3 Spatial prediction of flood-susceptible areas using frequency ratio and maximum entropy models Safura Siahkamari 1 , Ali Haghizadeh 11 , Hossein Zeinivand 1 , Naser Tahmasebipour 1 , Omid Rahmati 1 1- Department of Watershed Management Engineering, Faculty of Agriculture, Lorestan University, Iran (*: Corresponding author) Abstract Modelling the flood in watersheds and reducing the damages caused by this natural disaster is one of the primary objectives of watershed management. This study aims to investigate the application of the frequency ratio and maximum entropy models for flood susceptibility mapping in the Madarsoo watershed, Golestan Province, Iran. At first, a flood inventory dataset was obtained from documentary sources of Iranian Water Resources Department (IWRD) and multiple field surveys. In the locations of 70 flood events recorded in recent decades were; then these flood events were randomly divided into two groups including training (49 cases) and validation (21 cases). The flood conditioning factor layers including slope angle, slope aspect, altitude, plan curvature of the terrain, land use, lithology, distance from the river, topographic wetness index, soil type and drainage density were prepared from the spatial database using ArcGIS 10.2 software. Based on the maximum entropy and frequency ratio methods as well as analysis of the relationship between the flood events belonging to training group and the factors affecting on the risk of flooding, the weight of classes of each factor was determined in a GIS environment. Finally, prediction map of flooding potential was validated using receiver operating characteristic (ROC) curve method. ROC curve estimated the area under the curve for frequency ratio and the maximum entropy models as 74.3% and 92.6%, respectively, indicating that the maximum entropy model led to better results for evaluating flooding potential in the study area. Keywords: Flood susceptibility, Frequency ratio, Maximum entropy, GIS, Iran. 1haghizadeh.a@lu.ac.ir