Abstract

ARISTIZABAL, Edier; VASQUEZ, Mariana  and  RUIZ, Diana. The Susceptibility of Landslide Evaluation and Zoning by Statistical Methods. TecnoL. [online]. 2019, vol.22, n.46, pp.43-64. ISSN 0123-7799.  https://doi.org/10.22430/22565337.1247.

Evaluating and zoning mass movement landslide is a fundamental tool for land planning. There are different methods that help to establish in a regional scale landslide hazard. The most common methods are bivariate statistics and multivariate statistic, which need a landslide historical inventory. This study takes a place in the North of the Andes in a region of Colombia called Aburrá Valley, for evaluating and zoning landslide susceptibility by two methods, one of them is bivariate statistics called Weights of Evidence which is recommended by the Geological Service of Colombia for rural area, and the second one is a multivariate statistic method, called logistic regression, which is widely used worldwide. Both methods are supported in frequency histogram, Pearson correlation, Discriminant Analysis, and Principal Component Analysis. The accuracy of the landslide susceptibility maps produced from the two models is classified in high, medium and low by ROC analysis. The AUC plot estimation results showed that the susceptibility map using Logistic regression has a training accuracy of 76.5% and a prediction capacity of 77.5 %. The Weights of evidence method has the highest training accuracy of 77.8 % and a prediction of 77.5 %. This result allows to include the methods in territorial planning studies.

Keywords : Weight of Evidence; logistic regression; hazard; mass movements; Valle de Aburrá.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )