Landslides are a common natural hazard worldwide with greater socioeconomic impacts in developing and tropical countries. In Central America and Costa Rica, this phenomenon is mainly triggered by seismicity and extraordinary rainfall. In order to portray the damaging landslides, that caused human and material losses in Costa Rica, DesInventar disaster database was used to analyze damaging landslides reports from 1970 to 2018. Moreover, different generalized linear models (GLM) were performed to analyze the landslide hazard, vulnerability and risk in all the municipalities of the country. An Akaike Information Criterion (AIC) backward selection was used to contrast and determine the best hazard and vulnerability models. From a total of ten variables, terrain ruggedness index, 5 year intensity-duration-frequency precipitation curves and earthquakes distribution determined the landslide hazard. Otherwise, population, municipality area and Social Development Index are the most suitable variables to explain the landslide vulnerability. Subsequently, the multiplication between alternative landslide hazard and vulnerability indexes produced the risk index. Consequently, the highest risk values were obtained for large and rural municipalities (Pérez Zeledón, San Carlos, and Turrialba) as well as for densely populated and urban units (Alajuela, Desamparados, and Cartago. Results are critical for disaster risk reduction public institutions and academic stakeholders. Therefore, this methodology could be an interesting opportunity for different tropical and developing countries to achieve national or regional analyses of the most important risk component in each municipality and implement risk reduction strategies adapted for each municipality characteristics.
Read full abstract