18 August 2010

New Paper on Australian Bushfires

I am a co-author on a new paper on Australian bushfires, just accepted for publication in the journal Weather, Climate and Society of the AMS. Here is the abstract and citation:

This study re-evaluates the history of building damage and loss of life due to bushfire (wildfire) in Australia since 1925 in light of the 2009 Black Saturday fires in Victoria in which 173 people lost their lives and 2,298 homes were destroyed along with many other structures. Historical records are normalised in order to estimate building damage and fatalities had events occurred under the societal conditions of 2008/09. There are relationships between normalised building damage and the El Niño-Southern Oscillation and Indian Ocean Dipole phenomena, but there is no discernable evidence that the normalised data is being influenced by climate change due to the emission of greenhouse gases. The 2009 Black Saturday fires rank second in terms of normalised fatalities and fourth in terms of normalised building damage. The public safety concern is that of the 10 years with the highest normalised building damage the 2008/09 bushfire season ranks third, to the 1925/26 and 1938/39 seasons, in terms of the ratio of normalised fatalities to building damage. A feature of the building damage in the 2009 Black Saturday fires in some of the most affected towns – Marysville and Kinglake – is the large proportion of buildings destroyed either within bushland or at very small distances from it (<10 m). Land use planning policies in bushfire-prone parts of this country that allow such development increase the risk that bushfires pose to the public and the built environment.

Crompton, R. P., K. J. McAneney, K. Chen, R. A. Pielke Jr., and K. Haynes, 2010 (in press): Influence of Location, Population and Climate on Building Damage and Fatalities due to Australian Bushfire: 1925-2009. Weather, Climate, and Society.
While it is certainly interesting that we do not find any signal of long-term climate change in the loss record, that finding certainly is not entirely unexpected given the growing body of research in this area.  The more significant findings of this paper have to do with issues of land-use planning and the relationship of bushfire damage to the El Niño-Southern Oscillation and Indian Ocean Dipole phenomena.

In due course I'll post up a pre-publication version of the paper.


  1. It is interesting to see that while a climate signal is detectable, a climate signal from greenhouse gases is not.

  2. -1-Jason S

    Yes, but not surprising. The effects of ENSO and the IOD are much larger and on much shorter timescales than the expected effects of GHGs on climate.

    To take an obvious example, we can see climate signals from seasonality every year.

  3. Hi Roger,

    It looks like a very interesting paper.

    As I noted at the time of the bush fires, I don't accept that housing development *must* be altered in order to lower loss of life and property. Especially loss of life. As an engineer, I realize that multiple solutions exist for almost all problems. If only land-use planning measures are mentioned in the paper, that will unfortunately lead to inference on the part of the readers that altered land-use (development patterns) is the only solution.

    There were various instances where simple lucky circumstances saved lives in the Australian bush fires. I recall the story of a woman and her children who took shelter in a wombat hole. And there were two men who took refuge in a water tank. These instances lead me to think that there are relatively straightforward and inexpensive means to prevent loss of life, even if fires rage right over the top of the inhabitants. For example, my townhome has a crawlspace with an earthen floor. Fire isn't a big enough danger in NC to do anything, but if I were in the Australian bush or southern California, I'd definitely build something such that I could go into some sort of structure/excavation in the crawlspace on 1-2 minutes' notice. That way, even if I realized only at the last minute that I couldn't protect my home, my home could burn down literally on top of me, and I'd still survive.

    A similar argument could be made for property protection, though I think the solutions would be more costly than the solutions to protect life. For example, as I mentioned at the time, there are water-absorbing gels (as used in diapers) that are significant improvements on water alone. Looking farther into the future, I don't think that it would be unreasonable to expect autonomous or remotely operated firefighters/firefighting vehicles in the coming 1-3 decades. Robots/robotic vehicles have the obvious advantage that they don't need to breath oxygen-containing, uncontaminated cool air. They also can be left to fight fires in essentially "suicide" situations.

  4. Dear Roger,

    Do you know of any studies that go further back in time, e.g. way back to the time of settlement in Australia (or even earlier)? My understanding of history is that wildfire was a big problem for the early settlers in Australia but can't find any references for this.


  5. Alex

    I will ask around the office and if I find anything out I'll post it here next week.

  6. Hi Roger,

    You say the results are not surprising, but the 2009 Bushfires Royal Commission Final Report refers to climate change as a relevant factor in its introduction, so I find your papers results quite surprising.

    I am interested in any findings in your appear that may allow us to better predict bad bushfires seasons looking at El Niño-Southern Oscillation and Indian Ocean Dipole phenomena, although I am unsure how much notice we can get of these occurring.


  7. Kate

    Our result is not surprising, even in the context of the part of the Royal Commission report that I think you’re referring to. It is important to distinguish between the anthropogenic climate change effect on bushfire weather and the effect of it on bushfire damage. Our paper addresses the latter. We show that we cannot detect an anthropogenic climate change influence on damage after the historical records have been normalised. This is not surprising, as Roger mentioned in -2-, especially when you consider that bushfire damage is not solely a function of bushfire weather.

    Even with a gradual shift in bushfire weather, a bushfire still has to be ignited (and non-natural elements such as arson and power lines have a role to play here), and once a bushfire is ignited, it has to impact a populated area (the chances of which are less in Australia as compared to other countries). These factors produce a large amount of annual variability making detection all the more difficult. What our results show is that increasing building damage due to bushfire in Australia is largely being driven by societal factors.

    Damage prediction is particularly difficult for the reasons referred to above. Although bushfire weather is influenced by the phase of ENSO or the IOD, extreme bushfire weather does not always produce severe damage but severe damage is more likely to occur in extreme bushfire weather. This is what our relationships between building damage and ENSO and with the IOD show. There is a significant amount of variability in damage even within a phase of either cycle and this is what makes damage prediction extremely difficult. There is other research available if you are interested in the predictability of ENSO etc.

    I might also mention that we submitted a draft version of our paper to the Royal Commission on their request.

  8. Sad that the Australian media appear to have taken little interest. have you had any interviews with Oz media?