Written evidence from Hawkins and Associates Ltd (SLC 037)

This submission has been prepared by R B Newbery of Hawkins and Associates Ltd (Hawkins) on behalf of Hawkins.  Hawkins is a forensic science and engineering consultancy, which, amongst numerous other types of incident, investigates rail accidents; our investigations have included the Ladbroke Grove, Hatfield and Potters Bar derailments (I gave evidence at the Inquest).  We also investigate road traffic accidents, which has led us through our work with motor insurers to investigate a number of accidents at level crossings.  We have also investigated accidents involving pedestrians at level crossings.  We have given evidence on such cases in civil and criminal courts. 

I became aware of this inquiry relatively late in the proceedings, but was told that a submission from Hawkins could still be considered.  As a result, before preparing this submission I have had the benefit of viewing those prepared by others, so this submission consists of a number of observations from our work which do not appear to have been covered by others and I have tried not to repeat observations already made by others:

1.              Many users of automatic half-barrier crossings (AHBC) are not aware of the intentionally short time between the amber light coming on and the train arriving (as little as 27 seconds). 

2.              The meanings of signs and lights at level crossings are not well understood.

3.              It is likely that few pedestrians have any experience of deciding if it is safe to cross ahead of a ‘vehicle’ travelling at 100 miles per hour, as permitted at AHBC and user worked crossings (UWC).  In the case of footpath or bridleway crossings, speeds of up to 125 mph can be permitted.

4.              The removal of all level crossings from the network in the short term would be prohibitively expensive, therefore education is possibly the most effective way of improving safety at level crossings.

1.              Automatic half-barrier crossings (AHBC)

1.1              Many of the accidents that we have investigated involving vehicles have been at AHBCs.  I expect the committee will be aware of the report of the public enquiry into the accident at Hixon level crossing in 19681 and the report by British Railway Board on level crossing protection published in 19782. 

1.2              The latter report states that the aim of the report was: “to consider ways in which methods of level crossing protection can be further developed in Great Britain, taking into account the cost and need to maintain an adequate and publically acceptable standard of safety - and to make recommendations.”  The general background section started “It was immediately evident to us that progress towards automated level crossings in Britain had come to a virtual standstill since the Hixon enquiry in 1968 whereas progress in Europe had been most impressive.”  Therefore “We considered that our first task should be to determine how and whether AHB should be simplified to make them cheaper to install and maintain without affecting their present excellent safety records.  We had made a number of recommendations which we believe should satisfy both aims”.

1.3              Recommendations from these documents have been carried through into standards which are current at the time of writing this submission, but these standards do not explain the rationales behind them.  In particular, with respect to AHB crossings, the report recommended “A basic time-cycle of 27 seconds for crossings up to 15 m long”.  This was a reduction in the length of time from how such crossings were set up previously; the concern was that longer time cycles would encourage road users to zig‑zag between the barriers.  It seems that it was anticipated that if it became common experience that trains would arrive soon after the barriers descended, then drivers will be less willing to take the risk.

1.4              The information that the time between the barriers descending and the train arriving at AHBCs is deliberately low does not appear to be readily available to members of the public - it is certainly not mentioned in the Highway Code.  Neither is it mentioned in “Using level crossings safely” 3, a document prepared by the Office of Rail Regulation (the copy I have is undated, but it is not recent).

1.5              HMRI’s Guidance on Level Crossings4 includes the following: “The time elapsed between the amber lights of the road traffic light signals starting to show and the train arriving at the crossing should not be less than 27 seconds.  The train should pass as soon after 27 seconds as possible.  At least 95% of trains should arrive within 75 seconds and 50% within 50 seconds, once the sequence of events to close the crossing to traffic has begun”).

1.6              The short time between the barriers descending and the train arriving on AHBC is certainly in contrast to the long waits that are often experienced at full barrier crossings, which are usually operated from a ꞋlocalꞋ signal box (local does not mean that all such crossings are readily viewed from the signal box - many are viewed via CCTV).  The lack of understanding of the short time between the barriers descending and the train arriving, coupled with a possible route across the crossing by avoiding the half barriers has led to some well publicised near misses, including a recent one involving a cyclist at the Waterbeach crossing close to our offices in Cambridge (I investigated a fatal accident at that crossing some years ago). 

1.7              In addition, if some trains do take a long time to arrive at AHBCs after the crossing sequence starts, as I found to be the case during an investigation into an accident in South Wales (at least one train took over three minutes to arrive and others over two minutes) then the ꞋmessageꞋ that trains arrived soon after the crossing sequence starts, will be diluted.

1.8              One of the reasons for the difference in time between trains arriving at crossings after the amber light has come on is that most AHBC are operated by an electro-mechanical treadle at a fixed distance from the crossing.  The distance is defined as the distance that a train travelling at ‘line speed’, the speed limit for the track will travel in 27 seconds.  It can be appreciated that a goods train travelling at say 40 mph will take much longer to travel the same distance (about 68 seconds).  If there is a station close to the crossing at which some trains stop and some do not, the treadle needs to be set for the trains that do not stop.

1.9              There are two short stretches of track, one in Norfolk (Norwich to Cromer) and one which crosses between Bedfordshire and Buckinghamshire (Bedford to Bletchley), where a different system is used to operate the AHBCs, which takes into account the speed of an approaching train and delays the start of the crossing sequence accordingly.  I do not know why such a system has not been used elsewhere.

2.              The meaning of signs and lights at crossings are not well understood

2.1              The approaches to level crossings where a road crosses the track are required to include a multiplicity of signs.  One in particular, which advises the drivers of long or slow vehicles to stop and call the signal box, contains more information than it is possible for a driver to read whilst moving.  Few drivers are likely to stop and read it.  I am aware of three occasions where the contents of the notice were generally ꞋunderstoodꞋ, but were misinterpreted.  The signs say “Drivers of large or slow vehicles must phone and get permission to cross”.  The definitions of large and slow are given in smaller print further down the notice and ꞋslowꞋ is defined at 5 mph or less.  In all three cases, the drivers of the vehicles did not think that ꞋslowꞋ applied to them, because their vehicles were capable of travelling at more than 5 mph, although they had probably had to slow to about 5 mph or less because of local circumstances.

2.2              Another aspect of level crossings is there is no positive ꞋgoꞋ signal.  The absence of a signal effectively means ꞋgoꞋ.  Therefore a driver approaching a crossing and seeing no signal is likely to consider it safe to cross.  If a driver did not notice the signal because of bright sunlight he or she might continue across the crossing.  I consider that this was probably a factor in two of my investigations.  Network Rail is aware of the potential problem of a low sun shining directly onto the lights, i.e. from behind the driver, thus making the light less obvious.  Where this is likely to be a problem long hoods are put over the lights, but it is not clear to me that the effect on visibility when the sun is directly in front of the driver is so well understood, although it is referred to by RAIB in its submission.  A driver approaching traffic lights under such circumstances would know that a light of some colour would be displayed, so will probably check carefully to see what colour light was showing before proceeding.  In the case of level crossings, if no light is perceived that could be taken to mean a ‘go’ signal and the driver might then proceed onto the crossing. 

2.3              I have read a report where a driver apparently thought that the flashing red signal meant it was okay to proceed with caution!

2.4              I do not know how expensive it would be to include green lights in level crossing signals.  The changes to the electrical circuit that would be required would be very simple.

3.              Assessing train speed

3.1              Few people regularly cross roads where the speed limit is more than 40 mph and many people’s experience is limited to crossing roads where the speed limit is 30 mph.  Thus, many pedestrians who cross at AHBC, user worked crossings (UWC), or footpath or bridleway crossings, are poor at assessing the speed of an oncoming train.  Trains are permitted to pass over such crossings at up to 100 mph. 

3.2              The difficulty of assessing a speed of oncoming vehicles when they are travelling straight towards an observer are described in the text book “Forensic aspects of driver perception and response”, 3rd Edition by Paul L Olson, Robert Dewar and Eugene Farber, Lawyers and Judges Publishing Company Inc 2010.  This subject is dealt with at some length but includes the following observations: “When the path of the other vehicle is close to being directly towards or away from the observer, the primary closing speed cue, and perhaps the sole cue, is rate of change of image size.  That is, if the object seems to be getting larger fast, that indicates a high closing speed.  The difficulty here is that this cue to closing speed depends not only on closing speed, but also on separation distance.  At large separation distances, the apparent size changes slowly regardless of closing speed.”   To paraphrase an example given in the text:  If a train is observed when 1,000 feet (305 metres) away, by the time it has reached 500 feet (152 metres) away the image size of the train will have doubled.  When the train has covered half the distance to the observer so that is 250 feet (76 metres) away its image size will have doubled again, but in half the time it took to double in size before.

3.3              The size of an approaching vehicle also affects the perception of its speed: “One phenomenon at work is the “large object illusion” - the perception that large objects are moving more slowly than small ones traveling at the same speed.  Most of us have experienced the impression that a jumbo jet approaching an airport seems to be moving slowly, while a very small aircraft moving at the same speed appears to be traveling quite fast.  (Trains are large by comparison with road vehicles.)  In addition, there is virtually no lateral motion (a good cue to speed) in the perception of an approaching train when the driver (or pedestrian) is close to the tracks.  Drivers often fail to take the speed of trains into account, relying primarily on an estimate of their distance.”

3.4              There is also an effect of age on an observer’s ability to assess the speed of oncoming vehicles as described in a paper by Hoffmann published in the Journal of Perception in 19946. 

3.5              The situation can be even worse at night.  Trains often have only a single light on the front, equivalent to a dipped headlight on a car.  Assessing the speed of an approaching train if there is no ambient lighting then becomes almost impossible.

3.6              In summary, many pedestrians will not appreciate how quickly a train will arrive at their position so will attempt to cross the track when it is not safe to do so. 

4.              Conclusion

4.1              I understand that no new level crossings are being built and Network Rail is taking the opportunity wherever possible to remove crossings from the network, but to remove them all in the short-term would be prohibitively expensive. 

4.2              I find it surprising that the Highway Code has so little information about level crossings, and even more surprising that a document prepared by the Office of Rail Regulations does not mention the deliberately short time between the barriers coming down and the train arriving at AHBC.  Is the paucity of information in the Highway Code a deliberate attempt to try to keep the Highway Code short to make it easier to read?  I do not recall receiving any ꞋtrainingꞋ on level crossings when learning to drive or even when preparing for and taking my advanced training test.  It has to be acknowledged that few people read the Highway Code after passing their test and the driving test is insufficient to prepare people for ꞋrealꞋ driving, but that is another subject.

4.3              I recall Network Rail carrying out a publicity campaign about safety on level crossings fairly recently and I suspect that attempts to educate drivers through campaigns such as that are the most sensible way to proceed.  To be effective, any such publicity campaign needs to send the appropriate messages and based on my investigation of level crossing accidents these messages could perhaps include:

(a)              Trains arrive quickly at AHBC,

(b)              You are unlikely to be able to judge the speed of an oncoming train,

(c)              There is no ꞋgoꞋ signal at level crossings - if you cannot see the lights clearly, look again.

Documents referred to:

1.              The report of the public enquiry into the accident at Hixon level crossing on January 6th 19681, presented to Parliament by the Ministry of Transport by command of Her Majesty July 1968 (HMSO)

2.              The report by Officers of the Department of Transport of the British Railway Board on level crossing protection published by HMSO in 1978.

3.              Using Level Crossings Safely, Office of Rail Regulation, undated

4.              Railway Safety Principles and Guidance Part 2 Section E Guidance on Level Crossings HSE Books 1996, prepared by Her Majesty’s Railway Inspectorate.

5.              Forensic Aspects of Driver Perception and Response 3rd Edition by Paul L Olson Robert Dewar and Eugene Farber, Lawyers and Judges Publishing Company Inc 2010

6.              “Estimation of time for vehicle arrival - effects of age on use of available visual information” by Errol R Hoffmann published in the Journal Perception 1994 Volume 23 pages 947 to 955.

October 2013