Over at Penguin Unearthed.
Extreme distributions
20 February, 2009 by penguinunearthed
John Connor, CEO of the Climate Institute, made a speech today talking about bushfires. Ive been pondering one of his key points for the last two weeks, ever since the bushfires.
Climate change is not just about warmer weather, it is about wilder weather.
One of the lessons I learned early on as an actuary is that the extremes of any distribution do not behave the same way as the middle. So if the middle of a distribution of temperature rises by 1 degree, the effects on the extremes are not intuitive.
Something that was a 1 in 100 chance of occuring in the old distribution is not just slightly more likely. Moving the distribution up the curve can significantly increase the chance of that 1 in 100 event. For an extreme event that is three deviations away from the mean of a distribution, moving the mean up 20% of a standard deviation doubles the chance of that extreme event. So put it in temperature terms – say the summer mean maximum temperature is 25 degrees, with a 1 in 400 chance of a 40 degree day. Increase the mean maximum temperature to 26 degrees, and there is a 1 in 200 chance of a 40 degree day. And that is assuming that the climate follows a nice stable normal distribution model (or bell curve – the one that statisticians like, because its easy to work with mathematically).
You dont need complex models of extreme weather events following climate change to realise that extreme events become much more frequent if the average temperature moves up a little bit. But complex models of extreme weather events suggest that the extremities move even further than a very simple normal distribution would suggest.
The Climate Institute put out a very detailed report from in September 2007. They first of all looked at the increase in extreme fire danger in the two global warming scenarios – low global warming (0.4 degree by 2020 and 0.7 degrees by 2050) and high global warming (1 degree by 2020 and 2.7 degrees by 2050) – based on projections from the Intergovernmental Panel on Climate Change (IPCC). The number of extreme fire danger days in 2020 in the low global warming scenario increased by 5-25%. And the number in 2050 using the high global warming scenario increased by 100-300%.
Extreme fire danger: image credit Parkaboy
They developed two new categories beyond the Australian standard extreme fire danger days – very extreme and catastrophic. Catastrophic fire danger days, which at the time of the study had occurred 12 times in the more than 30 years of the temperature records studied at the sites studied, were projected to occur once every three years at 7 of the 26 sites studied, and once every 8 years in 12 more.
Given that the flash point of eucalyptus oil is 49 degrees, bushfire risk does not just increase in a linear way with temperature. The closer the air temperature gets to 49 degrees, the more likely that just some random heat point will set something off. The maximum temperature in much of Victoria on Saturday 7 February was around 46 degrees. And there had been little or no rain for the last month, with baking temperatures. Although that weather pattern is quite theoretically possible without global warming, global warming makes it much more likely.
John Connor is right. Time to add Very Extreme and Catastrophic to all those bushfire warning signs you see everywhere in the Australian bush.
La Griffe du Lion does cunning things with integrated tails of probability density functions too. How surprising the results might be largely depends on familiarity with the theory and the way it is presented.
For example, saying that a 1 degree increase in the mean temperature causes all the 39.5 degree days to turn into 40.5 degree days isn’t surprising at all. If I said we are going to double the probability of extreme weather events (in the form of days over 40 degrees) then both surprise and panic ensue. But really I’m saying the same thing in two different ways.
What people have always done is merely to define “extreme” to mean whatever doesn’t happen all that often (i.e. they move the threshold) then learn to live with the new normal. So the number of “extreme” days will never change.
All of which has nothing to do with the frequency or intensity of bushfires. Bushfires in Australia are nothing new. The frequency is determined by the rate at which fuel builds up (I’ve walked through National Parks where the stack of dry fuel was between half a meter and a meter high in large sections of the underbrush). The intensity is determined by how long the fuel has had time to build up, and whether it gets backburned during winter (no fuel = no fire). The loss of life and property is determined by how close people’s houses are to the bush and the availability of swimming pools, dams, water tanks, etc.
I don’t see global warming changing any of those factors.
This of course was the same argument that cost Larry Summers the Presidency of Harvard. He tried to excuse the lack of women in the Maths faculty at Harvard by saying that small differences in average maths fluency between men and women in the population translated to much larger differences in the number of male and female maths geniuses.
Tel, I understand that the worry for south eastern Australia is that average temperatures may increase by considerably more than the world average, due to changed ocean currents. Just as the UK may actually get colder.
Tel, I’m not sure that’s entirely true – the point is that just a 1 or 2 degree increase in mean temperature all on its own might well mean that events like Black Saturday become, say, twice as likely, all else being equal. And surely it’s not the temperature either – the inevitable consequence of increased greenhouse gas concentrations is that there’s more energy trapped inside the troposphere, which is almost certain to translate to stronger winds. It was the wind as much as anything that made the recent fires so devastating.
I understand the point, and it is wrong. Fires require fuel and the fuel buildup takes time. It is possible that a warmer mean temperature will cause more sudden fire outbreaks due to volatile oils and fuel/air combustion but without hard fuel (dry wood, dry grass, leaves, etc) you won’t get a “Black Saturday” type situation.
http://www.csiro.au/resources/Victorian-Bushfires-QA.html
Damn, I screwed up my quote marks yet again.
Tel_, I said “all other things being equal”. I’d guess less far than 1% of the total fuel available in Victoria for a major fire was burnt up on Feb 7. So the limiting factor for the frequency of such fires is unlikely to be the availability of fuel. Further there were fires on that day that ravaged areas that had been cleared recently and had little more than short grass available as fuel.
Of course, where fuel reduction is important is around inhabited areas. I agree we’re not likely to get 200+ dead in a day that much more often, for various reasons, including fuel concentrations around towns etc., but also because we’ll improve our ability to save lives in such events. The worst consequences might be more ecological than in the immediate threat to human lives and property.
I agree, that immensely higher and drier weather won’t, of themselves, cause more bushfires. Otherwise the Simpson Desert would be riddled with them. But we already had lots of conditions leading to bushfire in the South east of Victoria – fule loads, houses ready to be burned, eucalyptus trees etc etc – look at 1939 and the Ash Wednesday fires. If you add even higher temperatures and drier weather to the mix (and yes, I forgot the winds) then it gets worse.
It is true, that if you just add a degree of temperature to a distribution, then each point of the distribution goes up one degree. But the probability of a temperature of more than 40 (say) will nearly double (using my simple normal distribution example).
And from everything I’ve read about these fires, the weather conditions, more than the fuel load, made them an order of magnitude more catastrophic than the usual run of the mill extreme fire danger bushfire.
Royal Commission finding or not, I can envisage carefully proscribed green belts ringing equally well defined population centres with the volatile eucalypt now rated a noxious plant within that green belt. Anyone seeking to live outside the village, hamlet, town will have to use the most fire resistant building materials and have a registered escape plan and/or bunker, perennial farm dam etc.
Like the old castle moat, the green belt of exotics will be a real challenge to select and nurture but in the face of climate change and ‘catastrophic’ fire danger days it might be a good bet. At least it will save a lot of back burning.
pablo, as NPOV pointed out, the meme that “there wouldn’t have been these deaths if we’d done more autumn burns (and its the bloody greenies’ fault they didn’t)” is quite untrue – especially the bit in brackets. As much burning was done as weather permitted but it made little difference to this extreme – probably once in a millenium – event.
But neither can you confidently blame AGW. Once in a millenium events do happen every thousand years or so.
I dunno – I’d say you’re on reasonably firm ground to claim that if there had been no climate change over the last 100 years, the fires on Feb 7 wouldn’t have been nearly as bad. A large number of temperature/rainfall records were broken in the months leading up to that point.
But realistically, we aren’t in a position to change the climate back any time soon, so there’s not a lot of point blaming CO2 emissions, except perhaps to keep the potential damage from AGW in the public’s mind.
DD@10. I’m not generally opposed to back burning. My point was that with climate change and catastrophic possibilities – particularly the flash point of 49 degrees for eucalyptus – no amount of CFS prevention planning will ensure life and property survive. Therefore seemingly drastic measures need to be taken at population points and suitable exotic vegetation as green belts could be a milder version of these measures. I would also add that ridge top living probably should also be proscribed.
“Weather permitted” is a completely subjective measurement. Burning in any weather is better than waiting for the hottest day of the year. Fact is, there is significant political opposition to backburning and here is one example:
http://www.crikey.com.au/Politics/20081017-Bushfires-or-the-biggest-backburns-in-living-memory.html
The relevant question to ask is, for each area where people died, how many years ago was fuel reduction (by whatever method) done, IN THAT AREA? The other relevant question is, how close to bush were the houses that burned? Let’s see if the Greens are willing to research and publicize those points.
I can confidently state that none of the areas that burnt out this summer will burn again for at least the next ten years, probably longer. I can make that statement regardless of global warming or freak days because I know there is no fuel on the ground.
http://www.theaustralian.news.com.au/story/0,25197,25031389-7583,00.html
I know it’s the Australian which has generally low credibility but the author is a scientist not a journalist, certainly not one of Murdoch’s staff writers.
“Weather permitted”, Tel, means that “not so wet that we can’t light the bloody thing and not so windy that it won’t escape”. Even then, backburned fires do escape – and then there is legal hell to pay from affected property owners (who are generally anything but greenies, BTW).
This stuff coming from the barking mad Right about how its all the greenies fault is fact-challenged. Marysville HAD a well maintained firebreak built around it – no greenies objected when it was built, and it made no difference anyway. The guy who was fined for “clearing trees from his own property” turns out to have cleared 18,000 of them and sold the timber. Greenies groups have been trying to get councils to stop people building on ridgetops for decades – partly because of the fire risk (fire being a natural part of the forest), and partly because if you can see for miles from a house then the house can be seen for miles. Too much backburning in some areas can increase the risk of hot fires by replacing less flammable rainforest with very flammable open woodland. Etc, etc …
dd…where’d you get the info about having cleared 18000 trees? I thought it was only 250. As for selling the timber…fair enough, I’d think. That might just have recouped the cost of the licenses to cut them down.
NPOPV, a little googling shows you’re right. Dunno where that number came from – I should do some fact-checking in future before asserting such things. I gather he did flog the timber, but.
NPOV has already called it out, but just for a concrete reference: http://www.mitchellshire.vic.gov.au/Files/12_Sept_05_Minutes.pdf (page 66)
And I’ll just add that Eucalyptus mostly only lives for 30 years, so the definition of “old growth” is pretty weird but I’m no legal expert.
Look, let’s at least be clear about the terms we are dealing with. There are three things that you need if you want to start a fire and they are:
[1] Air
[2] Fuel
[3] Ignition (heat/spark/flame/ember/etc)
The fire break is primarily a device to prevent the spread of fire, in other words factor [3], a backburn is addressing factor [2], they are completely different things for different purposes. While ignition prevention is a good idea (especially in the hottest parts of the year), you can never win by depending on ignition prevention alone because there are so many sources of ignition.
Global warming and record-breaking hot days essentially increase factor [3] by winds that carry embers, and volatile oil that forms a fuel/air mix that can easily flash up and by pre-heating and pre-drying the wood. All of these are ignition-related risk factors. Plus you can be sure that people either careless or destructive will create their own ignition sources. Put this together and you will always have ignition… always.
Thus, when we understand that [1] and [3] are guaranteed to happen, the only limiting point is [2] which is fuel availability.
If you don’t believe me on this, then go an look at basic fire safety guides (all of which cover the “triangle” of [1], [2] and [3] above) and also specifically at the work done by David Packham and the CSIRO into bushfire management. Go and get some scientific references, or just find a family who has survived in a bushfire prone part of Australia for more than one generation and get them to explain it.
If you want to also go into the questions of property damage and loss of life, you also have a whole heap of research into distance between houses and the edge of the bushland and also the types of trees planted near houses.
Well now I’m calling you out Tel – “Eucalyptus has an average lifespan of 30 years”??
The Mountain Ashes that that area of Victoria is known for as an average lifespan of 400 years (http://en.wikipedia.org/wiki/Eucalyptus_regnans). I’ve no idea what species Mr Sheahan had on his property – the council records don’t state (but they do make interesting reading – apparently aborialists and fire strategists were brought in by the council to determine whether the clearing was a suitable firebreak, which if it were, would have been exempt from the planning laws prohibiting the clearing of frees. The specialists apparently determined that the firebreak was not suitable, though there’s no details as to why) – however I think it’s fair to guess that many of the trees would have been well over 30 years old. Tel, which eucalyptus species have a 30-year average life-span?
Hmmm, that Wikipedia page doesn’t provide a citation for the “400 year” figure. For those references that I did try following, I could not find one discussing the age of the trees. After more searching I found very few web references that do discuss the typical age of Australian trees. I’ll admit I was probably a bit low on the 30 year estimate but I’m still curious what the definition of “old growth” really is and who had the job of counting tree rings.
If you search on “Jann Elizabeth Williams, John Woinarski” you get a google book result, who come up with the conclusion:
Note, these are maximum ages, not typical ages. The great majority of trees don’t reach the maximum age so the average would be considerably lower. I’m not sure what counts as fair wear and tear on a tree.
There’s also a study of plantation timber here: http://www.fao.org/DOCREP/004/AC121E/ac121e04.htm
They rarely leave plantation timber in the ground longer than 20 years because essentially the tree doesn’t grow a lot bigger after that so the value diminishes for the plantation owner. That’s not really a fair comparison to trees in the wild, but it gives some idea of the maturity age. If you wanted to replace a forest this would be the time taken to grow “full size” trees. Also, plantation trees have been closely studied so more information is available.
I’ve just seen an awful lot of in-the-wild Australian native trees that died from borers and termites and just generally dropping branches off and stuff. You see plenty of dead gum trees sticking up with trunk no wider than your leg, or half dead gums with only a strip of living bark and the rest of it dead.
Probably there’s a job for someone wandering around Australia looking for fallen trees and still in the ground dead trees and then getting out the microscope and counting the tree rings. Looks like there aren’t a whole lot of studies available on the subject. Everyone seems determined to build taxonomy breakdowns like this one:
http://www.flora.sa.gov.au/cgi-bin/texhtml.cgi?form=speciesfacts&family=&genus=Eucalyptus&species=&iname=&submit=Search
Really useful for identifying a tree, not much information about the tree once you have identified it.