Landslide risk zoning using ANP network analysis process and ANN artificial network in GIS environment (case study: Rodbal river watershed)
Subject Areas :mohammad ebrahim afifi 1 * , Mahen Sadrizadeh 2
1 - Assistant Professor of Islamic Azad University, Larestan branch
2 - Senior expert in natural geography and principal of Lar high schools, Larestan, Iran
Keywords: Landslide zoning, neural network, fuzzy logic, AHP model, Bal river basin.,
Abstract :
Landslide is one of the most common geological hazards in the world, which always causes loss of life and financial damage. Identifying landslide-prone areas through hazard power zoning with suitable experimental models is one of the primary measures in reducing possible damages and risk management. Due to its geographic location and natural features, Rodbal Basin, which is located in Darab County, Fars Province, is one of the most prone areas in the country to the occurrence of landslides and other phenomena related to landslides. In this research, we evaluated the risk of landslides in Rodbal Darab basin by using neural network method, AHP, ANP models and fuzzy logic with OR, And, Sum, Product and Gamma operators. A total of 9 main criteria related to the occurrence of landslide phenomenon including lithology, distance from fault, slope, direction of slope, slope, distance from road, height, distance from waterway and land use were analyzed. These criteria were classified as factor maps, each one separately and valued using statistical methods. The final map produced for landslide risk zoning in the basin showed that a total of about 70% of the basin is at very high and high risk of landslides. Also, the existing landslides were mapped through ground observation as a ground witness (danger reference) which showed high adaptability with the neural network method and the combination of fuzzy logic with 0.3 operator and ANP.