An advanced coverage estimation method to quantify biological soil crust coverage using Sentinel-2 imagery in desert and sandy land of China

Monitoring the distribution and area change of biological soil crusts (BSCs) can enhance our understanding interactions between nonvascular plants environment in drylands. However, using only pixel-based binary classification methods results large-area estimation errors at large scales. The lack available calculation for directly measuring BSC coverage multispectral satellite images makes it challenging to obtain data further studies To address these issues, this study developed feature space conceptual models desert sandy land based on characteristics comprised normalized difference vegetation index (NDVI) combined with brightness (BI), encompassing moss, lichen, non-BSC. relied crust (BSCI) NDVI, including vegetation, mixed BSCs soil. Using Sentinel-2 imagery a spectral unmixing model, abundance was quantified four growth areas located Gurbantunggut Desert Mu Us Sandy Land China. Validation method indicated that root mean square error (RMSE) 10% 8% land, respectively (estimation accuracies 79% 81%, respectively). This demonstrated proposed effectively estimate subpixel scale. resulting provide possibility evaluate functions regional ecosystems.