Mapping Vegetation and Land Use Types in Fanjingshan National Nature Reserve Using Google Earth Engine

Researchers developed a robust method to map vegetation and land use types in Fanjingshan National Nature Reserve (FNNR), a biodiversity hotspot in China, using Google Earth Engine. The study demonstrated that multi-seasonal Landsat image composites combined with elevation data significantly improved mapping accuracy by addressing challenges posed by steep terrain and persistent cloud cover. By leveraging spectral vegetation indices (SVIs) and shade/illumination normalization techniques, the team achieved consistent classification accuracies above 70%, even with limited training data. Advanced machine learning routines, optimized through tuning techniques, ensured reliable monitoring of forest cover and land use in FNNR and similar mountainous, cloud-prone regions.

Tsai, Y. H., Stow, D., Chen, H. L., Lewison, R., An, L., & Shi, L. (2018). Mapping Vegetation and Land Use Types in Fanjingshan National Nature Reserve Using Google Earth Engine. Remote Sensing, 10(6), 927. https://doi.org/10.3390/rs10060927