20 m resolution maps of paddy rice cropping intensity and planting dates in Monsoon Asia from 2018 to 2021

Robust monitoring of paddy rice cropping intensity and planting dates across Monsoon Asia (MA) using satellite remote sensing is crucial for ensuring regional food security and sustainable agricultural development. Existing satellite-based paddy rice maps face limitations such as coarse spatial resolution, inadequate coverage, and poor data quality. To address these issues, we developed a sample-free (no training samples required) algorithm to generate paddy rice cropping intensity and planting date maps at a 20 m (1/5400°) spatial resolution for the entire MA between 2018 and 2021. Our innovative algorithm utilizes Sentinel-1 SAR data to extract potential transplanting signals of paddy rice fields, and verifies the authenticity of these signals using multispectral (optical), infrared, and passive microwave satellite data. In detail, we perform well-designed regional-specific checks on the width, prominence, sharpness, location, and value of all troughs in SAR backscatter time series, because some of them are induced by the disturbances of factors other than flooding. Validation against 2305 field reference samples and 1484 public dataset samples indicates an overall accuracy of 82‒84% in MA, outperforming most existing regional-scale datasets. Our regional mean planting dates also correlate well with the regional statistics-based RiceAtlas dataset (R2: 0.92, RMSE: 28 days).