Preparation of flood susceptibility map and its occurrence probability using Shannon entropy model (case study of Firoozabad river basin)

Among the natural hazards, without a doubt, the flood is known as a natural disaster. In this research, Shannon entropy model was used to prepare a flood sensitivity map. First, 34 flood watersheds were selected from Firoozabad basin, and then these 34 points were classified into two groups. With 22 points, 65 percent of the points for training and modeling, and 12 points, 35 percent of the locations that were not used in modeling were used for validation. First, a map of the status of the floods was developed and Then, 10 factors, slope, tilt, lithology, land use, NDVI, SPI, TWI, altitudes, rainfall and distances from the river were selected as flood factors in Firoozabad basin. Prioritizing the effective factors in the occurrence of flood by Shannon entropy index showed that the NDVI layers (2.03), rainfall (0.00), distance from the river (1.89), SPI (385.1), elevation classes (999 (0/19), gradient with weight (0,932), lithology (478/0), TWI (379/0), and land use (280/0), respectively (0/184) have the highest and the least impact Flood events. Based on the results of the ROC curve, the predicted surface area under the curve with 35% of the validation data is equal (91.42%) and for the success rate with 65% of the equal education data (92.53%).