Landslide Hazard Analysis Using Frequency Ratio Model

In the north part of Iran (Alborz Mountain belt), landslides occur frequently due to climatologic and geologic conditions and high tectonic activities, that results, annually, millions of dollars financial defect excluding casualties and unrecoverable resources. The reliable hazard map would help to mitigate the consequences of landslide occurrences by land-use management and other strategies. This paper evaluates the hazardous area in Marzan Abad (Central Alborz, North part of Iran) using probabilistic–Frequency ratio (PFR) model, Geographic Information System (GIS) and Remote sensing techniques. Hazardous areas have been analyzed and mapped using the landslide occurrence factors by frequency ratio model.In GIS platform, layers such as geology, geomorphology, soil, slope, aspect, elevation, annual precipitation, land use, distance from faults, lineaments, roads and drainages were displayed, manipulated and analyzed. The validation of hazard map has been estimated with the validation group of actual landslides and rate curves method. The Area Under the Curve (AUC) evaluates how well the method predicts landslides. The results have showen satisfactory agreement between prepared hazard map and existing data on total landslide locations (93.60%) and validation group of landslide locations (91.68%) So, the methodology used in this study was validated. Final hazard map classified in five hazardous classes (very high, high, moderate, low and non hazardous area). Receiver Operating Characteristic curve method (ROC curve) was used to validate the classification and based on its area under the curve value, final classification was evaluated as excellent classification (AUC=0.94). This study evaluates geology, soil and distance to road networks as the most effective factors on landslide analysis and deep valleys, old landslide traces, area near the roads and faults as the most hazardous areas for landslide occurrence in Marzan Abad area