Floods risk determination through a fuzzy logic system in developing countries. Case study Magdalena River, Colombia

Abstract

Natural disasters around the world, and specifically floods, can generate great economic, en- vironmental, and human life losses in short periods of time. Early warning systems have been developed as a tool to mitigate the impact of these floods, nevertheless, most of this information is dispersed. In the particular case of the Colombian context, the Institute of Hydrology, Meteorology and Environmental Studies of Colombia (IDEAM) has developed a platform called Famine Early Warning Systems Network (FEWS), which presents data of interest for the different stations in the basins of the territory, such as: level, flow, and rainfall. From the height level of water table data of each station, the platform presents three possible alerts: yellow, orange, and red, which are obtained y comparing with reference or threshold levels associated ith ach station. However, the method currently used to determine the alerts does not consider the ariation of he water table level as a function of ime, or the degree of danger represented y a value close to or ending to flood threshold levels. This work proposed a contribution to determine the flood risk level, based on the development of a fuzzy system, taking as inputs the water table level and the level variation as a function of time, to obtain as output the flood risk level in numerical and linguistic terms. For the determination of the risk level, the system applies inference logic rules, involving operations between fuzzy sets represented through mathematical membership functions. Finally, the operation of the fuzzy system was validated using data from the FEWS platform for the hydrological stations of the Bajo Magdalena basin. This project aims to serve as a reference to improve early warning systems of developing couries