Cloud-based mapping of potential peatlands in Mongolia using Coupled optical and microwave remote sensing and hierarchical decision tree

In Mongolia, peatlands are critical for regional hydrological regulation, biodiversity conservation, and carbon sequestration, yet their spatial distribution and extent remain poorly characterized due to limited large-scale surveys. As critical terrestrial carbon reservoirs, peatlands play a vital role in carbon cycling and ecological conservation services. Carbon emissions associated with peatland degradation in Mongolia have become an an increangly important concern in the context of climate change. However, rapid peatland identification across Mangolia remains challenging due to its complex topography, extensive land area, and the limitations of relying on a single remote sensing data source and limited local computational power. Existing studies are largely limited to either small-scale local surveys or large-scale assessment based on coarse-resolution data, constraining their applicability for comprehensive peatland mapping at the national scalle. Hence, this study proposes a national-scale peatland mapping methodology based on the Google Earth Engine (GEE) platform and a four-layer hierarchical decision tree. Multi-source datasets, including Sentinel-1, Sentinel-2, and Copernicus DEM GLO-30, were utilized to derive 14 key features using a quantile-based method. The quantile method was used to extract 14 key features. The results indicate that: (1) The p10–p90 quantile method effectively distinguishes peatlands from other land covers and avoids misclassifications in high-altitude permafrost zones and steep slopes, providing spatially explicit classification support; (2) The decision tree filter sequentially excludes permanent water bodies and Gobi Desert bare land, then screens potential wetlands before precisely identifying peatlands. Combined with GEE cloud computing, the analysis time for peatland extraction covering 1.566 million km 2 was reduced to just hours; (3) The study identifies a “dense in the north, sparse in the south” peatland spatial pattern in Mongolia. The hydrothermal regime, topography, and permafrost jointly influence peatland distribution. Our method combines interpretability and potential transferability by design, providing a technical reference for large-scale peatland extraction in data-scarce and remote regions.