RealClimate logo

Going to extremes

Filed under: — gavin @ 17 February 2011

There are two new papers in Nature this week that go right to the heart of the conversation about extreme events and their potential relationship to climate change. This is a complex issue, and one not well-suited to soundbite quotes and headlines, and so we’ll try and give a flavour of what the issues are and what new directions these new papers are pointing towards.

Let’s start with some very basic, but oft-confused points:

  • Not all extremes are the same. Discussions of ‘changes in extremes’ in general without specifying exactly what is being discussed are meaningless. A tornado is an extreme event, but one whose causes, sensitivity to change and impacts have nothing to do with those related to an ice storm, or a heat wave or cold air outbreak or a drought.
  • There is no theory or result that indicates that climate change increases extremes in general. This is a corollary of the previous statement – each kind of extreme needs to be looked at specifically – and often regionally as well.
  • Some extremes will become more common in future (and some less so). We will discuss the specifics below.
  • Attribution of extremes is hard. There are limited observational data to start with, insufficient testing of climate model simulations of extremes, and (so far) limited assessment of model projections.

The two new papers deal with the attribution of a single flood event (Pall et al), and the attribution of increased intensity of rainfall across the Northern Hemisphere (Min et al). While these issues are linked, they are quite distinct, and the two approaches are very different too.

The aim of the Pall et al paper was to examine a specific event – floods in the UK in Oct/Nov 2000. Normally, with a single event there isn’t enough information to do any attribution, but Pall et al set up a very large ensemble of runs starting from roughly the same initial conditions to see how often the flooding event occurred. Note that flooding was defined as more than just intense rainfall – the authors tracked runoff and streamflow as part of their modelled setup. Then they repeated the same experiments with pre-industrial conditions (less CO2 and cooler temperatures). If the amount of times a flooding event would occur increased in the present-day setup, you can estimate how much more likely the event would have been because of climate change. The results gave varying numbers but in nine out of ten cases the chance increased by more than 20%, and in two out of three cases by more than 90%. This kind of fractional attribution (if an event is 50% more likely with anthropogenic effects, that implies it is 33% attributable) has been applied also to the 2003 European heatwave, and will undoubtedly be applied more often in future. One neat and interesting feature of these experiments was that they used the set up to harness the power of the public’s idle screensaver time.

The second paper is a more standard detection and attribution study. By looking at the signatures of climate change in precipitation intensity and comparing that to the internal variability and the observation, the researchers conclude that the probability of intense precipitation on any given day has increased by 7 percent over the last 50 years – well outside the bounds of natural variability. This is a result that has been suggested before (i.e. in the IPCC report (Groisman et al, 2005), but this was the first proper attribution study (as far as I know). The signal seen in the data though, while coherent and similar to that seen in the models, was consistently larger, perhaps indicating the models are not sensitive enough, though the El Niño of 1997/8 may have had an outsize effect.

Both papers were submitted in March last year, prior to the 2010 floods in Pakistan, Australia, Brazil or the Philippines, and so did not deal with any of the data or issues associated with those floods. However, while questions of attribution come up whenever something weird happens to the weather, these papers demonstrate clearly that the instant pop-attributions we are always being asked for are just not very sensible. It takes an enormous amount of work to do these kinds of tests, and they just can’t be done instantly. As they are done more often though, we will develop a better sense for the kinds of events that we can say something about, and those we can’t.

178 Responses to “Going to extremes”

  1. 151

    Dan H 142: Is it a coincidence that the highest yield was observed in 1998?

    BPL: Sigh. Yes, Dan, it almost certainly is. Please read an introductory statistics book.

  2. 152
    Brian Dodge says:

    “But since the peak temperatures have changed much less than the average, the agricultural impact would be much less. As I stated previously.”
    [Response: Stating it doesn’t make it so. You would need to actually look at the data and the projections. – gavin]
    Dan H. — 24 Feb 2011 @ 11:08 AM

    I’m happy to oblige-

    “Here we report that annual mean maximum and minimum temperatures have increased by 0.35°C and 1.13°C, respectively, for the period 1979–2003 and a close linkage between rice grain yield and mean minimum temperature during the dry cropping season (January to April).”

    Looks like Dan nailed it, but –

    “Grain yield declined by 10% for each 1°C increase in growing-season minimum temperature in the dry season, whereas the effect of maximum temperature on crop yield was insignificant. ” ibid.

    “…the agricultural impact would be much less…” Well, much less rice, anyway.

    “Surprisingly, observed variations in average growing-season temperatures of ±2 °C in the main wheat growing regions of Australia can cause reductions in grain production of up to 50%. Most of this can be attributed to increased leaf senescence as a result of temperatures >34 °C.”

    “We conclude that the recipe for a world record includes a combination of cultivar and sowing date that will lead to grain growing through the solar radiation peak, a cool but sunny summer, and attention to agronomic detail so that no growth constraints apply.” (Leibigs Law)
    Future headline – “Lindzen Iris Effect predicted to reduce crop yields, by limiting Solar Radiation during critical growing period!!!” &:>)

    “Is it a coincidence that the highest yield was observed in 1998?”
    Comment by Dan H. — 24 Feb 2011 @ 2:13 PM

    according to Agricultural Outlook/August 1998, Economic Research Service/USDA
    “U.S. winter wheat plantings were down from a year earlier, suggesting a smaller crop in 1998. However, generally favorable weather, especially during harvest, will boost the winter wheat yield to a record 46.6 bushels per acre.” Less productive land was left fallow, which also drives up yield/ha.
    “..average protein content of hard red winter (HRW) wheat is reportedly below normal…” one of the observed “benefits” of more CO2, eh? I say, let them eat cake! (for those who don’t cook, pastry flour is lower in protein than bread flour).

    I downloaded wheat production and temperature stats from
    98-99 wheat crop yield was fourth out of the 6 year period 98-04, so although global T anomalies were high in 98, yields were not globally higher. in Oz, the three most productive years averaged 0.22 deg C temperature anomaly, the three least productive years averaged 0.40 deg C T anomaly.

  3. 153
    Ray Ladbury says:

    Dan H., Near as I can tell, people arrive at opinions in two ways–science and making crap up. Why not try science.

  4. 154
    Tim Joslin says:


    The Pall et al paper relates directly to a project I’m working on, so I’ve looked at it further since my initial comments on this thread. I noted in my #73 that the paper does not include evidence of the skill of the model in forecasting flooding in a particular year, i.e. the extent to which it reproduces the actual conditions in autumn 2000. It also fails to establish the bias of the model, i.e. how the flood threshold chosen in the modelling exercise (0.41mm/day runoff) relates to real world flooding levels.

    It now seems to me that the reason Pall et al did not attempt to quantify their model’s skill and bias is because they thought they didn’t need to.

    They write that:

    “Crucially, however, most runoff occurrence frequency curves in Fig 3 remain approximately linear over a range of extreme values, so our FAR estimate would be consistent over a range of bias-corrected flood values.”

    But Fig 3 has a logarithmic scale!

    In fact, a simple inspection of Fig 3 suggests to me that the increased risk of flooding due to AGW (their FAR or fraction of anthropogenic risk) is quite sensitive to the bias of their model suite.

    Do you agree?

    I’ve discussed this issue in more detail on my own blog.

  5. 155
    wili says:

    Well, it’s happened. A long time denialist on another blog has linked to this thread to claim that no extremes in weather can ever be connected to GW (which, of course, he doesn’t believe in anyway). This guy is a past master at misrepresenting articles and cherry picking data. Of course, other than this case, he describes RC as a completely biased, unscientific, unreliable, alarmist site. Thank you for not letting “what they might say” get in the way of presenting the clearest, most accurate account of unfolding events and science as you can. No matter what we say, they will always attempt to misrepresent and distort it one way or the other. We can’t let fears of how intentional mis-informers might warp our positions stop us from having these important discussions.

    Meanwhile, can we nominate posters for the borehole? If so, may I humbly nominate DanH. He seems to be the latest in a long series of people who manage to come here and hijack the thread so that all posters are responding to his idiocy rather than discussing the very real and enormously important issues, issues that are otherwise clearly and thoroughly presented and discussed only here and in a very few other locations.

    In the mean time, can we all resist the temptation to constantly feed such trolls?

    It only encourages them.

  6. 156

    Dan H. #142:

    One of this country’s staples is wheat production. Cehck out the Kansas wheat yield since 1979. Is it a coincidence that the highest yield was observed in 1998?

    Coincidence. 2003 was also a good year and the continental US was unusually cool that year. I haven’t tried correlating the data since it’s in the form of graphs in the PDF you pointed to and I’m not going to look up usable data. If you want to make a case do the stats yourself.

    wili #155: I agree with the general view that you don’t feed trolls but while they are still posting here well-considered rebuttals can be read by others who actually want to be informed. Remember always when you talk here, it’s not a private conversation like email (oops, that’s not private either).

  7. 157
    Joseph Sobry says:

    I guess we can conclude this whole blog with the old tired joke that when, for one hour, I have one leg in the deep freezer and the other leg in the oven I should on average be quite comfortable. Except that at the end of one hour I will have to have both legs amputated. Get a grip.

  8. 158


    Even if it were not such an old joke, the humor quotient would be rather low–poor old Eastern Australia is still rather in mind.

  9. 159

    Not much word on recovery that I could find with a quick search. But crop losses are estimated at a billion and a half.

  10. 160
    adelady says:

    Kevin, I thought your billion and a half was a bit low. They say that’s only for Yasi, but the wording makes me think it’s also for the SE Queensland flooding.

    So it doesn’t account for grain crops lost and damaged throughout NSW, Vic and SA. The big issue for this in light of the Russian and Chinese wheat harvest problems is that the moisture encouraged swelling and sprouting before the heads could be harvested. So the decreased value reflects the fact that these damaged crops are not just lower quality, they’re suitable only for animal feed. So the apparent “harvest” is quite a bit less than you might think because a large part is not suitable for flour making.

    And I’ve seen pictures of Vic vineyards with flood debris hanging off the wires – a month before harvest. Not an encouraging prospect for crushing if some of that filth stays stuck between the berries.

    I think the total losses for this summer’s mayhem will probably be more than some of the worst drought years. (At least in a drought year you don’t waste money on seeds, cultivation, fertilisers and the rest when you =know= no crop is possible.)

  11. 161
    Pekka Kostamo says:

    #123 Hank:

    Hmph. Current global average is +14 C.

    Mitigation target by 2100 is +16 C. Not many climatologists are confident it will be reached.

    Probable by 2100 is +18 C (i.e. UK Gov. “prepare for” statement, considering achievable political realities)

    Seems there are many climatologists who hold +20 C quite possible in a somewhat longer time frame.

    And then they say +24 C can not be excluded as a long term equilibrium temperature. Average global temperature, that is.

    My grandson, recently born, and his pals may well experience the year 2100 first hand. Will it be a happy time?

  12. 162
    Sou says:

    @ adelady #160. Don’t forget to factor in the fungal and other disease outbreaks in grapes and other fruit, particularly stone fruit, as a result of the unseasonal excessive humidity this summer. It’s not just the extensive areas under actual flood that suffered.

    Plus the costs of not being able to travel the most direct route to almost anywhere in Victoria and many places in Queensland. With so many roads closed for extended periods, the cost of travel and transport must have gone up quite a lot.

  13. 163

    #160–It may be too low for all I know; I can’t claim the estimate as in any real sense ‘mine’–just a figure that I found in a couple of minutes of searching. It did say “floods,” but with no more specificity than that. And the sources are pretty murky, too–“estimates from [unnamed] grower groups.” So how comprehensive the estimates are remains an open question.

    It’s clear, though, that the number is purely for *crop losses*: “Damage to machinery, infrastructure and processing factories will push the total damage bill higher.” (And this number, too, is presumably specific to growers, and would exclude, say, municipal infrastructure.)

  14. 164
    Hank Roberts says:

    Pekka, I agree. The change coming is enormous; whatsisname back there tried to pretend it’s not much by comparing before and after to a baseline zero Kelvin.

  15. 165
    wili says:

    On the topic of extremes, we still seem to be right on the edge of record ice extent lows for this time of year, just barely above the level of ’06.

    Of course, this year’s ice is much thinner and much more ‘rotten’ than that of five years ago.

    Given all this, this late in the year, is there any chance that we WON’T set a new record low extent (and of course total volume) this September? Can we rule out a major recovery this late in the season?

    And on possible causes of future extremes–is there any more new on the seabed methane front? A Shakhova and Semiletov slide presentation “Methane release from the East Siberian Arctic Shelf and the Potential for Abrupt Climate Change” from December circulating on the web mentions that ‘prorating’ exact measurements from one location to all other known ‘hot spots’ of methane release in the East Siberian Arctic Shelf would give a figure of 3.5 Gt methane release going on right now. But other sources talk about something more like 7 Mt of gradual release currently occurring. Are these merely upper and lower bounds of possible release? The upper one seems pretty scary. Is prorating in this case not intended as an actual prediction but rather a kind of ‘what if’ imaginative exercise?

    From the slide presentation (…/1A_Shakhova_Final.pdf)

    “Interpretation of acoustical data recorded with deployed multibeam sonar allowed moderate quantification of bottom fluxes as high as 44 g/m2/d (Leifer et al., in preparation). Prorating these numbers to the areas of hot spots (210×103 km2) adds 3.5Gt to annual methane release from the ESAS. This is enough to trigger abrupt climate change (Archer, 2005)” Figures seem to be taken from Leifer et alia (2009-in preperation).

  16. 166
    wili says:

    Sorry, this might be a better link to the symposium:

  17. 167

    #165–“Given all this, this late in the year, is there any chance that we WON’T set a new record low extent (and of course total volume) this September? Can we rule out a major recovery this late in the season?”

    Yes, and no. The variability means that early prediction is very tough–the long term is dominated by melt, but the short term is dominated by wind, more than anything else–though it’s not the only highly significant weather variable. (Cloudiness can have a big, big impact, to name just one.)

    All that said, it is (or should be) food for thought just how low the current (possibly annual maximum) extent is. (Of course, ice volume is most likely at an all-time low for this time of year. Bring on the Cryosat data–!) And one of the regulars at Neven’s sea-ice blog just calculated that the February mean was a whisker *below* ’06.

  18. 168
    Joseph Sobry says:

    #158 Kevin, I was afraid I would be misunderstood and I was.
    The joke was only directed at people who think that a small change in globally averaged temperatures can not have any serious consequences.
    I only tried to point out that a benign (comfortable) looking average may be computed from a set of data with some really extreme and horrendous data points.
    The whole point is that a small shift in global average temperatures may well have some extremes in local times, places and various aspects of the weather with consequences that we will not like. To wit all the heat waves, floods and cold spells (currently ongoing here locally), that we have seen in the last few years or so.
    The consequences will not only be in the local climate but also in many other areas such as the economy the environment etc. etc. I do not know what the final tally will be in Australia, Pakistan and Russia just to name a few. Perhaps we should ask Mr. Lomborg how much damage we can suffer and still increase the wealth so we can take on global warming in a few decades.

  19. 169
    Hank Roberts says:

    “Natalia Shakhova1, Igor Semiletov
    Methane release from the East Siberian Arctic Shelf and the Potential for Abrupt Climate Change

    Specific features of the Arctic shelf methane hydrates
    • More sensitive to warming:because only 1/3 of energy required to convert deep ocean hydrates to free gas (54.2 kJ/mol) requires to convert to free gas Arctic hydrates (18,2 kJ/mol);
    • More vulnerable because they have naturally been experiencing warming by as
    much as 17˚C while deep oceanic hydrates were warmed by less than 1˚C;
    • More significant in their accumulations:
    because their spatial concentration is many folds greater as well as
    pore occupancy (20-100% vs. 1-2% of deep oceanic hydrates);
    • More potential for abrupt releases …

  20. 170
    Hank Roberts says:

    Oh, and that Arctic permafrost? Here’s what’s happening to warm it:

    ” Snow and ice help control how much of the Sun’s energy Earth soaks up. Bright white snow and ice reflect energy back to space. Because that energy does not get absorbed, it does not go into Earth’s climate. As a result, snow and ice cool the planet. This effect is called a climate forcing because snow and ice directly influence the climate. …

    The image shows how the energy being reflected from the cryosphere has changed between 1979 and 2008. When snow and ice disappear, they are replaced by dark land or ocean, both of which absorb energy. The image shows that the Northern Hemisphere is absorbing more energy, particularly along the outer edges of the Arctic Ocean, where sea ice has disappeared, and in the mountains of Central Asia.

    “On average, the Northern Hemisphere now absorbs about 100 PetaWatts more solar energy because of changes in snow and ice cover,” says Flanner. “To put it in perspective, 100 PetaWatts is seven-fold greater than all the energy humans use in a year.” Changes in the extent and timing of snow cover account for about half of the change, while melting sea ice accounts for the other half.

    Flanner and his colleagues made both calculations by compiling field measurements and satellite observations from the Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Very High Resolution Radiometer, and Nimbus-7 and DMSP SSM/I passive microwave data. The analysis is the first calculation of how much the energy the entire cryosphere reflects. It is also the first observation of changes in reflected energy because of changes in the entire cryosphere.

  21. 171
    wili says:

    Thanks Kev and Hank.

    Sounds to me as though the death rattle (Cheyne-Stokes) has already set in. Time to inform the next of kin?

  22. 172

    #168–I got your point, and it’s a good one. Sorry for a response that perhaps came off more humorlessly than I intended.

    “Ask Mr. Lomborg. . . ?” Ah, you are a joker!

  23. 173
    guthrie says:

    Hank # 170 – hmm could it be that Arctic permafrost turning out like that is a positive feedback? Leading to warming greater than that caused simply by excess CO2?

  24. 174
    Dan H. says:

    I would not read too much into the recent dip in Arctic sea ice quite yet (I assume you are looking at the same DMI data I am). While we are currently barely above the low maximum of 2006, you should note that 2006 had the highest minimum of the past few years, resulting in one of the smallest overall melts from winter to summer. Conversely, 2008 saw about a million more square km of ice in the winter and a million less in the summer.
    The Arctic Ocean was quite warm through the end of 2010, but has cooled down in January, as has the Northern Pacific and Atlantic. The sudden decrease may just be a blip due to winds or circulation patterns, opposite to the spring jump last year. Changes in Arctic Ocean SST should give an indication of thingfs to come.

  25. 175
    John E. Pearson says:

    Hank posted this in 170.

    Anybody know why the forcing increased in the US southwest?

  26. 176
    wili says:

    Guthrie, it is definitely a feedback. That is why it is alarming, even if it is at much lower levels than the presumably upper-end estimate of 3.5 Gt/year of methane mentioned in the slide show (“The Bad News” slide toward the end).

    It seems our resident troll continues to have trouble reading, or perhaps is just addicted to picking convenient cherries. To repeat myself:

    Of course, this year’s ice is much thinner and much more ‘rotten’ than that of five years ago.

  27. 177

    Wili, thanks in turn for the link to the RS impacts issue. I’ll be perusing that. As to “Cheynes-Stokes” (or not), I suppose that there’s a definitional problem there, analogies being what they are. I wouldn’t claim we’re there yet; but I suspect, FWIW, that the demise of the Arctic icecap is all but unavoidable at this point.

    Just when it can be expected is another question, with recent projections/predictions (made with varying degrees of professionalism) ranging from the next couple of years out to about 2050. Also FWIW, I expect to see it, but I don’t necessarily expect to live to 2050.

  28. 178


    “Know?” Can’t claim that. But I’d presume that persistent drought in the southwest in recent years is part of the picture.

    Better informed thoughts?

    On southwestern drought and future water supply: