Mapeamento da suscetibilidade a escorregamentos translacionais rasos na bacia do Rio Pirajibú-Mirim, Sorocaba-SP: aplicação dos modelos SHALSTAB e de Rede Neural Artificial (RNA) na estimativa de atributos físicos do solo.
Melo, Carrie Peres
MetadataShow full item record
The susceptibility to mass movements on slope areas can generate risks of natural disasters. So the study is of great importance, with interest in mitigating damage to life, material, economic, environmental, and especially in populated places, where it can affect people, and the social environment in general. The objective of this research is to identify the classes of susceptibility to shallow translational landslides in the basin of the Pirajibú-Mirim River, city of Sorocaba - SP to achieve this goal was used the Artificial Neural Network (ANN) was to determinate soil physical properties. The morphometric maps were made from images of Alos Palsar and SRTM (Shuttle Radar Topography Mission). These images have a resolution of 12.5 and 30 meters, respectively. The software used is free and open-source SAGA GIS and QGIS. The procedures were: mapping of the relief attributes (slope, aspect, texture, convexity, and the catchment area). Secondly, the classification of the land topography generated sixteen terrain classes. To identify the landslide susceptibility areas was applied the SHALSTAB model, using attributes of the slope, catchment area, bulk density, cohesion, hydraulic conductivity, angle of friction, and soil thickness. To determine the angle of internal friction, and cohesion of soils the ANN technique was used. Like the permeameter of Guelph, has been created to determine the hydraulic conductivity of soils in situ. AHP (Analytical Hierarchy Process) was applied to evaluate and visualize the urban areas at risk. The results achieved by ANN techniques, AHP, and the SHALSTAB model were satisfactory for obtaining the map of the areas susceptible to shallow translational landslides, both in urban areas and in more withdrawn sites. The artificial neural network technique allowed for the spatialization of the SHALSTAB parameters, demonstrating the increase in data quality and with this results closer to reality. The contribution of this research concerns the increase of data and provision of results. Support the planning and operation of preventive and mitigating actions in these locations as well as stimulating the increased use of free and open-source software.
The following license files are associated with this item: