Improving landslide inventories by combining satellite interferometry and landscape analysis: the case of Sierra Nevada (Southern Spain)

Abstract

AbstractAn updated and complete landslide inventory is the starting point for an appropriate hazard assessment. This paper presents an improvement for landslide mapping by integrating data from two well-consolidated techniques: Differential Synthetic Aperture Radar (DInSAR) and Landscape Analysis through the normalised channel steepness index (ksn). The southwestern sector of the Sierra Nevada mountain range (Southern Spain) was selected as the case study. We first propose the double normalised steepness (ksnn) index, derived from the ksn index, to remove the active tectonics signal. The obtained ksnn anomalies (or knickzones) along rivers and the unstable ground areas from the DInSAR analysis rapidly highlighted the slopes of interest. Thus, we provided a new inventory of 28 landslides that implies an increase in the area affected by landslides compared with the previous mapping: 33.5% in the present study vs. 14.5% in the Spanish Land Movements Database. The two main typologies of identified landslides are Deep-Seated Gravitational Slope Deformations (DGSDs) and rockslides, with the prevalence of large DGSDs in Sierra Nevada being first revealed in this work. We also demonstrate that the combination of DInSAR and Landscape Analysis could overcome the limitations of each method for landslide detection. They also supported us in dealing with difficulties in recognising this type of landslides due to their poorly defined boundaries, a homogeneous lithology and the imprint of glacial and periglacial processes. Finally, a preliminary hazard perspective of these landslides was outlined.