Fire safety: reactions to a fire alarm
Mind how you behave
20 January 2012
There is no knowing how people will react to a fire alarm, says David Ware
Having been in fire safety for over 20 years, I have attended numerous incidents where people’s reactions to the fire alarm were different to those assumed by the designer/architect. This article explores the factors that need to be considered with regards to human behaviour when carrying out your fire risk assessment.
On one occasion, I visited a supermarket where the fire alarm was sounding because of a small fire. I asked the manager why it was still full of shoppers. It transpired they had refused to evacuate the building because they wouldn’t leave their shopping.
This echoes other high-profile fires such as the serious fire that erupted just after 1pm on 8 May 1979 at a Woolworths store in Manchester. The fire, which started in the second floor furnishing department, killed nine shoppers and one member of staff, of whom three were found just six feet away from an exit. Of the 12 calls made to the fire service, none had come from the store itself. The second floor was gutted by the fire, while the third floor suffered severe smoke damage.
Subsequent research showed a number of customers (predominately in the public restaurant area) refused to leave despite the sounding of alarms, requests from staff and even the smell and visibility of smoke. Some even continued to queue at an abandoned check-out. The majority of those who perished were in this area.
On average, two-thirds of the time from the onset of the alarm to reaching exits was spent by people not moving at all
In November 1987, a fire at King’s Cross Station resulted in 31 fatalities and 600 people being sent to hospital with smoke inhalation. Many of the deaths were attributed to delay in responding to the fire alarm, and, according to BBC TV News at the time, although smoking had been banned on underground trains since 1984, many commuters lit up on the wooden escalators on their way out.
How people react in a fire situation is a key element of any fire risk assessment. I have seen many fire risk assessments and little has been mentioned either about how people are likely to react, or what you can do to speed up this reaction to ensure prompt evacuation. There is no point in the assessor ensuring travel distances strictly comply with the various codes if everyone just sits there when the fire alarm operates.
So, how does the general public react when a fire alarm sounds in a public building?
Generally speaking, people do not run out of the building, but instead look around to try and get more information. They ask themselves: is it a test? Is it a false alarm? This is further compounded when staff are untrained in correct fire procedures and continue serving customers: if the public see the staff ignoring a fire alarm, they will too.
Research has been carried out on human behaviour with the findings being incorporated into fire standards such as BS 7974 and more recently BS 9999. One such report was carried out “to determine what factors may deter people who are escaping from a fire from using internal escape routes and, having established the factors, to assess their importance”.
The most important finding of this research is that the start-up time (i.e. people’s reaction to an alarm) is as important (if not more so) as the time it takes to physically reach an exit. On average, two-thirds of the time from the onset of the alarm to reaching exits was spent by people not moving at all. This is based on the timeline analysis, which is the main principle of BS 9999 (Figure 1).
BS 9999 Effective Fire Safety in the Design, Management and Use of Buildings allows you to increase travel distances when it can be demonstrated that you can reduce the pre-movement time by early warning along with good management. This is the time between occupants becoming aware of the fire and the occupants beginning the life safety strategy as shown in Figure 1.
Figure 1: Fire growth timeline
Do people die in fires because they panic?
Panic is an overwhelming feeling of fear and anxiety. The notion that people will act as inanimate objects running to exits is questioned by social scientists, who accept delay is the main factor that contributes to death in large-scale fires.
There are nine factors considered to create the model of human behaviour:
- Deaths in large-scale fires attributed to panic are far more likely to have been caused by delays in people receiving information.
- The fire alarm sirens cannot always be relied upon to prompt people to immediately move to safety. People in public buildings assume nearly ALL fire alarms are false alarms or fire drills. This is where management should back up the fire alarm with a voice system and ensure that all staff are trained to start the evacuation. Don’t forget this is the primary role of the fire warden.
- The start-up time (people’s reaction to an alarm) is just as important as time taken to reach an exit.
- Much of the movement in the early stages of fire is characterised by investigation, not escape.
- People tend to move to familiar exits: they like to go out the way they came in. You could argue that if all the occupants are likely to go to the main entrance in public buildings, then this should be wide enough to accommodate them: in fact, fire codes recommend that the designer makes the way in the way out as well.
- Individuals often move towards group members with whom they have emotional ties. (Human ties are an extremely important factor when considering your evacuation strategy. A hospice I visited had problems with their evacuation strategy because ALL visitors were advised to leave the building on the operation of the fire alarm while the staff investigated. The problem was that the parents would not leave their children and evacuate. Would you?)
- Fire exit signs are not always noticed (or recalled) and may not help with problems in orientation.
- People are often prepared to try and move through smoke, especially if visibility is over 10m. This is considered in fire engineering strategies for tenability and is a key parameter in justifying extended travel distances in buildings.
- People’s ability to move towards exits may vary considerably (e.g. a young fit adult versus an elderly or disabled person). All risk assessors must consider people with disabilities and carry out either Personal Emergency Evacuation Plans (PEEPs) or Generic Emergency Evacuation Plans (GEEPs).
Human behaviour in fires can also affect evacuation from high-rise buildings. In February 1972, a serious fire broke out in the Andraus Building, a 115m high, 32-storey building in downtown São Paolo, Brazil. It caused the deaths of 16 people who were trapped in the building: a further 330 were injured. Most of the survivors chose to climb to the roof of the building where they were evacuated by helicopters. The fire was heavily publicised on all TV news channels.
The start-up time (i.e. people’s reaction to an alarm) is as important (if not more so) as the time it takes to physically reach an exit
Two years later, another serious fire broke out in the 25-storey Joelma Building, again in downtown São Paolo. There were 756 employees present: it had no emergency exits, fire alarms or fire sprinkler systems. Initial efforts led to the successful evacuation of some 300 employees before the heat and smoke became too overwhelming. Many people did not try and evacuate but went straight to the roof: the problem on this occasion was that the roof was much smaller and helicopters could not gain access. There were 90 fatalities, all because of people remembering a similar incident.
In conclusion, human behaviour in fires and the likely reactions of people to a fire alarm sounding must be taken into account. Consider what can be done to ensure a prompt and safe evacuation from buildings: installing a voice alarm system, for example, should reduce the evacuation time from many public buildings and carrying out frequent fire drills with well-trained staff is essential if evacuation is to be successful.
David Ware is Managing Director of Fire Risk Consultancy
Related competencies include: M009