Determination of landslide susceptibility in linear infrastructure. Case: aqueduct network in Palacé, Popayan (Colombia)

This research aimed to predict the occurrence of mass movements in the aqueduct network of Palacé, in the municipality of Popayan (Colombia). We evaluated the quality of SRTM and ASTER digital terrain models by comparing them with contour lines using a map scale of 1: 25000. The landscape parameters...

Full description

Autores:
Correa Muñoz, Nixon Alexander
Higidio Castro, Jhon Fander
Tipo de recurso:
Article of journal
Fecha de publicación:
2017
Institución:
Universidad Nacional de Colombia
Repositorio:
Universidad Nacional de Colombia
Idioma:
spa
OAI Identifier:
oai:repositorio.unal.edu.co:unal/67574
Acceso en línea:
https://repositorio.unal.edu.co/handle/unal/67574
http://bdigital.unal.edu.co/68603/
Palabra clave:
62 Ingeniería y operaciones afines / Engineering
Geomorphometry
mass removal processes
landscape parameters
logistic regression
accuracy
Geomorfometría
procesos de remoción en masa
parámetros del terreno
regresión logística
exactitud
Rights
openAccess
License
Atribución-NoComercial 4.0 Internacional
Description
Summary:This research aimed to predict the occurrence of mass movements in the aqueduct network of Palacé, in the municipality of Popayan (Colombia). We evaluated the quality of SRTM and ASTER digital terrain models by comparing them with contour lines using a map scale of 1: 25000. The landscape parameters derived from the SRTM-DEM were analyzed with a multivariate procedure using algorithms implemented in free software, along with thematic information of the study area (coverage, distance to faults, rivers and precipitation). We selected non-redundant variables with the non-parametric ACP technique, and obtained a susceptibility prediction model using logistic regression, with two types of variables: dependent (landslides inventory from field observation) and independent (slope, slope length factor, topographic wetness index, flow path length, soil units and rate of convergence) resulting in a susceptibility map, reclassified into categories according to the values of probability. The prediction model could not be quantitatively assessed because of the absence of studies with a semi-detailed scale, but the estimation of the mean square error of elevation, from which the terrain parameters were derived, the level of detail and the performance of the classifier with ROC curve, yielded a zoning consistent with the findings of the field visits.