How extreme weather is testing Britain’s railways

The UK rail network is increasingly under pressure from extreme weather events, a situation intensified by climate change. From autumn leaf fall to summer heatwaves and winter snowstorms, each season brings unique operational risks. Experts such as Professor Brian Haddock, a former Network Rail specialist, and Professor John Beckford, a cybernetician and information specialist, have explored the practical impacts of these weather extremes and innovative strategies, including artificial intelligence, to mitigate them. Their insights reveal the complexity of rail operations and highlight the urgent need for systemic adaptation.

Seasonal impacts and infrastructure vulnerability

Seasonal variations in the UK create a wide range of operational challenges, with each part of the year stressing different assets and systems. Autumn, often underestimated, presents persistent safety and performance issues. Falling leaves are crushed by passing trains and combine with moisture to form a paste-like Teflon on the railhead. This reduces adhesion, making it harder for trains to accelerate or brake and increasing the risk of signals being passed at dangerous speeds. It can also lead to wrong side track circuit failures, where signalling systems fail to detect a train’s presence.

To manage these risks, the industry deploys railhead treatment trains, applies adhesion modifiers and uses remote monitoring to track adhesion levels. Despite this, the challenge remains seasonal and resource intensive. Haddock explains that mitigation efforts are concentrated into a narrow window: “Between the first of October and the 13th of December, that’s when we operationalise our mitigation.” Outside this period, the railway relies largely on reactive measures, leaving it vulnerable when weather patterns shift.

Winter brings a different set of hazards. Ice forming on conductor rails can interrupt the electrical supply, causing trains to lose power. Snow and ice can accumulate at points, preventing them from moving and restricting the ability to route trains safely. Rolling stock is affected too, with frozen doors, couplers and braking components causing delays and cancellations. While some modern trains have ice-detection electronics and heating systems, Haddock notes that these protections are not foolproof: “Trains can just cut themselves out to protect their electronic systems… there’s a lot of vulnerabilities there.” Extreme cold can also reduce staff access to remote locations, complicating recovery during severe weather.

Summer conditions introduce risks that are less visible but equally disruptive. Prolonged heat causes rails to expand, increasing the likelihood of buckling if temperatures exceed critical thresholds. Speed restrictions are often imposed, reducing network capacity. Overhead line equipment is also vulnerable, with high temperatures affecting wire tension and pantograph contact. Extended dry spells can destabilise embankments and alter track geometry. Haddock stresses that temperature is not the only factor: “It doesn’t necessarily need to be hot, it just needs to be very dry for a long period of time, [which] can lead to all sorts of problems with track geometry.” These effects may develop slowly, making them harder to detect before they impact operations.

Storms cut across all seasons and represent one of the most unpredictable threats. High winds and heavy rainfall frequently bring trees and vegetation down onto tracks and overhead lines. Managing this risk is complicated by environmental and planning constraints. Haddock highlights the scale of the challenge: “We’ve got millions of trees, and it’s very costly to make sure that we’re compliant with our standards.” Balancing ecological responsibilities with operational safety requires careful judgement, yet increasing storm intensity is making traditional vegetation management less effective.

The age of the UK’s rail infrastructure amplifies these seasonal stresses. Much of the network was built in the Victorian era, long before current weather patterns. Coastal routes, tunnels and earthworks are particularly exposed to erosion, flooding and storm surges. Haddock points to repeated failures of the Dawlish seawall as emblematic of the problem: “Are we just going to keep rebuilding it or are we going to think completely different about it?” Maintaining ageing assets while adapting them to modern climate risks remains one of the most significant operational and strategic challenges facing the railway.

Predictive modelling and AI

Haddock and Beckford emphasise the potential of predictive modelling and Artificial Intelligence (AI) to transform how rail networks respond to extreme weather. Traditional data systems focus on asset management rather than outcomes, limiting their usefulness in operational decision-making.

Beckford describes the Seasonally Agnostic Railway Model (SARM), designed to integrate asset data, weather forecasts and train performance: “If we can make a useful assertion three or four days out about the likely impact of expected weather events, then rail engineers could be deployed to [take] preventative action on assets, such as points at risk of freezing or [blocked/full] drains so that the track doesn’t flood.”

SARM is designed to learn from experience, improving predictive accuracy over time. Retrospective tests showed predictions of train delays within one minute of actual performance, even when accounting for unrelated delays. Beckford explains that such modelling allows proactive maintenance and operational decisions that reduce disruption, suggesting that “success might be measured by the reduction in service disruption due to asset failure.”

However, Haddock emphasises that data quality remains a critical barrier. “We are really poor at collecting weather data’s impact on the railway,” he notes, pointing to issues such as delay attribution, where cascading effects of extreme weather are often misclassified or obscured. Without accurate, integrated data, even sophisticated AI cannot deliver its full potential.

The real value of AI lies in supporting long-term strategic planning. Beckford explains: “If we take this sort of model and we look at 2035, what does the railway look like in 10 years? What does the weather look like in 10 years? What are the assets that will become vulnerable?” By anticipating future vulnerabilities, operators can plan maintenance, infrastructure upgrades and operational strategies. Haddock adds that international examples, such as Dutch and Spanish high-speed networks, provide lessons in dynamic, resilient infrastructure management. He states that we need “some of these systems to… break this cycle,” advocating a holistic approach that combines technology, operational practice and infrastructure planning.

Collaboration and systemic adaptation

Both experts stress that addressing extreme weather risks requires a systemic and collaborative approach. The UK rail network has traditionally managed challenges in isolation, focusing on individual assets or specific seasonal risks. Haddock argues this is insufficient: “We’ve got to stop looking at them individually. We’ve got to look at them as a system and until we do that, no one’s going to benefit, especially the passengers, the freight users.”

Collaboration between government agencies, infrastructure managers, train operators and technological innovators is crucial. Haddock also points to the potential role of start-ups and third-party vendors in improving operational resilience. Investment in AI, predictive modelling and digital tools could enhance decision-making and planning, but only if integrated with infrastructure management and operational strategy.

Climate change is intensifying extreme weather, making proactive adaptation essential. Haddock points to the 2020 Carmont derailment in Aberdeenshire as a stark illustration of the human cost of failing to address systemic vulnerability. The passenger train derailed after striking debris washed onto the track following heavy rainfall, killing three and injuring several others. While extreme weather was initially cited as the cause, investigations revealed rainfall alone did not explain the disaster.

“People blame it on rainfall… [but that] isn’t correct,” Haddock says. “The drainage design wasn’t right, and the amount of debris on the track was minimal compared to some of the stuff we’ve seen. Yet it caused such a horrific accident.” The official inquiry found that deficiencies in drainage, earthworks design and monitoring had allowed material to accumulate on the track during intense rainfall, creating conditions for the derailment. The site had a known history of instability, yet existing systems failed to identify or manage the risk.

For Haddock, Carmont exposed a deeper problem in how the industry understands and manages extreme weather. Rather than being treated as a predictable stress on infrastructure, weather is often categorised as an external, unavoidable cause, masking underlying engineering and organisational shortcomings. “It didn’t even have to be there,” he adds, referring to the debris. “And yet it led to an accident I’ve never seen anything like in my career.”

The crash highlighted how fragmented data, limited real-time monitoring and a lack of system-wide visibility can allow relatively small failures to align with catastrophic consequences. Haddock argues that better integration of asset data, drainage performance and weather impact modelling could have flagged the risk earlier and prompted preventative action. In a climate that is becoming increasingly volatile, Carmont should be seen not as an isolated tragedy, but as a warning of what can happen when ageing infrastructure, extreme weather and insufficient predictive capability intersect.

Conclusion

The UK rail network faces a critical juncture. Seasonal weather, ageing infrastructure and climate change pose significant operational risks, while traditional approaches to data management and asset oversight are inadequate. Experts such as Haddock and Beckford advocate for a holistic, systems-based approach that integrates predictive technologies, proactive maintenance and long-term strategic planning. With climate change intensifying weather extremes, the choices made today around infrastructure resilience, data integration and technology adoption will determine whether the rail system can withstand future challenges or continue to be vulnerable to disruption. As Haddock puts it, “We’ve got to look at them as a system, and until we do that, no one’s going to benefit.”

About the Interviewees 

Professor John Beckford is a cybernetician and management scientist with over 30 years’ experience helping organisations improve effectiveness through systemic, data-driven approaches. He is Visiting Professor at Loughborough University and University College London, President of the Cybernetics Society, and holds multiple board and trustee roles. Author of three books and numerous publications, John specialises in linking theory and practice, using holistic, collaborative methods to diagnose organisational challenges and implement sustainable solutions. He focuses on empowering teams to learn and adapt, combining mentoring, research and practical guidance to achieve measurable results.

Dr Brian Haddock is an experienced rail industry expert, specialising in strategies to improve the UK rail network’s resilience to extreme weather and climate change. He leads the UK Railway Research Innovation Network (UKRRIN) weather and climate workstream, bringing together universities, SMEs, and infrastructure owners to deliver practical, industry-focused solutions. Dr Haddock has mentored PhD studentships at Loughborough University, published influential research for government bodies, and facilitated collaboration between academia and industry through workshops and seminars. His work focuses on risk management, strategic planning, and sustainability, shaping a more resilient and future-ready rail network.