Situation awareness failures at level crossings

When rail accidents happen 'human error' is often blamed. Paul Salmon argues that in order to understand what's gone wrong, the whole system must be examined.

The tragic events at Kerang, a rural town in Australia, on the 5th June 2007 demonstrate the importance of optimising road users’ situation awareness at rail level crossings. Here, the driver of a loaded truck continued toward a rail level crossing, unaware that a passenger train was also approaching. The resulting collision killed 11 train passengers and injured a further 15 people, including the truck driver.

Following an exhaustive investigation into the incident, the Office of the Chief Investigator concluded that the train and train crew, the truck, the road and rail infrastructure, and the rail level crossing warning devices all played no causal role in the incident. They commented that, “For reasons not determined the truck driver did not respond in an adequate time and manner to the level crossing warning devices”. The truck driver was subsequently prosecuted on the basis that he had failed to keep a proper lookout. He pleaded not guilty to eleven counts of culpable driving causing death and eight counts of negligently causing serious injury and was acquitted by a jury. The rail level crossing involved was subsequently modified to include boom gates, Light Emitting Diode (LED) lights, rumble strips, and active advanced warning signs.

With terms such as ‘human error’ often bandied about in the aftermath of such incidents, it is tempting to think that the road user is the broken component that creates rail level crossing collisions and that the rest of the road and rail ‘system’ works just fine. However, current ergonomics thinking would suggest that there is more to the problem of degraded road situation awareness at rail level crossings. The aim of this article is to demonstrate, through deeper examination of the contributory factors involved in the Kerang tragedy, the role of wider systems failures in rail level crossing collisions.

The individual psychological explanation

At an individual level it is clear that the truck driver was unaware of the approaching train and of the flashing lights at the rail level crossing. There is no suggestion that the driver was impaired through distraction, fatigue, alcohol or drugs, or that he wilfully attempted to beat the train through the crossing. Using Neisser’s perceptual cycle model to explain the truck driver’s behaviour, one plausible explanation is that the truck driver’s failure to see the flashing lights was caused by a schema-driven looked-but-failed-to-see error whereby he scanned the rail level crossing’s warning lights, but did not perceive their flashing state. This failure in turn was likely caused by the activation of the wrong mental schema in the mind of the truck driver, namely a schema for the level crossing in a non-activated state. The truck driver had driven trucks along the same route approximately once a week for seven years prior to the incident but had never previously experienced a train at the crossing. His previous exposure to the crossing in a non-activated state shaped his cognition and expectancies in such a way that he behaved as if a train wasn’t approaching, just as he had done on numerous previous occasions. The warnings provided at the crossing failed to override his wrongly activated schema.

The systems explanation

Examining the road and rail system shows that various factors at the equipment and surroundings level played a role in the truck driver’s failure to become aware of the approaching train. Although the rail level crossing had flashing lights, road-based warnings and an early warning sign, it was not fitted with boom gates which would have provided a more conspicuous visual cue. At the time of the incident considerable sun glare may have impaired the driver’s ability to see the crossing warning signage and controls. The contrast between the train and its background is likely to have been reduced as a result of the truck-facing side of the train being shadowed. Trees in close proximity to the left hand side of crossing may have obscured the truck driver’s vision of the approaching train while the A-pillar of the truck also provided a potential momentary obscurement of a stationary vehicle located on the opposite side of the crossing.

The train driver sounded the horn twice on approach to the crossing, first at the whistle board, and then, continuously for 7 seconds from the point at which the train was 140 metres from the crossing. It is unlikely, however, that the first horn sounding would have alerted the truck driver to the presence of the train. Finally, the road speed limit at the time of the collision was 100km/h. A lower speed limit could have provided more time for evasive action once the truck driver became aware of the approaching train.

The physical processes and actor activities level of the system covers the chain of events that led to the truck colliding with the train. Included at this level is the truck driver’s lack of experience of the rail level crossing with a train approaching. On the day of the incident the driver was delayed in departing the depot due to freight loading issues. This delay ensured that he encountered the crossing at the same time as the train. From here on, activation of the wrong schema meant that the driver failed to notice the approaching train, and the crossing’s flashing light warnings, until it was too late.

The delayed loading of the truck is placed at the technical and operational management level of the system, as is an inspection of the rail level crossing conducted by an infrastructure manager. This inspection was undertaken in response to a series of near miss incidents at the crossing prior to the incident and a letter from the train operator to the track manager expressing concern over road user behaviour at the crossing.

It is apparent that the haulage company involved was not aware of the near miss incidents, a fact that is placed at the local government and company management level of the system.

At the regulatory bodies, state government and industry level various factors combined to ensure that the crossing was not upgraded to include boom gates. On the basis of the near miss incidents and letter to the track manager, various activities were initiated, including public education efforts and the addition of the crossing to the state government’s rail level crossing upgrade prioritisation list. As a result, the crossing was assessed the year before the incident using the Australian Level Crossing Assessment Model (ALCAM). It was assigned a risk score and was ranked 140 out of 143 rail level crossings on the prioritisation list, meaning it was not upgraded at the time of the incident.

Limitations associated with the ALCAM tool are placed at the Government/Parliament policy and budgeting level. Although ALCAM provides a risk assessment score for different rail level crossings, there are various issues that raise questions over its utility. For example, it does not currently take into account accident or near miss data or data on human performance and is heavily weighted towards exposure data, that is, the volume of road and rail traffic passing through the rail level crossing. These limitations means that rail level crossings such as Kerang, which have a low road and rail traffic volume, typically achieve lower risk scores than high traffic rail level crossings. In addition, financial and budgetary constraints limit the number of rail level crossings that can be upgraded to full active controls.

Taken together the individual psychological explanation and the systems explanation demonstrate that road user behaviour at rail level crossings is an emergent property of complex sociotechnical road and rail systems. In the case of Kerang, although the driver’s degraded situation awareness was the key causal factor, there were multiple interacting upstream factors that played a role in this degradation. The tragic incident was a systems accident driven by contributory factors across all levels of the road and rail system. The following are key messages for practitioners attempting to enhance safety at rail level crossings:
Multiple interacting factors shape the behaviour of road and rail systems.

There is a shared responsibility for rail level crossing safety across multiple actors and organisations. A shift away from individual blame and culpability to a learning culture for system improvement is needed. This shift can only occur when the complexities of human behaviour and the impact of the system on behaviour is understood. Modifying rail level crossing infrastructure alone will leave other behaviour-shaping factors free to play a role in future incidents.

By Paul Salmon, Associate Professor in the Faculty of Arts and Business at the University of the Sunshine Coast, Queensland, Australia.

This article was first published in issue 521 of The Ergonomist, November 2013.