ABSTRACT. The aim of DTs is to solve many tasks - which have been done empirically in the past, or with a high level of expertise or experience - more intuitively, quickly, and reliably using a digital world (Eidson et al., 2010). The use of numerical simulations to run scenarios for managing large dams and river basins is now a routine. However, in the context of future projection and coupled soil-atmosphere simulations, Nazemi & Wheater (2015) have shown that the current ability of far-field models to incorporate water needs and, above all, to represent them accurately, is not sufficient. Avesani et al (2021) pointed out that far-field hydrological models pose considerable demands in terms of memory allocation and CPU time, particularly when assessment of modelling uncertainty is required. The water management sector, like others, is experiencing a wave of digital innovation. Various initiatives have been taken to create and test more efficient platforms for water resource management, using advanced digital technologies. For example, Zeng et al (2023) demonstrated the potential of integrating the Internet of Things (IoT) with Blockchain technology for smart water resource management and water monitoring of cultivated fields; Stein et al (2023) reported on case studies conducted in Paris and Berlin, exploring the potential of digital solutions to raise public awareness of urban water management issues. Despite the numerous examples of digital twins (DT) in various fields, it is worth noting the scarcity of DT applications in the specific field of water resources management. Few attempts have been made to apply DT technology to the whole of a large dam or river basin or a large lake, for digital management of water resources and decision-making based on abundant and diverse data.

ABSTRACT. Managing a modern wastewater treatment plant (WWTP) is a complex task that requires a holistic approach towards multiple objectives, such as process, energy, resource efficiency and carbon footprint. While the increased use of online sensors has certainly enhanced process monitoring capabilities, it is still a challenge for operators and to transform data into actionable information on WWTP operation. Digital twins integrating online data capabilities and process simulators, thus creating a virtual replica of the WWTP, can effectively address this challenge while supporting the digital transformation of water utilities and helping them retaining knowledge within its workforce. Despite these obvious benefits, a handful of digital twin implementations in WWTP are in operation worldwide, and many useful lessons can be derived from existing digital twins.

Aarhus Water is the water utility responsible for WWTP operations in the city of Aarhus (Denmark) and has been a pioneer in the implementation of digital solutions to support and optimize operations throughout the integrated water cycle. Aarhus Water’s vision is to integrate digital twins in WWTP operation to address a number of goals, such as training of operators and decision-support for optimal facility operation, plant maintenance and process design/retrofitting. In this context, Aarhus Water and DHI have engaged in cooperation to implement a digital twin of Egå WWTP (120,000 PE) by using DHI’s technology TwinPlant. This abstract presents the full-scale digital twin application in Egå WWTP and lessons learnt during the implementation and early operational stages.

Egå’s TwinPlant digital twin integrates acquisition, processing and cleaning for real-time data, WWTP modelling using the process simulator WEST, input and output data handling using cloud applications (Microsoft Azure) and a web-based user interface for visualization of monitoring and simulation data and key performance indicators (e.g., energy efficiency, carbon footprint). TwinPlant also allows for 48-h forward prediction of WWTP performance through influent flow and load forecasting, providing for early warning in case of exceptional weather events. Furthermore, what-if scenario capabilities allow for virtual testing of alternative operational scenarios, such as changes in controller settings, and simulating process and equipment failures and the associated impact on effluent compliance.

With the transition from the implementation to the operational phase, it is more evident that a digital twin is a living tool, shaped by the continuous interaction between utility managers, solution providers and end-users (e.g., plant operators). This interaction is crucial to build trust towards a digital twin and support value creation within the utility.

ABSTRACT. The consideration of stormwater management is often overlooked as a potential lever to enhance urban sustainability. The intensive urbanization of our cities significantly contributes to the increase in runoff water which gives rise to issues such as pollution and flooding. By collecting raindrop where it falls, sustainable stormwater management succeeds in minimizing the spread of pollution while preserving the natural water cycle. Additionally, by accommodating water on the surface and making it fit the landscape, it promotes the return of nature to the urban environment. However, despite its undeniable advantages, numerous barriers continue to hinder the widespread diffusion and adoption of this approach in urban design practices. Several intricate factors contribute to explaining this persistent reality. Firstly, among design and construction professionals there remains a lack of deep understanding of this sustainable management, primarily due to the long-standing habit of considering these waters as a constraint to be discreetly managed underground far from the eyes of city-dwellers. The transition to the vision of urban stormwater as a valuable resource necessitates a paradigm shift as well as new collaborations among stakeholders not accustomed to working together. While sanitation historically fell under the purview of technical services, it's now imperative for landscape architects, architects, and urban planners to also get involved and take ownership of stormwater management. It quickly became evident that the most suitable solution to achieve these objectives would be the utilization of computer-aided design software. This software would enable all stakeholders involved in the urban layout project to collaborate, integrating and highlighting sustainable stormwater management solutions from the project's initial phases throughout its evolution. Our aim was to create a collaborative design platform that transcended the mere technical aspect traditionally associated with the sanitation engineer's role. This platform would encompass various other dimensions: hydrological (taking into account the watershed), landscape-oriented (attach great importance to the layout of the space), social (embracing a multi-purpose approach), and economic (reducing construction costs and expenses linked to stormwater transport and treatment). The project, initiated in 2016, aimed to create a tool that enhances the skills of designers, facilitates collaboration among various stakeholders, improves communication between designers and project owners, and raises awareness about sustainability while involving citizen participation. It also permits the generation of a digital twin of the layout project simulating sustainable stormwater management in urban areas and along road projects for prospective scenarios. This includes visualizing runoff, studying layout project variants and assessing their resilience in the face of extreme events. The platform was evaluated and validated on a redevelopment project for Charles de Gaulle Avenue and its surrounding structures, located in the municipality of Saint-Saulve (Valenciennes Métropole, Département du Nord, France) covering a total area of approximately 2.5 hectares.

ABSTRACT. Under a climate change scenario, and with the environmental pressures and regulations in the European Union, the correct management in a quasi-real-time and the forecast under different scenarios of the sewage water system (SWS) start to play a crucial role; they significantly impact urban floods and water quality treatments. This has recently been the case in Italy, where extensive and long-term measurement campaigns have occurred in different SWS intending to calibrate traditional numerical models. Nevertheless, those models require a lot of geometrical data, struggle with seasonal variations and are computationally expensive. We developed a Data-Driven Digital Twin (DT) using different Neuronal Networks (NN) for a small SWS basin in northern Italy. The basin under study consists of a 140 km long sewage network and a total of 22 Doppler sensors that measure every six minutes the water velocity, water pressure (depth) and water temperature and three rain gauges that measure every minute for a total of 1140 days of register. Due to the conditions in which the sensors work, it is typical to have low-quality measures. For this reason, using label data, we developed, trained and included in the DT a NN capable of detecting any anomaly value, assigning a possible cause to the problem (i.e. dirty sensor), and suggesting a correct potential value; this model shows an accuracy >90%. After this quality control, the data pass into the main DT.

This research evaluates two approaches: a convolutional layer NN and a Garph NN. Both models mimic the configuration of the SWS and use the same data to be trained. The data was divided into 70% for training and 15% for validation. The model was trained to forecast XX minutes of data up to 48 hours. The models can include or not real-time and a forecast of meteorological data. Moreover, the model was evaluated under scenarios of missing data (i.e., temporal or permanent sensor removal). Both models show a general accuracy of>90%; nevertheless, Graph NN offers a significant advantage due to the connection of the measurement points (nodes). Finally, this project shows DT's successful development and application due to collaboration between industry, government and academia. Moreover, several valuable lessons were acquired in point selection, data preparation, model selection, and calibration.

ABSTRACT. The concept of digital twin is originated from aerospace technology which aims to monitoring the physical processes of the machinery. After many years development, not only its concept but also its scope has been strongly extended which brings many opportunities for other industries. In theory, the digital technologies are able to represent the microscopic process of the object, which can improve the user’s understanding of the process and give chance to the user to control the occurrence and development of the object's process. However, for the water industry, especially the flood management in large catchment, how to add the digital technologies in the current working process and system and what kind of added values can it bring become hot tops in the many water committees. On one hand, in large catchments, many physical process in water cycle are still unclear and difficult to find the way to well represent in the computer, on another hand, the core point in the current working process of the flood defense in the large catchment is more focus on the monitoring of macroscopic state of the catchment, one BIM model or even several specific models cannot fix with the requirement of current managers. Therefore, how to balance the macroscopic demands and the microscopic outputs is the core challenges in the digital application in the flood management in large catchment which request an operational solution. With the application in Zhangweihe river basin (37,584km2) in China, a new system framework and modelling strategy are well presented in this paper, which indicates the added values produced by new digital technologies. According to the core difficult in the flood management of Zhangweihe river basin, two kinds of models and a new parallel computation approach has been created to help the user to understanding the flood process and forecast the impacts of different decisions. The system is just built and used in the daily work of current managers in China. The operational strategy presented in this paper can be migrated to other water system of large-scale catchments.

ABSTRACT. Digital twins, virtual replicas of real-world systems, have emerged as powerful tools in various domains, including engineering and environmental sciences [1,2]. In the context of river systems, real-time field observations and numerical data from high performance computing simulations form together a practical tool to design flood protection and mitigation plans. In order to design a good digital twin, a large and reliable database is needed. For this purpose, in Spain, the National Geographical Institute (IGN) provides a detailed DTM 5mx5m of the Spanish territory and the Ebro river authority (CHE) has a lot of gauge points in which water levels and/or discharge is measured every 15 min, as well as bathymetry in cross sections along this river. Besides that, recent advancements focused on efficiency of numerical models combined with computational tools for acceleration, such as Graphical Processing Units (GPU’s) allow the development of this digital twin. In this work, a digital twin for 230 km of the Ebro river, from Zaragoza to Mequinenza dam, in Spain is presented. This river, besides being the largest river in Spain, presents a particularity in the chosen reach, as it combines the presence of a flood-prone area where flood events frequently occur and water inundates crops and villages, including a huge reservoir in the lower part with much slower dynamics and a particular topography that channels the river and leads to a one-dimensional nature. A GPU-accelerated SWE model specifically designed for real-time flood events simulation is used in this work which enables an efficient flood forecasting. The numerical model used is based on a well-balanced explicit first-order upwind finite volume scheme, ensuring stability even in complex topographies and facilitating simulation over dry beds without numerical instabilities and wet-dry fronts. First, the digital twin is validated reproducing flood events that occurred in the past in the Ebro River. Additionally, an analysis is carried out to assess the most appropriate boundary conditions (BC) for modelling the dam, considering both a time-dependent level condition and a weir condition. The development process of the digital twin is detailed, encompassing mesh optimization and validation procedures. Simulation results are compared with field data spanning the entire duration of flood events (up to 20 days in any case), allowing for a comprehensive analysis of performance and time efficiency achieved using different GPU devices and BC configurations. The remarkable agreement between the observed and simulated results, coupled with the computational speed attained, further establishes the model as a promising digital twin and forecasting tool for effective river management and flood event prediction.

ABSTRACT. Hydrological and hydraulic models used for forecasting and providing reliable inputs are essential for effective water management. Throughout their application, models might produce results of unsatisfying accuracy due to many uncertainty sources. Considerable uncertainty stems from model parameters values, which are usually obtained through calibration, i.e., iterative adjustment of the parameter values to achieve the best possible fit between simulated and observed variables. Model calibration yields time-invariant parameter estimates, which can lead to poor-quality simulation outputs that cannot serve as decision support. Specifically, parameter values can be expected to vary due to secondary- or seasonal processes that are not explicitly accounted for in the model, or due to anthropogenic activity (e.g., land-use change). Therefore, models used for operational forecasting should be run with up-to-date parameter values. This necessitates frequent model recalibration, which can be quite impractical due to high time- and computational requirements of the calibration procedure. Therefore, developing fast(er) calibration algorithms could be a viable alternative. This paper explores the potential of control theory-based, tailor-made, data assimilation algorithm intended for continuous update of the parameters of hydrological and hydraulic models. The algorithm enables the parameter values to be regularly updated at each computational time step based on the dynamically assessed goodness-of-fit (GOF) performance indicator. This approach enables one-pass calibration procedure. Using this algorithm instead of traditional, iterative calibration procedure where models’ GOF is assessed at the end of the simulation, can improve efficiency and effectiveness of models’ calibration. The proposed one-pass calibration approach will be tested on two synthetic test cases, one example of a hydrological model and one example of a 1D hydraulic model.

ABSTRACT. An appropriate response of urban service infrastructures during extreme weather events such as heavy rainfalls is essential to ensure cities’ resilience. A malfunctioning of urban drainage can significantly impact services such as transportation systems. Herein, preserving a good response of these infrastructures during such events aims to maintain essential services while ensuring users’ safety. Due to recent dramatic flooding events in metro stations that have occurred around the world, this work highlights the initial activities of a PhD study that aims to improve the existing knowledge regarding the safety of users of this service during these kinds of episodes. The mentioned PhD thesis focuses on experimental and numerical activities regarding underground metro stations and, particularly, Paral·lel station, belonging to the Barcelona Metro. This station, located in a central and busy area of the city is shared by two main lines of the metro network and one of their entrances is a historically flood-prone area that is frequently inundated during extreme torrential floods. During such episodes a clear danger to metro users is posed, as the access stairs become the only evacuation routes and, at the same time, the preferred path for incoming water flow. This research aims to use the software Flow-3D to calibrate and validate a tridimensional numerical model that reproduces in detail, the hydraulic conditions that occur on the mentioned stairs during stormwater intrusion. A full-scale model was constructed at the labs of the Technical University of Catalonia, Spain. It consists of a 16 steps stair 1.0 m wide, with a tread of 0.34 m and a riser of 0.17 m. The system has two horizontal platforms, one at the bottom 2.0 m long, and the other, 4,5 m long, at the top before reaching a water tank that provides different discharges through a pumping system. Calibration data will be collected at the top platform in sections at 1.0 m, and 2.0 m from the beginning of the stair steps. On them, water depths and velocities are estimated to determine discharges based on the percentage of the used pumping power. The model geometry has been designed using a CAD software and exported to Flow-3D to be used as the surface interacting with the fluid. The model includes flow aeration analysis in most of the steps implementing software modules such as Air Entrainment, Bubble and Phase Change, and Density Evaluation. Thus, the numerical model will obtain flow variables in all the stair domain in a context where flow variability and interaction with air are extremely high. Once all these properties are defined in the model, the flow behavior in the staircase can be reproduced in both laboratory and untested scenarios, thereby facilitating the replicability and analysis of results.

ABSTRACT. The sudden stopping of the turbines used in hydropower production leads to the propagation of waves in both upstream and downstream directions. This phenomenon, referred to as triggering or disjunction, results in an elevation in water level in the headrace channel. The complex patterns of free-surface undulations in this channel depend on the triggering event characteristics and the geometry of the channel. Those patterns result from the combination of i) the water level adaptation to the abrupt change in the discharge regime and momentum distribution, ii) the pressure distribution within the wave, causing secondary undulations and iii) the channel configuration, including profile geometry, the presence of secondary channels or hydraulic structures like locks, etc. Classical 1D numerical approach fails to accurately reproduce these complex patterns. Shallow Water Equation, based on the assumption of hydrostatic pressure distribution, might prove inadequate to model secondary waves; non hydrostatic approaches such as Boussinesq equations can deal with this phenomenon [1], but both are unable to describe breaking bores phenomena. Furthermore, 1D modelling is unable to accurately capture wave reflection phenomena, as well as geometrical effects. In this study, we investigate the modelling of different disjunction test cases in a headrace channel, using various numerical and scale models. This paper describes the methodology used to correctly reproduce undular bores, compared with field measurements. A 1D computation with CNR's Crue10 code [2] is carried out as a first step. Then 2D modelling is performed with the 2-D code Basilisk [3], using three numerical solvers/approaches: Saint-Venant’s Shallow Water Equations, Green-Naghdi, and Multilayers. The results were validated against laboratory and field experiments [4]. For a « wave Froude number » Fro<1.09 and a near-uniform channel geometry, the study shows that the Saint-Venant solver can accurately capture the dynamics of the observed wave patterns. In that case, Saint-Venant, Green-Naghdi, and Multilayers solvers provide close results. Intercomparison of the 2D results with in-situ measurements (for two different triggers) revealed a trendy for this Saint-Venant approached to underestimate the maximum water levels, with an average bias of - 6cm.

ABSTRACT. Flood events, the most prevalent natural disaster, affect millions of people annually with over a billion potentially susceptible worldwide. In response, developing and implementing efficient forecasting and mitigation tactics is essential. Central to these strategies are hydraulic simulations, whose model parameters calibration is often impaired due to the absence of extensive data in floodplain areas for two-dimensional (2-D) simulations.

In this research, we focus our attention on the zoning of the roughness zones in a 2-D hydraulic model. The zoning of roughness in the floodplains often relies on delineated areas by land use from land cover dataset as Corine Land Cover. Within a wide hydraulic model, the number of roughness zones can be consequent (up to hundreds). Too many zones can lead to ill-posed problems for calibration, and before any attempt to calibrate these parameters on the floodplains, we reduce the dimension of the problem.

Then, we calibrate these zones through an inverse problem using satellite observations such as backscattering coefficients from Synthetic Aperture Radar (SAR) images. Leveraging the growing accessibility of satellite data, we aim to enrich our predictive capabilities in flood modelling.

ABSTRACT. Wateringues territory extends over 1500km2 between Calais and Belgium frontier along the north sea.This a low area, part of it is under water level. It is protected againt sea intrusion par a group of pumping stations with a total flow capacity of over 100 m3/s. This system ensures evacuation of AA river flood and direct runoff at high tide.

A Recent study conducted by Wateringues Intercity Institution (IIW) showed that climatic change will cause dramatic increase in pumping costs due the combined effects of incease in annuel pumping volume and electricity pricing.

Existing pumping system is operated by IIW. various solutions are under study in order to meet this challenge and to try to control these soaring costs. One of them aims at analyzing the potential flexibility provided by the pumping system in order to optimize pumping operation with the combined contribution of on site measurement data, hydraulic modelling et artificial intelligence.

This paper provides an assessment of potential benefits gained by using this approach. Il describes the architecture of such a system, emphasizes complementarity betwween components, and connexion with existing monitoring system.

ABSTRACT. Inundatio est un projet européen (2020-2022) développé dans le cadre du programme Interreg Sudoe, pour lequel s’associent des partenaires espagnols, français et portugais. Son principal défi est de créer un système capable de détecter et d’anticiper les crues soudaines de manière automatisée par la collecte et l’analyse des informations hydrométéorologiques et des prévisions, de les traduire en impact sur le territoire et de les croiser avec les enjeux. Les résultats sont publiés sur une plateforme Web qui propose également des simulations de scénarios de risque. Cette preuve de concept vise à compléter le maillage du territoire non couvert par les outils de prévention expertisés (par exemple par les SPC en France) et à évaluer la qualité des résultats pouvant être déduis de chaînes de traitement sans intervention humaine. Le projet aborde également d’autres méthodologies innovantes pour la prévention des inondations : utilisation de sources non systémiques (ici végétales) pour la reconstitution d’événements passés), détection automatique des ouvrants, instrumentation spécifique aux ouvrages patrimoniaux, … Les modélisations ont été déployées sur les bassins de la Nive avec exutoire à Ustarritz et du Gave de Pau avec exutoire à Lourdes. Sensiblement plus grand (environ 1000km²) que les bassins ciblés initialement par le projet, ils présentent l’avantage de disposer de données existantes conséquentes. Les performances de 5 algorithme de machine learning ont été évaluée sur la Nive et le Gave de Pau, sur 4 scénarios. Le modèle de machine learing basé sur régression linéaire a été le plus efficace sur tous les scénarios. 5 algorithmes de deep-learning ont été confrontés au meilleur modèle de machine learning sur le meilleur scénario et une variante. La meilleure configuration est obtenue avec l’utilisation d’un réseau de convolution neuronal multi-têtes (MH CNN), nécessitant en outre des ressources informatiques moindre que les algorithmes basés sur LSTM notamment. L’ensemble des configurations de deep learning est plus performant que la meilleure configuration de machine learning. L’analyse CBR (Case-based reasoning), basée sur une formule de distance intégrant les 18 indices de Sobol (3 variables, 6 pas de temps) aboutit à des situations suffisamment discriminantes pour permettre un choix de 10 cartes effectivement centré sur l’événement test à reproduire. Sans surprise, plus l’horizon est lointain, plus la précision est dégradée. Les gains de ressource informatique supposés par l’utilisation méthode de lattice Boltzmann en lieu et place d’un modèle numériques 2D classique basé sur les équations de Saint-Venant ont été en revanche décevants pour ce projet, en raison d’instabilités importantes nécessitant un raffinement du pas de temps de calcul.

ABSTRACT. 1. Introduction Climate change has altered hydrometeorological patterns. Flash flood disaster, as a kind of sudden flood disaster, has higher destructive power in a short time. In China, high frequency and high death toll account for 70% of the flood disaster losses. In order to mitigate and respond to extreme disaster events, it is urgent to select the driving factors affecting disasters, and develop reliable modeling techniques to identify the risk regionalization of flash flood disasters. 2. Objectives This project reveals the risks of flash flood disasters by analyzing the relationship between the disaster environment, rainfall, human activities and other aspects of flash flood disasters in Fujian province, so as to identify potential risk areas and provide solid foundation for early warning decision-making. 3. Methods According to the characteristics of rainfall, underlying surface and human activities in Fujian Province, a total of 54 alternative indicators were obtained such as rainfall characteristics, underlying surface characteristics. Principal component analysis was used to reduce the dimensionality of 54 alternative indicators and 10 main indicators were obtained. It includes the elevation of villages, distance between villages and rivers, short duration heavy rainfall in small watershed, topography, confluence time, flood peak modulus, wading projects, housing types, current flood control capacity, monitoring and warning facilities. According to the risk theory, the index system of flash flood disaster risk analysis is constructed by integrating the above indexes, and the index system is constructed and analyzed from three dimensions of risk, exposure and vulnerability. 4. Results The high risk areas are mainly concentrated in coastal areas with high typhoon frequency and inland mountainous areas with high rainstorm value. Meanwhile, through comparative analysis of historical flash flood disasters in Fujian Province, 80% of historical flash flood disasters fell in high-risk and medium-risk areas. The occurrence density of flash flood disaster in high-risk areas is 3 times than that in low-risk areas. 5. Conclusion (1) Flash flood risk identification based on small watershed can reflect the response of flash flood disaster to short-term heavy rainfall, the underlying surface and human activities. The risk assessment method based on watersheds has a certain physical mechanism. (2) The evaluation results show that flash flood disaster prone area in Fujian Province is concentrated in the coastal area affected by typhoon rainstorm and the high value area of severe convective rainstorm in inland mountainous area. The frequency of flash flood disaster is affected by both natural environment and human activities. (3) Through the verification of historical flash flood disaster data, the risk identification results are reliable, and risk assessment model can reflect the risks more accurately, which can provide reliable support for flash flood disaster prevention and accurate forecast and early warning.

ABSTRACT. Flash flood disaster is one of the primary types of disaster in China that cause casualties. It is mainly triggered by short-duration heavy rainfall, which rapidly converges and overflows river channels. Flash flood disaster is characterized by rapid rise and fall, high destructive power, and a wide-ranging influence on multiple locations. The amplification of the disaster is often caused by factors such as river channel diversion, sediment deposition, and blockage of bridge openings by floating debris, adding to the high level of uncertainty. To explore the amplification effect of flash flood disaster, the Zhai Gang River in Liannan County, Qingyuan City, Guangdong Province, was taken as an example. A hydrological-hydrodynamic model was used to simulate the impact of bridge blockage on the flooded area. The hydrological model was the China Flash Flood Hydrological Model (CNFF-HMS), and the hydrodynamic model was the Integrated Flood Modeling System (IFMS). The hydrological model was calibrated and validated using data from 9 rainfall events and 2 extreme flood events, with an average Nash-Sutcliffe efficiency coefficient of 0.76 during the calibration period and 0.81 during the validation period, indicating the applicability and computational accuracy of the hydrological model. The hydrological-hydrodynamic coupled model was then validated using data from a torrential rainstorm-induced flash flood event that occurred on June 21, 2022. The simulations were conducted for three scenarios: no bridge blockage, partial bridge blockage, and complete bridge blockage. The results indicated that the flooded area was 1.33 times larger for complete bridge blockage and 1.10 times larger for partial bridge blockage compared to the scenario without blockage. This provides an initial quantification of the amplification effect of the disaster, offering a more scientifically precise basis for early identification of flash flood disaster risk and guidance for personnel evacuation and relocation.

ABSTRACT. We present two mathematical models for weakly dispersive non-linear waves in prismatic channels of trapezoidal cross-sections. The first model, derived from a variational principle, is an extension of the well known Serre equations; the second one is a variation of Winckler and Liu’model (2015), which belongs to the family of Boussinesq-type equations. Both are valid for arbitrary cross-sectional channels, but the very common case of trapezoidal sections is investigated here in view of understanding and modelling Favre waves in such channels (Favre, 1935). The first model allows theoretical derivations in agreement with previous publications (including scale model measurements by Sandover and Taylor, 1962). The second model is well suited to numerical implementation and allows studying the influence of the channel cross-section shape on the propagation properties of the Favre waves (amplitude vs. celerity, for example).

References

Favre, H. 1935. Etude théorique et expérimentale des ondes de translation dans les canaux découverts. [In French.] Paris: Dunod.

Sandover, J., and C. Taylor. 1962. “Cnoidal waves and bores.” La Houille Blanche 48(3):443–465. https://doi.org/10.1051/lhb/1962045.

Winckler, P., and P. K.-F. Liu. 2015. “Long waves in a straight channel with non-uniform cross-section.” J. Fluid Mech. 770(May):156–188. https://doi.org/10.1017/jfm.2015.147.

ABSTRACT. Tidal rivers are complex systems at the interface between continental surfaces and the ocean. Water levels as well as river discharge are highly influenced by tidal dynamics. Therefore, the combination of a flood and tidal constraints on the water levels can be critical for flood protection. Additionally, the propagation of sediment particles or of a pollution can be difficult to apprehend due to the back-and-forth advection combined with dispersion. The River Saigon in Vietnam is a very specific example of tidal river. Indeed, its watershed is very flat and the Saigon net flow is generally negligible compared to the tidal flow (Camenen et al., 2021). The River Saigon flows through the Ho Chi Minh City (HCMC) megalopolis and understanding its hydrodynamics is crucial in terms of flood risk assessment, saline intrusion, pollution and eutrophication. Nevertheless, only very scarce measurements are available to assess the flow of the River Saigon and tidal fluctuations. The River Saigon recently experienced severe events. For example, during the passage of Typhoon Usagi in November 2018, torrential rains (up to 300 mm in a few hours; Rodrigues do Amaral et al., 2023) were observed in the catchment area and HCMC suffered severe flooding. However, water level measurements on the River Saigon only indicate a slight impact of the typhoon on the net discharge, only on the internal part of the river and with a time lag of more than 24 hours. These measurements alone do not allow us to understand the dynamics of the Saigon during this event. Numerical modelling appears essential to identify the main factors explaining the behaviour of the Saigon during Typhoon Usagi. The purpose of this paper is thus to show the potential of 1D numerical modelling to understand the flow dynamics in a complex tidally influenced network. We used the Mage code developed at INRAE Lyon, which has been already validated on other tidal river systems such as the Adour or the Lower-Seine river systems. The calibration of friction coefficients is quite challenging for tidal rivers due to the control of the downstream water levels on the flow dynamics. The River Saigon presents additional issues such as the lack of reliable data on the upstream net discharge and the potential effects of large canals within the HCMC megapolis on the flow dynamics. We present here the calibration of the model of River Saigon as well as an analysis of the flow dynamics during Typhoon Usagi using the model.

ABSTRACT. Dam break failures can lead to the rapid propagation of flood waves across large areas, causing significant urban or industrial flooding. Such events would cause catastrophic risks to downstream populations and result in severe structural damage. Therefore, it is essential to forecast the dominant physical phenomena involved in flood wave propagation over spaces containing macro-roughness, such as buildings, to assess and mitigate these risks. This study first aims to provide datasets from reduced scale physical experiments of transient flow through complex geometries. Specifically, the paper examines the propagation of flood waves over various macro-roughness configurations. The experiments are conducted in a rectangular horizontal open channel, where flow conditions are achieved by rapidly opening a gate holding a volume of water. To assess the impact of obstacles on flow behavior, different obstacle configurations and sizes are investigated and compared. Second, a preliminary comparison with numerical simulations is performed. The experiments provide complete water hydrographs upstream and downstream of the breach. Conductive and acoustic gauges are positioned at five locations to track the wavefront and water depth variation. Additionally, Large-Scale Particle Image Velocimetry (LSPIV) digital image technique is employed to obtain instantaneous free surface velocity profiles and water surface patterns. Furthermore, numerical simulations of the observed flows are performed using the open-source Navier-Stokes solver code_saturne. The free surface evolution is described using the Volume-Of-Fluid (VOF) method. The results are discussed, and the comparison between experimental data and numerical simulation results demonstrates a good agreement.

ABSTRACT. Ripàrian zones are an essential and active part of the ecological environment of a river, but there is no commonly accepted and precise academic definition for it. In the methods of riparian zone boundary delineation covered in academic articles or among the official regulatory documents of various countries, the basis for its delineation is generally the highest floodplain boundary or a fixed buffer zone width. This paper focuses on modelling based on topography, surface and groundwater hydrology, vegetation, and soil data to delineate the boundary of the riparian area and to analyze its response during the course of different flood events. A computing program is developed to simulate the river-riparian system under different flood events and to generate riparian zone maps, taking as starting point the 2.5, 10, 50 and 100-year return period flood extent obtained by means of IBER, a two-dimensional hydrodynamic model. Results show that changes in topography and land use, as well as discharge direction between river and groundwater influence the extent of riparian zones. Additionally, during floods the riparian zone area increases much slowly than that of the flood, it continues to increase until the flood subsides, but this extent increase at last is limited at a certain level. The results of the modelling show that the setting of the calculation mesh size and groundwater horizontal propagation speed affect the time step value and determine the stability and accuracy of model. A small pixel achieves fine simulation results with high computational cost while a large pixel size can be used with a large time step, and thus it is more convenient for long simulation times. The method skips the traditional task of visually interpreting and identifying plant types to delineate riparian zones, thus reducing errors caused by subjective factors. The code, based on the riparian dynamics is presented as a proposal of a new riparian delineation approach, useful for research and management applications.

ABSTRACT. The evaluation of the dispersion of liquid discharges in rivers by modelling relies heavily on the selection of appropriate models and methods combining hydrography, hydraulics and dispersion of discharges. To ensure the best possible solution and identify uncertainties, a comparison of the different available tools is an essential step to consider. This research aims to compare and characterize the optimal application areas of different tools at IRSN, considering their suitability for routine assessment (monitoring, site expertise, etc.) and accidental scenarios (situation assessment, decision support…), as well as their performance and operational constraints, including data requirements and response times. Specifically, three tools employed at IRSN are compared: one based on a 1D dynamic approach (SYMBIOSE, developed at IRSN and co-owned by IRSN and EDF), one based on a 2D analytical approach in steady-state conditions (CASTEAUR2D, developed at IRSN) and one model developed by EDF and originally applied by IRSN for flood risk assessment (TELEMAC 2D). The study focuses on the Loire River, in particular the downstream of the Dampierre-en-Burly nuclear power plant located 50 km upstream from Orleans and which is already used as an application area for the TELEMAC 2D model at IRSN. The study methodology is carried out in the following steps: the TELEMAC 2D model is adapted to a liquid discharge/effluent transfer context ; the tools are prepositioned, based on their representational domains and different groups of reference scenarios (routine and crisis situations and flood, module and low flow conditions) characterized by their constraints of implementation, time and spatial scales, and expected precision; The overlapping domains of the different tools are identified by a comparative analyses based on their applications to these scenarios and their confrontation to empirical datasets including liquid discharges and volume concentrations of tritium downstream of one or more nuclear installations. Finally, based on the obtained results, criteria for prioritizing the different tools are proposed according to these different situations.

ABSTRACT. Accurate modeling of urban flood events is a critical aspect of urban water management, particularly as the frequency and severity of such events continue to increase. Due to their computational efficiency and ability to capture important hydraulic phenomena, shallow water models (SWMs) find extensive application in the simulation of unsteady river flow. However, the accuracy of SWMs in representing flow in street networks during urban flood has not been exhaustively tested. Thus, this study test the capacity of an SWM to reproduce the flow during extreme flood in urban area by comparing numerical results to experimental ones. The latter have been collected from a physical model of an urban area with a flat bottom. The main focus of the study is based on the comparison of the numerical and experimental outflows. The proposed district geometry, aims to represent typical urban geometries. The experimental rig has an horizontal scale of 1/200 and a vertical one of 1/20. The district consists of a 5 m x 5 m transparent Plexiglas plan, with a network of streets with different widths and angles. The streets network has seven north-south streets and seven west-east streets. Each street has a rectangular weir at their downstream end. A volumetric pump is connected to each street and can control the upstream discharge with a precision of 1%. Water surface elevation measurements were conducted using aerial ultrasonic sensors. The 14 street outputs were monitored and each sensor is positioned above the outlet channel, 10 cm upstream to the weir. The outflow rate of each outlet is determined using a weir discharge equation. To simulate unsteady flows in the district geometry, a 2D shallow water model was developed using the k-epsilon turbulence model. An unstructured mesh with 387k cells and 208k nodes was created. The narrow and wide streets have 25 and 50 cells respectively across their width. Hydrographs with a symmetric and triangular shape are injected in the physical and numerical models. The base flow is equal to 14 L/s and the peak flow equal to 42 L/s. Seven different raising times, ranging from 3.5s to 55s, were tested. The SWM's performance were evaluated by comparing three indicators between the physical and numerical models: the outlet peak flow, the outlet peak flow duration, and the outlet hydrograph skewness. The results indicate that the SWM can reasonably reproduce the hydrograph evolution passing through a network of streets. The study provides valuable insights for engineers and researchers working in the field of urban water management and flood risk assessment.

ABSTRACT. Compound flooding refers to the phenomenon where multiple drivers such as rainfall, storm surge and river discharge, happen simultaneously or consecutively within a relatively brief timeframe (Guan et al., 2023). This is the case of hurricanes and events of heavy rainfall where the combination of drivers can increase the hazard and subsequent the impacts in comparison to a single driver. Therein lies the importance of compound flood events from a risk standpoint. However, its modelling is still in process as commercial models tend to be computational expensive or too simple . Recently, novel open-source models such as Super-Fast Inundation of CoastS (SFINCS) (Leijnse et al., 2021) have emerged with the aim of efficiently modeling compound flooding. These tools employ simplified versions of physically based models to achieve faster results. This study presents a framework to model compound flood using SFINCS and asses its ability on the past event Florence, a hurricane that impacted the eastern coast of the United States in 2018. The methodology employed to achieve the objective begins with establishing a systematic approach to prepare a data catalog, that includes the necessary inputs like topography and meteorological data . Then, the SFINCS model is configured to generate flood maps, which are validated with existing observations, literature and models. In conclusion, the study provides insights on the flood map outcomes, analyzes the model's capability to reproduce compound flooding, and presents the limitations identified in the current simulations. It sets the first steps of an automated workflow for storm surge impact assessment.

ABSTRACT. A Spur dike is an elongated artificial structure with one end on the bank of a stream and the other end projecting into the current, and it is the most cost-effective river training structure among other hydraulic structures that can be built at the channel's banks. At the same time, a series of spur dikes are usually more efficient in stabilizing the alluvial shores, whereas single spur dikes alter the local field. Nevertheless, the local scour surrounding the spur dike affects the strength of the foundation of the spur dike. Thus, the analysis of local scour phenomena surrounding hydraulic structures in rivers is crucial to minimize the hazard of foundation collapse. Therefore, experiments have been conducted to study the phenomenon of local scouring around a series of repelling spur dikes under clear water conditions where the angle of inclination of the non-submerged spur dike with the vertical wall was kept 600 during the study in the straight rectangular flume. In these laboratory experiments, flow velocities and bed deformation around the series of repelling spur dikes were measured using an Acoustic Doppler velocimeter and a high-resolution laser displacement meter respectively. Velocity measurements provide information on dominant agents responsible for the local scour. Measured local scour around the series of spur dikes is compared with available empirical and Lacey regime equations to determine their suitability.

ABSTRACT. Bridges are among the most crucial infrastructures of transportation networks. The vast majority of them are placed above rivers and waterways and have thus to resist various hydrological events throughout their lifetimes. Studies conducted on the causes of bridges failures have highlighted the importance of hydraulic processes, and scour in particular has been identified as one the main reasons for bridge failure. The presence of a pier in a stream locally disturbs the flow and as a result, an excess of shear stress is exerted on the sediment bed. The sediments can be excavated and transported by the flow which results in the digging of a scour hole. In the worst case scenario, the scour hole can reach the pier foundations which could ultimately lead to its failure. For these reasons, scour has been the subject of many studies over the past decades, but accurate predictions are still challenging. Indeed, scour involves a variety of coupled physical processes. In the first place, the complex hydrodynamics around the obstacle which needs to be properly solved to accurately compute the shear stress on the sediment bed. Then two kinds of sediment transport can be identified. The suspended load consisting of sediments being put into suspension in the water column. The bedload transport made up of particles rolling, sliding and saltating over the bed with more or less permanent contacts. Morever, when the bed slope becomes steeper than the angle of repose, sediment avalanches occur. All those sediment transport regimes control the evolution of the bed morphology and in turn affects the hydrodynamics. The ongoing development of a 3D sediment transport and morphodynamics model for scour study is presented. This new solver is developed under OpenFOAM and make use of the finite volumes method to solve both the hydrodynamics and the suspended load transport. The bedload transport is solved with the Exner equation using the finite areas method. To track the bed morphology changes, the arbitrary lagrangian eulerian method is used. It consists of having the mesh getting distorted to match the bed morphology. Various validation benchmarks are presented to illustrate the model such as dune migration and the settling of particles in laminar flow. The developed model will be disseminated as open source in a near future.

ABSTRACT. Flood events generating potential risk for scour formation in rivers containing bridge piers and abutments founded on bedrock may become more frequently in the future, especially because of increasing climate change. Any scour of the bedrock generated by flood events may potentially endanger the stability of the bridge and its appurtenant structures. Hence, pertinent assessment of this scour formation through space and time is becoming increasingly relevant to bridge safety. This paper presents an innovative 3D fluid-solid numerical coupling to compute scour of rock. The coupling runs on rocsc@r, a novel cloud-based digital platform. It allows predicting scour of unprotected bedrocks in rivers and channels, and particularly at hydraulic structures such as dams, bridge piers and abutments. The hydraulic modelling is based on LES simulations or RANS-based simulations as typically applied by CFD models, and accounts for air entrainment into the flow. The rock break-up is modelled and visualized within the dedicated rocsc@r digital on-line environment and interface. The fluid-solid coupling is performed sequentially via an API connection between the on-line rocsc@r platform and the FLOW-3D® CFD environment on a local PC. Detailed flow hydraulics are thereby computed by FLOW-3D® and sent to the rocsc@r platform, which computes the corresponding scour potential for a single layer of rock blocks or for a certain time duration during the fracturing process of the rock mass. In a next step, rocsc@r sends back the scoured bottom to FLOW-3D®, which automatically adjusts the flow hydraulics for the next iteration. Iterations continue sequentially until a dynamic scour equilibrium situation is obtained. A dedicated interface allows visualizing the progressive scour formation. Furthermore, the paper presents an application example of this novel 3D fluid-solid coupling to the development of bedrock scour in a river reach containing multiple bridge piers. The main flow parameters, such as the flow velocity, shear stress and stream power at the water-rock interface are presented. The influence of the bridge piers on these parameters is outlined, together with local bedrock scour potential in the vicinity of the piers as well as global scour potential of the riverbed up-and downstream of the piers. Bedrock scour is thereby computed by different rock break-up mechanisms, such as rock joint fracturing or hydraulic plucking of blocks. Finally, the obtained computational results are compared with standard approaches for rock scour at bridge piers.

ABSTRACT. In numerical river simulation, one-dimensional models are commonly used to study water level and discharge for large domains or long time series. These models are less time-consuming than two- and three-dimensional numerical models, and require fewer input parameters and allow ensemble runs. To build a one-dimensional hydraulic model, a pre- and post-processing tool is needed for creating reach geometry, specifying initial and boundary conditions, friction coefficients and other numerical parameters. Such a tool needs to ensure the consistency of the model and provide a user-friendly graphical user interface. In this article, we present Pamhyr2, the fully rebuilt version of the PAMHYR modelling platform. It is developed using Python, PyQt and Matplotlib. Pamhyr2 is free and open-source, multilingual, cross-platform (Linux, Windows) and is generic enough to accept various one-dimensional solvers. Pamhyr2 includes and enhances all the features from the previous version: multiple-reach modelling, geometry definition from cross-section data, initial and boundary conditions, friction coefficients, hydraulic structures, lateral inflow, punctual intake and the results visualization window. In addition, this version includes new features such as pollutants modelling, bed-load and bed evolution. We describe several windows: the creation of sedimentary layers in the river bed, the sediment characteristics for each layer, the sediment, and pollutant boundary conditions and the lateral inputs. These functionalities are illustrated with simple examples. We finally show the visualization windows for bed evolution.

ABSTRACT. Abstract In this study, sediment transport mechanisms and sediment control are explored, which are essential for fields such as geology, civil engineering, and environmental management. The Monte Carlo method is commonly used in sediment transport modeling for parameter calibration and provides a probabilistic assessment of model uncertainty. The integration of AI and machine learning algorithms in the calibration of 2D sediment transport models using the Monte Carlo method is also evaluated. The Iber software package is utilized for hydraulic and sediment transport simulations in the Riba-Roja Reservoir located in the northeastern Iberian Peninsula. Results of Monte Carlo simulations used to calibrate the Riba-Roja sediment transport model are presented. The best-fitted ASV curve, identified as ASV 1, closely approximates the experimental data, especially in the mid to downstream regions, while exhibiting some discrepancies in the upstream areas. The parameter values for ASV 1 indicate that the sediment transport rate is low, and the critical shear stress for erosion is also relatively low. The critical shear stress for deposition is higher, indicating that sediment particles require higher shear stress to be deposited than to be eroded. The settling velocity of sediment particles is slow. These parameter values suggest that the sediment transport process in Riba-Roja Reservoir during the 2008 flood was characterized by a low sediment transport rate and slow settling of sediment particles. Additionally, the evaluation of the performance of a Deep Neural Network (DNN) model in predicting acoustic saturation volume curves is discussed. The Root Mean Squared Error (RMSE) values decrease as the number of test datasets increases, indicating a reliable predictive performance of the DNN network. However, the rate of decrease in RMSE values slows down as the number of test datasets increases, suggesting that the performance evaluation may not benefit from using more test datasets beyond a certain threshold. Boxplots visually represent the distribution of Mean Absolute Error (MAE) for individual DNNs in predicting ASV curves. Overall, the study highlights the importance of accurate calibration of sediment transport models and the potential of AI and machine learning algorithms to improve the accuracy of the model results while reducing computational time. The study also provides insights into the sediment transport processes in Riba-Roja Reservoir, which are essential for effective sediment transport management to ensure the sustainability of the reservoir and reduce the risk of flooding. The results of this study can aid in planning and designing coastal management projects, structures, and protection of coastal resources.

ABSTRACT. The MEPELS system is a modular Open Source software system allowing the Modeling of the Evolution of Sandy Beaches and Coasts. Future developments may allow its use to be extended to pebble and gravel beaches, but this excludes muddy and rocky coastal areas. It is suitable for coastal applications such as marine flooding, coastal protection, and research on the evolution of the coastline. It is developed to characterize at any point of the world coastline the characteristics of the environment around the land-sea boundary, for a period of a few days. It is focused on a limited environment: an area of 1 to 2 km2, essentially marine, the terrestrial part being limited to the aerial beach. The domain is discretized with metric to decametric resolution. As amphibious operations are carried out in calm seas and light breezes, the unknown sought is the morphology of the place for a defined date. The main purpose of MEPELS is to predict as accurately as possible the depth and slope of the underwater beach on that date. Depending on the coastal locations, the duration of data obsolescence varies. For most of the world's coastlines, it only takes 2 weeks for the bathymetry to be obsolete and for currents and wave propagation to change. The MEPELS system is based on an initial Digital Terrain Model (DTM), updated by modeling the morpho-sedimentary dynamics induced by hydrodynamic conditions. The initial DTM is an input data that the system will use to calculate an updated bathymetry. This action is carried out by two approaches called MEPELS short-term (a few hours to a few days) and MEPELS long-term (up to a few years). MEPELS long-term consists in calculating the changes in morphology until the day of the last storm or the last gale, in order to have an updated the most probable DTM for the use of MEPELS short-term. MEPELS short-term is used to adjust the DTM to the most recent conditions. This consists of taking recent bathymetry (surveyed by boat or by satellite image processed out in the previous few days), or updated bathymetry using MEPELS long-term, and imposing the latest most energetic conditions that affected this environment. This modeling of the impact of a strong gale or storm is carried out over a period of a few hours up to a maximum of 72 hours. MEPELS can also be used to forecast currents and swells in the study area. This work can also be carried out using models dedicated to wave modeling such as PREVAG (Shom model).