A Global Landslide Non-Susceptibility Map_ variation and applicability

Guoqiang Jia 1, Massimiliano Alvioli 2, Stefano Gariano 2, Fausto Guzzetti 2, Qiuhong Tang 1, and Ivan Marchesini 2, 2020, A Global Landslide Non-Susceptibility Map_ variation and applicability, EGU2020 European Geosciences Union General Assembly 2020, online, 4-8/05/2020,
URL: http://www.cnr.it/prodotto/i/425521

Areas non susceptible to landslides are regions where landslides are not expected, or where susceptibility is negligible. Thus they can provide new insights into landslide hazard assessment and land use management and can be targeted as areas for urban planning and dwelling. Non-susceptible landslide areas can be determined with substantially less information as compared with landslide susceptibility. Previous works in Italy and the Mediterranean region and in the USA showed that only morphometric information is needed to distinguish non-susceptible landslide areas. We used 90-m digital terrain data (SRTM DEM V4.1) to calculate global slope and relief maps, and applied globally the quantile non-linear (QNL) model previously obtained in Italy. We define the output map a global landslide non-susceptibility map (GLNSM). The QNL model is a relationship between terrain relief and slope based on an Italian landslide inventory dataset with high completeness and accuracy. Results indicate that 82.89% of the landmasses are non-susceptible areas across the globe, which is more than the percentage of non-mountainous areas (73.6% based on GEO-GNOME). We further considered GLNSM in relation to global climate, elevation, geology, land use, precipitation and seismicity classifications. High percentage (more than 85.0%) of non-susceptible areas are detected in the tropical and arid, flat (low than 500 m), sedimentary, artificial and high vegetated, less rainy (less than 400 mm per year) and seismicity inactive (less than 0.4) regions. Our results of GLNSM was also validated with some well-represented regional landslide inventory datasets, for which we used four national (Austria, China, Ireland and USA) datasets and nine regional (Arizona, Missouri, Oregon, Utah and Washington in USA, Guangdong and Yunnan in China, and Koshi river region in Nepal) datasets. Applicability of GLNSM reveals that 0.7% of non-susceptible areas are covered by artificial structures, about three times of that in susceptible areas (i.e., not non-susceptible areas), while population density of non-susceptible areas are about twice of that in susceptible areas. About 90.5% of population resides in the non-susceptible areas.

Dati tratti da https://intranet.cnr.it/people/