Abstract

Background and aim:

Mass movements are among the most destructive geomorphological hazards in mountainous regions, causing substantial damage to transportation infrastructure, natural resources, and human settlements each year. Accurate mass movements inventory mapping and continuous monitoring of unstable slopes are fundamental to hazard assessment and risk management. Although field surveys and optical remote sensing techniques are widely used for mass movements detection, their effectiveness is often constrained by adverse weather conditions, cloud cover, and limited accessibility to rugged terrain. In contrast, Synthetic Aperture Radar (SAR) provides all-weather, day-and-night imaging capabilities, enabling precise detection of surface deformation. Differential Synthetic Aperture Radar Interferometry (DInSAR) has become a valuable tool for mass movements monitoring; however, image pairs are commonly selected using fixed temporal intervals, with limited consideration of rainfall events that frequently trigger slope failures. Therefore, this study aims to detect, monitor, and inventory active mass movements along the Panjab–Bayjan section of the Haraz Road using Sentinel-1 SAR data and DInSAR, while evaluating the effectiveness of an event-based image pair selection strategy.

Materials and methods:

The study area comprises a 16-km section of the Haraz Road between Panjab and Bayjan, one of the most landslide-prone corridors in northern Iran. Sentinel-1 SAR images acquired between 2019 and 2024 were employed. The main methodological innovation of this study is the selection of SAR image pairs based on significant annual rainfall events rather than conventional fixed temporal intervals, allowing rainfall-induced slope deformation to be captured more effectively. Selecting image pairs around significant rainfall events links observations to a plausible mass movements-triggering process and helps examine deformation during periods of increased slope activity. Image processing was conducted in the SNAP environment following the standard DInSAR workflow, including orbit correction, co-registration, interferogram generation, topographic phase removal, interferometric filtering, phase unwrapping, and phase-to-displacement conversion. A coherence threshold of 0.4 was subsequently applied to reduce noise and improve the reliability of deformation measurements. Finally, the detected mass movements were validated using geological evidence, existing mass movements inventory data, and field observations.

Result and discussion:

The results demonstrated that the Sentinel-1 datasets exhibited sufficient coherence over most of the study area, confirming their suitability for interferometric analysis. Time-series deformation analysis revealed considerable temporal variability in slope activity during the study period. The highest negative displacement was observed in 2022, reaching approximately −41.5 cm, indicating intensified landslide activity, whereas 2021 showed the lowest deformation rates, reflecting relatively stable slope conditions. In addition, the largest spatial extent of active mass movements was identified in 2024, affecting approximately 1.4 km of the roadway. Spatial analysis further revealed that most detected mass movements were concentrated within the Shemshak Formation and colluvial deposits, highlighting the controlling influence of geological conditions on slope instability. Validation results confirmed the robustness of the proposed approach, with mass movements maps achieving an overall accuracy greater than 80% and a Kappa coefficient exceeding 70%. Field observations, geological evidence, and existing inventory maps provide complementary references for assessing detected zones. However, DInSAR measures displacement along the satellite line of sight, so viewing geometry and coherence limitations should be considered when interpreting results. These findings indicate that the event-based image pair selection strategy not only improves the detection of active mass movements but also enhances the identification of slow-moving and previously unrecognized slope instabilities.

Conclusion:

The findings demonstrate that integrating Sentinel-1 SAR imagery with DInSAR and an event-based image pair selection strategy provides an effective and reliable framework for mass movements detection, surface deformation monitoring, and mass movements inventory mapping in mountainous environments providing a more realistic representation of mass movements dynamics. The proposed methodology can support continuous landslide monitoring, periodic updating of landslide inventories, landslide susceptibility and hazard assessment, and informed decision-making for infrastructure management and disaster risk reduction in mountainous transportation corridors. Event-informed radar analysis could support repeatable inventory updates where field access is difficult.