Road-rail interface safety performance report

Transcription

1 Road-rail interface safety performance report This report is issued by: Colin Dennis Head of Risk and Safety Intelligence Rail Safety and Standards Board If you would like to give feedback on any of the material contained in this report, or if you have any suggestions for future editions, please contact: Siona Pitman Safety Intelligence Analyst Rail Safety and Standards Board Evergreen House 160 Euston Road London NW1 2DX The report may be downloaded at no cost from the RSSB website: Rail Safety and Standards Board 2008 Printed January 2008

2 Executive summary 3 1 Introduction Scope and structure 4 2 Overall risk The risk at the interface Overall trends 6 3 Risk at level crossings Risk from level crossing interface 7 4 Level crossing safety performance Headlines Fatalities and injuries Train accidents at level crossings Near misses and misuse at level crossings 24 5 Risk at the road-rail interface The risk from the road-rail interface Post-Great Heck risk assessment 37 6 Road-rail incursions and bridge strikes Headlines Fatalities and injuries Bridge strikes Road-rail incursions by entry point Summary of safety performance risk for road-vehicle incursions 51 7 Initiatives 52 Appendix 1. Level crossing research and development 61 Appendix 2. Key safety facts 65 Appendix 3. Crossing types 66 Appendix 4. The development of public road level crossings 70 Appendix 5. Bridge strike classification 75 Appendix 6. Definitions 76 Appendix 7. Glossary 78 2

3 Executive summary Accidents occurring at the road-rail interface (including level crossings, vehicles and pedestrians, bridges access points and fences) account (on average) for 13.4 fatalities and weighted injuries (FWIs) per year, or approximately 7% of the total system risk. This excludes the risk from trespassing Level crossing risk There were 10 accidental fatalities at level crossings between January and tember Three road vehicle occupants were killed in accidents at level crossings (two in the same incident); 2006 saw no such events. The remaining seven related to pedestrians struck by trains. There were nine collisions between trains and road vehicles over the same period. This continues the improved performance seen in Fewer near misses with road vehicles were recorded in the first nine months of 2007, compared with Jan-t However, more near misses with pedestrians were reported than earlier in the decade. The recent trend for incursions via level crossings where there is no collision has remained fairly sporadic over the last five years. Almost all of these incident types result in the vehicle fouling the line. Bridge strike risk There was a 6% increase in the number of bridges struck between January and tember 2007, compared to the same period in The serious bridge strike count has risen every year, bar one (2005), since However, the first nine months of 2007 saw a 53% drop in the number of these incidents, compared to the same period in Road-rail incursion risk (excluding trespassers) Bridges Incursions via bridges, including those where a vehicle crashes through a fence on the approach to a bridge (resulting in it being struck by a train), seldom occur. Fences There are typically between 40 and 50 incursions by road vehicles through fences each year. Incursions through fences account for almost 70% of all incursions. Access points The numbers of vehicle incursions from access points fell between 2003 and 2006 by 52%. Despite this, the first nine months of 2007 have seen the count rise to three more than at the same stage in The number of vandalism incursions (deliberate access to the infrastructure) has fallen over the period by 58%, although 2007 has seen an increase of 56% over the first nine months. 3

4 1 Introduction This report investigates the full extent of the risk that arises where roads meet railways. Over the last three years, we have focused on level crossings alone; here, we also present data on incursion incidents involving bridges, boundary fences and access points. Trends have been identified, in order to highlight areas of changing safety performance, but the report also outlines the underlying causes that contribute to the current level of risk (along with the initiatives and research projects that are in hand to improve safety further). Fatalities and injuries are discussed throughout this document. Fatalities that occurred in a particular event, or group of events, are first considered separately. The major and minor injuries that arose are then taken into account. This enables injuries to be weighted in accordance with their relatively less serious outcome. The current weighting is 0.1 for each major injury and for each minor injury, the combined measure being deemed fatalities and weighted injuries (FWI). We would appreciate your views on the content of this report, along with any ideas about additional information that you would like to see in future editions. Please send your comments to Siona Pitman, whose contact details may be found on the title page. Or, alternatively, click here to access our new feedback form on the RSSB website. 1.1 Scope and structure This document extends the overview of road-rail interface statistics provided in the 2007 halfyear Safety Performance Report and include quarter 3 data. Specifically, it looks at: Injuries and fatalities to passengers, staff and members of the public (excluding trespassers who gain access via the road-rail interface). Near misses and misuse at level crossings. Trains striking road vehicles at level crossings and following incursions from bridges and through fences. 1 Road vehicles striking bridges. Research projects. Initiatives. This is covered in three main sections: Level crossings, Vehicle incursions at bridges and Vehicle incursions at fences. The report also seeks to analyse the impact of the Network Rail s ongoing campaign on level crossing safety, in addition to reviewing how the All Level Crossing Risk Model (ALCRM) is used in the risk management process. Incidents at level crossings include those involving both pedestrians and road vehicles. Note, however, that trespass at locations other than level crossings is beyond the scope of this report. For an illustrated list of the types of level crossing used in Great Britain, see Appendix 3; for more on the historical development of level crossings, see Appendix 4. 1 For the analysis of level crossings, we consider road vehicles from motorbikes through to lorries and tractors. For vehicle incursions, however, this ranges from cars through to lorries and aircraft. 4

5 2 Overall risk 2.1 The risk at the interface Version 5 of RSSB s Safety Risk Model (SRMv5) estimates the risk at the road-rail interface (including level crossings, bridges and fences, but excluding trespass), to be 13.4 FWIs per year; this is 7% of the total system risk. Chart 1 reveals that most of this risk involves members of the public, predominately pedestrians, being struck by trains. Such accidents are likely to prove fatal, largely because the mass and speed of a train are such that a collision is more likely to kill a person than cause injuries. The second largest grouping involves collisions between trains and road vehicles. Both types of accident are most likely to occur at level crossings. Crossing users can also be harmed if they are hit by, or collide with, barriers or other equipment or if they slip, trip or fall while traversing the crossing. Note that train occupants are exposed to less than 10% of the risk. Chart 1. Overall risk at road rail interface (SRMv5) Total system risk (excludes suicides) FWI 73.1 fatalities weighted injuries Risk associated with road rail interface Road vehicle occupants in collisions with trains 26% Train occupants 9% 7% 13.4 FWI 12.2 fatalities 1.2 weighted injuries Members of the public 55% Slips, trips & falls 4% Accidents with crossing equipment 3% Bridge bashes <1% Passengers on station crossings 3% (Excludes trespassers) Pedestrians struck by trains Chart 2 confirms that most of the risk at the road-rail interface does indeed occur at level crossings, with less than 10% arising from bridge strikes, incursions at fences, bridges and access points. Such incursions often result from road traffic accidents (for example where a vehicle crashes through a boundary fence). Note that road vehicle occupants who are killed in road traffic accidents are excluded, as they are not under the direct control of the railway. 5

6 Chart 2. Overall risk at road rail interface by site type (SRM v5) Total system risk (excludes suicides) FWI 73.1 fatalities weighted injuries Risk associated with road rail interface 7% Fences 4.8% Bridges 2.3% Access points 2.7% 13.4 FWI 12.2 fatalities 1.2 weighted injuries Level crossings 90.2% 2.2 Overall trends A key issue is how a particular area of risk changes (improves or deteriorates) over time. RSSB has developed a precursor incident model (PIM) to assess changes in the risk from train accidents caused by, among other factors, public behaviour at level crossings and this covers the last five years. Although most areas of activity show markedly improving trends, the risk associated with level crossings has remained fairly static (see Chart 3). Much of this report analyses the main drivers behind this risk profile and actions being taken to tackle it. Chart 3. Overall PIM indicator value Risk in de x ( Ma rch = 10 0) 110 Trains & rolling stock Objects on the line 100 Infrastructure failures 90 Irregular working SPADs 80 Public behaviour at level crossings

7 3 Risk at level crossings Level crossings date back to the earliest horse-drawn tramway systems. Indeed, some of Tyneside s eighteenth-century wooden wagonways had crossing points from which a keeper would oversee manoeuvres. Though the interface became safer as more legislation (and technology) was introduced in the nineteenth and twentieth centuries, it still remains one of the most hazardous elements of the railway. While crossings are designed to be safe when operated correctly, abuse, error, incompetence or mechanical failure can result in trains striking pedestrians or road vehicles. When this happens, the crossing user runs a greater chance of being killed or seriously injured. Although crossing users bear the majority of the risk, those on the train can also be affected. This occurred at Ufton Nervet in November 2004, when a train derailed after hitting a car that had been parked on the crossing by a suicidal motorist; not only was the car driver killed, but the train driver and five passengers also died. Fatalities at level crossings make up a very small proportion of the total annual road death toll. The count for 2006 stood at 3,172 fatalities to all road users, a drop of 259 since The last five years has seen an annual average of four fatalities from collisions between trains and road vehicles. However, the majority of fatalities at the road-rail interface take place while pedestrians are crossing the line, with around seven being killed per year (on average). 3.1 Risk from level crossing interface Level crossing risk in context The risk profile for level crossings is based on the quantification of risk resulting from hazardous events occurring on the mainline railway that could lead to fatalities, major injuries or minor injuries to passengers, workforce members or members of the public. This report utilises SRMv5, which shows an estimate of 12.5 FWI per year on average, or 6% of the total risk on the entire railway (excluding suicides). Chart 4 presents the breakdown of the risk to level crossing users. Most of the harm arises from pedestrians and cyclists being struck by trains: such incidents generally prove to be fatal. Chart 4. Fatalities and weighted injuries at level crossings (SRMv5) Total system risk Risk associated with level crossings Road vehicle occupants in collisions with trains 22% Train occupants 6% Slips, trips & falls 5% 6% Accidents with crossing equipment 4% Members of the public 59% Pedestrians struck by trains Passengers on station crossings 4% 7

8 3.1.2 Risk by crossing type Crossings differ in the protection they offer users, their degree of usage, and in the speed and frequency of the trains that pass over them. It is therefore not surprising that there are considerable differences in the levels of risk that each type presents Detailed look at the different crossing types Level crossings have great variation in their use, ranging from sparsely traversed farm crossings to interfaces with urban or trunk public highways. They may be split into two broad groups: Active crossings where the road vehicle user or pedestrian is given a warning of a train s approach in the form of a barrier or warning system (audible or visual). Active crossings are either triggered via an automatic system, such as a treadle operating a barrier, or manually by a signaller, crossing keeper or member of the train crew. Passive crossings where no warning system is provided, the onus being on the road user or pedestrian to determine whether it is safe to cross the line. This includes, where appropriate, using a supplied telephone to call the signaller. At these types of crossing, the responsibility for detecting the presence of an approaching train and deciding whether or not it is safe to cross is left entirely to the user. The latest census (31 December 2006) states that there are 6921 level crossings on Network Rail managed infrastructure (NRMI), all of which are within the scope of this report. Table 1 shows these crossings by type and Network Rail territory. 2 Illustrated descriptions of each type may be found in Appendix 3. Table 1. Level crossing population by region London North East London North West Scotland South East Western TOTAL MCG MCB MCB-CCTV Total active manual AHB ABCL AOCR AOCL UWC-MWL Total active automatic UWC UWC-T OC footpath Total passive TOTAL The table excludes disused crossings on mothballed lines and sleeping dogs (for explanation, see Appendix 6). The category of footpath crossings comprises footpath crossings (86%), bridleway crossings (7%) and station foot and barrow crossings (7%). These are analysed as a single category throughout most of this section because the data in SMIS is not always precise enough to differentiate between them. They have been collectively grouped under passive crossings, but in reality some have automatic protection: 4% (including some at stations) have miniature warning lights and 1% are station crossings with white lights. A further 3% are equipped with telephones and 28% have whistle boards. 8

9 3.1.3 Train accident risk at level crossings Collisions at level crossings account for around one-third of all train accident risk (where the greatest risk is to the public). While it is rare for people on board trains to be killed or seriously injured, 13% of the underlying train accident risk for passengers and workforce is predicted to occur at level crossings. Over the last 50 years, there have been three level crossing collisions resulting in passenger fatalities: 3 Ufton Nervet (2004): Five passengers and the train driver died when a train derailed after hitting a car that had been deliberately parked on the crossing. Lockington (1986): Eight passengers were killed when a diesel multiple unit struck a van which had been driven onto the automatic open crossing against red lights. Hixon (1968): Eight passengers and three train crew were killed when a 12-coach express collided with a heavy road transporter travelling at slow speed across an automatic half-barrier crossing Causal factors of train accident risk Chart 5 breaks down the causes of train accident risk at level crossings. Most of the risk arises from the behaviour of road vehicle drivers, whether deliberate (such as weaving around automatic barriers) or in error. It is very rare (1%) that technical or mechanical failures cause accidents at crossings. Indeed, the underlying causes are basically no different from those that dominate road accidents more generally: violating road traffic controls, misunderstanding signs, complacency and underestimating risk. 5 Managing the risk that results from user behaviour requires close co-operation between the railway, highway agencies, and local planning authorities and the police. As only about 5% of the residual risk is under the railway s direct control, this means that the majority is made up of errors, violations or other actions by members of the public. Chart 5. Causes of train accident risk at level crossings (SRMv5) Road vehicle driver error 68% Road vehicle driver deliberate action (including suicide) 17% SPADs, overspeeding & train driver error <1% Signaller, crossing keeper & track worker errors 3% Equipment failure 1% Environment 2% Vandalism <1% Road accidents & stranded vehicles 7% 3 A further four accidents in this period resulted in the death of train crew members: Naas (1979), Chivers Number 1 (1976), Shalmsford Street (1970) and Stobswood (1959). In addition, an incident at Nairn s user-worked crossing in 1982 resulted in a large number of serious injuries, but no on-board fatalities. 4 Both the transporter crew and the police who were escorting it were unaware of the warning time that would be provided for the onset of a train (24 seconds). Neither had they observed the Emergency Notice or the telephone provided for them to contact the signaller. 5 First Report of the Working Party to the National Level Crossing Safety Group. tember

10 Risk from trains striking road vehicles on level crossings, by crossing type Chart 6 shows the risk from trains striking road vehicles using three metrics: overall risk, risk per crossing, and risk per billion moments (a measure of the number of road vehicles using the crossing per day multiplied by the number of trains passing over it). Overall risk reflects the number of crossings of each type on the network, the hazards associated with them, and the amount of traffic that passes over them. Automatic half-barrier crossings (AHBs) and user-worked crossings with telephones (UWC-T) account for most of the overall risk. There are a large number of user-worked crossings, but most are installed on private roads. AHBs are typically more heavily used and are often sited on busier, highspeed main lines. The barriers on AHB crossings span only half the highway. This was designed in to provide an escape route for stranded motorists, but additional risk arises from road vehicle drivers deliberately zigzagging in an attempt to beat the train. 6 Research into the possible use of median strips or lane separators to dissuade zigzagging is currently in progress. Risk per crossing is a useful metric for safety management purposes, as it can help identify relatively high-risk sites. Automatic open crossings have the highest risk per crossing because they are mostly located on the public highway with warning lights, but no physical barrier to road traffic (although train speeds are lower than for AHBs and the train driver is responsible for ensuring that the crossing is clear; automatic open crossings locally monitored (AOCLs) only). From the crossing user s perspective, the risk per billion moments is the most pertinent measure. User-worked crossings now present the highest level of risk (especially those with miniature warning lights), although fatalities are not common, due to low traffic levels at these interfaces. Crossing users have a responsibility to ensure that it is safe to cross. Users must also alight from their vehicles to open gates or barriers on both sides of the crossing before driving across, presenting opportunities for delay and distraction. Research continues into user-worked crossings and how they are signed and protected. Chart 6. Risk from trains striking road vehicles by crossing type (SRMv5) 10 Fatalities and weighted injuries MCB/MCG AHB ABCL AOCL/R UWC-MWL UWC UWC-T OC MCB/MCG AHB ABCL AOCL/R UWC-MWL UWC UWC-T OC MCB/MCG AHB ABCL AOCL/R UWC-MWL UWC UWC-T OC Overall risk per year Risk per 1000 crossings per year Risk per billion RV-train moments 6 Full barriers provide a more significant obstacle, but require traffic to be stopped for up to six times longer per train traverse than a half-barrier. 10

11 Risk from trains striking pedestrians, by crossing type Chart 7 confirms that footpath crossings account for most of the overall risk to pedestrians, partly because there are more of them than other crossing types. Automatic half-barrier crossings account for the highest level of risk per crossing (approximately 3.7 FWI per thousand crossings per year). Half-barriers tend to be installed away from large conurbations and are often without pavements in rural areas. They are not designed to physically prevent pedestrians from crossing when a train is approaching if they choose to do so (although flashing lights and yodel alarms are provided, in part, to warn them). Chart 7. Risk from trains striking pedestrians by crossing type (SRMv5) Manual AHB ABCL AOCL/R 0.03 UWC-MWL 0.03 UWC 0.18 UWC-T 0.18 FP Fatalities and weighted injuries Manual AHB ABCL AOCL/R UWC-MWL UWC UWC-T FP Overall risk per year Risk per 1000 crossings per year 11

12 4 Level crossing safety performance 4.1 Headlines In the first nine months of 2007: There were 10 accidental fatalities at level crossings. There were nine collisions between trains and road vehicles. This is low by historical standards and continues the improved performance seen in Collisions between trains and road vehicles resulted in fatal injuries to three road vehicle occupants, two in the same incident; 2006 saw no such events. Near misses involving trains with road vehicles at level crossings remain at a low level, comparable with the end of Conversely, near misses with pedestrians are being reported more frequently than in the early years of the decade. 4.2 Fatalities and injuries Fatalities: January tember 2007 There were nine collisions between trains and road vehicles at level crossings during the first nine months of Three resulted in fatalities to the road vehicle occupants. On 2 February, an Inverness-Wick train struck a car at Delny AOCL (automatic open crossing, locally monitored), near Barbaraville (Scotland). Two young men in the car died; the vehicle driver received major injuries. The crossing was operating correctly at the time of the accident. On 23 February, an empty coaching stock formation collided with a van on Gailes AHB level crossing (Scotland), killing the driver, who was standing near the road vehicle (a Network Rail van). 7 The crossing was working correctly at the time of the incident and the event is being treated as a suspected suicide. On 1 March, a train struck a car on Swainsthorpe AHB level crossing (South East), fatally injuring the driver. The barriers were down when the car attempted to zig-zag between them. The same crossing was the scene of a fatal collision in In addition, seven pedestrians died in accidents on level crossings: On 1 January, a male was struck on Paggetts footpath (FP) level crossing (South East). It was reported that he may have been playing chicken under the influence of alcohol. On 15 January, a 14-year-old girl was struck at Johnstown FP level crossing, near Ruabon (Western), whilst trying to retrieve her shoe. She was part of a group of who appear to have been playing chicken. On 24 January, a passenger was struck and fatally injured at Wokingham manually controlled barrier (MCB) level crossing (South East). The young male climbed over the barriers to join a train in the Up platform and was struck by a Down train. On 25 January, an elderly woman was struck on Doddington Road FP crossing (London North Eastern). 7 Though the victim in this case was outside his van at the time of the incident, he is still classed as a road vehicle occupant as he remained with said van whilst it was on the crossing. 12

13 On 31 July, an elderly woman walking her dog was struck at Sandringham Avenue UWC-T level crossing (London North West). On 16 August, a young male jumped over the barriers at Horsham Road CCTV level crossing (South East). On 16 tember, a member of the public was struck on Windwhistle FP crossing near Western-super-Mare (Western). Considering the number of crossings on the railway, the number of fatalities remains low, especially in relation to the number of pedestrians and vehicle occupants killed on the roads (821 pedestrians and cyclists and 2,321 car occupants in ) Trends in fatalities and injuries Fatalities on Network Rail managed infrastructure For a fatality to be classed as occurring on a level crossing, the deceased must have been on board a train, in a road vehicle, or have been a pedestrian on the interface at the time of the incident. Fatalities to trespassers who gained access to the running line are thus not included. It can often take time for a decision to be made regarding the exact location where a fatal accident occurred; this may also change as more information becomes available. Chart 8 shows the accidental fatalities that have occurred at level crossings since The data has been broken down by pedestrians, road vehicle occupants and train occupants. The majority of fatalities over the last 10-years (80%) involve pedestrians. (Note that suicides and suspected suicides are not included.) Chart 8. Fatalities at level crossings (excluding suicides) 16 Pedestrian Road Vehicle On board train Total Number of level crossing fatalities (Jan-) 8 Department for Transport (DfT). Road Casualties in Great Britain: 2006 Annual Report [Table 6c]. tember (Excludes 30 cases where the road user type was not reported.) 13

14 The first nine months of 2007 saw a fairly high number of accidental level crossing user fatalities. The total of 10, comprising seven pedestrians and three road vehicle occupants, is double the total for the whole of All the fatalities in the period plotted in Chart 9 resulted from accidents involving trains, with the exception of an incident in 2003 (in which an elderly man died after falling and striking his head on a crossing post). The chart shows that most fatalities and injuries occur to crossing users, as opposed to train occupants. 9 On 24 January, a passenger was struck on Wokingham MCB level crossing, being fatally injured after climbing over the barriers to catch a train waiting on the opposite platform. Chart 9. Fatalities and weighted injuries Fatalities and weighted injuries at level crossings (excluding suicides) Weighted injuries Public Workforce Passenger (on crossing) Passenger (on train) Fatalities Public Workforce Passenger (on crossing) Passenger (on train) (Jan-) While many of the high-profile incidents at level crossings involve motor vehicles, a large number of fatalities at the interface are due to the behaviour of those who are on foot. Indeed, most of the problem lies with misuse that is, a deliberate transgression of the rules (violation) or making of an honest mistake (error) Injuries by types of user behaviour As discussed previously the term trespass is now confined to situations where people go where they are never authorised to be, rather than where they behave inappropriately (either from error or violation) at places where they are allowed to be at certain times and under certain conditions. This revision was brought about by a review of the circumstances surrounding recent level crossing accidents. In several cases, correctly functioning equipment was unable to prevent some users from crossing in error. To describe the victims of such accidents as trespassers is not appropriate, as this behaviour is very different from taking a short cut along the trackside, for example. Thus, accidents occurring to users at 9 Most users are members of the public; the girls fatality injured at Elsenham (2005) are classified as passengers because of their intent to travel; the youth killed at Gomshall (2004) is also deemed a passenger as he had alighted from a train just prior to the incident. 14

15 level crossings are now classified as one of three types: proper use, misuse (error), or misuse (violation). The three categories concerned may be defined as follows: Proper use: users begin to cross entirely legitimately, but unforeseen events lead to a transgression (as when a motor vehicle breaks down half-way across a crossing, or the level crossing fails due to an error outside the user s control). Misuse: error - users cross when a train is imminent, but are honestly mistaken about its proximity and the warnings given by signs, sirens etc. Misuse: violation - users cross when a train is imminent, but deliberately disregard warnings and signage (as when users weave between half-barriers, or play chicken with oncoming trains). Chart 10 shows accidental fatalities and weighted injuries at level crossings, with a focus on user behaviour. Note that more than 90% of harm in the previous five years has resulted from misuse. Of this, error, as opposed to violation, constitutes the greatest proportion. 10 The large number of fatalities and injuries attributed to violation in 2004 is due to the incident at Ufton Nervet, where a train struck a car that had been deliberately parked on the crossing. The chart also shows that, in the first nine months of 2007, misuse has more than doubled, compared to the total for the first nine months of last year. Chart 10. Accidental injuries at level crossings by user behaviour Misuse: error Misuse: violation Proper use Total Fatalties and weighted injuries (Jan-) Table 2 lists the number of accidental injuries to pedestrians at level crossings and includes slip, trip and falls, along with pedestrians struck or trapped by crossing equipment. The total for 2006 was below the five-year average; in 2007, there have already been seven fatalities. However, this number is subject to change as new information (eg, through coroners 10 The figures may understate the proportion of injuries arising from violations: unless there is strong evidence that the crossing user deliberately disregarded signs or warnings, the event is classified as an error. 15

16 verdicts) becomes available. Four of the seven pedestrians were struck on footpath crossings as seen in Chart 11. Footpath crossings require the user to check whether it is safe to cross. They provide basic protection, most restricting access with a manual gate (which can be opened at any time). Some also provide miniature warning lights (as seen in Appendix 3). Where there is insufficient sighting for oncoming trains, a whistle-board is displayed and drivers must use their horn between the hours of 07:00 and 23:00 to warn pedestrians of their approach. There are several research initiatives published or in progress which are designed to address these issues. Table 2. Accidental injuries to pedestrians and cyclists at level crossings Year Fatalities Major Minor FWIs (Jan-) (Jan-) The Elsenham Station incident on 3 December, 2005 which involved the deaths of two young people, gained a high public profile. The causes received as much attention within the industry as they did in the media. Ultimately, the issues centred on the management of risk and whether the industry could or should have done more to demonstrate that it had managed the safety risk at the interface to a level that is as low as reasonably practicable. The infrastructure at Elsenham includes the station itself, along with a footpath level crossing. Safety at any such crossing will always depend to a degree on public compliance with the rules for its use. However, even with the best designs of crossings, there is always likely to be a degree of public non-compliance. From monitoring at Elsenham before and after the accident, it was evident that the level of non-compliance was significant. Despite this, responsibility for managing the risk to users rests with industry duty holders and local authorities. To this end, one of the formal inquiry recommendations involved the provision of a footbridge (which is now in use). 11 Chart 11 shows fatalities to crossing users in incidents involving trains (ie, excluding slips, trips and falls and similar accidents). It shows pedestrian and road vehicle occupant fatalities separately, and identifies the type of crossing involved. The figure for accidental fatalities to road vehicle occupants contrasts sharply, given that there were no such deaths in The fatality at Wokingham MCB level crossing on 24 January is the first at this crossing type for at least ten years. 11 RSSB, Formal Inquiry final report: Elsenham, p. 46, rec. 2 (RSSB, 2006). 16

17 Chart 11. Fatalities to crossings users struck by trains (excluding suicides) Pedestrians (public and passengers) footpath UWC-T UWC UWC-MWL Fatalities Road vehicle occupants (excluding suicides) 5 AOCL ABCL AHB MCB-CCTV MCB MCG Train occupants (staff & passengers) Jan Jan Jan - All the fatalities in Jan-t 2006 occurred in the London North East and South East territories (see Table 3). In the first nine months of 2007 the fatalities are more spread across the territories with at least one in each, apart from Scotland. Historically, London North East and South East have each averaged between two and three pedestrian fatalities per year; combined, these figures account for around two-thirds of all pedestrian fatalities. The two territories in question also contribute towards the highest fatality rate per thousand crossings. Table 3. Pedestrian fatalities at level crossings (by territory) London North East London North West Scotland South East Western GB Total 2006 (Jan-) per thousand crossings (Jan-) per thousand crossings Annual average per thousand crossings

18 4.3 Train accidents at level crossings The following section looks at the risk from road vehicles being struck at level crossings. Unlike incidents involving pedestrians, road vehicle accidents at level crossings have potentially much higher consequences, due to the possibility of derailing a train and harming its occupants. Note, however, that it is generally the car drivers or their passengers who suffer the most when collisions occur Trends in casualties resulting from road vehicles struck at level crossings Table 4 provides a breakdown of the number of fatalities (excluding suicides), major injuries and minor injuries that occur when trains strike road vehicles at level crossings. It also shows that 2006 was the first year to have no accidental fatalities from train striking road vehicles since The three accidental fatalities in 2007 (to tember) occurred in two incidents: On 2 February, an Inverness-Wick service struck a car at Delny AOCL crossing, near Barbaraville in Scotland. Two young men in the car died; the vehicle driver sustained major injuries. The other incident occurred when a train struck a car on Swainsthorpe AHB level crossing, fatally injuring the driver on the 1 March. The barriers were down when the car attempted to zig-zag between them. Table 4. Accidental injuries resulting from trains striking road vehicles at level crossings Year Fatalities Major Minor FWIs (Jan-) (Jan-) Note that there have been very few major and minor injuries suffered over the last couple of years. Between January and tember 2007, 3.1 FWIs were recorded. This is just over the tenyear average. 18

19 4.3.2 Trends in the number of road vehicles struck by trains at level crossings The number of collisions between road vehicles and trains decreased in There were nine such collisions during the first nine months of Three of these collisions resulted in at least one fatality (one is a suspected suicide). Most of the train accident risk at level crossings arises from the behaviour of road vehicle occupants. Chart 12 shows that, despite the rise in the crossing fatality count (compared with Jan-t 2006), the number of collisions has remained approximately the same. The tenyear average is around 18 collisions per year. User-worked crossings (ie, UWCs, UWC-Ts and UWC-MWLs) have accounted for 37% of collisions over the last ten years. This is in line with the highest risk crossings shown in Chart 6. Research on UWCs is continuing to discover how they are signed and protected. A detailed list of research can be found in Appendix 1. Chart 12. Collisions between trains and road vehicles at level crossings Number of events Footpath OC UWC-T UWC UWC-MWL AOCL/R ABCL AHB MCB-CCTV MCB MCG Jan- Since January 2002, there have been 92 collisions at level crossings, 18% of which have resulted in at least one fatality. Historically, most vehicle strikes occur at AHB, AOCL and user-worked crossings. Chart 13 confirms that the highest number of collisions take place at AOCLs. The fact that there are only 128 of them on the network illustrates the level of risk at this type of interface. Note that the consequences at these crossings are generally low, due to lower train speeds. This is because train drivers are required to check that the crossing is clear before proceeding. The second highest number occurs on AHB crossings, with 22 collisions having occurred in the last five-and-a-quarter years. AOCL and AHB crossings make up over half the total 19

20 number of collisions at level crossings, the latter often being sited on busy lines, in many cases with high road traffic levels. Chart 13. Collisions at level crossings by crossing type (Jan-) Fatalties Number of incidents Footpath OC UWC UWC-T UWC-MWL AOCL/R ABCL AHB MCB-CCTV MCB MCG Collisions leading to one or more fatalities have generally occurred at AHB level crossings over the last five years. UWCs may present the highest risk, but fatalities are not as common at this interface, as they tend to be sited in areas with lower traffic levels Derailments at level crossings Derailments at level crossings are the mainly the result of third-party activity, involving public road users, users of private crossings (such as agricultural machinery operatives or those herding cattle). Errors or other actions by these groups can lead to injuries or fatalities to train passengers or crews because a collision between a train and a vehicle (or, exceptionally, a large animal) can lead to a derailment and ultimately a major train accident. This type of accident is not unique to Britain. There have been a number of high-profile train accidents at level crossings across the world in recent years; the most significant example in 2007 occurred in Hatav, Turkey, when a freight train hit a farm truck. Seven people were killed. Another serious derailment occurred in Kerang, Australia, when a train hit a truck, killing 11. Chart 14 shows that: 2006 was the second successive year that none of the collisions at level crossings resulted in a derailment. This good news story has continued into tember There have been no derailments at MCB level crossings in the last 10-years. 20

21 Chart 14. Level crossing collisions causing derailments Number of events Footpath OC UWC-T UWC UWC-MWL AOCL/R ABCL AHB MCB-CCTV MCB MCG Jan- Table 5 lists the location and crossing type involved in collision and derailment accidents at level crossings since January There has not been such an incident since the end of 2004 the first time in nine years that there has been a period of this length between incidents of this kind. Table 5. Derailments caused by trains striking vehicles at level crossings (since 1991) Date Location Level crossing type Train type Vehicle Fatalities 28/05/1991 Cilyrychen 02/12/1991 Yalding Major injuries Minor injuries Automatic open crossing locally monitored (AOCL) Passenger multiple unit Car User-worked crossing with telephone (UWC-T) Passenger multiple unit Car /02/1992 Dimmocks Cote Automatic half barriers (AHB) Passenger multiple unit Lorry /04/1992 Mucking Automatic half barriers (AHB) Passenger multiple unit Lorry /11/1992 Star Lane Automatic half barriers (AHB) Passenger multiple unit Car Protected by minature warning 21/09/1994 Gardners lights (UWC-MWL) Passenger multiple unit Car /10/1995 Hauxton Automatic half barriers (AHB) Passenger multiple unit Car /03/1998 Swineshead Automatic half barriers (AHB) Passenger multiple unit Lorry /10/2000 Pooley Green Automatic half barriers (AHB) ECS multiple unit Bus /02/2001 Lancing User worked- no phone (UWG) Passenger multiple unit Van /07/2001 Tisbury Quarry Protected by minature warning lights (UWC-MWL) Passenger multiple unit Lorry Automatic open crossing locally 23/08/2001 Ardrossan Harbour monitored (AOCL) Passenger multiple unit Bus Automatic open crossing locally 11/09/2001 Tregoss Moor monitored (AOCL) Passenger multiple unit Lorry /04/2002 Blaxhall Automatic open crossing locally monitored (AOCL) Passenger multiple unit Tractor /12/2002 Six Mile Bottom Automatic half barriers (AHB) Passenger multiple unit Car /01/2003 New Fishbourne Automatic half barriers (AHB) Passenger multiple unit Car /09/2003 Chapel User-worked crossing with telephone (UWC-T) Passenger HST Tractor /08/2004 Coswarth Automatic open crossing locally monitored (AOCL) Passenger multiple unit Tractor /11/2004 Ufton Automatic half barriers (AHB) Passenger HST Car /12/2004 Pumphouse User worked crossing (UWC) Passenger multiple unit Van

22 The most recent collision to lead to a derailment at a level crossing occurred when a passenger train stuck a van at Pumphouse UWC (near Blotoft Siding) on 6 December 2004 (London North Eastern). The two vehicle occupants received fatal injuries Incursions via level crossings Level crossing incursions are different from the collisions mentioned previously in this report, in that the vehicles involved generally enter the infrastructure via a level crossing and are not struck by trains. This can occur either in a deliberate or mistaken act, in the aftermath of a road traffic accident, or when a road vehicle turns off the road and drives along the track, as occurred on 23 March at Stone AHB in Faversham (South East). Here, the road vehicle driver was passing from the Up to the Down side and turned right along the track, blocking both lines. The train driver was shaken by the incident, but no collision occurred. Occasionally at automatic crossings, heavy traffic prevents a crossing from being cleared in time for a passing train; similarly, road vehicles can break down, becoming stranded on crossings causing delays and near miss incidents. Chart 15 shows that the recent trend for incursions via level crossings is reasonably sporadic. Almost all of these event types result in the vehicle fouling the line. Note that there have been two incidents where a vehicle has ended up not foul of the line the first time this has occurred since Chart 15. Incursions via level crossings Not foul of line Foul of line RV incursions via level crossings Jan Jan - Many incursions where a train strikes a road vehicle occur away from level crossings. During the last five years, the rate has remained fairly static, at around four accidents per annum. The chapter on Road-rail incursions and bridge strikes looks at this type of incident in more detail. 22

23 4.3.5 Trains striking gates at level crossings Chart 16 shows the number of trains striking level crossing gates or barriers. Passive manual crossings, including user-worked crossings and trainman-operated crossings, contribute most to the overall trend. There were no incidents of trains striking gates or barriers in However, on 15 November 2004, a passenger train struck the gates at Rowston MCG level crossing (London North Eastern) resulting in one major injury to a passenger. This is the first incident of a train striking a level crossing gate or barrier that has resulted in a major injury since at least In general, trains only strike barriers when a road vehicle has pushed them out of alignment that is, foul of the running line immediately prior to the train s appearance. Chart 16. Trains striking level crossing gates or barriers 9 8 Passive manual Active manual Active automatic 7 Number of incidents (Jan-) 2007(Jan-) The three incidents in 2007 involved two incidents of trains striking gates and one barrier strike: High winds caused the crossing gates at Pleasants Lane level crossing (UWC-T) to land on the track; they were subsequently struck by an empty coaching stock (ECS) train. A passenger train struck a gate at Caersws MWL crossing (Western). The gate had come open due to defective clasp. A lorry jumped the lights and struck a barrier at Gatehead AHB level crossing and caused the barrier to foul the line. A freight train subsequently struck the barrier. 23

24 4.4 Near misses and misuse at level crossings Due to the comparatively small number of accidents that occur at level crossings, it is difficult to recognise trends using accident data alone. Therefore, the industry also collects information on near misses and misuse. Since any collision has the potential to cause a derailment and result in a major train accident, these incidents need to be looked at closely Level crossing near misses Chart 17, which shows the changes in near misses at level crossings, paints a mixed picture of trends in level crossing safety. Reported near misses involving motor vehicles fell to a rate of around 150 per annum, which is relatively low compared with previous years. However, near misses with pedestrians remained at the fairly high levels seen during 2006, and well above the numbers recorded earlier in the decade. Viewed in conjunction with the PIM, it may be that Network Rail s Don t run the risk campaign, which focussed mostly on drivers in 2006, has had some beneficial effects on driver behaviour (although it is probably too early to say definitively). Note, however, that according to Network Rail, the public recall rate for the advertisements has been exceptionally high at around 15-25% for the general population, rising to 74% for authorised users 12, who received private users packs. The campaign has now started to focus more on pedestrian misuse. For more detailed information, see section 7 (Initiatives). Chart 17. Near misses with motor vehicles and pedestrians at level crossings 250 Motor vehicles 250 Pedestrians Near misses (annual moving average) Passive Automatically protected Manually protected These include local farmers with access to private crossings. 24

25 Near misses with road vehicles Near misses are usually reported by train drivers who have taken action to avoid a collision, or feel that they came close to striking a road vehicle or pedestrian. This type of incident reporting can be somewhat subjective and conditions do not always allow for accuracy. In general, between 20 and 30 near misses are reported for every pedestrian hit by a train, while around ten near misses are reported for every road vehicle struck. Because of the difficulty in ensuring objectivity, overall trends are of greater significance than the actual numbers. Chart 18 shows the total number of near misses with trains at level crossings by month, along with the associated annual moving average (AMA). In , the pattern was more seasonal in nature, with peaks evident in the summer months and troughs in the winter. This has altered slightly from mid-2005, when the numbers were more sporadic; 2006, for example, was well below the five-year average, with the first half of 2007 showing a similar trend. As noted on page 25, there has been a downward trend since the start of the campaign. Chart 18. Road vehicle near misses chart Near miss AMA Road vehicle near miss per month Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug

26 Chart 19 presents the average number of monthly near misses for each type of level crossing. The data is normalised by the number of level crossings of each type. Automatic (AOCL) and open crossings (OC) have relatively high rates of near misses with road vehicles. This is due to the lack of barriers, which makes it easier for road users to cross (deliberately or inadvertently) in front of a train. It is not surprising that gates controlled by railway personnel present the lowest near miss figures. Chart 19. Road vehicle near misses per crossing by crossing type 0.04 Average number of monthly incidents (Jan-) ABCL AHB AOCL/R Footpath MCB MCB- CCTV MCG OC UWC UWC- MWL UWC-T Other Research into what flashing red lights mean to motorists (as compared to fixed red lights) is in preparation, following a survey undertaken in Germany, which showed that one-third of respondents did not know what they signified ADAC (München) at SELCAT Workshop, Lille, May

27 Near misses with pedestrians The number of near misses for the period January 2002 tember 2007 is shown in Chart 20. Note the peaks in the summer months: these are due to increased crossing usage and the longer daylight hours, which allow more incidents to be witnessed. The AMA shows that there has been a steady increase in the numbers of near misses since However, while the reports filed in 2006 dropped slightly from February onwards, an upward trend was evident at the beginning of this year which then began to level out towards the end of quarter 3. Note that the Don t run the risk campaign has focused on pedestrian safety since May 2007 (see section 7 for further details). Chart 20. Pedestrian level crossing near misses Near miss AMA Pedestrian near miss per month Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug

28 Chart 21 shows the number of near misses with pedestrians for each type of level crossing, with data normalised by the number of crossings of each type. The highest levels of pedestrian near misses occur at AOCL, UWC-MWL and closed circuit television (CCTV) crossings. Chart 21. Pedestrian near miss per crossing type (Jan-t) Number of near miss incidents normalised by crossing numbers (Jan-) MCG MCB MCB- CCTV AHB ABCL AOCL/R OC UWC UWC-T UWC- MWL footpath As we saw with road vehicle users, AOCLs do not have barriers and require the user to obey the warning signs. Open crossings (like all crossings) are generally safe when used correctly; those who are impatient or try to beat the train can sometimes underestimate the time it takes for a train to arrive at the crossing (or underestimate how long it actually takes to get to the other side) Near miss incidents by time of day After considerable research into the noise nuisance and safety implications of train horns, a new standard was introduced. This took effect in April 2007 and required the industry to implement three changes: The introduction of a night-time quiet period, between 23:00 and 07:00, when trains will no longer routinely sound their horns at whistle boards. Where the technology is available on the train, drivers will only use the low tone horn at whistle boards. For all new or replacement horns on trains capable of travelling up to 160kph, a much lower minimum sound pressure level has been established (a maximum sound level has also been introduced (Min 101dB and Max 106dB)). 28

29 To attempt to understand the effects of this change, we have looked at the number of near misses by time of day. Chart 22 compares the number the occurred between April-t 2006 and April-t The numbers of near miss incidents during the night have risen from four to 12 since the quiet period was introduced. Given the relatively small number of incidents, it is too early to say if there has been a statistically significant change in the number of near misses. However, we will continue to monitor this data in the coming months. It is also worth noting that there has been a general increase (25%) in the number of near miss incidents during the times when horn usage is permitted. Chart 22. Comparison near miss incidents April-t 2006 and April-t tember August 16 Number of near miss events July Jun May April :00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 00:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 Time of day Time of day Chart 23 shows a breakdown of near misses by time of day. The peak between the hours of 15:00 and 18:00 coincides with the close of school day. The quiet period (between 23:00 and 07:00) saw approximately 6% of near miss incidents recorded over the five-and-aquarter years in question. Note, however, that this figure may not be directly comparable with the non-quiet period because, as darkness falls, train drivers are less likely to see (and thus report) near miss incidents. 29

30 Chart 23. Pedestrian level crossing near miss by time of day (Jan 2002 t 2007) 120 number of near miss events (Jan ) Level crossing misuse Misuse refers to a variety of situations in which crossing users attempt to traverse a crossing when it is unsafe to do so, or otherwise fail to use it correctly. Typical reports involve road users jumping red lights or trying to beat the barriers, pedestrians ignoring warnings (including cases of children playing chicken ), and, at user-worked crossings, gates being left open, telephones left off the hook, and users failing to report when they are clear of the tracks. Misuse is thought to be considerably under-reported because the majority of crossings are not monitored. Again, in the light of these difficulties, overall trends are more significant than the numbers themselves. Chart 24 shows that the rising trend since 2002 has been reversed (from mid-2006). Indeed, a promising decrease in the number of reported misuse incidents is now evident. This coincides with the Network Rail s Don t run the risk campaign, which was first aired in May 2006, suggesting that the initiative may well be having a beneficial effect. 30

31 Chart 24. Road vehicle misuse 120 Misuse AMA 100 Road vehicle misuse Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Chart 25 presents the rate of level crossing misuse per crossing per territory. The same seasonal variations in misuse are apparent. Chart 25. Road vehicle misuse by territory South East London North Eastern Western Scotland London North West Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug

32 Chart 26 shows the number of road vehicle misuse incidents for each type of level crossing. MCB and MCB-CCTV crossings record the highest number of incidents per crossing, while AOCL show the next highest. MCB and CCTV crossings are the two of the limited number of monitored crossings and therefore have a higher likelihood of an act of misuse being detected and reported. Chart 26. Road vehicle level crossing misuse per crossing (Jan-t) Average monthly numbers of incidents (Jan-) ABCL AHB AOCL/R Footpath MCB MCB- CCTV MCG OC UWC UWC- MWL UWC-T Other Misuse by user behaviour and crossing type Level crossings with gates that are left open to the road present a danger to road vehicle occupants, who are likely to believe that the crossing is safe to traverse even when it is not. The frequency and nature of reported crossing misuse varies by crossing type (see Chart 27). User-worked crossings account for almost two-thirds of the total number of misuse incidents that are recorded in SMIS. Those with telephones present the most misuse and many of these relate to the user leaving the gates open or failing to report that the crossing is clear. This is perhaps due to the fact that these types of crossings are monitored and are witnessed by railway personnel. 32

33 Chart 27. Level crossing misuse by category and crossing type (Jan-t) footpath UWC-MWL UWC-T UWC OC AOCL/R 3.1% 9.2% 45.7% 12.0% 0.6% 0.9% ABCL AHB MCB-CCTV MCB MCG 0.4% 3.0% 5.0% 9.6% 10.4% Other Gate / barrier left open / raised Vandalism Phone left off the hook User fails to report clear Road vehicle strikes / struck by crossing equipment 0% 10% 20% 30% 40% 50% 60% 70% Percentage of misuse reports (Jan-) Misuse is generally referred to as poor behaviour by level crossing users and does not necessarily endanger the user first hand. However, such actions have the potential to endanger future users of the crossing. There are also issues relating to performance delays due to users failing to report that the crossing is clear when it required to do so. Typical signpost at a UWC-MWL crossing 33

34 4.4.4 Near miss and misuse by crossing type It is difficult to determine the true levels of level crossing misuse on the railways as the majority of interfaces are not monitored. However, factors which are likely to increase reporting rates include: Local and national initiatives (to raise awareness, improve reporting or increase conviction rates for crossing misuse). Accidents. Changes in personnel at the local level. Chart 28 shows the number of near miss and misuse incidents when a pedestrian or road vehicle uses the crossing when it is unsafe to do so. It is clear that the highest percentage of misuse occurs on user-worked and manually controlled crossings. Road vehicle users account for around 56% of misuse and near miss level crossings. Chart 28. Misuse and near miss by crossing type (Jan-t) footpath UWC-MWL UWC-T 2.1% 8.0% 18.4% UWC OC AOCL/R 1.3% 6.1% 11.1% Pedestrian crosses when unsafe Near miss with pedestrian Road vehicle crosses when unsafe ABCL 0.7% Near miss with motor vehicle AHB 11.3% MCB-CCTV 17.5% MCB 19.6% MCG 3.9% 0% 5% 10% 15% 20% 25% 30% Percentage of misuse reports (Jan-) At the more serious end of the misuse spectrum are incidents where a pedestrian or vehicle occupant uses the crossing when it is unsafe to do so. Examples include crossing users trying to beat the lights or pedestrians running across when the barriers are lowering. Such incidents are not always classified as a near miss with a train. The number of motorists using crossings when it is unsafe to do so fell slightly in 2006 a trend which continued into the early half of 2007, as seen in Chart 29. Again, Network Rail s media campaign, which focused on driver behaviour (commenced in May 2006), could very well have played a part in this improvement. The AMA fell by 14% from the end of April 2006 to tember this year. 34

35 Chart 29. Serious misuse (road vehicle crosses when unsafe) includes near misses RV crosses when unsafe Monthly average (Jan-) Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug It can be seen from Chart 30 that the number of pedestrians crossing when it is unsafe to do (including near miss) so has increased year on year. This could indicate increased reporting. Pedestrians misusing manually controlled crossings account for almost 40% of all incidents and 26% of incidents on user-worked crossings. Chart 30. Serious misuse (pedestrian crosses when unsafe) by year and crossing type Number of incidents footpath OC UWC-T UWC UWC-MWL AOCL/R ABCL AHB MCB-CCTV MCB MCG (Jan-) 35

36 5 Risk at the road-rail interface A typical day sees around eleven reports of road vehicles colliding with underline and overline rail bridges. Most do not cause any significant damage to the bridge and pose little threat to the safety of trains. However, around one incident per week is classed as serious or potentially serious due to the extent of the structural damage involved. Incursions from bridges can also lead to serious train accidents, as at Great Heck on 28 February This accident began with a Land Rover and trailer (itself carrying another vehicle) leaving the hard shoulder of the M62 immediately before the protecting barrier of a bridge across the East Coast Main Line. It then plunged down the steep road embankment before continuing down the railway embankment, into the path of a passenger train. The impact caused the train to derail, though it remained substantially in line and upright until it hit a set of points. This deflected it into the path of a northbound freight, which was travelling along the Down line at that moment. Incidents like Great Heck are a reminder that the interface between the road and railway can have significant safety implications. 5.1 The risk from the road-rail interface Chart 31 reveals that the residual risk (1%) at the road-rail interface involves two main groups: road vehicle occupants in collisions with trains and train occupants (including workforce). The chart also includes the risk from bridge strikes. Note that level crossings are excluded. Road vehicle occupants who are killed in road traffic accidents are also excluded, as they are not under the direct control of the railway. Chart 31. Fatalities and weighted injuries at road rail interface (excludes level crossings) (SRMv5) Total system risk Risk associated with road rail interface (excluding level crossings) Train occupants 35% 1% Road vehicle occupants in collisions with trains 65% 36

37 5.2 Post-Great Heck risk assessment There can be no doubt that Great Heck was a turning point for our industry. After the accident, Network Rail, the Highways agency and local authorities set about mitigating the risk from road vehicle incursion incidents by carrying out a risk assessment of all road bridges and roads running parallel to the railway (neighbouring sites). This was developed in a three-stage process: 1. Identify the number of road bridges and neighbouring sites and their locations. 2. Risk asses each location. This was carried out as a joint exercise between Network Rail and the individual local authority; a ranked score was then given. 3. Implement risk mitigation measures to reduce the risk at those sites identified as being high risk. The risk assessment is based on a risk ranking methodology. Those sites which returned a risk ranking in excess of 90 were taken on to Stage 3. Following a more detailed assessment, the type of risk mitigation measure is considered. Funding for mitigation measures is provided by Network Rail and the local authority equally. Table 6 shows the number of sites identified in each of Network Rail's five territories and those that have been risk assessed thus far. Of the total number of sites identified on the network, 9,469 have been risk assessed (18% have a score above 90). The proportion of sites that have a score of 90 or more after Stage 2 is around 56%. Approximately 9% of these do not require further mitigation as it was not reasonably practicable. Finally, around 177 sites have been agreed but have not been implemented and require further consideration for the implementation of risk mitigating measures. Table 6. Number of road-rail incursion sites that have been risk assessed and modified Territory Description Bridge Sites Neighbouring Sites Total Sites identified Risk ranked Score > South East Modified Sites identified Risk ranked Score > Western Modified Sites identified Risk ranked Score > London North West Modified Sites identified Risk ranked Score > London North East Modified Sites identified Risk ranked Score > Scotland Modified Sites identified Risk ranked Stage 1 score > Network total Stage 2 score > Stage 3 not practical Stage 3 agreed mod M odified to date Source: Network Rail. Data recorded to Period 5 (August 2007). 37

38 6 Road-rail incursions and bridge strikes 6.1 Headlines There was a 6% increase in the total number of bridge strikes between January and tember 2007, compared to the same period in However, serious bridge strikes fell by 53% in the first nine months of Around two bridge incursions occur each year. An example of an incident that could have caused a derailment occurred in April 2007, when a motor vehicle flipped off Deganwy Road overline bridge. Incursions through fences account for almost 70% of all incursions on the network. Vehicle incursions through access points have fallen in recent years, although the first nine months of 2007 have seen the count rise to three more than at the same stage in There were no derailments from incursions between January and tember The most recent event was at Copmanthorpe on 25 tember 2006, when a vehicle crashed through a boundary fence and onto the line. 6.2 Fatalities and injuries Over the last five-and-and-a-quarter years, there has been a total of 4.5 FWIs contributing to the risk at the road-rail interface (this excludes level crossings). In the last five years, two road vehicle occupant fatalities have occurred in incidents where an incursion has led to a derailment. 6.3 Bridge strikes Bridge strikes present a level of risk to train passengers and railway workers, and damage to bridges can cause major disruption to train services. They often result in the death or serious injury of the driver or passengers in the road vehicle involved, as well as other road users. Whilst a derailment from a bridge collapse has not happened recently in Britain 14, there is still an underlying level of risk from road vehicles striking such structures. In 2006, there were, on average, over five road vehicle strikes at road-under-rail (underline) bridges each day; this represents a marginal increase in the total number compared with the previous four years. The consequences of a bridge strike can vary from minor paint scrapes, to more serious where the bridge and track may get shifted sideways by the collision with the potential for an oncoming train to be derailed. Currently all trains are stopped when approaching a bridge that has been struck, until a nominated engineer can examine the structure, unless local instructions allow otherwise. 14 Such incidents have occurred in the past, however. For example, the swelling of the River Towy, near Llandelio, Carmarthenshire, washed the Glanrhyd Bridge away from its supports, with fatal consequences: a Swansea-Shrewsbury service plunged into the torrent on 19 October 1987, killing four people. See Report on the Collapse of Glanrhyd Bridge on 19 th October 1987 (HMSO, 1990). 38

39 6.3.1 Performance delays due to bride strikes While the SRM shows that the risk from bridge strikes is relatively low (as seen in Chart 31), the associated performance delays incur great cost to the industry. The total train delays caused by bridge strikes for the last three years is shown in Table 7. Although the number of bridge strikes has gradually increased over the years (as seen in Chart 32) the average delay per incident (DPI) and cost per incident (CPI) has reduced in the last two years. Table 7. Performance delays due to bridge strikes Source: Network Rail (TRUST database) Year Incidents Minutes Cost DPI CPI ,242, , ,653, , to ,080, , Total number of bridge strikes Chart 32 shows the annual number of bridge strikes from 2002 and includes both overline and underline bridges. There was a 6% increase in the number of bridges struck between January and tember 2007, compared to the same period in As the number of incidents recorded last year was higher than any other in the review period, this figure (2198) may be surpassed by December. Network Rail has a severity classification scheme for bridge strikes and are classified as; serious, potentially serious or not serious. A description of each can be found in Appendix 5. Chart 32. Total bridge strikes by severity Unknown Serious Potentially serious Not serious Number of bridge strikes per year (Jan-) 2007 (Jan-) 39

40 Chart 33 is cyclical in nature, with the highest level of bridge strikes occurring in the third quarter of each year. At present, we do not have any firm evidence as to why this might be the case. However, it may relate to the activities surrounding the harvest season, during which more heavy goods vehicles carrying farming equipment and crops would be active. Chart 33. Monthly bridge strikes Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug The serious bridge strike count has risen every year, bar one (2005), since This is evident in Chart 34. However, from January to tember 2007 there was a 53% drop in the number of these incidents, compared to the same period in This fact brings hope that the upward trend will be reversed by the end of Chart 34. Serious bridge strikes Serious bridge strikes (Jan-) 2007 (Jan-) 40

41 As we can see in Chart 35, 94% of all bridge strikes are deemed not serious from a railway operations point of view. This indicates that the risk from such events is low, although potentially high-consequence accidents can occur on occasion. This is illustrated by the incident at Oyne (Scotland) in 1978, when a rail-over-road (overline) bridge was hit by an item of construction plant causing severe twist to the track just a few minutes before a passenger train was due. The train driver saw what had happened too late to stop short of the bridge and the train was derailed. There were five minor injuries as a result. Underline bridges account for 92% of all bridge strikes since 2002 (see Chart 36). Road vehicles striking underline bridges are a concern for the railway, as they are often struck by lorries or buses that are too tall to pass beneath. When a bridge is struck, all trains are stopped in the area until an engineer is called out to examine the structure for any damage. This causes substantial disruption to the rail industry through delays and damage. Chart 35. Bridge strike severity (Jan-t) Potentially serious 2% Serious 1% Unknown 3% Not serious 94% Table 8 provides a breakdown of the total number of bridges that exist at the road-rail interface. 15 The table shows that there are around 41% more underline bridges in the roadrail interface scope than their overline counterparts. Table 8. Breakdown of road-rail interface bridges by Network Rail territory South East London North Eastern London North Western Western Scotland Unknown Area Overline Bridge - Viaduct Overline Bridge Total overline bridges Underline Bridge Underline Bridge - Intersection Underline Bridge - Viaduct/Intersection Total underline bridges Note that side bridges and bridges only carrying pipes and electrical equipment are not included within the total numbers as they have no impact on the road-rail interface and are therefore outside the scope of this report. 41

42 Chart 36 shows that, between 2002 and 2007, around 7% of all bridge strikes occurred at overline bridges. Such accidents present an obvious risk to the railway. Although the greatest hazard comes from a vehicle or its load (or both) fouling the line, debris from dislodged masonry could also lead to a derailment. Chart 36. Bridge strikes by bridge type 2002 t 2007 Rail over water 0.24% Road over rail (overline) 6.36% Rail over road (underline) 91.92% Side Bridge 0.07% Viaduct 0.16% Footbridge 0.11% Unknown 0.98% Rail over rail 0.16% Table 9 shows the number of bridge strikes broken down by seriousness and territory. The regional totals show that a fairly even rate has been returned between 2002 and The South East remains the worst-affected area in terms of the total number of strikes, probably because it has the highest population density and hence road traffic. Note, however, that Western has seen the greatest number of serious strikes in the first nine months of this year. Table 9. Bridge strikes by seriousness, broken down by territory South East London North East London North West Western Scotland Unknown (incl. not entered) Seriousness to Total Not serious Potentially serious Serious unknown SE total Not serious Potentially serious Serious unknown LNE total Not serious Potentially serious Serious unknown LNW total Not serious Potentially serious Serious unknown W total Not serious Potentially serious Serious unknown SCT total Not serious Potentially serious Serious unknown Unknown total Totals

43 6.3.3 Underline bridge strikes Table 10 below lists the 20 most commonly struck bridges since Grantham fares particularly badly, with three different structures making an appearance in the top 12. Entry no.4 is the famous White House Bridge in Swindon, which, despite being painted bright yellow and adorned with the slogan LOW BRIDGE in huge black letters, still suffers incidents involving double decker buses and high-sided lorries with frequency. All of those listed are rail-over-road (underline) bridges. Table 10. Top 20 bridges struck between 2002 and 2007 (tember) Rank Location of Bridge Bridge Number Total strikes 1 Grantham Lower Sydenham Grantham Swindon Latchmere Jcn Lichfield City Hitchin Hullavington Ely Wilton Landor Street Jcn Grantham Cook Street Underbridge Wimbledon (Main Line) Dalry Walkden Shortlands Homerton =19 Woking =19 Brookwood =19 Ashford (Rye Line) =20 Cleland =20 Langley Network Rail has had in place a national bridge strike initiative since 2004, to reduce the consequence of bridge strikes. The robustness of underline bridges to withstand a strike by road vehicles is being assessed by route. This has led to Signal Box Special Instructions (SBSI) for each bridge whereby the signaller can instruct train drivers as follows. The assessment results in four categories for a bridge: 1. GREEN for such massive bridges that a strike has no consequence to the structure. 2. DOUBLE AMBER the first train approaches the bridge under caution from the signaller at 5mph and reports on any debris or damage; if there is no debris, the line returns to normal speed. 3. AMBER all trains approach under caution, the first at 5mph and if no debris is seen following trains at 20mph, until the bridge has been examined. 4. RED all trains are stopped until the bridge has been examined. 43

44 6.3.4 Overline bridge strikes leading to debris on the line An example of an incident which could have been so much worse occurred on 6 November 2006, when an overline bridge at Llangaffo was struck by a road vehicle. As a precaution, all train movements were stopped, but a Chester service ran through the debris caused by the collision, despite receiving an emergency Stop message. This led to damage to the train s sander and lifeguard; the driver also reported an air leak, which had to be repaired temporarily before the train could proceed. Chart 37 shows the number of incidents where a bridge strike has resulted in debris falling on to the track or infrastructure below. The proportion of these incidents where a train has struck the debris is also shown (around 14%). Though these incidents are infrequent around one two per year, they are still cause for concern. The chart also shows the number of bridge strikes that have caused part of the bridge or road vehicle to overhang the railway line. It includes incidents where a bridge structure has become unsafe after a strike, with materials being left to hang precariously. Under these circumstances, a bridge would require reinforcement or repair to prevent further debris from falling on to the track. It is therefore necessary to stop all train movements as further vibrations could cause blocks or debris to become dislodged. Chart 37. Overline bridge strikes resulting in debris on the line 180 Number of bridge strikes Debris struck by train Debris Overhang Bridge struck to Network Rail is reviewing all bridge strike documentation and procedures, including initiating a change to the Rule Book whereby train drivers will be allowed to examine structures after a strike on an overline bridge to make an initial assessment of the severity of the strike. This is to allow the trains to start running again earlier than under the current rule. The proposal is that, when a road vehicle strikes a road-over-rail bridge, trains may be driven under caution up to the bridge at 5mph and pass under the bridge if the line appears to be clear. This proposal is being risk assessed by Network Rail. 44

45 6.4 Road-rail incursions by entry point Road vehicles and aircraft are considered in this analysis have gained access onto railway property via bridges, fences, access points and at level crossings. In previous reports, it was not possible to distinguish between a vehicle that had crashed through a boundary fence and a vehicle that had gained access to railway property via a gate. Such incidents were grouped into the fence entry point category. Chart 38 shows the total number of vehicle incursions and those that ended up foul of the line, along with the FWIs for this type of incursion. It does not include vehicles incursions at level crossings (which are described earlier in this report). However, it does include incursions onto the railway after road traffic accidents, road vehicles abandoned or driven onto the railway and railway/contractor owned road vehicles via access points. The chart also shows the number of incidents where an incursion has led to a road vehicle being struck by a train (shown in yellow). The total numbers of incursions leading to train strikes by incursion point can be seen in Chart 43. On average, there are around 66 road rail incursions each year; of these, almost half end up foul of the line. This is a high proportion of incursions that have the potential to cause catastrophic train accidents. Chart 38. All types road-rail incursions (excluding level crossings) All road-rail incursions (fences, bridges and access points) Struck by train Not foul of line Foul of line FWI (Jan-) 0 45

46 6.4.1 Incursions from bridges The Great Heck incident remains the most recent bridge incursion where a road vehicle caused a derailment. It occurred when a vehicle left a motorway and fouled the track. It was subsequently struck by a passenger train, which derailed and collided with an oncoming freight train. Ten fatalities (including four members of staff), 38 major injuries and 36 minor injuries resulted. Although, this type of incursion is rare, such incidents present risk where the road meets the rail. The number of incursions via bridges and level crossings is much lower than those via fences. The trend via bridges is difficult to determine, as the numbers of incidents are so low. However, quarter has already seen the same number of incidents as the whole of This is portrayed in Chart 39. Incursions via bridges, including those where a vehicle crashes through a fence on the approach to a bridge (resulting in it being struck by a train), are sporadic. Indeed, only around two incidents occur each year, although a potentially serious bridge incursion occurred in April 2007, when a motor vehicle flipped off Deganwy Road (North West) onto the railway. Fortunately, however, there were no trains in the area at the time. Chart 39. Road Vehicle incursions from bridges Not foul of line Foul of line RV incursions via bridges Jan Jan - Another example of a serious incident occurred on 6 November 2003 at Poton (South East). A previous accident on 21 October had left the bridge parapet and safety railings severely damaged and unable to withstand a further vehicular impact. Thus when a vehicle struck the bridge on 6 November, there was nothing to impede its progress; the momentum caused by its speed carried it across the parapet and onto the running lines below. Fortunately, there were no trains passing at the time of the incident. 46

47 6.4.2 Incursions through fences Where roads are concerned, the landowner is responsible for providing fencing at the boundary point. For the railway, it is the infrastructure manager (IM) on whom the onus lies. Monitoring the risk at this interface is thus of great importance to the industry. This is evident from the following incident, which occurred on 25 tember Shortly before 21:00, a car left a country road in Copmanthorpe, North Yorkshire, and crashed through the railway boundary fence onto the line. The driver, a local man, was killed when his vehicle was struck by a 100mph passenger train bound for Edinburgh Waverley. The Class 221 Super Voyager unit derailed, but remained upright although extensive damage was caused to the half-mile of track beyond the point of collision. None of the traincrew or 74 passengers were harmed. The Rail Accident Investigation Branch (RAIB) has released a report into the incident. It made no recommendations in relation to the cause of the accident, although two other recommendations came out of the investigation. 16 Note that there are usually between 40 and 50 incursions by road vehicles through fences each year, with the number fluctuating. A closer look at the data in Chart 40 reveals that there is no pattern in the monthly number of incidents. Incursions through fences account for more than 70% of all incursions. The number of fence incursions (Jan-t 2007) has fallen by 33%, compared to the same period last year. Chart 40. Incursions through fences Road vehicle incursions through fences Not foul of line Foul of line Jan Jan

48 6.4.3 Incursions through access points In previous road incursion reports, access points were grouped with fences, thus it was not possible to distinguish between the numbers of incursions via the two interfaces. However, reclassification means that, for the first time, we have a clearer picture of the numbers for each specific group. Note that, where the railway is reached via an access point, they are generally spilt into three subgroups: Rail-owned vehicles which have legitimate access to the track in order to carry out works. Occasionally these can end up fouling the line after getting stuck or parking in an inappropriate place. Vandals/criminals who have gained access to the lineside in order to perpetrate a crime by either leaving a vehicle (usually stolen) on or next to the track. Sometimes rail owned vehicles are stolen and driven onto the line and set ablaze before the criminals leave the scene. Members of the public wanting to take a shortcut via the railway, who willingly disobey the rules and end up being struck by a train. Chart 41 shows that the numbers of vehicle incursions from access points have been falling in recent years. Despite this, the first half of 2007 has seen the count rise to three more than at the same stage in This means that an overall increase in these incidents may be evident by the close of Chart 41. Incursions through railway access points Not foul of line Foul of line 25 RV incursions via access points Jan Jan - 48

49 6.4.4 Vandalism Chart 42 shows that the number of vehicles being left lineside by vandals fell between 2002 and However, the number of incidents that have occurred between January and tember 2007 already equals the count for the whole of Note that incursions of this type can occur via access points and level crossings (see Table 11). More vehicles are placed intentionally foul of the line than not foul of the line. Once again, this highlights the danger of these crimes, which have the potential to result in a serious derailment. In February this year, for example, some youths stole a dumper truck and drove it along the line before overturning it on the permanent way at Blandford Road. Such a vehicle could have derailed any train that struck it. The most recent act of vandalism that caused a collision and subsequent derailment occurred in the Bucksburn area of Aberdeen in tember A dumper truck was set rolling down an access road and burst through a fence before coming to rest foul of the line. A passenger train struck the JCB, causing major injuries to the train driver. Chart 42. Road vehicles left on or near the line purposefully Crime and vandalism 30 Not foul of line Foul of line Crime - vehicles left on the lineside Jan Jan Table 11. Road vehicles left on or near the line by location - Crime and vandalism Year Access point Fence Level crossing Total (Jan-) (Jan-)

50 6.4.5 Incursions resulting in train strike Incursions on the railway are relatively rare occurrences. However, they have the potential to lead to serious consequences if a train collides with a vehicle. Chart 43 indicates that, of the 66 vehicles which break through the railway s boundary year-on-year, only 6% are actually struck by a train. Most incidents involving trains striking road vehicles stem from fence incursions. This may be due to the fact that fences essentially cover the entire railway; thus they interface with much more of the lineside than bridges and access points. In 2003, a road traffic accident caused two fatalities and one major injury in the same incident. A Land Rover careered off a motorway at Groveley Lane bridge in Longbridge and ended up on the running line. A southbound passenger train clipped the vehicle as it passed. Chart 43. Road vehicles struck by train after incursion 6 Fence 5 Bridge Access point Incursions resulting in train strike (Jan-) 2007 (Jan-) Table 12 shows the location and access point of the incidents in 2007 where at train has struck a road vehicle. Table 12. Location and access point of road vehicle incursions leading to train strike Date Location Train Incursion site 17/09/2007 Markinch Passenger multiple unit Fence 20/05/2007 Heaton Norris Jcn Passenger multiple unit Access point 29/05/2007 Keith Passenger multiple unit Fence 09/03/2007 Normanton LC Passenger multiple unit Fence 19/08/2007 Garnqueen North Jcn Passenger multiple unit Access point 50

51 6.4.6 Road vehicle incursions leading to derailment (other than at level crossings) In addition to the road vehicle occupant risk and, as for the level crossings, there is the potential for a derailment to lead to multiple fatalities. The number of road vehicles that have gained access to the railway and led to a derailment can be seen in Table 13. The most recent event was at Copmanthorpe. As noted in section 6.4.2, this incident occurred when a car left a road and crashed through the boundary fence onto the line. The driver was killed as his vehicle was struck by a 100mph passenger train. (None of the traincrew or 74 passengers were harmed.) The number of derailments from incursions is relatively low; indeed, only four incidents have occurred in the past 10 years. Table 13. Road rail incursions leading to derailment and fatalities Date Location Incursion site Train type Vehicle Fatalties 14/09/1999 Aberdeen Fence Passenger multiple unit Rail plant (JCB) 0 28/02/2001 Great Heck Bridge Loco hauled/push-pull passenger Car 10 28/02/2002 Metheringham Fence Passenger multiple unit Car 1 25/09/2006 Copmanthorpe No.2 LC Fence Passenger multiple unit Car Summary of safety performance risk for road-vehicle incursions The majority of incursions are the result of road traffic accidents, where a motorist has lost control over their vehicle and impinged on the permanent way. Chart 44 represents the number of fatalities and weighted injuries to those involved in railway incursions, including level crossings and those on board trains and road vehicle occupants. It does not include pedestrians or passengers on foot. The data also excludes fatalities at level crossings. The chart clearly indicates that the total number of fatalities due to incursions is fairly low. The exception is 2001, when the incident at Great Heck caused ten fatalities (including four members of staff), 38 major injuries and 36 minor injuries. Chart 44. Fatalities and harm due to vehicle incursions (excluding level crossings) Fatal Major Minor FWIs Fatalities and injuries Fatalties and weighted injuries to

52 7 Initiatives National Level Crossing Safety Group The National Level Crossing Safety Group (NLXSG) acts as a forum where the rail industry (principally RSSB, Network Rail and ORR) meets with representatives of groups that regularly use level crossings, and other parties interested in the safe operation of crossings, to raise the awareness of safety matters in all considerations of the interface. In particular, the group seeks to: Improve behaviour of pedestrians, motorists and other users at level crossings through communication and education programmes to ensure that road and railway safety risks are understood. Examine public policy and make recommendations to simplify and consolidate regulatory matters covering safety at level crossings, including road traffic and highway matters, planning guidelines for development and the effective prosecution of offenders in the interest of public safety. The NLXSG adopts the approach of the 4 Es, Enabling, Education, Engineering and Enforcement. The group will also evaluate the impact of any initiatives it puts in place for the improvement of safety at level crossings. The group acts as a stakeholder in relevant research projects carried out by RSSB. NLXSG is managed within the Community Safety section of RSSB s National Programmes department and is directed by a level crossing steering group. The group reports regularly to the Community Safety Steering Group (CSSG). Recent work The group consulted with DfT and the Driving Standards Agency (DSA) during the preparation of the recently published revised Highway Code (2007 edition). Coverage of level crossings in the code was improved with regards to; Clearance of OHL over crossings. Prevention of motorists reversing onto a crossing. Use of railway telephones when provided. The group has made a submission to the DSA consultation on Driver Certificate of Professional Competence, to reinforce the teaching of level crossing skills for lorry, bus and coach drivers. In future these drivers will need to gain a certificate of competence in line with European Directive 2003/59/EC. A working party of the group produced a report requested by the Secretary of State for Transport. The report (dated tember 2006) covered actions needed to be taken to reduce and manage level crossing risks more effectively. It concluded that there is a pressing need to modernise the approach taken in Britain to optimise the management arrangements for level crossings from both road and rail perspectives. It also states that dealing with safety risk and road/rail congestion should be an integral part of any future policy or road/level crossing safety legislation. 52

53 This report led the group to work with the CSS, formerly the County Surveyors Society, to set up Road-Rail Partnerships with Network Rail (more detail below). These meetings, managed by Network Rail, share information, safety assessments, consultation on future development planning and road maintenance. The group helped Network Rail lobby for the acceptance of amendments to the Road Safety Act, These amendments impose requirements on local traffic authorities and highway authorities for the provision, maintenance and operation of protective equipment at crossings on the public highway. The group consulted with DfT on post-court road safety interventions for convicted traffic offenders. It suggested that there should be a module which tackles level crossing safety and the effects of misbehaviour that endangers human life and the railway. This would serve to highlight the negative consequences for the driver, their family, friends, other road users and train occupants. This would form part of the examination for traffic offenders who later want to replace their driving licence after a period of disqualification. The NLXSG is the means by which RSSB is fulfilling its duty to initiate, with ORR/DfT, a review of the Rail Safety Principles and Guidance on level crossings. The review is to produce crossing guidance, warning and access systems that embrace current human factors knowledge. This was a recommendation placed on RSSB and Network Rail by the Formal Inquiry into the accident at Elsenham in December 2005, when two young girls were fatally injured on the station footpath crossing. The ORR Board will be asked to proceed with the review after receiving estimates of the cost and time for the work. Contact: Alan Davies alan.davies@rssb.co.uk Network Rail initiatives There is currently considerable research and practical action being taken to raise awareness and change user behaviour around level crossings to reduce misuse and fatalities. Network Rail s advertising initiative, Don t run the risk This campaign, now in its second year, has used a combination of television radio, press, leaflet and computer media to impress on target audiences the danger of running red lights or disobeying traffic signs at level crossings. The hard-hitting advertisements are an element of the most extensive public safety campaign ever run by the rail industry. The message is simple level crossings are safe if used correctly but if you run the risk, the cost can be fatal. More than 5.5 million has been spent so far on this sustained campaign, which aims to change the behaviour of level crossing users to protect themselves and reduce the danger to train passengers. 53

54 In 2007, the campaign has been specifically targeting the groups known to be at highest risk of crossing misuse: year olds and the over 65s. In addition to the regular advertisements, Network Rail has also produced an online viral music video, which aims to educate people on the dangers of misusing level crossings. The production, which stars glamour model Lucy Pinder, targets year-old males. The video s shocking ending is designed to resonate with the audience to bring home the dangers of running the risk at level crossings. The latest viral campaign may be found at The broadcast and online aspects are just part of a campaign specifically aimed at hot spot areas places where there are higher numbers of level crossing incidents. All media channels will be used including television, radio, cinema, online, outdoor posters, newspaper advertisements and direct marketing. Level crossing risk co-ordinators An interactive stand featuring a level crossing quiz is also touring the country to highlight the issues raised in the Don t run the risk campaign. The stand, which includes real level crossing warning lights, centres around What s the risk?, an interactive video game that tests the knowledge of both drivers and pedestrians on the correct use of level crossings. Targeting hot spot areas, the awareness days are attended by Level Crossing Risk Control Co-ordinators and other operational staff who can give help and advice to visitors. Bridge strike prevention Network Rail chairs the Bridge Strike Prevention Group. Other members of the group include DfT and road transport organizations. The group develops and promotes measures to prevent and reduce the number of bridge strikes. It has produced good practice guides for use by professional drivers and transport managers to help raise awareness of the risks and consequences of bridge strikes, and to give guidance on how they can be prevented. The All Level Crossing Risk Model Launched in January 2007, the All Level Crossing Risk Model (ALCRM) was developed as a way of understanding the relative risk presented by each of the 7,000 crossings on the network. The model provides a standard database for collecting, storing and recalling details on the features of each individual interface. This enables the selection of the best risk control solutions, with a view to ensuring risk is reduced so far as is reasonably practicable. 54

55 The ALCRM was jointly developed between RSSB and Network Rail. RSSB took the lead role for the research part of the project, devising the complex and extensive algorithms/formulae which form the model s core workings; Network Rail shared its operating experience and knowledge of level crossings to ensure the ALCRM closely reflected reality. Network Rail took the lead role in translating the algorithms into a software-based tool on a central Oracle database system; it also provided the training to its own Area-based teams, who are now responsible for data collection, input to the model and consideration of the results produced. The ALCRM recently won the Institute of Engineering Technology (IET) 2007 Award for the Advancement of Railway System Safety. The target dates populating the model are end-december 2007 for all public road crossings and January 2010 for all other crossing types. To date (tember 2007), over 2,700 crossings have been entered and it is already showing benefits. For example, the model has eased the process of identifying those crossings which possess greater risk levels, such that focused consideration may be given to providing appropriate mitigation where reasonably practicable. The level crossing risk management toolkit The level crossing risk management toolkit was launched in tember 2006, with an update coming in June It is used by the industry to supplement level crossing risk assessments by providing detailed information about the human factors issues underlying crossing risk, and suggesting appropriate risk mitigation measures. For further details, see Road rail partnerships The traditional approach to level crossing risk sees it as a railway problem with little involvement from road authorities except where difficulties have occurred. Over the last four years representatives of the Driving Standards Agency and the CSS have become active members of NLXSG and great progress has been made with amendments to the Highway Code and driving test procedures. The latest initiative is to set up local road / rail partnerships where all those involved in managing risk at level crossings in, typically, a county area, can work together to identify and prioritise activities, based on the three Es of engineering, enforcement and education. This has become a good example of the fourth E engagement and will be rolled out over the country after a trial in Lincolnshire, West Sussex, North Yorkshire and Dorset. Twenty partnerships are now in place and others being set up include one to cover the Highlands of Scotland. New planning processes in Network Rail In support of its Level Crossing Strategy, Network Rail has already centralised the design process for level crossing replacements and enhancements and is drawing up a national programme, based on the expected dates when crossing equipment at each location will become due for renewal. The next stage is the setting up of a Level Crossing National Specialist Team which, among other activities, will create a single plan for all these crossings in plenty of time to identify which ones could be closed, going through the planning process in sufficient time to allow this to happen. The new organisation went live in October The options for each crossing will be: Closure 55

56 Enhancement Replacement on a like-for-like basis SELCAT Project SELCAT (Safer European Level Crossing Appraisal and Technology) is an action from the European Commission's 6th Framework Programme involving 19 partners from nine European countries and Japan, Russia, China, India and Morocco. This project aims to collect, structure, analyse and disseminate existing world-wide research results and to stimulate new knowledge exchange in the area of level crossing safety. The research, having conducted a literature review, is currently evaluating existing and new technologies, and risk modelling techniques in use or available to help manage risk at level crossings in Europe and for partner nations. RSSB are leading Work Package 3 which will deliver a framework for a generic level crossing risk model. This model will provide each member state with the tools and techniques to better understand their risk profile. This information can then be used to ensure that investment is effectively targeted to reduce risk. Contact: Jackie Barrett jackie.barrett@rssb.co.uk RSSB Research and Development RSSB manages a programme of research and development (R&D) on behalf of government and the railway industry. The programme is funded by the Department for Transport and aims to assist the industry and its stakeholders in achieving key objectives: Improving performance in terms of health and safety, reliability, and punctuality. Increasing capacity and availability. Reducing cost. Integrating all of these to compete effectively with other transport modes (or complement them as appropriate) and deliver a sustainable future for the railway. The RSSB R&D programme focuses on industry-wide and strategic research that no individual company or sector of the industry can address on its own. It therefore includes research covering systems issues across the whole railway, and the engineering interfaces within the railway, as well as the interfaces with other parts of the community. The programme is also instrumental in supporting the development of a future vision for the railways and assessing how that vision can best be delivered. The level crossing topic covers R&D into the causes of misuse, the effectiveness of existing and new technologies in reducing risk, and ways of improving the management of crossings, including improved risk profiling. The scope of this topic includes both footpath and station crossings, as well as road vehicle crossings. It includes technical, human factors and economic issues that affect the railway (and roads) at level crossings. To date the research programme has not investigated vehicle incursions at other than level crossings, but if stakeholders identified the need for such research it would be taken forward in the usual way. 56

57 The NLXSG is the principal client for level crossing research, reporting to the Community Safety Strategy Group; it also reviews level crossing legislation and encourages closer relations between the rail industry and local authority planning and highway departments and crossing user organisations. As the asset owner and duty holder for level crossings on the national rail network in Great Britain, Network Rail is the principal stakeholder in NLXSG and in the research that informs and supports Network Rail s Level Crossing Strategy. Research in this area is underpinned by the Level Crossing Topic Research Plan a copy of which is attached as an appendix to this report. It is designed to be a living document and will be updated on a regular basis. The areas of research required to address this topic are presented below: Area 1 - Understanding the risk at level crossings to enable prioritisation of remedial actions In order to understand the risks associated with automatic level crossings, a risk model was developed in This work was followed by the development of a risk model for all types of level crossings, which has now being implemented. (See separate note above). The work carried out to date has helped industry s understanding of the causes of level crossing risk; particularly that it is largely the result of undesirable public behaviour rather than failures on the part of the railway. Based on this knowledge, it is important that industry gain a more detailed understanding of human behaviour at level crossings, for example the way pedestrians perceive risk at crossings and the relationship between risk perception and actual behaviour. Initial work in this area is expected to be completed during To date there has been little systematic collection and collation of usage and near-miss data. This is mainly because crossings are remote from supervision; and because the use of CCTV data, while getting cheaper, is still costly and requires considerable analysis. This gap is unhelpful in understanding changes to vehicle user and pedestrian risk particularly why the former seems to have reduced and the latter remained unchanged over time. Research is required to address this gap in data. Research is required to understand the split between user error and violations, in particular to what extent do users choose to disobey or ignore the rules, and to what extent are they unaware of or mistaken about them? Countering this problem requires action by the industry and by the highways sector in understanding the problem, identifying research needs and providing possible solutions on several fronts. This work is being driven by NLXSG and specific issues which require research are being identified on a rolling basis. Research in this area will look at the societal and attitudinal changes to risk taking and to authority in general, and how they may lead to positive or negative effects in terms of safety at level crossings. Different social groups may behave, and change their patterns of behaviour in different ways. Practical application of knowledge in this area would include, for instance, improving the design of safety communications. Research is also required to evaluate the impact of repeat accidents at a level crossing (even if causes were outside rail's control) and whether this impact should be taken into consideration when risk-rating a crossing. The industry require a tool to decide where to target efforts, especially where crossings are subject to repeat offenders and where children are involved. 57

58 Area 2 Identifying & sharing good practice in Britain and overseas to facilitate the adoption of appropriate solutions Research into good international practice on level crossing issues could help deliver safety improvements, which are reasonably practicable and fundable within current industry arrangements. Current research is identifying good practice in the application of red light cameras, car number recognition, CCTV and other systems and fixed penalty regimes, as deterrent and enforcement mechanisms at level crossings. Other good practice, which requires investigation, is the experience gained within Canada regarding second train coming and the disability research carried out in Australia regarding the problems associated with flange way gaps. In addition, an understanding of what is considered good practice in the education of the risks associated with level crossings is required. Area 3 - Identifying new technical and operational solutions to prevent errors and misuse of the crossings RSSB recently examined the options of using obstacle detection systems, based on radar technology, to improve safety at level crossings. Obstacle detectors at level crossings are already used (in Germany and Italy and are on trial elsewhere) to detect obstructions capable of causing significant damage to a train, or to assist the signaller in charge of a closed-circuit television (CCTV)-controlled crossing. In order to avoid unacceptably high levels of safe-side (false) trips. However, such a system has to be sensitive enough to distinguish a significant threat to a train (such as a car), from an insignificant one (like a shopping basket or a fox) in order to avoid unacceptably high levels of safe-side (false trips). The opportunities and challenges posed by these detectors need to be fully understood if they are to be used successfully in a cost-effective way in Britain. The project investigated the technicalities, logistics, cost and safety benefits of the most promising potential solutions for AHBs and for CCTV crossings, and made several recommendations, which are being assessed for implementation by Network Rail. The illegal and dangerous activity of vehicles violating red lights by zigzagging across automatic half-barrier level crossings has contributed to at least four deaths between 1994 and 2005 and could be the precursor to a major train accident. Current research aims to assist the industry by investigating possible solutions, which may prevent or deter violator and error behaviour by road users and pedestrians at level crossings, eg by assessing the merits of: Installing median strips (a form of physical barrier, separating the adjacent lanes) on the road approaches to the crossings. Evaluating alternative technological remedies, such as signage, cameras, road markings, vehicle-activated signs and traffic calming to understand what works best. Due to complaints about noise pollution generated by driver-operated train horns, a need was identified to examine the use of train horns and alternative technologies and operational processes. This work has now been implemented. At user worked crossings, users are currently required to traverse the crossing four times on foot to open and close gates. Research is under way to examine technical solutions to risk reduction for these types of crossings. It is examining the merits of employing different types 58

59 of gates or raising barriers, including manually pumped or electrically driven options and automatically lowering or user lowered options, and the potential use of miniature warning lights. By adopting a cost benefit approach the research will the help focus further work by the industry on the most worthwhile options. Further work on signage at, and lights used to warn users on the approach to crossings is in development. In particular, the provision of miniature red and green warning lights at crossings, the installation of which generally predated the application of human factors science to railway users, is to be examined on a systematic basis. Area 4 - Understanding the costs of level crossings and the benefits of adopting alternatives to optimise societal benefits Research in this area has reviewed how the costs associated with maintaining, operating, upgrading, and renewing level crossings may be reduced. This research is informing future developments of new crossing types and look at the benefits of closures and conversions. Earlier reports have concluded that in certain circumstances, there may be a case for replacing a crossing with a bridge over (or exceptionally, under) the railway. A comprehensive model to aggregate the various whole-life costs of a public road crossing should allow the industry to decide, using cost benefit analysis, whether it is economic in the long term to replace any particular crossing by a bridge or by diverting traffic to other routes. This model would assist in the development of business cases for crossing closures or conversions to bridges. It is currently on trial in several local authority areas. There is a need for wider evaluation of the economics of level crossings, rather than just the costs, and on a societal level instead of the railway industry s viewpoint alone. Research in this area should therefore not only focus on the immediate costs and benefits associated with closing a particular crossing but also the benefits for the whole rail route in terms of increased line speed and line availability. It should look at the potential benefits within local authority geographical areas, by evaluating the potential impact to road travellers and the local community in terms of value of time associated with delays at level crossings, and the environmental impact of waiting motor vehicles at crossings both during normal usage and for special one-off or seasonal events. Similar research is being carried out in the USA under the North Carolina Sealed Corridor Project. Area 5 - Working in collaboration with highway and planning authorities to design out safety risk and reduce the overall cost to society Research could contribute to the development of better relationships with external agencies representing public authorities and stakeholder groups. In particular, a better appreciation of highway issues and working together with planning authorities may help address the overall residual safety risk, design risk out of new schemes at source, reduce road congestion / delay issues and save money through integrated planning. Issues that have already been identified as requiring research support under this area include: The use of level crossing signals on unrestricted roads. The timing of amber lights and whether it is sufficient. The impact on the road network with the introduction of extra and faster trains. The impact of smart re-routeing instead of bridges on the road network. The potential use of cheap bridge crossings and footpaths at level crossings. 59

60 The potential benefits of this research are that level crossing safety becomes an integral part of government transport and land use planning processes, and there is a more system authority approach to decision-making on issues such as bridge substitution and enforcement strategies. Area 6 - Understanding the needs of vulnerable users at level crossings to facilitate social inclusion Research is needed to inform future developments of new crossing types (or modifications to existing designs) to take account of the need of vulnerable users, particularly pedestrians and other non-motorised users at level crossings. A start has been made via NLXSG s disability working party, which is expected to help identify areas of concern that need to be better understood. The first two research ideas looking at the risk to physically impaired and hearing-impaired people at level crossings are being developed into remits for discussion. This research should also look at how to reduce the impact of any potential vulnerable-user solutions on 'able-bodied' users. Area 7 - Review and overhaul of the legislative framework for level crossings to identify legal requirements & consolidate disparate regulations Members of NLXSG have identified the need for a review and overhaul of the legislative framework for level crossings, possibly consolidating disparate laws/regulations into a single act covering road, rail and planning issues. If, as it appears likely, Government agrees to such an initiative going forward, research could also assist in defining the needs and areas for change, and help secure a legislative, consultative and standards framework that engages all parties and facilitates adequate risk control by duty holders. Area 8 Research to support inquiry recommendations, government and regulatory policies, proposed and new legislation From time to time, research is required to ensure recommendations relating to the management of relevant risks from investigations of recent major incidents, from Rail Accident Investigation Branch (RAIB) investigations as well as other HMRI and industry investigations, are satisfactorily addressed. Recent and current initiatives have led to several research projects resulting from the events at Ufton in 2004 and Elsenham in The impact on the railways from changes to road traffic acts and the increase in the use of congestion charging may also need to be investigated in the future. Contact: Michael Woods michael.woods@rssb.co.uk 60

61 Appendix 1. Level crossing research and development Project No. T000 Title User worked and footpath level crossing research Description Understanding risk relating to user-worked and footpath level crossings by surveying 300 crossings, interviewing users and analysing accident data. Proposing improved risk control measures, incident reporting and data collection. Published Status / / Target Contractor 15/08/2001 Published T028 Development of a universal level crossing risk tool Developing a risk tool to encompass all level crossings and locating the tool in a web browser environment in line with Network Rail Information Management strategy. 16/07/2007 Published T032 Trials of median strips / lane separators at level crossings Determining whether the provision of road median strips at the approaches to automatic half barrier crossings would assist in the prevention of accidents caused by drivers violating the crossing controls. T105 Wayside horns at level crossings Establishing the best way to deliver audible warnings to crossing users should noise bans make it impossible to rely on horns operated by train drivers. 25/07/2007 Published 04/05/2005 Published T232 Improving level crossing information systems Improving Network Rail's national level crossing information database - ALRMS. Assessing user and stakeholder requirements, the strengths, weaknesses, and opportunities to upgrade the system, to manage level crossing risk better. 23/02/2006 Published T269 T332 T333 T336 Human factors risk at user worked crossings Understanding the risk at station and barrow crossings Evaluating best practice deterrence and enforcement mechanisms at level crossings Modelling the economics of level crossing closures and conversions Gaining a better understanding of human behaviour at user worked crossings and identifying contributory factors to risk at the crossings. Evaluating potential risk reduction measures. Examining the requirements for the use of station and barrow crossings to traverse the track. Assessing current safety controls, and the scope of procedural or technical innovation to reduce risk. Understanding the cost-effectiveness of, and issues surrounding, red light cameras and investigating alternative deterrent and enforcement strategies used in Great Britain and overseas. This project has developed an economic model to give an initial, robust assessment of the business case for converting or closing a public road level crossing. 21/06/2004 Published 25/11/2005 Published 18/05/2007 Published 11/06/2007 Published T364 T521 The cost of level crossings - an international benchmarking exercise Developing enhanced consequence algorithms for level crossing risk models This research benchmarked the costs of upgrading level crossings in Britain and overseas. It aimed to discover how to achieve increased safety at crossings for lower cost. Developing enhancements to the automatic level crossing risk model and the specification of the all crossing risk model. 25/09/2006 Published 23/02/2006 Published 61

62 (cont.) Appendix 1. Level crossing research and development Project No. Title Description T522 Obstacle detection at level crossings Understanding the potential opportunities and challenges posed by installing obstacle detectors at level crossings in Great Britain by examining the technicalities, logistics, cost and safety benefits of potential solutions. Published Status / / Target Contractor 13/11/2006 Published T524 Use by other railways of risk models and risk assessments for level crossings Assessing the benefits of various risk models and risk assessments for level crossings used by railway administrations in Great Britain, Ireland and overseas. Many countries have risk models for the level crossings on their railways. This project compared some of these models with the All Level Crossing Risk Model (ALCRM) and assessed their benefits to see if any could be applied to the ALCRM. 18/05/2007 Published T527 Analysis of research ideas from recent international level crossing conferences The research captured research ideas raised at recent level crossing conferences and ranked them in order of priority. 25/06/2007 Published T528 Attitudes to, processes and funding for, crossing closures in other countries The project investigated the difference between national approaches to closing level crossings, and identified lessons that may be learned that would make the closure of crossings in Great Britain less complicated. 25/06/2007 Published T561 Evaluating safety benefits from miniature warning lights at level crossings This project aimed to evaluate the safety benefit of installing miniature warning lights at user worked crossings where no warning light system was previously installed. 18/06/2007 Published T668 Research into the safety benefits provided by train horns at level crossings To evaluate the risk at whistle board crossings and determine what level of mitigation is reasonably practicable to implement. 31/12/2006 Published T680 T681 Mapping the extent of the train horn noise problem Understanding the problems that train horn noise causes to neighbours This project was designed to understand the geographical extent of properties and residents affected by the sounding of train horns at whistle boards protecting level crossings. This project was designed to gain an understanding of the effect train horns have on residents who live in the vicinity of whistle boards that are sited to protect pedestrians on level crossings. 30/11/2006 Published 30/11/2006 Published 62

63 (cont.) Appendix 1. Level crossing research and development Project No. T334 Title Reducing the risk to motorists traversing user worked crossings on foot Description Examining technical solutions to risk reduction at user worked crossings where users are currently required to traverse the crossing four times on foot to open and close gates. Published Status / / Target Contractor 2008 Halcrow T335 Improving road user and pedestrian behaviour at level crossings Understanding offender, violator and error behaviour by road users and pedestrians at level crossings. Evaluating alternative technological remedies, such as signage, cameras, road markings and traffic calming Human Engineering T652 Examining the benefits of 'another train coming' warnings at level crossings Examining the benefits that special warnings that 'Another train is coming' can give to users of level crossings and how these warnings may best be delivered to the user AD Little T653 T707 Safer European level crossing assessments and technology (SELCAT) Realising the potential of vehicle activated signs at approaches to public road level crossings To provide an overview of existing and planned level crossing research and a standard method of reporting level crossing accidents in European countries. Vehicle activated signs (VASs) are widely used within the British road network. The project seeks to establish the efficacy of VASs in preventing blocking back of road traffic over automatic half barrier (AHB) level crossings Consortium / RSSB in house 2009 Human Engineering T719 Monitoring motorists' behaviour at level crossing median strip trial sites This project is collecting and analysing CCTV recordings of driver behaviour at three level crossings, where road median strips, also known as lane separators, are to be installed, to determine whether median strips prevent accidents WSP Civils T737 T738 Documenting the All Level Crossing Risk Model Trialling the national roll out of the level crossing cost model (AXIAT) using road rail partnerships Supplying a history of the All Level Crossing Risk Model (ALCRM) and a detailed referenced explanation of its construction and functionality and the important parameters used in its formulation. Trialling the national roll out of the level crossing cost model (AXIAT) using road rail partnerships A D Little 2008 Halcrow 63

64 (cont.) Appendix 1. Level crossing research and development Project No. T650 Title A review of risks affecting disabled pedestrians at level crossings Description Indenting the risk issues facing pedestrians with disabilities at level crossings and assessing what can be done to reduce them. Published Status / / Target Contractor 2009 In development T654 Specific risks faced by hearing impaired pedestrians at level crossings Evaluating the needs of hearing impaired level crossing users at level crossings in terms of warning systems and messages In development T706 Developing a code of practice for the design of road approaches to level crossings Level crossings are a unique type of intersection between road and rail. The project aims to produce guidance so that road engineers can design the road approaches to such intersections to the latest standards and to conform with good practice In development T729 Determining whether it is 'ALARP' to restrict the size of objects to be detected by obstacle detectors at level crossings Following the work done by AD Little on T522 on obstacle detection, Network Rail is interested in determining the minimum size of object for which a cost effective case can be made for detection at level crossings In development T730 Understanding human factors and developing risk reduction solutions for pedestrian crossings at railway stations Undertaking a comprehensive review of existing guidance related to the design of station pedestrian crossings as recommended by the Rail Accident Investigation Branch s inquiry into the fatal accident at Elsenham In development T756 Research Into traffic signs and signals at level crossings Evaluating the effectiveness and comprehensibility of traffic signs and signals on the approach to level crossings In development 64

65 Appendix 2. Key safety facts Annual totals Jan - t Fatalities and injuries (Level crossings) (Fatalities) (8) (14) (15) (13) (5) (3) (10) (Major injuries) (5) (13) (16) (9) (6) (6) (2) (Minor injuries) (37) (33) (72) (38) (40) (28) (19) (Fences) (Fatalities) (1) (0) (0) (0) (1) (0) (0) (Major injuries) (1) (0) (0) (0) (0) (0) (0) (Minor injuries) (0) (0) (0) (0) (0) (0) (0) (Bridges) (Fatalities) (0) (2) (0) (0) (0) (0) (0) (Major injuries) (0) (1) (0) (0) (0) (0) (0) (Minor injuries) (0) (0) (0) (0) (0) (0) (0) (Access points) (Fatalities) (0) (0) (0) (0) (0) (0) (0) (Major injuries) (0) (0) (0) (0) (0) (0) (0) (Minor injuries) (0) (0) (0) (0) (1) (1) (0) Level crossing misuse and nearmisses Pedestrian crosses when unsafe Pedestrian level crossing near misses Road vehicle crosses when unsafe Road vehicle level crossing near misses Road vehicle incursions 69 (82) (83) (65) (Level crossings) (7) (9) (13) (5) (14) (13) (15) (Not foul of the line) (0) (0) (1) (1) (0) (0) (2) (Foul of the line) (7) (9) (12) (4) (14) (13) (13) (Fences) (46) (42) (49) (42) (51) (42) (28) (Not foul of the line) (24) (29) (28) (25) (34) (31) (15) (Foul of the line) (22) (13) (21) (17) (17) (11) (13) (Bridges) (4) (4) (1) (1) (2) (2) (2) (Not foul of the line) (2) (1) (1) (1) (0) (0) (1) (Foul of the line) (2) (3) (0) (0) (2) (2) (1) (Access points) (12) (27) (20) (17) (13) (12) (15) (Not foul of the line) (3) (11) (7) (8) (9) (8) (7) (Foul of the line) (9) (16) (13) (9) (4) (4) (8) Collisions with road vehicles (Level crossings) (16) (24) (16) (16) (11) (10) (9) (Fences) (1) (1) (2) (4) (3) (2) (3) (Bridges) (1) (1) (0) (0) (0) (0) (0) (Access points) (2) (0) (3) (1) (1) (1) (2) Derailments (Level crossings) (2) (2) (3) (0) (0) (0) (0) (Fences) (1) (0) (0) (0) (1) (1) (0) (Bridges) (0) (0) (0) (0) (0) (0) (0) (Access points) (0) (0) (0) (0) (0) (0) (0) 65

66 Appendix 3. Crossing types ACTIVE CROSSINGS Manual crossings and opened when required if no train is approaching. Manually controlled gate (MCG) This crossing is equipped with gates, which are manually operated by a signaller or crossing keeper either before the protecting signal can be cleared, or with the permission of the signaller or signalling system. At the majority of these crossings, the normal position of the gates is open to road traffic, but on some quiet roads the gates are maintained closed to the road Manually controlled barrier (MCB) MCB crossings are equipped with full barriers, which extend across the whole width of the roadway, and are operated by a signaller or crossing keeper before the protecting signal can be cleared. Road traffic signals and audible warnings for pedestrians are interlocked into the signalling system. Manually controlled barrier protected by closed circuit television (MCB-CCTV) Similar to MCB crossings, except that a closed circuit television (CCTV) is used to monitor and control the crossing from a remote location. 66

67 Automatic crossings Automatic half-barrier (AHB) AHB crossings are equipped with barriers that only extend across the nearside of the road (so that the exit is left clear if the crossing commences operation when a vehicle is on it). Road traffic signals and audible warnings are activated a set time before the operation of the barriers, which are activated automatically by approaching trains. The barriers rise automatically when the train has passed, unless another train is approaching. Telephones are provided for the public to contact the signaller in case of an emergency or, for example, to ensure it is safe to cross in a long or slow vehicle. These crossings can only be installed where the permissible speed of trains does not exceed 100mph. Automatic barrier locally monitored (ABCL) As far as the road user is concerned, this crossing looks identical to an AHB crossing. The difference is that train drivers must ensure that the crossing is clear before passing over it. Train speed is limited to 55mph or less. Automatic open crossing remotely monitored (AOCR) The AOCR is equipped with road traffic signals and audible warnings only: there are no barriers. It is operated automatically by approaching trains. Telephones are provided for the public to contact the signaller in an emergency. Only one crossing of this type remains on NRMI, at Rosarie in the Scottish Highlands. 67

68 Automatic open crossing locally monitored (AOCL) Like the AOCR, this crossing is equipped with road traffic signals and audible warnings only and is operated automatically by approaching trains. The only difference is that no telephone is provided for crossing users: train drivers must ensure that the crossing is clear before passing over it and train speed is limited to 55mph or less. If a second train is approaching, the lights continue to flash after the passage of the first train, an additional signal lights up, and the tone of the audible warning changes. User-worked crossing with miniature warning lights (UWC-MWL) This crossing have gates or full lifting barriers, which the user must operate prior to crossing. Red/green miniature warning lights, operated by the approach of trains, inform the user whether it is safe to cross. PASSIVE CROSSINGS User worked crossing (UWC) This crossings has gates or, occasionally, full lifting barriers, which the user must operate prior to crossing. The user is responsible for ensuring that it is safe to cross; hence there must be adequate visibility of approaching trains. Once clear, the user is required to close the gate or barriers. These crossings are often found in rural areas, for example providing access between a farm and fields. They often have an identified user, some of whom keep the crossing gates padlocked to prevent unauthorised access. 68

69 User-worked crossing with telephone (UWC-T) These are similar to the standard user worked crossing, but a telephone is provided. In some circumstances (for example when crossing with livestock or vehicles) the user must contact the signaller for permission to cross, and report back when they are clear of the track. They are provided where visibility of approaching trains is limited, or the user needs to move livestock over the railway on a regular basis. Open crossing (OC) At open crossings, which are sited when the road is quiet and train speeds are low, the interface between road and rail is completely open. Signs warn road users to give way to trains. Road users must therefore have an adequate view of approaching trains. The maximum permissible speed over the crossing is 10mph or the train is required to stop at a stop board before proceeding over. Footpath crossing These are designed primarily for pedestrians and usually include stiles or wicket gates to restrict access. The crossing user is responsible for making sure that it is safe to cross before doing so. In cases where sufficient sighting time is not available, the railway may provide a whistle board, instructing drivers to sound the horn to warn of their train s approach, or miniature warning lights. A variant is the bridleway crossing, which is usually on a public right of way, although some are private and restricted to authorised users. Some footpath crossings are in stations and these can be protected by a white light generally used by railway staff only (which extinguishes when a train is approaching). All these crossing types, some of which clearly have automatic protection, are analysed as a single group in this report because of concerns over the accuracy of crossing type data in SMIS. 69