Identificação De Áreas Suscetíveis a Deslizamentos Com Lógica Fuzzy: Estudo De Caso No Município De Teófilo Otoni, Brasil

L Landslides cause significant socio-environmental damage. The complexity of mitigating these events can be reduced by geotechnologies such as Fuzzy logic. Therefore, this study aimed at developing a landslide susceptibility map for the municipality of Teófilo Otoni using GIS and Fuzzy logic. The variables considered were slope, geomorphology, lithology, precipitation, proximity to roads, land use and land cover, landslide scars, and residential areas. The results indicated that 63.89% of the municipality’s area is classified as high or very high risk, with 80.06% of residences located in these areas. Model validation showed that 84.17% of the landslide scars overlapped with high or very high-risk areas, demonstrating strong predictive accuracy. Comparisons with other models confirmed greater reliability using Fuzzy logic, with an AUC-ROC of 0.9143, standing out as the most accurate in predicting landslide-susceptible areas, when compared to other models developed by IBGE and CPRM. Therefore, it is concluded that Fuzzy logic proved to be an effective tool in identifying landslide-prone areas, enabling more precise interventions.