Three girls were killed in a club stampede last weekend in my home town, Budapest, and speculations are all over the Hungarian interwebs ranging from stabbing to racist accusations to asthma.
I am no expert in pedestrian traffic flows but a few of my colleagues are so I have absorbed some basic ideas on how stampedes work and why they are sometimes almost inevitable.
If you think pedestrian flows are quite distant from railway engineering, think again: what about designing stations, platforms, train doors and interiors? Trains and stations have to be evacuated sometimes, and even in regular operations, it is essential to have an idea of what affects dwell times (i.e. the time needed for alighting and boarding).
The basic phenomenon is the following: when people are condensed to some critical density, marginal physical forces of each individual add up to a lethal pressure and this is even true without the panic which inevitably bursts in such occasions. That means, although in many cases, it feels more comfortable for the outsider to imagine people getting stabbed, a bomb or a bomb threat, often the only cause of the events is the density and some conflicting (but not necessarily aggressive) movements. Often, victims are not even stepped on, they are still standing but pushed from all directions. (See the Love Parade video at the end of the post.)
The simple way to avoid these events is to estimate the amount of people safely fitting into a venue and do not allow more people. This easy (and by the way already legally required) measure would have saved the victims both in Budapest and at the Love Parade in Duisburg in 2010. You don’t even have to use very elaborate calculations: just like designing bridges to safely bear the weight of hundreds of cement trucks, you just need to use a generous factor to err on the safe side.
The hard way, if many people already pile up, is less obvious, and there is still much to work on as far as I understand. There are requirements to offer much more emergency exits than normal doors, but there are some less trivial measures which are proven to help: counterintuitively, installing some barriers in the way of the crowd can speed up evacuation and make it safer, as dividing people to smaller groups reduces the forces and conflicts.
Perhaps the most common such a barrier is the security path between the left and right sides of a standing concert audience — not only the most crowded middle front parts are at least separated to two now, but security guards can easily help at these densest areas.
A crowded London Underground station (photo: LaN_Luis)
Why can’t we always go for the simple way? Well, you can always limit a number of people into a venue (assuming, fairly enough, that the capacity of the outside world is infinite) but this is not true for unforeseen events such as political protests or during natural disasters. Also, at large festivals significant people can gather at a stage although the total festival capacity is still unused. Finally, at large transfer stations (imagine a subway interchange in London, Paris or virtually anywhere in Asia), the number of attendants are undefined — there is literally an infinite flow of people.
By the way, there is almost certainly another dynamic of such accidents at a much larger level. At this time, everyone seems very concerned about crowd safety and luckily I am not the only one mentioning the forces of the crowds as a cause besides knifes etc. But this is essentially hindsight wisdom — I am pretty sure (although not totally) that there will be no similar accident in Germany, or in Budapest, in the next decade. To give another example, it took Sweden the Gothenburg disco fire to introduce stricter regulations, or at least stricter enforcement. To paraphrase a bonmot about financial crises, maybe the next stampede happens when the last person remembering the previous one retires?
Further resources:
- Prof. Serge Hoogendoorn (TU Delft) explains what happened at the 2010 Love Parade in this video, the talk is in Dutch but original recording and a simulation is shown at 00:37:17 (no shocking footage here).
- Prof. Keith Still is the most known specialist in crowd management, so read on more on his site.
Disclaimer: As I said earlier, I am no expert in this domain and just wanted to list some pointers to others who are — nevertheless if you have any corrections or can recommend more online resources, please leave a comment.
(source: ProRail)
