New Expert System for Short, Medium and Long- Term Flood Forecasting and Warning

Floods are among the most powerful forces on the planet. Therefore, the need for a system that predict and warn about the flood occurrence is required. The combination of several approaches into a single system was our concern to design and develop an intelligent system that meets the demands of our research. After many studies, we decided to work with multi-agent systems to benefit of its advantages in terms of the distributed artificial intelligence, and we worked with expert systems to benefit of the concept of logic programming and the concept of facts and rules. In this paper, we present an expert system for real-time flood forecasting and warning that consists of two levels of processing. A first level for short-term forecasting and warning ie. for a time that does not exceed two to three days, using the proposed model, which is based on coefficients that will be calculated to do the flood forecasting and warning. A second level for medium (the time could be up to 10 years) and long-term (the time could be up to 1000 years) forecasting and warning via the empirical model of Hazan-Lazarevic.