Elsevier

Applied Ergonomics

Volume 95, September 2021, 103433
Applied Ergonomics

Loud and clear? Train horn practice at railway level crossings in Australia

https://doi.org/10.1016/j.apergo.2021.103433Get rights and content

Highlights

  • Little is known about the safety benefits of train horns.

  • Field observations were conducted at 54 railway level crossings around Australia.

  • Train horn practice is very variable, resulting in inconsistent warnings.

  • A large proportion of train horn blasts are insufficiently loud to warn road users.

  • Further standardisation and studies on road user perceptions are warranted.

Abstract

The road environment has changed markedly over the years. Train horns are ostensibly used to alert road users and are typically mandatory on approach to railway level crossings. However, they have increasingly been seen as a nuisance. Beyond their negative aspects, a study has yet to comprehensively evaluate train horn effectiveness and understand if they remain beneficial and relevant in the contemporary environment. Hence, this study aimed to provide evidence on the actual use of train horns. Field observations were conducted at 54 railway crossings across four Australian States. The effects of level crossing type, location, and environmental conditions were investigated in relation to train horn loudness as objectively measured at the crossing. Results revealed that train horns were not always used, presenting an issue for passive level crossings. However, when sounded, train horn use was redundant and insufficiently loud at level crossings equipped with bells. Taken together, current train horn practice was found to be highly variable and dependent on crossing type, remoteness, and individual train drivers, thus resulting in inconsistent warnings and raising important implications for standardisation.

Introduction

Trains have an extremely limited and delayed ability to react to sudden intrusions on rail tracks, and train drivers can only apply train brakes, reduce speed, and blow the train horn to warn those in danger of a collision. The use of train horns is detailed in the Codes of Practice (e.g. Australian Rail Track Corporation, 2015; Brookfield Rail, 2016), which instructs train drivers on the scenarios requiring train horn use, including the intensity, length, and repetition necessary based on the circumstances and distances over which warnings need to be heard. The train horn is a mandated safety device used at known high-risk locations, such as railway crossings, bridges and tunnels, as well as when a train moves from a stationary position (e.g. on station departure) or when approaching track workers. Horn use is also mandated when train drivers foresee a dangerous situation (Queensland Rail, 2020). High-risk locations are often marked with whistle boards that indicate to train drivers when to use their horn, with guidelines also stipulating that train horns should not be sounded without a valid reason. Similar guidelines are used in the United Kingdom (Hardy and Jones, 2006).

According to Australian standards, train horn decibel readings are set for urban and non-urban locations (Rail Industry Safety and Standards Board, 2016). An acceptable range for a warning device being sounded in urban locations is 96–101 dBA, with the minimum decibel limit in non-urban locations at 106 dBA (measured 30 m away from the locomotive). Trains are often equipped with two horns with different volumes: the main horn, and a low (or city) horn with a lower volume. Prior to 2016, train horn volumes were set in the National Railway Code of Practice at a maximum of 130 dBA for the main horn when measured 1 m from the locomotive, and a minimum of 88 dBA 200 m from the locomotive. The low horn has a set range of 85–90 dBA 100 m from the locomotive (Blutstein, 2015). Similar operating requirements exist for Australian freight locomotives which have been documented by providers in the CRN Engineering Standard – Rolling Stock (John Holland Rail, 2011), as well as in the United States (Meister and Saurenman, 2000).

The use of train horns is regulated to balance safety considerations and the excessive noise in surrounding residential areas. Several jurisdictions have quiet sections/times where train horn use is prohibited except in emergencies. For instance, train horns can be banned in Canada between 10pm and 6am, subject to a safety audit and installation of automatic gates (Blutstein, 2015). The United States implemented train horn bans on track sections rather than based on time of day and, in 2007, the United Kingdom enforced quiet times (11pm – 7am) or zones within which train drivers are required to use the low horn at whistle boards (Blutstein, 2015). In Australia, TasRail (Tasmania) have not implemented a quiet time, but have instructed train drivers in 2014 to only use their lower volume horn during the hours of 10pm and 6am (Blutstein, 2015).

The literature has primarily focused on the negative impacts of train horns. Such research identified two significant effects: residential noise and land value. Traffic noise and other transport-related noise negatively impacts residents' wellbeing (Hardy and Jones, 2006). For instance, noises associated with railway operations have been found to increase sleep fragmentation and cardiovascular arousal and distress (Micheli and Farné, 2016). Additionally. residents surveyed in Brazil documented symptoms of irritability, headaches, poor concentration and insomnia associated with excessive train noise, with train horns considered to be the loudest noise and night-time disturbances from train noise, including horns, found to be most distressing (Zannin and Bunn, 2014). Furthermore, noise mapping models have been developed to quantify the spread of train noise at and around railway crossings in surrounding residential areas (Huang and Warner, 2010). Specifically, Bunn and Zannin (2016) showed that sounding train horns resulted in noise levels of 80–90 dB at residences close to the railway tracks whilst not sounding them reduced noise levels by 10–30 dB. Somewhat unsurprisingly, train horn noise was also found to reduce property value, with Bellinger (2006) noting that train horns decreased property value by 4.1% for each 10 dB increment in a small US town.

Previous research has shown road user behaviour to be a prominent factor in risk of collisions between trains and road users at railway level crossings (Edquist et al., 2009; McCollister and Pflaum, 2007). Therefore, level crossing safety research has extensively examined the effects of warning devices on the behaviours of different road user groups in the field (Larue et al., 2019; Larue and Wullems, 2017; Liu et al., 2015) and in controlled environments such as driving simulators (Larue et al., 2018a, Larue and Naweed, 2018; Lenné et al., 2011; Tey et al., 2011). However, such research does not consider train horns, and limited information exists on train horn uses, road-user audibility of train horns, and their impacts on behaviour.

Recently, Yeh, Multer, and Raslear (2016) used Signal Detection Theory to examine the effects of a range of driver behaviour factors at level crossings. Specifically, this theory considered how drivers undertook decisions based on attitudinal factors and the detectability of the train, including cues leading to the detectability of trains. Train horns were included in the cues leading to the trains' detectability, but their specific effects were not evaluated. However, the study revealed that the recent safety improvements at level crossings are largely due to driver attitudes towards stopping at the crossing and improvements in the cues leading to train detection. This finding suggests that research should specifically investigate the effects of train horns.

Research on the effectiveness of train horns has currently focused on the minimum loudness necessary to ensure detection by road users and reduce the negative effects on nearby communities. Train horns were found to be effective at warning drivers of an approaching train when they can be heard above ambient noise (Landry et al., 2016). Rapoza and Fleming (2002) evaluated the train horn sound level required for a driver with normal hearing to detect this warning with a 95% likelihood when detection must occur to avoid a collision. This research was conducted for an average motor vehicle and an average maximum locomotive speed and suggested that the optimal range for train horn sound levels ranges from 106 dB to 112 dB, depending on the train speed.

The safety benefits of train horns tend to be supported by the negative consequences of train horn bans and hours of limited train horn use. For instance, a review of crashes at railway crossings in the United States found that collisions increased when horn bans were implemented but decreased when rescinded (Rapoza et al., 1999). Train horns were found to lead to a 38–69% reduction in train-vehicle collisions, depending on railway crossing type (active or passive) and location (city or rural) (Rapoza et al., 1999). Coleman and Stewart (1990) examined the effect of train horns on reducing collisions at active crossings in Florida, United States, and found train horn bans to be associated with a tripling of level crossing collision rates at gated crossings, despite the presence of bells, flashing lights and barriers. However, the studies that have evaluated the effects of train horns are limited, dated and predominantly focus on US practices which substantially differ from other countries and settings where train horns are used for longer durations and at louder levels.

Pedestrians and cyclists are more likely to rely on auditory information (i.e. primarily bells, followed by hearing the train) than visual information (flashing lights and behaviour of other road users). On the other hand, drivers are more influenced by visual information such as booms or flashing lights (Beanland et al., 2016). This suggests that train horns are potentially more beneficial to pedestrians’ and cyclists' safety than other road users. The increasing prevalence of pedestrian distractions, particularly auditory ones (e.g. talking on a mobile phone, listening to music) (Goodman, 2018; Larue et al., 2020; Mwakalonge et al., 2015) may diminish the effectiveness of railway crossing signals for road users who are dependent on most of these behaviours. This is particularly concerning given that pedestrians have the highest rates of non-compliance, whether self-reported (Mulvihill et al., 2016) or observed (Larue et al., 2018), and that both pedestrians and cyclists are more likely to access information outside of the warning systems, such as the position of the train (Mulvihill et al., 2016). Such concerns have led the rail industry to trial new signage at level crossings to mitigate distraction issues around level crossings (Hirsch et al., 2017; Larue et al., 2021; Larue and Watling, 2021).

Cushing-Daniels and Murray (2005) examined the trade-off between house values and the safety of crossing users. Their cost-benefit analysis suggests that the costs associated with requiring the use of train horns outweigh the safety benefits (measured as lives saved), raising questions regarding the relevance of train horns as a necessary warning in the current environment. However, such outcomes may result from a poor evaluation of the safety benefits of train horns given the lack of research attempting to quantify such benefits.

Rules and guidelines surrounding the use of train horns are not restrictive and, to a great extent, allow train drivers to exercise their own judgements on how to use their horns. Rail is a regulated environment where collisions result in severe consequences to drivers and occupants of trains and road users, engendering driver focus on safety and compliance. This is reflected by the relative rarity of occurrences of major safety events in Australia, such as Signals Passed At Danger (SPAD) in the rail environment: 260 and 190 SPADs were reported for passenger and freight trains respectively (corresponding to 1.87 and 2.22 SPADs per million train kilometres respectively) in Australia between July 2019 and June 2020 (Office of the National Rail Safety Regulator, 2020). It is also reflected in the emphasis placed on their mitigation (Naweed et al., 2015).

Train driving, in both traditional and modern forms, carries substantial cognitive workload demands, for instance through continuous visual monitoring of the environment (Naweed, 2014). As train drivers develop and rely on mental models to reduce their cognitive workload as their driving skills increase (Moray et al., 2017), expectation bias and habituation in driving can also lead to errors, for instance through strong-but-wrong assumptions about the environment in which they are driving (Moray, 1990).

In the absence of specific evidence surrounding train horns, it is likely that train drivers comply with the use of train horns, especially since the onboard locomotive data recorder captures the use of the horn as a binary record. However, it is unknown how and when drivers use their horn when approaching crossings (i.e., how many times, how long, how loud), as no research till date has investigated how train horns are used in practice in the field or the mental models used by train drivers when deciding whether to use their horn.

Despite continuous safety improvements of rail networks (Fraszczyk et al., 2016; Office of the National Rail Safety Regulator, 2019), the number of vehicle-train collisions at level crossings have become stagnant (Operation Lifesaver, 2020). The road environment's social-technical properties have changed markedly since the last studies on train horns were undertaken (Keller and Rickley, 1993; Multer and Rapoza, 1997; Rapoza et al., 1999). Taking into consideration the current, noisier environments (Bunn and Zannin, 2016), the emergence of better soundproofing of vehicles (Brach, 2009), the distraction of pedestrians with headsets and mobile devices, and the expansion of rail networks and traffic (Love et al., 2017) resulting in bells ringing at level crossings for extended periods (Larue and Naweed, 2018; Zannin and Bunn, 2014), it is evident that a contemporary evidence-based study is required to understand whether train horns remain beneficial and indeed, relevant, for safety at level crossings.

There is currently a lack of research on how train horns are used in practice and whether they are beneficial to road users' safety. This research, therefore, aimed to provide evidence on the actual use of train horns in Australia in a variety of contexts (i.e., level crossing protection, location), and how loud they are where road users are dependent on them. Such research is a stepping-stone towards understanding the relevance of train horns for safety.

Section snippets

Study design

Field-based observations were conducted at Australian level crossings with various characteristics to record the practice of train horn use while taking into consideration (i) the protection at the crossing and (ii) the geographical location. The type of protection at the level crossing was sampled to represent a large variety of possible configurations, including passive crossings (with a stop sign), active crossings with flashing lights, active crossings with flashing lights and boom gates,

Number of train horn sounding per approach

A total of 305 travelling trains were observed (Table 1). Out of these level crossing traversals, 459 train horn blasts were recorded. Train horns were observed to be used during the approach of level crossings between 0 and 5 times. Most trains used it either once (133 times; 44%) or twice (109 times; 36%). A total of 27 traversals were observed without any train horn use (9%).

Statistical analyses showed that train horns were significantly less likely to be used at level crossings with a stop

Train horn practice

The methodology used in this research consisted of field observations covering major Australian States for rail networks and a variety of crossings types and locations. As per guidelines and Standards within an Australian context (Blutstein, 2015; Rail Industry Safety and Standards Board, 2016), the train horn was widely used at level crossings at least once as an alerting signal (i.e. to inform others of the approach of the train). Often, it was sounded multiple times to reinforce the

Conclusion

This observational field study has identified that train horns are not always used when a train approaches a crossing, particularly at passive crossings. Observations have also revealed that train horns are often used multiple times, resulting in train horns being provided to road users too early and at reduced loudness due to attenuation. Importantly, a significant proportion of train horns that are sounded at active level crossings are insufficiently loud enough to warn road users due to the

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The research team would like to acknowledge the assistance of Queensland Rail and Arc Infrastructure and financial support from the Australasian Centre for Rail Innovation (ACRI) project LC17.

References (66)

  • G.S. Larue et al.

    The road user, the pedestrian, and me: investigating the interactions, errors and escalating risks of users of fully protected level crossings

    Saf. Sci.

    (2018)
  • G.S. Larue et al.

    Frustration at congested railway level crossings: how long before extended closures result in risky behaviours?

    Appl. Ergon.

    (2020)
  • G.S. Larue et al.

    Can road user delays at urban railway level crossings be reduced? Evaluation of potential treatments through traffic simulation

    Case Studies on Transport Policy

    (2020)
  • M.G. Lenné et al.

    Driver behaviour at rail level crossings: responses to flashing lights, traffic signals and stop signs in simulated rural driving

    Appl. Ergon.

    (2011)
  • J. Liu et al.

    What are the differences in driver injury outcomes at highway-rail grade crossings? Untangling the role of pre-crash behaviours

    Accid. Anal. Prev.

    (2015)
  • P.E.D. Love et al.

    Off the rails: the cost performance of infrastructure rail projects

    Transport. Res. Pol. Pract.

    (2017)
  • G.J.L. Micheli et al.

    Urban railway traffic noise: looking for the minimum cost for the whole community

    Appl. Acoust.

    (2016)
  • C.M. Mulvihill et al.

    Using the decision ladder to understand road user decision making at actively controlled rail level crossings

    Appl. Ergon.

    (2016)
  • J. Mwakalonge et al.

    Distracted walking: examining the extent to pedestrian safety problems

    J. Traffic Transport. Eng.

    (2015)
  • A. Naweed

    Investigations into the skills of modern and traditional train driving

    Appl. Ergon.

    (2014)
  • L.-S. Tey et al.

    Measuring driver responses at railway level crossings

    Accid. Anal. Prev.

    (2011)
  • Australian Bureau of Statistics

    Australian standard geographical classification - remoteness area

  • ALCAM in detail

  • Code of Practice for the Defined Interstate Rail Network

    (2015)
  • V. Beanland et al.

    Variability in decision-making and critical cue use by different road users at rail level crossings

    Ergonomics

    (2016)
  • H. Blutstein

    Inappropriate use of train horns in rail corridors within Melbourne

  • R.B.M. Brach

    Insertion Loss: Train and Light-Vehicle Horns and Railroad-Crossing Sound Levels

    (2009)
  • Brookfield Rail

    Network Safeworking Rules and Procedures. Rail Traffic Whistles

    (2016)
  • M. Burgess et al.

    Review of Alternatives to ‘beeper’ Alarms for Construction Equipment

    (2009)
  • J. Coleman et al.

    Florida Whistle Ban

    (1990)
  • M.L. Cummings et al.

    Effects of single versus multiple warnings on driver performance

    Hum. Factors: The Journal of Human Factors and Ergonomic Society

    (2007)
  • J. Edquist et al.

    A Literature Review of Human Factors Safety Issues at Australian Level Crossings

    (2009)
  • L. Goodman

    KiwiRail Distracted Users Survey

    (2018)
  • Cited by (2)

    View full text