Coseismic landslide susceptibility analysis using LiDAR data PGA attenuation and GIS: The case of poás volcano, Costa Rica, Central America

Abstract

A landslide susceptibility model for Poás volcano was created in response to the most recent event that triggered landslides in the area (the Mw 6.2 Cinchona earthquake, which occurred on the 8th of January, 2009). This earthquake was the sixth event related to destructive landslides in the last 250 yr in this area and it severely affected important infrastructures. This chapter refers to a study, which consisted of three phases, as follows: (1) creation of a post-Cinchona earthquake landslide catalog, which was done manually based on a set of high resolution orthophotos and LiDAR data and it includes 4,846 landslides; (2) a landslide susceptibility model, based on the Mora-Vahrson-Mora method, the data from our landslide inventory, and a new modeling of earthquake triggering indicators based on the attenuation of the peak ground acceleration of the event, and (3) an evaluation of the methodology used, which for the Cinchona case resulted in an overlap of the actual landslides and the higher susceptibility zones of ~97%. Based on our new methodology, four landslide susceptibility models were simulated: the Cinchona earthquake, the Mw 5.5 Sarchí earthquake 1912, and two hypothetical earthquakes: one on the Angel fault (Mw 6.0) and the other one on the San Miguel fault (Mw 7.0). The Toro and Sarapiquí river canyons, the non-vegetated corridor located west from the main crater of Poás and the areas where the La Paz Andesites Unit are located are always the zones with the highest susceptibility to slide values. Meanwhile, the northern part of the study area, where the Río Cuarto Lavas unit outcrops, always presented the lowest susceptibility values due to both the low slope angles and weathering level of its rocks.