"Our greatest responsibility is to be good ancestors."

-Jonas Salk

Wednesday, January 5, 2011

Getting Uncertainty Backwards

Since we are already warming and we expect to be warming too much, the only argument against policy is strong new evidence that the warming is overestimated. Arguments for large uncertainty are definitely of scientific interest, but do not weigh against vigorous policy. Rather, quite the contrary.

Interestingly, there is a large overlap between those arguing for uncertainty and those arguing for laissez faire greenhouse gas policy. That position is fundamentally incoherent.

Lindzen and Spencer do not fall into this trap - they advocate for a low sensitivity with high confidence, a confidence that does not seem justified but at least is consistent with their policy positions. Broecker on the other hand argues for a very high uncertainty and very vigorous policy response. So climate scientists align coherently.

But people from outside the field often take the position that "this science is unsound, so how can we take action?" This nonsensical position is common among the people sidetracked by the millenial data sideshow.

The point is that there is a real sensitivity in the system that is clearly nonzero, and that there are various ways to establish a confidence spectrum as to what it might be.

That sensitivity propagates through various impact risks, each with its own confidence spectrum.

And whether we accept a dollar denomination or not, the aggregate risk is some combination of the component risks.

Much is made of whether the physics has a fat tail, but the fact remains that the impacts do have one-sided fat tails. If there are benefits to the coming changes, they are smallish compared to our worst fears (methane feedback, new deep ocean circulation, ice sheet collapse). This is why, the greater the emphasis on uncertainty, the greater the risk.

However often I say this, it seems to fall on deaf ears. Please make the effort to understand.

If the IPCC has understated uncertainties then it has understated risks.

I am amazed that those least concerned about uncertainty tend to be the ones talking about the risks the most while those most concerned about it tend to be the ones who are unconcerned.

This is clearly the result of professional denialist talking points playing up uncertainty. The thing is, they have managed to confuse just about everybody outside of science.

There is a great deal we don't know. The recent breakdown of the Arctic vortex with consequent severe winter weather in the eastern USA and western Europe is not something I had heard much emphasized in the predictions. The (amazingly immature) crowd of naysayers is now crowing about how little we know, and defenders of science are casting about for excuses. Wrong wrong wrong.

Things nobody saw coming are starting to happen. Already. This is bad news. Uncertainty is really very high. This means we had better get this carbon thing under control.

You keep using this word uncertainty. I do not think it means what you think it means.


Anonymous said...

You say:
"But people from outside the field often take the position that "this science is unsound, so how can we take action?" This nonsensical position is common among the people sidetracked by the millenial data sideshow."

True, but isn't this is also the position of several of your skeptical climate science colleagues. Curry, Loehle, Micheals, Pielke Sr. etc. There is also the entire lukewarmer contingency within the statisticians who want to take over as citizen scientists, and who hold this same position.

dhogaza said...

Yes, gryposaurus, one can excuse as ignorance the misunderstanding of the significance of uncertainty by lay people, but professionals such as those you list don't share that ignorance.

They are being (trying to find a kind word here), disingenuous.

Curry, especially, who harps on uncertainty as being an argument for inaction far more stridently than the others, IMO, at least recently.

Nick Palmer said...

Just to save people wondering, the last sentence is a reference to the "inconceivable" theme in "The Princess Bride"

link to montage of "inconceivable" moments

Steve Bloom said...

Speaking of changes in deep ocean circulation.

Michael, now I'm wondering about a possible connection between this Labrador Current warming and the previously observed leakage of the Agulhas Current into the South Atlantic. Shouldn't we be looking to see some increase in Atlantic meridional flow? Or are the error bars on that too large for now?

Unknown said...

As I mentioned in some previous occasions, pure large uncertainty does not lead to any policy action. Perhaps we should say we have a particular kind of practical certainty (combined with various uncertainties). MT mentioned it as "real sensitivity is nonzero". I want to mention it a little more specifically e.g. "It is very unlikely that the equilibrium response to CO2 doubling is less than +1.2 deg C".

Steve Bloom said...

Kooiti, IMHO it's time to relegate the Charney sensitivity to the ash heap of history. Real (Earth system) equilibrium sensitivity is known to be ~5C, and I think we need to start emphasizing that (and as you probably know some scientists are alredy advocating for just that). In effect, advocates for a low Charney sensitivity are arguing that the slow (Earth system) feedbacks are comparatively high (either that or they reject the rock-solid paleo results), but with the continued emphasis on the Charney sensitivity they're never made to defend the point.

Aaron said...

Loss of the Arctic vortex is in various GCM runs, it just at the end of the runs, so nobody talked about it very much – they thought it was way out in the future and they did not want to be, “Alarmist”.

Welcome to IPCC climate model “Year 2083”, i.e., 3 years after the start of the final Arctic Sea Ice break up.

Loss of the Arctic vortex changes our whole basis of weather prediction and hence our basis of engineering design. In short, loss of the Arctic vortex puts all of our engineered structures and infrastructure at risk. Most of these engineered structures are sureties for loans. Thus, AGW poses a huge risk for the mortgage and loan industry, but they seem not to understand these additional risks.

Since the 2007 sea ice melt, the Atlantic seaboard seems to have become a preferred storm track with a trend of increasing frequency of extreme events in summer and winter. I expect this trend to increase as there is less and less sea ice in the Arctic. Washington, D.C. Is well placed to get a first hand view of these events, and the current session of congress is there at the right time to see the biggest weather show in human history. Now, if they can just believe their eyes.

John Mashey said...

Thanks! I recognized the quote from one of my favorites, but hadn't seen the montage. Perhaps we should start a collection of recognizable clips that apply to climate science / anti-science. My entry (in honor of Inhofe, Barton, Cuccinellui, and maybe some members of the new House is Monty Python witch scene.

Anonymous said...

I'm also curious as to why Weitzman's work isn't more influential in killing this confusion. Hasn't he demonstrated that the economics argues for action moreso if the uncertainty is large?

It seems like normal-course cost-benefit-analysis is built more for uncertainty along the lines of "should we commit our workforce to Blackberries or iPhones or Androids now, or should we wait several quarters to have more "certainty" as to how the market plays out?" But in that case, "uncertainty" doesn't dominate the decision-making process the way that Weitzman suggests it does for climate...

This paper attempts to explain why science and economics cannot resolve these questions to anywhere near the degree of accuracy that we have come to expect from more traditional applications of cost-benefit analysis (CBA), because there is so much deep structural uncertainty associated with climate change. The "unknown unknowns" of climate change make CBA signicantly more fuzzy in this arena than in more traditional applications - like constructing roads, strengthening bridges, or setting building codes in earthquake-prone zones. The paper tries to make sense of this anomalous situation and explores what might be done in terms of actionable alternatives under such fuzzy circumstances.

I don't remember any killer rebuttals to Weitzman... more like bobbing and weaving. And certainly Weitzman hasn't backed down any...

Michael Tobis said...

Steve (especially your latest), Aaron: references?

Steve: "Michael, now I'm wondering about a possible connection between this Labrador Current warming and the previously observed leakage of the Agulhas Current into the South Atlantic. Shouldn't we be looking to see some increase in Atlantic meridional flow? Or are the error bars on that too large for now?"

elephino. "leakage" sounds like small beer in the global continuity scheme of things but this is the first I've heard of it.

David B. Benson said...

Climate sensitivity is no less than the transient response. The transient response is (approximately( estimated via a zero dimensional, zero reservoir model, Decadal Global Warming, to be greater than 2 K for 2xCO2.

Steve Bloom said...

David, it's a pernicious error to in any way imply that real sensitivity could be that low. Of course the questions of how fast we get there and what happens along the way are crucially important, but in discussing them we shouldn't lose track of the long-term big picture.

Michael Tobis said...

Steve, "pernicious error" is strong language. Please provide evidence.

Michael Tobis said...

Grypo, Pat Michaels is anything but a colleague and Pielke Sr. is not worth taking seriously. I'm not very familiar with Loehle. As for Curry, who know what is making her tick, but she is certainly among those sidetracked by the millenial data sideshow.

David B. Benson said...

Steve Bloom --- You didn't bother to even read the linked study, much less delve into it, it does appear.

But completely independently of any other study I arrived at a transient response which is at the low end of the IPCC range of Charney equilibrium climate sensitivity.

Moreover, the equilibrium sensitivity must be higher (ordinarily) or at least equal to (special systems) the transient response, As I stated more briefly in my prior comment.

Steve Bloom said...

Michael, as best I can establish the question of the Agulhas Current "leakage" (maybe analogize to a dab of arsenic?) is the hot topic in the physical oceanography community, at least in terms of resources being put into getting obs in a hurry. This recent paper is how I became aware of the issue. (Bryant and Gareth @ HT covered it a couple of months ago.) The breadth of scientific activity can be seen by checking the AGU FM abstracts.

I'm trying to understand as much as I can about oceanic and atmospheric circulation chnages, as they seem to be the most critical thing to watch. Understanding the key inter-connections is difficult, but I have the impression that's the case for scientists too.

Steve Bloom said...

I apologize if that sounded harsh, David. I was reacting to your use of a 2K figure as a basement for sensitivity. What you said was true, but saying it in the way you did (similarly and more commonly people will talk about 1.2C for CO2-only) works to shift the Overton window toward the low side. And yes, I'm well aware of the difference between transient and equilibrium.

Michael, I said pernicious more or less because of the foregoing, although I could go into more detail if that's not clear. I didn't check the dictionary definition, but to me it has implications of bad, persistent and hard to get rid of. It and "global warming" are the two leading forms of climate terminological kudzu; the latter may be with us forever (although perhaps we can affect the reaction people have when they hear it), but I hold out some hope for the former.

Of course none of this is to say that Charney and his committee made a mistake, since any hope of establishing ESS was beyond the capabilities of the time, but having thought about this rather a lot lately I'm convinced the scientific and activist communities (hopefully with the journalisitic one to follow) need to abandon use of the Charney sensitivity ASAP. (Of course it's still fine as a benchmark for models.)


Steve Bloom said...

Driving the nail in even farther, let's not forget that the Charney sensitivity can never occur in the real world. More recent science has made it clear that not only won't the things it holds constant stay constant, but that carbon feedbacks in particular won't be waiting around for the Charney equilibrium to be reached.

These folks were at AGU making the case for using ESS rather than paleo-analogues. I'm inferring that they like the Charney sensitivity even less, but that's not too much of a leap. Their key point seems to be that feedbacks will behave differently (i.e. worse) under a fast transient, meaning that we should be so lucky as to see climate conditions like the mid-Pliocene. I'll quote the whole thing since it's (IMHO) such an important point:


Steve Bloom said...


"Given the inherent uncertainties in predicting the climate and environmental response to anthropogenic emissions of greenhouse gases, it would be highly beneficial to society if science could identify geological analogues to the human races current grand climate experiment. Unsurprisingly this has been a focus of the geological and palaeoclimate communities over the last thirty years, with many scientific papers claiming that intervals in Earth history can be used as an analogue for future climate change. Using a coupled ocean-atmosphere modelling approach we test this assertion for the most likely pre-Quaternary candidates within the last 100 million years with a particular focus on warm 'interglacial' events within the Pliocene epoch. In all instances these intervals, including Pliocene interglacials, fail as true analogues since they either represent equilibrium climate states to a long-term greenhouse gas forcing, whereas anthropogenic emissions of greenhouse gases provide a progressive (transient) forcing on climate, or the sensitivity of the climate system itself to greenhouse gasses was different in the past due to tectonic and other factors. Therefore, we conclude that there are no satisfactory pre-Quaternary geological analogues for a future greenhouse warming. References to geological analogues should be abandoned and replaced by the investigation of processes operating during warm climate states. The relevance of warm intervals in Earth history, including the Pliocene, is not in the assessment of climate sensitivity, nor is it in the vain search for a direct analogue for late 21st century climate. The strength of pre-Quaternary palaeoclimate studies, and Pliocene research in particular, is in (a) the assessment and calculation of the response of global temperatures to increasing greenhouse gas concentrations in the longer term (multiple centuries), what we term the Earth System Sensitivity, and (b) in the assessment of the abilities of climate and Earth System models used to predict climate change in both the past and future."

Notice that they got a bit shouty with the title.

Steve Bloom said...

Loehle is a McIntyre sycophant. He has some sort of sciencey PhD, but no background in climate research. IIRC a few years ago he produced a recent climate recon that turned out to be very poorly done (I think published in E+E, so no surprise there).

David B. Benson said...

I opine that the Miocene is a better analogue given that CO2 concentrations now are around what they were during the Miocene.

The impact of Miocene atmospheric carbon dioxide fluctuations on climate and the evolution of terrestrial ecosystems
Wolfram M. Kürschner , Zlatko Kvaček, and David L. Dilcher

Anonymous said...

@ Steve Bloom: I must say that when Alan Haywood presented those conclusions about Earth System Sensitivity and the Pliocene during his lecture at UofToronto last March - even thought I kinda casually understood this before - I really struggled to "believe" it. The implications are grim.

I am increasingly of the opinion that a large* part of the relevant** populations will have to individually go through some sort of trajectory of "Oh Shit!!!" and initial despair before we get anywhere...

* "Large" doesn't mean anywhere near 50%. A much smaller critical percentage of influencers can and will shift broad perceptions remarkably rapidly. But I think that they/we'll need to really get it viscerally and intellectually...

** A better word than "relevant" will occur to me after I hit "post", but what I am referring to are populations in the nations that are doing the damage... As much as the people in other nations on the receiving end damages are also "relevant" (and how!), I just mean that even if 100% of them "got it", it wouldn't matter if too few of the US, Can-Aus-GBR, "BASIC", FSR, etc. "got it"... and I think that "getting" it - breaking through a lifetime's paradigm of "normal" - is going to necessarily include some combination of initial shock and despair before moving on...

Nothing original in my comments. Coming to terms with post-2100 outlook for SLR is the same kind of "Oh shit!" moment.. just sayin' it seems like a necessary part of the shift. Maybe I read too much Clive Hamiltion...

Unknown said...

I mentioned +1.2 K "Charney" sensitivity not as a likely value, but as the low end of its possible values.

My message is that we need something which practically everyone (including real skeptics) can consider certain in order to agree some policy option about climate change. I take that the total feedback to CO2 doubling, including water vapor and ice albedo, is unlikely to be negative. It is translated to that the "Charney" sensitivity is unlikely to be less than +1.2 K. (Perhaps we need to exclude Lindzen and Paltridge from "everyone", but perhaps no more leading scientists.)

joe said...

Re: Paleo-analogues, a recent paper in Geology says:

"The consensus among these proxies suggests that Arctic temperatures were ~19 °C warmer during the Pliocene than at present, while atmospheric CO2 concentrations were ~390 ppmv."

Ballantyne et al., GEOLOGY, July 2010 38; no. 7; p. 603–606; doi: 10.1130/G30815.1

dhogaza said...

steve bloom:

"I was reacting to your use of a 2K figure as a basement for sensitivity. What you said was true, but saying it in the way you did"

He stated this was for the transient response. I read that and understood that the equilibrium sensitivity on real live planet earth must be higher, not equal to, 2K.

"pernicious error" seems to be an unwarranted response to a statement that boils down to saying "equilibrium sensitivity must be greater than 2K".

Steve Bloom said...

Sure, and then denialists happily take away the 2C (or 1.2C, or whatever basement figure is being mentioned), combine with the still-extant assessment that 2C is a "safe" level, and depart in triumph. The general public will just be confused by an excess of numbers. I'll repeat my request that whenever these lower numbers are mentioned (and I think there's no need to mention the Charney sensitivity at all outside technical discussions) it be emphasized that the Eemian, Pliocene and Miocene paleo-analogues all point to ~5C sensitivity. The science behind that conclusion is rock solid at this point. (Oddly enough it seems to have become that solid in just the last few years, but hardly anyone noticed.)

Kooiti, please explain how it's useful to concede the non-existence of the water vapor feedback, which is what happens when we agree with denialists about 1.2C.

David, the Miocene certainly adds to the evidence, but as that abstract tries to make clear it can't be considered an analogue. There is no analogue, not even the PETM. We are forcing the climate system at a much greater rate than can occur naturally. Force it hard enough and even that 5C sensitivity will cease to be any sort of guide. (Hansen has a nice discussion of this latter point in his book.)

Thanks for that, Joe, although note that the indicated temperature is consistent with a global temp of ~3C at that time. The tropics hardly warmed at all. Polar amplification is serious business for anyone who likes ice.

Everyone please look at the figure on page 7 of this paper. This is what happened to the far North Atlantic during the Pliocene as a consequence of much-increased meridional flow. Toasty. Californians may also note something interesting -- think warm Mediterranean climate, as in Libya.

Steve Bloom said...

BTW, David, transient sensitivity can be than equilibrium sensitivity if a too-fast transient enhances existing feedbacks or triggers entirely new ones. Hansen points out that something like that must have happened during the PETM.

dhogaza said...

steve bloom:

"Sure, and then denialists happily take away the 2C (or 1.2C, or whatever basement figure is being mentioned), combine with the still-extant assessment that 2C is a "safe" level, and depart in triumph."

The fact that you don't care for the way in which David worded his perfectly correct statement doesn't make his perfectly correct statement a "pernicious error".

There are, perhaps, better ways to offer constructive criticism, if constructive criticism is the point.

Anyway, my last word on the matter ...

Robert said...

MT - While I fully agree that uncertainty is not a reason to undertake action, I am less certain that your statement "If the IPCC has understated uncertainties then it has understated risks" is necessarily accurate. I think "may have understated the risks" would be more accurate.

For example, even though there is greater uncertainty, the current best estimates may in fact prove to be right, in which case the currently stated risks would not understated. Or IPCC may have understated the uncertainty only on the low end or the high end. In the former case, the risk may actually be overstated while in the latter case, the risks would most certainly be understated.

Steve Bloom - Perhaps I missed it, but I still would be interested in a citation for your statement that we now "know" that climate sensitivity is 5C since this flies in the face of what I have read which still puts in the 2.5 to 4.5 range.

Anonymous said...

@ Robert. I'm not Steve, but I suspect he is referring to the following work (amongst others...):

Earth system sensitivity inferred from Pliocene modelling and data

From the abstract:
... Taking these lines of evidence together, we estimate that the response of the Earth system to elevated atmospheric carbon dioxide concentrations is 30–50% greater than the response based on those fast-adjusting components of the climate system that are used traditionally to estimate climate sensitivity...

Steve Bloom said...

dhogaza, I did indeed mean that it's a pernicious error to mention these low numbers, and the Charney sensitivity in particular, without at least noting the much higher ESS (plus peak Pliocene CO2/temp, although as noted that should not be presented as an analogue). David himself, of course, is in no way pernicious. Repeating, again: ESS sets the parameters for the climate response we (or our descendants) will actually experience. Note how Robert just nicely illustrated the problem.

Steve Bloom said...

Robert, at this point there's a massive amount of evidence about CO2 and temperatures at the warm peaks of the Pliocene and Miocene, the former being especially valuable since much climate-relevant geological data remains available and it's recent enough as to be nearly identical with modern conditions.

The temp range you mentioned is (more or less) for the Charney sensitivity, which to explain a little more cannot ever happen in the real world since it holds constant or ignores too many feedbacks that will kick in (many already are kicking in) long before the CS fast feedbacks can reach equilibrium. The paper rustneversleeps linked to had to calculate ESS using a model since Pliocene CO2 peaked at no more than 400 ppm, far short of the 560 ppm doubling (relative to the pre-industrial level) used for equilibrium sensitivity. (The paper cites a long list of other papers describing the evidence for CO2 and temps during the Pliocene, although to get the general idea it's probably sufficient to read the Robinson paper I linked above.) IIRC 560 ppm CO2 hasn't been seen since at least the Miocene or Oligocene and not for any extended period since the Eocene (~40 mya), by which time climate proxies become rather thin under the ground. I should say thin in terms of detail, but there's still plenty of basis to conclude that with doubled and higher CO2 the Eocene saw the warmest temperatures of the entire Phanerozoic (the last ~600 my since the end of the last snowball epoch), including the very nasty (as we would perceive it) PETM. Note that the "clathrate gun" key to such episodes (there have been others but the PETM was the worst, at least in the latter half of the Phanerozoic) is more loaded now (thanks to the cold temperatures of the last few my) than it was at the start of the PETM.

I'll repeat the point that a too-fast CO2 excursion will result in the ESS having to be revised considerably upward, mainly because negative feedbacks are too slow. During the Pliocene those negative feedbacks were very much in play, meaning that the 5C ESS derived in the paper is a best case.

Finally, for perspective, I should just note that current CO2 forcing is increasing at a rate approximately 100 times faster than has occurred under natural conditions. To quote the late, great Richard Pryor: "We in trouble."

David B. Benson said...

rustneversleeps --- Thank for the link.

All --- Even with just the transient response (so far) of about 2.3 K for 2xCO2 we are in very deep trouble due to This is very alarming because if the drying is anything resembling Figure 11, a very large population will be severely affected in the coming decades over the whole United States, southern Europe, Southeast Asia, Brazil, Chile, Australia, and most of Africa.
[Link is a pdf.]

Robert said...

Steve Bloom - Thank you for your long reply. While I believe I generally understand the concept of ESS, I am not sure these recent studies are even close to the last word on CO2 climate sensitivity or allow one to state with conviction that we know it is 5C, giving the many uncertainties in Tertiary climate and temperature reconstructions and the various other climate forcings in effect millions of years ago. Unfortunately, the paper cited by Rust is behind a paywall and the abstract you cited is for a poster session so I won't be able to get much more there, but will see what freely availabe papers I can find on this topic.

Steve Bloom said...

Robert, apparently you neglected to click on the links I provided above to two relevant open-access papers, one pertaining to the Eemian (Turney) and the other to the Pliocene (Robinson). Both discuss evidence for real-world sensitivity substantially higher then the CS. A public-access version of the Lunt et al. modeling paper is here.

To clarify a key point, ESS for present CO2 levels is ~5C, but since we won't be seeing those levels at equilibrium that 5C is *not* what we can expect. If we stopped emitting CO2 very soon we might hope for a bit better than that for peak temperature, but it seems increasingly likely that it's going to be worse, in part because people just refuse to believe that such a thing is possible.

Of course, the climate system may bite us in our collective butt based on what we've already done, as with the methane lurking (and now observed to be leaking at an increasing rate) under the other ESS. This is one chamber of that clathrate gun I mentioned above. Other equally unpleasant things are possible.

BTW, pdfs for a majority of recent papers can be found easily using Google Scholar. For others, email the corresponding author and request a copy; my experience is that most will respond promptly.

Steve Bloom said...

Thanks for mentioning the Dai peper again, David. It's a reminder that even a relatively small increase in global temperature can have very large effects, on us if not the climate system.

Robert may want to note that such changes are possible only because of the expansion of the tropical troposphere (in both directions) and the resultant compression of the rest of the atmospheric circulation toward the poles. Of course, it's not a problem unless you're a farmer who likes to plant things in areas with good soil that get reliable rain, or someone who likes to, you know, eat. Ask the Australians how the other thing is working out for them, and be sure to check current reports on what last year's nasty weather did to agricultural commodity productivity and prices.

As was mentioned above, we are also seeing significant and unnaturally rapid changes in the oceanic circulation. What could possibly go wrong?

Here's an open-access review paper on the atmospheric changes, Robert. Note that it's already moved rather a lot.

Robert said...

Steve - I had scanned the paper on the Eemian and the alghus(sp?) current leakage but admittedly did see the link to the Robinson paper. While it is not clear to me how the Eemian paper indicates a climate sensitivity of 5C, I appreciate you providing the other links and will check them out.

Steve Bloom said...

Robert, as I noted above the last time 5C happened was a very long time ago. The evidence we have for sensitivity is at lower CO2 levels and thus lower temps. If, for example, we see a peak Pliocene temperature of ~3C associated with CO2 levels of <400 pmm and no other cause for the warming, and bearing in mind the logarithmic relationship between the two, that indicates to sensitivity in the 5C range. When the Eemian, the Pliocene and the Miocene all point to the same answer, it's time to pay attention.

Bear in mind that the technical ESS is a bit of a sideshow, especially since it can only be a guide to the future rather than a guarantee. Look instead at the vast amount of evidence we have for the parameters of peak Pliocene climate. An abrupt transition to such a climate would be very unpleasant, especially since it's unlikely to stop there (unless we change our ways quickly).