30 m-resolution annual crop type maps in Northeast China from 2001 to 2022

作者：Yuanyuan Di, Jinwei Dong, Nanshan You, Zhichao Li, Álvaro Moreno-Martínez, Emma Izquierdo‐Verdiguier, Jing Sun, Ping Fu · 发表于：Scientific Data · 年份：2026 · DOI：10.1038/s41597-025-06516-1 · 被引用次数：4 · 研究领域：Remote Sensing in Agriculture、Soil Geostatistics and Mapping、Satellite Image Processing and Photogrammetry

Northeast China, a crucial agricultural region contributing one-third of China’s commodity grain production, lacks detailed, high-resolution crop maps pre-2013 due to limited satellite observations. To bridge this gap, this study developed annual 30 m crop type maps for Northeast China (2001–2022) using all available Landsat and MODIS imagery and the Highly Scalable Temporal Adaptive Reflectance Fusion Model (HISTARFM) algorithm. The accuracy of crop-type maps was assessed using three complementary approaches. First, validation against ground-truth data (2017–2022) yielded overall accuracies of 80.7%–91%. Second, validation using ‘trusted pixels’ from existing crop-mapping products (2001–2020) produced overall accuracies of 85%–95%. Third, comparison with government statistics (2001–2022) showed average R 2 of 0.98 (paddy rice), 0.83 (maize), and 0.90 (soybean). The generated maps were found to be the most consistent with government statistics compared to pre-existing crop maps, while providing comprehensive spatial and temporal details. This dataset is an important contribution to long-term fine-resolution crop mapping at the regional scale in China, which provide valuable guidance for sustainable agriculture practices in China’s primary grain-producing region.