Votre recherche
Résultats 2 ressources
-
Urban flooding threatens Indian cities and is made worse by rapid urbanization, climate change and poor infrastructure. Severe flooding occurred in cities such as Mumbai, Chennai and Ahmedabad. This has caused huge economic losses and displacement. This study addresses the limitations of traditional flood forecasting methods. It has to contend with the complex dynamics of urban flooding. We offer a deep learning approach which uses the network Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to improve flood risk prediction. Our CNN-LSTM model combines spatial data (water table, topography) and temporal data (historical model) to classify flood risk as low or high. This method includes collecting data pre-processing (MinMaxScaler, LabelEncoder) Modeling, Training and Evaluation. The results demonstrate the accuracy of flood risk predictions and provide insights into flexible strategies for urban flood management. This research highlights the role of data-driven approaches in improving urban planning to reduce flood risk in high-risk areas. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
-
Climate change has increased the frequency and intensity of extreme floods in the Lower Mekong River Basin (LMB). This study leverages the Long Short-Term Memory (LSTM) model to evaluate its performance in predicting river discharge across the LMB and to identify the key variables contributing to flood prediction through SHapley Additive exPlanation (SHAP) and Universal Multifractal (UM) analyses, in a scale-dependent and scale-independent manner, respectively. The performance of the LSTM model is satisfactory, with Nash–Sutcliffe Efficiency (NSE) values exceeding 0.9 for all subbasins when using all input features. The model tends to underestimate the largest peak flows in the midstream subbasins that experienced extreme rainfall events. According to SHAP, soil-related variables are important contributors to discharge prediction, with their impacts partially manifested through interactions with precipitation and runoff. Furthermore, the dominant contributing variables influencing flood prediction vary over time: soil-related variables and vegetation-related variables played a more significant role in earlier years, whereas hydrometeorological variables became more dominant after 2017. The UM analysis investigates the scaling behaviours of contributing variables, showing that hydrometeorological-related variables have a greater influence on predicting extreme discharge across the small temporal scales. Additionally, the UM analysis indicates that the model's performance improves as the temporal variability in extremes of the combined features decreases across 1 to 16 days. Overall, this study provides a comprehensive assessment of the LSTM model's performance in discharge prediction, emphasising the impact of the variability in the extremes of combined features through the scale-independent interpretation. These findings will offer valuable insights for stakeholders to improve flood risk management across the LMB. © 2025 The Authors