Landslide Susceptibility Zonation Based on Slope Units and Grid Units Using the Weight of Evidence and Frequency Ratio Methods in the Kulon Progo Mountains, Indonesia

The Kulon Progo Mountains in the south of Java Island have complex geological conditions and are highly susceptible to landslides, resulting in significant material losses and threats to human safety. In recent years, the frequency of landslides in Kulon Progo has increased markedly, from 208 events in 2020 to 622 in 2022, with estimated material losses reaching 450 million USD in 2022. To reduce these risks, it is essential to identify landslide susceptibility zones. This study uses grid and slope units to perform landslide susceptibility zonation. The weight of evidence (WoE) and frequency ratio (FR) methods were applied to compare model accuracy. Nine landslide-controlling factors, including slope, aspect, elevation, slope material, lineament density, Normalized Difference Vegetation Index (NDVI), distance to rivers, distance to roads, and rainfall, were analyzed using spatial methods and landslide event data. The results show that slopes steeper than 40° and high lineament density (> 1.61 km/km²) are the most significant factors influencing landslide susceptibility in both units. Validation using the Area Under Curve (AUC) indicates that the slope unit-based WoE method provides better accuracy, with AUC values of 0.812 for the success rate and 0.827 for the prediction rate. The study produced four landslide susceptibility zonation maps, classifying areas from very low to high susceptibility. High susceptibility zones are concentrated in the northern part of the study area, especially on steep slopes, with some areas in the southern part.