Abstract:

When examined in terms of the number of occurrences in our country, one of the natural disasters which we encounter frequently is landslides. slope, aspect, elevation, curvature, topographic wetness index, NDVI, annual total precipitation, proximity to fault lines, lithology and land use status parameters were used in the study. Parameters were correlated and analyzed with the landslide inventory map using the frequency ratio method, the weight of evidence method and the logistic regression method. landslide susceptibility maps were produced as a result of the analysis. The maps produced are divided into 5 classes as "very low, low, medium, high and very high". The ROC curve and AUC values were used to evaluate the performance of the models created as a result of the analysis, the compatibility of the old landslide events with the risk groups was examined in the evaluation of the model accuracy. As a result of the study, accuracy rates of 82.9% for the Frequency Ratio Method, 83.5% for the Weight of Evidence Method and 77.1% for the Logistic Regression Method were obtained.

Keywords: