Performance of Chaos theory on natural systems in landslide hazard zonation in Fahlian river basin

Catchments are natural systems in which natural phenomena like landslides are considered as natural disasters. As a chaos factor, they have a main role in forming unstable condition, tackling energy, accelerating erosion and generating sediment. So factors causing slope instability are reaction of the system to positive feedback. This study aims to employ theory of natural system chaos, in the form of Shannon entropy index, to find the factors causing landslide and its hazard zonation in Fahlian basin. First, input layers including geology, rainfall, slope, aspect, land use, distance to river, distance to fault, and elevation were digitized using GIS techniques. Then occurred landslides were detected using satellite images and field study. Landslide hazard zonation based on defined weights of each parameter was generated. In order to run model and study its accuracy, receiver or relative operating characteristic (ROC), was used with 70 and 30 per cent of data as training and test, respectively. Results show that slope and aspect have the maximum effect on landslide occurrence with ultimate weight of 0.662 and 0.308, respectively. Landslide susceptibility zonation map show that more than half of study area (56.97 percent) have very high to high susceptibility. Disaggregation of areas with SCAI method show the high accuracy of the model in detection of area with average, low and very low susceptibility. Frequency ratio of hazard classes deals with high accordingly, area under curve (AUC) of ROC was estimated 0.87 with 0.026 standard deviation which is known as very good accuracy of model.