Next Article in Journal
Communal Land Titling and New Geographies of Development in Northern Thailand
Previous Article in Journal
Spatial Evolution of Grassland Ecological Carrying Capacity and Low-Carbon Development Pathways for Animal Husbandry in Inner Mongolia
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 14
Issue 5
10.3390/land14051093
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Evolution of Cropland Slope Structure and Its Implications for Fragmentation and Soil Erosion in China

by
Guangjie Liu
1,2,
Yi Xia
1,2 and
Li Bao
1,2,*
1
College of Resources and Environment, Yunnan Agricultural University, Kunming 650201, China
2
Yunnan Soil Fertilization and Pollution Remediation Engineering Research Center, Kunming 650201, China
*
Author to whom correspondence should be addressed.
Land 2025, 14(5), 1093; https://doi.org/10.3390/land14051093
Submission received: 15 April 2025 / Revised: 10 May 2025 / Accepted: 15 May 2025 / Published: 17 May 2025
Browse Figures
Versions Notes

Abstract

Cropland slope structure is a key factor influencing agricultural sustainability and ecological risk, especially in topographically complex regions. This study proposes a novel framework that integrates slope spectrum analysis with H3 hexagonal grid partitioning to examine the spatiotemporal dynamics of cropland slope across China from 1990 to 2023. Using 30 m CLCD land cover data, we derived key indicators, including the T-value, upper slope limit (ULS), peak area proportion (PaP), slope at maximum area (SMA), and cropland slope change index (CSCI). This grid-based, multi-indicator approach enables the fine-scale detection of slope structure transitions. Results show that the average slope of cropland fluctuated at around 4.12°, peaking at 4.18° in 2003, while the ULS remained stable at 17°, with 95% of cropland below this threshold. Regionally, cropland in southwest and northwest China was concentrated on steeper slopes (ULS > 26°, PaP < 10%), whereas flatter areas in north and south China had cropland mainly below 15°. From 1990 to 2023, upslope expansion was evident in south China (CSCI > 10), while downslope shifts aligned with high-slope cropland in the western regions. Geographically weighted regression revealed significant positive correlations between increasing ULS and CSCI and elevated cropland fragmentation and soil erosion in hilly areas. These findings highlight the ecological risks of cropland expansion into steep terrain. The proposed framework offers a spatially explicit perspective of cropland slope evolution and supports targeted strategies for land management and ecological restoration.

1. Introduction

Cropland resources form the foundation of food security and ecological security within a national land space. Their sustainable utilization is a key factor in harmonizing human–land relationships and ensuring high-quality development [1,2]. The global population is projected to continue growing over the next 50 to 60 years [3], leading to higher demands for food production, which leads to further encroachments upon land previously used for agriculture [4,5]. At the same time, there exists a fundamental tension between agricultural production and ecological conservation [6,7,8]. The United Nations’ 2030 Sustainable Development Goals (SDGs) call for a balance between increasing agricultural output and maintaining ecosystem services [9]. Under constraints such as a limited total cropland area, urban development pressure, and ecological space protection, the spatial configuration of cropland is undergoing significant restructuring [10,11,12,13].
Urban expansion, agricultural restructuring, and ecological programs have altered the slope characteristics of China’s cropland, even as its total area remains relatively stable [14,15,16,17]. Large areas of low-slope, high-quality cropland have been lost to construction or reforestation [18,19,20,21,22,23], while new cropland is often reclaimed in hilly or mountainous regions with steeper slopes [24,25]. This has led to an overall increase in cropland slope, with implications for cultivation conditions, soil conservation, and sustainability [26,27,28,29]. Understanding these changes is essential to assess land use risks and improve quality-oriented cropland management.
Previous research has mainly focused on the spatiotemporal patterns of cropland [10,30,31], changes in land quality [32,33], and the mechanisms of cropland loss [34,35]. Terrain-based analyses have largely emphasized variations in soil physicochemical properties [36,37], while systematic assessments of slope, a key topographic attribute, remain limited. Although some studies have noted the shift in new cropland towards sloped terrain [24,38], detailed quantitative evaluations of slope variation trends within analysis units, spatial heterogeneity, and regional differentiation mechanisms are still lacking.
The slope spectrum is an effective tool for capturing terrain undulation through illustrating the frequency distribution of slope values across a study area [39], offering a more precise understanding of slope structure compared to conventional slope classification. While widely applied in urban studies to assess uphill expansion and its implications [40,41,42,43,44], its use in cropland research remains limited. Existing studies on cropland slope mostly focus on national or watershed scales, and often rely on administrative units [16,24,45], which fail to reflect intra-unit slope variability and spatial heterogeneity. Furthermore, how changes in slope structure drive cropland fragmentation and soil erosion, and whether such effects vary regionally, remains insufficiently explored.
China, with a population of 1.4 billion, has only 0.36 hectares of arable land per capita—just 60% of the global average [46]—and over one-third of this land is sloping cropland [29,47]. To safeguard food security, strict policies such as requisition–compensation balance and rural–urban land conversion have been enforced to maintain the 120 million ha redline. As a key parameter in national land surveys, cropland slope is crucial; however, systematic, fine-scale analyses are lacking. Topography strongly influences both soil erosion and land fragmentation [48]. In China, areas where soil loss exceeds 25 t ha−1 yr−1 are primarily found on sloping farmland [49], and severe fragmentation occurs in mountainous and rapidly urbanizing regions. However, most studies use static indicators (e.g., mean slope), limiting their ability to reflect spatiotemporal slope dynamics.
Given China’s complex terrain and diverse socioeconomic conditions, cropland slope dynamics exhibit strong spatial heterogeneity. Traditional administrative-unit-based analyses are inadequate for detecting such variability. Therefore, refined spatial frameworks are urgently needed to more accurately reveal slope evolution and its environmental consequences more accurately.
To address these research gaps, this study integrates slope spectrum analysis with the high-resolution H3 grid system to quantify cropland slope structure through a set of novel indicators: the spectrum intersection (T-value), peak area proportion (PaP), upper limit of slope (ULS), and cropland slope change index (CSCI). Rather than focusing solely on mean slope, slope structure refers to the composition, proportion, and spatial organization of the slope values within each analysis unit, capturing the microtopographic complexity and diversity of cropland. Using this framework, we conduct a fine-scale, national assessment of the evolution of cropland slope structure in China from 1990 to 2023, identifying dominant patterns and regional differentiation. In addition, we apply a geographically weighted regression (GWR) model to explore the heterogeneous impacts of slope structure on cropland fragmentation and soil erosion. This study not only provides empirical evidence for understanding cropland use risks and supporting precision management strategies, but also contributes a novel methodological approach for terrain analysis in land use change research. Furthermore, it enhances the theoretical insights into the interactions between physical geography and human activity, offering valuable support for sustainable land management and regional development.

2. Materials and Methods

2.1. Research Area

China spans a vast territory with significant topographic variation and diverse landforms, ranging in elevation from −154 m to 8848 m, encompassing plains, hills, mountains, and plateaus. Influenced by terrain, water-heat conditions, and population distribution, the spatial distribution of cropland is highly uneven (Figure 1; the abbreviations of Chinese provinces are shown in Appendix A. All maps in this article are based on the approval number GS (2024) 0650 for China’s map shp file, with data sourced from the National Geoinformation Public Service Platform). According to the national land survey data [50], China has a total of 127.86 million hectares of cropland. Of this, 61.93% is located on land with slopes of 2° or less, 15.32% on land with slopes between 2° and 6°, 13.40% on land with slopes between 6° and 15°, 6.04% on land with slopes between 15° and 25°, and 3.31% on land with slopes greater than 25°. As urbanization accelerates, cropland resources in the plains are under increasing pressure, prompting the gradual expansion of cropland into higher-slope areas. This has led to profound changes in land use patterns, fragmentation, and ecological risks. The diverse topographic gradients and significant regional differences make China a representative region for analyzing the evolution of slope structure and conducting spatial zoning studies.

2.2. Data Sources and Preprocessing

2.2.1. Cropland Dataset

This study employs the China Land Cover Dataset (CLCD) [51], which provides consistent, high-temporal-resolution land use data across China from 1985 to 2023 [52,53]. The CLCD has been widely used in various research fields, including land use change detection, ecological and environmental monitoring, and sustainable development assessments [54,55,56]. For this study, CLCD data from 1990 to 2023 were selected and analyzed at the original spatial resolution of 30 m to ensure comparability and analytical robustness. Based on the land use classification system provided by the CLCD, cropland classes were extracted and used to delineate the spatial distribution of cultivated land within the study area.

2.2.2. Digital Elevation Model and Slope Calculation

The NASADEM dataset, developed by NASA, was employed as the elevation data source for extracting and analyzing slope information. With a spatial resolution of approximately 30 m, it provides significant improvements in the accuracy, consistency, and completeness of terrain representation, particularly in areas previously affected by SRTM voids [57]. Slope values were computed on the Google Earth Engine platform and then converted to integer values via truncation at 1° intervals to facilitate the construction of slope spectra. All datasets were aligned to the CLCD coordinate system (Albers_Conic_Equal_Area) for consistent spatial analysis.

2.2.3. Soil Erosion Dataset

The 30 m annual soil water erosion dataset in the Chinese mainland from 1990 to 2022 [58,59] provides high-resolution hydraulic erosion estimates (unit: t/(hm2·a)) for the Chinese mainland from 1990 to 2022. The dataset was developed using the Google Earth Engine platform and is based on the Revised Universal Soil Loss Equation (RUSLE). Since the soil water erosion datasets of each province are stored separately, GDAL (version 3.10.2) [60] was used to mosaic the provincial datasets and unify their coordinate systems to generate a complete, spatially consistent dataset for the study area. Using the H3 hexagonal grids as spatial units, zonal statistics were then applied to aggregate the soil erosion volumes within each grid cell.

2.3. Research Methods

2.3.1. Distribution Characteristics of Cropland Slope

This study employs slope spectrum analysis to examine the spatial distribution characteristics of cropland within different slope intervals. The slope spectrum is a smoothed histogram used to illustrate the distribution of slope within a spatial region [39]. Specifically, a defined slope interval (1° used in this study) is used to calculate the proportion of cropland area within each slope interval relative to the total cropland area. This results in a curve representing the proportion of cropland area across various slope intervals, which can also be interpreted as the distribution frequency of cropland within each slope interval. The x-axis of the frequency curve represents the slope, and the y-axis indicates the proportion of the area within the corresponding slope interval (Figure 2). Using this method, both the cropland slope spectrum and the background terrain slope spectrum (which reflects the proportion of land area within each slope interval relative to the total land area and serves as the terrain slope’s background information) are calculated [16].
To quantify and describe the morphological characteristics of the cropland slope spectrum, this study defines the concept of cropland slope structure. This reflects the structural features of cropland distribution across slope gradients and is characterized by a set of key indicators, as follows: (1) T-value: This refers to the slope value at which the cropland slope spectrum intersects with the regional background slope spectrum. It represents the critical threshold where the distribution trend of cropland shifts from favoring gentle slopes to avoiding steeper ones. (2) ULS: Defined as the slope class at which the cumulative cropland area reaches 95% of the total, ULS reflects the degree of cropland adaptation to slope gradients, and its temporal trend helps reveal the potential and limits of cropland expansion into steeper terrains. (3) PaP: The percentage of cropland area corresponding to the peak value of the slope spectrum. (4) SMA: Slope at Maximum Area (SMA) refers to the slope class at which the cropland slope spectrum reaches its peak value, indicating the slope most extensively occupied by cropland. (5) PaT: Proportion above T-value represents the proportion of cropland area located on slopes steeper than the T-value, reflecting the extent of the high-slope cultivation pressure.
The concept of cropland slope structure serves to reveal how cropland is distributed across different slope gradients, offering an integrated perspective on the spatial adaptation and expansion of agriculture in response to terrain conditions. Rather than focusing solely on cropland quantity or average slope, this structure captures the full shape of the slope’s distribution curve, highlighting both concentration patterns and marginal expansion tendencies.

2.3.2. Cropland Slope Change Index

Slope spectrum analysis has been widely applied in urban studies, particularly in assessing uphill expansion, as urban areas typically do not exhibit downslope development. Consequently, slope-based indicators derived from multi-temporal spectrum analyses are often termed climbing indices, such as the Built-up Land Climbing Index [40] and the Cropland Climbing Index (CCI) [45]. However, the term ‘climbing index’ may overemphasize the notion of upward expansion, potentially neglecting downslope trends—such as the aggregation of cropland toward lower slope intervals within the spectrum. To more accurately capture the bidirectional nature of cropland slope dynamics, we adopted the term Cropland Slope Change Index (CSCI), which better reflects the full range of slope structure evolution.
The CSCI is calculated as the difference between PaT at two different time points. The formula for PaT at time j is given by the following:
P a T j = A j ( > T ) A j × 100 %
where A j ( > T ) is the area of cropland at time j located on slopes steeper than the threshold T .   A j is the total cropland area at time j .
Subsequently, the CSCI is defined as the difference between the PaT values at time j and time i , as follows:
C S C I = P a T j − P a T i
In this formula, P a T j is the proportion of cropland area on slopes steeper than T at time j . P a T i is the proportion of cropland area on slopes steeper than T at time i . A positive CSCI value (CSCI > 0) indicates an increased proportion of cropland in steeper slope areas, reflecting an upslope expansion trend. Conversely, a negative CSCI value (CSCI < 0) suggests a shift in cropland toward gentler slopes, reflecting a downslope contraction pattern.

2.3.3. Geographically Weighted Regression

To explore the spatially varying relationships between cropland slope structure changes and their ecological impacts, this study employs Geographically Weighted Regression (GWR) [61]. GWR is a local regression technique that extends traditional regression models by allowing for coefficients to vary across geographic space, thus capturing the spatial heterogeneity in the relationships among variables. The general form of the GWR model is as follows:
y i = β 0 ( u i , v i ) + ∑ k = 1 p   β k ( u i , v i ) ⋅ x i k + ε i ,
where y i is the dependent variable at location i , x i k represents the k -th independent variable, ( u i , v i ) are the coordinates of location i , β k ( u i , v i ) is the spatially varying coefficient for the k -th variable, and ε i is the error term. GWR estimates parameters using spatially weighted least squares, with weights being assigned based on the geographic distance between observations.

2.3.4. Grid-Based Spatial Partitioning with H3

To more precisely capture the spatial variation in cropland slope characteristics and to overcome the inconsistencies caused by administrative boundaries, this study adopted the H3 hexagonal hierarchical spatial indexing system for the spatial partitioning of the study area. H3 is a global, multi-resolution grid system based on regular hexagons, offering advantages such as spatial consistency, a high indexing efficiency, and support for multi-scale analysis [62]. Its uniform grid structure helps avoid the analytical bias introduced by map projections or irregular region shapes, making it particularly suitable for large-scale and cross-regional slope analyses. The H3 grid at resolution level 5 is employed, where each hexagonal cell covers an average area of approximately 252.90 km2. A total of 42,157 cells are generated to cover the entire study area. This spatial scale is sufficient to ensure the smoothness and continuity of the slope spectrum while avoiding local fluctuations caused by overly fine grids. It also balances computational efficiency and data redundancy, ensuring the stability and reliability of the results. Therefore, higher-resolution grids were not considered.
The implementation steps are as follows: first, a complete coverage of level-5 hexagonal cells were generated for the study area using the H3 tool. Then, cropland data and slope data were overlaid onto each grid cell to calculate indicators such as total cropland area, average slope, and the proportion of cropland across different slope classes within each cell. Finally, based on the grid-level statistics, the spatial patterns and temporal trends in the cropland slope distribution were analyzed.

3. Results

3.1. Nationwide Cropland Slope Structure Variation Characteristics

From 1990 to 2023, the average slope of cropland in China exhibited a fluctuating trend at around 4.12° (Figure 3). A notable increasing phase occurred between 1995 and 2003, peaking at 4.18°, followed by a gradual decline with fluctuations. In terms of slope distribution, due to the large total area of cropland in China, the ULS remained stable at 17° throughout the 30-year period, indicating that 95% of cropland was consistently located within areas below 17° in slope. Although this threshold continues to dominate at the national scale, changes within specific slope intervals can still be observed, reflecting nuanced shifts in cropland distribution across different terrain gradients. Such internal adjustments within a stable upper limit suggest a structural reshaping process, in which cropland becomes more dispersed and stretched toward higher slopes under land use pressure.
Figure 4 illustrates the slope spectrum distributions of terrain background and cropland from 1990 to 2023 at five-year intervals. The cropland slope spectrum curves generally exhibit an increase-then-decrease trend, with a distinct peak observed in the [1°, 2°) interval. The intersection point between the cropland and terrain background spectra occurs around a slope of 6°. Although the turning point (T value) shifts only slightly over time, a consistent rightward movement is evident, indicating a gradual expansion of cropland into steeper areas. In regions with slopes of less than 6°, the proportion of cropland decreased slightly, from 76.1% in 1990 to 75.66% in 2023. This subtle downward trend is also visible in the magnified view of the spectrum peak, reflecting a marginal reduction in cropland concentration in low-slope zones over time.
To further clarify the spatial changes in cropland across slope gradients, we generated heatmaps of the absolute cropland area change and relative change (the proportion of change relative to total land area in each slope interval) within the [0°, 20°) range. As shown in Figure 5, significant cropland retreat occurred between 1990 and 2010 within the 0° to 4° slope range. In contrast, during the more recent period (2010–2023), cropland expansion was observed in the 6° to 13° slope range, although the extent of this expansion is considerably smaller than the retreat observed in low-slope areas. This reflects a stage-wise transformation: early land consolidation and ecological restoration programs may have reduced low-slope farmland, while later rural revitalization and land development in western China drove moderate slope utilization.

3.2. Slope Spectrum Changes in Cropland from the Perspective of the H3 Grid

3.2.1. Analysis of Slope Spectrum Characteristics

By analyzing the slope spectra of all cropland grids, a more accurate understanding of the changes in the slope characteristics of cropland at the grid scale can be obtained. In order to better describe the state of slope distribution, the SMA (Figure 6) and PaP (Figure 7) can reflect the peak distribution characteristics in the slope spectrum. From the figure, it can be seen that cultivated land in China is mainly concentrated in the low-slope zone of 0–6°, and is affected by the topography, with the peaks in the slope spectra in the regions of southwest China (YN, GZ, SC, CQ), the central and western provinces (SX, SN, NX), and part of the southwestern provinces (the central part of FJ and the southern part of ZJ) being significantly shifted upward, and distributed in the range of 7° to 25°. The peak difference map further reveals the direction of change in the dominant values of the slope of the cultivated land, in which the peaks shifted upward by 3–8°, concentrated in the Yunnan–Guizhou Plateau, are the most significant. This upward shift is consistent with the region’s complex mountainous terrain and reflects a response to limited low-slope land resources.
From the spatial distribution of the PaP values in 1990 and 2023, most areas in the southwestern, southeastern, and central provinces of China exhibit generally low PaP values (<20%), indicating that cropland is more dispersed across different slope classes and that the proportion of cropland concentrated at the peak of the slope spectrum is relatively low. Such dispersion patterns often occur in fragmented mountainous landscapes where agricultural land is scattered and constrained by topographic heterogeneity. In contrast, eastern provinces (HE, SD, HA, AH, and JS), as well as northwestern (NM and XJ) and northeastern (HL) regions, show a more pronounced concentration pattern, with a large proportion of cropland (20–70%) being distributed around the dominant slope. This suggests a more stable cropland slope structure, a feature consistently observed in both time periods.
From the spatial distribution of T values in 1990 and 2023 (Figure 8), T values are generally low (<6°) in regions such as the Northeast Plain, North China Plain, and the middle and lower reaches of the Yangtze River Plain, indicating that cropland is primarily concentrated in low-slope areas with minimal cultivation on sloped land. In contrast, higher T values, with some exceeding 25°, are observed in areas such as the Yunnan–Guizhou Plateau, the margins of the Sichuan Basin, and the southern hilly regions, reflecting a greater concentration of cropland in high-slope areas and indicating some degree of pressure from slope-based agriculture. The T value change map (ΔT) reveals the evolving trends in cropland slope structure between 1990 and 2023. In most regions, the change in T values is relatively small, indicating a stable cropland slope structure. However, in southwestern regions and the Zhejiang–Fujian hilly areas, T values significantly increased (>3°), suggesting that cropland is expanding into steeper areas, likely due to growing pressure on cropland resources, the limited flat land available for development, and the expansion of slope-based agriculture. These findings imply an adaptive adjustment to terrain constraints, particularly in ecologically fragile zones where marginal land is increasingly utilized.
The ULS of cropland represents the maximum slope at which cropland is distributed within a given area, serving as a key indicator of the extent to which cropland has expanded into steeper terrains. By calculating the ULS within each grid cell, we generated spatial distribution maps of ULS for the years 1990 to 2023, along with a corresponding map depicting the temporal changes in ULS over this period (Figure 9).

3.2.2. Analysis of Cropland Slope Changes

Based on Formula 1, we calculated the CSCI at 10-year intervals (1990–2000, 2000–2010, and 2010–2020). Additionally, to reflect long-term changes in the study area, we also computed the index from 1990 to 2023. The results are shown in Figure 10.
From a spatial perspective, an upward trend in the CSCI is observed in parts of the eastern coastal regions (e.g., FJ and ZJ) and the Central China plains (e.g., JX, HN, and HB), particularly in the Yangtze River Delta, the Pearl River Delta, and the Beijing–Tianjin–Hebei (Jing-Jin-Ji) region. This trend was especially pronounced during the 1990–2000 period and is likely associated with urban expansion and the occupation of low-slope cropland for construction, which pushed agricultural activities toward more topographically complex peripheral areas, thereby increasing the average slope level of cropland. This shift also reflects the spatial mismatch between urban growth and agricultural protection, where prime farmland is often sacrificed first. In contrast, the southwestern region—especially Yunnan, Guizhou, and the periphery of the Sichuan Basin—showed an overall decreasing trend in CSCI from 1990 to 2023, with a particularly significant decline after 2000. Notably, some areas of the Loess Plateau, such as southern Shaanxi and western Henan, exhibited a phased increase in CSCI. Meanwhile, large parts of the Qinghai–Tibet Plateau, areas north and south of the Tianshan Mountains, and the arid regions of northwest China experienced limited changes in CSCI, suggesting relatively stable cropland spatial patterns—possibly due to the inherently low proportion of cropland in these regions. Additionally, strict ecological zoning and the marginal economic returns further limit slope agriculture expansion in such highland and arid zones.
In summary, the changes in the CSCI not only reflect the dynamics of slope conditions but also reveal, to some extent, the profound impacts of regional agricultural development strategies, urbanization processes, and ecological governance policies on cropland spatial patterns.

3.2.3. Analysis of Slope Structure Patterns

This study identifies the typical patterns of cropland slope structure using five key indicators: T value, PaT, SMA, PaP, and ULS. A K-means clustering method was employed to classify the slope structure types. Prior to clustering, all variables were subjected to missing value removal and Z-score normalization to ensure consistency and comparability across different indicators.
The number of clusters was set to five, balancing classification distinction with ecological interpretability. To enhance the comparability between different years, the K-means model was first trained on data from 1990, and the cluster centers were extracted. The resulting classes were then re-ordered based on the mean T value to achieve semantic consistency across categories. Using the same imputation approach, normalization parameters, and cluster centers, data from 2023 were classified accordingly. This procedure avoided the potential classification bias caused by model differences and ensured the temporal consistency and stability of the analysis. The clustering results are shown in Figure 11.
Since K-means clustering lacks semantic interpretability, we extracted the overall slope spectrum characteristics of each cluster in 1990 (excluding Cluster 5, which corresponds to scattered cropland patches or outliers, often characterized by 100% dominance within a single slope class). Figure 12 illustrates the slope-wise area distribution of different clusters, highlighting the notable heterogeneity in cropland slope structures. Cluster 1 is mainly concentrated within the 1–2° range, showing a pronounced preference for flat terrain, with a peak area proportion exceeding 30% and a rapid decline beyond 5°, representing typical plains cropland. Cluster 2 peaks between 2 and 5°, with a broader slope distribution than Cluster 1, indicating its location in low hills or gently sloping areas with favorable farming conditions. Cluster 3 shows a more evenly distributed area proportion across 3–10°, reflecting moderate suitability under an undulating topography. Cluster 4 is dominated by medium to high slopes, maintaining a relatively high area proportion above 10° without a distinct peak, suggesting its distribution in hilly or mountainous regions where farming is more difficult. Overall, the distinct differences in slope structure across clusters demonstrate that slope spectra can serve as an effective classification basis for identifying cropland’s spatial structure and its temporal evolution.
By comparing the clustering results of 1990 and 2023, the transformation of cropland slope patterns over the past 30 years was obtained. From Figure 13, it is evident that the cropland slope structure in most regions of China remains relatively stable, with changes occurring in scattered and isolated areas. The regions with concentrated changes mainly include the northern part of Shaanxi Province (SN), which shifted to Cluster 3, and much of southern China, which transitioned to Cluster 3 (as shown in the figure, the dominant transition is from Cluster 1 to Cluster 3). In southwestern China, the dominant transition is towards Cluster 4, particularly in the border areas of Guangxi (GX) and Guizhou (GZ). Other type transitions, such as towards Cluster 1, are sporadically distributed in regions like Ningxia (NM) and Xinjiang (XJ).
To more clearly analyze the transformations between cluster types, three typical regions were selected for detailed analysis (Figure 14). In the case of Shanxi (SX) and Shaanxi (SN) Provinces, a significant transformation from Cluster 4 to Cluster 3 was observed in northern Shaanxi and western Shanxi. This change is primarily attributed to the implementation of the ‘Grain for Green’ policy, which led to a marked reduction in cropland, particularly in high-slope areas. In northern Shaanxi, cropland area decreased from 28,100 km2 in 1990 to 19,700 km2 in 2020. In the border region of Yunnan (YN), Guizhou (GZ), and Guangxi (GX), the transformation was mainly characterized by a shift from Cluster 3 to Cluster 4, indicating an evident trend of cropland expansion toward steeper slopes. This area lies in the transitional zone of the Yunnan–Guizhou Plateau and is a representative karst landscape region in Southwest China, where slope variation is particularly prominent. For Guangdong (GD) Province, changes were mainly concentrated in the Pearl River Delta (especially Guangzhou), where a shift from Cluster 1 to Cluster 3 was observed. As one of the fastest-urbanizing regions in China, rapid urban expansion has driven cropland retreat toward higher-slope areas.

3.3. The Impact of Cropland Slope Structure Changes on Cropland Fragmentation and Soil Erosion

Based on the information provided by the slope spectrum, slope structure status indicators (including T-value, SMA, PaP, PaT, and ULS) were constructed to reflect spatial differences. Dynamic slope structure indicators (including ΔT, ΔSMA, ΔPaP, CSCI, and ΔULS) were used to reflect temporal differences. A GWR analysis was conducted using both status and dynamic indicators in relation to fragmentation, soil erosion, and their differences, to analyze the correlation between different indicator characteristics.
For the calculation of the cropland fragmentation index, we adopted the method proposed by Zhao et al. [63], using CLCD data with the H3 grid partition as the unit. Soil erosion was obtained through processing with Soil Erosion Dataset, and the specific method is described in the dataset introduction section. Due to the unavailability of 2023 data in the Soil Erosion Dataset, soil erosion values from 2022 were adopted in this study. Through geographically weighted regression, we obtained the correlation (R2) between 1990 and 2023 and the state index of the corresponding years, and also calculated the correlation between the amount of change and the change index for the period 1990–2023; the results are shown in Table 1.
According to the spatial correlation patterns between the cropland fragmentation index and slope structure dynamic indicators (Figure 15), CSCI and ΔT exhibit strong positive correlations in South China (e.g., JX, FJ), Southwest China (e.g., GX, GZ, YN), and parts of Northwest China (e.g., SX, SN, NM). This indicates that in these regions, once croplands begin to exhibit an ‘upslope expansion’ trend in slope structure (i.e., increasing CSCI and T values), fragmentation tends to intensify, possibly reflecting the disruptive effect of slope farming on cropland connectivity. In contrast, ΔPaP shows a significant negative correlation near the border between HE and SX, suggesting that in these areas, an increased proportion of cropland located at the slope spectrum peak (i.e., higher PaP) is associated with reduced fragmentation. This implies that the concentration of cropland within a specific slope range may help maintain the integrity and cohesion of the cropland patches. Meanwhile, ΔULS displays a consistent positive correlation in North China (e.g., BJ, TJ, HE, and SD) and the central provinces, indicating that as the ULS of cropland increases—i.e., cropland gradually expands to a steeper terrain—the degree of fragmentation significantly increases. This may result from a combination of factors, such as slope constraints, topographic dissection, and marginal land development, which collectively lead to a more dispersed cropland spatial pattern.
According to the spatial correlation patterns between soil erosion and slope structure dynamic indicators (Figure 16), CSCI and ΔT show significant positive correlations in Southwest China (e.g., GZ, SC, and YN) and in central hilly regions (e.g., HN, CQ, and western HB). This indicates that once croplands in these areas undergo a noticeable “upslope” trend or a general shift in slope structure toward steeper gradients, the risk of soil erosion tends to sharply increase. ΔPaP also exhibits a positive correlation in the transitional zone between the Huang-Huai Plain and the middle–lower reaches of the Yangtze River, suggesting that an increase in the proportion of cropland at the slope spectrum peak is associated with increased soil erosion. The correlation of ΔSMA is most pronounced in the mountainous areas of the southwest and the central hilly regions, where increases in the slope corresponding to the maximum cropland area ratio significantly contribute to the intensification of soil erosion. In contrast, ΔULS generally displays a positive correlation across most parts of the country, especially in regions such as YN, GZ, CQ, SN, and SX, where the increase in upper limit slope (ULS) is notably associated with greater soil erosion, particularly in mountainous and hilly landscapes.
While the GWR results reveal clear spatial correlations between cropland slope structure changes and both fragmentation and soil erosion, these relationships may be influenced by confounding factors, such as rainfall, vegetation cover, land use intensity, and policy interventions (e.g., Grain for Green). Due to limitations in the data and model, this study does not explicitly control for these factors, which may affect causal interpretation. Future research could incorporate control variables or adopt multivariate and structural modeling approaches to improve explanatory power.

4. Discussion

4.1. Spatial Heterogeneity of Slope Structure Changes and Its Underlying Causes

The cropland slope structure in China exhibits pronounced spatial heterogeneity during its spatiotemporal evolution, which is largely shaped by the complex geomorphic patterns and regionally differentiated land use pressures. Specifically, in the hilly and mountainous regions of Southwest China—such as the eastern Yunnan–Guizhou Plateau (YN, GZ) and the Xiang-Gui Hills (HN, GX)—shortages of flat arable land and high-intensity human–land conflicts have driven the continuous expansion of cropland into steeper terrains. This has led to significant increases in both the ULS and the CSCI. This ‘steep-slope expansion’ trend reflects a spatial compensation mechanism in cropland development at topographic boundaries and is emblematic of the progressive encroachment toward the ecological carrying capacity in mountainous areas. In contrast, relatively flat regions such as the North China Plain (HB, HN, SD), the Northeast China Plain (LN, JL, HL), and the Songnen Plain (HL) exhibit minimal changes in slope structure. These areas are characterized by a concentration of cropland on low-slope land, with stable peak values and a generally flattened slope spectrum. This suggests a mature and stable cropland configuration, with limited room for new development and a slope structure that has entered a plateau phase.
Further analysis reveals that this spatial divergence is not only governed by topographic constraints but is also closely related to regional socioeconomic development stages, land consolidation policies, and cropland protection strategies. For instance, in southern hilly areas (e.g., JX, FJ, and GD), ecological restoration programs such as the Grain-for-Green Project have curbed excessive cropland expansion on steep slopes. However, these interventions have also triggered the redistribution and reorganization of marginal cropland, resulting in new inflection points and differentiation in slope structure.
In summary, the spatial heterogeneity of cropland slope structure in China is a product of the interplay between natural geomorphic conditions and anthropogenic activities. It reflects a typical ‘Southwest Mountain–Eastern Plain’ spatial gradient, which is both constrained by terrain and shaped by dynamic land use policies. This pattern underscores the coupled mechanisms of topographic adaptation and ecological constraints in the cropland development process.

4.2. Comparing the Scale Effects of H3 Grids and Administrative Units

Traditional land use studies often rely on administrative boundaries as spatial statistical units. However, this approach is prone to the Modifiable Areal Unit Problem (MAUP) [64], which introduces scale- and shape-related biases into the statistical outcomes, thereby obscuring critical intra-regional spatial heterogeneity. In this context, our study employs the H3 hexagonal grid system as the spatial unit for analyzing cropland slope structure and its ecological implications.
Compared to the administrative-unit-based approaches [16,24], the H3 grid framework demonstrates stronger explanatory power in capturing spatial sensitivity and structural detail. Although both administrative units and H3 grids exhibit similar spatial patterns in cropland slope indices—such as the CSCI—H3 grids more effectively reflect the intra-regional spatial variation at finer scales. For instance, W. Chen et al. [45] conducted a national-scale analysis using prefecture-level administrative units and found that 86.76% of regions had a consistently positive Cropland Climbing Index (CCI). However, our fine-resolution grid-based analysis reveals that the actual proportion is significantly lower. The irregular geometry of administrative boundaries can introduce statistical distortion, such as overestimations of regional slope trends due to extreme values in a small subset of cropland. Regularly shaped H3 grids help mitigate such errors.
Furthermore, the H3 system supports a nested, multi-scale structure, which facilitates scale decomposition and cross-regional comparisons, thereby enhancing the flexibility and generalizability of spatial analyses. Of course, using H3 grids also involves challenges, such as the need to select appropriate resolutions and manage more complex data workflows. Nevertheless, by mitigating MAUP-related biases and offering a standardized framework for spatial comparison, H3 grids provide a technically robust alternative for analyzing cropland slope structure and the associated ecological effects.

4.3. Slope Structure, Cropland Fragmentation, and Soil Erosion Risk

Slope structure significantly affects cropland fragmentation and soil erosion risk. Higher ULS and CSCI values are linked to steeper, more fragmented cropland, especially in topographically complex areas. This study, using geographically weighted regression, identifies spatially non-stationary relationships between slope structure and fragmentation—patterns that are obscured in traditional analyses based on global regression or aggregate indicators. This finding supports that of Zhao et al. [63] while extending the analysis with higher spatial precision. In areas such as Southwest China and the margins of the Sichuan Basin, the influence of slope on fragmentation is particularly strong. By contrast, flatter regions such as the North China Plain exhibit weaker correlations, underscoring the role of the terrain background.
Steep-slope cultivation also correlates with high soil erosion risk [29,65], with high-ULS/CSCI zones overlapping with erosion hotspots. Importantly, our study shows that not only are static slope metrics crucial, but so is temporal slope evolution (e.g., T-value changes). Sharp T-value fluctuations often align with land consolidation or ecological restoration activities, and may signal risky cropland expansion toward slope limits. Therefore, policymakers should incorporate dynamic slope metrics when planning land development and restoration, to ensure a balance between agricultural productivity and ecological sustainability.

5. Conclusions

This study developed a multi-indicator framework to investigate the spatiotemporal evolution of cropland slope structure in China from 1990 to 2023. Based on a slope spectrum analysis and H3 hexagonal grid partitioning, the framework integrates key indicators—including T-value, SMA, PaP, PaT, and ULS—to comprehensively quantify the changes in slope structure across diverse agricultural landscapes. By combining spatial clustering with a temporal trend analysis, the study reveals pronounced regional disparities in the transformation of cropland slope patterns.
The findings indicate a relatively stable slope structure in most cropland areas nationwide. However, notable spatial heterogeneity was observed: Southwest China (e.g., YN, GZ, GX) exhibited a significant downward trend in CSCI, reflecting the impact of ecological restoration programs and the retreat from steep-slope cultivation. In contrast, eastern urban agglomerations experienced CSCI increases, driven by urban encroachment on low-slope farmland and the resulting displacement of agriculture to more topographically complex peripheries. Additionally, local upslope cultivation trends were identified in transitional zones such as the Loess Plateau and the periphery of the Sichuan Basin. The results of the Geographically Weighted Regression analysis reveal notable spatial heterogeneity in the relationships between slope structure changes and landscape outcomes. Increases in ΔT, ΔULS, and CSCI are strongly and positively associated with cropland fragmentation and potential soil erosion in parts of Southwest China (e.g., Yunnan, Guizhou, and the Sichuan Basin periphery) and the Loess Plateau. These findings indicate that slope expansion in ecologically sensitive and topographically complex regions significantly contributes to landscape fragmentation and ecological degradation, emphasizing the need for regionally differentiated land use management and erosion control measures.
This study contributes to the methodological advancement of land system science by integrating slope spectral analysis with grid-based spatial statistics and regression modeling. It provides empirical support for the incorporation of slope structure metrics into cropland monitoring and ecological risk assessment frameworks. However, this research is subject to several limitations, as follows: (1) physical drivers such as precipitation, vegetation cover, and conservation engineering were not fully disentangled from land use dynamics; (2) the lack of high-resolution socioeconomic data limited the attribution analysis of cropland expansion mechanisms; (3) the applied indicators, while robust at macro-scales, may overlook localized micro-topographic variations. Future research should incorporate structural equation modeling (SEM), multi-source remote sensing data, and finer-scale slope units to construct a more complete causal network.
Looking forward, this study offers a scientific basis for formulating differentiated cropland management strategies. We recommend strengthening slope-based zoning control, especially in ecologically fragile and erosion-prone areas; promoting slope-appropriate conservation practices such as contour farming, agroforestry, and terracing; integrating slope structure indicators into national land use planning and early warning systems; and aligning cropland consolidation with ecological restoration to achieve synergy between agricultural productivity and environmental sustainability. These policy insights will support the long-term goal of developing resilient and sustainable cropland systems in China.

Author Contributions

Conceptualization, G.L. and L.B.; methodology, G.L.; software, G.L. and Y.X.; data curation, Y.X.; writing—original draft preparation, G.L.; writing—review and editing, G.L.; visualization, G.L. and Y.X.; supervision, L.B.; funding acquisition, G.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by Yunnan Fundamental Research Projects (grant NO. 202201AT070257).

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Acknowledgments

DeepSeek was used to generate the full names and English abbreviations of each province in China presented in Appendix A, as well as to assist in reviewing the English language of the manuscript. The authors would like to thank the anonymous reviewers for their valuable comments and suggestions on this article.

Conflicts of Interest

The authors declare no potential conflicts of interest with respect to the research, authorship, or publication of this article.

Abbreviations

The following abbreviations are used in this manuscript:
CLCDChina Land Cover Dataset
CSCICropland Slope Change Index
GWRGeographically Weighted Regression
ULSUpper Limit of Slope

Appendix A

Table A1. Administrative divisions of China and their English abbreviations.
Table A1. Administrative divisions of China and their English abbreviations.
Full Name AbbreviationFull Name Abbreviation
Beijing MunicipalityBJHunan ProvinceHN
Tianjin MunicipalityTJGuangdong ProvinceGD
Hebei ProvinceHEGuangxi Zhuang Autonomous RegionGX
Shanxi ProvinceSXHainan ProvinceHI
Inner Mongolia Autonomous RegionNMChongqing MunicipalityCQ
Liaoning ProvinceLNSichuan ProvinceSC
Jilin ProvinceJLGuizhou ProvinceGZ
Heilongjiang ProvinceHLYunnan ProvinceYN
Shanghai MunicipalitySHTibet Autonomous RegionXZ
Jiangsu ProvinceJSShaanxi ProvinceSN
Zhejiang ProvinceZJGansu ProvinceGS
Anhui ProvinceAHQinghai ProvinceQH
Fujian ProvinceFJNingxia Hui Autonomous RegionNX
Jiangxi ProvinceJXXinjiang Uygur Autonomous RegionXJ
Shandong ProvinceSDHong Kong Special Administrative RegionHK
Henan ProvinceHAMacao Special Administrative RegionMO
Hubei ProvinceHBTaiwan ProvinceTW

References

  1. Zhou, Y.; Li, X.; Liu, Y. Cultivated Land Protection and Rational Use in China. Land Use Policy 2021, 106, 105454. [Google Scholar] [CrossRef]
  2. Chen, S.; Pei, B. Spatial-Temporal Variations of Cultivated Land Compensation and Its Compensation Mechanism in Mainland China. Environ. Impact Assess. Rev. 2025, 110, 107712. [Google Scholar] [CrossRef]
  3. World Population Prospects 2024: Summary of Results; United Nations Department of Economic and Social Affairs, Population Division: New York, NY, USA, 2024.
  4. White, J.C.; Gardea-Torresdey, J. Achieving Food Security through the Very Small. Nat. Nanotechnol. 2018, 13, 627–629. [Google Scholar] [CrossRef] [PubMed]
  5. Potapov, P.; Turubanova, S.; Hansen, M.C.; Tyukavina, A.; Zalles, V.; Khan, A.; Song, X.-P.; Pickens, A.; Shen, Q.; Cortez, J. Global Maps of Cropland Extent and Change Show Accelerated Cropland Expansion in the Twenty-First Century. Nat. Food 2022, 3, 19–28. [Google Scholar] [CrossRef] [PubMed]
  6. Koiry, S.; Huang, W. Do Ecological Protection Approaches Affect Total Factor Productivity Change of Cropland Production in Sweden? Ecol. Econ. 2023, 209, 107829. [Google Scholar] [CrossRef]
  7. Liu, J.; Jin, X.; Xu, W.; Gu, Z.; Yang, X.; Ren, J.; Fan, Y.; Zhou, Y. A New Framework of Land Use Efficiency for the Coordination among Food, Economy and Ecology in Regional Development. Sci. Total Environ. 2020, 710, 135670. [Google Scholar] [CrossRef]
  8. Jin, G.; Peng, J.; Zhang, L.; Zhang, Z. Understanding Land for High-Quality Development. J. Geogr. Sci 2023, 33, 217–221. [Google Scholar] [CrossRef]
  9. United Nations, U.N. Transforming Our World: The 2030 Agenda for Sustainable Development; Working Paper; United Nations: New York, NY, USA, 2015. [Google Scholar]
  10. Li, H.; Song, W. Spatial Transformation of Changes in Global Cultivated Land. Sci. Total Environ. 2023, 859, 160194. [Google Scholar] [CrossRef]
  11. Kang, J.; Duan, X.; Yun, R. The Impact of Urbanization on Food Security: A Case Study of Jiangsu Province. Land 2023, 12, 1681. [Google Scholar] [CrossRef]
  12. Cai, J.; Ma, E.; Lin, J.; Liao, L.; Han, Y. Exploring Global Food Security Pattern from the Perspective of Spatio-Temporal Evolution. J. Geogr. Sci. 2020, 30, 179–196. [Google Scholar] [CrossRef]
  13. Lee, C.-C.; Zeng, M.; Luo, K. The Impact of Urbanization on Food Security in China. Int. Rev. Econ. Financ. 2024, 93, 1159–1175. [Google Scholar] [CrossRef]
  14. Ouyang, X.; Xu, J.; Li, J.; Wei, X.; Li, Y. Land Space Optimization of Urban-Agriculture-Ecological Functions in the Changsha-Zhuzhou-Xiangtan Urban Agglomeration, China. Land Use Policy 2022, 117, 106112. [Google Scholar] [CrossRef]
  15. Tu, Y.; Chen, B.; Yu, L.; Xin, Q.; Gong, P.; Xu, B. How Does Urban Expansion Interact with Cropland Loss? A Comparison of 14 Chinese Cities from 1980 to 2015. Landsc. Ecol. 2021, 36, 243–263. [Google Scholar] [CrossRef]
  16. Pan, S.; Liang, J.; Chen, W.; Peng, Y. Uphill or Downhill? Cropland Use Change and Its Drivers from the Perspective of Slope Spectrum. J. Mt. Sci. 2024, 21, 484–499. [Google Scholar] [CrossRef]
  17. Lu, J.; Peng, Q.; Song, Y.; Lyu, L.; Chen, D.; Huang, P.; Peng, F.; Liu, Y. Characteristics and Effects of Global Sloping Land Urbanization from 2000 to 2020. Sci. Total Environ. 2024, 937, 173348. [Google Scholar] [CrossRef] [PubMed]
  18. Zhou, Y.; Zhong, Z.; Cheng, G. Cultivated Land Loss and Construction Land Expansion in China: Evidence from National Land Surveys in 1996, 2009 and 2019. Land Use Policy 2023, 125, 106496. [Google Scholar] [CrossRef]
  19. Zhou, Y.; Chen, T.; Feng, Z.; Wu, K. Identifying the Contradiction between the Cultivated Land Fragmentation and the Construction Land Expansion from the Perspective of Urban-Rural Differences. Ecol. Inform. 2022, 71, 101826. [Google Scholar] [CrossRef]
  20. Wang, M.; Yuan, M.; Han, P.; Wang, D. Assessing Sustainable Urban Development Based on Functional Spatial Differentiation of Urban Agriculture in Wuhan, China. Land Use Policy 2022, 115, 105999. [Google Scholar] [CrossRef]
  21. Li, W.; Chen, Z.; Li, M.; Qiu, X.; Zhao, Q.; Chen, Y. Spatial Conflict Identification and Scenario Coordination for Construction-agricultural-ecological Land Use. Environ. Dev. Sustain. 2025, 27, 1933–1961. [Google Scholar] [CrossRef]
  22. Zhao, S.; Yin, M. Change of Urban and Rural Construction Land and Driving Factors of Arable Land Occupation. PLoS ONE 2023, 18, e0286248. [Google Scholar] [CrossRef]
  23. Li, W.; Wang, W.; Chen, J.; Zhang, Z. Assessing Effects of the Returning Farmland to Forest Program on Vegetation Cover Changes at Multiple Spatial Scales: The Case of Northwest Yunnan, China. J. Environ. Manag. 2022, 304, 114303. [Google Scholar] [CrossRef]
  24. He, T.; Li, J.; Zhang, M.; Zhai, G.; Lu, Y.; Wang, Y.; Guo, A.; Wu, C. Uphill Cropland and Stability Assessment of Gained Cropland in China over the Preceding 30 Years. J. Geogr. Sci. 2024, 34, 699–721. [Google Scholar] [CrossRef]
  25. Khurshid, M.; Nafees, M.; Khan, A.; Yin, H.; Ullah, W.; Rashid, W.; Han, H.; Lashari, A.H. Off-Season Agriculture Encroachment in the Uplands of Northern Pakistan: Need for Sustainable Land Management. Land 2022, 11, 520. [Google Scholar] [CrossRef]
  26. Huo, J.; Yu, X.; Liu, C.; Chen, L.; Zheng, W.; Yang, Y.; Tang, Z. Effects of Soil and Water Conservation Management and Rainfall Types on Runoff and Soil Loss for a Sloping Area in North China. Land Degrad. Dev. 2020, 31, 2117–2130. [Google Scholar] [CrossRef]
  27. Fang, H. Effect of Soil Conservation Measures and Slope on Runoff, Soil, TN, and TP Losses from Cultivated Lands in Northern China. Ecol. Indic. 2021, 126, 107677. [Google Scholar] [CrossRef]
  28. Yu, P.; Yang, T.; Zhang, Z.; Zhou, X.; Qi, Z.; Yin, Z.; Li, A. Soil and Water Conservation Effects of Different Tillage Measures on Phaeozems Sloping Farmland in Northeast China. Land Degrad. Dev. 2024, 35, 1716–1733. [Google Scholar] [CrossRef]
  29. Wang, Y.; Hu, X.; Yu, S.; Wang, Z.; Zhao, J.; Fang, N.; Xiao, H.; Wang, L.; Shi, Z. Soil Conservation of Sloping Farmland in China: History, Present, and Future. Earth-Sci. Rev. 2024, 249, 104655. [Google Scholar] [CrossRef]
  30. Du, G.; Wang, X.; Wang, J.; Liu, Y.; Zhang, H. Analysis of the Spatial–Temporal Pattern of the Newly Increased Cultivated Land and Its Vulnerability in Northeast China. Land 2023, 12, 796. [Google Scholar] [CrossRef]
  31. Li, D.; Chen, G.; Xi, Z.; He, P.; Xin, X.; Chen, J.; Yu, H.; Kang, G. Spatiotemporal Changes of Cultivated Land Utilization in Black Soil Region, China Based on Geo-Information Tupu. Sci. Rep. 2025, 15, 10306. [Google Scholar] [CrossRef]
  32. Han, S.; Shao, Q.; Ning, J.; Jin, S. Spatio-Temporal Changes of Arable Land and Their Impacts on Grain Output in the Yangtze River Economic Belt from 1980 to 2020. Remote Sens. 2024, 16, 1672. [Google Scholar] [CrossRef]
  33. Ye, S.; Song, C.; Kuzyakov, Y.; Cheng, F.; Kong, X.; Feng, Z.; Gao, P. Arable Land Quality in Developing China: An Integrated Exploration from Global Challenges to Localized Solutions. Land 2024, 13, 350. [Google Scholar] [CrossRef]
  34. Junaid, M.; Gokce, A. Global agricultural losses and their causes. Bull. Biol. Allied Sci. Res. 2024, 2024, 66. [Google Scholar] [CrossRef]
  35. Dai, K.; Cheng, C.; Kan, S.; Li, Y.; Liu, K.; Wu, X. Impact of Arable Land Abandonment on Crop Production Losses in Ukraine During the Armed Conflict. Remote Sens. 2024, 16, 4207. [Google Scholar] [CrossRef]
  36. Asmamaw, L.B.; Mohammed, A.A. Effects of Slope Gradient and Changes in Land Use/Cover on Selected Soil Physico-Biochemical Properties of the Gerado Catchment, North-Eastern Ethiopia. Int. J. Environ. Stud. 2013, 70, 111–125. [Google Scholar] [CrossRef]
  37. Geremew, B.; Tadesse, T.; Bedadi, B.; Gollany, H.T.; Tesfaye, K.; Aschalew, A. Impact of Land Use/Cover Change and Slope Gradient on Soil Organic Carbon Stock in Anjeni Watershed, Northwest Ethiopia. Environ. Monit. Assess. 2023, 195, 971. [Google Scholar] [CrossRef]
  38. Chen, H.; Tan, Y.; Xiao, W.; Li, G.; Meng, F.; He, T.; Li, X. Urbanization in China Drives Farmland Uphill under the Constraint of the Requisition–Compensation Balance. Sci. Total Environ. 2022, 831, 154895. [Google Scholar] [CrossRef] [PubMed]
  39. Tang, G.; Li, F.; Liu, X.; Long, Y.; Yang, X. Research on the Slope Spectrum of the Loess Plateau. Sci. China Ser. E-Technol. Sci. 2008, 51, 175–185. [Google Scholar] [CrossRef]
  40. Zhou, L.; Dang, X.; Mu, H.; Wang, B.; Wang, S. Cities Are Going Uphill: Slope Gradient Analysis of Urban Expansion and Its Driving Factors in China. Sci. Total Environ. 2021, 775, 145836. [Google Scholar] [CrossRef]
  41. Shi, K.; Liu, G.; Zhou, L.; Cui, Y.; Liu, S.; Wu, Y. Satellite Remote Sensing Data Reveal Increased Slope Climbing of Urban Land Expansion Worldwide. Landsc. Urban Plan. 2023, 235, 104755. [Google Scholar] [CrossRef]
  42. Shi, K.; Cui, Y.; Liu, S.; Wu, Y. Global Urban Land Expansion Tends to Be Slope Climbing: A Remotely Sensed Nighttime Light Approach. Earth’s Future 2023, 11, e2022EF003384. [Google Scholar] [CrossRef]
  43. Duan, J.; Peng, Q.; Huang, P. Slope Characteristics of Urban Construction Land and Its Correlation with Ground Slope in China. Open Geosci. 2022, 14, 1524–1537. [Google Scholar] [CrossRef]
  44. Zhang, H.; Zhao, X.; Ren, J.; Hai, W.; Guo, J.; Li, C.; Gao, Y. Research on the Slope Gradient Effect and Driving Factors of Construction Land in Urban Agglomerations in the Upper Yellow River: A Case Study of the Lanzhou–Xining Urban Agglomerations. Land 2023, 12, 745. [Google Scholar] [CrossRef]
  45. Chen, W.; Xie, Y.; Yuan, J.; Zeng, J.; Yang, L.; Gu, T.; Lei, F. Slope-Climbing of Cropland and Its Effects in China. J. Mt. Sci. 2024, 21, 2754–2769. [Google Scholar] [CrossRef]
  46. Faostat. Available online: https://www.fao.org/faostat/en/#home (accessed on 7 May 2025).
  47. Wen, A.; Tang, Q.; Ouyang, C.; Zhu, B.; Wang, Y.; Li, A.; Li, S.; Zhu, W.; Liu, L. Mountain Protection and Mountain Development in China: Review and Prospect. Bull. Chin. Acad. Sci. (Chin. Version) 2023, 38, 376–384. [Google Scholar]
  48. Borrelli, P.; Robinson, D.A.; Fleischer, L.R.; Lugato, E.; Ballabio, C.; Alewell, C.; Meusburger, K.; Modugno, S.; Schütt, B.; Ferro, V.; et al. An Assessment of the Global Impact of 21st Century Land Use Change on Soil Erosion. Nat. Commun. 2017, 8, 2013. [Google Scholar] [CrossRef] [PubMed]
  49. Liu, B.; Xie, Y.; Li, Z.; Liang, Y.; Zhang, W.; Fu, S.; Yin, S.; Wei, X.; Zhang, K.; Wang, Z.; et al. The Assessment of Soil Loss by Water Erosion in China. Int. Soil Water Conserv. Res. 2020, 8, 430–439. [Google Scholar] [CrossRef]
  50. Xinhua News Agency Bulletin of Main Data of the Third National Land Survey. Available online: https://www.gov.cn/xinwen/2021-08/26/content_5633490.htm (accessed on 10 May 2025). (In Chinese)
  51. Yang, J.; Huang, X. The 30 m Annual Land Cover Datasets and Its Dynamics in China from 1985 to 2023. 2024. Available online: https://zenodo.org/records/12779975 (accessed on 14 May 2025).
  52. Zhang, C.; Dong, J.; Ge, Q. Quantifying the Accuracies of Six 30-m Cropland Datasets over China: A Comparison and Evaluation Analysis. Comput. Electron. Agric. 2022, 197, 106946. [Google Scholar] [CrossRef]
  53. 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, 3907–3925. [Google Scholar] [CrossRef]
  54. Pan, Y.; Tian, Y.; Wu, Y.; Fu, M. A Comprehensive Approach for the Spatial Optimization of the Biodiversity Conservation Network in the Qinling Mountains, China. Landsc Ecol 2025, 40, 11. [Google Scholar] [CrossRef]
  55. Feng, Z.; Yang, X.; Li, S. New Insights of Eco-Environmental Vulnerability in China’s Yellow River Basin: Spatio-Temporal Pattern and Contributor Identification. Ecol. Indic. 2024, 167, 112655. [Google Scholar] [CrossRef]
  56. An, X.; Jin, W.; Zhang, H.; Liu, Y.; Zhang, M. Analysis of Long-Term Wetland Variations in China Using Land Use/Land Cover Dataset Derived from Landsat Images. Ecol. Indic. 2022, 145, 109689. [Google Scholar] [CrossRef]
  57. Crippen, R.; Buckley, S.; Agram, P.; Belz, E.; Gurrola, E.; Hensley, S.; Kobrick, M.; Lavalle, M.; Martin, J.; Neumann, M.; et al. NASADEM global elevation model: Methods and progress. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2016, XLI-B4, 125–128. [Google Scholar] [CrossRef]
  58. Yan, J.; Wang, S.; Feng, J.; He, H.; Wang, L.; Sun, Z.; Zheng, C. New 30-m Resolution Dataset Reveals Declining Soil Erosion with Regional Increases across Chinese Mainland (1990–2022). Remote Sens. Environ. 2025, 323, 114681. [Google Scholar] [CrossRef]
  59. Yan, J.; Wang, S.; Feng, J.; He, H.; Wang, L.; Sun, Z.; Zheng, C. The 30 m Annual Soil Water Erosion Dataset in Chinese Mainland from 1990 to 2022. 2024. Available online: https://www.scidb.cn/en/detail?dataSetId=9d14070a664f4d368ca107c5e9d6b746 (accessed on 10 January 2025).
  60. GDAL/OGR Contributors. GDAL/OGR Geospatial Data Abstraction Software Library; Open Source Geospatial Foundation: Beaverton, OR, USA, 2025. [Google Scholar]
  61. Fotheringham, A.S.; Brunsdon, C.; Charlton, M. Geographically Weighted Regression. In The Sage Handbook of Spatial Analysis; Sage Publishing: Thousand Oaks, CA, USA, 2009; Volume 1, pp. 243–254. [Google Scholar]
  62. Brodsky, I. H3: Uber’s Hexagonal Hierarchical Spatial Index. Available from Uber Engineering Website. Available online: https://eng.uber.com/h3/ (accessed on 5 January 2025).
  63. Zhao, N.; Chen, K.; Wu, X.; Zhang, L.; Wang, W. Cropland Fragmentation Change across China over the Last Two Decades. Agric. Syst. 2024, 218, 104010. [Google Scholar] [CrossRef]
  64. Wong, D.W.S. The Modifiable Areal Unit Problem (MAUP). In WorldMinds: Geographical Perspectives on 100 Problems: Commemorating the 100th Anniversary of the Association of American Geographers 1904–2004; Janelle, D.G., Warf, B., Hansen, K., Eds.; Springer: Dordrecht, The Netherlands, 2004; pp. 571–575. ISBN 978-1-4020-2352-1. [Google Scholar]
  65. Li, K.; Wang, L.; Wang, Z.; Hu, Y.; Zeng, Y.; Yan, H.; Xu, B.; Li, C.; Cui, H.; Yu, S.; et al. Multiple Perspective Accountings of Cropland Soil Erosion in China Reveal Its Complex Connection with Socioeconomic Activities. Agric. Ecosyst. Environ. 2022, 337, 108083. [Google Scholar] [CrossRef]
Land 14 01093 g001
Figure 1. Cultivated land distribution (2023) and topography in the study area.
Figure 1. Cultivated land distribution (2023) and topography in the study area.
Land 14 01093 g001
Land 14 01093 g002
Figure 2. Cropland slope spectrum and key indicators.
Figure 2. Cropland slope spectrum and key indicators.
Land 14 01093 g002
Land 14 01093 g003
Figure 3. Average slop of cultivated land from 1990 to 2025.
Figure 3. Average slop of cultivated land from 1990 to 2025.
Land 14 01093 g003
Land 14 01093 g004
Figure 4. The slope spectrum of the terrain background and cropland from 1990 to 2023.
Figure 4. The slope spectrum of the terrain background and cropland from 1990 to 2023.
Land 14 01093 g004
Land 14 01093 g005
Figure 5. Heatmaps of absolute cropland area changes and relative changes.
Figure 5. Heatmaps of absolute cropland area changes and relative changes.
Land 14 01093 g005
Land 14 01093 g006
Figure 6. Spatial distribution of SMA and its changes.
Figure 6. Spatial distribution of SMA and its changes.
Land 14 01093 g006
Land 14 01093 g007
Figure 7. Spatial distribution of PaP and its changes.
Figure 7. Spatial distribution of PaP and its changes.
Land 14 01093 g007
Land 14 01093 g008
Figure 8. Spatial distribution of T values in 1990 and 2023 and its changes.
Figure 8. Spatial distribution of T values in 1990 and 2023 and its changes.
Land 14 01093 g008
Land 14 01093 g009
Figure 9. Spatial distribution of ULS in 1990 and 2023 and its changes.
Figure 9. Spatial distribution of ULS in 1990 and 2023 and its changes.
Land 14 01093 g009
Land 14 01093 g010
Figure 10. Temporal trends in CSCI in China (1990–2023).
Figure 10. Temporal trends in CSCI in China (1990–2023).
Land 14 01093 g010
Land 14 01093 g011
Figure 11. Slope characteristic classification of China in 1990 and 2023.
Figure 11. Slope characteristic classification of China in 1990 and 2023.
Land 14 01093 g011
Land 14 01093 g012
Figure 12. Slope spectrum of cropland with different clustering results.
Figure 12. Slope spectrum of cropland with different clustering results.
Land 14 01093 g012
Land 14 01093 g013
Figure 13. Transformation of cropland slope patterns over the past 30 years.
Figure 13. Transformation of cropland slope patterns over the past 30 years.
Land 14 01093 g013
Land 14 01093 g014
Figure 14. Typical areas of cropland slope pattern transformation.
Figure 14. Typical areas of cropland slope pattern transformation.
Land 14 01093 g014
Land 14 01093 g015
Figure 15. Spatial correlation patterns between the cropland fragmentation index and slope structure dynamic indicators.
Figure 15. Spatial correlation patterns between the cropland fragmentation index and slope structure dynamic indicators.
Land 14 01093 g015
Land 14 01093 g016
Figure 16. Spatial correlation patterns between soil erosion and slope structure dynamic indicators.
Figure 16. Spatial correlation patterns between soil erosion and slope structure dynamic indicators.
Land 14 01093 g016
Table 1. Correlation analysis results (R2).
Table 1. Correlation analysis results (R2).
Indicator1990
(R2)		2023
(R2)		Change Index
(R2)
1990 Cropland Fragmentation Index0.5486————
2023 Cropland Fragmentation Index——0.5650——
1990–2023 Change in Cropland Fragmentation Index————0.3546
1990 Soil Erosion Amount0.5622————
2023 Soil Erosion Amount——0.5882——
1990–2023 Change in Soil Erosion Amount————0.5076
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.

Share and Cite

MDPI and ACS Style

Liu, G.; Xia, Y.; Bao, L. The Evolution of Cropland Slope Structure and Its Implications for Fragmentation and Soil Erosion in China. Land 2025, 14, 1093. https://doi.org/10.3390/land14051093

AMA Style

Liu G, Xia Y, Bao L. The Evolution of Cropland Slope Structure and Its Implications for Fragmentation and Soil Erosion in China. Land. 2025; 14(5):1093. https://doi.org/10.3390/land14051093

Chicago/Turabian Style

Liu, Guangjie, Yi Xia, and Li Bao. 2025. "The Evolution of Cropland Slope Structure and Its Implications for Fragmentation and Soil Erosion in China" Land 14, no. 5: 1093. https://doi.org/10.3390/land14051093

APA Style

Liu, G., Xia, Y., & Bao, L. (2025). The Evolution of Cropland Slope Structure and Its Implications for Fragmentation and Soil Erosion in China. Land, 14(5), 1093. https://doi.org/10.3390/land14051093

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop