Coupling FR-AHP with Information Value Model for Landslide Susceptibility Assessment in the Yanhe River Basin, China

The Yanhe River Basin on the Loess Plateau suffers frequent landslides, threatening sustainable development and ecological security. This study employs correlation and multicollinearity analyses to select evaluation factors and construct an index system. The information value (IV) model determines indicator values, while the Analytic Hierarchy Process (AHP), frequency ratio (FR), and their combination, FR-AHP, establish relative weights of influencing factors, thereby constructing weighted IV models validated by Receiver Operating Characteristic (ROC) curves. Results demonstrate that: (1) slope, aspect, curvature, lithology, average annual precipitation, NDVI, distance from rivers, distance from roads, and land use type are key influencing factors; (2) compared with IV, AHP-IV, and FR-IV models, the FR-AHP-IV model achieves the highest prediction accuracy with an AUC = 0.818; and (3) areas of low, moderate, high, and very high susceptibility account for 32.01%, 40.69%, 18.49%, and 8.80%, respectively, with higher susceptibility along both banks of the middle and lower reaches. The landslide-prone hotspot in the southern Baota District aligns with actual survey conditions. The methodological framework proposed in this study can serve as a reference for basins with similar geographic settings and hazard backgrounds, yet the specific zoning results should be adapted and validated in accordance with the actual conditions of individual basins.