The flooding of Gorg station on Barcelona Metro Line 2, which took place on 16 December 2025, was a reminder of a challenge facing many cities: underground transport networks are highly exposed to intense rainfall, and even short-lived events can cause major disruption.
In the Autodesk Water webinar Flood Resilience in Barcelona Metro: From Urban Flood Modelling to Risk-Informed Transport Planning, held on 9 July 2026, researchers from the Universitat Politècnica de Catalunya’s Flumen Research Institute and specialists from Veolia explained how hydraulic modelling can transform flood risk from an abstract possibility into measurable information that supports better decisions.
The panel brought together Eduardo Martínez-Gomariz, Lecturer and Researcher at the Flumen Research Institute, Universitat Politècnica de Catalunya (UPC); Dr Edwar Forero-Ortiz, Researcher at UPC; Beniamino Russo, Professor of Hydraulics and Hydrology at UPC; Àlex de la Cruz Coronas, Project Manager at Veolia; and Paloma B. Akerman, Products Accounts Executive at Autodesk Water.
Akerman opened the discussion by placing Barcelona’s experience in a wider context. Rainfall patterns are changing, and cities are facing increasing pressure to understand and manage climate-related risks. Autodesk Water’s digitalisation approach moves from predictive asset management through InfoAsset Manager, to coupled 1D-2D hydraulic modelling with InfoWorks ICM, and ultimately towards real-time forecasting through InfoWorks ICM Live.
Her key message was that data alone can indicate where a problem may occur, but physical modelling helps explain why it happens. This distinction is what turns modelling into a decision support platform rather than simply a monitoring tool.
Martínez-Gomariz then introduced the work of the Flumen Research Institute, founded in 2012 by UPC and CIMNE, which combines numerical modelling with physical laboratory testing. The institute’s approach brings together experimental research and computational methods to better understand how water behaves in complex urban environments. Using the IPCC risk framework, he explained that flood risk exists only when three elements come together: hazard, exposure and vulnerability.
DataWorks and advanced technologies are always powerful enablers, but the real strength of engineering transformation lies in the teams who turn vision into reality every day - Paloma Akerman
This distinction is important because rainfall alone does not create risk. A hazardous event only becomes a risk when it affects an exposed asset or population that has a certain level of vulnerability. In the case of Barcelona Metro, the infrastructure itself becomes the receptor of risk, meaning that understanding the interaction between rainfall, drainage systems, stations and operational constraints is essential.
For the Barcelona Metro study, this meant developing hydrodynamic models capable of calculating water depths and velocities across the urban environment. These results were then translated into hazard maps, combined with vulnerability information to identify risk hotspots, and used to test possible adaptation measures. The process is iterative: models help identify weaknesses, proposed solutions can be tested, and the results can then guide further improvements.
The objective was not only to understand where flooding could happen, but also to understand its consequences for metro operations. By linking urban flood behaviour with transport infrastructure, the research team could move from a simple description of flood exposure towards a more practical assessment of how services could be disrupted and how resilience measures could be prioritised.
Understanding how floods affect the metro network
Forero-Ortiz presented the findings of the EU RESCCUE project, carried out between 2016 and 2020 with Cetaqua and Transports Metropolitans de Barcelona (TMB).
Historical disruption records showed that flooding problems were concentrated in low-lying areas such as Paral·lel, Espanya and Liceu, where rainfall can exceed drainage capacity. These locations are particularly sensitive because surface water can accumulate quickly and find pathways into underground infrastructure.
We still have a lot to do, but we have already established the basis for a good strategy for increasing the metro resilience in Barcelona - Eduardo Martínez-Gomariz
One critical operational threshold, specific to the Barcelona metro, occurs when water reaches around 15 centimetres above the rail bed. At this point, the track circuits can short-circuit, causing the signalling system to read the section as occupied and bringing the line to a stand. This means that even relatively shallow flooding can have a disproportionate impact on metro operations here, disrupting services before water levels become visibly extreme.
The team modelled all 26 stations on Line 3 under a 20-year storm scenario combined with climate change conditions. The results showed that 15 stations could reach high-hazard levels and 11 could reach high-risk levels. The analysis demonstrated that the impact of flooding is determined not only by rainfall intensity but also by local topography, drainage capacity, and the characteristics of each station.
The model incorporated Barcelona’s drainage system, including more than 2,100 kilometres of pipes and a 660,000-cell surface mesh. Tunnels and access points were represented within the model to understand how water moves from streets into underground infrastructure. By coupling the surface and underground environments, the researchers identified where water enters the network and how conditions evolve once it reaches stations and tunnels.
Roughly 25% of the underground accesses are at high or medium risk of being flooded during extreme precipitation events - Àlex de la Cruz Coronas
One of the most important findings was that relatively small changes at street level can create much larger consequences underground. Between current conditions and the future climate scenario, surface water depth at entrances increased by only around four centimetres, while tunnel water depth increased by 54%. This highlighted a key principle for adaptation: preventing water from entering stations can be more effective than managing flooding after it has already reached underground spaces.
The modelling therefore supports a shift in approach, from reacting to flood events towards anticipating them. Instead of focusing only on emergency response after flooding occurs, cities can use these tools to identify vulnerable locations in advance, evaluate different interventions and improve long-term planning.
Questions on modelling, validation and future applications
During the Q&A session, participants from engineering and consulting organisations across Europe, the Gulf and Asia asked about modelling choices, validation methods and practical applications.
The panel explained that metro infrastructure was represented in InfoWorks ICM through detailed components such as ventilation grilles, stairs and lifts, modelled as inlet structures with their real geometry and connected dynamically to the surface model through a coupled 1D-2D approach.
When asked why Espanya station was particularly vulnerable, Forero-Ortiz explained that the combination of intense rainfall, limited drainage capacity and local topography channels water towards this area.
Russo added that climate uncertainty was addressed by analysing multiple climate model scenarios based on the IPCC’s sixth assessment report. The approach considered a range of possible futures rather than relying on a single prediction.
Validation was carried out at two levels: using historical flow measurements, sensors and rainfall data for the drainage model, and comparing simulated flood depths against evidence from past events, including photographs and other records.
The panel also discussed the difficulty of relying solely on permanent pumping or storage solutions at stations where flooding may be relatively infrequent. Instead, modelling helps identify where targeted interventions and operational measures can provide the greatest benefit.
We considered a wide range of rainfall return periods in our simulations, from very low values (1-2 years) up to very extreme events (500 years in the worst case - Beniamino Russo
The webinar concluded with a discussion on what other cities can learn from Barcelona’s experience. Martínez-Gomariz highlighted that the foundations for a metro resilience strategy are now in place, although further work remains. Russo emphasised that, despite the scale of the challenge, cities now have the knowledge and tools needed to address flood risk, and that the Barcelona approach can be replicated elsewhere.
Forero-Ortiz stressed the importance of combining urban drainage modelling with transport infrastructure planning and operational measures such as early warning systems.
De la Cruz Coronas added that modelling only creates value when operators and authorities are involved throughout the process, from defining the questions to implementing the solutions.
The main lesson from Barcelona is that flood resilience begins before the water arrives. By combining physical modelling, digital twins and operational knowledge, cities can understand their vulnerabilities, test possible solutions and make better decisions for the future.
