Wildfires and global warming

Off on an email list we’ve been discussing wildfires and their connection to global warming. I’ve always been somewhat cynical about the connection, which appears in the popular press [1] to amount to “fires are hot, global warming is, err, warm, so they must be connected”. However, I really do know well enough to ignore the popular press.

This kind of thing: There are many reasons for any particular fire, but basically the (wildfire) pattern is reflection of two things related to higher temperatures – earlier spring snow melt and also higher spring and summer temperatures,” he added. [2] isn’t the popular press though. Or The increase in outdoor fires in England andWales between 1965 and 1998 may be attributable to a trend towards warmer and drier summer conditions (Cannell et al., 1999). [3] (thanks Eli).

I’m entirely happy with dryness causing fire, and I’m entirely happy with warmer temperatures causing earlier snowmelt causing summer dryness causing fires. What I’m dubious about is fire risk from pure higher temperatures unmediated by drying. Maybe this is a strawman. Maybe no-one believes it; though the quote above suggests not.

JF says that Westerling and Swetnam [4] is a key paper; and that says what the quote above said, so is probably the source for the above. However it is hard to disentangle the effects of temperature and dryness, since dryness often leads to heat (since evaporative cooling is then lost, as in the Great European Summer of 2003). Indeed, W&S say “Temperature affects summer drought” though it seems odd to make it so one-way, when the system is clearly coupled, and I would have thought that the influence was more largely in the other direction. W&S make no effort that I can see to disentangle the two effects.

But does anyone actually believe that warmer temperatures, without and change in dryness, causes wildfires? Is a fire really (significantly, measureably) more likely to start and / or spread at 32 oC than 30 oC? It seems rather unlikely to me, though I suppose possible.

[Very late update: http://bbickmore.wordpress.com/2012/07/31/deseret-news-hotter-climate-equals-more-wildfires/ -W]

39 thoughts on “Wildfires and global warming”

  1. What I know a little bit about, for N. California — When it’s warmer there’s more rain (rather than snowfall) both fall and spring; this favors the invasive annual grasses that have roots spread wide and shallow catching every drop of dew, over the natives that are deep rooted and well adapted to winter dormancy. So ‘just warming’ means a very different pattern of wet ground and sunlight, versus frozen ground below snowpack, for weeks on both sides of wintertime. This favors those annuals, and they are exquisitely adapted to fire.

    I’ve done little prescribed burns on private land, and watched a fire front move across a meadow full of medusahead and cheatgrass (this search finds them both, as will anyone in the field):

    These plants are already dry with seeds fully formed by late June, and will burn before the native grasses are even big enough to suffer much harm from a mild fire.

    But the fire moving through this stuff is incredibly noisy, it sounds like popcorn popping. The result of a burn is a black sooty field — covered with little green seeds.

    The invasives _love_ fire and prepare the site to burn better. When the seeds germinate and spread shallowly they’re prepared to grab the flush of minerals from the ash at the first rainfall.

    [Ah, this is a good point. It is in the papers I cited but I didn’t think about it. Veg pattern matters, and mildly warmer can potentially change the veg a lot -W]


  2. Nerst equation states that chemical reactions speed up exponentially with temperature. Not sure that really has anything to do with the question for a fire starting, but it might for fire spreading.


  3. Ah, the key paper [4] is talking about the western US forests, and it is something like a one-way relationship — really big timber is “thousand hour fuel” and takes that long to dry out. And _that_ is a cumulative moisture deficit.

    “we found that an advance in spring produces a relatively large percentage increase in cumulative moisture deficit by midsummer”

    Yeah, I’d say ‘purely temperature’ is a strawman-wannabe, the IPCC must assume people know you can’t change just one thing.

    Might be worth comparing your local wildfire fuel moisture information — it may be you live in an area where most of the fuels are smaller lighter material and respond much more quickly to changes in humidity with temperature.

    Remember sunlight will heat up grasses and twigs and dry them by forcing moisture out even when the air around them is also humid.


  4. Well, here’s a mechanism (in the Pacific Northwest/British Columbia) that might not be dryness-related:
    Moderated winter temperatures kill off fewer pine beetles, the beetles proliferate and kill many pine trees which then burn more easily. I guess the trees themselves are drier (no transpiration), and I suppose that would result in reduced humidity, but the direction of the interaction in this story has temperature being the driver.


  5. Australia’s drought is believed to be linked to AGW. When comparing the current drought (that is last 10 years) to previous droughts, the rainfall deficiency is not unprecedented, we have had similar dry periods early in the 20th century. However the pattern of the drought seems to roughly match AGW predictions, in that it is biting hard in the south of the continent, and the north has if anything had above average rain. AGW models suggest a drying trend mostly in the south, with possible increases in rain for the far north. Previous severe droughts hit both north and south.

    What is unprecedented for Australia’s drought is the combination of lack of rainfall, and high temperatures, which seems to be an obvious recipe for increased fire risk – at least until the vegetation dies off. The worst fire risk in Australia is near the coast where it is normally wet, and then at high risk of high fuel loads and big fires when it is unusually dry, so I’d guess that if there is a general drying trend then there might not be an increase in fire risk, but rather a decrease due to decreasing fuel loads.

    The current drought in Australia is a dramatic variation on a decadal scale – the 80s and 90s were much wetter, so it is not simply a long term AGW trend, and I would suggest is either extremely unusual weather, or mostly weather helped a long a little by AGW. We’ve had some very serious fires recently, but we’ve had very serious fires in past decades as well.

    I would also guess that increased temperature with constant dryness would mean increased rainfall (to offset the increased drying effect of the heat). This could mean more vegetation, so more fuel. Then given the same period of drought, more drying out due to higher temps, so more flammable.

    The other big factor is wind. I can’t recall anything on AGW changing winds, but if the speculated increase in severe weather is right, then this might mean more strong wind events, which could increase the risk of the most dangerous high wind fanned fires.


  6. I think Australia’s recent experience is instructive. Perhaps there isn’t a huge risk difference in 30 vs 32 oC. But when you’re pushing 50, things change pretty dramatically. Tinder dry biomass, even the eucalyptus oil evaporating from the trees ahead of the fire front.

    Prof. Karoly points out that the FFDI reached unprecedented levels on Feb 7 – almost double what was supposed to be it’s maximum.



  7. maybe higher t° tend to cause more thunderstorms (more wildfire starts) ?

    [Yes I wondered about that. But thunderstorms ought to be quite humidity-dependent too -W]


  8. IN Australia it seems to be a combination of features; high load, lengthy dry spell, very hot day and high wind. I don’t think that it’s such a stretch in Australia to link this to AGW. I think that CSIRO modelling indicated that the conditions for these hot, dry, windy days will be more common due to AGW.

    [Ah, don’t get me wrong: I’m not saying there is no link to AGW -W]


  9. Also in Australia it seems most wildfires aren’t all that wild. The terrible fires just outside Melbourne earlier this year were either deliberately lit or caused by high voltage power lines (due to the additional dust load, so I hear).


  10. I don’t know about more likely at 32° as opposed to 30° for a single day, but many days in a row at 32° as opposed to many days at 30° could make it more likely.


  11. Speaking from ignorance here, but I assume that for a fire to spread from Object A to Object B, you need 1. Object B being contacted with a flame from Object A; and 2. heating of Object B to a certain temperature at which it will combust when exposed to flame. So if Object B starts out at a 2C higher temp, it seems that it would be marginally easier for the fire to spread even if humidity stays constant.

    [OK, agree so far -W]

    As for whether it’s “really (significantly, measureably) more likely to start and / or spread at 32 oC than 30 oC”, that might depend on whether the question is Pielkeian (prove from actual wildfire data that the temp difference made a difference in fire outcome with a 95% level of certainty) or a more normal question of whether fire spread is a reasonably-derived effect of AGW independent of humidity.

    [Less happy at this point. I imagine this would be suscpetible to laboratory study. My feeling (which I can in no way back up) is that a few degrees background temperature change is so small as to be lost in the noise compared to the humidity forcing -W]

    Finally, local winds in fire-prone areas might be affected by increased temps. I know global winds are expected to decrease from AGW, but local areas, like here in California where temp differentials between land and sea can drive winds, might be different.

    [I’m sure you find both increases and decreases in wind, either with warming or cooling -W]


  12. In some respects this is like planting zones, where the issues are not the average temperature but the coldest one in the winter (see pine bark insects also) and the date of the last freeze.

    In this case the hottest day may have an inordinate effect.


  13. Before reading this this i had never heard about the connection between fires and global warming but after reading both sides of the argument make sense. I see how people wuld think that because they are both warm or hot that they must be related but i am as well skeptical that that are directly related. I agree with the dry brush starts on fire easier but this dry brush is not global warming, Now was this brush directly related to the global warming that is going on? Possibly but possibly not.


  14. On the temperature scale you probably need to think in terms of Kelvin instead of Centigrade. So in the given examples the difference might be between 303 and 305.

    However to the extent that drought can be connected to GW then that could be quite a significant factor.

    I remain quite skeptical of the ability to connect local conditions like drought with changes in global temperature. More logical is to connect things like smaller snow packs, or reduced glaciers, but it is hard to see a wildfire connection there.

    Many of the wild fires in California are man made.


  15. Nicolas

    “I remain quite skeptical of the ability to connect local conditions like drought with changes in global temperature.”

    I wouldn’t always call drought a ‘local’ condition. Most droughts in Australia are regional, for example, currently covering most of the southern half of Australia. I also don’t think you can say AGW causes only a rise in temp, it does seem to cause other effects (Like drought in southern Australia).

    “Many of the wild fires in California are man made.”
    Same in Australia, but the conditions to make the fire catastrophic (like in Melbourne this year) were entirely temperature dependent. Those fires certianly wouldn;t have been so bad without the setup caused by increased air temp, wind, and drought.


  16. Nathan,

    At the scale of Southern Australia it is quite clear that no one is capable of attributing these types of conditions to global warming. The models just aren’t that good yet.


  17. “However to the extent that drought can be connected to GW then that could be quite a significant factor.

    I remain quite skeptical of the ability to connect local conditions like drought with changes in global temperature. More logical is to connect things like smaller snow packs, or reduced glaciers, but it is hard to see a wildfire connection there.”

    If this is merely saying that a specific local condition is very difficult to directly attribute to global warming then yes.

    However, I think the relationship where global warming means more water vapour can be held in atmosphere and limits on the rate at which heat can be lost to space are well established. The effects being that there are more extreme heavy downpours and more droughts are reasonably well accepted.

    So attributing one specific drought to global warming is very difficult but the idea that more of some types of drought and heavy downpour events are expected to become more common is reasonably well accepted.

    Myles Allen, William Ingram & David Stainforth, Constraints on future changes in climate and the hydrologic cycle, Nature Insight article, Nature, 419, pp.224-232, 12 September 2002.

    Click to access nature_insight_120902.pdf


  18. Hi William,

    Another report for you:

    http://www.cmar.csiro.au/e-print/open/hennessykj_2005b.pdf (PDF 2MB)

    This one, from CSIRO here in Oz, assesses the changes resulting in Fire Danger Indices (FDIs), used in the preparation of fire weather forecasts. FDIs take into account temperature, wind, vegetation dryness (called curing) ect…. It also assesses the likely impact arising from changes in frequency and magnitude of extreme weather conditions (something I’m a little more sceptical towards than changes in mean conditions).

    The gist of it all is that fire danger is likely to increase in Oz with climate change, mean FDIs are projected to be higher and the frequency of extremes are likely to increase… and as we found to our detriment in February, extreme events are where the real danger lies… (see above note about extreme event prediction).


  19. I live in Eastern Australia and would like to add some observations of my own – whether the extreme heat waves that fuelled last summers firstorms can be related to AGW or not (I think they were), fire risk management is affected.
    I believe that cool season fuel reduction burning is hindered by rising minimum overnight temperatures; the old practice around here was to light up during coolish weather with the expectation that nights would be cool enough for dew to form and the fires would self extinguish. Warmer overnight temperatures of even one degree can mean large areas simply don’t get cool enough for dew to form and opportunities to ‘cool’ burn to reduce fuel loads are reduced. Burning only within cleared firebreaks simply isn’t possible for much inaccessible bushland and experienced help can be spread too thinly during the short windows of opportunity to do so safely; a reasonable reluctance to light fires that could easily break containment leads to fuel reduction burning not happening at all and consequently fuel loads keep growing. (For those who don’t live in the fire adapted Australian bush, please note that leaf litter and vegetation breaks down slower than it builds up and it’s mostly extremely flammable, even without prolonged dry weather).

    [OK, but notice that even this is a moisure-mediated effect -W]


  20. I think you’re on a looser trying to treat heat and moisture independently 🙂

    Fires depends on flammable material and enough initial heat to get things burning (ignition) … plus oxygen. It’s a “no-brainer” that it doesn’t matter how hot it is, you wont have a fire if you have no combustible material. Conversely, fire, as you know, is the number one risk in the Antarctic … where it’s a tad cold … but things are rather burnacious around human habitations …

    You know all that stuff. You’ve just raised a strawman, but most of the comments above, explicitly or implicitly invoke both factors, and that’s because that’s the real world.

    No one believes 32C will make the ocean burn more than it would at 30C. Everyone believes that the number of *dry* 32C days over the number of 30C days will lead to more combustible material in a lot of situations … ergo … AGW likely, in some places, increases the risk of fire. In those places with different vegetation (or no vegetation) or more precip, maybe not …


  21. William, higher overnight minimum temperatures outside the high fire danger season may be an indirect influence but I think it does have an impact. Dew formation is moisture and temperature mediated – higher overnight minimums will directly influence dew formation as will reduced humidity. Both appear to be changing due to climate change. Changed fuel reduction practices are contributing factors to wildfires. Of course 50 degree celsius on top of prolonged dry conditions and strong winds will be ideal for firestorms and once underway even areas which have had understory recently burned will have firestorms through Eucalypt forest canopies. Are such heatwaves AGW related? Sure we’ve had heatwaves in the past and even record ones, but isn’t a warming trend making hotter extremes a bit more likely and cold extremes a bit less likely? The heatwaves through south east Australia broke all records by big margins. It would be comforting to think that they are one-off weather events but the previous records broken were the previous year – not exactly 1 in 100 year events when they happen twice in 2 years. The frequency of extreme weather events is probably very hard to fix to clear climate change related effects but I can’t help but think these are being influenced by global warming.

    [Broadly speaking I think you are right, but perhaps your “hotter extremes a bit more likely and cold extremes a bit less likely? ” is the key, and is subtly wrong, as I’ve been trying to say. In fact it makes dryer more likely and wetter less so, and that is the key point. Linked to that is: GW makes the world warmer on average – this is clear – but it doesn’t need to make it dryer on average – indeed, it will likely make it wetter. So the dryer bit is harder to work out. I’m not claiming this is any major flaw or insight: just that the straightforward: warmer => more fire is not correct -W]


  22. I’m probably thinking closer to home where regional impacts on climate are thought to include less precipitation – at least that has been the trend for SE Australia and I’ve read that modelling supports that. Elsewhere even in Australia there has been more rain – but my impression is it’s been from relatively few extreme events but I’d need to check Bureau of Meteorology records. If that latter is correct that increased rainfall doesn’t necessarily mean less likelihood of wildfires as rapid vegetation growth after storms and flooding will almost certainly be followed by ‘normal’ seasonal dry conditions; increased fuel loads will lead to increased fire risk. Not directly related to temperature of course.
    I do appreciate that warmer temperatures alone would affect fire frequency and severity differently according to regional and local factors and warmer temperatures causing more wildfires isn’t universal or even more likely on balance over the globe.


  23. “In fact it makes dryer more likely and wetter less so, and that is the key point. Linked to that is: GW makes the world warmer on average – this is clear – but it doesn’t need to make it dryer on average – indeed, it will likely make it wetter.”

    William, I think you need to look more closely at this statement and think about how fire regimes work. It seems a lot of people her pick Australia as an example, and I think it is the best example. My understanding is that we expect (from AGW) less rainfall and more hot days in the south, and more hot days and more rainfall in the north. Both are expected to lead to more fires, becuase fires happen in particular seasons (I think this is outlined in that earlier CSIRO report I linked to). It may also be necessary to look at where rainfall deficits occur. Extra rainfall in the north will happen in the wet season (when fires don’t normally happen), and extra heat in the dry season (when fires do happen). The drying and warming of the south has already led to worse bushfires. I guess in summary, in Australia, a wetter world doesn’t decrease the chance of fire as the fire season won’t get more rain.

    [That is entirely possible. What I’m trying to say – clearly with a great lack of success – is that if “warmer => more fires” then “GW => more fires” is easy to say and is indeed often said. However, since it is vastly more complicated than that, “GW => more files” isn’t easy to demonstrate at all. “more fires in Oz” perhaps, or perhaps S Oz, or perhaps more for a while and then less, or… -W]


  24. I had never really given any thought to the connection between wildfires and global warming. After reading this article I kind of thought it made sense. If the earth is warming then everything with become dryer and cause fires. It really is an interesting concept to think about.


  25. No to harp on about this, but here’s a better discussion of the recent heat in Australia

    It doesn’t matter, the US is colder than normal.

    Therefore climate science is proven false. WUWT tells us this two or three times a week.

    (I’m being cynical …)


  26. Dhogaza

    That’s right! I forgot it all cancels out…

    Still, was the hottest November day on record in Adelaide today (and most of South Australia… and western Victoria… and western New South Wales) at 42.8C! I think it’s probably all due the UHI in the unoccupied desert.


  27. Re: #2.

    The point might be right but don’t you mean the Arrhenius equation? In that case the dominant increase with temperature would be carried by the Boltzmann factor exp(-const/T).


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s