Scholarlink

<p>Winter wheat is a major food crop in China, with the North China Plain (NCP) serving as the primary production region and playing a critical role in national food security. However, long-term, high-resolution winter wheat maps for this region remain limited. This study presents a 30 m resolution winter wheat distribution dataset for the NCP (WheatMapNCP) from 2000 to 2024, generated by combining automated training sample generation with a random forest (RF) classifier. A stratified random sampling framework was used to estimate unbiased winter wheat area. Validation results show that the produced maps achieved an average overall accuracy (OA) of 95.98 ± 1.15%, and an average F1 score of 86.4%. The estimated planted areas show a consistent temporal trend with official statistics, and the mapped areas show strong correlation to statistics at the provincial (R² = 0.97–0.98) and municipal (R² = 0.84–0.96) levels. The dataset not only provides long-term, high-quality support for monitoring winter wheat dynamics, but also offers a reliable approach for generating timely, transparent wheat area estimates based on satellite data.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A 30 m winter wheat distribution dataset for the North China Plain from 2000 to 2024

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月14日

查看全文
  • Fangjie Li,
  • Inbal Becker-Reshef,
  • Josef Wagner,
  • Yuval Sadeh,
  • Jean Rehbinder,
  • Françoise Nerry

摘要

Winter wheat is a major food crop in China, with the North China Plain (NCP) serving as the primary production region and playing a critical role in national food security. However, long-term, high-resolution winter wheat maps for this region remain limited. This study presents a 30 m resolution winter wheat distribution dataset for the NCP (WheatMapNCP) from 2000 to 2024, generated by combining automated training sample generation with a random forest (RF) classifier. A stratified random sampling framework was used to estimate unbiased winter wheat area. Validation results show that the produced maps achieved an average overall accuracy (OA) of 95.98 ± 1.15%, and an average F1 score of 86.4%. The estimated planted areas show a consistent temporal trend with official statistics, and the mapped areas show strong correlation to statistics at the provincial (R² = 0.97–0.98) and municipal (R² = 0.84–0.96) levels. The dataset not only provides long-term, high-quality support for monitoring winter wheat dynamics, but also offers a reliable approach for generating timely, transparent wheat area estimates based on satellite data.