PICO
Emergencies such as storms, wildfires, floods, or earthquakes can affect various interconnected environmental and critical infrastructure systems, potentially causing failures that may cascade from one system to another. For instance, such events can contaminate water sources, disrupt the operation of water treatment, supply, disinfection, and distribution, cause overflows in sewerage and drainage systems, and disrupt services in power grids, telecommunication networks, and transportation networks. Especially in cases of contamination or disruption of disinfection, such events can result in severe risks to public health, the economy, and the environment.
Addressing these challenges requires a unified Cyber-Physical-Socio-Environmental System (CPSES) approach that models the interactions and dependencies among the various components. We propose an Integrated Digital Twin architecture as a holistic framework that incorporates and coordinates different Digital Twins modelling the different Cyber, Physical, Social and Environmental systems, to capture the propagation of contaminants and estimate their impact.
The CPSES framework incorporates real-time sensor data, geographical information systems (GIS), computational models, and state-estimation algorithms to dynamically model events and enable proactive planning and real-time decision support for local authorities, first responders, utility operators, and public health officials.
For example, a sudden storm can increase water levels in a reservoir, causing an overflow that significantly raises the water level in a downstream river. This, in turn, can lead to sewage overflow from a nearby manhole, potentially affecting first responder operations, and flooding a power substation, which disrupts its operation and, in turn, disconnects a pump supplying water to a central tank.
A core technology for implementing this framework are the Data Spaces, which serve as secure, standardized environments for ingesting and sharing data among multiple stakeholders and infrastructure operators. Moreover, State Estimation is critical for producing realistic assessments of the current and near-future states of the system. State Estimation can be extended by combining physics-based models with machine learning, to estimate unobserved system states and continuously update parameter values. As a result, data spaces, integrated with GIS, computational models, state estimation, and machine learning, provide a Digital Twin that serves as a single point of reference. This allows risk analysts to assess vulnerabilities, estimate the spread of events, and model cascading effects on other systems.
This integration, facilitates rapid and precise interventions, such as rerouting water supplies, isolating at-risk sewer lines, or reconfiguring power distribution. The HPC-based urgent-computing paradigm can also be considered to ensure stakeholders receive risk assessments, contamination maps, and infrastructure failure forecasts within the strict timeframes required for crisis response.
To demonstrate real-world applicability, we discuss the Cyprus Digital Twin, an innovative platform where a simulated emergency triggered a contamination/overflow event in the Yermasogia Reservoir. This event threatened the aquifer and the extraction of potable water from boreholes. By integrating contamination propagation models, public health models, flood hazard models, geospatial data, power network fragility curves, and real-time sensor measurements, the Digital Twin and its tools were able to provide comprehensive situational awareness, assess the potential impact of the event, and support the rapid decision-making process.
How to cite: Eliades, D. G., Vrachimis, S., Kyriakou, M., Laoudias, C., Panayiotou, C., and Polycarpou, M.: Interconnected Digital Twins for Water Contamination Management, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13629, https://doi.org/10.5194/egusphere-egu25-13629, 2025.
The management and strategic planning of urban Water Distribution Systems (WDS) face new challenges indicative of a more profound transformative shift across all dimensions of urban life.
External factors and environmental influences on the urban context have expanded significantly. Historically, fluctuations in water demand and urban development were primarily driven by localised system dynamics. Today, however, the reliability of WDS is shaped by (i) deep uncertainties spanning national and continental scales and (ii) considerations for both short- and long-term horizons. These include persistent trends such as demographic and migration shifts, evolving climatic conditions, and transient events with varying levels of predictability, including seasonal droughts, abrupt changes in governmental policies, and economic volatility, particularly in the transitioning energy sector.
Consequently, there has been a shift towards a more holistic perspective of the urban WDS. New approaches encompass the WDS infrastructure along with its operation, management, and integration with the wider energy grid, representing a significant paradigm change from the traditional approaches focusing primarily on the network of pipes.
As the problem's description complexity increases, the development and use of benchmarks become more relevant. These tools are essential for rigorous and reproducible testing of our solutions and to guide an evidence-based decision-making process. Historically, the academic research field of urban WDS design optimisation has been a prime testing ground. Numerous open problems have been introduced in the literature since the late '80s/early '90s, with notable examples being Anytown, Hanoi, and the New York Tunnels. However, the evolving complexity of the problem indicates that historical benchmarks may no longer suffice, while their adapted versions lack a unified framework, with multiple problem formulations scattered across the literature.
In this work, we explore the coupled operation-design optimisation problem for Water Distribution Systems (WDS). Building on a critical review of the literature, we identify the strengths and limitations of existing benchmark formulations, paving the way for a discussion on the key attributes that next-generation benchmarks should embody. Our work aims to establish a comprehensive problem framework for joint operation and staged design (planning) optimisation, ensuring it addresses the complex and evolving challenges faced by WDS globally. Particular emphasis is placed on capturing the dynamic interplay between viable policy interventions and the variability of critical factors, such as water demand, electricity prices, energy mix, and timing.
How to cite: Zanutto, D., Castelletti, A., and Savic, D.: Rethinking the Coupled Operation-Design Optimisation Problem in Water Distribution Systems under Deep Uncertainties, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16168, https://doi.org/10.5194/egusphere-egu25-16168, 2025.
Continuous drinking water supplied by water distribution networks (WDNs) is essential for social well-being and economic development. WDNs are often divided by isolation valves into segments containing various elements, including nodes, pipes, tanks, and pumps. In the event of an element failure (e.g., pipe failure) within a segment, that segment can be isolated to facilitate repair. Effective failure management requires identifying critical valves and segments to minimize the number of affected users in such an event. Traditional hydraulic-based criticality analysis requires an hydraulic model to assess the criticality of isolation valves and segements, which can be time-consuming particularly for complex systems. Placing new valves also often requires data-intensive and time-consuming optimization methods thatare typically impractical for small and medium-sized WDN operators. To address these challenges, this study introduces a graph-based method to assess and improve WDN resilience by evaluating the criticality of valves and segmentsand the placement of new valves. The approach first evaluates the criticality of isolation valves based on their impact on network performance. Then, it reduces segment criticality by strategically adding new valves to minimize unmet water demand during isolation events. To achieve this, the mathematical graph of a WDN is constructed based on GIS data where valves are considered as nodes and segments as weighted edges. Subsequently, graph-based segment failure magnitudes are calculated, and eigenvector centrality is used to rank valves based on their influence, considering the importance of connected valves. Then critical segments are identified, and the Louvain-based community detection technique is used to determine the optimal placement of additional isolation valves. The method iteratively reassesses critical values to progressively reduce the criticality of both: segments and valves. The method was applied to a benchmark case study and a real WDN in an Alpine municipality in Austria. Results show a strong correlation (>0.9 Spearman) with hydraulic-based approaches. The developed approach effectively identified the most critical segments and valve, reducing the segment criticality by at least 40% and the number of critical valves to one fourth. These findings highlight the efficiency of community detection in valve placement and its potential to reduce both segment and valve criticality. This graph-based methodology is particularly beneficial for small-to-medium-scale WDNs lacking resources for hydraulic models.
Funding: The project “RESIST” is funded by the Austrian security research programme KIRAS of the Federal Ministry of Finance (BMF).
How to cite: Sitzenfrei, R., Satish, R., Rajabi, M., Hajibabaei, M., and Oberascher, M.: Graph-Based Methodology for Segment Criticality Assessment and Optimal Valve Placements in Water Networks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1254, https://doi.org/10.5194/egusphere-egu25-1254, 2025.
In this era of growing population, urbanization and relevant environmental issues, an adequate management of water distribution networks (WDN) is needed to deal with current and future challenges and ensure an efficient water supply. In this context, techniques relying on the use of WDN hydraulic models (i.e., model-based approaches) can be a useful tool to support water utilities in decision-making processes, spacing from leak detection to the definition of maintenance actions. However, model-based approaches generally rely on the availability of a well calibrated hydraulic model of the WDN, which depends on detailed information on WDN features (e.g. topology, pipe characteristics, etc), that may not be available or perfectly known to water utilities.
In the last decades, the process of WDNs digitalization resulted in the availability of a large amount of data (e.g. discharge in pipelines, pressure at nodes, etc.). Among these data, pressure measurements may be particularly easy to obtain, due to the lower costs and limited intrusiveness of pressure sensors compared to flow meters, with the possibility of installing them in a significant number of WDN sections. In light of the above, this study proposes a new method for the identification of anomalous events occurring in a WDN exclusively based on the use of pressure data collected through a series of pressure sensors installed in the network. Even without requiring detailed information on WDN characteristics or the use of the hydraulic model, the method allows detecting both hydraulic anomalies (e.g. anomalous consumption) or mechanical anomalies (e.g. significant leakage events, or unknown gate valves status after maintenance actions) which can significantly impact system functioning and whose prompt identification can improve the quality of the service provided by water utilities.
To effectively detect anomalous events while excluding other potential causes of pressure variations (e.g. changes in the WDN inlet pressure due to modifications in the controls of pumping systems), the method is based on the analysis of pressure differences calculated for all possible couples of sensors located in the WDN, which are expected to deviate from the ordinary range of values only in the case of anomalies because of the local alteration of the pressure-head distribution produced.
The proposed method is tested on a real case study in Northern Italy, featuring around 300 users and provided with a system of pressure sensors collecting data with hourly temporal resolution. The application of the method to the above case study revealed its effectiveness in detecting a series of anomalous events with different magnitude throughout the day. In addition, the method was demonstrated to be capable of identifying anomalies occurring simultaneously in different areas of the WDN. Overall, it is believed that the developed method can provide a solid indication based on which the water utility can promptly act and verify the possible presence of anomalous events of different nature.
How to cite: Alvisi, S., Mazzoni, F., and Marsili, V.: A pressure-based approach for the identification of anomalous events in water distribution networks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4368, https://doi.org/10.5194/egusphere-egu25-4368, 2025.
Water distribution networks are increasingly equipped with measurement devices for real-time monitoring of hydraulic parameters, including permanent pressure sensors distributed in the network. Information from these devices can be utilised, among others, for pressure-based leakage localisation, which aims to identify a specific area of the network where a leak might have occurred, and assist refined on site leakage pinpointing. The effectiveness of leak localisation methods is thereby influenced by several factors and their associated uncertainties. For example, errors in available network data affect first the optimal sensor placement, second the hydraulic model calibration, and finally the accuracy of spatial localisation of a leakage. Yet, a systematic analysis including a quantification of the propagation of errors through the sub-processes included in pressure-based leakage localisation is still missing in literature.
The aim of this work is to combine different types of uncertainties in pressure-based leakage localisation to systematically investigate the effects of error propagation through the sub-processes. The following sub-processes are implemented in the error propagation analysis (the considered uncertainties of each subprocess are added in brackets): (1) creation of a hydraulic model or network graph based on GIS data (network topology, pipe diameters, pipe roughness, nodal demand), (2) selection of sensor placements (number of sensors), (3) model calibration during a leakage-free period (measurement errors), and (4) leakage localisation (measurement errors). In this work, both data-driven (i.e., graph-based state interpolation, differential pressure analysis) and model-based (i.e., sensitivity matrix, graph-based genetic algorithm) leak localisation methods are implemented for comparison. Both the L-Town benchmark network from the “Battle of the Leakage Detection and Isolation Methods” and a real-world WDN with engineered leakage events are utilised as demonstrative case studies. The leak localisation performance is evaluated by the pipeline distance between the assumed leakage location and the real leakage location.
The preliminary results show that model-based methods are substantially more accurate than data-driven methods under perfect conditions, i.e., perfectly calibrated hydraulic model and no measurement errors. However, model-based methods are also more affected by errors in the GIS data, as an accurate hydraulic model has a major influence on the accuracy. In the next steps, other uncertainties will be systematically added across all defined sub-processes in bandwidths defined by literature values, and their joint influence on the effectiveness of pressure-based leakage localisation will be analysed. These findings can then be used to optimise the quality of data collection strategies based on their relative importance, ultimately leading to an improvement in pressure-based leak localisation in science and practice.
FUNDING
This publication was produced as part of the “FOUND” project. This project is funded by the Federal Ministry of Agriculture, Forestry, Regions and Water Management (BML) (Austria) (Project C300198).
How to cite: Oberascher, M., Steins, E., Diao, K., Cominola, A., and Sitzenfrei, R.: Effects of error propagation of uncertainties on pressure-based leakage localisation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5966, https://doi.org/10.5194/egusphere-egu25-5966, 2025.
Water distribution systems (WDS) face increasing challenges from climate change, urbanization, and growing populations, making efficient and accurate modeling crucial for their management. While traditional physics-based hydraulic models provide reliable results, they are computationally intensive, limiting their application in real-time decision-making and large-scale optimization. Our multi-year research journey demonstrates a progressive evolution in WDS modeling, successfully combining data-driven efficiency with fundamental physical principles.
Our investigations evolve from conventional neural networks to increasingly sophisticated physically-informed approaches. We demonstrate how Graph Neural Networks can leverage network topology to improve prediction accuracy, but more importantly, how reformulating the problem in the edge space allows direct embedding of mass conservation principles. This novel Edge-Based Graph Neural Network (EGNN) architecture not only achieves superior performance but also demonstrates remarkable zero-shot generalization capabilities across different network configurations.
Building on these insights, we further reformulate steady-state estimation as a diffusion process on graph edges, incorporating both mass and energy conservation laws. This physics-based reformulation enables direct GPU acceleration without relying on machine learning approximations, achieving near-perfect accuracy on multiple benchmarks while maintaining substantial computational speedups compared to traditional solvers.
We then extend this approach to handle uncertainty by developing a topological Gaussian Process framework, where the covariance structure naturally encodes the physical conservation laws. This probabilistic extension enables rapid uncertainty quantification under variable demands, providing analytical uncertainty bounds without the computational burden of Monte Carlo sampling, while preserving the physical consistency guaranteed by our diffusion-based formulation.
How to cite: Taormina, R., Kerimov, B., and Tscheikner-Gratl, F.: Accelerating Steady-State Analysis in Water Distribution Systems with Physics-informed Deep Learning and Topological Signal Processing, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17576, https://doi.org/10.5194/egusphere-egu25-17576, 2025.
Climate change, demographic development, increasing urbanization and growing tourism pose complex challenges for the water supply. To address these challenges, this study carried out a scenario analysis for a case study in Austria. For this purpose, different scenarios were derived that consider different trends in climate change, population growth, urban development, and tourism development. To estimate the change in peak water demand due to climate change, a random forest regression model was derived. The model uses climate indices such as hot days and mean air temperature as explanatory parameters for varying water demand. The model was trained and tested using historical water demand records. The quality of the modeling approach was evaluated using common methods such as the mean absolute percentage error. To predict the future water demand under climate change, climate projections for Austria from the RCP4.5 and RCP8.5 climate change scenarios were used. Different population trends were considered for the scenarios, ranging from a decline to a scenario with strong growth. In the tourism sector, a range from minimal growth in the number of overnight stays to a significant increase was assumed. For the industry, a range from minimal growth to a significant increase in the future was also considered. Various scenarios were developed that take into account the different developments of the individual factors and their range. By including the range of factors, the uncertainties in their future development are also represented. For each scenario, changes in peak water demand were derived for the period 2031-2060. The scenarios provide a variety of outlooks, ranging from minimal to significant changes in peak water demand. The results show a range of possible water demand developments for each scenario, given the uncertainties in the underlying factors. Several factors have been identified as critical to the development of future peak water demand. These include population growth, climate change, and, in some areas, seasonal tourism. The development of the housing situation is also an important factor, as water consumption differs significantly between residents of single-family homes and those of apartment buildings. Depending on the development of the housing situation, water demand can fluctuate considerably. Depending on the scenario, the future average water demand in one study site may increase by 3% to 15%. Population growth combined with urban development and the effects of climate change have been identified as the key factors for demand increase. By considering a variety of possible developments, the analysis provides a good basis for long-term water supply planning and can be used to raise awareness in a region for sustainable housing development
How to cite: Stelzl, A., Kudaya, F., and Fuchs-Hanusch, D.: Future water demand forecast by integrating climate and socio-economic scenarios, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16969, https://doi.org/10.5194/egusphere-egu25-16969, 2025.
Water demand-side management strategies are increasingly recognized as key for urban water conservation, complementing supply-side operations. Recent studies demonstrate that consumption-based feedback can effectively encourage water conservation behaviors. Smart meters, supported by digital innovations like the Internet of Things (IoT) and advanced data analytics, have become central to enabling personalized feedback and reinforcing behavioral changes. These advancements highlight the need for Non-Intrusive Water Monitoring (NIWM) algorithms capable of estimating individual water end uses from aggregate household consumption recorded by single-point smart meters. Existing research offers heuristic and machine learning algorithms to address two primary tasks in NIWM: disaggregation of concurrent end uses and automatic classification of the resulting water end-use data. While many algorithms have been designed, calibrated, and validated using high-resolution temporal data—often synthetically generated or inaccessible due to closed datasets—reproducibility in a realistic environment remains a challenge. Furthermore, most algorithms are tested in virtual settings, overlooking real-world concerns related to data transmission, end users’ privacy, and the intrusiveness of centralized analyses by the water utility or a third party.
In this study, we present a novel cyber-physical testing facility for edge-based, real-time classification of residential water end uses. This facility replicates typical residential water use scenarios and employs machine learning algorithms for on-site, edge computing. Its physical components are modular and include a water tank, two circulation pumps, and piping and valves to simulate flow rate trajectories of various end-use categories. Water consumption is measured using a digital flow meter, with data processed by PyNIWM, an open-source Python framework for NIWM, operating in near real-time on a local computer. By integrating physics-based simulations of water use with edge computing, our test stand supports (i) benchmarking and reproducibility of NIWM algorithms in realistic conditions, (ii) privacy-compliant end-use classification and analysis, (iii) near real-time reporting of NIWM outcomes to users, and (iv) modularity to test various soft- and hardware setups. This approach bridges the gap between virtual testing and practical implementation, addressing key challenges in modern water management while advancing privacy-conscious, user-oriented solutions for smart water metering.
How to cite: Kunze, F., Bölter, C., Haueisen, S., Tilcher, D., Gross, M.-P., and Cominola, A.: A cyber-physical testing facility for edge-based automatic classification of residential water end uses, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15869, https://doi.org/10.5194/egusphere-egu25-15869, 2025.
Unprecedented global population displacement in recent years has increased the burden of waterborne illnesses in refugee and internally displaced person (IDP) settlements. Unlike contexts where water is piped directly to the home, in urban-scale refugee and IDP settlements, water users manually collect water from public tapstands and transport it to their dwellings where they store and use it over several hours. This creates the potential for recontamination, increasing waterborne illness risk. Humanitarian responders need to optimize water treatment to minimize waterborne illness risk at the household. Quantitative microbial risk assessment (QMRA) has been used to assess health risk from drinking water in a variety of contexts. However, conventional QMRA approaches rely on pathogen enumeration data, which is too slow, expensive, and logistically challenging to respond to rapid fluctuations in water quality (WQ) in humanitarian contexts.
We propose a novel hybrid machine learning (ML)-QMRA approach that links operational WQ data to QMRA using probabilistic ML models for responsive risk assessments. The ML-QMRA model uses a two-stage probabilistic ML approach: first we forecast WQ from tapstand to household via a deep composite quantile regression neural network (DCQRNN) and then we link household WQ to E.coli data using a support vector quantile regression (SVQR) model. This predicted E. coli becomes an input to an QMRA model designed based on WHO QMRA guidelines.
We tested this ML-QMRA modelling approach using operational WQ data from the Kyaka II refugee settlement in Uganda to assess daily probabilities of infection for pathogenic E. coli and rotavirus. The ML-QMRA model forecasted a mean infection risk for pathogenic E. coli ranged of 4.5x10-2 and 0.19x10-4 for rotavirus. The ML-QMRA model also determined that to keep the risk of infection from pathogenic E. coli within 5% of the minimum daily risk of infection, 0.8 mg/L of FRC was needed at the tapstand at a turbidity of 1 NTU. The FRC requirement increased with turbidity, up to 1.25 mg/L at a turbidity of 20 NTU. This water quality was also sufficient to manage rotavirus infection risk.
Our study shows how hybridizing process-based QMRA health risk assessment with probabilistic ML models can enable integration of operational data for more rapid risk assessment than conventional approaches using pathogen data. The ML-QMRA model also enables us to set multi-parameter water quality targets for routine monitoring data that are based on health-risk, not arbitrary guidelines. The ML-QMRA approach has applications in a range of contexts outside of humanitarian contexts in urban water management to make QMRA more responsive to rapid WQ fluctuations.
How to cite: De Santi, M., Ali, S. I., Khan, U. T., Brown, J. E., Heylen, C., String, G., Naliyongo, D., Ogira, V., Lantagne, D., Fesselet, J.-F., and Orbinski, J.: Rapid and responsive water quality risk assessment using a hybrid machine learning integrated quantitative microbial risk assessment model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3709, https://doi.org/10.5194/egusphere-egu25-3709, 2025.
Increasingly frequent and severe droughts challenge urban water supply managers worldwide. Measures to guarantee urban water supply reliability and security, such as infrastructure investments and temporary restriction and conservation, can exacerbate financial risk for water utilities. In response, water utilities often utilize surcharges - temporary increases in water prices - to increase revenues during droughts and mitigate their financial risk. However, in cities where water supply challenges are compounded by high social inequality, these measures may raise the costs of water services to socially disadvantaged communities to an intolerable level. This study explores opportunities and tradeoffs of integrating index-based insurance into water supply portfolios for the Federal District of Brazil (FDB), a region that exemplifies many challenges present in cities of the Global South. We compute and comparatively analyze different water supply management pathways using WaterPaths, a state-of-the-art open-source exploratory modeling software designed to support water supply portfolio management. We test four different policy architectures across 999 realizations of flow, demand, and evaporation for a 5-year horizon. (1) no financial mitigation: policy considering only water supply restrictions and water transfer during drought scenarios; (2) surcharges: policy considering restriction measures, transfers, and surcharges; (3) index-based insurance: policy with restriction measures, transfers, and index-based insurance payment; (4) hybrid policy: policy with restriction measures, transfers, surcharges, and index-based insurance. Results indicate that incorporating index-based insurance into water supply portfolios can minimize the financial risk for water utilities while lowering the financial burden on vulnerable water users. Considering that insurance companies are risk neutral, these results indicate that integrating index-based insurance in current industry practices and water management portfolios can bring financial relief to households without imposing a substantial cost to water utilities.
How to cite: Silva, G., Gold, D., Albuquerque Alves, C. D. M., Kreibich, H., Cominola, A., and Mendiondo, E. M.: Integration of index-based insurance into water supply portfolios to support equitable urban water management in the Global South, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13705, https://doi.org/10.5194/egusphere-egu25-13705, 2025.
The concept of smart water supply grids has emerged as a promising response to the evolving challenges faced by traditional water supply systems. Traditional water supply systems are grappling with sustainability challenges such as increasing water main breaks, diminishing freshwater resources, untraceable non-revenue water usage, and escalating water demand. The condition worsens in arid- or semi-arid regions which also suffer from water scarcity due to harsh climate effects, surface water inaccessibility, water loss, and over-exploitation. Smart water infrastructure consists of online monitoring sensors and smart meters at the physical systems layer, various analytical tools, and visualization platform for optimal decision-making by the management authorities. The tighter integration of these components through IoT devices and advanced technologies including artificial intelligence and machine learning, tailored hybrid models for water networks hydraulics and water quality, is crucial in achieving digital twin for water management. This integration allows improved leakage reduction, pressure management, water quality protection, and overall system resilience. We present a case study of the development of a smart water supply system in the IIT Jodhpur campus situated in a semi-arid region of the state Rajasthan, India. We elucidate all phases of the study, i.e., sensor placement, data collection, mathematical modeling, and validation study analysis, addressing several challenges with real-time data and field deployment of digital twin technology. This study underscores the ongoing and incremental digital transition towards smart water systems, which is expected to yield significant benefits if collaboration among academia, industry, and government is effectively fostered. The present approach offers a robust framework for tackling water scarcity challenges in arid- or semi-arid regions and to create sustainable and resilient urban water infrastructures that are capable of adapting to both current and future challenges.
How to cite: Singh, S., Tewari, P. K., and Agarwal, A.: On the Potential of Smart Water Supply Grid in Semi-Arid Region, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-686, https://doi.org/10.5194/egusphere-egu25-686, 2025.
ABSTRACT
The increasing frequency of extreme rainfall caused by climate change raises flood risks in Korea, provoking demands for new, effective methods for early flood detection. Recently, deep learning-based waterbody segmentation methods using surveillance camera images have become a focal point for effective flood detection. Many studies have developed waterbody segmentation models using benchmark datasets collected from diverse regions that represent their unique geographical and environmental characteristics. While benchmark datasets offer valuable insights, they often lack sufficient data to generalize environmental features on a universal scale. To handle the limitations of the generalization, models should learn regional features of the applied environment to ensure reliable flood detection performance. Although the need for developing region-specific datasets designed for local application has risen, little efforts have been devoted to constructing Korean river datasets due to the lack of resources and time. To address these challenges, this study introduces a novel region-specific waterbody segmentation dataset called KU River Dataset and proposes automated prompting, an advanced model training technique. First, KU River Dataset consists of 280 river and stream images and is specifically designed to reflect the diverse characteristics of the environment in Korea under various light conditions and surrounding landscapes. Second, an automated prompting method adapting a foundation model enhances the model’s performance using limited data. We employed Segment Anything Model 2 (SAM 2), a foundation model for image segmentation tasks. The automated prompts, generated from SAM 2’s image encoder, guide the model to focus on features of the waterbody. As a result, SAM 2 trained with KU River Dataset achieved 5% improvement in Intersection over Union (IoU) score on the test set compared to SAM 2 trained with a benchmark dataset of the same size. These results demonstrate the effectiveness of applying a region-specific dataset and an automated prompting method for improving regional flood detection. To improve the model’s robustness across diverse environmental conditions, including low-light and flood scenarios, we plan to gather more images of night vision and inundated riversides. Through the further development of our dataset, we expect to enhance the precision of early flood detection systems.
ACKNOWLEDGEMENT
This work was supported by Korea Environment Industry & Technology Institute (KEITI) through R&D Program for Innovative Flood Protection Technologies against Climate Crisis, funded by Korea Ministry of Environment (MOE)(RS-2023-00218873).
How to cite: Kim, J. and Jung, D.: KU River Dataset for Waterbody Segmentation in South Korea: Application of Foundation Model with Auto-prompting, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14024, https://doi.org/10.5194/egusphere-egu25-14024, 2025.
Urban Drainage Networks (UDNs) are essential for managing flood risks in cities facing intense rainfall, exacerbated by climate change. Many researchers focus on reconstructing and retrofitting UDNs to enhance resilience to urban flooding. Additionally, there is increasing interest in developing real-time flood assessment systems for early warning. These efforts necessitate advanced modeling to accurately measure key hydraulic variables, such as inflow rates and water depths. Conventional tools like the Storm Water Management Model (SWMM) use hydrodynamic simulations but often require extensive calibration and can be computationally intensive. As an alternative, surrogate models provide faster simulations with reasonable accuracy; however, they rely on large datasets and can be case-specific.
This study addresses these research gaps for modeling UDNs by proposing a physics-informed graph network for hydraulic analysis. Unlike traditional surrogate models that derive hydraulic variable data from observed measurements, this framework is fundamentally based on physical laws, such as the conservation of mass and energy. The methodology consists of two main stages. First, the UDN is converted into a directed weighted graph, where nodes represent manholes and edges represent conduits. The edge weights reflect the physical properties of the conduits, helping the model mimic the network's hydraulic behavior. In the second stage, flow routing in the UDN is done using customized graph theory metrics to route the flow within conduits. The flow path from inlet nodes to the network's outfall is determined based on the weighted shortest path principle. Using these flow paths and the pipe capacities, the inflow for each pipe is calculated using a new index called modified runoff edge betweenness centrality.
The developed methodology is applied to two real branched networks in Alpine cities. The physics-informed graph network model was evaluated under 77 rainfall scenarios with varying durations and return periods. The maximum inflow results in the conduits were compared to those obtained by the hydrodynamic model SWMM. The correlation coefficients (R²) ranged from 0.76 to 1 for the first case study and from 0.91 to 1 for the second, demonstrating strong agreement between the surrogate and SWMM models across diverse rainfall scenarios. This physics-informed graph network model is effective in research with limited data and high computational demands, such as real-time UDN assessments and optimization tasks.
Funding: The project “RESTORE” is funded by the Austrian Science Fund (FWF) P 36737-N.
How to cite: Rajabi, M., Hajibabaei, M., and Sitzenfrei, R.: Urban Drainage Network Modeling Based on Physics-Informed Graph Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1890, https://doi.org/10.5194/egusphere-egu25-1890, 2025.
Urban flooding has become a worldwide catastrophic disaster including in Spain due to
increased urbanization, reduced infiltration capacity, and climate change. Though the annual
average rainfall is low in Badalona, a city located in eastern Catalonia and a part of the Barcelona
metropolitan area, it experiences pluvial and flash floods due to the intense rainfall that occurred
in a short interval of time brought by the Mediterranean climate. The combined sewer system of
the drainage network in Badalona City acts as the conveyor of the urban sewer system,
stormwater runoff, and industrial wastewater system. To minimize the surcharging of the
drainage networks, it is necessary to predict the surface runoff and forecast the floods as
accurately as possible. Drainage models such as MOUSE, Infor Works, and SWMM were
developed for such applications for the city. However, the manual calibration results in a long and
tedious process, primarily based on the educated guess of the modeler which could lead to a
possibility of missing the optimum parameter sets during calibration. This makes an automatic
process preferable. Additionally, the optimization done using multi-objective function strategies
has been shown to provide more reliable results compared to the traditional methods. This
project aimed to develop and compare the single and multi-objective function strategies to
optimize the urban drainage model parameters using genetic algorithms. Upon the comparative
analysis of single and multi-objective optimization strategies, it was demonstrated that the multi
objective optimization provides more robust and versatile model compared to single objective
approach providing a balanced trade-off between the multiple objectives. This aids in providing
a holistic approach for drainage network management of an area providing resiliency and
efficiency through a robust framework for addressing various issues such as flood preventions,
water quality management and model performance.
How to cite: Karki, N. and Medina Iglesias, V. C. D.: COMPARISON BETWEEN SINGLE AND MULTI-OBJECTIVE STRATEGIES FOR URBAN DRAINAGE MODEL OPTIMIZATION USING GENETIC ALGORITHMS: A case study of Badalona Urban drainage network., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7209, https://doi.org/10.5194/egusphere-egu25-7209, 2025.
Urban stormwater networks (USNs) are essential in safeguarding urban areas from pluvial flooding. To better understand USNs’ performance and manage them effectively, these infrastructures often undergo various simulations and analyses. However, a significant challenge arises from the quality and availability of data required for such simulations, particularly network data such as sewer slopes and diameters. Due to inconsistent and incomplete documentation, network data are often missing or of poor quality, especially in USNs constructed decades ago. Thus, reliable methods for retrieving these data are crucial to ensure a solid foundation for hydrodynamic analyses in USNs.
This study proposes a data retrieval model to reconstruct missing sewer diameter and slope information. The model is built on graph theory, leveraging the topological and connectivity patterns of USN components. Unlike other retrieval approaches, it is fully automated, computationally efficient, and does not require detailed information. The model comprises four modules: uniformity, hierarchy, elevation, and hydrodynamic, explained as follows: 1) Uniformity Module: In this module, missing data between sewers with identical diameters connected along a directed path are filled with the same diameter information using the shortest path metric, retrieving part of the missing sewer diameter. 2) Hierarchy Module: This module employs a modified centrality metric called runoff edge betweenness centrality to reproduce transition patterns in USNs, where sewer diameters progressively increase from upstream to downstream and accordingly fill the gaps in sewer diameter data. 3) Elevation Module: Missing sewer slope information (invert elevations) is retrieved in this module by considering available slopes of neighbouring sewers and incorporating minimum slope requirements derived from the retrieved diameters. This allows for the approximation of the so-called underground slope surface. 4) Hydrodynamic Module: After filling the data gaps, a hydrodynamic model of the USN is assembled by converting the graph of the network to a Stormwater Management Model (SWMM). The aim here is to ensure that the reconstructed USN can meet actual operational conditions (e.g., by investigating capacity discrepancies in terms of the flow depth-to-diameter ratio in reconstructed USNs). In case of any flow discrepancies, the reconstruction procedure is repeated for specific sewers with the flow depth-to-diameter ratio exceeding a specified threshold.
The proposed model was validated using two real-world USNs. Data gaps were artificially generated by randomly eliminating sewer diameter and slope information, ranging from 10% to 90%, and repeating each data gap scenario 100 times, resulting in 1,800 incomplete USN scenarios. The model was applied to these incomplete networks, and the retrieved USNs were compared to those with complete datasets in terms of hydrodynamic properties (e.g., flow rates, flooded nodes, and flood volumes) and physical characteristics (e.g., diameters and invert elevations). The results demonstrate that the model provides highly promising outcomes, successfully retrieving missing sewer diameter and slope information even with up to 90% data gaps while preserving the hydrodynamic behaviour of the original networks. This graph-theory-based model can be used as a practical tool for water utilities, offering a reliable method for retrieving missing or unavailable data.
How to cite: Hajibabaei, M., Hesarkazzazi, S., and Sitzenfrei, R.: Retrieval of Missing Data in Urban Stormwater Networks Based on Graph Theory, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6806, https://doi.org/10.5194/egusphere-egu25-6806, 2025.
The digital transformation of the water sector presents a potential for addressing critical challenges posed by climate change and the growing need for sustainable practices. Central to this transformation is data, which underscores the necessity for robust frameworks and tools that enable efficient sharing, interoperability, and smart decision-making. Our work highlights the important aspects of digitalization in the water sector, including the evolvability of data models and APIs, dataset discoverability, scalability in data publishing, and the ability to perform cross-domain queries through semantic interoperability. Tackling these challenges is key to unlocking the potential of digital transformation in fostering a resilient and sustainable water economy.
Despite significant advancements in the digital transformation of domains such as air quality, energy, mobility, and traffic, the water sector—particularly in circular water management—remains underserved. Existing initiatives, such as the UK National Digital Twin program and Digital Flanders' DUET project, demonstrate the utility of dataspaces, ontologies, and dynamic knowledge graphs but have yet to be extended adequately to address the specific needs of the circular water economy. Current digital landscapes in the water domain lack adaptive, future-proof pathways capable of integrating diverse data sources into a cohesive framework. To bridge this gap, a holistic approach is required, transitioning from isolated systems to interconnected and dynamic scenarios.
We propose a framework tailored to the specific challenges of the circular water sector. The framework aims to collect, standardize, and integrate the unstructured data generated by various initiatives, thereby fostering interoperability. Unified models and an open dataspace will be central to this framework, enabling seamless interaction between stakeholders. Achieving this vision necessitates collective efforts from data producers and consumers to establish standardized semantics, terms, and definitions, resulting in a unified representation of the water domain.
By facilitating standardized data collection and integration, the proposed framework will enable the creation of adaptive ontologies and dynamic knowledge graphs, which are crucial for modeling the complexity of the water domain. The Open Dataspace will serve as an ecosystem for data sharing across sectors, such as water and energy, enhancing cross-domain interoperability and expanding the utility of shared data assets. These efforts will ensure compliance with laws and regulations, providing equitable treatment to all stakeholders while supporting decision-making processes at various levels.
The implementation of this framework will pave the way for resilient and sustainable water management systems, promoting an interconnected digital ecosystem that can address the unique challenges of water reuse and sustainability. Our study underscores the transformative potential of digital tools and standards in fostering a sustainable water economy and provides a roadmap for future research and development in the circular water sector.
How to cite: abdelfatah, S., Alferes, J., and Colpaert, P.: Opportunities and challenges of interoperable Data sharing in the field of Circular water, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16919, https://doi.org/10.5194/egusphere-egu25-16919, 2025.
