Scholarlink

<p>Accurate and timely information on planting intensity and crop rotation is essential for guiding agricultural policies and ensuring food security. However, reliable and up-to-date maps for major crops—wheat, maize, rapeseed, soybean, and potatoes—are lacking in the Loess Plateau, a key grain-producing region in western China. To address this gap, this study aims to generate a high-resolution (10 m) crop planting pattern dataset for the Loess Plateau from 2018 to 2022. The research methodology involved four key steps: (1) Enhancing the sample dataset using phenological indices and the Dynamic Time Warping (DTW) algorithm; (2) Identifying crop planting intensity based on phenological growth curves; (3) Developing independent random forest classifiers tailored to agricultural climate zones; and (4) Constructing an optimal feature subset for crop classification. The resulting maps demonstrated high overall accuracies (OA) is greater than 0.81, with satellite-based estimates showing strong agreement with municipal statistical data (R<sup>2</sup> ≥ 0.60). These results provide crucial insights for the management of agricultural ecosystems in the Loess Plateau and can support more informed decision-making in regional agriculture.</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

High-resolution (10 m) dataset of multi-crop planting structure on the Loess Plateau during 2018–2022

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月10日

查看全文
  • Xining Zhao,
  • Jichao Wang,
  • Yelu Ding,
  • Xiaodong Gao,
  • Changjian Li,
  • Hongwei Huang,
  • Xuerui Gao

摘要

Accurate and timely information on planting intensity and crop rotation is essential for guiding agricultural policies and ensuring food security. However, reliable and up-to-date maps for major crops—wheat, maize, rapeseed, soybean, and potatoes—are lacking in the Loess Plateau, a key grain-producing region in western China. To address this gap, this study aims to generate a high-resolution (10 m) crop planting pattern dataset for the Loess Plateau from 2018 to 2022. The research methodology involved four key steps: (1) Enhancing the sample dataset using phenological indices and the Dynamic Time Warping (DTW) algorithm; (2) Identifying crop planting intensity based on phenological growth curves; (3) Developing independent random forest classifiers tailored to agricultural climate zones; and (4) Constructing an optimal feature subset for crop classification. The resulting maps demonstrated high overall accuracies (OA) is greater than 0.81, with satellite-based estimates showing strong agreement with municipal statistical data (R2 ≥ 0.60). These results provide crucial insights for the management of agricultural ecosystems in the Loess Plateau and can support more informed decision-making in regional agriculture.