Next Article in Journal
Environmental Justice in the Green Transition of Rural Post-Industrial Waterfronts: A Villagers’ Perspective—A Case Study of the Waterfront Area in Jiangsu Province, China
Previous Article in Journal
Spatial–Temporal Characteristics and Influencing Factors of Shrinking County Towns’ Resilience in China
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Continuous Monitoring of Cropland Abandonment in China Since the 21st Century: Interpreting Spatiotemporal Trajectories and Characteristics

1
Rural Development Institute, Chinese Academy of Social Science, Beijing 100732, China
2
College of Land Science and Technology, China Agricultural University, Beijing 100193, China
*
Author to whom correspondence should be addressed.
Land 2025, 14(11), 2203; https://doi.org/10.3390/land14112203 (registering DOI)
Submission received: 10 October 2025 / Revised: 3 November 2025 / Accepted: 4 November 2025 / Published: 6 November 2025
(This article belongs to the Section Land Use, Impact Assessment and Sustainability)

Abstract

Farmland abandonment poses a significant threat to China’s food security by contributing to inefficient land use. Utilizing remote sensing data and the multiple cropping index extraction method, this study extracts abandonment cropland information and analyzes its spatiotemporal patterns across China, with its findings validated against the “China Rural Revitalization Survey” (CRRS) data. The results indicate that since the 21st century, China’s cropland abandonment rate has fluctuated around 5.86%, affecting an average of 7.6 million hectares annually. Spatially, cropland abandonment is more severe in southern China, with hotspots clustered around 25° N and 30° N latitudes. This southward shift exacerbates the spatial mismatch between water resources and cropland. Furthermore, abandonment is particularly pronounced in grain production—marketing balance areas and main marketing areas, intensifying pressure on national food self-sufficiency. Slope and fragmentation also drive abandonment, with steeper (>15°) and more fragmented plots showing higher susceptibility. These complex patterns are uncovered through the study’s systematic innovations—a dual-indicator quantification method, a multi-source validation framework, a dynamic spatiotemporal atlas, and a novel interpretation of abandonment multifunctionality, which also positions farmland reuse as a buffer against unemployment risks. We thus recommend addressing land fragmentation as a core strategy, through high-standard farmland construction, innovative contract rights certification, and expanded agricultural socialized services to promote moderate-scale farming. Finally, we urge the adoption of region-specific and category-based recultivation approaches, supported by clear governance priorities.

1. Introduction

Safeguarding national food security is a critical strategic issue concerning the national economy and people’s livelihood [1]. Currently, China feeds nearly 18% of the world’s population with only 9% of the world’s arable land. In the short term, efforts to increase yield by enhancing production capacity are constrained by limitations in scientific and technological progress. Therefore, achieving food security goals requires the assurance of more arable land resources. Despite the imminent shortage of land suitable for agricultural expansion, the phenomenon of farmers abandoning cultivated land remains frequent [2]. Since abandoned cropland holds significant potential for increasing grain production [3], promoting the reclamation of such land represents an important short-term option to boost grain output. This would effectively mitigate food security risks and holds great significance for achieving China’s food security strategy.
Farmland abandonment, as a common phenomenon in the process of modernization, is widespread globally. A key branch of research focuses on quantifying its status, yet the exact extent remains unclear due to the dynamic nature of abandonment and reclamation, definitional differences, and data scarcity. Understanding this status is fundamental for both academic research and government-led remediation efforts. Social surveys and remote sensing analysis are the two primary investigative methods [4]. While surveys are valuable for quantitatively exploring the root causes of abandonment, they are prone to significant estimation biases. For instance, an overestimation bias arises when studies focus on villages in mountainous areas with poor agricultural conditions [5,6,7,8,9]. Conversely, underestimation occurs because migrant worker households (who are more likely to abandon cropland) are often excluded from surveys, and because farmers or local officials may conceal abandonment to secure subsidies or meet performance metrics. The high cost of surveys also limits large-scale, long-term tracking studies.
These spatio-temporal limitations of survey-based research hinder robust cross-country comparisons and longitudinal analysis of abandonment’s relationship with socioeconomic development, making it difficult to form a dialectical view of the issue—whether it is an alarming threat or an inevitable byproduct of economic development. In contrast, remote sensing-based macro-studies, despite challenges in detecting hidden abandonment [10], offer advantages in avoiding the selection biases of surveys and overcoming their spatiotemporal constraints. Therefore, remote sensing and household survey data each possess distinct strengths and weaknesses in acquisition methods, coverage, accuracy, and applicability. Integrating these approaches can complement their respective limitations and significantly enhance the comprehensiveness and reliability of research on farmland abandonment. Another research branch focuses on the causes of farmland abandonment, aiming to analyze farmers’ motivations for leaving agriculture and formulate measures to incentivize the recultivation of abandoned land. The reasons for farmland abandonment can be categorized into natural drivers such as land fragmentation [11], poor topographic conditions [8], environmental degradation (e.g., soil erosion and pollution), and natural disasters (e.g., volcanic eruptions, earthquakes, and floods) [12]; economic drivers such as low comparative returns from agriculture and high costs of land transfer [13]; social drivers such as urbanization and the shift from agricultural to non-agricultural sectors [14]; and institutional drivers such as the “equal distribution” system under the household responsibility system [15], inefficient land transfer mechanisms [16], and deviations in the implementation of farmland replenishment policies [17]. International studies have also examined factors such as socioeconomic and political transformations (e.g., the dissolution of the Soviet Union) [18], conservation area programs, forest protection policies [19], wars, and conflicts [20,21,22]. Evidently, the causes of farmland abandonment are complex and multifaceted. Excluding uncertain factors like natural disasters and political upheavals, it becomes increasingly challenging to identify the root causes of farmland abandonment dialectically, scientifically, and systematically from the multitude of deterministic factors.
This study leverages multi-source remote sensing data, including land cover data and the Normalized Difference Vegetation Index (NDVI), to identify abandoned cropland using a multiple cropping index extraction method. By capitalizing on the reproducibility and ease of correction offered by Stata 17 and R 4.3.3 software in spatial data extraction, processing, and computation, we derive spatiotemporally continuous distribution data of abandoned cropland across China from 2002 to 2020. This approach provides a new pathway for understanding the extent of cropland abandonment at the national level. Furthermore, by integrating data from the China Rural Revitalization Survey (hereinafter referred to as “CRRS”), this study quantitatively analyzes household-level cropland abandonment patterns for the years 2019, 2021, and 2023. The results are validated using the remote sensing monitoring outcomes, thereby enhancing the reliability of the research conclusions.
This study presents significant refinements and expansions to existing research, yielding a series of innovative contributions to the understanding of cropland abandonment. Methodologically, it proposes a dual-indicator system: cropland abandonment probability (based on pixel counts) and cropland abandonment rate (adjusted by area weighting). This system not only constitutes a methodological advance but also functions as a powerful diagnostic tool for quantifying how cropland fragmentation drives abandonment risks. In data validation, a novel cross-verification of remote sensing results with household survey data effectively mitigates single-source limitations and substantially improves finding’s reliability. Regarding spatiotemporal patterns, the study marks a paradigm shift from a static “binary map” to a “multi-stage temporal atlas” of cropland abandonment, revealed through 20 years of continuous data. This refined perspective allows for the precise pinpointing of hotspot regions and, crucially, provides the temporally grounded data necessary to evaluate recultivation potential and inform the design of mitigation strategies. Particularly noteworthy is the pioneering interpretation of abandonment multifunctionality: by analyzing pre- and post-COVID-19 pandemic dynamics, the study frames abandoned cropland as an “employment buffer” and “livelihood safety valve,” offering critical empirical evidence for refining governance policies. Collectively, this work not only advances theoretical insights into abandonment patterns but also delivers a systematic framework for their targeted management.

2. Materials and Methods

2.1. Data Sources

This study integrates multi-source data from 2001 to 2020. We used monthly MODIS NDVI data at 1 km resolution, generated using the Maximum Value Composite method, which captures crop growth dynamics through its temporal profile. Land use data (30 m resolution) for 2001–2020 were obtained from the dataset of Wuhan University in China interpreted from Landsat imagery [23]. Elevation and slope data at 500 m resolution came from the Resource and Environmental Science and Data Center, Chinese Academy of Sciences. Cultivated land area statistics were compiled from the China Land and Resources Statistical Yearbook (2002–2017, with 2018 missing) and the Ministry of Natural Resources (2019–2020), noting a methodological discontinuity after 2009 that affects comparability with pre-2008 data.
For empirical validation, we used three waves of the CRRS (2020, 2022, 2024). Conducted using a multi-stage stratified random sampling method, these surveys collectively cover 14 provinces and over 300 villages. A balanced panel of 2255 households tracked across all waves provides detailed data on farmland management—including contracted and abandoned areas—enabling robust cross-validation with our remote sensing analysis.

2.2. Identification Methodology and Criteria

This study identifies abandoned cropland through a two-stage process: (1) extracting cropland areas from land use data, and (2) calculating the Multiple Cropping Index (MCI) for each parcel using monthly NDVI data to determine cultivation status. Complete abandonment is defined as an MCI of zero for a given year, differs from land cover conversion, which involves a change in land cover type [24,25,26,27]. We focus specifically on complete abandonment, as the associated degradation of productive assets makes recultivation more difficult.
There is no global consensus on defining cropland abandonment. Aligning with Chinese legal frameworks and the CRRS standard—all of which use two consecutive years of non-cultivation as the threshold—this study defines abandoned cropland as parcels that maintain an MCI of zero for two or more consecutive years [16]. This ensures policy consistency and facilitates direct validation with household survey data.

2.3. Processing Methodology

Extracting the Multiple Cropping Index (MCI) based on changes in NDVI peaks and troughs is the most critical step in this study. A complete cropping season was defined by a ‘trough–peak–trough’ cycle in the NDVI time series. Peaks and troughs were identified as the local maxima and minima, respectively, based on the first-order difference method [28], and the number of peaks is counted as the initial MCI. To mitigate noise caused by cloud cover during image acquisition, the initial NDVI data was denoised and smoothed using the Savitzky–Golay filter (S-G filter) prior to MCI calculation [29]. Optimal fitting was achieved using a moving window of 3 observations and a 1st-order polynomial [30].
Furthermore, minor fluctuations in NDVI can lead to “pseudo-troughs” and “pseudo-peaks”. To address this, minimum thresholds for the NDVI value at a peak and for the difference in NDVI between a peak and its adjacent troughs were set to eliminate these false signals, thereby identifying “true peaks” and “true troughs” for calculating the final MCI. The specific criteria for removing pseudo-troughs and pseudo-peaks are as follows: (1) Remove a “pseudo-trough” if its NDVI difference from both adjacent peaks is <20% of the annual amplitude [28]. (2) Remove a “pseudo-peak” if its NDVI difference from either adjacent trough is <50% of the annual amplitude [28]. (3) Retain only peaks with an NDVI value ≥ 0.4 [31,32]. (4) Consider only peaks occurring between March and October [31]. (5) Classify a parcel as abandoned cropland if NDVI > 80% of the annual maximum persists for >7 months.
Identifying abandoned cropland based on the multiple cropping index extraction method (MCI) is the focus of this study. The core of the methodology involves applying a cropland mask to the MCI to exclude interference from non-cropland pixels. A cropland mask was created by resampling and reclassifying land use data, retaining pixels with >50% cropland area. Slope data were similarly resampled and classified according to standard categories (Technical Regulations for the Current Land Use Status Survey). The final identification of abandoned cropland (MCI = 0 for two consecutive years) was performed by overlaying the annual cropland mask with the MCI raster (Figure 1).

2.4. Measurement of Abandonment Extent

This study employs two complementary metrics to quantify cropland abandonment. The abandonment probability reflects the likelihood of abandonment at the pixel level, calculated as the ratio of abandoned pixels to total cropland pixels. In contrast, the abandonment rate represents the proportion of abandoned cropland area, using the cropland percentage within each pixel as a weighting factor to ensure area-based accuracy. The formulas are as follows:
F A P t = i I ( F A t i = 1 ) i I ( F A t i = 1 ) + i I ( F A t i = 0 )
F A R t = i P t i × I ( F A t i = 1 ) i P t i × I ( F A t i = 1 ) + i P t i × I ( F A t i = 0 )
F A t = F A R t × F t
In the equations: F A P t and F A R t represent cropland abandonment probability and rate in year t (2002–2020); i denotes pixel identifier; F A t i equals to 1 if pixel i was abandoned in year t, and 0 otherwise; I ( ) is an indicator function that returns 1 if the condition is true and 0 if false; P t i represents the cropland percentage within pixel i in year t ; F A t and F t denote the abandoned and total cropland area in year t , respectively.
Based on the China Rural Revitalization Survey (CRRS) data, the relevant indicators are operationalized as follows: the cropland abandonment rate is measured as the proportion of permanently abandoned cropland area to the total cropland area operated by the respondent households; the probability of cropland abandonment behavior at the household level is defined as the ratio of households that engaged in abandonment to the total number of surveyed households.

3. Results

3.1. The Overall Trend and Characteristics

3.1.1. The Cropland Abandonment Rate in China Has Shown Fluctuations, with an Average of Around 5.86%

Cropland abandonment is a rational decision made by farmers based on the principle of profit maximization, which involves the disengagement of capital from land. Research findings indicate that cropland abandonment has been widespread in China since the beginning of the 21st century, with an abandonment rate fluctuating around 5.86%, corresponding to an average of 7.6 Mha of abandoned farmland annually (Figure 2). In terms of temporal variation, the abandonment rate peaked at 7.39% in 2009, with an abandoned area of 10 Mha. In 2019, the abandonment rate was 6.83%, meaning that 8.7 Mha were abandoned out of a total cultivated area of 127.9 Mha. However, according to CRRS data, in 2023, a total of 411 households engaged in abandonment, accounting for 8.90% of all surveyed households, while the total abandoned area represented 1.32% of the total cultivated area managed by these households. Since household surveys inherently exclude migrant worker households, who are more likely to abandon land, the abandonment rate derived from survey data is lower than that identified by remote sensing. It is essential to adopt a dialectical perspective on the advantages and disadvantages of abandoning over 6.7 Mha of cropland annually. Large-scale abandonment undoubtedly represents a significant waste of capital and land resources, contradicts China’s current situation of insufficient cropland resources and heavy reliance on imported agricultural products, thereby undermining national food security. Conversely, it also implies substantial potential for increasing grain production and adjusting agricultural structure in China.

3.1.2. Reclaiming Abandoned Cropland Has Become an Important Means for Farmers to Mitigate Unemployment Risks

The ongoing economic restructuring in China is inevitably accompanied by structural unemployment, leaving migrant workers in traditional industries such as construction vulnerable to job losses. With two-thirds of migrant workers not covered by unemployment insurance [33], reclaiming abandoned cropland to compensate for income loss has emerged as a critical strategy for them to counter unemployment risks. The large-scale reclamation of abandoned cropland following the COVID-19 outbreak exemplifies how such abandoned cropland can serve social security functions, such as stabilizing employment and sustaining livelihoods. Remote sensing data analysis reveals that after the pandemic began in 2020, extensive reclamation efforts led to a decline in the abandonment rate to a 20-year low of 4.87%, with approximately 2.5 Mha of abandoned cropland being reclaimed. According to CRRS data, during the pandemic control period in 2021, the proportion of abandoned cropland in major labor-exporting provinces such as Sichuan, Shaanxi, and Guizhou was 4.79%, 3.51%, and 2.62%, respectively—lower than the pre-pandemic levels of 12.75%, 6.23%, and 17.35% in 2019, and also lower than the post-control levels of 7.36%, 8.50%, and 10.99% in 2023. Thus, in the face of unprecedented challenges and external shocks such as economic restructuring, reclaiming abandoned cropland will continue to be a vital approach for farmers to combat unemployment risks and enhance income.

3.1.3. The Abandonment Rate of Supplemented Cropland Has Consistently Exceeded That of Existing Cropland

The term “supplemented cropland” is defined in contrast to “existing cropland.” Existing cropland refers to parcels that have maintained their agricultural use over consecutive years without being converted to other land types. Supplemented cropland, on the other hand, is primarily created through the reclamation or consolidation of other land types under policies such as the “balance of cropland occupation and compensation” or the “linkage between increase in urban construction land and decrease in rural construction land.” The abandonment rate of supplemented cropland is defined as the proportion of newly supplemented cropland in a given year that is abandoned in the following year. Research shows that the abandonment rate of supplemented cropland has consistently been higher than that of existing cropland. The average annual abandonment rate for supplemented cropland is 13.84%, which is 7.92 percentage points higher than the national average. In 2019, the abandonment rate of supplemented cropland reached 11.93%, 2.47 times that of existing cropland (Figure 3). This disparity can be attributed to the insufficient emphasis on quality improvement in cropland protection policies. Supplemented cropland is often located in areas prone to waterlogging, soil degradation, poor hydrothermal conditions, or steep slopes. Typically, such cropland is supplemented in small, scattered, and fragmented parcels. Due to poor cultivation conditions and low accessibility, these areas are directly susceptible to abandonment. Furthermore, the fitted trend line of the abandonment rate for supplemented cropland shows an initial increase followed by a decrease, reflecting the evolution of China’s cropland protection policies from a focus solely on quantity to a balanced emphasis on both quantity and quality.

3.2. Spatial Patterns and Regional Heterogeneity

3.2.1. The Spatial Distribution of Abandoned Cropland in China Has Shown a Tendency to Concentrate in Lower Latitude Regions

Throughout the 21st century, abandoned cropland has been predominantly located in areas such as the Shaanxi-Gansu-Ningxia border region, the Yunnan-Guizhou Plateau in the southwest, the middle reaches of the Yangtze River, the coastal areas of South China, and the Yangtze River Delta and Pearl River Delta (Figure 4). Among these, the Yunnan-Guizhou Plateau exhibits the most extensive distribution of abandoned cropland. The Shaanxi-Gansu-Ningxia border region lies in an arid and semi-arid zone, where severe water scarcity significantly constrains agricultural production and leads to pronounced land marginalization. Although the Yunnan-Guizhou Plateau receives ample rainfall, its karst topography has poor water retention capacity, and there are few sites suitable for reservoir construction, severely impacting crop cultivation and growth and resulting in a high risk of cropland abandonment. Historically known as the “land of fish and rice,” the middle Yangtze River region, the coastal areas of South China, and the Yangtze River Delta and Pearl River Delta have high levels of economic development, providing farmers with greater access to non-agricultural employment and income. Driven by the income disparity between agricultural and non-agricultural sectors, farmers in these regions often choose to abandon their cropland. Additionally, cropland abandonment occurs in secondary key areas such as the Sichuan Basin and the Xinjiang Basin. Even the Huang-Huai-Hai Plain, a crucial grain-producing region, has witnessed cropland abandonment, mainly concentrated in the Yellow River estuary in Dongying (Shandong), the Shandong-Jiangsu border area, and central Henan. On a positive note, except for sporadic abandonment in the Greater Khingan Mountains, cropland abandonment is rare in Northeast China, which helps maintain the region’s role in safeguarding national food security.
In terms of spatial distribution changes, cropland abandonment is more severe in southern China than in the north, showing a tendency to concentrate in lower latitudes. As seen in the boxplot in Figure 5, the median latitude of abandoned cropland in China gradually shifted southward during 2002–2020. In 2002, the median latitude was 32.65° N. From 2002 to 2007, it fluctuated near the Qinling–Huaihe Line. After 2008, abandonment intensified in the south, and by 2019, the median latitude had decreased to 26.40° N—a southward shift of 6.25° compared to 2002. Furthermore, the shortening length of the boxplot’s interquartile range (IQR) and the median’s proximity to the lower quartile also indicate an increasing concentration of abandoned cropland in lower latitudes. Combining the information from Figure 4 and Figure 5, at the beginning of the study period, hotspots of abandonment were observed around 35° N and 25° N. Over time, abandonment in the Shaanxi-Gansu-Ningxia border region was effectively controlled, causing the peak near 35° N to gradually weaken and eventually disappear. Meanwhile, abandonment intensified in the middle reaches of the Yangtze River, forming a new peak around 30° N. At the same time, increased abandonment in southwestern China and the Pearl River Delta region strengthened the peak near 25° N. CRRS data confirm that cropland abandonment in southern China is significantly higher than the national average. In 2023, abandoned cropland accounted for 10.43% of total cropland area among households surveyed in Guizhou, while Zhejiang, Guangdong, and Sichuan reached 7.24%, 6.21%, and 5.54%, respectively—far exceeding the national average of 1.32%. In contrast, cropland in northeastern China (Heilongjiang, Liaoning, and eastern Inner Mongolia) was fully utilized, with abandonment rates of only 0.01%, 0.38%, and 0.25%. Henan, Shanxi, Anhui, and Shandong also had abandonment rates below 1%. China’s inherent spatial mismatch—where cropland resources are concentrated in the north, but water resources are more abundant in the south—is further exacerbated by large-scale abandonment of existing cropland in the south. This intensifies the spatial imbalance between water and cropland resources, posing challenges to national food and ecological security.

3.2.2. Higher Cropland Abandonment Rates in Grain Balance and Main Marketing Areas

As an extreme form of “non-grain production,” cropland abandonment is prevalent, but its extent varies in main production areas, production-marketing balance areas, and main marketing areas. Research findings (Figure 6) indicate that the average annual cropland abandonment rates in these three types of regions are 2.89%, 12.58%, and 14.07%, respectively. The abandonment rates in balance areas and main marketing areas are 4.35 and 4.87 times higher than that in main production areas. Such high abandonment rates intensify the pressure on achieving the national grain self-sufficiency strategy and are inevitably shifting additional pressure onto the main production areas. Although the main production areas bear the dual responsibility of ensuring regional self-sufficiency and supplying grain to the production-marketing balance and main marketing areas. These balance areas and main marketing areas should also undertake the task of achieving self-sufficiency in staple grain consumption. Specifically, balance areas should aim for self-sufficiency for their resident population, while main marketing areas should at least strive for self-sufficiency in wheat and rice for their rural resident population. Many balance areas have transitioned into marketing areas, the grain production capacity in marketing areas has been severely compromised, and main production areas face challenges such as the “inversion of grain production and financial revenue” and being “high-yield yet impoverished counties”. The concentration of food security pressure on the main production areas implies a concentrated and cumulative effect of risks from external crises and natural disasters, which undermines the country’s resilience in ensuring food security. Therefore, it is imperative to take multiple measures to continuously enhance the grain production capacity and self-sufficiency rates in marketing and balance areas, fostering a new grain security framework where main production areas serve as the primary force, with strong support from marketing and balance areas.

3.2.3. Increased Abandonment Risk with Cropland Fragmentation and Slope

Cropland fragmentation is a significant factor influencing farmers’ land use decisions. A consensus has gradually emerged in academia regarding the definition of land fragmentation, identifying small scale and dispersed layout as its most essential characteristics [34,35]. The cropland data used in this study reflect the proportion of cropland within each unit grid, which can, to some extent, serve as an indicator of fragmentation—a smaller proportion indicates a higher degree of fragmentation. The results show that as the cropland proportion decreases (i.e., fragmentation increases), cropland abandonment intensifies (Figure 7a). When the cropland proportion is between 50% and 60%, the abandonment rate reaches its highest level, with a multi-year average of 11.36%, and even peaked at 13.20% in 2009. Conversely, when the cropland proportion increases to 90–100%, the multi-year average abandonment rate drops to its lowest level of only 2.80%. To further validate this finding, we compared the annual cropland abandonment probability with the abandonment rate. Throughout the study period, the former consistently exceeded the latter by an average of 4.32 percentage points (Figure 7b). This indicates that cropland parcels most likely to be abandoned are generally small in scale, and when scale is incorporated as a weighting factor, the calculated abandonment rate is lower than the abandonment probability. Currently, the average cropland area per household in China is less than 0.53 ha, significantly lower than that of agricultural powers such as the United States and European countries. Even in land-scarce countries like South Korea and Japan, the average cropland area per household is 2.7 and 3.8 times that of China, respectively [36]. It can therefore be inferred that China’s basic national context—characterized by a large population and limited arable land—predisposes its cropland to a high risk of abandonment.
Slope, as a critical natural endowment of cropland, directly affects input-output efficiency by influencing water retention, fertilizer retention, and farming convenience, forming a fundamental basis for farmers’ economic assessments and decisions regarding abandonment. The results indicate that cropland abandonment is highly sensitive to slope (Figure 8). Flat land with a slope below 2° offers superior water and fertilizer retention and convenient farming conditions, consistently maintaining the lowest abandonment level, with an average annual rate of 4.81%, 1.06 percentage points lower than the national average. As the slope increases, the cropland abandonment rate gradually rises. Cropland with slopes between 6° and 15° is already categorized as sloped farmland, where disadvantages such as soil erosion begin to manifest. The average annual abandonment rate for such land is 6.08 percentage points higher than the national average. Sloped cropland exceeding 15° and steeply sloped cropland suffer from severe soil erosion and limited production potential. Whether due to an active withdrawal strategy driven by low agricultural returns (disengaging capital from land) or passive exit behavior prompted by external forces like the “Grain for Green” program, these areas experience the highest degree of abandonment. Their average annual abandonment rates are 17.43% and 12.74%, respectively—3.62 and 2.65 times that of flat land. This pattern is largely consistent with abandonment characteristics observed in Europe, the United States, and Japan. Japanese statistics show that even with government subsidies provided to farmers in mountainous and semi-mountainous areas, their abandonment rates remain about 3 times and 2.5 times higher than those in plain areas [37].

4. Discussion

The results regarding cropland abandonment, derived from both remote sensing data and household survey data, are largely consistent, enhancing the reliability of the research findings. Since the beginning of the 21st century, cropland abandonment has been widespread in China, with an abandonment rate fluctuating around 5.86%, corresponding to an average of 7.6 Mha abandoned annually. The reclamation of abandoned cropland emerged as a crucial livelihood strategy for returning migrant farmers during the COVID-19 pandemic. This result is supported by two consistent lines of evidence: remote sensing data revealed a 19-year low in abandonment rates in 2020 with about 2.5 Mha of abandoned cropland being reclaimed, while CRRSs confirmed that abandonment in major labor-exporting provinces was more contained during the 2021 lockdowns than in either 2019 or 2023. In terms of spatial distribution changes, cropland abandonment is more severe in southern China than in the north, showing a trend of concentration towards lower latitudes. Areas around 30° N and 25° N are hotspots for abandoned cropland. The large-scale abandonment of existing cropland in the south further exacerbates the spatial imbalance between water resources and cropland, posing challenges to national food and ecological security. Abandonment is increasingly severe for supplemented, fragmented, sloped, and steeply sloped cropland. Overall, the higher abandonment rates in grain production-marketing balance areas and main marketing areas increase the pressure on achieving China’s grain self-sufficiency strategy.
This study represents a valuable contribution to the field of cropland abandonment. Firstly, building upon traditional cropping intensity extraction techniques, it ensures the accuracy of abandoned cropland identification through a set of rigorously defined parameters. Secondly, leveraging the advantage of extensive CRRS social survey data, the research pioneers a mutual validation between remote-sensing-based and survey-based datasets, thereby significantly enhancing the credibility of its findings. Thirdly, by comparing cropland abandonment patterns before and after the COVID-19 pandemic, the study empirically verifies the function of cropland as a buffer for employment and a safety net for livelihoods. This finding provides a compelling evidence-based counterargument to the prevailing Chinese regulation that stipulates the revocation of land contract rights after two consecutive years of abandonment. Finally, utilizing 20 years of longitudinal data, the study constructs a “multi-stage temporal atlas” of cropland abandonment, which effectively reveals the spatial transition of abandonment hotspots and offers a robust foundation for formulating targeted, region-specific policies.
However, this study is subject to several limitations. The primary limitation lies in the distinction between cropland abandonment and fallowing, a well-recognized challenge. While our methodological design addressed this by implementing a two-year non-cultivation threshold—a criterion that effectively filters out single-year fallow cycles and noise while capturing sustained abandonment signals—certain ambiguities persist. It is important to note that our model does not classify managed multi-year fallowing, such as land intentionally planted with cover crops or hyperaccumulators for phytoremediation, as abandonment, as these practices maintain a distinguishable land surface state. However, interpretative challenges remain in scenarios involving two-year natural fallowing with spontaneous weed recovery and no crop cultivation, which blurs the line between abandonment and managed rest. Nonetheless, such extended natural fallow periods are uncommon under prevailing Chinese agricultural policies, which predominantly promote seasonal or single-year fallowing. Thus, we contend that our operational definition remains a robust and effective proxy for identifying cropland abandonment.
A further limitation concerns the potential for systematic underestimation inherent in both the remote sensing and household survey data. The underestimation in the survey data primarily stems from two sources: (1) the sampling frame’s inadequate coverage of migrant worker households, who demonstrate a higher propensity to abandon cropland, and (2) potential underreporting by respondents due to concerns over subsidy eligibility or local performance evaluations. Meanwhile, the underestimation in the remote sensing data originates from a technical constraint: our cropland mask extraction criterion, which retained only pixels with cropland coverage exceeding 50%, inevitably excluded highly fragmented parcels with lower cropland coverage. Our findings indicate that these excluded parcels are precisely the areas at the highest risk of abandonment. Consequently, the actual extent of cropland abandonment is likely more severe than that captured by either dataset alone. Furthermore, a key limitation is that the accuracy of remotely sensed cropland abandonment estimates is influenced by parameter settings. The vastness of China necessitates regional calibration of these parameters to achieve higher precision.
Based on this study, we have distilled four key directions for future research. The primary direction involves quantifying vegetation succession trajectories on abandoned cropland and associated ecosystem services such as carbon sequestration. The core direction focuses on revealing the transmission mechanisms through which land fragmentation drives abandonment by affecting mechanization efficiency, labor costs, and land transfer pathways. The crucial direction requires integrating behavioral economics theories to systematically investigate the socio-psychological and economic determinants of farmers’ abandonment decisions. The macro direction needs to explore the complex feedback relationships between cropland abandonment systems and rural-urban transformation, providing systematic solutions for sustainable land use.

5. Conclusions

This study systematically reveals the pervasiveness, spatiotemporal dynamics, and socioeconomic drivers of cropland abandonment in China. Based on our empirical evidence, we propose the following targeted policy recommendations.
Our spatial analysis clearly identifies southern China (around 25° N and 30° N) as hotspots of cropland abandonment. This region not only suffers from high land fragmentation but also experiences accelerated vegetation succession due to favorable hydrothermal conditions, significantly increasing the difficulty of recultivation. Governance policies should therefore prioritize these areas. We recommend revising the standards for high-standard farmland construction to explicitly include reclaimable abandoned cropland within its scope. By improving supporting infrastructure and implementing rapid soil fertility enhancement measures, production capacity can be promptly restored.
The study finds significant topographic differentiation in the drivers of abandonment. Accordingly, we propose zonal management strategies: (1) For abandoned cropland in plain areas, priority should be given to grain production to expand the grain-sown area. (2) Hilly villages, characterized by high topographic relief, face significant challenges in implementing appropriately scaled farmland management and promoting mechanization. On the basis of not destroying the topsoil, these areas should be permitted to develop specialty fruits, medicinal herbs, high-quality forage grass, and other production. This approach helps prevent cropland abandonment that would otherwise occur when farmers face the trade-off between being unable to grow cash crops and the low profitability of grain crops [38]. (3) For abandoned villages situated at high altitudes with poor farming conditions and severe labor outflow, the primary approach should be ecological restoration (returning farmland to forest or grassland). This reduces ecological risks such as soil erosion and avoids ineffective investment in “hollowed-out villages.”
The analysis of recultivation patterns during the pandemic provides direct evidence of abandoned cropland’s role as an employment buffer. This finding strongly challenges the current rigid regulation of revoking land contracts after two years of abandonment. Therefore, we propose establishing a data-driven tiered fallowing system: implementing a “buffer fallow” mechanism and exploring flexible management models such as land contract rights shareholding in population-outflow areas, while maintaining strict controls in core grain-producing areas. This balances the land’s safeguard function with production discipline.

Author Contributions

T.L. conceived and designed the structure of this paper, processed the data and wrote part of the paper. Y.W. wrote part of the literature review, results and discussion section. C.L. participated in the framework discussion and data processing. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Social Science Fund of China (grant no. 21BJY132, 24&ZD145).

Data Availability Statement

The authors do not have permission to share data.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Chen, X.W. Several important issues in current agriculture and rural areas. Chin. Rural. Econ. 2023, 2–17. [Google Scholar] [CrossRef]
  2. Li, S.F.; Li, X.B. Progress and prospect on farmland abandonment. Acta Geogr. Sin. 2016, 71, 370–389. [Google Scholar] [CrossRef]
  3. Zheng, Q.; Ha, T.; Prishchepov, A.V.; Zeng, Y.; Yin, H.; Koh, L.P. The neglected role of abandoned cropland in supporting both food security and climate change mitigation. Nat. Commun. 2023, 14, 6083. [Google Scholar] [CrossRef]
  4. Wang, Y.H.; Yang, A.X.; Yang, Q.Y. The extent, drivers and production loss of farmland abandonment in China: Evidence from a spatiotemporal analysis of farm households survey. J. Clean. Prod. 2023, 414, 137772. [Google Scholar] [CrossRef]
  5. Han, Z.; Song, W. Abandoned cropland: Patterns and determinants within the Guangxi Karst Mountainous Area, China. Appl. Geogr. 2020, 122, 102245. [Google Scholar] [CrossRef]
  6. Shi, T.C.; Li, X.B.; Xin, L.J.; Xu, X. The spatial distribution of farmland abandonment and its influential factors at the township level: A case study in the mountainous area of China. Land Use Policy 2018, 70, 510–520. [Google Scholar] [CrossRef]
  7. Li, S.F.; Li, X.B.; Xin, L.J.; Tan, M.; Wang, X.; Wang, R.; Jiang, M.; Wang, Y. Extent and distribution of cropland abandonment in Chinese mountainous areas. Resour. Sci. 2017, 39, 1801–1811. [Google Scholar] [CrossRef]
  8. Li, S.F.; Li, X.B. Economic characteristics and the mechanism of farmland marginalization in mountainous areas of China. Acta Geogr. Sin. 2018, 73, 803–817. [Google Scholar] [CrossRef]
  9. He, Y.F.; Xie, H.L.; Peng, C.Z. Analyzing the behavioural mechanism of farmland abandonment in the hilly mountainous areas in China from the perspective of farming household diversity. Land Use Policy 2020, 99, 104826. [Google Scholar] [CrossRef]
  10. Xiao, G.F.; Zhu, X.F.; Hou, C.Y.; Xia, X. Extraction and analysis of abandoned farmland: A case study of Qingyun and Wudi counties in Shandong province. Acta Geogr. Sin. 2018, 73, 1658–1673. [Google Scholar] [CrossRef]
  11. Luo, B.L.; Wan, Y.L.; Hong, W.J.; Zhong, W. Fragmentation, service outsourcing and farmland abandonment: An empirical analysis based on questionnaires from 2704 households in 9 provinces. Econ. Rev. J. 2019, 63–73. [Google Scholar] [CrossRef]
  12. United Nations, Department of Economic and Social Affairs, Population Division. World Urbanization Prospects: The 2018 Revision; United Nations: New York, NY, USA, 2019. [Google Scholar]
  13. Jin, X. An economic perspective on the abandonment of new farmland. Rural. Econ. 2013, 25–26. [Google Scholar]
  14. Hu, M.; Wang, C.C. Impacts of rural labor’s non-agricultural transfer on cropland abandonment. J. Subtrop. Resour. Environ. 2013, 8, 56–63. [Google Scholar] [CrossRef]
  15. Wang, K.X.; Xu, Y.J. Institutional constraints and policy relief of land abandonment. New Heights 2017, 36, 55–59. [Google Scholar]
  16. Shao, J.A.; Zhang, S.C.; Li, X.B. The role of rural farmland transfer in preventing farmland abandonment in the mountainous areas. Acta Geogr. Sin. 2015, 70, 636–664. [Google Scholar] [CrossRef]
  17. Chen, H.; Tan, Y.Z.; Xiao, W.; He, T.; Xu, S.; Meng, F.; Li, X.; Xiong, W. Assessment of continuity and efficiency of complemented cropland use in China for the past 20 years: A perspective of cropland abandonment. J. Clean. Prod. 2023, 388, 135987. [Google Scholar] [CrossRef]
  18. Kurganova, I.; De, G.V.L.; Six, J.; Kuzyakov, Y. Carbon cost of collective farming collapse in Russia. Glob. Change Biol. 2014, 20, 938–947. [Google Scholar] [CrossRef]
  19. He, Y.; Amintas, B.; Johanna, B.; Helmers, D.; Iuliano, B.G.; Kimambo, N.E.; Lewińska, K.E.; Razenkova, E.; Rizayeva, A.; Rogova, N.; et al. Radel off, Monitoring cropland abandonment with Landsat time series. Remote Sens. Environ. 2020, 246, 111873. [Google Scholar] [CrossRef]
  20. Victor, M.O.; Rasmus, F.; Pontus, O.; Bonifacio, R.; Butsic, V.; Druce, D.; Ray, D.; Prishchepov, A.V. The impact of conflict-driven cropland abandonment on food insecurity in South Sudan revealed using satellite remote sensing. Nat. Food 2021, 2, 990–996. [Google Scholar] [CrossRef]
  21. Ma, Y.; Lyu, D.M.; Sun, K.; Li, S.; Zhu, B.; Zhao, R.; Zheng, M.; Song, K. Spatiotemporal Analysis and War Impact Assessment of Agricultural Land in Ukraine Using RS and GIS Technology. Land 2022, 11, 1810. [Google Scholar] [CrossRef]
  22. Gergana, N.D.; Johannes, K. Abandoning land transforms biodiversity. Science 2023, 380, 581–583. [Google Scholar] [CrossRef] [PubMed]
  23. 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]
  24. Li, S.Z.; Xiao, J.; Lei, X.Y.; Wang, Y. Farmland abandonment in the mountainous areas from an ecological restoration perspective: A case study of Chongqing, China. Ecol. Indic. 2023, 153, 110–124. [Google Scholar] [CrossRef]
  25. Guo, A.D.; Yue, W.Z.; Yang, J.; Xue, B.; Xiao, W.; Li, M.; He, T.; Zhang, M.; Jin, X.; Zhou, Q. Cropland abandonment in China: Patterns, drivers, and implications for food security. J. Clean. Prod. 2023, 418, 138154. [Google Scholar] [CrossRef]
  26. Liu, B.; Song, W. Mapping abandoned cropland using Within-Year Sentinel-2 time series. Catena 2023, 223, 106924. [Google Scholar] [CrossRef]
  27. Ma, Y.C.; Wang, J.S.; Xiong, J.H.; Sun, M.; Wang, J. Risk assessment for cropland abandonment in mountainous area based on AHP and PCA—Take Yunnan province in China as an example. Ecol. Indic. 2024, 158, 111287. [Google Scholar] [CrossRef]
  28. Zhu, X.L.; Li, Q.; Shen, M.G.; Chen, J.; Wu, J. A methodology for multiple cropping index extraction based on NDVI time-series. J. Nat. Resour. 2008, 23, 534–544. [Google Scholar]
  29. Savitzky, A.; Golay, M. Smoothing and differentiation of data by simplified least squares procedures. Anal. Chem. 1964, 36, 1627–1639. [Google Scholar] [CrossRef]
  30. Liu, L.; Xiao, X.M.; Qin, Y.W.; Wang, J.; Xu, X.; Hu, Y.; Qiao, Z. Mapping cropping intensity in China using time series Landsat and Sentinel-2 images and Google Earth Engine. Remote Sens. Environ. 2020, 239, 111624. [Google Scholar] [CrossRef]
  31. Xu, Q.Y.; Yang, G.J.; Long, H.L.; Wang, C.; Li, X.; Huang, D. Crop information identification based on MODIS NDVI time-series data. Trans. Chin. Soc. Agric. Eng. 2014, 30, 134–144. [Google Scholar]
  32. Fan, J.L.; Wu, B.F. A methodology for retrieving cropping index from NDVI profile. J. Remote Sens. 2004, 8, 628–636. [Google Scholar] [CrossRef]
  33. Wang, Z. Impact of COVID-19 on China’s employment protection and social security system. Econ. Rev. J. 2020, 7–15. [Google Scholar] [CrossRef]
  34. Zhao, K. Land Fragmentation and its quantitative measurement method. China Land Sci. 2011, 25, 35–39+88. [Google Scholar] [CrossRef]
  35. Hu, X. Effects of Land Fragmentation on Agricultural Production in Guangxi by Empirical Study; Guangxi University: Nanning, China, 2016. [Google Scholar]
  36. Cai, J.M. How to firmly maintain the red line of 1.8 billion mu of farmland. Chin. Cadres Trib. 2022, 43–48. [Google Scholar] [CrossRef]
  37. Ministry of Agriculture, Forestry and Fisheries, Japan. The Current State of Abandoned Farmland in Japan; Ministry of Agriculture, Forestry and Fisheries: Tokyo, Japan, 2011.
  38. Xu, Q.; Yin, R.L.; Zhang, H. Economies of Scale, Returns to Scale and the Problem of Optimum-scale Farm Management: An Empirical Study Based on Grain Production in China. Econ. Res. J. 2011, 46, 59–71+94. [Google Scholar]
Figure 1. Flow chart of data processing. * PT and PP represent “pseudo-troughs” and “pseudo-peaks”, respectively; TT and TP represent “true troughs” and “true peaks”, respectively; M C T t and M C T t 1 represent the multiple cropping index in year t and t–1, respectively; V represents the annual amplitude of NDVI; D t b p represents the difference in NDVI between the trough and the preceding peak; D t a p represents the NDVI difference between the trough and the subsequent peak; D p b t represents the NDVI difference between the peak and the preceding trough; D p a t represents the NDVI difference between the peak and the subsequent trough; N D V I p and M p represent the NDVI value and the occurrence month of the peak, respectively; Max represents the annual maximum value.
Figure 1. Flow chart of data processing. * PT and PP represent “pseudo-troughs” and “pseudo-peaks”, respectively; TT and TP represent “true troughs” and “true peaks”, respectively; M C T t and M C T t 1 represent the multiple cropping index in year t and t–1, respectively; V represents the annual amplitude of NDVI; D t b p represents the difference in NDVI between the trough and the preceding peak; D t a p represents the NDVI difference between the trough and the subsequent peak; D p b t represents the NDVI difference between the peak and the preceding trough; D p a t represents the NDVI difference between the peak and the subsequent trough; N D V I p and M p represent the NDVI value and the occurrence month of the peak, respectively; Max represents the annual maximum value.
Land 14 02203 g001
Figure 2. The rate and area of abandoned farmland in China from 2002 to 2020. The dashed line represents an average cropland abandonment rate of around 5.86%.
Figure 2. The rate and area of abandoned farmland in China from 2002 to 2020. The dashed line represents an average cropland abandonment rate of around 5.86%.
Land 14 02203 g002
Figure 3. The abandonment rate and abandonment area of supplementary farmland and existing farmland from 2002 to 2020.
Figure 3. The abandonment rate and abandonment area of supplementary farmland and existing farmland from 2002 to 2020.
Land 14 02203 g003
Figure 4. Kernel density distribution of abandoned farmland in China from 2002 to 2020.
Figure 4. Kernel density distribution of abandoned farmland in China from 2002 to 2020.
Land 14 02203 g004
Figure 5. Density distribution and box plot of abandoned farmland in latitude direction in China from 2002 to 2020. The line plots the annual median of the latitudinal distribution of cropland abandonment across the years.
Figure 5. Density distribution and box plot of abandoned farmland in latitude direction in China from 2002 to 2020. The line plots the annual median of the latitudinal distribution of cropland abandonment across the years.
Land 14 02203 g005
Figure 6. The average abandonment rate of farmland in China’s main production areas, production—marketing balance areas, and main marketing areas from 2002 to 2020.
Figure 6. The average abandonment rate of farmland in China’s main production areas, production—marketing balance areas, and main marketing areas from 2002 to 2020.
Land 14 02203 g006
Figure 7. Relationship between the degree of fragmentation of farmland and abandonment: (a) The FAR in different proportion intervals of farmland in key years; (b) Comparison of FAP and FAR from 2002 to 2020.
Figure 7. Relationship between the degree of fragmentation of farmland and abandonment: (a) The FAR in different proportion intervals of farmland in key years; (b) Comparison of FAP and FAR from 2002 to 2020.
Land 14 02203 g007
Figure 8. The trend of cropland abandonment rate versus slope gradient.
Figure 8. The trend of cropland abandonment rate versus slope gradient.
Land 14 02203 g008
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

Li, T.; Liu, C.; Wang, Y. Continuous Monitoring of Cropland Abandonment in China Since the 21st Century: Interpreting Spatiotemporal Trajectories and Characteristics. Land 2025, 14, 2203. https://doi.org/10.3390/land14112203

AMA Style

Li T, Liu C, Wang Y. Continuous Monitoring of Cropland Abandonment in China Since the 21st Century: Interpreting Spatiotemporal Trajectories and Characteristics. Land. 2025; 14(11):2203. https://doi.org/10.3390/land14112203

Chicago/Turabian Style

Li, Tingting, Changquan Liu, and Yanfei Wang. 2025. "Continuous Monitoring of Cropland Abandonment in China Since the 21st Century: Interpreting Spatiotemporal Trajectories and Characteristics" Land 14, no. 11: 2203. https://doi.org/10.3390/land14112203

APA Style

Li, T., Liu, C., & Wang, Y. (2025). Continuous Monitoring of Cropland Abandonment in China Since the 21st Century: Interpreting Spatiotemporal Trajectories and Characteristics. Land, 14(11), 2203. https://doi.org/10.3390/land14112203

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

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.
Back to TopTop