A long-term paddy rice distribution dataset in Asia at a 30 m spatial resolution

Area of Science:

• Agricultural Science
• Environmental Science
• Remote Sensing

Background:

• Rice is a global staple food and a significant source of anthropogenic methane emissions.
• Accurate rice distribution mapping in Asia is vital for food security and climate change monitoring.
• Existing remote sensing data is limited by cloud cover, hindering long-term, high-resolution rice mapping.

Purpose of the Study:

• To develop the Global Crop Dataset-Rice (GCD-Rice) for mapping rice cultivation.
• To provide a long-term (1990-2023) high-resolution dataset for 16 Asian countries.
• To improve monitoring of food security and greenhouse gas emissions from rice paddies.

Main Methods:

• Utilized Landsat and Sentinel-1 satellite imagery.
• Employed a phenological approach combined with a random forest model.
• Validated maps using 258,547 field samples and statistical area comparisons.

Main Results:

• Achieved high validation accuracies: 89.88% user accuracy, 88.52% producer accuracy, and 88.85% overall accuracy.
• Statistical area comparisons showed strong agreement with R²=0.768, slope=0.874, and RMSE=0.346.
• Generated a comprehensive dataset of rice cultivation across three seasons for 16 Asian countries.

Conclusions:

• The GCD-Rice dataset offers a reliable, high-resolution tool for tracking rice distribution.
• This dataset is essential for informed decision-making regarding food security and climate change mitigation.
• The methodology provides a robust framework for future agricultural monitoring systems.