Land Use/cover dataset of the Guangdong-Hong Kong Macao Greater Bay Area at 30 m Resolution from 1976 to 2020

High spatiotemporal resolution, large-scale, multi-category land use data are crucial for studying regional ecological and environmental changes, urbanization impacts, and sustainable development planning.
In this study, the Google Earth Engine platform, Landsat satellite imagery, and a substantial number of manually interpreted samples were used to develop a dataset of annual land use in the Guangdong-Hong Kong Macao Greater Bay Area (GBA) at a 30 m resolution from 1976 to 2020. This dataset, termed Annual Land Use/Cover of the Greater Bay Area (LUC-GBA), was used to analyze the annual land use variation in 11 cities within the GBA. The high level accuracy achieved with the LUC-GBA dataset was evidenced by an overall accuracy (OA) of 93.98% and a kappa coefficient of 0.92 in 2020. The OA of interannual classification models ranges from 83.9% to 93.9%, and the kappa coefficients from 0.805 to 0.923, thus demonstrating the good performance of the model in multi-category land use classification within the GBA.
Resolution: 30m
Time span: 1976-2020 (excluding 1981-1985)