Parameters and methods used in flood susceptibility mapping: a review

Abstract A correct understanding of the parameters and methods used in flood susceptibility mapping (FSM) is critical for identifying the strengths and limitations of different mapping approaches, as well as for developing methodologies. In this study, we examined scientific publications in the literature using WoS. Although the number of methods used is quite high, the number of parameters used in these methods varies, with a maximum of 21 and a minimum of 5 parameters preferred. It was found that the most commonly used parameter has a preference rate of 97%, but there is no common parameter in 100% of the studies. The methods used for determining flood susceptibility include multi-criteria decision-making (MCDM) methods, physically based hydrological models, statistical methods, and various soft computing methods. Although the use of traditional statistical methods and MCDM methods is already high among researchers, the methods used in flood susceptibility analysis have evolved over the years from traditional human judgments to statistical methods based on big data and machine learning. In the reviewed studies, it was observed that machine learning, fuzzy logic, metaheuristic optimization algorithms, and heuristic search algorithms, which are soft computing methods, have been widely used in FSM in recent years.