Nature has an interesting editorial this week on the state of the science for attributing extreme events. This was prompted by a workshop in Oxford last week where, presumably, strategies, observations and results were discussed by a collection of scientists interested in the topic (including Myles Allen, Peter Stott and other familiar names). Rather less usual was a discussion, referred to in the Nature piece, on whether the whole endeavour was scientifically worthwhile, and even if it was, whether it was of any use to anyone. The proponents of the ‘unscientific and pointless’ school of thought were not named and so one can’t immediately engage with them directly, but nonetheless the question is worthy of a discussion.
This workshop was a follow-up to one held in 2009, which took place in a very different environment. The meeting report was typical of a project that was just getting off the ground – lots of potential, some hints of success. Today, there is a much richer literature on the topic, and multiple approaches have been tried to generate the statistical sample required for statements of fractional attribution.
But rather than focus on the mechanics for doing this attribution, the Nature editorial raises more fundamental questions:
One critic argued that, given the insufficient observational data and the coarse and mathematically far-from-perfect climate models used to generate attribution claims, they are unjustifiably speculative, basically unverifiable and better not made at all. And even if event attribution were reliable, another speaker added, the notion that it is useful for any section of society is unproven.
Both critics have a point, but their pessimistic conclusion — that climate attribution is a non-starter — is too harsh.
Nature goes on to say:
It is more difficult to make the case for ‘usefulness’. None of the industry and government experts at the workshop could think of any concrete example in which an attribution might inform business or political decision-making. Especially in poor countries, the losses arising from extreme weather have often as much to do with poverty, poor health and government corruption as with a change in climate.
Do the critics (and Nature sort-of) have a point? Let’s take the utility argument first (since if there is no utility in doing something, the potentially speculative nature of the analysis is moot). It is obviously the case that people are curious about this issue: I never get as many media calls as in the wake of an extreme weather event of some sort. And the argument for science merely as a response to human curiosity about the world is a strong one. But I think one can easily do better. We discussed a few weeks ago how extreme event attribution via threshold analysis or absolute metrics reflected a view of what was most impactful. Given that impacts generally increase very non-linearly with the size/magnitude of an event, changes in extremes frequency or intensity have an oversized influence on costs. And if these changes can be laid at the feet of specific climate drivers, then they can certainly add to the costs of business-as-usual scenarios which are then often compared to the cost of mitigation. Therefore improved attribution of shifts in extremes (in whatever direction) have the potential to change cost-benefit calculations and thus policy directions.
Additionally, since we are committed to certain amount of additional warming regardless of future trends in emissions, knowing what is likely in store in terms of changing extremes and their impacts, feeds in directly to what investments in adaptation are sensible. Of course, if cost-effective investments in resilience are not being made even for the climate that we have (as in many parts of the developing world), changes to calculations for a climate changed world are of lesser impact. But there are many places where investments are being made to hedge against climate changes, and the utility is clearer there.
Just based on these three points, the question of utility would therefore seem to be settled. If reliable attributions can be made, this will be of direct practical use for both mitigation strategies and adaptation, as well as providing answers to persistent questions from the public at large.
Thus the question of whether reliable attributions can be made is salient. All of the methodologies to do this rely on some kind of surrogate for the statistical sampling that one can’t do in the real world for unique or infrequent events (or classes of events). The surrogate is often specific climate simulations for the event with and without some driver, or an extension of the sampling in time or space for similar events. Because of the rarity of the events, the statistical samples need to be large, which can be difficult to achieve.
For the largest-scale extremes, such as heat waves (or days above 90ºF etc), multiple methodologies – via observations, coupled simulations, targeted simulations – indicate that the odds of heat waves have shortened (and odds for cold snaps have decreased). In such cases, the attributions are increasingly reliable and robust. For extremes that lend themselves to good statistics – such as the increasing intensity of precipitation – there is also a good coherence between observations and models. So claims that there is some intrinsic reason why extremes cannot be reliably attributed doesn’t hold water.
It is clearly the case that for some extremes – tornadoes or ice storms come to mind – the modelling has not progressed to the point where direct connections between the conditions that give rise to the events and climate change have been made (let alone the direct calculation of such statistics within models). But in-between these extreme extremes, there are plenty of interesting intermediate kinds of extremes (whose spatial and temporal scales are within the scope of current models) where it is simply the case that the work has not yet been done to evaluate whether models suggest a potential for change.
For instance, it is only this year that sufficient high frequency output has been generically archived for the main climate models to permit a multi-model ensemble of extreme events and their change in time – and with sufficient models and sufficient ensemble members, these statistics should be robust in many instances. As of now, this resource has barely been tapped and it is premature to declare that the mainstream models are not fit for this purpose until someone has actually looked.
Overall, I am surprised that neither Nature or (some?) attendees at the workshop could find good arguments supporting the utility of attribution of extremes – as this gets better these attributions will surely become part of the standard assessments of impacts to be avoided by mitigation, or moderated by adaptation. We certainly could be doing a better job in analysing the data we have already in hand to explore whether and to what extent models can be used for what kinds of extremes, but it is wrong to say that such attempts are per se ‘unverifiable’. As to whether we are better off having started down this path, I think the answer is yes, but this is a nascent field and many different approaches and framings are still vying for attention. Whether this brings on any significant changes in policy remains to be seen, but the science itself is at an exciting point.
Professional publicists find hype attractive because human attention is hardwired to attraction by extremes.
The distasteful fact is that is equally true on all sides. Regardless of how much is said about signal-to-noise ratios and framing all sides in the climate debate lust after contrast.
PBS pundits, Beeb reporters and scientific court historians all love inflanmatory quotes no less than Murdoch’s minions, and Nature cover designers would be as remiss as political spot producers if they failed to keep up with the eye catching state of the art of advertising.
Captcha certainly will reduce spam when it includes extended scientific notion character set.
Great idea !
Laplace:
“The most important questions of life are indeed, for the most part, really only problems of probability.”
I rarely have an opinion since I lack the scientific credentials, but this question has a wider social/political angle. Modeling that can give a probability of the increase over baseline of unusual events over time should most certainly interest policy makers and money matters. As the models improve the geographic locale will narrow from global to e.g a continent.
“… the odds of heat waves have shortened (and odds for cold snaps have decreased)”
Perhaps a clearer statement would be that odds of heat waves have increased, while odds of cold snaps have decreased?
The quoted text mixes directions, as “shortened” and “decreased” appear at first blush to say the same thing…
Having been at the meeting, I think that most of the skepticism was directed at the pseudo-real-time aspect of extreme event attribution. The ‘users’ (apart from perhaps the media and lawyers) did not identify a valuable use for the attribution of specific individual events (such as the EU or Russian heatwave). The vast majority thought that the attribution of classes of extreme events (e.g. increase in 90th percentile hot days) is a valuable scientific endeavour.
Ed.
[Response: Ed, thanks. I don’t see a huge distinction between the notion of fractional attribution of a single event and the attribution of a class of events very like the single event in question (though I appreciate that I might not have thought about this particularly deeply). Obviously, one can envisage singular events that have attributable singular causes (i.e. a dramatic cool spell after a huge tropical eruption), but that doesn’t seem particularly relevant for the general class of things that people have looked at. – gavin]
[Response: A further thought. The real-time attribution part goes much more to the role of science in answering questions that the public has. And as I mention, extreme events provoke a big increase in questions. To the extent that a public service must, in some sense, serve the public, it is useful to have near-real time analysis while people are interested. Not all scientists have this mission, but those working for a putative ‘Climate services’ dept. might well have…. – gavin]
As far as I know, the utility point seems to be more psychological than rational. Yes, most adaptation policies are or should be based on projections, but the attribution story makes the projections easier to convey. I think it links to a fundamental way of learning – the same why everybody wants to know how and why an accident happened so that we can avoid it in the future.
As to the quality of the models, Nature follows the argument that was made by many speakers that an attribution study should include a verification section showing that the model represents the relevant mechanisms well enough. The does not follow immediately from the right resolution; the relevant mechanisms should also be shown to be simulated well enough. This is not necessarily the case for current models for all interesting extremes…
[Response: Sure, but it also isn’t necessarily true that it is never the case. – gavin]
For members of the public, and hence policymakers, a statement like “event X that we have seen is largely attributable to global warming” is actually more meaningful, in terms of plans for future action, than “events of class X are largely attributable…” etc. A good story tends to trump general statements, except maybe for scientists like us.
I am reminded of this story.
“In the history of ideas, there are examples of questions being answered that had earlier been judged forever out of science’s reach. In 1835 the celebrated French philosopher Auguste Comte wrote, of the stars: ‘We shall never be able to study, by any method, their chemical composition or their mineralogical structure.’ Yet even before Comte had set down these words, Fraunhofer had begun using his spectroscope to analyse the chemical composition of the sun. Now spectroscopists daily confound Comte’s agnosticism with their long-distance analyses of the precise chemical composition of even distant stars.” – Richard Dawkins
Those who say science cannot know are usually wrong. Even as people say we cannot attribute extreme weather events to climate change, some already are being attributed. We’ve got the sun, now let’s reach for the stars.
1) I am not a climate scientist –just a common citizen — but I think Nature has got it completely backwards. There is already a climate model in existence that is being used to design our infrastructure — NOAA Atlas 14 –and that model grossly underestimates rainfall in US coastal areas.
2) As a result, real estate developers are getting EPA/state/local government approvals to lay down massive amounts of impervious cover (roads, parking lots etc) and building stormwater systems –which have an operational live of a 100 years — that are inadequate the very day on which they become operational. People die in floods, you know.
3) Plus NOAA’s precipitation data is updated about every 40 years or so. Atlas 14 — dated 2006 — only applies to some states. Technical Paper 40 ( 1961) is still the standard precipitation data source for many states:
http://www.nws.noaa.gov/oh/hdsc/currentpf.htm
4) The taxpayers fund science so that it can contribute knowledge and wisdom to solution of our national problems. If the only thing you do is confirm the obvious 30 years hence, then why should the voters fund your hobby?
5) Do you people agree with NOAA’s Atlas 14 , Volume 2? If not, what are you doing to get it corrected?
If scientists refuse to answer public questions about attribution of specific extreme weather events to climate change on the grounds that those questions are poorly posed, then others, less constrained by knowledge or scruples, will step in and provide the answers.
It’s admittedly difficult, maybe impossible, to answer such questions with the rigour and confidence that scientists usually insist upon. And, when it comes to probabilities and uncertainties, communication with the general public is fraught with pitfalls. That’s all the more reason not to leave attribution commentary to the uninformed.
1) For example, I live in an area subject to flooding about 15 miles west of Philadelphia. The Pennsylvania Turnpike is being widened to 8 lanes and they are using the following NOAA Atlas 14 rainfall data to compute stormwater runoff: 2year storm event: 3.16 inches, 5Y: 3.91 inches, 10Y: 4.57, 25Y: 5.60, 50Y: 6.53, 100Y: 7.63 inches. Basins designed to reduce the 2Y and 5Y peak rates and keep 10Y rate from increasing.
2) Only problem is , we have had 4 storms in August-September over the past 13 years that gave 8 inches, 9.23 inches, 9.4 inches, and 5.13 inches. Three 200 year storms within a decade. August 2011 broke Philadelphia’s 100 year old rainfall record BEFORE Hurricane Irene arrived on August 28 and dumped 8 inches of rain.
3) And it is not just Philadelphia — Atlanta received a massive rainfall in Sept 2009 that NWS characterized as in excess of a 10,000 year storm event.
4) I suspect NOAA’s ocean model is faulty. Or perhaps they have no ocean model — Atlas 14 Volume 2 was based on statistical analysis of weather station observations for the past 100 years and to the best of my knowledge NOAA did not have weather stations sited on the Atlantic Ocean every 10 miles and collecting rainfall data over the past century.
5) NOTE: One problem is that Pennsylvania uses the mean in Atlas 14 rainfall estimates for storm events instead of the upper bound of the 90 percent confidence interval. But I think a problem still exists even if you do that.
5) But I have seen no researchers looking into this anomaly and its causes. Can anyone shed any light on this?
Gavin, I’m surprised you were surprised: this would not be the first time scientists organize a meeting to check that something they have planned or even done is really useful. Science is about knowledge not necessarily utility.
[Response: Indeed. But most of the time it goes without saying (i.e. everyone takes it for granted). – gavin]
More seriously, I’m also somewhat surprised by your response about mitigation and adaptation. It looks as if you consider that attribution to anthropogenic forcings is a prerequisite or a synonymous for reliable 21st century projections. Is it that obvious?
[Response: No, I didn’t say that. But in assessing future vulnerabilities it is much easier if we are already seeing effects, even though the latter does not preclude the former. – gavin]
Just think about the formerly attributed late 20th century positive NAO trend and look at the recent CMIP5 projections: you might be also surprised. Moreover and more importantly, the lack of successful attribution does not preclude the validity of the projections: just think of the early global warming projections.
If statisticians really want to address people curiosity, they might also think of clarifying our ideas about model reliability and use attribution as an objective tool for better constraining climate projections.
I can say with 100% confidence that if only humanity had increased GHG emissions, then Katrina could have been prevented.
You get the random weather you get. Period. Specific attribution is: change anything significant about the past, and you can prevent not just any event, but all events. Completely different events will take their place.
“We’re gonna get x% more floods” is a good scientific statement.
“Global Warming caused specific event Y” is one of those true by default statements. That wonderful day last week? AGW supplied.
Don Williams,
this is not the site of NOAA, and there is a world with climate and climate science outside of north america. A climate map is not a climate model, and the purpose of science is to gain knowledge und understanding how things work
@ Marcus
What happened to rigorous peer review? Isn’t that the essence of science?
[Response: Sure, but posting extremely specific off-topic issues that only you know anything about on a blog comment thread is not the way to do it. – gavin]
Down on the farm, attributing extreme events is worth $billions or much more. Consider the corn situation right now. The corn belt covers Ohio through Iowa and Missouri through Minnesota. I don’t know how far it goes into Canada. The area is huge and so is the money involved. From Kansas into Saskatchewan they grow wheat. Wheat requires less water. Cotton requires even less water and is called a dryland crop.
Farmers are very aware of the weather and its history, but they don’t speak “Scientific.” If you want to communicate with farmers, you had better talk like a farmer.
An extreme event well worth talking about is the weather we have had for the last 2 years. Missouri corn died early this year because of the drought and was chopped up to make silage. [Silage is “grass or other green fodder compacted and stored in airtight conditions, typically in a silo, without first being dried, and used as animal feed in the winter.”] Silage is definitely not what you want to make out of your corn. This year, there was insufficient rain when it was needed. Last year there was too much rain when dryness was needed at harvest time. The harvest was delayed because the fields were too muddy for a combine. The corn is fine in Minnesota. Illinois land is still a foot short of soil moisture.
Should farmers switch to a dryland crop? Should they install irrigation? Where will irrigation water come from? Should the government subsidize crop insurance for those who don’t switch? How will this affect price and availability of foods? Was the land in Iowa really worth $10,000/acre?
Attributing the change in the rain pattern is absolutely critical when talking to farmers and congressmen. Communicating to them is just as critical for getting enough votes in congress to mitigate GW and to adapt to GW. If RealClimate is about anything, RC has to be about attributing the change in the rain pattern.
Hi Gavin,
Yes – I agree that the greatest demand for real-time attribution comes from the public and the media. Does this mean we *have* to provide answers? This is a tough question.
Quantitative statements can of course be made soon after any event, but by definition these will be less rigorous than a more lengthy look, which may take a year or more and come to a different conclusion based on more evidence. There was a great deal of discussion about the benefits and risks of such an approach.
And, as Geert Jan has said, the model evaluation part is absolutely crucial. For example, evidence was presented that the recent Texas summer heatwave was at least partly due to the preconditioning of a dry spring and hence dry soils – the extent to which the models used for attribution capture this process needs to be assessed to be confident about any conclusions. Could this be done in real-time? I doubt it!
cheers,
Ed.
[Response: If the public have valid science questions, and scientists refuse to answer, you will find the void filled by people who are not scientists for their own purposes. Thus the choice is not between answer or no answer but between answer with scientific backing and answer unhinged from any connection to the facts. Whether you personally should be mandated to do this – well, of course not, but if the government sets up a climate service to provide information to the public about climate, then, yes, they should as best as they are able.
Whether data that exists can be analysed in real-time is not a question of science, but a question of tools. There are enough high frequency output data sets of climate model simulations that a large class of ‘extremes’ come with the purview of the multi-model ensemble. That in practice no-one is able to query them for a specific kind of extreme statistic in real time is an argument for investment in appropriate analysis software, not a statement that it is scientifically invalid. And of course for each analysis thought needs to be given to what the “Level of Scientific Understanding” is for each conclusion. Are you really implying that statistics of heat waves say have to take a year of work for every individual event? Maybe the first or second time it might take a while, but it gets much easier after the first dozen no?
I agree that setting up a PPE for a new target is time consuming if that is what you want to do, but that is not the only way to do this, and it wouldn’t be first choice for most events. – gavin]
I think posters #8 and #11 have a point. Andy S succinctly sums the points I was about to make: witness some of the discussion and media reporting generated in response to the 2012 North American heatwave (caveat, much is still woeful). I’m not suggesting statements are crafted to garner media attention, but in response to climate/weather extremes the average person will look for cues to help interpret such events.
Gavin, thanks for this post. I think this subject is extremely difficult but also important. It is the extreme events that are often most damaging, and if we can develop a probabilistic methodology for attribution, it’s predictive value will far exceed its explanatory value.
Don Williams has an important point–Agencies like NOAA, the Army Corps, etc., and insurance companies need to be making decisions based on the new climate regime we are creating or the result will be unsustainable loss of life and property.
Extreme value statistics is still a relatively young field. However, despite the rarity of events in the tail of the distribution, in some ways, these events may actually offer more insight into how the distribution is changing than those near the mean. It is one time where the Central Limit Theorem isn’t our friend.
It’s the slippery slope argument, I think. Where do you draw the line? I would suggest you draw the line where there is some value of prediction.
Prediction goes hand in hand with understanding attribution. If volcanic eruptions are predicted very well, then thats all well and good.
How about the Texas heat wave? Did the numerical models over predict or under predict the event. Or did they do it after the fact (like predicting a cool ocean in the wake of a hurricane -not really a prediction at all).
An analysis of models verse obs is required for such discussions. Is there any data for Texas Heat Wave?
[Response: Specific prediction longer than a seasonal forecast for a large scale drought like this is impossible, and it’s tough for seasonal forecasters as well. It has very little to do with fractional attribution to specific drivers which is a statistical argument, not a specific one. i.e. what are the odds for event like this and are they changing? Not that global warming caused this event to happen on Tuesday. There are plenty of interesting questions in prediciton of course, but they aren’t really related. – gavin]
The collapse in Arctic sea ice that we are currently witnessing is an extreme event that appears to be validly attributed to anthropogenic climate change as it occurs.
The Nature editorial referred to the Arctic sea ice collapse in the first paragraph before moving to other extreme events like hurricanes and heat waves but what is happening in the Arctic right now seems to answer the question “can we attribute extreme events to climate change in real time?” at least for one form of extreme event.
Geert (@ 7) made the point that “an attribution study should include a verification section showing that the model represents the relevant mechanisms well enough.” That appears to be simple enough to do for the Arctic sea ice in real time given the existing literature and the clear trend in the data.
Let me give it a try….
By studying, and therefore proving, Climate Change, we can ‘inform’ business that, with 70-80% of our developed industrial infrastructure within the zone that said research’s have thus been proven will soon be innundated by the rising seas that said industry (and our lifestyles) is/are responsible for, it’s Great Time to think like Sir Arthur C Clarke, when he was writing “Fountains Of Paradise”.
Now, while the ‘space elevator’ may not ever come to pass, Stratolaunch Systems – which will use a ‘White Knight-like’ carrier, to carry SpaceX’s Falcon 9 Rocket to the edge of space; driving the cost of Earth to LEO launchs down to pennies on the dollar of what it costs today – Is close to being a REALITY.
That taken….just THINK ABOUT IT!
The Lunar Surface is quite actually an IDEAL Industrial Environment!
Why, there’s probably enough Iron-Titanium-Aluminum in the top Meter of Regolithe to PLATE THE MOONS SURFACE several centemeters thick!
Why….Just Image! No EPA! Why, you could have a Chernobyl AND a Bhopal BLOWING UP EVERY WEEK, and – No Biggie!
Think about it……One Piece (Seamless, as the seams are the ‘weakest link’) Composite Aircraft, cured in the Giant Vacuum Furnace that is the Lunar Day!!!
And, well, since you’re goping to have to build it ALL OVER AGAIN – SOMEWHERE! – Any-Old-Way…..Why NOT?!?
There….there’s my “Best shot du jour”.
“Therefore improved attribution of shifts in extremes (in whatever direction) have the potential to change cost-benefit calculations and thus policy directions.”
That is IT… no other comes close as a reason to do this stuff.
This is the most important reason why SOME people will claim that this is scientifically useLESS and not worthy of examination. Stop being so charitable.
Denialists HAVE to fear attribution becoming more accurate and damage being attributed more often. It is the one thing that WILL get the public to turn on them and recognize their lies.
It is the “end game” in which the obfuscators finally lose control over the outcome of court battles and “conflicting studies” and the “experts disagree” argument and, as the Tobacco Companies had to do, finally retreat.
BJ
Yeah I don’t know. I wouldn’t be at all surprised to see AGW factored into a 10 day forecast. Maybe the blocking high will have slightly higher pressure if it were not for global warming. Likely in my opinion, since obs are always being fed into weather models.
But this is the slippery slope argument, because then a big dump of snow and blizzard-like conditions gets pinned on global warming, since melting arctic ice and increased water vapor made the snow storm “even worse”.
Pakistan gets washed away because of higher sea temps, global warming again.
I think if you want to attribute these types of extremes to AGW, then you need to make a prediction before hand. The point being is that there are literally thousands of events going on all the time and picking “some” of the bad ones out before they occur is a clear sign that your attribution is credible.
Whether you think this is impossible or not is not the issue. It’s far more scientific (and useful) to at least try than being captain hindsight. Applying the climate dice to the specific real world events before they occur is the best measure of attribution. If there is any evidence of more extremes then you will not only get some statistical validation, you will also be in a position to deduced what happen and why.
When are you ever going to make a mistake (and challenge your own understanding) if you simply attribute things after they occur. Never.
Any chance of obtaining a complete list of attendees? Is there also any chance that the people who made the cynical remarks referred to in the Nature editorial will step forward to expand here on their comments? It would be a fruitful and interesting discussion to have “in person” instead of being relayed by an editorial.
Litigation is rather important in dealing with climate issues. The EPA endangerment finding for carbon dioxide had to be dragged out of it through court, for example. However, this has been a part of mitigation. Adaptation is resting with legislative bodies so far, such as the decisions to ignore sea level rise in North Carolina or to build stronger levies for New Orleans.
I feel that the Nature piece misses the main place where attribution will be used, and that is to make people whole when disaster strikes. We still don’t have a farm bill, but a carbon tariff applied to Chinese imports could certainly help to compensate farmers for lost crops. It is in the international arena that attribution will be used the most. The standard of proof is not high there. That China has the largest emissions is reason enough to extract compensation. And it should be. For mitigation we must all work together, but for adaptation, polluter pays and China is the polluter that has pushed climate change into the dangerous regime.
If recent extreme events are not at all due to global warming, we can assume they were aberrations and simply wait for a return to normal.
But if these events are due in large part to global warming, then we need to spend a lot of money now to change our infrastructure to prepare for a new normal.
How could anyone claim it’s not useful for society to know which is the case?
Three papers I found influenced my thinking on this point:
L. A. Smith, N. Stern, “Uncertainty in science and its role in climate policy”, Phil. Trans. R. Soc. A (2011) 369, 1–24, doi:10.1098/rsta.2011.0149
K. E. Trenberth, “Framing the way to relate climate extremes to climate change”, Climatic Change, DOI 10.1007/s10584-012-0441-5
K. E. Trenberth, Attribution of the human influence on climate Volume 1, 2011 2011 John Wiley & Sons, Ltd. 1, http://www.cgd.ucar.edu/cas/Trenberth/trenberth.papers/WIREspaper-mockup-2.pdf
I find Smith & Stern to be the most sensible, although I agree with Trenberth, and have advocated his position in my personal contacts. There is a scientific narrative, but few can understand that narrative unless it unfolds dramatically before their eyes. Fewer still understand the need and ramifications of economic adaptation as well as energy adaptation.
I think a more compelling story can be had in the logs of corporations like Swiss Re (see http://www.swissre.com/rethinking/climate/) who have logged tangible impacts of climate. Ultimately, climate change will have an economic impact, and, in one sense, economic gains in the present are being obtained at penalty of large economic costs in the future. The case needs to be made that, if nothing else, we are ECONOMICALLY tied to Nature.
> I never get as many media calls as in the wake of an extreme weather event of some sort.
That says it all, there is a question and scientists are asked. If an answer leads to decisions, good. If it does not pushes forward science at first sight, get better glasses or wait. One can never know what comes behind next corner.
Aren’t there more urgent questions having a deadline worth to discuss and finally answer? Can’t we postpone science inspects science until climate isn’t a concern any longer?
Gavin said, ” but if the government sets up a climate service to provide information to the public about climate, then, yes, they should as best as they are able.”
Then I’d say the primary goal should not be the communication of the science, but the creation of a non-partisan official group to communicate the science.
Oops. Did I just describe the IPCC?
If extreme weather events are going to be attributed to climate change with such immediacy after the fact, then shouldn’t the methodology used to support that attribution be something that’s applicable for all such extremes? For example… same methodology for extreme flood AND drought, and in particular… extreme heat AND cold?
[Response: Sure – the methodologies will be the same, though the results will vary from strong to non-existent attributions, and with error bars that vary too, all as a function of whether (and how well) the model(s) work. – gavin]
I still haven’t heard a peep from attribution studiers into the extreme cold anomalies that were persistent in Europe and Alaska over the winter. No doubt the deteriorating Arctic Ice sheet would have some role in there, but if there isn’t a consistent methodology capable of incorporating all extreme weather events of the same type, then is the exercise really robust? You can’t just rule some prolonged extreme weather events as “not related to climate change” and others as “definitely a signature of climate change”, or can you?
[Response: This has been the subject of a number of papers from Judah Cohen, more recently Jennifer Francis, and has been frequently speculated on. However, compared to the ones related to heat waves, this is a much more tenuous claim (involving an indirect mechanism of the summer sea ice extent (or Siberian snow cover) impacting the phase of the NAO and the ‘waviness’ of the Jet Stream in the winter). There is simply not the same weight of evidence to support it. – gavin]
So then, applying the “Moscow Warming Hole” methodology (or any of the recent methodologies that’s been usd) to the European cold snap this past Winter (because that is a similarly extreme and similarly long duration event that would similarly demand attribution)– what are the odds that our warming climate (as represented by the “much richer” statistical modeling in those studies) would produce such an event?
When(if) yet another similar event shows up this winter, is anyone going to be willing to re-visit the methodology, or are we going just stay with declaring “significant” and “useful” the studying of only extreme warm events as statistically more likely in a warming climate.
[Response: Lot’s of people are looking at this – and if convincing results are found, you can bet that it will get a lot of attention. My opinion is that if there is a signal, it is likely to be *much* smaller than the year-to-year variability and very difficult to detect with any confidence. – gavin]
Hi Gavin,
An initial scientific response to any event could continue to be ‘this type of event is projected to become more/less likely under continuing climate change’. Does a ‘climate service’ really need to be able to say ‘the risk of this event was increased/decreased by X% due to climate change’ in real-time?
Maybe the media and public do want to know this, but there is a risk that when more evidence arrives the value of X will change significantly and our credibility is damaged, even if we had made the appropriate caveats about level of understanding at the time. The media will always lose those caveats in the reporting (e.g. the infamous UK BBQ summer).
Or, do you think that we can/will be able to provide a reliable value of X in real time?
I think attributing extremes in general is a good idea, but I am not so sure whether the investment required to do this in real time is worth the risk. What decisions are altered by our ability to attribute straight away?
cheers,
Ed.
[Response: You are correct that it is possible that nuance will be lost by the time that any result hits the headlines (Duh!). But that is true for any result and so should not determine what science gets done. However, as I stated above, the main reason to provide real time access and analysis of *existing* data sets is to provide information that is based on something, rather than having the same discussions with information based on nothing. A side effect of enabling this, better systems will need to be created that will lead to more interesting and more accessible science that will overall end up improving our understanding of how models simulate extremes, what the robust results are (if any) and provide a target against which further progress can be made. – gavin]
It isn’t the mean that kills, it is the extreme. Knowing how the extremes are changing is critical to scale climate science to people. And, knowing the trends and how they affect the extremes is critical information, for insurance, for disaster relief, and for planning ahead. I don’t understand why one wouldn’t consider this important.
I greatly appreciate your responses.
“[Response: Sure – the methodologies will be the same, though the results will vary from strong to non-existent attributions, and with error bars that vary too, all as a function of whether (and how well) the model(s) work. – gavin]”
“[…if convincing results are found, you can bet that it will get a lot of attention. My opinion is that if there is a signal, it is likely to be *much* smaller than the year-to-year variability and very difficult to detect with any confidence. – gavin]”
…But you do not think that this issue casts a shadow on current attribution attempts? To me, not only does it NOT seem like ‘we’re there yet’, but worse… there is some sort of growing view that these currcent methodologies are acceptable/publishable/headlining, despite their inability to capture our full climate/weather representation (I realize that is just my opinion). A methodology that renders warm events as ‘clear signals’ and cold events written off as ‘not significant over year-to-year variability’ would seem (to me) to be a biased or incomplete methodology, particularly if such seemingly “impossible” cold weather anomalies occur more frequently. In fact, to me it would point more to the need to better understand future feedbacks of a reduced Arctic ice content in returning anomalously cold air climatologically/teleconnectedly into the mid-latitudes– which could greatly affect the temperature record in time-scales on which these insta-attribution studies have been restricting themselves.
If the true ‘right’ methodology is still not there, then why continue rushing to press with present studies with a non-comprehensive methodology? It seems like every heat-wave or drought will feature a race by various scientists to take their ‘warming/drying’ climate dice to press, while every cold snap will feature crickets (for now).
I realize the “final” methodology (whenever that comes) will “confirm” the previous results that our changing climate will generate more warm extremes, but why should this eventuality truly absolve these early forays despite their inability to do much better than simple probabilities with non-linear trendlines, while completely ignoring many other extreme event types?
[Response: I’m not sure I follow you. First, there is never going to be a ‘final’ methodology – that’s not how science works. Second why is it problematic that some results are going to be more robust and clearer than others? Every model (more or less) shows increased heat waves, but whether they show increased variability in the jet stream leading to increased cold snaps is much less clear (at first cut the models show the opposite). So for the second case the details are going to matter more, and the more something depends on the details the more variance (and uncertainty) will increase too. – gavin]
Different topic — has anyone noticed the British Environmental Audit Committee Report, issued this morning. http://www.publications.parliament.uk/pa/cm201213/cmselect/cmenvaud/171/17102.htm
Oops — the date on the British Environmental Audit Committee Report says Sept 12 (not this morning).
Salamano
Huh? Your characterization resembles nothing that is occurring on the planet where I currently reside. When looking at extreme events, what determines whether they are within normal variability or not is the amount by which they deviate from past extremes. A record cold event that deviated by six standard deviations or that persisted for a month and a half or that covered an abnormally large portion of the globe would be recognized as out of the ordinary just as surely as a record hot event.
Climate change does not mean the end of weather. We still get record lows, just as we get record highs. However, the fact that record highs are far outpacing record lows is certainly evidence of a warming trend. I would recommend reading some of what Tamino has written on his Open Mind blog.
One industry that cares very much about translating this knowledge into action is the insurance industry. The frequency of severe weather events is a money issue to them. And as they figure that out and translate that knowledge into higher insurance premiums, businesses and individuals will start to see what AGW really means.
Ed: The ‘users’ (apart from perhaps the media and lawyers) did not identify a valuable use for the attribution of specific individual events (such as the EU or Russian heatwave). The vast majority thought that the attribution of classes of extreme events (e.g. increase in 90th percentile hot days) is a valuable scientific endeavour.
I dont’ understand this. With “fractional” attribution, these two things are one and the same thing.
Suppose we compute the expected increase in Nth percentile events, for all N (the “valuable endeavour”). When a given extreme even occurs, we can find its percentile, and then simple arithmetic immediately tells us “this event has been made X% more likely by AGW”, right?
I suppose by “final” methodology, I’m thinking about a consistent and prevailing scientific understanding about how our past/present warming manifests itself in future weather (much in the same way we’re trying to establish what the various expectations are from El Nino on local/regional 90-day weather experiences). Science eventually settles on various understandings and expectations in this light. So, when I speak of ‘methodology’ I guess I’m talking about a uniform process by which every extreme of a certain “type” are able to be examined for “Global Warming” fingerprints and signatures.
If you go around with a different methodology for extreme warmth attribution than you do with extreme cold, in my opinion you don’t have a robust approach, but one that’s open to criticism. If results are “Robust” in one sense (indicating the expectation of a __-sigma warm event is much more likely in [global warming model] vs. [persistence]), but then that same methodology renders __-sigma cold events that actually occur as utterly impossible, that to me is a problem. If you’re really trying to identify the change in trend/variance in climate experience, then your approach needs to be able to handle both sides of the extremity coin. [Not just apply statistics + model runs to a declared non-linear temperature trendline when it comes to warm events, but then ditch that and go for discussion about jet streams and all that to dismiss cold events] … If “loaded climate dice” studies are what make it through peer-review and get headlining status on major publications, then those same dice need to be rolled in explanation of the extreme cold events as well, no?
I don’t see it as being able to separate warm extreme events from cold extreme events and developing different caveats for each. If we’re not quite there yet, than these initial forays suffer somehow from the ability to arrive at a quasi-correct expectation in the ultimate sense, while the whole time doing so in the least portable manner, and offering the least understanding about the entire feedback system.
If all observed events of type A are declared ‘impossible’ or ‘opposite’ of what a model postulates, then addressing the model or the understanding becomes more important than merely publishing more results each time other events happen to score a hit.
The baited-breath world that seems so eager to jump at publishing these single-event, right-after-the-fact attribution studies– and the blog/media outlets that rush to promote them, might come across as more measured and reasonable (to me) if they wait and study until they develop a methodological answer that can encapsulate either (a) BOTH extremes into their statistical/model/climate-dice option, or (b) A physical ‘jet stream’ discussion/model/feedback that can explain BOTH extremes in light of the same warming world.
Nobody’s saying nothing should be researched/published ever until we’re absolutely sure of everything, but (to me) SOMEBODY should be out there cooling the engines on how quickly these studies have been rushed into the literature, promoted all around, while their very real criticisms, cautions, and caveats receive scant mention even though that’s where the real effort should be directed. Maybe that’s what NATURE felt like talking about.
Salamano,
What are you talking about? You’re making vague insinuations that verge on accusations of malfeasance. I know of nothing that answers to the description you are giving. Give concrete examples.
The same models are being used on both record-hot and record-cold events. Have you even read any of the studies being done?
Consider sealevel rise. If we build critical infrastructure with an expected life span of 100 years we need to know the likely bad case scenario. Looking at those sorts of figures over the last 20 years, and how the worst case keeps rising, I suspect that the logical figure is now about 2 meters. Were my county to add some expensive infrastructure in our frequently flooding flood plane again it needs to be 2-3 feet higher than our two alltime record floods during the last decade.
So, to follow up on what Ed Hawkins said in #6, Eli wandered (literally, it’s a very long story) into the conference and sat in on one day’s sessions. There were three or four take homes
1. There is a huge disjunction between what climate scientists (want to ) work on and what the operators, planners and policy people need.
2. Aligning the the groups has to start with the operators (of things like water supply systems) if it is to have any effect.
3. Some (Chatham house rules) proposed that there be a formal organization charged with providing real time attribution to policy makers and the public. Most (at least as the Rabett heard) were dubious about the possibility of establishing such or whether there was a reliable scientific base for doing this at the skill level necessary.
4. Some interesting information about improvements to data bases, local attribution studies, etc. in the posters.
5. Oh yes, the magic of meeting people who only knew Eli through his electronic image.
I think that if there had been policy people or corporate types at the conference the response would have been quite different.
We have had more huge fires this past decade than over the previous century combined. If you include drought, many large cities nearly ran out of water, and some towns did run out of water. We’ve got an appalling plan for Australia’s biggest inland water system that doesn’t appear to factor in extreme drought (or flood). Flash floods are much more frequent – and causing lots of damage. That’s just locally – similar problems are being felt in lots of places around the world.
Insurance companies have put up rates sending out letters that blame it on more weather disasters (past and expected) because of climate change. Confirmation by scientists would add weight.
Local councils and governments need to let the public know why infrastructure is going to cost more (and why the cost of repairs after damage will keep rising). Storm water drainage, desal plants, coastal erosion barriers, bridges, ports etc are all going to be built to different specs. That is, they will cost more than they would have without global warming, and we are all going to have to pay somehow.
I wonder if the public are becoming confused by or dismissive of, the mixed messages about extreme weather. For example Dr Christy has noted that the IPCC TAR stated “milder winter temperatures will decrease heavy snowstorms” but after the winters of 2009-10 and 2010-2011, advocates of climate change, the Union of Concerned Scientists, stated “Climate Change makes major snow storms more likely”. So here we have climate change making snow storms both less likely and more likely. Here in Australia our Climate Change Commissioner, Dr Tim Flannery, stated in 2007 during a long period of drought “hotter soils meant even rain that falls will not fill our rivers and dams”. Since then the East coast of Australia has been inundated with rain, dams are full, rivers have flooded and Australia has been pronounced drought free. I’m sure you can imagine the attitude of the average Australian looking at his flooded living room, to the pronouncements of the Climate Commissioner
[Response: The TAR reference is an unsourced statement in WG2 – not WG1 where you would expect it. The WG1 discussion on extremes doesn’t mention heavy snowstorms at all (as far as I can tell). I’m not at all sure where any info on heavy snow events could have explicitly come from in 2001 – even now we don’t have good robust model results on this, so it is conceivable that the WG2 authors just made an assumption that warmer means less snow and therefore less heavy snowstorms. But this is a very loose chain of reasoning and there are plenty of reasons why it might not work out – and how it might work out differently in different regions. On the other hand, claims that more heavy snowstorms are a specific prediction can’t be sustained either. More intense precipitation is predicted (and has been observed), but how that trend and the trend towards less snow in mid-latitudes intersect will be subtle – and again, a definitive study on this is lacking. The situation with droughts and floods are similarly complex, though there are more reliable studies there with results that depend quite a lot on region and season. Note that in all cases these projections are statistical – not absolute, so looking at one flood, drought or snow storm as if it provided some kind of conclusive statement is never going to be sensible. Your overall thrust supports my opening contention though – we need to have more real science on these topics in near real time otherwise the gap gets filled with content-less speculations from people with agendas. – gavin]
I haven’t read all the comments up to this point, but I would like to mention that if extremes attribution could improve the Intensity-Duration-Frequency Curves used by local planners and engineers trying to plan for climate impacts this work could be quite useful. I am surprised that no one brought this up. I work in local government and currently we are struggling with how to take climate change forecasts and make them useful for our engineers in the design and maintenance of our stormwater and sewer systems. Typically they are designed to peak events based on 50, 100, and 200 year return periods which are reflected in the IDF curves. These curves are based on historical data, but if the future is going to be defined by changes in climate then we need to adjust the curves. the problem with the climate data is that it has such a large range of possible outcomes (based on different models/scenarios) in our region it is difficult narrow it to something useful. We are looking at using median values, but we are not sure our engineers who like to work in a world (real or imagined) of more certainty. If extreme event attribution could help provide more certainty that would be really be useful.
I just thought that, should my previous post get published, that the comment could be made that drought and flood in Australia are manifestations of extreme weather events thus proving the point that climate change is causing extreme weather events. However there are not unique in Australia as there was a much longer drought in the 1890s and indeed a famous Australian poem written in 1906 about the authors love of Australia has the line “of droughts and flooding rains”.
RobLL,
Considering the past two deacdes of sea level rise, the highest rate was ~4mm/yr from 1996-2006. Prior to and subsequent to that time frame, the rate has been ~2mm/yr. The average rate is ~3mm/yr, and shows no sign of accelerating. Therefore, building an infrastructure to last 100 years, needs to be able to withstand about half of a meter of SLR.
NOAA’s Atlas 14 is the most important and influential climate model in the USA today.
Tens of thousands of people who probably have never heard of James Hansen but who are building
our cities, our roads and our infrastructure are guided by Atlas 14. It seems to me that
it is worthy of critical peer review by the climate research community at Realclimate.org. See
http://www.nws.noaa.gov/oh/hdsc/PF_documents/Atlas14_Volume2.pdf and search for
“model”.
If you can improve its forecasts for coastal areas, then please do so. The value-added would
be worth a 1000 times whatever federal grant would be needed.
If you can’t improve Atlas 14 at the present time –but think that the historical data on which it is
based is becoming obsolete due to climate change and that extreme events outside its
forecasts are becoming more likely– then please present a case to NOAA for why they
should add a warning in Atlas 14 so that our builders, engineers and government
officials/planners will start taking that into account in their current designs.
Incidentally, the creators of Atlas 14 did test for signs of climate change in the data — they didn’t find
any of significance rising above the noise. But they only looked up to year 2000.