and 2013, cropland, forest, shrub, water, snow/ice, and barren land exhibited the highest dynamics, at −0.65%, 0.91%, −2.13%, 2.08%, 3.42%, and −2.01%, respectively. The dynamics of impervious surfaces showed a decreasing trend, with rates of 4.3%, 3.94%, and 2.10% over the three periods, indicating that impervious surfaces continued to expand from 1990 to 2021, but the rate of change slowed down. In the upper reaches of the Yellow River, the dynamics of grassland were low and the changes were not obvious. Although the area decreased slightly, the change rate was gentle. The dynamics of barren land were −0.81%, −1.20%, and 0.39%, respectively. Before 2013, the area of barren land decreased and was continuously developed and utilized at an accelerated speed, but it showed signs of recovery afterward. Since wetlands are primarily distributed in the upstream area, the dynamics of upstream wetlands are the same as those of the entire basin. The dynamics of impervious surfaces reached a peak value of 6.88% in 2002–2013. In the middle reaches of the Yellow River, the dynamics of forest were 0.40%, 0.96%, and 0.85%, respectively, with a rapid growth rate and obvious expansion. The dynamics of shrub were 0.57%, −5.16%, and −2.6%, respectively, showing a shrinking area. Grassland expanded before 2013 and then contracted. In the lower reaches of the Yellow River, the dynamics of cropland were −0.33%, −0.59%, and −0.56% over the three periods, indicating an accelerated reduction rate and large-scale occupation. The dynamics of impervious surfaces were 2.82%, 2.67%, and 2.14%, respectively, showing continuous expansion, albeit at a slower pace.

The land use data from two periods were spatially superimposed and rasterized in ArcGIS, and a Sankey diagram of land use transfer was created using Origin visualization, as shown in Figure 4. From the basin-wide perspective, cropland, grassland, barren land, and impervious surfaces changed most actively. From 1990 to 2002, the largest net decrease was in cropland, which decreased by 7474 km², followed by barren land, which decreased by 5967 km². The reduced cropland was mainly converted to grassland (83.17%), and the reduced barren land was primarily converted into grassland (92.97%). Grassland saw the largest net increase of 5224 km², mainly coming from cropland (67.45%) and barren land (26.48%). This change was related to the launch of the project of returning cropland to forest and grassland in 1999, converting large areas of sloping cropland unsuitable for cultivation into herbaceous plants to restore the grassland ecosystem. From 2002 to 2013, the largest net decrease was in barren land, with a reduction of 7846 km², mostly converted to grassland (94.14%). Grassland experienced the largest net increase of 8115 km², converted from cropland and barren land. From 2013 to 2021, the largest net decrease was in grassland (9262 km²), with the reduced grassland mainly transitioning to cropland (60.08%) and barren land (18.09%). This shift was due to the rapid population growth in China during this period, leading to a sharp increase in food demand and the expansion of cropland. The largest net increase during this period was in forest (5688 km²), which was transferred from grassland (72.9%) and cropland (20.40%). The rapid expansion of forest was linked to the project of returning cropland to forest and grassland, as well as the construction of the shelterbelt system in the middle reaches of the Yellow River starting in 1995 and subsequent forest ecological protection measures [33]. Additionally, cropland was increasingly occupied by impervious surfaces during all three periods, accounting for 12.37%, 15.01%, and 12.03%, respectively. This was due to rapid economic development, urbanization, and industrialization in the Yellow River Basin, which led to a significant increase in impervious surfaces and a corresponding decrease in cropland. In the upper reaches of the Yellow River from 1990 to 2021, land use changes were mainly conversions between cropland, grassland, and barren land. The area of wetlands decreased from 915.6 km² to 168.9 km², with 83.56% converted to grassland. From 1990 to 2002, the cropland area increased by 3102 km², primarily from grassland (93.54%), and barren land area decreased by 2967 km², mainly from grassland (91.62%). From 2002 to 2013, grassland area significantly increased (3490 km²), mainly from cropland (49.74%) and barren land (44.93%), while barren land continued to decrease (3638 km²), mainly transitioning into grassland (92.53%); the largest net decrease in grassland from 2013 to 2021 was 3101 km², mainly from cropland (47.50%) and barren land (39.32%). On the one hand, extreme weather events, such as high temperatures and heavy rains, occurred frequently during this period, accelerating soil moisture evaporation and causing soil erosion, which hindered grassland growth and recovery. On the other hand, overgrazing due to the development of animal husbandry weakened the grassland ecosystem, leading to grassland degradation [52]. The increase in cropland mainly came from grassland (93.67%). Wetland area increased slightly to approximately 223.6 km², but still decreased by 383.8 km² compared to 1990, indicating that despite recent ecological measures, wetlands still require continued attention. In the middle reaches of the Yellow River, land use changes mainly involved the conversion of cropland to grassland, grassland to forest, and barren land to grassland. From 2002 to 2013, the largest net increase was in forest area (7140 km²), and the largest net decrease was in barren land (4186 km²). The increase in forest was mainly from cropland (25.85%) and grassland (66.37%), while the decrease in barren land was mainly due to conversions to grassland (98.28%). In the lower reaches of the Yellow River, land use changes mainly involved the conversion of cropland to impervious surfaces. From 1990 to 2002, 2002 to 2013, and 2013 to 2021, cropland area decreased by 1036.48 km², 1615.23 km², and 1051.68 km², respectively. The proportion of reduced cropland converted to impervious surfaces was 75.77%, 76.99%, and 75.03%, respectively.

3.3. Analysis of Land Use Spatial Pattern Change

Three land use types that were sensitive to dynamic changes in typical years were analyzed separately in different watersheds. The results of the standard deviation ellipse and center of gravity migration trajectories are shown in Figure 5, with parameters detailed in Table 4. The standard deviation ellipse oblateness of shrubs in the Yellow River Basin continues to increase, indicating an enhanced significance of directional distribution. The center of gravity shifted to the southeast before 2002, and then moved to the west with a significant increase in rate, suggesting that the overall distribution of shrubs has shifted westward and spread out. For impervious surfaces, the center of gravity shifted southwest before 2002, and then northwest, with a decreasing migration rate and gradually decreasing ellipse oblateness. This indicates that the center of gravity for construction land moved westward and its distribution became more concentrated. China’s Western Development Strategy, launched in 2000, promoted economic development in the western region, urbanization, and the expansion of impervious surfaces in the Yellow River Basin [53], causing the center of gravity for impervious surfaces to shift westward and become more concentrated. The center of gravity for wetlands first moved northwest and then southeast, with the migration rate initially accelerating and then slowing down. The standard deviation ellipse oblateness changed significantly, from 0.74 in 1990 to 0.34 in 2021, indicating a shift from a scattered to a more concentrated wetland distribution. During this period, many wetlands were converted into grasslands due to decreased precipitation in the upper reaches [54], causing water levels to drop and wetlands to dry up. Additionally, the construction of reservoirs and water diversion projects in the upper reaches intercepted significant water sources, reducing water supply to the middle and lower reaches and converting wetlands into grasslands. The standard deviation ellipse oblateness for forest, water, and impervious surfaces in the upper reaches of the Yellow River, and forest, shrubs, and impervious surfaces in the middle reaches, remained relatively unchanged. The average oblateness was 0.573, 0.738, 0.688, 0.353, 0.455, and 0.315, respectively, indicating a directional and relatively stable spatial distribution. The center of gravity for forest and impervious surfaces in the upper reaches shifted south, while in the middle reaches, the center of gravity for forest and impervious surfaces shifted north, opposite to the upstream trends. The migration speed of the center of gravity for impervious surfaces first increased and then slowed. The center of gravity for water in the upper reaches shifted southwest before 2013, and then northeast with a sharp increase in migration rate. This change was influenced by the adjustment of the Xiaolangdi Reservoir and the implementation of the East Route Project of the South-to-North Water Diversion, affecting downstream water distribution [55,56]. The center of gravity for shrubs in the middle reaches shifted north overall, with the speed first increasing and then decreasing. Due to the narrow and elongated shape of the lower reaches of the Yellow River, the standard deviation ellipse oblateness for grassland, water, and impervious surfaces was relatively high, averaging 0.595, 0.848, and 0.803, respectively, showing a pronounced directional distribution. The center of gravity for grassland shifted southwest at a stable rate, while the center of gravity for water also shifted southwest. Before 2013, the center of gravity for impervious surfaces shifted northeast at a slower rate, then shifted southwest at an accelerated rate.

3.4. Analysis of Driving Forces of Land Use Change

In different periods, land use type was used as the dependent variable, while population, GDP, temperature, precipitation, slope, and DEM were used as independent variables to detect the impact of each driving factor on the change in land use type. The results are shown in Table 5. Except for the slope factor during 2013–2021, which did not pass the significance test, all other factors in the remaining periods had a significant p value of 0, indicating they passed the significance test and could be analyzed as influencing factors. From 1990 to 2002 and 2002 to 2013, the driving factors were ordered from largest to smallest q value as follows: population > GDP > DEM > temperature > precipitation > slope. From 2013 to 2021, the order changed to population > GDP > temperature > DEM > precipitation > slope. The explanatory power of temperature on land use exceeded that of DEM. Population had the strongest explanatory power on land use in all three periods, followed by GDP. This indicates that economic factors have a significant impact on land use in the Yellow River Basin, with the driving effect of utilization change being greater than that of topographic and climatic factors. However, over time, except for precipitation, which showed an increase in q value after 2013, the q values of other factors exhibited a decreasing trend. This suggests that the influence of economic, topographic, and climatic factors on land use change in the Yellow River Basin is gradually weakening. With economic development and structural adjustments, the economic development model in various regions of the Yellow River Basin has gradually shifted from high-speed growth to high-quality development, emphasizing sustainable development and environmental protection. Consequently, the influence of economic factors on land use has diminished. Since 1990, the state has implemented several major ecological projects in the Yellow River Basin, such as the Taihang Mountains Greening Project (1994), the Return of Farmland to Forests and Grasslands (1999), the Second Phase of the Three-North Shelterbelt (2002), and the Landscape, Forest, Farmland, Lake and Grass Restoration Project (2016) [33]. These projects have had an important impact on the land use structure in the Yellow River Basin. The driving role of policy factors in land use changes has gradually increased, thereby weakening the relative influence of economic, topographic, and climatic factors.

Interactive detection of the six driving factors was conducted, and the results are shown in Figure 6. Each driving factor significantly interacts with land use change, demonstrating either a double-factor enhancement or non-linear enhancement relationship. This indicates that land use in the Yellow River Basin is driven by the coordinated influence of multiple factors, with a close relationship between them. From 1990 to 2002, the interaction between precipitation and DEM had the strongest explanatory power, with a q value of 0.303, which is greater than the sum of q values for each factor individually, indicating a non-linear enhancement relationship. This was followed by the interaction between precipitation and population, with a q value of 0.3, showing a double-factor enhancement relationship. From 2002 to 2013, the interaction between precipitation and GDP had the strongest explanatory power, with a q value of 0.287, also indicating a non-linear enhancement relationship. From 2013 to 2021, the interaction between precipitation and population had the strongest explanatory power, with a q value of 0.269, showing a double-factor enhancement relationship. The single-factor explanatory power of slope was the weakest in all three periods. However, its explanatory power was significantly enhanced after interaction with other factors, especially with the population factor, followed by its interaction with GDP. Additionally, the explanatory power was stronger after the interaction of precipitation and population with other factors.

3.5. Prediction of Future Land Use Changes

This study used the FLUS model to predict land use changes in the Yellow River Basin for the year 2030. We used the 2010 land use data to simulate the 2020 land use situation; the model achieved a Kappa coefficient of 0.726. This indicates good predictive performance and high applicability for the study area, making it suitable for predicting future land use changes. The predicted land use for the Yellow River Basin in 2030 is shown in Figure 7. The changes in land use are illustrated in Figure 8. Compared to 2020, the land use types in the Yellow River Basin in 2030 will still be dominated by cropland, forest, and grassland. The main changes will occur in the middle and lower reaches, affecting cropland, forest, grassland, and impervious surfaces. The expansion trends of forest and impervious surfaces are significant, with increases of 6136 km² and 4891 km², respectively. Conversely, cropland and grassland will experience substantial decreases, losing 4723 km² and 6824 km², respectively. This suggests that in the coming years, due to human activities and reforestation efforts, there will be a rapid growth in impervious surfaces to meet socio-economic development needs, and an increase in forest area to support ecological protection and restoration in the Yellow River Basin. Consequently, the significant loss of cropland and grassland will pose a threat to food security. This issue requires the attention of relevant authorities, as future policies should not sacrifice cropland to achieve economic development and ecological protection goals.

4. Conclusions and Policy Recommendations

4.1. Conclusions

This study reveals the spatiotemporal characteristics of land use types and their driving factors in the Yellow River Basin, and predicts land use changes for the year 2030. This provides important insights for understanding the ecological environment changes and land resource management in the Yellow River Basin, contributing to the sustainable development of the region. The main conclusions of this study are as follows:

1. According to the year of mutation, the period from 1990 to 2021 is divided into three stages: 1990–2002, 2002–2013, and 2013–2021. The predominant land use types in the Yellow River Basin are grassland (around 57%), cropland (around 25%), and forest (around 10%). From 1990 to 2021, the areas of forest, grassland, water, and impervious surfaces increased significantly, while the areas of cropland, shrubs, barren land, and wetlands decreased significantly. The middle reaches of the Yellow River exhibit the highest comprehensive land use dynamics and the most active changes in land use types.

2. Cropland, grassland, barren land, and impervious surfaces in the Yellow River Basin are sensitive to changes. The upper reaches of the Yellow River are characterized by conversions between cropland, grassland, and barren land. In the middle reaches, conversions primarily involve cropland to grassland, grassland to forest, and barren land to grassland. In the lower reaches, cropland is primarily converted into impervious surfaces. Shrub and impervious surfaces show a westward shift in their centers of gravity, while wetland distribution changes from being dispersed to concentrated. The other land use types exhibit stable directional distributions.

3. Economic factors are the primary driving force for land use changes in the Yellow River Basin. However, with a shift towards high-quality sustainable development and the implementation of various ecological projects since 1990, policy factors have become dominant in influencing land use changes. Consequently, the impact of economic, climatic, and topographic factors has weakened. Land use change in the Yellow River Basin is driven by the coordination of multiple factors, with significant interaction between economic and climatic factors enhancing the driving force of land use change.

4. In 2030, the Yellow River Basin’s predominant land use types will still be grassland, cropland, and forest. However, forest land and impervious surfaces are expected to expand significantly, while cropland and grassland will see substantial reductions. It is crucial to focus on food production while balancing economic development and ecological protection in the Yellow River Basin.

4.2. Policy Recommendations

Based on the findings of this paper, we propose the following recommendations in order to provide reference for planning and policy making:

1. Promoting sustainable land use and ecological balance: In view of the reduction in cropland and grassland, we recommend the adoption of sustainable land use planning techniques to maximize land use efficiency and balance the needs of food security, economic development, and ecological protection. At the same time, ecological projects and policy interventions such as protection, ecological restoration, and afforestation should be strengthened to address land degradation and promote environmental sustainability.

2. Ensuring food security and high-quality development: In the face of the potential threat to food security from the increase in impervious surfaces, we recommend that the protection of cropland be a priority task and strategies be developed to mitigate the impact of industrialization and urbanization on agricultural resources. In addition, it is encouraged to shift from rapid economic expansion to a high-quality, sustainable development model to reduce the negative impact of economic activities on the environment.

3. Adapting to climate change and multi-factor coordination: In view of the impact of climate on land use change, we recommend strengthening environmental change adaptation strategies, such as implementing water resource management plans to support wetland protection and reduce the impact of extreme weather events. At the same time, multi-factor coordination in land use management should be promoted to improve the efficiency of policy implementation and land use planning.

4. Continuous monitoring and innovative technology application: In order to more accurately track land use changes and improve predictions, we recommend continuous monitoring using advanced remote sensing and modeling technologies. This will provide decision makers with decision support based on current trends and data, thereby improving the scientific and forward-looking nature of land use management.

Through the implementation of these comprehensive recommendations, we expect to provide more effective strategies for land use management in the Yellow River Basin and provide valuable references for planners, policy makers, scholars, and researchers in related fields.

Author Contributions

Conceptualization, methodology, software, validation, formal analysis, writing—original draft preparation, Y.C. (Yali Cheng). Supervision, funding acquisition, writing—review and editing, Y.C. (Yangbo Chen) All authors have read and agreed to the published version of the manuscript.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. Location of the Yellow River Basin.

Figure 2. Distribution map of land use types in the Yellow River Basin in typical years from 1990 to 2021.

Figure 3. Land use type proportion in each river basin in typical years.

Figure 4. Sankey diagram of land use change in the Yellow River Basin from 1990 to 2021.

Figure 5. Standard deviation ellipses and centroid migration trajectories of typical land use types in different river basin sections from 1990 to 2021.

Figure 5. Standard deviation ellipses and centroid migration trajectories of typical land use types in different river basin sections from 1990 to 2021.

Figure 6. Interactive detection results of land use change in the Yellow River Basin from 1990 to 2021.

Figure 7. (a) The actual land use types for the year 2020; (b) the predicted land use types for the year 2030.

Figure 8. The changes in land use types in the Yellow River Basin from 2020 to 2030.

Table 1

Neighborhood factor parameters of different land use types.

Land Use Type	Neighborhood Factor Parameters
Cropland	0.4
Forest	0.6
Shrub	0.2
Grassland	0.3
Water	0.2
Snow/ice	0.1
Barren	0.1
Impervious	0.8
Wetland	0.3

Table 2

Area change trend and mutation test results of various land use types in the Yellow River Basin from 1990 to 2021.

Land Use Type	M-K Trend Test		Mutation Point Test
	Statistics Z Value	Trend (α = 0.05)	M-K Test (α = 0.05)	Sliding t Test(α = 0.05)
Cropland	−7.119	significant decrease		1996, 2002, 2015
Forest	8.027	significant rise		2013, 2017
Shrub	−5.141	significant decrease		1996, 2002, 2005
Grassland	4.914	significant rise		2002, 2004, 2009, 2013
Water	5.530	significant rise		1996, 2008, 2013
Snow/ice	0.795	insignificant rise	2003	1993, 2005, 2013
Barren	−6.697	significant decrease		1998, 2002, 2011, 2013
Impervious	8.027	significant rise		2002, 2013
Wetland	−2.514	significant decrease	1995	2001, 2015

Table 3

Dynamics of land use in different stages of each river basin from 1990 to 2021.

Area	Period	K (%)									LC (%)
		Cropland	Forest	Shrub	Grassland	Water	Snow/Ice	Barren	Impervious	Wetland
Entire Yellow River	1990–2002	−0.29	0.46	−0.14	0.10	−0.40	−0.19	−1.20	4.30	−6.80	0.95
	2002–2013	−0.65	0.91	−2.13	0.16	2.08	3.42	−2.01	3.94	7.50	0.98
	2013–2021	−0.03	0.82	−0.84	−0.25	0.68	−3.68	0.09	2.10	9.06	1.04
Upper Yellow River	1990–2002	0.59	0.87	−0.43	−0.02	−0.15	−0.19	−0.81	4.23	−6.80	0.74
	2002–2013	−0.47	0.64	−0.72	0.10	1.96	3.42	−1.20	6.88	7.50	0.76
	2013–2021	0.28	0.61	−0.46	−0.13	0.51	−3.68	0.39	2.74	9.06	0.85
Middle Yellow River	1990–2002	−0.55	0.40	0.57	0.34	−0.97	- *	−2.29	5.47	−7.69	1.21
	2002–2013	−0.73	0.96	−5.16	0.28	2.41	27.03	−4.95	4.12	72.73	1.24
	2013–2021	−0.04	0.85	−2.60	−0.47	0.16	−11.90	−1.51	1.91	11.11	1.26
Lower Yellow River	1990–2002	−0.33	0.94	−6.59	−2.16	−0.39	-	−2.45	2.82	-	0.64
	2002–2013	−0.59	0.94	−5.69	−2.02	2.07	-	−1.03	2.67	-	0.74
	2013–2021	−0.56	0.29	48.31	−3.19	2.17	-	−9.74	2.14	-	0.81

* Indicates that there is no such land use type.

Table 4

Standard deviation ellipses and centroid migration parameters.

Area	Land Use Type	v * (km/a)			α *
		1990–2002	2002–2013	2013–2021	1990	2002	2013	2021
Entire Yellow River	Shrub	2.67	5.09	6.22	0.62	0.64	0.65	0.67
	Impervious	3.61	2.21	2.11	0.37	0.27	0.27	0.26
	Wetland	1.56	3.10	1.12	0.74	0.82	0.25	0.34
Upper Yellow River	Forest	0.65	0.31	0.56	0.57	0.58	0.58	0.56
	Water	1.29	2.84	7.47	0.74	0.73	0.74	0.74
	Impervious	0.12	0.57	0.39	0.70	0.69	0.68	0.68
Middle Yellow River	Forest	1.57	2.01	2.83	0.33	0.35	0.35	0.38
	Shrub	1.01	0.72	5.28	0.46	0.42	0.44	0.50
	Impervious	0.67	0.86	0.51	0.30	0.31	0.32	0.33
Lower Yellow River	Grassland	0.80	0.81	0.70	0.63	0.58	0.60	0.57
	Water	3.53	3.47	4.87	0.82	0.85	0.86	0.86
	Impervious	0.68	0.04	0.73	0.81	0.81	0.80	0.79

v * is the migration rate of the center of gravity, in kilometers per year; α * is the flattening of the ellipse.

Table 5

Detection results of land use change factors in the Yellow River Basin from 1990 to 2021.

Impact Factor	1990–2002		2002–2013		2013–2021
	q Statistic	p Value	q Statistic	p Value	q Statistic	p Value
population	0.2423	0	0.2034	0	0.1930	0
GDP	0.1680	0	0.1584	0	0.1448	0
temperature	0.1512	0	0.1317	0	0.1143	0
precipitation	0.1090	0	0.0899	0	0.0980	0
slope	0.0164	0	0.0135	0	0.0058	0.3816
DEM	0.1573	0	0.1336	0	0.1057	0

References

1. Yang, J.; Huang, X. The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019. Earth Syst. Sci. Data; 2021; 13, pp. 3907-3925. [DOI: https://dx.doi.org/10.5194/essd-13-3907-2021]

2. Xu, G.; Wang, Z.; Hu, Z.; Zhang, J. Evaluation of land use/cover change and land use grade in Qaidam basin, China. J. Hydroelectr. Eng.; 2019; 38, pp. 44-55.

3. Zhang, Z.; Zhu, J. Spatial-temporal changes of land use and hot zone diagnosis analysis in the past 25 years of the Great Xi’an area. J. Xi’an Univ. Technol.; 2020; 36, pp. 159-169.

4. Scientific Steering CommitteeInternational Project Office of LUCC Nunes, C.; Augé, J.I. Land-Use and Land-Cover Change (LUCC): Implementation Strategy. Environmental Policy Collection; IGBP Secretariat: Stockholm, Sweden, 1999.

5. Turner, B.L.I.; Skole, D.L.; Sanderson, S.; Fischer, G.; Fresco, L.; Leemans, R. Land-use and land-cover change. Science/Research plan. Glob. Change Rep.; 1995; 43, pp. 669-679.

6. Li, J.; Zhao, G. A Study of Driving Forces of Land Use and Land Cover Change. Sci. Technol. Manag. Land Resour.; 2004; 21.

7. Yang, H.; Jiang, J. Analysis of land use change and driving force in Jinta county. Nat. Resour. Inf.; 2023; pp. 1-8.

8. Hu, Y.; Zhang, Y. Spatial–temporal dynamics and driving factor analysis of urban ecological land in Zhuhai city, China. Sci. Rep.; 2020; 10, 16174. [DOI: https://dx.doi.org/10.1038/s41598-020-73167-0]

9. Gong, P.; Wang, J.; Yu, L.; Zhao, Y.C.; Zhao, Y.Y.; Liang, L.; Niu, Z.G.; Huang, X.M.; Fu, H.H.; Liu, S. et al. Finer resolution observation and monitoring of global land cover: First mapping results with Landsat TM and ETM+ data. Int. J. Remote Sens.; 2013; 34, pp. 2607-2654. [DOI: https://dx.doi.org/10.1080/01431161.2012.748992]

10. Chen, J.; Chen, J.; Liao, A.P.; Cao, X.; Chen, L.J.; Chen, X.H.; He, C.Y.; Han, G.; Peng, S.; Lu, M. et al. Global land cover mapping at 30 m resolution: A POK-based operational approach. ISPRS J. Photogramm. Remote Sens.; 2015; 103, pp. 7-27. [DOI: https://dx.doi.org/10.1016/j.isprsjprs.2014.09.002]

11. Sulla-Menashe, D.; Gray, J.M.; Abercrombie, S.P.; Friedl, M.A. Hierarchical mapping of annual global land cover 2001 to present: The MODIS Collection 6 Land Cover product. Remote Sens. Environ.; 2019; 222, pp. 183-194. [DOI: https://dx.doi.org/10.1016/j.rse.2018.12.013]

12. Defourny, P.; Bontemps, S. Land Cover CCI Product User Guide—Version 2.0; UCL-Geomatics: London, UK, 2017.

13. Wu, Q.; Zhang, X.; Jiang, J. Land-Use and Land-Cover Change (LUCC) Detection Based on MODIS and CLCD Data and Its Impact on the Simulation of Vegetation Gross Carbon Assimilation in the Loess Plateau. Geogr. Geo-Inf. Sci.; 2023; 39, pp. 30-38.

14. Shao, Z.; Li, N. Consistency Analysis and Accuracy Evaluation of Mutiple Land Cover Products: Taking the Huai River Basin as an Example. Mod. Agric. Sci. Technol.; 2024; 7, pp. 170-175.

15. Wu, Z.; Cai, Z. Accuracy evaluation and consistency analysis of multi-source remote sensing land cover data in the Yellow River Basin. Chin. J. Eco-Agric.; 2023; 31, pp. 917-927.

16. Liu, F.; Yang, G.; Han, X.; Jia, G.; Wang, N. Spatial-temporal Evolution of Land Use and Spatial Autocorrelation Analysis in Horqin Sandy Land--A Case Study of Naiman Banner. J. Northwest For. Univ.; 2020; 35, pp. 148-157.

17. Chen, X.; Zhu, M.; Gong, J. Analysis on changes and driving factors of land use in Dongguan City. J. Guangzhou Univ. (Nat. Sci. Ed.); 2023; 22, pp. 32-39.

18. Zhang, Z.; Kang, H. Spatiotemporal variability characteristics and driving forces of land use/land cover in Wuwei City from 2000 to 2020. Sci. Technol. Eng.; 2023; 23, pp. 8579-8587.

19. Liu, J.; Wu, B. Analysis on the Characteristics and Driving Forces of Land Use Change in Luoyang City. J. Anhui Agric. Sci.; 2023; 51, pp. 63–65+68.

20. Liao, M.; Fang, X. Spatiotemporal Characteristics of Land Use/Cover Changes in the Yellow River Basin Over the Past 40 Years. J. Soil Water Conserv.; 2024; 38, pp. 165–177+189.

21. Ding, J.; Pu, Z. Land Use Change in Huocheng County Based on Geomatics Information Mapping. J. Shenyang Univ. (Nat. Sci.); 2023; 35, pp. 380-387.

22. Wang, X.; Liu, P.; Geng, M.; Wang, M.; Lu, J. The key driving factors of land use patio-temporal change and its influence trendA case study of Wuwei. China Environ. Sci.; 2023; 43, pp. 6583-6591.

23. Xia, L.; Liu, J.; Wei, F.; Xu, Z.; Long, X.; Zhang, Y. Spatiotemporal pattern change of cultivated land in Weibei Dryland of Shaanxi Province. Trans. Chin. Soc. Agric. Eng.; 2021; 37, pp. 256-264.

24. Ren, H.; Ren, Y. Evolution of land use landscape pattern and influencing factors in rapid urbanization areas: A case study of Suzhou city. J. Northeast Norm. Univ. (Nat. Sci. Ed.); 2024; 56, pp. 143-153.

25. Wang, H.; Zhao, R. Changes of Wetland Landscape Pattern and Its Driving Forces in the Central Urban Area of Chongqing from 2010 to 2020. Ecol. Environ. Monit. Three Gorges; 2024; pp. 1-17.

26. Cuideng, X.; Chen, Y. Land Use Change and Its Driving Mechanism in Dongjiang River Basin from 1990 to 2018. Bull. Soil Water Conserv.; 2020; 40, pp. 236–242+258+331.

27. Hu, W.; Wei, X. Analysis of Land Use Change and Driving Forces in Dazhou City Based on GIS. China For. Econ.; 2022; 4, pp. 34-37.

28. Sun, S.; Zhang, X.; Zhang, Z. Spatiotemporal Changes and Driving Forces of Landuse in Shaanxi Province During 2000–2018. Bull. Soil Water Conserv.; 2021; 41, pp. 339-349.

29. Li, M.; Han, Y.; Zhao, H.; Wang, Y. Analysis on Spatial-temporal Variation Characteristics and Driving Factors of Fractional Vegetation Cover in Ningxia Based on Geographical Detector. Ecol. Environ. Sci.; 2022; 31, pp. 1317-1325.

30. Ge, J.; Ge, J. Study on spatiotemporal evolution and driving mechanism of land use in Zhengzhou city based on geographical detectors. Stand. Surv. Mapp.; 2023; 39, pp. 72-77.

31. Zhang, R.; Liu, W.; Song, Z. Spatio-temporal evolution and driving factors of national development zones of China based on geodetectors. J. Nat. Resour.; 2021; 36, pp. 2672-2683. [DOI: https://dx.doi.org/10.31497/zrzyxb.20211015]

32. Chen, L.; Cai, H.; Zhang, T.; Zhang, X.; Zeng, H. Land use multi-scenario simulation analysis of Rao River Basin based on Markov-FLUS model. Acta Ecol. Sin.; 2022; 42, pp. 3947-3958.

33. Tian, M.; Feng, C.; Wang, S.; Tian, Y.; Niu, Q. Ecological restoration of the Yellow River basin in the last 70 years and systematic restoration thinking. J. Environ. Eng. Technol.; 2023; 13, pp. 1787-1797.

34. Lou, J.; Dang, X. Land Use Change and Driving Force Analysis of the Ten Tributaries Basin in the Yellow River Basin from 1986 to 2020. J. Soil Water Conserv.; 2024; 38, pp. 319–327+336.

35. Xiao, D.; Yan, H. Spatiotemperal evolution of land use pattern in the Yellow River Basin (Henan section) from 1990 to 2018. Trans. Chin. Soc. Agric. Eng.; 2020; 36, pp. 271–281+326.

36. Yang, Y.; Guo, J. Analysis of Land Use Changes and Driving Forces in the Riparian Zone of Ning Xia-Inner Mongolia Section of the Yellow River. Yellow River; 2023; 45, pp. 81-86.

37. Liang, K.; Gong, L. Regional differentiation and future patterns of land use changes in the water conservation zone of the Yellow River. Adv. Water Sci.; 2024; 35, pp. 325-337.

38. Chen, Z.; Xu, Q. Analysis of Precipitation Characteristics in Jinhua by Mann-Kendall Test Method. Bull. Sci. Technol.; 2016; 32, pp. 47-50.

39. Wu, Y.; Guo, L. An Analysis of Multi-time Scales Characteristics of Runoff in Mainstream of Mekong River. China Rural Water Hydropower; 2021; 10, pp. 42-49.

40. Fu, L.; Chi, Y. Characteristics and Driving Forces of Land Use Change in the Yellow Basin from 2000 to 2020. Ecol. Environ. Sci.; 2022; 31, pp. 1927-1938.

41. Mao, H.; Jia, K. Spatiotemporal pattern of land use change in Yinchuan Plain from 1980 to 2018. Sci. Technol. Eng.; 2020; 20, pp. 8008-8018.

42. Wang, J.; Liu, X. Recognizing and measuring spatial conflict of land use via “risk-effect”. Trans. Chin. Soc. Agric. Eng.; 2022; 38, pp. 291-300.

43. Li, M.; Qin, X.; Zhen, F. Spatial feature of returning people flows in the Spring Festival and its impact on urban consumption economy: An empirical study of 93 major cities in China. World Reg. Stud.; 2024; pp. 1-12.

44. Li, S.; Wang, T.; Yan, C. Spatial-temporal evolution of human activity intensity in the Mu Us Sandy Land from 1965 to 2015. J. Lanzhou Univ. Nat. Sci.; 2021; 57, pp. 330-337.

45. Wang, J.; Xu, C. Geodetector: Principle and prospective. Acta Geogr. Sin.; 2017; 72, pp. 116-134.

46. Li, C.; Wu, Y.; Gao, B.; Wu, Y.; Zhen, K.; Li, C. Spatial Differentiation and Driving Factors of Rural Settlement in Plateau Lake:A Case Study of the Area Around the Erhai. Econ. Geogr.; 2022; 42, pp. 220-229.

47. Fu, Y.; Xie, W.; Liu, J.; Wei, F.; Guo, Y.; Su, F.; Yang, Y. Spatio-temporal Pattern Evolution and Prediction of Ecological Well-being in Jiangsu Province Based on GeoSOS-FLUS Model. Resour. Environ. Yangtze Basin; 2023; 32, pp. 71-82.

48. Yang, P.; Tian, J.; Gao, Y. Characteristics of spatio-temporal changes and future trends forecast of vegetation cover in the Yellow River Basin from 1990 to 2022. Acta Ecol. Sin.; 2024; 44, pp. 1-12.

49. Liu, T.; Zhang, X.; Lin, C. Functional Analysis of Water Conservation in Zhungeer Banner Based on InVEST and FLUS Models. Acta Agrestia Sin.; 2023; 31, pp. 3831-3840.

50. Ren, Y.; Liu, X.; Xu, X.; Sun, S.; Zhao, L.; Liang, X.; Zeng, L. Multi-scenario simulation of land use change and its impact on ecosystem services in Beijing-Tianjin-Hebei region based on the FLUS-InVEST Model. Acta Ecol. Sin.; 2023; 43, pp. 4473-4487.

51. Li, X.; Li, D.; Liu, X. Geographical Simulation and Optimization System (GeoSOS) and Its Application in the Analysis of Geographic National Conditions. Acta Geod. Cartogr. Sin.; 2017; 46, pp. 1598-1608.

52. Guo, J.; Zhang, Q.; Song, M.; Shi, Y.; Zhou, B.; Wang, W.; Li, Y.; Zhao, X.; Zhou, H. Status and Function Improvement Technology of the Grassland Ecosystem in the Upper Yellow River Basin. Acta Agrestia Sin.; 2020; 28, pp. 1173-1184.

53. Jiyuan, L.I.U.; Zengxiang, Z.; Xinliang, X.U.; Wenhui, K.; Wancun, Z.; Shuwen, Z.; Rendong, L.I.; Changzhen, Y.A.N.; Dongsheng, Y.U.; Shixin, W.U. et al. Spatial Patterns and Driving Forces of Land Use Change in China in the Early 21st Century. Acta Geogr. Sin.; 2009; 64, pp. 1411-1420.

54. Huang, W.; Gao, Y.; Xi, C.; Wu, C.; Huang, C. Correlation Analysis of Wetland Evolution and Measured Runoff in the Lower Yellow River. J. Hydroecol.; 2022; 43, pp. 1-7.

55. Liang, X. Temporal and Spatial Variations and Driving Forces of Wetlands in the Yellow River Basin Based on GEE. Master’s Thesis; Henan University: Kaifeng, China, 2023.

56. Wang, W.; Zhao, Z. Analysis on Ecological Water Conservancy Construction of the Yellow River Basin in the New Era. China Resour. Compr. Util.; 2022; 40, pp. 153-155.