International Research Center of Big Data for Sustainable Development Goals, School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai, 519082, Guangdong, China.
Faculty of Geographical Science, Beijing Normal University, Beijing, 100875, China.
Sci Data. 2023 Sep 26;10(1):658. doi: 10.1038/s41597-023-02573-6.
China is the world's second-largest maize producer, contributing 23% to global production and playing a crucial role in stabilizing the global maize supply. Therefore, accurately mapping the maize distribution in China is of great significance for regional and global food security and international cereals trade. However, it still lacks a long-term maize distribution dataset with fine spatial resolution, because the existing high spatial resolution satellite datasets suffer from data gaps caused by cloud cover, especially in humid and cloudy regions. This study aimed to produce a long-term, high-resolution maize distribution map for China (China Crop Dataset-Maize, CCD-Maize) identifying maize in 22 provinces and municipalities from 2001 to 2020. The map was produced using a high spatiotemporal resolution fused dataset and a phenology-based method called Time-Weighted Dynamic Time Warping. A validation based on 54,281 field survey samples with a 30-m resolution showed that the average user's accuracy and producer's accuracy of CCD-Maize were 77.32% and 80.98%, respectively, and the overall accuracy was 80.06% over all 22 provinces.
中国是世界第二大玉米生产国，贡献了全球 23%的产量，对稳定全球玉米供应起着至关重要的作用。因此，准确绘制中国玉米的分布地图对于区域和全球粮食安全以及国际谷物贸易具有重要意义。然而，由于现有的高空间分辨率卫星数据集存在云覆盖导致的数据空白，特别是在潮湿和多云地区，因此仍然缺乏具有精细空间分辨率的长期玉米分布数据集。本研究旨在制作一张中国长期、高分辨率的玉米分布图（China Crop Dataset-Maize，CCD-Maize），该图可识别 2001 年至 2020 年 22 个省、直辖市的玉米。该图是使用高时空分辨率融合数据集和一种基于物候的方法（称为时间加权动态时间规整）制作的。基于分辨率为 30 米的 54281 个实地调查样本的验证表明，CCD-Maize 的平均用户精度和生产者精度分别为 77.32%和 80.98%，在所有 22 个省的整体精度为 80.06%。
