Land Use Dataset in the Turpan-Hami Basin from 1990 to 2020

The Turpan-Hami (Tuha) Basin, as an ecologically fragile zone in the arid region of northwestern China, is an important energy base and fruit production area in Xinjiang. With the development of energy extraction and the fruit industry, land use types have undergone significant changes in recent decades, profoundly impacting regional ecological security, water resource allocation, and the sustainable development of the Silk Road Economic Belt. To thoroughly reveal the characteristics of land use evolution in the region over the past 30 years, this study, based on the Google Earth Engine (GEE) cloud platform, integrated Landsat series remote sensing images (30 m resolution) from 1990 to 2020. By combining field survey samples and multi-dimensional remote sensing indices, the first high-resolution 30-year time-series land use dataset for the Tuha Basin was constructed. The dataset employed a supervised classification using the random forest algorithm, integrating original spectral bands with typical remote sensing indices (such as NDVI, NWI, NDBBI, etc.). Training data were obtained through field sampling of land type sample points and interpretation results from high-resolution images, effectively improving classification accuracy and regional adaptability. Accuracy assessment results show that the dataset has a high overall accuracy (87.2%) and Kappa coefficient (0.84). This dataset can provide scientific data support for land use simulation, ecological risk assessment, carbon storage estimation, and spatial planning optimization in the region.