Spatial Mapping of Extreme Precipitation Events Using Artificial Neural Networks

In recent decades, floods have emerged as significant concerns with profound implications for human lives and infrastructure. In order to mitigate potential casualties, the development of early warning systems is crucial for both urbanized and rural regions. However, the task of developing accurate models for complex hydrosystems with heterogeneous and nonlinear precipitation patterns can be challenging and not always straightforward. In this context, the development of an Artificial Neural Network (ANN) model to predict flood risk is of great importance. ANN models have gained significant attention in hydrological studies due to their ability to capture complex non-linear relationships between input variables and output predictions. Therefore, this work aims to propose an accurate and reliable spatial mapping of extreme precipitation events in order to identify high risk areas, with a specific application to a rural semi-arid area in Tunisia.