Global cropland at 30m resolution for 2022 (GCL_30m_2022)

Global cropland at 30m resolution for 2022 was developed by leveraging multi-source data integration and machine learning classification algorithms, supported by a cloud computing platform.Cropland data layer from existing products were first collected and investigated. Based on independent validation samples. Multi-source cropland datasets were integrated in the criteria of higher accuracy and higher spatial resolution (Nabil et al., 2021). For regions with high uncertainty, crowd-sourced in situ measurement, massive remote sensing data, artificial intelligence classifiers were integrated to reproduce the cropland map for each agroecological zones (Bofana et al., 2020). Based on the optimized cropland dataset, Big Earth Data techniques were applied to all available Gaofen-1 imageries in China and Landsat & Sentinel-2 imageries at global scale from 2021 to 2023 to identify cultivated and uncultivated cropland based on time series analysis on annual basis from 2021 to 2023. Annual cultivated cropland maps were composited to generate global cropland product for the three continuous years at high spatial resolution (30m).