<b>A </b><b>G</b><b>lobal </b><b>R</b><b>eview of Monitoring Cropland Abandonment </b><b>U</b><b>sing Remote Sensing methodology: </b><b>Temporal</b><b>-spatial </b><b>P</b><b>atterns, Causes, Ecological Effects, and Future Prospects</b>

Using remote sensing methodologies to uncover the temporal-spatial patterns of cropland abandonment (CA) offers significant advantages at bothmacro scales and in real time. However, the current literature lacks a systematic review of specific typologies and methods regarding the application of remote sensing technology to CA monitoring. To address this knowledge gap, we systematically review remote sensing-based methods for monitoring CA, its causes, and ecological effects. Our results show thatthe methods for monitoring abandoned cropland can be classified into two major categories: those based on image spectral featuresand thosebased on land covertemporal trajectories and vegetation phenologydynamics. Among the eight subcategories, vegetation phenology and dynamic methods exhibit the highest average overall accuracy at 89.33±3.37%, compared to the other methods. It is crucial to assess the causes of CA through remote sensing observations, such as road density, spatial information of agricultural infrastructure, and the ecological effects resulting from abandonment, including food loss risks, carbon sequestration, wildfire risk, evapotranspiration, wilderness quality, biodiversity, and climate change. Through this systematic review, we argue that remote sensing has greater potential for monitoring CA in the future, with room for further progress in the classification of abandoned cropland types, the observation of fragmented and temporally unstable parcels, and the ecological effects in different scenarios. More importantly, we presenta trinity CA monitoring framework based on the cause-pattern-effect pillars, which offers a novel perspective for comprehensive research on CA. Overall, our work provides a systematic and insightful perspective for advancing remote sensing research on CA采用遥感方法揭示耕地撂荒（cropland abandonment, CA）的时空分布特征，在宏观尺度与实时监测层面均具备显著优势。然而，当前学界尚未针对遥感技术应用于耕地撂荒监测的具体分类体系与方法开展系统性综述研究。为填补这一研究空白，本文系统梳理了用于耕地撂荒监测的遥感方法、撂荒成因及其生态效应。研究结果表明，耕地撂荒监测方法可分为两大类别：一类基于影像光谱特征，另一类基于土地覆盖时空轨迹与植被物候动态。在八大子类别中，植被物候与动态方法的平均总体精度最高，达89.33±3.37%，优于其余各类方法。通过遥感观测手段评估耕地撂荒的成因至关重要，例如道路密度、农业基础设施空间分布等；同时撂荒引发的生态效应亦需关注，涵盖粮食损失风险、碳固存、野火风险、蒸散发、原生地质量、生物多样性以及气候变化等维度。通过本次系统性综述，本文认为遥感技术在未来耕地撂荒监测领域具备更大应用潜力，可在撂荒耕地类型分类、破碎化与时态不稳定地块监测以及不同场景下的生态效应评估等方面实现进一步突破。更为重要的是，本文提出了一套基于‘成因-格局-效应’三维框架的耕地撂荒监测体系，为耕地撂荒的综合性研究提供了全新视角。总体而言，本研究为推动耕地撂荒遥感监测领域的发展提供了系统性且富有深度的研究视角。