Large-Scale Populus euphratica Distribution Mapping Using Time-Series Sentinel-1/2 Data in Google Earth Engine

Accurate and efficient large-scale mapping of P. euphratica distribution is great importance for managing protecting forests, policy making, realizing sustainable development goals in the ecological environments desert areas. In large regions, numerous types vegetation exhibit spectral characteristics that closely resemble those euphratica, such as Tamarix, artificial allée trees, posing challenges accurate identification euphratica. To solve this issue, paper presents a method mapping. The geographical were first utilized to rapidly locate appropriate region interest further reduce background complexity interference from other similar objects. Spectral features, indices, phenological backscattering features extracted all available Sentinel-2 MSI Sentinel-1 SAR data 2021 regarded input random forest model used classify GEE platform. results then compared with using only index added by visually quantitatively referencing field-surveyed samples, UAV data, high-spatial-resolution Google Earth Data Map World. comparison indicated proposed method, which adds both time-series could correctly distinguish have information rates omission errors (OEs), commission (CEs), overall accuracy (OA) 12.53%, 11.01%, 89.32%, respectively, representing increases approximately 9%, 17%, 13% features. significantly improved classification terms and, especially, commission.