SEA_Cropland10: Annual 10-m high-resolution cropland maps for Southeast Asia since 2019

SEA_Cropland10 is a high-resolution (10-meter) annual cropland dataset covering Southeast Asia from 2019 to 2025, achieving overall accuracies exceeding 92%. Developed within Google Earth Engine, it synergizes Sentinel-1 SAR and Sentinel-2 optical imagery with embeddings from the AlphaEarth foundation model. This integration effectively mitigates challenges posed by persistent cloud cover and complex terrain, significantly improving the detection of fragmented smallholder farms and sloping agriculture compared to existing global products. The dataset is provided as binary raster layers: cropland pixels—encompassing paddy fields, drylands, and sloping agriculture—are assigned a value of 1. All non-cropland categories (e.g., forests, water, built-up areas) and regions outside the study area are masked as NoData, ensuring the data strictly represent cultivated land. Specifications & Usage: Format: GeoTIFF (WGS 84, EPSG:4326) Resolution: 10 meters Temporal Coverage: Seven annual layers (2019–2025) This dataset is optimized for regional food security monitoring, analyzing land-use change dynamics, and assessing agricultural expansion into highland areas. It is fully compatible with standard GIS software (QGIS, ArcGIS) and spatial analysis environments (Python, R). Google Earth Engine App: [Link]. [Version Update Note]: This release updates the dataset with the newly processed 2025 data. Citation: Cai, Y., Li, B., Liu, X., Jiang, X., Zhu, Y., Luo, S., ... & Zeng, Z. (2026). Annual 10-m high-resolution cropland maps for Southeast Asia since 2019 using AlphaEarth embeddings. Earth System Science Data Discussions, 2026, 1-30.