Spatio-temporal cluster analysis of recent Italian landslides

The present study addresses the pattern distribution of recent landslides in Italy. The main objective is the detection and mapping of spatio-temporal clusters of landslides that occurred in the period 2010-2017 in the country. To this aim, a subdivision of the national area into 158 warning zones, as identified by the 21 civil protection regional centres to deal with weather-induced hydro-geological hazards, is adopted. Information on landslides comes from FraneItalia, a geo-referenced catalogue developed consulting online news sources. Analyses are performed both at national scale and at a regional scale, focusing on the Campania region. The space-time permutation scan statistics model is applied to detect statistically significant clustering, accounting for the geographical spatial dimension and for the temporal dimension. Two types of analyses are performed: annual, considering each single year; and multi-annual, encompassing the entire 8-year study period. In both cases, spatio-temporal cluster analyses are able to detect areas and frame-periods characterised by relevant and recurrent landslide activity. Finally, the obtained results are compared with a standard landslide density map, highlighting the complementarity of the two approaches.

Automated summary

This study investigates the pattern distribution of recent landslides in Italy. The space-time permutation scan statistics model is used to detect statistically significant clusters of landslides. The results are compared with a standard landslide density map, demonstrating the complementarity of the two approaches.