Advanced Flood Classification Using Rapid Machine Learning Techniques: Insights from Saint-Charles Station, Quebec

Flood forecasting is critical for effective water resource management, particularly in regions prone to flooding. This study presents an innovative approach to flood-type forecasting at the Saint-Charles station, located in Quebec, Canada, using an Extreme Learning Machine (ELM) methodology. Our objective was to develop a robust classification model to predict flood types, enhancing preparedness and mitigation efforts accurately. The dataset used spans from 2008 to the end of 2023, encompassing various hydrological parameters, meteorological data, and historical flood records. The ELM, known for its rapid learning speed and minimal computational burden, is applied to classify flood types based on the input features extracted from the dataset. The data preprocessing involves the normalization and handling of missing values to ensure model accuracy. Feature selection is performed to identify the most influential variables contributing to flood occurrences, including precipitation, river discharge, and soil moisture levels. The ELM model is trained and validated using a cross-validation technique to avoid overfitting and ensure generalization. The results indicate that the ELM model demonstrates high classification accuracy and efficiency in predicting flood types at the Saint-Charles station. The model's performance is evaluated using standard metrics such as accuracy, precision, recall, and F1-score. A comparative analysis with other machine learning models highlights the superiority of ELM in terms of speed and predictive capability. This study highlights the potential of ELM as a valuable tool for flood forecasting, providing actionable insights for water resource managers and policymakers. The findings contribute to the body of knowledge in flood management and AI applications in hydrology, paving the way for further research and the implementation of advanced machine learning techniques in environmental monitoring.