Assessing landslide susceptibility using ANN and ANFIS to forecast landslides in Sumatera Indonesia

Abstract

The purpose of this study was to use data collected from actual landslide events between 2008 and 2018 in models to assess landslide susceptibility and to accurately forecast landslides in Sumatra, Indonesia. Artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) modeling were compared. A digital elevation model (DEM) was used to generate data on elevation and slope. The neural network simulations were tested using a dataset from 2019, yielding a match greater than 80% with actual landslides. The accuracy and compatibility of ANN and ANFIS were compared using the 2019 landslide. Seismic activity, a parameter indirectly impacting landslides that are often ignored in probability models, was used. Precipitation, soil type and texture, and land cover were also used. The resulting landslide susceptibility map for 2008 to 2018 divides Sumatra into three zones; (1) high risk, (2) intermediate-risk, and (3) low risk.