Building Damage Characterization for the 2016 Amatrice Earthquake Using Ascending–Descending COSMO-SkyMed Data and Topographic Position Index

Abstract

The August 24, 2016, M 6.2 Amatrice earthquake in central Italy was one of the most destructive earthquakes in recent years, and it caused considerable human losses. The goal of this study is to distinguish between collapsed and standing buildings in the town of Amatrice using synthetic aperture radar (SAR) data acquired from the COSMO-SkyMed satellite. Data from both ascending and descending orbits in single HH polarization were evaluated separately. Two main representative change indices, SAR intensity (SI) and SAR coherence, were combined using multitemporal discriminant and fuzzy analyses. To enhance the contribution of SI in the discriminant analysis, the aggregate analysis, in a raster-based environment for each intensity dataset, was taken into account. By integrating the ascending and descending datasets, two models based on discriminant analysis and fuzzy logic were created, and different damage proxy maps (DPM) for the region were presented. The accuracy comparison of single-path discriminant analysis and the new integrated models indicates improvement, which implies that the combination of west-to-east and east-to-west observations can reduce some limitations of SAR imagery (e.g., shapes of roofs, shadows) on building classification. In addition, topographic position index (TPI) analysis was conducted as a spatial-based scale-dependent approach, and it revealed more aspects of possible damage in the town. The results of DPM and TPI indicate that 55% of the collapsed buildings are located in valleys, which may have caused seismic wave amplification in this region.