Displacement Prediction of Slow-Moving Landslides Using InSAR and Ensemble Regression Models based on Slope Units

Ground displacement is a key indicator of slope instability, crucial for mitigating landslides amid climate-driven triggers. Interferometric Synthetic Aperture Radar (InSAR) has become a key tool for detecting and characterizing large-scale, slow-moving displacements. This study aims to (i) characterize ground deformation in an Andean region with known landslide activity using the Small Baseline Subset (SBAS) InSAR technique, and (ii) propose a novel predictive framework for slow-moving displacements. Line-of-Sight (LOS) displacement time series (TS) from 2021–2023 were aggregated based on mean and extreme values at the slope unit (SU) level and described using static and dynamic variables, with the latter computed over 7-28-day intervals. The decomposed TS (trend and periodic terms) were modeled using Extreme Gradient Boosting (XGBoost). The characterization of the study area identified three zones with slow-moving deformation, with LOS velocities ranging from − 68 to 388.6 mm/year (ascending) and − 245.7 to 165.1 mm/year (descending). The predictive framework showed best performance in Zone 1, where MaxAbsDtsdesc predicted the trend term with RMSE = 3.76 mm, R² = 1.00, MAPE = 3%. The poorest performance occurred in Zone 3, with periodic errors reaching up to 262.90 mm. Elevation, fault proximity, and groundwater storage (GWS) were key predictors for the trend term, while GWS dominated in the periodic term. Overall, mean-based TS outperformed maximum-based ones for the periodic term, while no consistent advantage was found between TS types for the trend term or between ascending and descending geometries. This approach offers valuable insights for territorial planning and risk management in landslide-prone Andean regions.