South China Sea Geomorphological Features Dataset

The factor dataset contains 15 key topographic factors (e.g., slope, curvature, roughness) derived from GEBCO_2023 bathymetry, which served as features for training the final XGBoost classification model. The contribution of these factors was quantified via SHAP analysis. The data, standardized and provided per terrain unit, excludes the Compactness Index (CI).The classification result dataset presents the final submarine geomorphological map of the South China Sea, classifying ~69k terrain units into a hierarchical system of 7 major and 13 sub-types. Each polygon contains the landform class. This result was generated using the terrain factors dataset with an XGBoost model, achieving an overall accuracy of 94.46%.