The Carmont Train Crash

What it means for Infrastructure Climate Resilience

Six years ago, on the morning of 12 August 2020, a ScotRail passenger train derailed near Carmont, south of Stonehaven in Aberdeenshire. The derailment was caused by the train striking debris washed onto the track following an intense rain storm. The accident claimed three lives, injured six others and became the deadliest railway accident in Britain for almost two decades. Yet the tragedy was more than an isolated engineering failure. The investigation conducted by the independent Rail Accident Investigation Branch (RAIB) concluded that it exposed wider weaknesses in the way Britain's railway understood and managed the risks posed by increasingly severe weather[1]. As climate change alters rainfall patterns and increases the potential for more frequent episodes of intense rainfall, Carmont has become a defining case study in the importance of building climate resilience into ageing infrastructure.

Before considering the technical lessons, it is important to remember the human cost. Train driver Brett McCullough, conductor Donald Dinnie, and passenger Christopher Stuchbury lost their lives in the derailment. Six other passengers were injured, some seriously. The numbers killed and injured would likely have been much higher if not for Covid lockdown restrictions limiting people's travel.  The tragic loss of life remains central to the accident investigation and serves as a solemn reminder that infrastructure resilience is ultimately about protecting lives, not simply safeguarding assets.

What actually happened on the day?

The circumstances leading to the crash illustrate how rapidly extreme weather can overwhelm established systems. Overnight on 11-12 August, Storm Francis brought exceptionally heavy rain across eastern Scotland. The rail network experienced widespread disruption on the morning of the 12th as a result, with many lines south of Perth closed due to flooding and landslips, and, separately, a section of canal bursting its banks closing the main Edinburgh-Glasgow route.

The 06:38 ScotRail service from Aberdeen to Glasgow initially travelled south without incident, and was expected to terminate at Dundee due to the flooding further south. However, it was stopped by a signaller due to a landslip further down the route. This forced the train to reverse direction and return north, planning to terminate at Stonehaven due to subsequent flooding further north.  After a delay to allow the train to safely switch tracks, the train then passed through the Carmont area a second time. By this time the rain had cleared and the sun was shining.  At approximately 9:37 a.m., travelling at around 73 mph, the train struck gravel and other material that had been washed from a drainage system onto the track. The leading power car derailed before colliding with a bridge parapet and embankment, while the coaches jack-knifed behind it with devastating consequences. 

Aftermath of the Carmont Train Crash. Image from RAIB report [1].

What were the root causes?

The RAIB investigation identified several contributing factors but the immediate cause of the derailment was failure of a drainage improvement scheme carried out several years earlier to collect runoff from neighbouring farmland and divert it safely across the railway.  It had not been constructed in line with the design. Modifications during construction formed an earth bund which concentrated runoff towards a gully above the railway. Under ordinary rainfall this presented little difficulty, but during the exceptional downpour, with over 50mm of rainfall in the preceding three hours, the concentrated flow eroded the trench, washing gravel and surrounding material down onto the railway line. This forced the train from the track as it approached at speed.  The intensity of rainfall had an estimated return period[2] of between 100 and 140 years depending on methodology, with the more recent methodology predicting a shorter return period due to climate change.

The investigation also identified important organisational factors which contributed to the accident. Related to the failed drainage system, Network Rail had not adequately managed changes made during construction, nor ensured that the completed asset continued to perform as intended. Operationally, route controllers lacked sufficient information, procedures and training to make fully informed decisions during a fast-moving weather event involving multiple incidents across the network. Existing weather forecasts identified areas at risk but could not pinpoint the highly localised rainfall that arises in summer convective storms and ultimately caused the failure. Consequently, the train continued at normal line speed through an area where the infrastructure had already been compromised.

Graph of rainfall over the three hours preceding the incident. This graph comes from the RAIB report [1] and presents analysis of rainfall radar undertaken by the Met Office following the incident.

What lessons can we take from this tragedy?

The RAIB concluded their investigation with twenty recommendations aimed at strengthening railway safety. They can be grouped into infrastructure, operational and management improvements. For infrastructure, recommendations include:

• Improved management of construction work.
• Stronger design standards for railway drainage systems.
• Better management of drainage assets.
• Improved understanding and management of earthwork and slope risks.
• Improved risk assessment linking weather forecasts with infrastructure vulnerabilities.

Operational and management recommendations included 

• Enhanced operational procedures for responding to extreme rainfall.
• Better use of modern weather forecasting and rainfall monitoring technology.
• Enhanced training and capability for Network Rail route control staff.

In addition to the RAIB investigation, Network Rail set up Earthworks Management and Weather Advisory task forces led by Lord Robert Mair and Dame Julia Slingo respectively.  Focusing on the physical infrastructure side, Mair made many recommendations including the importance of: integrating earthworks, drainage, and vegetation management; using intelligent infrastructure data to improve earthworks asset management; and better monitoring technologies to detect failures and predict possible failures[3].

How did Network Rail respond?

In response, Network Rail has embarked on an extensive programme of work intended to address these recommendations and improve resilience across the network. The events at Carmont highlight that these changes cannot rely upon historical experience alone. Britain's railway comprises thousands of miles of embankments, cuttings, bridges and drainage systems, many dating back more than 150 years, with limited maintenance budgets for many years. Infrastructure that performed adequately under twentieth-century conditions may face significantly greater stress as extreme weather events become more frequent. Climate adaptation therefore requires continuous reassessment of risks rather than assuming that past performance guarantees future safety.

Particular emphasis has been placed on understanding the condition of embankments, cuttings and drainage systems that may be vulnerable during prolonged or intense rainfall. To this end, Network Rail has embarked on one of the most comprehensive programmes of weather resilience improvements undertaken on Britain's railway. A central focus has been the systematic assessment of earthworks (embankments and cuttings) that support over 20,000 miles of railway. The company has expanded the use of detailed geotechnical surveys, airborne LiDAR mapping, and remote condition monitoring to identify assets at greatest risk of failure during prolonged or intense rainfall.  Assets are then prioritised according to the consequences of failure, enabling maintenance and investment to be directed towards the highest-risk locations. Drainage management has also been significantly strengthened. Thousands of culverts, drains and catchpits have been inspected, mapped and, where necessary, repaired or enlarged. 

This represents a shift in approach from the more traditional cycle of inspection, assessment, and repair or renewal to a more mature, data-driven approach to risk identification and management.  This is a transition from verifying that an asset meets design standards to assessing what will happen if the design parameters are exceeded: will it fail, how will it fail, what would be the consequences and are they acceptable?

Operational procedures have also evolved alongside engineering improvements. Network Rail has invested in enhanced weather forecasting capability, combining Met Office forecasts with higher-resolution rainfall modelling and real-time monitoring to provide controllers with a more detailed understanding of local conditions. Dedicated Weather Management Teams now assess the developing risk during severe weather events taking account not only of the weather conditions but giving greater consideration to combinations of forecast rainfall, known earthwork vulnerabilities and drainage conditions. These operational changes mitigate risk while the physical infrastructure remains vulnerable.

Conclusion 

The immediate cause of the derailment was not simply heavy rain but the interaction between extreme weather and infrastructure and operational vulnerabilities.  Although weather forecasts had warned of widespread rainfall, the nature of summer convective storms makes it difficult to forecast the location and map them onto the rail network, and climate projections indicate higher probability of such storms. This is then combined with in-built limitations and vulnerabilities of the infrastructure.

The Carmont investigation became about far more than one incident and one defective drain. It highlighted the growing challenge of managing infrastructure built largely during the Victorian era under climatic conditions that are changing from the basis that many original designs anticipated. 

The investigation emphasised that climate resilience is not achieved through a single engineering project but through a continual process of understanding, monitoring and adapting infrastructure to changing environmental conditions. The response has therefore extended beyond repairing one failed drainage system to embedding climate adaptation into railway asset management, operational planning and investment decisions. As heavy rainfall events are predicted to become more intense, the lessons from Carmont have relevance well beyond the railway. Roads, bridges, power networks and water infrastructure face similar challenges. Identifying vulnerable assets, designing for future climatic conditions rather than historical averages, and using better data to inform operational decisions are practical steps that can improve resilience and ultimately save lives. Carmont remains a tragic reminder of the consequences of underestimating climate risk, but it also provides a constructive example of how organisations can learn, adapt and build safer infrastructure for the future.

The legacy of Carmont is therefore both tragic and constructive. Nothing can undo the loss suffered by the families of Brett McCullough, Donald Dinnie and Christopher Stuchbury. However, their deaths have prompted a fundamental reassessment of how Britain's railway manages the intersection between ageing infrastructure and a changing climate. Carmont illustrates that resilience is not achieved through a single engineering solution but through the combination of better asset management, improved operational judgement, modern weather intelligence and a willingness to learn from failure.

Notes

[1] The full report is publicly available: Rail Accident Investigation Branch (2022). Report 02/2022: Derailment of a passenger train at Carmont, Aberdeenshire, 12 August 2020. GOV.UK. RAIB Carmont Report

[2] The return period is a method for defining the probability that a certain intensity of storm will occur. A 100 year return period signifies that it will occur once in a hundred years, or there is a 1% probability of the storm occurring in any given year. The relationships between intensity, duration and return periods are defined in the Flood Estimation Handbook (FEH). At the time the carmont drainage system was designed, the 1999 version of the FEH applied. The current 2013 version differs from the earlier version, in part because climate change has made heavier rainfall more likely to occur, so a storm of a particular duration and intensity now has a shorter return period than in the 1999 version.
Met Office analysis of the rainfall radar data shows 51.5 mm of rain fell between 05:50 hrs and 09:00 hrs at the Carmont accident site. Based on this amount of rain falling over a 1 km2 area, the return period for this event is 144 years using the 1999 FEH methodology and 100 years using the 2013 FEH methodology.

[3] A Review of Earthworks Management, Prepared by a Task Force led by Professor Lord Robert Mair, CBE FREng FICE FRS: Earthworks Review Final Report