Land use optimization plays pivotal roles in bridging regional sustainable development with national strategic priorities, yet persistent methodological disconnects between diagnostic evaluation and prescriptive optimization hinder practical implementation. Addressing this critical gap, this research develops an integrated “Evaluation-Diagnosis-Simulation-Optimization” (EDSO) framework through synergistic coupling of multi-dimensional functional indices, PLUS modeling, and random forest. The framework utilizes multi-source geospatial data with Lingbao City, Henan Province, China, as the case study area. The results indicate that the transition marked by 1.40% farmland loss due to forest regeneration and urban-rural construction expansion, while non-agricultural output value surged by 23.09% alongside improved infrastructure accessibility (2012–2020). Rural depopulation paralleled the degradation of commercial and public services, exposing vulnerabilities in rural governance. The simulation demonstrates the comprehensive optimized scenario's superiority in landscape efficiency, with 280 million CNY higher ecological value than alternatives. The optimized schedule identified 200.22 km² of land returned to forest and 4.32 km² of fragmented rural low-efficiency built-up land. The schedule identified forest carbon sink peaking at 35-year tree growth cycles, critical insights informing 30m-resolution rotation scheme. The EDSO framework's ability to integrate diagnostic analytics with spatially adaptive interventions offers a transformative toolkit for territorial planning, effectively bridging the science-policy gap.

Rural-urban income inequality (RUIE) remains a persistent challenge across many rapidly developing regions, undermining socio-economic welfare and hindering sustainable development. The ecological transition of cropland use (ETCU), a management paradigm rooted in the principles of sustainable intensification, is widely advocated to enhance agricultural sustainability. While localized case studies suggest that such practices can bolster rural incomes through improved resource efficiency or product diversification, robust empirical evidence on whether large‑scale ETCU adoption mitigates regional RUIE remains scarce, and the underlying systemic mechanisms are poorly understood. This study develops a multi-scale "cropland-agriculture-rural" framework to analyze how ETCU influences RUIE in China's intensively cultivated Huang-Huai-Hai Region. By integrating plot-scale biogeochemical modeling (PEST-DNDC), county-level panel data (2000–2020), and Structural Equation Modeling, we systematically quantify the spatiotemporal patterns and causal pathways of this transition. The results demonstrate that: (1) ETCU and RUIE exhibit spatially heterogeneous co-evolution; improvements in ETCU are not consistently associated with reduced inequality, and in some coastal counties, they even coincide with widening gaps. (2) ETCU affects RUIE primarily by restructuring agricultural multifunctionality, which serves as the core mediator. (3) The effects are highly dimension-specific. Enhancing resource-use efficiency exerts a modest yet significant direct mitigating effect on RUIE. Conversely, improvements in productivity and environmental performance exhibit no direct effect; instead, they indirectly exacerbate inequality by weakening the agricultural product supply function—which our model identifies as a strong and direct suppressor of RUIE. The economic value-added and ecological conservation functions do not exhibit significant direct equalizing effects. These findings indicate that the equity outcomes of ETCU are not automatic but depend critically on how it reshapes the multifunctional structure of agriculture. Policy interventions must therefore adopt an integrated approach that strategically guides the agricultural transition toward balanced multifunctionality, safeguards diversified production, and ensures that efficiency gains translate into broadly shared benefits for both rural and urban households.

The man-land relationship and its dynamics were core themes in modern geography. Complex interactions exist between the natural environment and human activities. However, current research focused on the impacts of human activities on ecosystems, ignoring the linear or non-linear characteristics and interaction thresholds. In this study, multi-source data and self-organizing feature map (SOFM) were used to classify man-land relationship areal types in southern Shaanxi, China from 2000 to 2020. The constrain lines method was employed to identify the interaction thresholds between human activities intensity (HAI) and various natural indicators. Results showed that: 1) higher and high natural environment index (NEI) accounted for more than 75% of the southern Shaanxi. 2) Low and lower HAI covered over 85% of the total area, with significant decreases and significant increases in HAI occurring in 45.61% and 24.53% of the study area, respectively. 3) The agricultural type decreased in southern Shaanxi, while the urban and comprehensive type increased. 4) There was a negative convex relationship between HAI and NEI. Thresholds for the effect of HAI on habitat quality were 0.67, 0.37, and 0.64 in the Qinling Mountains, Hanjiang Basin and Daba Mountains, respectively. This study emphasizes that one or more thresholds should be considered for ecological restoration and human management in different regions, and provides a new perspective for understanding the complex man-land relationship in underdeveloped mountainous regions and offered strong support for coordinating ecological protection with socio-economic development.

Balancing food security with environmental sustainability represents a fundamental challenge for global agriculture, necessitating a comprehensive assessment of cropland use sustainability to inform science-based decision-making. Current methodologies present two main limitations: while integrated indicator frameworks are effective for diagnosing agricultural development patterns, their analytical perspectives tend to be static, and they are often constrained by data limitations in capturing key parameters of agroecosystem processes. Conversely, process-based models excel at simulating detailed agroecosystem dynamics but seldom translate these simulations into spatially differentiated management strategies. To address this gap, this study developed an integrated framework coupling process-based simulation with diagnostic assessment, centered on a dynamic, multidimensional Ecological Cropland Use Index (ECUI) that synthesizes productivity, environmental impact, and resource use efficiency into a composite sustainability metric. We applied this framework to China's Huang-Huai-Hai region, a critical grain-producing area facing acute sustainability pressures, using outputs from a regionally calibrated PEST-DNDC model. Our analysis from 2001 to 2020 indicates that regional grain output increased by 128.1%, yet this gain coincided with a fundamental shift from a net carbon sink to a persistent net source, with croplands incurring an annual SOC loss of 79.88 Tg C in 2020, alongside rising greenhouse gas emissions. Spatially, over 70% of croplands experienced ECUI degradation, with high-sustainability zones shifting from traditional agricultural cores to the Bohai Rim. Key systemic constraints include low nitrogen-use efficiency, soil carbon decline, and suboptimal yields. Based on diagnostic analyses, we propose targeted optimization strategies for identified constraint zones. This study demonstrates how linking mechanistic modeling with diagnostic indices can elucidate sustainability trade-offs and generate actionable management insights. The framework offers a transferable pathway for guiding the sustainable intensification of agricultural systems worldwide.

Prolonged and unidirectional rural-to-urban migration has intensified rural hollowing and led to the emergence of numerous "hollow villages", posing a significant threat to the sustainability of rural development. However, the complexity of human-earth coupling makes it difficult to achieve a fine-scale assessment of this phenomenon at the national scale. To address this gap, a bottom-up and region-adapted assessment framework based on human-earth system theory and multiple open-access data sources is proposed to assess the dynamic patterns of rural settlement (RUS) hollowing risk in China over the past two decades. The results show that the vitality of RUS in China declined from 4.12 in 2000 to 3.12 in 2020, with approximately 50 % of RUS concentrating 80 % of the rural population. The area of rural residential land with high potential for land consolidation increased from 74,109.21 km2 in 2010 to 95,420.31 km2 in 2020, primarily concentrated in the Huang-Huai-Hai Plain and the Northeast Plain. The intensification of RUS hollowing risk is mainly driven by the spatial spillover effects of urbanization and the level of socioeconomic development within counties. By leveraging publicly available data with a cost-effective assessment framework, this study not only provides scientific support for spatially refined rural planning and governance but also offers policy implications for promoting rational county urbanization under China's evolving development context.

Aboveground biomass (AGB) is a critical indicator for assessing carbon sequestration and ecosystem health in transboundary ecologically fragile areas. High-precision estimation and spatiotemporal inversion of AGB are the key to investigating transition zones. However, inadequate feature selection and complex parameter tuning limit accuracy and spatiotemporal representation in the estimation model. An AGB estimation model that integrates SHAP-based feature selection with a particle swarm optimization-enhanced random forest model (RF_PSO) was proposed. Then AGB trajectory clustering was used to characterize the grassland change pattern. The method was applied to grasslands across the China–Mongolia–Russia (CMR) border area from 2000 to 2020. The results show that (1) the SHAP-RF_PSO model achieved the highest accuracy (R2 = 0.87, RMSE = 45.8 g/m2), outperforming other estimation models. (2) AGB improvements were observed in 72.13% of the area, mainly in MN_EA, MN_CE, and CN_NMG, while 27.39% showed degradation, concentrated in CN_NMG and MN_CE. The stable area accounts for 0.48%, which is scattered in RU_BU and RU_ZA.CN_NMG. (3) Four change patterns, namely Fluctuating Low, Stable Low, Fluctuating High, and Stable High, were identified, with major shifts in 2007, 2012, and 2014. (4) Projections indicate that 80% of the region may maintain current trends, 13% may reverse, and 7% remain uncertain, requiring targeted interventions. This study offers a robust tool for high-precision AGB estimation and supports dynamic monitoring in the CMR border area.

Plastic-covered greenhouse (PCG) is widely used in agricultural production due to its temperature control, water conservation, and wind protection characteristics, significantly enhancing crop yields and economic benefits. However, its long-term and extensive use can lead to environmental issues, such as the accumulation of local toxic gases and the degradation of soil physicochemical properties. Therefore, obtaining a comprehensive distribution of PCGs is essential. To monitor PCGs on a large scale, this study developed a novel approach for producing the first global 10 m PCG dataset (Global-PCG-10) with high-quality. Firstly, the globe was divided into multiple 5 degrees grids, and grids for classification were organized based on global cropland layer. Then, multi-temporal Sentinel-2 data and initial labels of PCGs were obtained through Google Earth Engine (GEE) to create a training set for deep learning. Next, initial labels were optimized with the active learning strategy combined with the deep learning model, APC-Net. Finally, the PCG classification results were predicted, spatially analyzed, and compared with publicly released land use and land cover (LULC) datasets. Experimental results indicate that the proposed Global-PCG-10 dataset (Niu et al., 2024) has a high overall accuracy of 98.04 % +/- 0.12 %. The global area of PCGs is 14 259.85 km(2), and 69.24 % of PCGs are located in Asia, covering around 9874.51 km(2). China has the largest PCG area of 8224.90 km(2), accounting for 57.67 % of the globe and 83.29 % of Asia. Comparisons with other LULC datasets revealed that PCGs, which should be classified as cropland, are often misclassified as bareland, impervious surfaces, ice/snow, etc.

The increasing public availability of long-term remote sensing imagery, combined with advancements in algorithms and cloud computing, has enabled the development of large-scale cropland mapping datasets at national and global levels. These datasets are critical for assessing the ecological and environmental risks of agricultural production in the Anthropocene era. While previous studies have primarily focused on improving the accuracy of cropland mapping, the implications of dataset discrepancies for evaluating cropland-induced ecological impacts remain insufficiently explored. Here, we integrate seven publicly available cropland datasets covering China and systematically analyze their spatiotemporal discrepancies, along with associated uncertainties in assessing habitat and biodiversity impacts. Our analysis incorporates key conservation areas, protected zones, species richness maps, and Dynamic Habitat Indices time series data. The results indicate that although all datasets indicate a general stabilization of cropland area in China over recent decades, significant discrepancies persist—particularly at finer spatial scales and in ecologically sensitive regions. These discrepancies not only influence the diagnosis of spatial conflicts between cropland and habitat areas but may also result in an underestimation of the threats posed by cropland expansion to habitats of threatened species. Furthermore, regions with greater cropland mapping discrepancies tend to exhibit lower environmental stability. Our findings highlight the need for selecting regionally appropriate cropland datasets when assessing historical and current cropland distributions and their ecological consequences. This study also offers insights into how remote sensing technologies can be better leveraged for ecological and agricultural sustainability research.

The dramatic expansion of agricultural greenhouses (AGs) in China has raised concerns about its environmental impacts. But our knowledge in this area is still limited, especially the temporal and scale impacts of AGs. To fill this gap, we utilized multiple remote sensing data, time-series segmentation algorithm, and the Mann-Kendall test to analyze the spatiotemporal evolution patterns of AGs in Shandong Province, China from 2001 to 2018. We then explored the impact of AGs on albedo, land surface temperature (LST), and Enhanced vegetation index (EVI) from annual and seasonal analysis perspectives. Results indicated that the total area and number of patches of AGs in Shandong initially grew, then declined, and subsequently grew again. Smaller AGs areas showed lower spatial aggregation and survival rates. Additionally, this study found that AGs have a significant impact on the environment. Increased spatial concentration and longer durations of AGs were linked to more significant reductions in albedo and EVI, along with more pronounced increases in LST. In summer and spring, AGs significantly boosted LST, while in autumn and winter, they significantly reduced albedo. AGs play a crucial role in supporting crop growth during autumn and winter. Moreover, the paper proposed several sustainable AGs management strategies to address these challenges. This study provides observational evidence of the environmental impacts of AGs for promoting sustainable agriculture.

Rural settlements (RUS) serve as spatial carriers for both residential and industrial activities in rural regions, reflecting the regional cultural and productive paradigms of rural communities. These settlements are undergoing substantial transformations driven by rapid urbanization, yet the patterns of evolution and underlying driving mechanisms of RUS remain unclear. In this study, we utilized 30-m resolution RUS data from 1990 to 2020 to investigate the morphological evolution of settlements in the Beijing-Tianjin-Hebei (BTH) region across regional, county, and urban-rural gradients. The impact of urbanization on RUS morphology was assessed using landscape pattern indices and structural equation modeling (SEM). Our results showed that the number of RUS increased by 1308 and the area expanded by 2559.9 km2, representing growths of 2.76% and 19.66%, respectively. RUS in the vicinity of Beijing and Tianjin exhibited a higher rate of morphological change, highlighting the significant influence of location on settlement evolution. A comparative analysis across urban-rural gradients indicated that RUS near urban center have experienced an increase in average size, accompanied by a decline in total size and number. This evolution has led to a shift toward more complex morphology and a more concentrated distribution. Cross-validation of the SEM across various sample sizes confirmed its reliability, demonstrating that urbanization's direct impact on RUS morphology is generally negative and insignificant, while its indirect effects through socio-economic, environmental, and infrastructural mediators are positive. This study provides valuable insights into the evolutionary characteristics of RUS across urban-rural gradients and clarifies the mechanisms through which urbanization shapes their morphological evolution, contributing to our understanding of rural transformation under rapid urbanization.

The world has experienced a rapid expansion of human settlements in both urban and rural areas in recent decades, yet the unequal impacts of this construction on global food security remain unclear. In this study, we delineated the global-scale expansion of urban-rural settlements at a fine resolution from 1985 to 2020 and quantified their uneven impacts on food security, focusing on the relationships between settlement types, cropland categories, and disparities in crop production. Our results showed that despite dramatic urbanization, rural settlements still constituted the majority of human settlement areas in 2020. Globally, cropland loss due to the expansion of rural settlements was 1.2 times greater than that caused by urbanization, while the associated yield loss was 1.5 times higher. Notably, urban-rural settlement expansion in Asia accounted for 61% of cropland loss and 64% of yield loss. Moreover, future scenarios predicted that Asia's urban-rural settlement expansion will continue to have the most significant impacts on the loss of cropland and yield throughout the 2030s. These results provide systematic evidence of the unequal impacts of urban-rural settlement construction on global cropland and food security.

Terrestrial ecosystem carbon stock (TECS) is critical to socioeconomic development and ecosystem services and is jointly affected by land use and cover and climate change. However, the dynamics of long-term annual TECS levels in urban agglomeration remain largely unknown, and research mostly ignores the spatial heterogeneity of climate factors, compromising sustainable environmental management and land planning strategies. To this end, we integrated field observations of carbon density, land use, and climate factors to map the annual distribution of TECS and analyzed their spatiotemporal variations and policy implications in the urban agglomeration of the middle Yangtze River Basin in China from 1990 to 2020. The results showed that 43,855.47 km2 of the land of the urban agglomeration changed from 1990 to 2020, accounting for 12.54% of the study area. The farmland and forest land area fluctuated and reduced, and the construction land area increased significantly. The increase in construction land was mainly from farmland and forest land. The TECS in urban agglomerations underwent a remarkable change, the overall trend fluctuated downward, and the maximum interannual variation was 1560 Tg. The transfer of construction land, farmland, forest land, shrubs, grassland, and other land mainly caused the change in carbon storage. Due to abnormal climate change, the urban agglomeration in some areas illustrated carbon storage with a spatially aggregated distribution. When considering the impact of climate change on carbon density, the TECS changes of land types other than forest land were found to be consistent with the area change but more significant due to climate change. The research results can provide reference data for regional land management policy formulation and realization of “dual carbon” goals.

The human-earth system (HES) is a complex adaptive system that embodies the interconnection and interaction between human activities and the geographical environment, characterized by a variety of features such as comprehensiveness, regionality, complexity, openness, and dynamics. Socio-economic development is facing increasingly serious regional problems such as land degradation, environmental pollution, and biodiversity loss. In essence, it is the result of the intensified impact of human activities on the earth system, resulting in the coupling disorder and functional imbalance of HES. Therefore, the spatial identification, type diagnosis, and intensity evaluation of HES are critical topics in modern HES science and comprehensive geography. These themes form the foundation for scientifically cognizing the evolution process and mechanisms of HES, as well as for supporting decision-making aimed at HES coordination and sustainable development. Based on the theory of human-earth relationship areal system and the geographic methodology of “main function-oriented zoning, dominant type classification, and principal purpose grading”, this study constructed a top-down scientific cognition and method system for modern HES identification-diagnosis-evaluation. The study integrated multi-source spatial data of land use, population density, nighttime light index, and point of interest, adopted multiple quantitative methods of decision tree, spatial clustering, and human footprint intensity index, and conducted the geospatial identification, dominant type diagnosis, activity intensity evaluation of the HES in China from 2000 to 2020. The results indicated that: (1) The area proportion of China’s HES has increased from 53.9% to 54.1%, stabilizing at around 54%. Spatially, it was characterized by a differentiation feature of high-values in the southeast and low-values in the northwest. The HES proportion in plain areas, cultivated land, and concentrated urban areas was significantly higher than that in mountainous and sparsely populated areas. The proportion of living functional zone in the HES increased from 3.61% to 5.24%, and the proportions of production and ecological functional zones decreased from 35.19% and 61.20% to 34.66% and 60.10%, respectively. (2) Rapid urbanization and rural revitalization have resulted in an increase in the area of urban HES and rural HES by 135.45% and 9.59%, respectively. However, these increases were mostly from agricultural HES. The agricultural HES and ecological HES have been affected by the Grain for Green Project and cultivated land expansion, resulting in mutual transformation between the two types and the decrease of 1.06% and 1.37%. (3) The human footprint intensity of China’s HES has increased from 9.28 to 10.25, with an increase of 10%. This increase was characterized by the expansion of high-value areas and the reduction of low-value areas, indicating the growing and spatial clustering of human activities. The findings of this study have provided hierarchical answers to key questions such as where are the distributions, what are the types, and what are the grades for HES. The scientific cognition and detection methods of modern HES can deepen the scientific understanding of the coupling process, mechanism, and pattern of HES, and support decision-making for HES coordination and sustainable development.

The systematic decline of rural areas in the process of rapid urbanization has become a global trend, creating greater challenges for sustainable rural development. As the spatial projection of socio-economic development and living environment in rural areas, the continuous tracking of rural settlements (RUS) is crucial to quantify the imbalance of rural development. However, consistent information on RUS is highly needed but is quite deficient in current research. In this study, a cost-effective mapping model was proposed to produce an annual RUS dataset in the rapid urbanization region of Beijing-Tianjin-Hebei (BTH) in North China during 1990-2020, and the temporal-spatial regularity of RUS changes was further analyzed. The location-based and the area-based comparison verified the effectiveness of our model, with a mean overall accuracy of 85% and a mean correlation value of 0.88, respectively. The total area of RUS in the BTH region increased by 2561 km2 from 1990 to 2020, while the average size of RUS remained stable after 2005. The annual change trends in RUS appeared with increasing and decreasing accounting for 76.33% and 23.67%, respectively. The centroids of RUS in Tianjin and Hebei have moved closer to Beijing, while those in Beijing have moved away from the former. Notably, we have identified 56.3% counties in the BTH region belong to the "Convex-I" change type in RUS. In general, our work can help to consistently quantify the spatiotemporal patterns of RUS in a cost-effective way, providing more explicit spatial information and continuous temporal information for rural residential land management.

通过引入多源多时相卫星遥感数据,提出了一种基于多核主动学习的农田塑料覆被分类算法,实现农业塑料大棚和地膜的精准分类.首先基于多时相Sentinel-1雷达和Sentinel-2光学遥感影像,提取其光谱特征、纹理特征等,以构建多维特征空间.然后构建多核学习模型,实现多源、多时相特征的自适应融合.最后构建基于池的主动学习策略,通过引入训练样本的淘汰机制,进一步提升分类模型的泛化能力.试验结果表明,本文所提分类方法的总体精度为95.6％,Kappa系数为0.922,相较经典支持向量机、随机森林、K近邻、决策树、AdaBoost模型,多核学习模型精度提高5.7、12.1、11.4、22.3、10.3个百分点;且在相同分类精度下,主动学习较被动学习可减少一半以上的标签数据;同时相较仅使用单时相及单传感器遥感影像而言,精度分别提高3.7、12.7个百分点.结果 表明,多核主动学习能够有效进行多传感器、多时相数据融合,并可以在小样本条件下取得更高的分类精度,从而为农田塑料覆被的遥感监测提供模型参考.

Rapid urbanization and economic development have led the diversified food production and consumption. In this context, as a highly efficient and intensive cultivated land use form, Greenhouse-led cultivated land (GCL) has continuously increased in recent decades worldwide. Previously works have documented the irrational expansion of GCL in challenging the ecological environment and sustainable agricultural development. However, these studies either have been short-term and point-based studies or have not revealed the long-term causes, process and patterns in a large-scale. In this study, long-term annual remote sensing-based and statistical data were used to investigate the spatiotemporal dynamics of GCL and its drivers in Shandong province, China from 1989 to 2018. The results showed that: 1) GCL in Shandong was toward continuous clustering dominated by medium-low and medium densities, showing the same trend as the increase of its total area; 2) GCL with a cumulative duration of more than 15 years and a demolition frequency of less than 0.2 were mainly distributed in the industrial clustering regions and roughly formed a circular expansion pattern around the central mountainous area with the most expansion period appeared in the mid-2010’s; 3) Budget expenditure for rural development, local retail sales and average earnings of local farmers were the most important local driving factors of the GCL expansion in Shandong. 4) The competition of external vegetable supply and the consumption demand from Beijing were the main external driving forces of the expansion of GCL in Shandong. These findings can enhance the comprehensive understanding of typical component of “Human-Nature” interaction and support the sustainable development of regional agriculture.

• A deep learning model is proposed for crop mapping from UAV hyperspectral data. • The model could make full use of spatial and spectral features simultaneously. • The model yields an accuracy of 86% and a Kappa index of 0.8347. • The dataset, UAV-HSI-Crop, has been released to promote future studies. UAV hyperspectral imagery (HSI) has the unique merits of both a very high spatial and spectral resolution, which provides a high-quality data source for automatic crop mapping. Recently, deep learning has been widely used in crop classification, however, the design of an accurate crop mapping model for HSI data still remains a challenging task. Therefore, this paper aims to propose a novel semantic segmentation model (HSI-TransUNet) for crop mapping, which could make full use of the abundant spatial and spectral information of UAV HSI data simultaneously. Specifically, the proposed HSI-TransUNet belongs to an improved version of TransUNet, and we have made four important modifications for HSI data. Firstly, a spectral-feature attention module is designed for spectral features aggregation in the encoder. Afterwards, a series of Transformer layers with residual connections are designed to learn global contextual features. In the decoder part, sub-pixel convolutions are adopted to avoid the chess-board effect in the segmentation results. Finally, we design a hybrid loss function to further refine the predictions for boundaries. Experiment results indicate that the proposed HSI-TransUNet has achieved good performance in crops identification with an overall accuracy of 86.05%. Ablation studies have been conducted to verify the effectiveness of each refined module in the HSI-TransUNet. Comparison experiments also show that HSI-TransUNet has outperformed several previous semantic segmentation models. The dataset in this paper, UAV-HSI-Crop, is publicly available. http://doi.org/10.57760/sciencedb.01898.

Agricultural land abandonment and retirement are important and lead to different types of land use and cover change. Generally, abandonment and retirement are caused by different social and environmental factors and result in different ecological and economic benefits and costs. Faced with the complexity of agricultural land change over time, this study aims to develop a new framework to distinguish between agricultural land abandonment and retirement and detect the extent and exact year of abandonment and retirement using Google Earth Engine (GEE). We tested our approach for three typical regions in the northern China crop-pasture band, where agricultural land abandonment and retirement are widespread. First, based on the spectral features obtained from Landsat images, annual land-cover maps were obtained with sample migration and random forest from 1998 to 2019 (0.87 overall accuracy). Second, a temporal consistency check method was proposed to further improve the classification performance (0.92 overall accuracy). Third, a trajectory-based change detection approach was developed to identify abandonment and retirement (F1 score for abandonment: 0.74, retirement: 0.83). Our results indicate that the spatiotemporal patterns of abandonment and retirement in the study area greatly differed. Overlapping of the topography and climate data showed that agricultural land with steep slopes ( $> 10^{\circ }$ ) was more likely to be retired and that abandonment was more likely to occur in areas with less precipitation. Overall, the methods used herein are robust for agricultural land abandonment and retirement monitoring and may be extended to other land-cover change studies.

本数据集为依托Google Earth Engine（GEE）云计算平台,基于10米空间分辨率的Sentinel-2遥感影像和随机森林模型,所生成的2019年全国农业塑料大棚空间分布专题数据。具体而言,首先通过野外调查和目视解译进行地面样本采集,并随机划分为训练集和测试集;然后进行地物光谱、纹理等特征提取,从而构建多维特征空间;最后构建随机森林分类模型,并利用训练好的模型对全国遥感影像进行并行计算,从而得到全国农业塑料大棚分类数据。精度测试表明,该数据集的平均分类精度为87.45%,能够正确反映农业塑料大棚在全国的空间分布情况。此外,为了更好对全国大棚分布数据进行可视化,本文同时计算了全国5公里格网内的大棚面积占比。本数据集为第一个公开发布的全国农业塑料大棚空间分布专题数据,可为相关领域的科研人员提供数据参考。