Landslide InSAR monitoring is crucial for understanding the evolutionary mechanisms of geological disasters and enhancing risk prevention and control capabilities. However, for complex terrains and large-scale landslides, satellite-based SAR monitoring faces challenges such as a low observation frequency and limited spatial deformation interpretation capabilities. Additionally, two-dimensional monitoring struggles to comprehensively capture multi-directional movements. Taking the post-disaster monitoring of the landslide in Yunchuan, Sichuan Province, as an example, this study proposes a method for three-dimensional deformation dynamic monitoring by integrating dual-view MIMO ground-based synthetic aperture radar (GB-InSAR) data with high-resolution digital elevation model (DEM) data, successfully reconstructing the three-dimensional displacement fields in the east–west, north–south, and vertical directions. The results show that deformation in the landslide area evolved from slow accumulation to rapid failure, particularly concentrated in the middle and lower regions of the landslide. The average three-dimensional deformation of the main slip zone was approximately 60% greater than that of the original slope, with a maximum deformation of −100 mm. These deformation characteristics are highly consistent with the topographic structure and sliding direction. Field investigations further validated the radar data, with observed surface cracks and accumulation zones consistent with the high-deformation regions identified by the monitoring system. This system provides a solid foundation for geological disaster early warning systems, mechanism research, and risk prevention and control.

Landslides are common in mountainous, hilly, and steep-slope regions, posing serious threats to human life, infrastructure, and ecological systems. Precipitation is a major triggering factor that can rapidly destabilize slopes by increasing moisture content and reducing soil shear strength. This study proposes an integrated landslide monitoring framework that combines satellite-borne InSAR for regional-scale detection, GB-InSAR for high-density dynamic monitoring, GNSS for precise three-dimensional ground deformation constraints, and rain-gauge observations for precipitation forcing characterization, with the aim of improving monitoring accuracy, spatiotemporal resolution, and early-warning capability. The results reveal a clear temporal correspondence between rainfall events and landslide displacement: during precipitation peaks (0.4 mm and 0.3 mm), displacement increased to approximately +25 mm and +35 mm, indicating direct rainfall-driven deformation responses. GB-InSAR further captures pronounced spatial heterogeneity across the landslide body, with maximum cumulative displacement reaching +656.3 mm and minimum displacement −1098.17 mm. GNSS measurements corroborate these deformation patterns, particularly during the acceleration stage when the trends derived from GNSS and GB-InSAR are highly consistent. Overall, multi-source evidence indicates that both short-duration intense rainfall and long-term precipitation accumulation contribute to landslide activity, with precipitation capable of inducing not only rapid displacement surges but also slower deformation under low-rainfall conditions. These findings provide a robust basis for landslide early warning and risk assessment in regions frequently affected by intense rainfall and sustained wet periods.

Slope displacement interval prediction quantifies uncertainty in future deformation and supports slope monitoring and early warning. This study extends the point-prediction FA-VAG backbone previously reported by Gan et al. rather than presenting that backbone as a new model. The new contribution is an interval-prediction layer that transfers the pretrained FA-VAG hidden states, reduces them by Principal Component Analysis (PCA), fuses them with multi-step point forecasts, and maps the fused representation to non-crossing bounds using a Random Vector Functional Link Network (RVFLN). A prior-informed improved Sparrow Search Algorithm (SG-ISSA) then optimizes the RVFLN output weights. For the prespecified target point 7, the framework achieved PICP/MPIW/CWC values of 0.90/5.97/5.97 for five-step prediction and 1.00/8.74/8.74 for seven-step prediction. Independent external validation yielded 0.90/58.82/58.82 for five-step prediction. These results support the interval-extension framework on the two evaluated datasets; additional probabilistic baselines and harmonized reruns of the ablation study remain necessary.