100 years of 'stay or go'

This article first appeared in the Winter 2009 issue of Fire Australia magazine

I was fairly sure that we would die there, as I didn’t think we could stand up to the flames and the heat. I don’t know how long we remained in the gutter. It couldn’t have been more than an hour. Throughout the time that I was there I felt I was expecting help from some outside quarter. It wasn’t until the time that I decided to get up that I realised fully that our only help lay with ourselves…” - Survivor and witness of the Ash Wednesday Bushfires.

Before the tragic February 2009 bushfires in Victoria Bushfire CRC researchers had looked at Australian bushfire fatalities over the past 100 years. Lead researcher Dr Katharine Haynes is a Bushfire CRC researcher at RMIT University and Macquarie University’s Risk Frontiers, was summoned to provide detailed evidence on the study to the Victorian Bushfires Royal Commission in May.

Australian bushfire policy for community safety is unique. Rather than attempting to evacuate all those who may be in the path of a bushfire, fire authorities in all States allow the public to make a choice: either get out of the area early, or prepare to stay and defend homes and property from the fire.

However, apart from a handful of post-fire investigations, no detailed research has ever been carried out into the circumstances of all recorded bushfire deaths in Australia.

A detailed database of Australian bushfire fatalities has been created by Risk Frontiers as part of a larger database of Australian natural hazards. This was augmented through a thorough documentary analysis of forensic, witness and police statements within coronial inquest reports. The database includes details of 552 civilian fatalities over the last 100 years. It provides a unique opportunity to assess the circumstances in which people perished and the suitability of the ‘stay and defend or leave early’ policy.

Analysis of the data clearly shows the dangers of being caught outside during a bushfire, with the majority of fatalities occurring whilst victims flee the flames during late evacuations followed by those caught outside defending wider property. The minority of deaths that occurred inside buildings were due to the inhabitants sheltering in place and not defending from ember attack.

In addition to substantiating the ‘stay and defend or leave early’ policy, the analysis has demonstrated the heightened vulnerability of women, children and the elderly. This is due to their propensity to evacuate late and their greater reliance on others for assistance.

In particular, there is evidence of a gendered division of roles and responsibilities during bushfires that contribute to these vulnerabilities. While men are most often killed outside, while at work or attempting to protect assets, most female fatalities occur while sheltering in the house or attempting to flee. The number of men killed by bushfires has decreased; however, this is not the case for women and children, who in recent years have died in relatively high numbers. Where possible, the decisions leading to fatalities have been examined based on people’s awareness of the fires and the effectiveness of the actions that they took to reduce their risk.

This Bushfire CRC study covers the main trends emerging from an analysis of Australian bushfire fatalities over the past 100 years. Attention is paid to the last 50 years and, in particular, to large fires within this time period.


In Australia in the past century, bushfires have rated as the fourth most hazardous type of disaster in terms of death after heatwaves, tropical cyclones and floods. With the advent of climate change and predictions that the bushfire season is likely to become longer and bushfires more frequent and more intense, it is crucial we are equipped with the best knowledge of how to reduce the risks to life presented by bushfires.

Australian fire authorities in all States allow the public to make a choice: either get out of the area early, or prepare to stay and defend homes and property from the fire. In other countries with a high bushfire risk, such as the USA and parts of Europe, evacuation is still seen as the safest emergency management approach.

The Australian Fire and Emergency Service Authorities Council (AFAC) ‘prepare, stay and defend or leave early’ policy is based on evidence from reports into some of Australia’s worst bushfire tragedies, including the Hobart bushfires in February 1967 in which 64 civilians were killed and the Ash Wednesday fires in 1983 in which 60 civilians were killed across South East Australia.

(A number of volunteer fire-fighter personnel were killed in the Ash Wednesday fires in 1983, including the first female volunteer to be killed. This paper does not deal with fire-fighter deaths as we are interested in civilian actions: however the great sacrifice made by all fire-fighters is recognised.)

Investigations into these deaths have revealed that (a) people were more likely to be killed by radiant heat or a vehicle accident while evacuating and (b) well-prepared houses can be successfully defended from bushfires and can provide safe refuge for people during the main passage of the fire front.

However, despite the apparent robustness of the AFAC policy position, until this study no detailed study has examined the causes and circumstances of civilian deaths during a number of different bushfires over a significant time-period.

A number of reports, books and website sources document the number of people, livestock and properties destroyed in various bushfires and the financial costs. However, none of these examine the circumstances of these deaths in any useful detail.Reputable electronic examples of these sources include the Emergency Management Australia (EMA) website www.ema.gov.au/ema/emaDisasters.nsf and the United States (US) National Interagency Fire Centre website www.nifc.gov/fire_info/historical_stats.htm. There have been few detailed investigations of the circumstances of these deaths. Limited examples include an investigation by the US Texas Department of State Health Services (TDSHS) after the Texas wildfires on 12 March 2006, which killed eleven civilians (TDSHS, 2007).

This study identified five separate incidents that led to the deaths: eight were killed directly or indirectly because of vehicular accidents, two because of late evacuation and one from refusal to evacuate (sheltered passively in the home).

An investigation in Australia by Krusel and Petris (1999), which examined the 1983 Ash Wednesday civilian deaths in Victoria, is the only study to examine coronial records in order to carry out a detailed examination of the actions of bushfire victims before their deaths.

This study identified that the majority of victims died during a late evacuation as the fire front arrived or because they were incapable of implementing a safe strategy due to inadequate warning, age and infirmity.

 Policy relevance

Translating the stay or go message into practice is complex, with a great deal of ambiguity in the ‘leave early’ advice and gaps in peoples’ perceptions of the actions they should take.

Many people still consider late evacuation as a valid last resort option: waiting to see how the fire develops and then fleeing at the last minute. Evidence also shows that people frequently retain a ‘fall-back plan’, with children, valuables or pets being loaded into cars so that they can be evacuated if the situation is deemed too dangerous. In other cases, people expect the emergency services to provide help, information, warning, guidance or assistance.

Recent inquiries have focused on whether people had enough official warnings to make a decision to stay or leave. However, a clear intention of the AFAC policy position is one of community self reliance, where people are encouraged to prepare both mentally and physically for a situation in which they may have no warnings and therefore no choice but to stay and defend their homes.

Our study demonstrated that some victims were not mentally or physically prepared to stay and defend their properties, often not knowing what an effective survival strategy would be or misunderstanding the messages (for example, people sitting in baths of water or sheltering passively in their house). Many others underestimated the ferocity of the fire.

Although these are issues the fire agencies now try very hard to rectify, this study yields some difficult examples for the ‘prepare, stay and defend’ argument. For example: how do we know how much preparation to fight the fires is enough? Even those who plan to stay are often not well-prepared: they do not expect to lose electricity or water, have no back-up plan and do not wear adequate protective clothing.

Of primary importance is the need to recognise and target vulnerable groups such as men who die attempting to defend property outside and women and children who die fleeing or sheltering in an undefended home.

Although this study has shown that men’s lives are equally endangered, these risks have been considerably reduced through a male-orientated gender blind risk management campaign.

Policy, outreach and long term community based projects must also now be ‘gendered’ and ‘aged’ in order to that women and children can equally reduce their risks and vulnerability to disaster.



Release date

Wed, 01/07/2009