Using remote sensing methodologies to uncover the temporal-spatial patterns of cropland abandonment(CA)offers substantial advantages at both macro scales and in real time.However,the current literature lacks a systema...Using remote sensing methodologies to uncover the temporal-spatial patterns of cropland abandonment(CA)offers substantial advantages at both macro 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 reviewed remote sensingbased methods for monitoring CA,its causes,and ecological effects.Our results show that the methods for monitoring abandoned cropland can be classified into 2 major categories:those based on image spectral features and those based on land cover temporal trajectories and vegetation phenology dynamics.Among the 8 subcategories,vegetation phenology and dynamic methods exhibit the highest average overall accuracy at 89.33%±3.37%.Remote sensing plays a crucial role in assessing the causes of CA,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.Further advancements are needed in classifying abandoned cropland types,observing fragmented and temporally unstable parcels,and assessing ecological effects across different scenarios.More importantly,we presented a trinity CA monitoring framework based on the cause-pattern-effect pillars,which offers a novel perspective for comprehensive research on CA.Overall,our work provided a systematic and insightful perspective for advancing remote sensing research on CA.展开更多
基金supported by the National Key R&D Program of China(grant number:2022YFE0195900).
文摘Using remote sensing methodologies to uncover the temporal-spatial patterns of cropland abandonment(CA)offers substantial advantages at both macro 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 reviewed remote sensingbased methods for monitoring CA,its causes,and ecological effects.Our results show that the methods for monitoring abandoned cropland can be classified into 2 major categories:those based on image spectral features and those based on land cover temporal trajectories and vegetation phenology dynamics.Among the 8 subcategories,vegetation phenology and dynamic methods exhibit the highest average overall accuracy at 89.33%±3.37%.Remote sensing plays a crucial role in assessing the causes of CA,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.Further advancements are needed in classifying abandoned cropland types,observing fragmented and temporally unstable parcels,and assessing ecological effects across different scenarios.More importantly,we presented a trinity CA monitoring framework based on the cause-pattern-effect pillars,which offers a novel perspective for comprehensive research on CA.Overall,our work provided a systematic and insightful perspective for advancing remote sensing research on CA.
