Hydrography30: A 30 m-resolution global hydrography dataset derived from FABDEM using terrain–observation fusion

Hydrography maps constitute a fundamental dataset for hydrological modeling and analysis and are commonly derived from digital elevation models (DEMs). Despite the growing availability of 30-m DEMs, a corresponding global hydrography dataset at 30 m resolution is still lacking, resulting insufficient representation of micro-scale drainage networks and watershed boundaries across a wide range of landscapes. To address these limitations, we develop a terrain–observation fusion framework that integrates Forest And Buildings removed Copernicus DEM (FABDEM) with multi-source surface water data to guide the identification of endorheic basins. We further employ multiple algorithms for flow-direction correction to adjust flow patterns within depressions. Based on this framework, we construct Hydrography30, a 30-m resolution global hydrography map that shows strong consistency with existing observational datasets. Spatial overlap of watershed boundaries exceeds 0.8 for 91% of Global Runoff Data Centre (GRDC) gauges, indicating high geometric accuracy and robust representation of both exorheic and endorheic basins. Validation against observed endorheic data further shows that 80% of the endorheic outlets in the Qiang-tang Basin are captured by the derived hydrography. Moreover, by providing globally consistent 30 m flow direction and flow accumulation data, this dataset establishes a high-resolution hydrography foundation that enables explicit, structure-resolving hydrological analysis at the global scale.