Assessment and Comparison of CART and TreeNet models to Landslide Susceptibility Mapping using SPM Software and Geographic Information System (GIS), (Case study: Kameh Watershed, Southern of Isfahan Province)

Landslide is one of the environmental hazards that every year cause loss of life and financial losses in large parts of the mountainous regions of Iran. Therefore, identifying important and effective factors for the occurrence of this phenomenon can be used as a practical tool in reducing possible losses.  Modeling the susceptible area into landside hazards can be a useful tool for the police makers for protecting the prone areas. The Kameh watershed in the south of Isfahan province has been selected as one of the areas prone to landslides due to its special topography, geology, lithology, climate, and social-economic situation. This research aims to identify the effective factors in creating the phenomenon of landslides and to determine the areas with a high risk of landslides by applying the CART and TreeNet models with using of SPM and ArcGIS 10.8 software in the studied area. We have used 11 factors that affect the occurrence of landslides including geology, amount of precipitation, slope, distance from the river, TWI, land use, NDVI, elevation, aspect, distance from the faults, and distance from roads, as the independent variables and 311 landslides (as a polygon) collected from the study area were used as dependent variables. For running our models, 70% of the landslides were used as training and the remaining 30% of the landslide points were used for validation. The results of this study, according to the various statistical indicators that have been applied, show a higher accuracy of the CART model than the TreeNet model. The output of the model is a landslide peril map with 5 classes: very high, high, medium, low, and very low. It would be worth the managers and policymakers to use these results for protecting the susceptible areas.