Susanna Dazzi

The extreme rainfall event that affected the Emilia-Romagna region (Italy) between 16 and 17 May 2023 triggered widespread flooding and multiple levee breaches along the Senio River, leading to extensive inundation of agricultural and peri-urban areas. This study presents a high-resolution two-dimensional (2D) hydrodynamic model based on the parallel GPU-accelerated solver PARFLOOD, coupled with a morphodynamic model to explicitly simulate multiple levee breach formation and evolution during the May 2023 event. The model reconstructs the flood wave using observed hydrometric data and consistent hydraulic boundary conditions. A grid resolution of 1 m enables detailed representation of breach widening, crest lowering, erosion dynamics, and inundation processes. The modelling framework allows quantitative assessment of overflow volumes, flood extent, water depth distribution, and flood arrival times, and is supported by a comprehensive dataset including hydrometric records, high-resolution pre- and post-event topography, flood extent mapping, and eyewitness observations. Results show that morphodynamic feedback strongly controls flood volume partitioning and inundation timing. A substantial portion of the total flood volume is released through breaches; however, return-flow breaches are also observed, whereby part of the water previously spilled into the inundated areas re-enters the river channel. This behaviour leads to complex hydraulic interactions between the river and the surrounding areas, including non-monotonic discharge evolution along the river reach. The analysis highlights the importance of multiple simultaneous levee breaches, a condition that is rarely documented in the literature and challenges traditional flood hazard assessment approaches based on independent breach scenarios. The study shows that a physically based breach approach, coupled with a 2D shallow water model, provides a robust framework for flood reconstruction and offers new insights for residual flood risk assessment in embanked low-gradient river systems.

1. In the context of flood risk management, real-time forecasting can support Early Warning Systems, which are useful to reduce casualties and economic losses. However, in case of levee-breach inun...

The extreme rainfall event that affected the Emilia-Romagna region (Italy) between 16 and 17 May 2023 triggered widespread flooding and multiple levee breaches along the Senio River, leading to extensive inundation of agricultural and peri-urban areas. This study presents a high-resolution two-dimensional (2D) hydrodynamic model based on the parallel GPU-accelerated solver PARFLOOD, coupled with a morphodynamic model to explicitly simulate multiple levee breach formation and evolution during the May 2023 event. The model reconstructs the flood wave using observed hydrometric data and consistent hydraulic boundary conditions. A grid resolution of 1 m enables detailed representation of breach widening, crest lowering, erosion dynamics, and inundation processes. The modelling framework allows quantitative assessment of overflow volumes, flood extent, water depth distribution, and flood arrival times, and is supported by a comprehensive dataset including hydrometric records, high-resolution pre- and post-event topography, flood extent mapping, and eyewitness observations. Results show that morphodynamic feedback strongly controls flood volume partitioning and inundation timing. A substantial portion of the total flood volume is released through breaches; however, return-flow breaches are also observed, whereby part of the water previously spilled into the inundated areas re-enters the river channel. This behaviour leads to complex hydraulic interactions between the river and the surrounding areas, including non-monotonic discharge evolution along the river reach. The analysis highlights the importance of multiple simultaneous levee breaches, a condition that is rarely documented in the literature and challenges traditional flood hazard assessment approaches based on independent breach scenarios. The study shows that a physically based breach approach, coupled with a 2D shallow water model, provides a robust framework for flood reconstruction and offers new insights for residual flood risk assessment in embanked low-gradient river systems.

This contribution presents a systematic framework for managing residual flood risk in embanked fluvial systems, focusing on the Enza River (Italy), a right-bank tributary of the Po River. Even with planned structural and maintenance measures, the fluvial system cannot safely convey extreme flood events (e.g., 500-year floods). Under these conditions, controlled overflows implemented through engineered spillways offer a robust risk-mitigation strategy, enabling the controlled release of floodwaters and reducing the consequences associated with accidental levee failure.The proposed approach integrates two-dimensional hydrodynamic simulations with the PARFLOOD model to delineate levee segments susceptible to overtopping, support the iterative optimization of spillway location and design parameters, and simulate flood inundation resulting from both uncontrolled levee breaches and controlled overflow conditions. Impact analyses are carried out using advanced tools developed under the MOVIDA project to quantify potential damage to population, infrastructure, and economic assets.The analysis of multiple flood scenarios (ranging from uncontrolled breaches to controlled overflow configurations, with and without complementary mitigation measures) demonstrates the strong potential of controlled overflows through engineered spillways to reduce flood impacts. The results indicate that controlled overflows can reduce inundated areas by up to 80% and direct economic losses by up to 96%, while substantially decreasing population exposure from approximately 7,900 to 64 individuals.These findings highlight the effectiveness of controlled overflows as a key element of residual flood risk mitigation, particularly when combined with conventional structural interventions. Such an approach enhances system adaptability and supports anticipatory, risk-informed floodplain management, representing a shift from passive flood defense toward proactive resilience-based planning.

In lowland areas that are susceptible to flooding, for example due to river levee breaches, the presence of linear elements such as levees of minor channels or road/railway embankments can significantly influence the inundation dynamics. Being more elevated than the surrounding land, these embankments can provide an obstacle to flood propagation, reducing the flood extent, but sometimes increasing the water depths. To obtain accurate flood simulations, and hence reliable flood hazard maps, the correct representation of embankments in the computational domain is therefore crucial.Numerical simulations for flood hazard assessment usually rely on the unrealistic assumption that the minor embankments in the floodable area are able to withstand the interaction with floodwaters without collapsing. However, especially when overtopped, these elements can be eroded or collapse, no longer limiting the flood propagation. The assumption of “non-erodible” embankments can thus lead to misestimating the flood hazard, but studies in literature about this issue are lacking.In this work, a preliminary analysis on how flood hazard mapping in lowland areas can be influenced by the possible failure of minor embankments is carried out, focusing on a case study in Italy. The results of two-dimensional numerical simulations of flood scenarios performed considering the minor embankments either as non-erodible or as erodible are compared. The most important problem in simulating the case of erodible embankments is the difficulty in including failure criteria to predict their collapse. While this is indeed a limitation for roads and railways, for which the large variability in building materials and structure prevents the definition of reliable failure criteria, it can be presumed that the available models developed to predict the breaching of earthen dams/levees can be applied to the levees of irrigation and drainage channels as well. For this reason, the study focuses on an area that is crossed only by the earthen levees of minor channels, and a simple erosion model that can automatically predict the breaching of these elements due to overtopping is adopted. The scenario considered for the analysis is the inundation induced by a levee breach in a nearby river.Results show that, when assuming “non-erodible” embankments, the flood extent can be underestimated, while the maximum water depth and the flood hazard classification can be overestimated upstream of the erodible embankments and underestimated downstream. Moreover, the flood arrival time can be anticipated in the downstream areas. Overall, despite being case-specific, the analysis suggests that the unphysical assumption of “non-erodible” embankments in lowland areas can significantly influence the flood hazard assessment, possibly underestimating it, and this limitation should be kept in mind by flood risk management authorities.This work is part of the project "MORFLOOD" (PNRR-M4C2 - I1.1-Avviso MUR n.104 del 02-02-2022 - PRIN2022 - Project code 2022SJ2NJ9 - CUP Code D53D23004860006 - Funded by the European Union-NextGenerationEU).

Physics-informed neural networks (PINNs) have recently been developed as a novel solution approach for physical problems governed by partial differential equations (PDEs). Compared to purely data-driven methods, PINNs have the advantage of embedding physical constraints in the training process, thus increasing their reliability. Compared to traditional numerical methods for PDEs, PINNs have the advantage of being “meshless”; they are in general less accurate and more computationally expensive, but also more suitable to sparse-data assimilation and to inverse modelling, which is increasing their popularity in many scientific fields. However, hydraulic applications of PINNs in the context of free-surface flows are still in their infancy.In this work, the effectiveness of PINNs to model one-dimensional free-surface flows over non-horizontal bottom is tested. The governing PDEs are the shallow water equations (SWEs), which represent the mass and momentum conservation in free-surface flows. The inclusion of a spatially variable topography in a meshless method such as PINNs is not trivial. Here, the idea of solving the augmented system of SWEs with topography is exploited. Augmentation consists in treating the bed elevation as a conserved variable (together with water depth and unit discharge) and adding a fictitious equation to the system, which states that this variable is constant in time (i.e., its time derivative is null), while it can be variable in space (its space derivative is included in the bed slope source term). In this way, bed elevation can be easily provided with other initial conditions, and the fixed-bed constraint preserves its value in time.Different cases of unsteady flows with flat and non-flat bottom are considered, and the accuracy obtained using PINNs with augmented SWEs is checked by comparing PINNs predictions with analytical solutions. Results show that a fair accuracy for depth and velocity can be obtained, even for some challenging test cases such as the dam-break over a bottom step and the planar flow over a parabolic basin (Thacker’s test case). Moreover, it is shown that, if PINNs are applied to a case with horizontal bottom, for which topography is not strictly necessary, similar accuracy and computational time are obtained when PINNs solve standard SWEs or augmented SWEs. It can therefore be concluded that the augmentation of SWEs is a simple but promising strategy to deal with flows over complex bathymetries using PINNs, which paves the way for applications to flows over more realistic topographies.

Among the non-structural strategies for mitigating the huge economic losses and casualties caused by floods, the implementation of early-warning systems based on real-time forecasting of flood maps is one of the most effective. The high computational cost associated with two-dimensional (2D) hydrodynamic models, however, prevents their practical application in this context. To overcome this drawback, “data-driven” models are gaining considerable popularity due to their high computational efficiency for predictions. In this work, we introduce a novel surrogate model based on the Transformer architecture, named FloodSformer (FS), that efficiently predicts the temporal evolution of inundation maps, with the aim of providing real-time flood forecasts. The FS model combines an encoder-decoder (2D Convolutional Neural Network) with a Transformer block that handles temporal information. This architecture extracts the spatiotemporal information from a sequence of consecutive water depth maps and predicts the water depth map at one subsequent instant. An autoregressive procedure, based on the trained surrogate model, is employed to forecast tens of future maps.As a case study, we investigated the hypothetical inundation due to the collapse of the flood-control dam on the Parma River (Italy). Due to the absence of real inundation maps, the training/testing dataset for the FS model was generated from numerical simulations performed through a 2D shallow‐water code (PARFLOOD). Results show that the FS model is able to recursively forecast the next 90 water depth maps (corresponding to 3 hours for this case study, in which maps are sampled at 2-minute intervals) in less than 1 minute. This is achieved while maintaining an accuracy deemed entirely acceptable for real-time applications: the average Root Mean Square Error (RMSE) is about 10 cm, and the differences between ground-truth and predicted maps are generally lower than 25 cm in the floodable area for the first 60 predicted frames. In conclusion, the short computational time and the good accuracy ensured by the autoregressive procedure make the FS model suitable for early-warning systems.