Landslide susceptibility mapping (LSM) in mountain–basin transition zones remains challenging because conventional approaches rely mainly on historical inventories and static conditioning factors, whereas independent deformation evidence is seldom incorporated to refine susceptibility zonation. This study proposes an integrated LSM framework for the Xining Basin by coupling a Mamba-based model (Mamba-LSM) with SBAS-InSAR-based deformation-informed bidirectional reclassification, with the key innovation lying in the use of independent deformation evidence to refine susceptibility zonation after model prediction. Specifically, Mamba-LSM integrates six-channel neighborhood patches, CNN-based local spatial encoding, and Mamba-based latent feature transformation to improve the representation of local terrain context for landslide susceptibility assessment. Results show that Mamba-LSM achieved the highest AUC among the evaluated models, reaching 0.9011 with an F1-score of 0.7431. After deformation-informed bidirectional reclassification, the high- and very-high-susceptibility classes occupied only 25.31% of the study area but contained 69.84% of the mapped landslides, and were concentrated mainly in valley–mountain transition belts, river-incised slopes, and engineering-disturbed sectors where SBAS-InSAR deformation hotspots were also preferentially distributed. These findings demonstrate that integrating independent SBAS-InSAR deformation evidence can improve both the spatial concentration of landslides in high-susceptibility zones and the physical interpretability of susceptibility zonation.

Accurate mapping of landslide-affected land surfaces from very-high-resolution optical imagery is essential for mountainous land monitoring and hazard-related land management, yet it remains difficult in complex terrain because landslide bodies are fragmented, elongated, shadowed, and easily confused with bare soil, terraces, roads, and erosion features. This study presents TM-Net as a task-oriented decoder-centric architecture for landslide-affected land surface mapping. The Inception Token Mixer is adopted from InceptionNeXt as an encoder-side component, whereas TM-Block is newly designed as a decoder-side twin-path refinement module for boundary recovery, slender target preservation, and complex-background suppression. On the fused public benchmark, TM-Net achieved the highest IoU (73.56%) and F1 (84.76%) under the identical evaluation protocol. On XBLD, TM-Net achieved the highest IoU (44.94%) and F1 (62.01%). These results indicate that TM-Net provides a favorable balance between missed detections and false-positive predictions for mapping visually identifiable landslide-affected land surfaces in heterogeneous mountainous landscapes.

Conditioning factor reduction can decrease redundancy and modelling complexity in landslide susceptibility mapping. These factors describe the environmental conditions associated with slope instability. Statistical screening, redundancy diagnostics, importance rankings and internal validation do not establish that the remaining conditioning factors can be safely removed. Whether aggressive reduction preserves the environmental information required for reliable regional landslide susceptibility mapping therefore remains uncertain. This study developed an integrated framework for evaluating the reliability of conditioning-factor reduction in the Hualong–Xunhua region of the Upper Yellow River, China, combining statistical diagnostics, model interpretation, GeoDetector analysis, and random and spatial validation. The compact system selected during development was further compared with the complete system using a reserved evaluation subset, followed by regional susceptibility mapping and spatial comparison. The results showed that, first, precipitation showed weak univariate but strong model and interaction evidence, illustrating their complementarity. Second, Top5, comprising normalised difference vegetation index (NDVI), distance to rivers, precipitation, lithology and terrain ruggedness index (TRI), met development criteria: mean area under the receiver operating characteristic curve (AUC) values were 0.866 under random validation and 0.846 under spatial validation, compared with 0.875 and 0.853 for the complete 17-factor system (Full17), respectively. Third, in the post-selection reserved comparison, mean AUC was 0.865 for Top5 and 0.895 for Full17, with paired uncertainty supporting this decline; favourable development performance did not ensure reliable substitution. Finally, Full17 was retained for final mapping; absolute score differences greater than 0.10 covered 32.02% of mapped area, indicating potential spatial information loss. Similar predictive performance did not imply spatial equivalence. These findings show that factor reduction requires evaluation of predictive performance, generalisation, model dependence, and spatial information preservation.