Coseismic and Post-Seismic Landslide Mapping in Mila, Northeast Algeria, Using Sentinel-1 DInSAR and Coherence Change Detection

Abstract

On August 7, 2020, an Mw 5.0 earthquake struck the Mila region in northeastern Algeria, triggering widespread slope failures that damaged urban areas, agricultural land, and road networks. This paper presents a Sentinel-1A Differential Interferometric SAR (DInSAR) and Coherence Change Detection (CCD) analysis of the resulting earthquake-induced landslides, using a 12-day pre- and post-event image pair (August 3–15, 2020). The differential interferogram reveals complex deformation patterns across multiple municipalities. Line-of-Sight (LoS) displacement mapping shows a maximum displacement of ~2.5 m in El Kherba, with significant deformation (>0.5 m) spanning ~2.5 km². Two-dimensional decomposition of the LoS vector using SRTM-derived slope and aspect resolves a vertical displacement component of up to 7.5 cm. Coherence loss (γ < 0.2) covers ~15 km² and provides spatial validation independent of phase unwrapping. West-to-east displacement profiles across El Kherba and Grarem Gouga show a clear topographic threshold: displacements increase abruptly above ~20° slope, reaching 9.5–10.2 cm in Zone A (slope > 30°). The spatial agreement among DInSAR deformation zones, low-coherence areas, steep marl slopes, and field-documented surface cracks confirms the proposed methodology's reliability for rapid post-seismic landslide assessment in regions with sparse ground-based monitoring.