The spatial differentiation of alpine wetlands on the eastern Tibetan Plateau using multi-source remote sensing images

The alpine wetlands on the eastern Qinghai-Tibet Plateau (EQTP) serve as a critical global ecological barrier. Under the dual pressures of climate change and human activities, these wetland systems face environmental challenges such as retrogressive succession, aridification, and desertification. Based on the Google Earth Engine (GEE) cloud computing platform, this study integrates high-resolution imagery, multi-source geoscience datasets, and field survey samples. Object-based image analysis (OBIA), logistic regression, and species distribution models (SDMs) were employed to systematically assess the spatiotemporal differentiation characteristics and key driving factors of alpine wetlands in EQTP. The results indicate that: (1) When applying OBIA classification to alpine wetlands, as image resolution increased from 30 m to 5 m, classification accuracy exhibited an improvement–saturation–fragmentation pattern. At a resolution of 10 m (Scale = 26), marsh wetland structures and spatial distribution characteristics were accurately identified, with a total wetland resource area of 17,454.56 km 2 . (2) Wetland distribution is driven by multiple factors, including climate (temperature, precipitation), topography (elevation, slope), and human activities (road density, settlement distribution). The best explanatory performance for driving forces was observed at a 500 m spatial scale (AUC = 0.81), confirming that climate factors predominantly govern long-term changes, while human activities significantly influence ecological patterns. (3) During 2021–2040, under a low-emission scenario, the area of highly suitable wetland zones was larger than under a high-emission scenario, with warming causing very high suitability zones to shift toward higher elevations. From 2041 to 2060, as regional warming intensified, the area of excellent suitability wetlands decreased. Between 2081 and 2100, the high-carbon emission scenario increased temperature in the high-altitude central study area, improving wetland suitability. This study proposes a GEE-based OBIA method for estimating alpine wetland resources, integrating logistic regression and SDMs to reveal the spatiotemporal differentiation mechanisms of alpine wetlands. The findings provide an effective technical framework for wetland research on the Qinghai-Tibet Plateau.