Determining the priority of effective factors in flood occurrence and preparing a zoning map of flood occurrence using a new random forest algorithm(Case study: Lavasanat watershed)

Floods are one of the three main natural hazards in Iran, and it is safe to say that significant floods occur at least once a year in this part of the country. According to studies, 40 small and large floods occur annually in all parts of the country. This highlights the need to study the factors influencing the occurrence of floods. The purpose of this study is to zoning flood prone areas in the region and determine the priority of factors affecting flood occurrence using a random forest algorithm. Thus, according to the studies and research background, 9 land use indicators, distance from the river, slope, height, humidity index, river flow, waterway capacity, precipitation, and land curvature were selected. After determining the variance inflation factor and tolerance coefficient, in the next step by entering the data related to the effective factors to the R software, modeling was performed using a random forest algorithm and the role of effective factors in flood occurrence in each sub-basin was determined. Flood risk zoning was prepared in four very dangerous, dangerous, medium and low risk zones in ARC / MAP 10.2 environment. The results show. . According to the research results in all sub-basins, rainfall is the most important factor in the occurrence of floods in the region. In the northern and western basins of the Dubai River region and the slope of the next ranks have played a role in the occurrence of floods, while land use in the southern sub-basins of the region is an important factor in the occurrence of floods. Also, according to the results, more than half of the area is located in a very dangerous and dangerous area. Most of the dangerous and very dangerous zones cover the northern, northwestern and western parts of the region, which include the sub-basins of Fasham, Meghun and Ahar.