Spatial distribution mapping of Populus euphratica in the Tarim River Basin using multi-source remote sensing data and deep learning

The Tarim River is the longest inland river in China, and its basin represents a typical ecologically fragile area within arid regions, highly susceptible to human activities and climate change. These factors contribute to increased regional desertification and the deterioration of the ecological environment. As the sole community-forming tree species in this basin, a timely understanding of the spatial distribution and growth status of Populus euphratica is essential for maintaining the ecological balance in the desert and ensuring the security of the oasis ecosystem. Currently, there is no spatial distribution map of Populus euphratica in the basin, primarily due to significant variations in stand density and tree branch architecture, along with a lack of high-spatial-resolution data. This study addresses the gap by constructing a comprehensive dataset of single-tree parameters for Populus euphratica through the integration of LiDAR and GF-2 satellite imagery. We developed a deep learning model tailored to different densities and crown architecture of Populus euphratica, enabling accurate quantification of the spatial distribution of these forests in the Tarim River basin. This study provides valuable insights for the inversion of large-scale fine forest structure parameters and serves as a crucial foundation for the management and conservation of forest resources in arid regions.