Assessment of Shallow Landslides Susceptibility Using SHALSTAB and SINMAP at Serra Do Mar, Brazil

The Serra do Mar mountain range in Brazil is highly susceptible to shallow landslides initiation, due to its relief and high rainfall index. The aim of this study is the application and comparative performance analysis of SHALSTAB and SINMAP in the landslide susceptibility assessment of the Pereque watershed at the Serra do Mar escarpments. The application of physically based models is an objective method used in landslide susceptibility to predict slope failure under different geotechnical scenarios, supporting hazard assessment studies. Model calibration is based on the landslide scars inventory of the 1985 and 1994 events, while topographic parameters are DEM sourced and geotechnical parameters obtained from soil samples. Using performance classifiers based on a contingency table to assess model performance, the results indicate that SHALSTAB is the best-fit model at the chosen scale due to higher accuracy and higher concentration of scars in unstable areas (>62%). SINMAP was less accurate, also exhibiting higher false positive results and lower density of landslides scars in unstable areas. For future studies, the compartmentalization of the study area according to geology is suggested as an improvement in the representativeness of modeling results.