Landslide susceptibility assessment in mountainous and tropical scarce-data regions using remote sensing data : a case study in the Colombian Andes
Fecha
2017
Autores
Ruiz Vásquez, Diana
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad EAFIT
Resumen
Landslides triggered by rainfall are one of the most frequent causes for natural disasters in the
tropical and mountainous countries, such as Colombia. However landslide susceptibility
assessments are often limited due to the scarcity of reliable observations and available information,
particularly in remote high-mountain regions. Although Colombia is a tropical and mountainous
terrains dominated by landslide prone region, it has little availability of data for landslide
susceptibility assessment. This study presents the application of a logistic regression model to
assess landslide susceptibility in the La Liboariana catchment. It is a basin on a tropical inaccessible
terrain in northern Colombian Andes, where on May 18th, 2015, more than 40 landslides and an
associated flash flood and debris flow afterwards killed 104 inhabitants. The applied approach is
based on free access remote sensing tools to complete and obtain the missing landslide causative
factors. To select key factors related to landslide occurrence the prediction and successes
performance of the susceptibility maps for each combination of landslide causative factors was
estimated using the Receiver Operating Characteristics (ROC). The results show that only three
factors gave the best predicting accuracy. All the factors were obtained by free remote sensing
tools, indicating they can provide enough information to achieve a successful approach to landslide
susceptibility assessment in complex terrains as the study area. This suggests that the proposed
approach could be implemented in several tropical regions with similar characteristics based only in
free access information.