Integrated catchment classification across China based on hydroclimatological and geomorphological similarities using self-organizing map and fuzzy c -means clustering for hydrological modeling

Catchment classification supports regionalisation and runoff prediction in data limited regions by organising basins into hydrologically coherent classes. China spans strong gradients in moisture availability, temperature regime, snow influence, and terrain, yet discharge observations remain sparse. We develop and evaluate an integrated climate-landscape classification for 13 487 HydroBASINS catchments using a hierarchical self-organizing map and fuzzy c -means (SOM–FCM) framework. Six hydroclimatic indices delineate climate regions on a 0.25° grid, and catchments are classified within each region using geomorphological and drainage network descriptors. The framework yields six climate regions and 35 classes, with fuzzy memberships characterising transitional areas. Hydrological relevance is assessed using seasonal hydrographs and event scale flow duration curves (FDCs) for ten gauged headwater catchments, and 13 flow signatures for 722 headwater basins matched to a discharge reanalysis product. Seasonal regimes are organised mainly by climate regions, whereas event response and high flow behaviour are modulated by landscape classes. Flow magnitude and high flow frequency signatures discriminate classes most strongly, while duration metrics show weaker contrasts. The resulting typology provides a transferable basis for selecting donor basins and constraining model parameters, thereby improving runoff prediction and regionalisation in ungauged catchments across China.