Multi-scale InSAR analysis of slow-moving landslides: assessing two-pass interferometric stacking with multi-temporal products

Interferometric Synthetic Aperture Radar (InSAR) is a powerful technique for detecting and monitoring landslides, offering wide-area coverage and high sensitivity to slow ground motion. However, terrain, land cover, and viewing geometry strongly influence measurement reliability, limiting the direct transferability of results. This thesis evaluates the potential and limitations of InSAR-based approaches through three complementary studies spanning regional mapping, lithological comparison, and site-specific analysis, with the aim of improving landslide assessment for infrastructure-related applications. The first study investigates regional-scale landslide detection in the Northern Apennines using Sentinel-1 two-pass interferometric stacking (2015–2022) from ascending and descending orbits. Monthly and annual stacks are analysed alongside geological and geomorphological datasets across Clay Shales, Flysch, and Over-Consolidated (OC) Clays. Results show strong lithological contrasts: clay-rich units exhibit dense, persistent deformation at moderate rates, whereas Flysch shows spatially discontinuous but locally higher, episodic motion. Downslope-projected velocities indicate sustained deformation in clay units and intermittent behaviour in Flysch, with stronger rainfall correlations in clay-dominated terrains. The second study maps 641 deformation signals in the Secchia–Ronco basin and compares them with EGMS Level-2 PS data (2019–2023). Using a ≥5 PS and ≥2 mm/yr rule, EGMS detects only 120 signals (18.7%), with detection decreasing under stricter thresholds. Stack-derived velocities exceed EGMS values, reflecting limited PS coverage in rural areas. EGMS performs best on engineered surfaces, while stacking captures widespread deformation. The third study applies a multi-source InSAR framework to the Calciano landslide, integrating stacking, PSIG time series, EGMS, and GNSS/inclinometer validation. Results show consistent temporal evolution, including pre-2017 acceleration followed by persistent lower-rate deformation. Overall, interferometric stacking provides robust spatial detection, while multi-temporal InSAR and ground data enhance temporal interpretation, supporting integrated workflows for landslide risk assessment.