Obviously the driver can pass information on verbally but how does he get to coach 12 on a packed train? Even if the comms are: Hang on we are going to evacuate the train we have to be able to pass the message to the passengers.
Interesting you mention that. What you say to the passengers regarding an evacuation can be quite risky. The advice given to me during my train evacuation training was that you do not announce simultaneously to all passengers that the train is to be evacuated - if it's a controlled evacuation. (Obviously there are very rare scenarios when you need everyone to get out via emergency egress at the nearest door, and you also somehow have time and a power supply to use the PA to instruct them to.)
There were several reasons given for this. In no particular order:
- it can make people impatient ("I want to go first!") and if they don't get their way, meaning they self-evacuate but not safely
- people can mis-hear or misunderstand words or phrases with words such as "evacuate", and then panic, causing an unnecessary stampede for random doors
- people tend to worry and have lots of questions even if they hear the full message, so it's best to help small groups face-to-face.
The last bit is quite pertinent but very hard on busy peak time services. Getting a smooth and relatively quick evacuation in poor conditions is challenging enough, but you also need to work out what to say in the mean time. It is often left up to the train crew to work out what to announce between the initial incident and any evacuation, which can vary hugely depending on other tasks such as fault finding or endless phone calls to Control, environmental factors, tiredness/frustration/morale, and their creativity to put a helpful spin on the situation.
Therefore you need to sort out a way to explain that all passengers will be escorted from the train, but you are working through the best option to help so many people. You can actually be explaining the situation to half the train whilst the other half is also some way through evacuation; to those leaving the service, they won't particularly care what's going on over the PA, but will be more worried about how the bloke in hi-vis is going to help you get your three-ton suitcase off the train. (Which they will now manage, by the way - in a controlled evacuation, you usually help passengers leave with their luggage unless it's hopelessly impractical, as it helps manage security and quite rightly prevents people fretting and panicking.)
The curious thing is that in a very crowded and rapidly deteriorating situation such as the one we're all talking about at Lewisham, the evacuation needs to be controlled but it is blatantly obvious that timing needs to be as swift as possible. The scenario was rapidly changing from needing a controlled evacuation to passengers performing an emergency evacuation. Whether this is trespass is, shall we say, unclear.
In terms of definitions, the rule of thumb I have always used for evacuations is this - if the train is immediately uninhabitable, such as through severe fire or smoke, then the crew need to stop any other train movements which may be affected, get the power switched off and perform an emergency evacuation to the nearest safe place. So this is an emergency evacuation.
If the train is
not immediately uninhabitable, but the most (or only) practical way for people to continue their journey is to leave the train away from a station, you need to perform a controlled evacuation. Theoretically this can actually be done with just the train crew, but this would be in dire circumstances indeed. Normally you agree this with TOC Control, Network Rail and the signaller, and gather together as many resources as possible to assist. You then arrange for a safe perimeter by means of line blocks, get the power switched off or preferably properly isolated, and use staff resources to shepherd passengers from a limited number of controlled exits.
The Lewisham incident was something of a hybrid of these, in terms of what needed to be done and what the passengers evidently needed done (and which they basically did for themselves). The environment was initially very uncomfortable, but as the Victoria, Central or Northern Tube lines show, not immediately uninhabitable at first. In fact, if anything, it would have been a very controlled and safe space for most passengers for the first 20-30 minutes, in comparison only with circumstances which would clearly have been so difficult and dangerous outside. As the comments from the Buzzstream blogger indicated!
However, once the trains had been stationary for longer than this, and particularly when the power was turned off over a wider area, the conditions would have become their own bespoke "immediately uninhabitable" environment. This is very different from what the railway would normally consider an emergency - for example, a train accident encompasses things like a collision or a fire, but not a crush-loaded train of cold passengers soiling themselves, which is degrading and highly unpleasant from a human but not technical point of view.
At this point, I can understand people effectively trying to do an emergency evacuation, and by this point ideally it should have been realised that people should be leaving these trains. But as others have said, the focus may well have been far too closely tuned onto the task of fixing the original fault. This is not something you'd usually do straight away if you'd hit another train, and neither should it be if you have freezing passengers urinating down the aisles. This is, hopefully, a pertinent point.
So it's very difficult to know how the information should have been portrayed over the PA, and also extremely difficult to know how a driver working on their own should be asked to do that, too. The case is probably mounting for dedicated welfare checking roles using onboard CCTV, and with those personnel able to recommend an immediate change in tack when recovering passengers from stranded services. But there must be simpler solutions, and I'm sure these can come out with some creative thinking after the heads have perhaps been banged together by the ORR and RAIB (which I hope they will be).
It's also clearly a fine line as to when you need to change between "perhaps thinking about controlled evacuation as a last resort" and "we must retrieve everyone to a station or access point right now". Fleet Controllers use timers to time-out fault finding, and have procedures known as "Cut and Run" and "Fix or Fail" to ensure action after a set period of time. I sense a marginally more complex timer solution is needed for those controlling the overall service recovery and incident response, able to time-out recovery attempts, based on a maximum time on a stranded train based on whether power is turned on or off, the outside conditions, likely or confirmed passenger numbers per coach, and the local area. We'll see, I suppose, if something like this is ever implemented!