Open AccessIntegração de dados espaciais em SIG para avaliação da susceptibilidade de ocorrência de deslizamentos
Author(s) -
Eusébio Reis,
José Luı́s Zêzere,
Gonçalo Vieira,
Maria Luı́sa Rodrigues
Publication year - 2012
Publication title -
finisterra
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.242
H-Index - 10
eISSN - 2182-2905
pISSN - 0430-5027
DOI - 10.18055/finis1569
Subject(s) - landslide , hazard , cartography , scale (ratio) , geographic information system , data mining , geography , geology , computer science , geomorphology , ecology , biology
SPATIAL DATA INTEGRATION IN GIS FOR LANDSLIDE SUSCEPTIBILITY PREDICTION. Improper land use is an important factor for the occurrence and intensification of natural hazards. Therefore, a correct hazard zonation is of extreme significance for the adequate planning of human activities. The main objective of this study, applied to a test site in the North of Lisbon (Fanhões-Trancão), is the application of a comprehensive methodology for landslide susceptibility evaluation at a regional scale. The methodological framework is supported by a relational database for GIS processing that includes maps of landslides and of conditioning factors (i.e. slope, aspect/geological structure, lithology, superficial deposits and land use). These maps were produced from detailed field surveys or derived from existing documents. Since the occurrence of distinct types of landslides is influenced differently by the conditioning factors, shallow translational landslides were chosen as an example for the application of the methodology. The susceptibility model is based on statistical and probabilistic algorithms that establish the relationships between the landslides and the conditioning factors (favourability functions), and allow the evaluation and validation of the map layers (independent variables) to be incorporated in the model. The spatial data integration in GIS environment is based in the bayesian interpretation of the favourability function, a technique that ranks the susceptibility values of landslide occurrence from 0 to 1. For model evaluation, following a cross-validation procedure, a random sample of the shallow translational landslides was used.