Assessment of Advanced Artificial Intelligence Techniques for Flood Forecasting

Flooding is a natural calamity that can destroy people’s lives, infrastructure, and the economy. Forecasting floods is critical for providing people with long-term flood risk management. Flood forecasting is essential in providing early information and knowledge to decision-makers to reduce the impact of flooding. The warning can also be given to potential flood victims and locations, and necessary action, such as mitigation and evacuation, can be taken. With current estimates showing increasing future scenarios, comprehensive flood risk management measures, including flood modelling, are needed. This publication aims to analyses flood risks worldwide. Various AI techniques have been developed and deployed to predict floods and take preventative actions. The primary goal of this study is to assess current improvements in flood forecasting utilizing artificial neural networks (ANNs), adaptive neuro-fuzzy inference systems (ANFIS), support vector machines (SVMs), and k-nearest neighbors (KNNs). As a result, this research presents the most effective short and long-term flood modelling techniques. ANNs, ANFIS, and SVMs are the most successful solutions for forecasting floods. Finally, new research and development directions are suggested to predict floods and take preventative actions.