No abstract

<p>Both frequency and severity of urban pluvial floods have been increasing due to rapid urbanization and climate change. Hydrological and two dimensional (2D) hydrodynamic models are still too computationally demanding to be used for real-time applications for large urban areas (i.e. flood management scale). As an alternative, data-driven models could be used for flood susceptibility mapping. This study evaluated and compared the performance of image-based models represented by a convolutional neural network (CNN) and point-based models represented by an artificial neural network (ANN), a random forest (RF) and a support vector machine (SVM) with regard to the spatial resolution of the input data. We also examined model transferability. Eleven variables representing topography, anthropogenic aspects and precipitation were selected to predict flood susceptibility mapping. The results showed that: (1) all models were skilful with a minimum area under the curve AUC = 0.87. (2) The RF models outperformed the other models for all spatial resolutions. (3) The CNN models were superior in terms of transferability. (4) Aspect and elevation were the most important factors for flood susceptibility mapping for image-based and point-based models respectively.</p>

The presence of impermeable surfaces in urban areas hinders natural drainage and directs the surface runoff to storm drainage systems with finite capacity, which makes these areas prone to pluvial flooding. The occurrence of pluvial flooding depends on the existence of minimal areas for surface runoff generation and concentration. Detailed hydrologic and hydrodynamic simulations are computationally expensive and require intensive resources. This study compared and evaluated the performance of two simplified methods to identify urban pluvial flood-prone areas, namely the fill–spill–merge (FSM) method and the topographic wetness index (TWI) method and used the TELEMAC-2D hydrodynamic numerical model for benchmarking and validation. The FSM method uses common GIS operations to identify flood-prone depressions from a high-resolution digital elevation model (DEM). The TWI method employs the maximum likelihood method (MLE) to probabilistically calibrate a TWI threshold (τ) based on the inundation maps from a 2D hydrodynamic model for a given spatial window (W) within the urban area. We found that the FSM method clearly outperforms the TWI method both conceptually and effectively in terms of model performance.

Abstract. Data-driven models have been recently suggested to surrogate computationally expensive hydrodynamic models to map flood hazards. However, most studies focused on developing models for the same area or the same precipitation event. It is hence not obvious how transferable the models are in space. This study evaluates the performance of a convolutional neural network (CNN) based on the U-Net architecture and the random forest (RF) algorithm to predict flood water depth, the models' transferability in space and performance improvement using transfer learning techniques. We used three study areas in Berlin to train, validate and test the models. The results showed that (1) the RF models outperformed the CNN models for predictions within the training domain, presumable at the cost of overfitting; (2) the CNN models had significantly higher potential than the RF models to generalize beyond the training domain; and (3) the CNN models could better benefit from transfer learning technique to boost their performance outside training domains than RF models.

Abstract. Data-driven models have been recently suggested to surrogate computationally expensive hydrodynamic models to map flood hazards. However, most studies focused on developing models for the same area or the same precipitation event. It is thus not obvious how transferable the models are in space. This study evaluates the performance of a convolutional neural network (CNN) based on the U-Net architecture and the random forest (RF) algorithm to predict flood water depth, the models' transferability in space and performance improvement using transfer learning techniques. We used three study areas in Berlin to train, validate and test the models. The results showed that (1) the RF models outperformed the CNN models for predictions within the training domain, presumable at the cost of overfitting; (2) the CNN models had significantly higher potential than the RF models to generalize beyond the training domain; and (3) the CNN models could better benefit from transfer learning technique to boost their performance outside training domains than RF models.