follow CCP

Recent blog entries
popular papers

Science Curiosity and Political Information Processing

What Is the "Science of Science Communication"?

Climate-Science Communication and the Measurement Problem

Ideology, Motivated Cognition, and Cognitive Reflection: An Experimental Study

'Ideology' or 'Situation Sense'? An Experimental Investigation of Motivated Reasoning and Professional Judgment

A Risky Science Communication Environment for Vaccines

Motivated Numeracy and Enlightened Self-Government

Making Climate Science Communication Evidence-based—All the Way Down 

Neutral Principles, Motivated Cognition, and Some Problems for Constitutional Law 

Cultural Cognition of Scientific Consensus
 

The Tragedy of the Risk-Perception Commons: Science Literacy and Climate Change

"They Saw a Protest": Cognitive Illiberalism and the Speech-Conduct Distinction 

Geoengineering and the Science Communication Environment: a Cross-Cultural Experiment

Fixing the Communications Failure

Why We Are Poles Apart on Climate Change

The Cognitively Illiberal State 

Who Fears the HPV Vaccine, Who Doesn't, and Why? An Experimental Study

Cultural Cognition of the Risks and Benefits of Nanotechnology

Whose Eyes Are You Going to Believe? An Empirical Examination of Scott v. Harris

Cultural Cognition and Public Policy

Culture, Cognition, and Consent: Who Perceives What, and Why, in "Acquaintance Rape" Cases

Culture and Identity-Protective Cognition: Explaining the White Male Effect

Fear of Democracy: A Cultural Evaluation of Sunstein on Risk

Cultural Cognition as a Conception of the Cultural Theory of Risk

« Is ideologically motivated reasoning rational? And do only conservatives engage in it? | Main | The equivalence of the "science communication" and "judicial neutrality communication" problems »
Saturday
Apr062013

More on the political sensitivity of communicating the significance of climate model recalibration

I posted something a few days about the political sensitivity of communicating information about scientists' critical assessments of the performance of climate models.  

In fact, such assessments are unremarkable. The development of forecasting models for complex dynamics (as Nate Silver explains in his wonderful book Signal and Noise) is an iterative process in which modelers fully expect predictions to be off, but also to become progressively better as competing specifications of the relevant parameters are identified and calibrated in response to observation.  

In this sort of process at least, models are not understood to be a test of the validity of the scientific theories or evidence on the basic mechanisms involved. They are a tool for trying to improve the ability to predict with greater precision how such mechanisms will interact, a process the complexity of which cannot be reduced to a tractable, determine algorithm or formula. The use of modeling (which involve statistical techniques for simulating "stochastic" processes) can generate tremendous advances in knowledge in such circumstances, as Silver documents.  

But such advances take time -- or, in an case, repeated trials, in which model predictions are made, results observed, and models recalibrated. In this recursive process, erroneous predictions are not failures; they are a fully expected and welcome form of information that enables modelers to pinpoint the respects in which the models can be improved.

Of course, if improvement fails to occur despite repeated trials and recalibrations, that's a serious problem. It might mean the underlying theory about the relevant mechanisms is wrong, although that's not the only possibility. There are phenomena that in their nature cannot be "forecast" even when their basic mechanisms are understood; earthquakes are probably an example--our best understanding of why they happen suggests we'll likely never be able to say when.

Usually, none of this causes anyone any concern.  The manifest errors and persistent imprecision of earlier generations of models didn't stop meteorologists from developing what now are weather forecasting simulations that are a thing of wonder (but that are still being improved!). Our inability to say when earthquakes will occur doesn't cause us to conclude that they must be caused by sodomy rather than shifting tectonic plates after all--or from using the scientific knowledge we do have about earthquakes to improve our ability to protect ourselves from the risks they pose.

Nevertheless, on a culturally polarized issue like climate change, this iterative, progressive aspect of modeling does create an opportunity to generate public confusion.  If one's goal is to furnish members of the public with reason to wonder whether the mechanisms of climate change are adequately understood--and to discount the need to engage in constructive action to minimize the the risks that climate change poses or the extent of the adverse impacts it could have for human beings--then one can obscure the difference between the sort of experimental "prediction" used to identify mechanisms and the sort of modeling "prediction" used to improve forecasting of the complex ("stochastic") interplay of such mechanisms.  Then when the latter sort of models generate their inevitable--indeed, their expected and even welcome failures--one can pounce and say, "See? Even the scientists themelves are now having to admit they were wrong!"

Silver highlights this point in the chapter on Signal & Noise devoted to climate forecasting, and discusses (with sympathy as well as discernment) the difficult spot that this puts climate scientists and climate-risk communicators in.

As I discussed in my post, this dilemma was posed by an article in the Economist last week that reported on the state of scientific engagement with the performance of climate model predictions on the relationship between CO2 emissions and surface temperatures.  Such engagement takes the form of debate -- or as Popper elegantly characterized it "conjecture and refutation," in which alternative explanations are competitively interrogated with observation in a way calculated to help isolate the more-likely-true from the vast sea of the plausible.

In fact, there was nothing in the article that suggested that the scientists engaged in this form of inquiry disagreed about the fundamentals of climate science. Or that any one of them dissents from the propositions that

(1) climate change (including, principally, global warming) is and has been occurring for decades as a result of human CO2 emissions;
(2) such change has already and will (irreversibly) continue to have various adverse impacts for many human populations; and
(3) the impacts will only be bigger and more adverse if CO2 emissions continue.

(These propositions, btw, don't come close to dictating what policy responses -- one or another form of "mitigation" via carbon taxes or the like; "adaptation" measures; or even geoengineering-- makes sense for any nation or collection of them.)

Maybe (1)-(3) are wrong?

I happen to  think they are correct, a conclusion arrived at through my exercise of the faculties one uses to recognize what is known to science. My recognition faculties, of course, are imperfect, as are everyone else's, and, like everyone else's are less reliable in a polluted science communication environment such as the one that engulfs the climate change issue.  

But the point is, whether those propositions are right or wrong isn't something that the debate reported on in the Economist article bears on one way or the other. The scientists involved in that debate agree on that. Any scientist or anyone else who disagrees about these propositions has to stake his or her  case on things other than the performance of the latest generation of models in predicting surface temperatures.

Well, what to add to all of this?

Surveying responses to the Economist article, one will observe some skeptics (but in fact not all; I can easily find internet comments from skeptics who recognize that the debate described in the Economist article doesn't go to fundamentals) are nevertheless trying to cite the debate it describes as evidence that climate change does not pose risks that merit a significant policy response.  They are trying to foster confusion, in other words, about the nature of the models that the scientists are recalibrating. Unsurprising.

But it is also clear that some climate-change policy advocates are responding by crediting that same misunderstanding of the models. These responses are denigrating the Economist article (which did not get the point I'm making about models wrong!) as a deliberate effort to mislead, and are defending the predictions of the previous generation of models as if the credibility of the case for adopting policies in response to climate change really does turn on whether the predictions of those models "are too!" correct.

I guess that's not surprising either, but it is depressing.

The truth is, most citizens on both sides of the climate debate are not forming their sense of whether and how our democracy should respond to climate change by following scientific debates over the precision of climate models.

What ordinary citizens do base their view of the climate change issue on is how others who share basic moral & cultural outlooks seem to regard it. The reason there is so much confusion about climate change in our society is that what ordinary citizens see when they take note of the climate change issue is those with whom they share an affinity locked in a bitter, recriminatory exchange with those who don't.

But all the same it is still a huge mistake for climate-change risk communicators to address these perfectly intelligent and perfectly ordinary citizens with a version of the scientific process here that evades and equivocates on, or outright denies that, climate scientists are engaged in model calibration.

In an open society--the only sort in which science can actually take place!--this form of normal science is plain to see.  Indignantly denouncing those who accurately report that it's taking place as if they themselves were liars embroils those who are trying to communicate risk in a huge, disturbing spectacle ripe with all the information about "us vs. them" that makes communicating science here so difficult.  

I admit that I believe that is wrong, in itself, to offer any argument in democratic debate that denies the premise that the person whom one is trying to persuade or inform merits respect as a self-governing individual who is entitled to use his or her reason to figure out what the facts are and what to do in response.

But I think it is not merely motivated reasoning on my part to think that the best strategy for countering those who would distort how science works is to offer a reasoned critique of those doing the distorting--not to engage in countervailing distortion.

One reason I believe that is that I have in fact seen evidence of it being done effectively.

Check out Zeke Hausfather's very nice discussion of the issue at the Yale Forum on Climate Change and the Media. It was written before the publication of the Economist article, but my attention was drawn to it by Skeptical Science, which discerningly that its thoughtful discussion of the recent debate furnishes a much more constructive response to the Economist news report than an attempt to deny that scientists are doing what scientists do.

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (27)

I'm a little surprised at your continuing to ignore the whole issue of scale or magnitude when you speak about so-called "adverse impacts" of climate change (not to mention the neglect of any mention of possible beneficial impacts, since in the general debate that's routine by now). Given that neglect, then of course it's possible to just wave away the question of climate sensitivity to CO2 levels. And then you're just left with your expressed puzzlement or depression with those policy advocates who get upset with mainstream media reports of reduced sensitivity. Whereas to my mind their reaction is understandable, predictable, and rational given what I take to be their primary goal, which is getting electorates to support the drastic reductions in carbon emissions necessary to have any real "mitigating" effect. But, if the climate is found to be less sensitive to CO2 levels than previously thought, it would be reasonable to expect that the adverse impacts would also be of a lower magnitude, hence involving a lower risk. The policy advocates see this as plainly as does everyone (or almost everyone) else, and they understand what an average person would understand about any kind of risk -- as the risk reduces, to whatever extent, so to does the level of concern, to a roughly equivalent extent. They can see, in other words, that their favored policies are becoming steadily less likely -- hence their unsurprising fuss.

April 6, 2013 | Unregistered CommenterLarry

@Larry:

A. Yes, I am fixating on the thoughtlessness of agreeing to play a symbolic/rhetorical game in which those on both sides agree to treat the performance of the models as a decisive test of their positions.Having made the bad decision to depict the models this way in the first place, those playing it will feel a sense of crisis when scientists do what they do w/ models, namely, update them in light of discrepancies between prediction & observation. Seems like an opportune time to say: "stop simplifying in a way that makes it harder, not easier, to engage in reasoned discussion."

B. B/c I am fixating on that (probably pointlessly; who listens to me, anyway?), I am definitely ignoring something of policy significance that is going on in the model recalibration. As the best prevailing understanding of what to expect, the models are of policy consequence; one can expect & hope the models will improve in the future but still recognize that the failure to act in the interim could be more costly--much more--than waiting for even better informatoin about what the situation is.

So now help me figure out what I admit I am not addressing. Is "magnitude" of impacts really what's at stake?

I get -- or see that many seem to assume-- that (1) the performance of the models could mean that warming will occur more slowly than anticipated. But that doesn't necessarily mean -- does it? -- that impacts will be less; just that (2) some impacts will likely occur later than they might otherwise have happened. No? (3) The performance of the models in this regard doesn't bear on what the impacts are likely to be conditional on warmng. Disagree?

If (1)-(3) are correct, what is the policyamking upshot? Right now there is certainly a practical debate about how to apportion policymaking focus between mitigation (CO2 reduction) & adaptation, & to a lesser extent geoengineering. Given that it is very costly economically & very difficult (geo)politically to constrain CO2, the argument for focusing more on adaptation & not fixating on mitigation seemed to be gaining strength (and also being ignored in a determined way by some--not all-- climate-change policy advocates).

But I'd think now either that (a) nothing has changed or that (b) perhaps there is potentially greater profit to be gained by mitigation after all. The basis for (b) would be that "less sensitivity" means that an economically feasible reduction, in a politically feasible amount of time, might not be as pointless as some had argued.

But I really don't know. Am eager to have the implications made clear to me.

But I don't expect a better grasp of the policy implications will in any way make sense of the mistake of denying that model calibration is normal. I think the inertia of thoughtlessness is great enough to impel continuation of the same regardless. But surprise me on that too, if I'm mistkaen!

April 6, 2013 | Registered CommenterDan Kahan

Dan

All the "bad" with AGW is predicated on the models, which are loaded with a high sensitivity. If the model are "wrong", and they are so being shown, then there is no "bad".

There are currently no positive trends for tornados, temp increase, hurricanes, increased rate of sea level rise, drought, or fire. The models all postuate an increase in these sometime in the future using high sensitify. The future postulated by these models is increasingly becoming doughtful. This is the reason you see such gnashing of teeth over the E article by the activists. The activists point to the models as a glimpse into the future and even Joe Public can see that such a future as shown by the models is getting dimmer as time goes on.

April 6, 2013 | Unregistered CommenterEd Forbes

Re: "model calibration" and the scientific method, both "normal" and paradigm changing. As I understand it the process is to come up with a theory for a particular phenomenon, and then construct a model of said phenomenon on the basis of that theory. Using that model, you make predictions, and then you judge how well the model did against the test of those predictions. The theory, in other words, is not insulated from the model it gave rise to. It's true that most models will require some degree of tweaking, but some of that at least will have at least some effect on the theory that generated the model in the first place. But what you don't do is just perpetually tinker with the model to make it conform to reality after the fact and call it "model recalibration", and never question the theory that it's based on, as though the theory itself were immune to any test of its validity. That's not science -- that's mere ideological fixation. At some point real scientists would want to start at least to re-examine some elements of the theory behind the models that seem to require so much "recalibration" -- would want to consider, e.g., the possibility of other, natural causes for climate change than human emissions.

But, to be a little more specific, and just accepting the underlying theory in this case, the performance of the models at least seems to me to imply not that warming will occur more slowly -- though that in itself would be of significance -- but rather that warming will be less than formerly predicted (because there is already less warming than expected for a given amount of CO2 in the atmosphere). That does mean, therefore, that the impacts will be less, not merely later than expected (though even if merely later, that obviously lessens the immediate urgency of the problem). So, yes, I would think that "magnitude of impacts" really is a part of what's at stake here.

(In any case, especially if you think that the magnitude of impact is still likely to be large, I don't understand the "lesser extent" characterization of the geoengineering approach (in both CO2 extraction and active cooling versions), since that's the only strategy that would seem to me to have a real chance of success, and would have spin-off benefits no matter how the climate itself goes, but that's a side issue.)

April 6, 2013 | Unregistered CommenterLarry

"But that doesn't necessarily mean -- does it? -- that impacts will be less; just that (2) some impacts will likely occur later than they might otherwise have happened. No?"

Strictly speaking, it means the impacts are *unknown*. They might occur later, they might occur sooner, they might not occur at all. The impacts depend on predictions of long term rises that are made by the same models that made the failed prediction of a short term rise. The models made a prediction, with an uncertainty bound, and within a decade it's already wrong. On what basis, therefore, do you claim the long term prediction to be reliable?

That there will be *some* impacts is fairly certain, but this does not imply that all the *other* impacts predicted will therefore also come true.

--

"Given that it is very costly economically & very difficult (geo)politically to constrain CO2, the argument for focusing more on adaptation & not fixating on mitigation seemed to be gaining strength"

Economic predictions say that as the world gets wealthier and as technology moves on, alternative sources of energy will become cheaper and we will eventually switch naturally, without having to be forced, because it will be in our immediate interests to do so. Just as we moved from wood to coal, and then from coal to oil, and then from oil to gas, so we will move to the next technology. Projections suggest costs will converge around 2050.

So the big question is not whether we stop using fossil fuels, but when? Do we do it now, when it is incredibly expensive and damaging because we're not ready? Or do we do it in 50-70 years time when it will be cheap and easy and something we'd probably do anyway?

If we start the switch in 50 years time, will the climate damage done be more or less costly than the economic costs if we switch now?

The calculation of the climate damage is done using the model predictions as inputs. There's a very real possibility they're over-estimating the temperature rise, perhaps by a factor of 2. The economic impact of this is estimated to be that it will be beneficial up to 2 C, and then rapidly damaging above that - the cost is fitted to a quadratic function of temperature change. (However, the uncertainty bands are broad.) Thus the cost we are comparing to the economic costs of mitigation are quite sensitive to the uncertainty.

Also, it is probably the case that the climate damage done depends not so much on the eventual magnitude, but on the rate of change of climate. So a lower sensitivity implies a lower rate of change for all time, and hence less difficulty adapting. Many things will be able to be replaced as part of the natural maintenance and renewal cycle.

Similarly, is it economically better to do the adaptation now, when it is expensive, or later, when it is needed? Again, it depends on how much and how fast. And the fundamental problem is that without validated models, we don't/can't know.

--

It's a big and complicated subject and I'm struggling to condense it. I'll try to say a bit more later, but please do keep asking questions about this.

April 6, 2013 | Unregistered CommenterNiV

@NiV & @Larry:

This distinction may seem hard to maintain, but separate (a) what you might think the significance of the modeling recalibration is from (b) what those doing the recalibrating understand the significance of what they are doing to be.

Consistent with Hausefather post I linked to and the Economist article itself (part of the reason I find the attack on it odd), I take it that the model recalibration is agnostic on impacts except as to timing. Specifically,

(a) Nothing in the debate bears on the impact of climate change conditional on warming (i.e., sealevel rise, severe weather, droughts, etc.). It's all about when & how much warming.

(b) Existing warming -- how much average global temperatures have already increased due to anthropogenic warming in the past century -- will have irreversible impacts (short of geoengineering), many of which will occur or unfold as a result of processes that will continue even if the temperature stays the same at this point. See the PNAS article I linked in the discussion of the post before last. Those assessments might themselves rest on contestable models, but those models are different from the ones that relate to the pace at which surface temperatures will increase.

(c) If previous models overstimated the level of warming conditional on how much CO2 would be emitted in the last decade or so, then that most likely means that the models are neglecting conditions (such as greater heat absorbing capacities in the deep ocean, etc.) that slow warming. Those conditions, however, only slow warming and are themselves limited in their potential to do so. Accordingly, if we continue to release the CO2 associated with existing fossil fuel sources, warrming will occur -- just later & at a slower pace than expected.

(d) That additional warming when it occurs will have additional adverse impacts.

If so, then, as Hausefather concludes, as the Economist article repeatedly repeatedly repeatedly! stresses (amazing; you'd never get this impression from reading the reactions to it) the policy significance of the model recalaiblration is not that big. We're still facing an unfolding series of cliamte impacts in real time; we'll still experience impacts in the future that are even worse if CO2 emissions are not constrained. Therefore, the imperative to adopt responsive policies is not diminished -- although possibly the "breathing room" for doing so is grdater than thought (I think the Economist editorial might use this term, in fact).

Again, don't challenge these conclusions on the basis of evidence or interpretations of them that are not related to forecasting of increased surface temperatures. *Just* focus on the understandings that exist among the scientists engaged in the scholarly exchange described in the Economist article.

Do you agree that scientists engaged in the model calibration don't see themselves as disputing anyting having to do with (a)-(d)? That's certainly my impression. For one thing, Hansen is depicted as recognizing that model recalibration should likely reflect the conclusion that the previous models overestimated how much warming there'd be conditional on CO2 release in the last decade; I don't think he has changed his mind at all about the "fundamentals" on cliamte change -- either the facts, or the policy implications.

If so, then I think the negative reaction to the Economist article among climate-policy advocates is better understood as connected to the rhetorical/symbolic logic of a debate in which the "models" have become identified as decisive tests of the fundamental claims of those on one or another side of the "cliimate debate." That's a position, I'd think, that the advocates should be trying to back out of, not deepen their commtiment to, b/c (a) it is false & (b) it is a position that will make them look absurd, and all the more so as time goes by.

April 8, 2013 | Registered CommenterDan Kahan

Dan -

Could you elaborate on the term "climate-policy advocates?" Who does it include, exclude?

April 8, 2013 | Unregistered CommenterJoshua

@Dan:
I see the point you want to make but I continue to think that the people you refer to as "climate policy advocates" and who have been upset at the Economist article have a better sense of its political, public relations implications than you appear to have. You come across as simply trying to downplay the significance of an admission that models constructed on the basis of a theory of catastrophic (more or less) AGW have failed, or look to be about to fail, to make predictions accurate even at their 95% probability range, and failed on the low side, which is the dominant message of the article. Yes, the scientists quoted are doubtless still concerned about climate change, as is the writer of the article itself. Yes, it's possible -- though this seems to be just a guess at this point -- that warming is merely delayed, but in the context that guess too seems like a rather feeble effort to "keep alarm alive" so to speak; it may also be simply that the models have overestimated the feedback effects that enhance the sensitivity of climate to CO2 levels, or underestimated negative feedback effects. So, quite apart from my own beliefs in the matter*, as you suggest, I would hazard a guess that anyone reading the article and not already heavily invested in the politics and policies that have been built on top of the accuracy of said models in the first place, would feel that their concerns about the issue have diminished to some degree, especially as an immediate or "urgent" political matter. And that, in turn, is what understandably upsets those who are heavily invested in said politics and policies.

That leaves them in a dilemma, as you point out, but I don't see that it helps to pretend that this article and the situation it reports on simply leave the issue unchanged. The wiser course, to my mind, esp. for those focused on the science as distinct from the politics of all this, would be to accept a diminished level of immediate concern, and use the political "breathing space" to re-examine the underlying theory, improve the models, and perhaps even re-think the array of policy responses to an obviously still very uncertain phenomenon, especially policies that have the potential to reverse the irreversible.

(*Fwiw, my own sense of the issue is that climate change is real enough, and that human emissions are currently its primary driving force. I don't cling to this as an article of faith, however, and as with all other theories of empirical phenomena I'm ready to adjust that sense in the light of new evidence. Claims of the immanence of "catastrophic" climate change -- almost always associated with political hype to "mitigate" the catastrophe by forcing down carbon emissions -- I regard as most likely ideologically-motivated fear-mongering.)

April 8, 2013 | Unregistered CommenterLarry

"Computer-generated models are used to predict future climates, but how much faith should we put in them to guide future actions? Should we treat their predictions as fact or fiction? With the hot topic of climate change ever current, can we wait to find out? Join climate scientist Tamsin Edwards, sceptic Jonathan Jones and policy adviser Claire Craig. Temperatures could rise in this session…"

http://www.cheltenhamfestivals.com/science/whats-on/2013/can-we-trust-climate-models

The models are being debated publicly now, which is a major change in the media. This opening of debate on CC is a major change that will have a profound affect on public opinion, which drives the finial political results.

April 8, 2013 | Unregistered CommenterEd Forbes

Dan -

I'm guessing that you've already seen this site, but in case not...

http://allmodelsarewrong.com

It touches on some themes that seem to me to be related to this discussion (in particular, some of the earlier posts Edwards put up).

It is where I first saw the quote:

“essentially, all models are wrong, but some are useful”
-statistician George E. P. Box

that I find most helpful as a frame for these discussions when I read something like the following:

You come across as simply trying to downplay the significance of an admission that models constructed on the basis of a theory of catastrophic (more or less) AGW have failed, or look to be about to fail, to make predictions accurate even at their 95% probability range, and failed on the low side, which is the dominant message of the article. Yes, the scientists quoted are doubtless still concerned about climate change, as is the writer of the article itself.

April 8, 2013 | Unregistered CommenterJoshua

" If previous models overstimated the level of warming conditional on how much CO2 would be emitted in the last decade or so, then that most likely means that the models are neglecting conditions (such as greater heat absorbing capacities in the deep ocean, etc.) that slow warming."

That's an unproven and highly speculative hypothesis. It might be lower sensitivity, or it might be an increase in negative forcing from other contributors, or it might be the loss of an alternative positive forcing, or it might be cyclic, or it might be that the 'noise' of random weather and chaotic internal variation superimposed on top of climate is larger than previously believed, or it might be the accuracy of the (extensively adjusted) measurements, or it might be that they've missed something more fundamental and the behaviour/effects are not as they supposed.

We don't know. I can't tell you that such speculations are definitely *wrong*, but it will take more than speculation about *possible* reasons before we conclude that your newly tweaked model is any better. It still has to prove itself.

You appear to be taking one *particular* alternative hypothesis/model - that there will be the same amount of warming only slower - as a given, and drawing policy conclusions from that. But rejecting the current models just means the current models are rejected, it doesn't mean anything else is thereby accepted, or can be taken as the default.

April 8, 2013 | Unregistered CommenterNiV

Dan, one of the real problems with these articles is something that Joshua has repeatedly and correctly pointed out about certain sceptic theories: they are not broad, deep, and self re-enforcing.

Whether one likes the IPCC approach, there was the effort to make the information with one section consistent with the others. Although Zeke usually does a good job, the first paragraph that starts "Ultimately..." starts poorly and quickly dips into nonsense. Assume a linear cost attribution, it will mean spending twice as much money for a 2C versus a 4C problem. Saying it does not matter to the public or policy makers whether it is $2 or $4 is unsupportable. It was not supported.

The next paragraph makes a claim of something that is known as untrue as true. It matters both about the CS as it does how much CO2 is emitted. Both are endemic to the equation.

Worse if you go to Knutti and Hegerl and decide you don't want that simplistic approach, all that discussion about the sting is due to estimates, simplifications, and not putting the graphs on semi-log basis. If you look at the development, it is stating that it is as easy to have a 2C to 1C drop in CS as a 1C to 0C. But for a log CO2 function 0C is an undefined CS, and impossible to reach. Yet, this artifact from simplification is used as if the long sting has meaning.

I do not know who is more correct, NiV or Larry, or perhaps both, since they bring out different problems of the same source.

April 8, 2013 | Unregistered CommenterJohn F. Pittman

Dan -

Did my previous post get lost? Did you monitor it out?

I wanted to point you to this site:

http://allmodelsarewrong.com/

In case you hadn't seen it before. It has some good discussion of some issues I see as related to this discussion. That site was the first where I ran across this quote:

“essentially, all models are wrong, but some are useful”
George E.P. Box

Which is what I think of when I read comments like the following:

You come across as simply trying to downplay the significance of an admission that models constructed on the basis of a theory of catastrophic (more or less) AGW have failed, or look to be about to fail, to make predictions accurate even at their 95% probability range, and failed on the low side, which is the dominant message of the article. Yes, the scientists quoted are doubtless still concerned about climate change, as is the writer of the article itself.

Larry -

Please note the bold:

models constructed on the basis of a theory of catastrophic (more or less) AGW have failed,

Now I don't want to get bogged down in semantics, so maybe I'm reading too much into that, but I see your statement as in error....

In theory, the models were based on scientific evidence, and produced a modeled outcome of a likelihood (within an error range) of a certain magnitude of warming - which some assess as having the potential (in proportion to the range of probabilities) to create significantly harmful outcomes.

There is a large gap between your construction and mine.

Now of course, as a die-hard believer in the ubiquity of motivated reasoning, I think it isn't likely that some degree of a presumption of dangerous warming did not worm it's way into model construction. But to categorize the entire model construction process as being based in assumed outcome - as your statement suggests, seems considerably less likely to me.

--------------------

Not directed at anyone in particular --

What I find interesting about some "skeptics"- who are so critical of climate models - is the selectivity of their criticism of models. IMV, it is impossible for us to reason without using models in the analytical process, and all models are wrong.

The conceptualization of climate models being wrong is, inherently, based on modeling. Although that modeling is of a different sort (more intuitive and less empirical), and as such less is theoretically coherent - although importantly not necessarily more "wrong" as the result of being less empirical - it is by definition, wrong.

It isn't the criticism of climate models per se, that I find a product of selective reasoning in the criticism of some "skeptics" about climate models, but the high level of certainty imbedded in much of the criticism I read. IMV, such over-certainty in outcomes of "skeptics'" "modeling" only undermines its validity - just as over-certainty in outcomes undermines the validity of climate modeling.

April 8, 2013 | Unregistered CommenterJoshua

A P.S.:

Roger Pielke Jr. has a post up on the Economist article that in part at least supports your notion, Dan, that policy advocates have put too much emphasis on climate models as a way to generate support: "Climate Predictions as Double-edged Sword".

First, though, I wanted to be more specific about one of your assertions re: the Economist article. Here's what you wrote:

If previous models overstimated the level of warming conditional on how much CO2 would be emitted in the last decade or so, then that most likely means that the models are neglecting conditions (such as greater heat absorbing capacities in the deep ocean, etc.) that slow warming. Those conditions, however, only slow warming and are themselves limited in their potential to do so. Accordingly, if we continue to release the CO2 associated with existing fossil fuel sources, warrming will occur -- just later & at a slower pace than expected.

And here's what the Economist wrote:

The mismatch might mean that—for some unexplained reason—there has been a temporary lag between more carbon dioxide and higher temperatures in 2000-10. Or it might be that the 1990s, when temperatures were rising fast, was the anomalous period. Or, as an increasing body of research is suggesting, it may be that the climate is responding to higher concentrations of carbon dioxide in ways that had not been properly understood before. This possibility, if true, could have profound significance both for climate science and for environmental and social policy.

Finally, here's Pielke's prescient warning from over six years ago, that I think matches well with your own concerns re: science communication, Dan:

There is no greater danger to support for action on important issues of human impacts on the environment than an overselling of what climate science can provide. If the climate behaves in ways that are unexpected or surprising it will be more than just credibility that is lost. Advocates for action should think carefully when gambling with the unknown predictive abilities of climate models. The human influence on the climate system is real, but the climate may not always cooperate.

April 8, 2013 | Unregistered CommenterLarry

@Joshua:
"Climate-policy advocates" is too broad for the purposes for which I'm using it. Here I mean only "those climate-change policy advoctes who are *slamming* the Economist article"-- or will leave it at that. Bet you can find them easily if you try! There is definitely an organized effort to respond to article by characerizing it is inacurate & motivated by desire to down0play climate; that's manifestly untrue. I think this effort is conseuence of bad decisions made in past to mischaracterize how modeling works. But it's not as if the there's an easy answer on communication issues here: maybe I'm wrong to think past decisions were mistake or that there were clearly better approaches; also, as I indicated, I *am* eager to be sure to know the Economst article is accurate (I have concluded, for now, that it certianly is! See Revkin's comments on same debate from 2 mos. ago. The only thing I'm firm on: it is not accepable to paint a misleading picture of how science works.

April 9, 2013 | Unregistered Commenterdmk38

@Larry:
Thanks for Pielke links. I agree that they are in same spirit-- in fact, they are better versions of what I am trying to say.
As for how I "come across": I *do* mean to come across as saying that those on either side who see the model recalibration as having large signifiance are misapprehending the normality of the such activity. It's *not* a test of whether climate change is something that is unfolding & having tremendous consequences for human socieitties. Believe that or not, what is going on w/ the model calibration is consistent w/ what you believe. LR = 1. I agree, of course, that model recalbiration can change assessment of *what* to do & *when*-- that is the whole point of having models. But it's not clear to me that the calibration won't actually strengthen the hand of those who think it *does* pay to invest more in mitigation, which the science literature had been seeming to suggest would be of limited efficacy given previous estmates of sensitivity & politically feasible estimates of carbon limits. But in any case, I don't think policy responses are dictated by any view of what the facts are here; when we move from "what is going on" to "what to do," values become dominant.

April 9, 2013 | Unregistered Commenterdmk38

Dan - I wasn't looking for you to name-names....but from where I sit, climate-policy advocates are abundant on both sides of the debate: Those on one side attack the Economist article because, they say, it is motivated by a desire to downplay the dangers of climate change. Those on the other side say that the article is motivated by a desire to keep the fires of alarmism burning. My view is that mis-representation of the models, and misuse of the scientific function of modeling, is part and parcel with motivated reasoning.

My comment was asking you to address that issue: "climate-policy advocate" describes just about everyone engaged on this issue.

April 9, 2013 | Unregistered CommenterJoshua

"All models are wrong" only in the sense that all models are simplifications of reality, not in the sense that all models make wrong predictions of parameters that are the focus or point of the model. Why do you think that scientists make models in the first place? It's not just so that they can recalibrate them.

The recalibration can change assessment of what to do and when, of course, but it can also lead to a reassessment of the theory that generated that models in the first place. To repeat a quote from the Economist article:

Or, as an increasing body of research is suggesting, it may be that the climate is responding to higher concentrations of carbon dioxide in ways that had not been properly understood before. This possibility, if true, could have profound significance both for climate science and for environmental and social policy.

You can simply ignore that possibility if you like, but then you "come across" not as someone defending the normality of science, but as someone defending an article of IMR faith -- like those pre-Copernican astronomers endlessly "recalibrating" their epicycles.

April 9, 2013 | Unregistered CommenterLarry

Larry -

"All models are wrong" only in the sense that all models are simplifications of reality, not in the sense that all models make wrong predictions of parameters that are the focus or point of the model.

I disagree. All models are wrong in their predictions, also. The questions become the practical implications of error range, the accuracy of quantification of error, views on the magnitude of error, or views of the implications of error.

Please note the bold in my excerpting your earlier comment:

....models constructed on the basis of a theory of catastrophic (more or less) AGW have failed,...

Now I don't want to get bogged down in semantics, so maybe I'm reading too much into that, but I see your statement as in error....

In theory, the models were based on scientific evidence, and produced a modeled outcome of a likelihood (within an error range) of a certain magnitude of warming - which some assess as having the potential (in proportion to the range of probabilities) to create significantly harmful outcomes.

There is a large gap between your construction and mine.

Now of course, as a die-hard believer in the ubiquity of motivated reasoning, I think it isn't likely that some degree of a presumption of dangerous warming did not worm it's way into model construction. But to categorize the entire model construction process as being based in assumed outcome - as your statement suggests, seems considerably less likely to me.

This, IMV, is connected to assumptions you are making about error and modeling, and assumptions you are making about the functionality of modeling. All models are wrong because all models make errors in predictions because all models have errors in their algorithms and/or have errors in their measured parameters/inputs.

--------------------

Related, but not directed at anyone in particular --

What I find interesting about some "skeptics"- who are so critical of climate models - is the selectivity of their criticism of models. IMV, it is impossible for us to reason without using models in the analytical process, and all models are wrong.

The conceptualization of climate models being wrong is, inherently, based on modeling. Although that modeling is of a different sort (more intuitive and less empirical), and as such less is theoretically coherent - although importantly not necessarily more "wrong" as the result of being less empirical - it is by definition, wrong.

It isn't the criticism of climate models per se, that I find a product of selective reasoning in the criticism of some "skeptics" about climate models, but the high level of certainty imbedded in much of the criticism I read. IMV, such over-certainty in outcomes of "skeptics'" "modeling" only undermines its validity - just as over-certainty in outcomes undermines the validity of climate modeling.

April 9, 2013 | Unregistered CommenterJoshua

For the sake of argument, I'm happy to delete the word "catastrophic" from the phrase you quoted, Joshua, even though from everything I've come across in the mitigation-advocacy camp that seems to have been a motivating assumption. Nevertheless, anthropogenic global warming is inherent in the theory that underlies the models, and the investigation of said phenomenon is the whole point of constructing the models in the first place -- models need a theory, in other words, they don't spring ready-made from "evidence".

There's not much to be said about your apparent notion that all models are wrong and therefore their errors don't matter -- or are to be "fixed" just as an end in itself, without consequence for either theory or policy -- except that it's certainly one way to insulate beliefs from any danger of being disconfirmed by reality. (And the fallback to a tu quoque argument doesn't really help.)

April 9, 2013 | Unregistered CommenterLarry

Larry -

For the sake of argument, I'm happy to delete the word "catastrophic" from the phrase you quoted, Joshua, even though from everything I've come across in the mitigation-advocacy camp that seems to have been a motivating assumption.

But the camp and the models are not one and the some. Obviously, there is overlap - a setup for motivated reasoning in the model development.

Nevertheless, anthropogenic global warming is inherent in the theory that underlies the models, and the investigation of said phenomenon is the whole point of constructing the models in the first place -- models need a theory, in other words, they don't spring ready-made from "evidence".

The physics whereby increased CO2 warms the climate is a construct upon which the models lie. I don't agree that mitigation of the GHE, is "the whole point of constructing the models in the first place,' at least theoretically. Theoretically, the "whole point of constructing the models" is to measure the impact of ACO2 on the climate. There is, of course, a gap between theory and reality, and motivated reasoning can certainly bridge that gap.

There's not much to be said about your apparent notion that all models are wrong and therefore their errors don't matter -- or are to be "fixed" just as an end in itself, without consequence for either theory or policy -- except that it's certainly one way to insulate beliefs from any danger of being disconfirmed by reality.

Hmmm. Well, I certainly didn't intend to convey that was my notion.

I think that errors absolutely do matter - errors allow for calibration. Yes, in a sense calibration is "fixing" the models, but fixing the models would certainly have consequence to both theory and policy. Of course, re-calibration could be, and IMO in some cases is, used as a protective mechanism against looking at those consequences, as you describe.

And the fallback to a tu quoque argument doesn't really help

No doubt. So you may have a point - I need to think about that: My first response is that motivated reasoning on one side does not justify such on the other side. Motivated reasoning does not necessarily imply incorrectness in conclusions. In other words, the existence of motivate reasoning amongst "skeptics" does not negate their argument that it exists amongst "realists." (of course it does, motivated reasoning is an attribute of human cognition).

I think my point was that an inability of some "skeptics" to recognize the necessarily fallible nature of models - all models including their own - leads them to specious conclusions regarding the presence of errors in climate modeling. But again, you may have a point - perhaps it was a tu quoque argument on my part.

April 9, 2013 | Unregistered CommenterJoshua

@Larry-- on "coming across" & permanent model recalibration-- I agree, as have said from beginning, that failure of models ever to progress furnishes reason to doubt that the theory they rest on is right. This is characeristic of econometrics (when the practioners of it even bother to check whether the models they posit are getting things right). But I don't think that climate science forecasting models are in that embarrassing state or anything close to it. Meteorology has made tremendous progress in last six decades or so. If someone had said, "ha! you are fundamentally misguided about the dynamics you are modeling" in response to the imprecision of models in 1960s, that would have been silly; the imprecision & mispredictions of those models furnished the raw materials for fine tuning that resulted in better models of today. I want to acknowledge that I don't have a degree of comprehension & knowledge here that would warrant anyone adjusting their own, but my sense is that the recalibration going on w/ climate models is perfectly consistent w/ recalibration that occurred at earlier stage of forecasting in meterogology; maybe we will learn that the situation is more like the sad one in the Samuelson example I linked to-- but that won't happen for a while. One can have reason to doubt the premises of the models & find the performance of the IPCC models in predicting surface temperatures consistent w/ that doubt; but the performance is consistent with confidence too, since it is in the nature of modeling of this sort that it advances progressively through calibration. So for now, the performance of the models is LR = 1 in that debate. I think this is a much more realistic assessment than the ones being offered either by those who are digging in & defending the past generation of models ("still within the 0.95 CI" is their embarassingly simplistic refrain-- as if one couldn't draw the inference that a model was overesetimating something from a string of outcomes at the lower bound of a 0.95 CI) or those who are popping champaign corks in anticipation of declaring that climate science is "bunk" b/c the IPCC models overestimated surface temperatures over a pariticular period. *These models* are not where the action is in the debate over what to make of climate change -- the fundamental debate turns on other things. I think you agree w/ this & are only objecting to my "tone" & how I "come across"? True?

BTW, anyone who believes that sealevels aren't rising & won't continue to do so as a result of *existing warming* (sea level rise continues even after temperatures stabilize; this is part of the "irreversibility" thesis), should be offering to sell insurance to coastal property owners in Fla, since right now commercial insurers won't underwrite new policies; the reinsurance mkt has dried up, so the primary insurers have pulled out; thus the field is wide open to you! (your only competitor will be the state residual pool, premiums for which are skyrocketing. (My understanding is that damage from hurricaines is conditional on sea level, which determines both extent of storm-related flooding & saline intrusion, the latter being of potentially even greater damage b/c it can have long-term or permanent impact on drinking water supply.) There are some great opportunities to make $ if you believe that the market -- not "group think" scientists or ideologically zealous advocates-- are overestimating impacts. The good news, too, if you want to exploit this mkt irrationality (if that's what you think it is!) is that it won't abate as a result of the model recalibration, b/c the impacts involved are "insensitive" to the sorts of adjustments in "sensitivity" that are at stake there.

Again, I am not taking a position on who is right. I'm pointing out how odd it is that those advocating policies to address anticipated climate impacts think the debate being reported in the Economist article matters. My hunch is that they are throwing "good" credibilty $ after "bad" by continuing to stake their position on a simplified version of what normal science looks like.

April 10, 2013 | Unregistered Commenterdmk38

@Dan:

No, I'm not one who says that climate science is bunk. I do think, however, and unfortunately, that's it's currently infected with a non-rational alarmist bias along the lines of the "limits to growth" crowd (or the current "planetary boundaries" crowd) -- the same people who were predicting massive human die-offs, the end of oil and other essential resources, and the general breakdown of global civilization by the 70's or 80's just a few decades ago, well before global warming became the signature issue for "the cause". And it just seems naive in anyone not a himself fervent supporter of said "cause" -- or even fervent supporters that are honest -- not to see that. But that in itself doesn't make arguments and evidence in support of AGW wrong, and you can find a few people, like Pielke, who can deal with and present the evidence in a calm, rational, and skeptical manner that makes human-induced climate change seem likely. That said, however, it remains the case that 1) timing and scale or magnitude matter, 2) appropriate policy responses are not determined by the science, as you've said, and 3) the likelihood of AGW, and particularly its magnitude and timing, is itself reduced by empirical findings that conflict with expectations built on its basis, as a Bayesian (or even just common sense) approach would indicate.

So I don't question the judgment of the insurers who face what they see as a risk that's difficult to quantify. And, needless to say, I agree with your assessment of the predicament of the advocates who get in a knot over mainstream media attention being given to the slowdown in climate warming -- except that, since I view them as primarily driven by an ideology rather than science in the first place, I don't think that predicament is "odd".

April 10, 2013 | Unregistered CommenterLarry

("still within the 0.95 CI" is their embarassingly simplistic refrain-- as if one couldn't draw the inference that a model was overesetimating something from a string of outcomes at the lower bound of a 0.95 CI) or those who are popping champaign corks in anticipation of declaring that climate science is "bunk" b/c the IPCC models overestimated surface temperatures over a pariticular period.

This, fits with my view as someone not smart or technically/mathematically -skilled enough to understand the statistics - and seems like an important point that both sides dismiss. Please excuse the ignorance I am about to expose -- and help me to see if there's any way I can understand this better.

As I see it, apparently the belief held by many is that by virtue of temps over the next three years or so, the statistical modeling will either be "validated" or "invalidated." How could that be? How could three years of temperature records cause a 180 degree in the swing of assessment? If over the next three years, temps jump back well into the CI range, would that mean that they couldn't drop back out after 5 years more of data? If they drop out after 3 more years, does that mean that they couldn't rise back in after 5 years more of data? And even further, wouldn't we expect that the temps would fall out of that 95% CI range some 5% of the time even if the modeling were valid (since the models essentially predict that would happen).

Seems to me that all we have (as a species) to look forward to is 50 or 100 or perhaps more years of refinement and calibration of the models, and probably endless bickering about whether the models have been validated or invalidated during that period. More same ol' same ol' (Actually, it's very unlikely that I'll be around for that 50 years, and none of us will be around for that 100 years). And in the end, the statistical analysis is not what will be most salient. If in 50 or 100 years, people's views on AGW will be shaped by the climate, not the models. Either there will be obvious changes or their won't. The arguments about the models will be largely irrelevant. And people will move onto yet another venue to realize their motivated reasoning.

April 10, 2013 | Unregistered CommenterJoshua

Joshua, there are several aspects to this. One is: All models are wrong, some are useful. People will disagree as to "useful."

The GCM model is about a phenomena that is noisy and also autocorrelated. The autocorrelation can cause either side to be "wrong" from time to time. Think of the IPCC taking smaller and smaller segments and claiming that warming is accelerating, and now skeptics doing the same in the opposite direction. Which brings us to the point some don't understand.

As such a model is close to being invalidated as to magnitude, one expects times when it will be valid and times when it will be invalid. Depending on the nature of the phenomena, we could see thirty years with the 2C being invalidated, and then 60 years from present going back to validation. That the odds are extremely low for this as presently modeled should tell one something. If it happens,it should tell one that the model has simplified characteristics that render its usefulness to that of the eye of the beholder, if you are making claims such as we have 150 days to save the world. It should definitely tell one to expect a lot of heat and opposition for something that is expensive and bothersome, since your point that most people expect a simple explanation, IMO, is typically true. Add in the fact that there will be times when the opposition can point out you have been wasting time and money better spent elsewhere, you have a PR nightmare.

This is one of the many reasons persons such as myself are luke warmers. There are pseudo cycles that were not included in the models because of their frequency. Are they important or not? At present, it appears that they are.

This is why I support "no regret" policies. It is also why I take the IPCC to task for being overconfident. They have set the stage for their own failure. IMO, YMMV.

As to motivated reasoning, it is simply a fact that has not stopped millennia of advancement, I see little reason to expect it to stop us now. I agree with your sentiment that "people's views on AGW will be shaped by the climate, not the models". The irony is that is what should have happened with our system. It used to be mandated that harm be proved. The courts and the regulators took the position that models could replace proof. As you have seen, models are not proof, they are an extensions of what we believe, and all too often in complex situations, they are simply "wrong." Or as modellers are apt to point out, not useful.

April 11, 2013 | Unregistered CommenterJohn F. Pittman

JFP -

I figured you might be one who would take pity on me...

For sure, a determination of "usefulness" is largely subjective - as is, I would add, the determination of what is "expensive and bothersome."

As for the importance of pseudocycles - there again, much uncertainty exists, IMO - especially given the descriptor of being cyclical Will we be back to the strong trend of increasing surface land temps once those cycles are completed? Are the models based on mis-timed assumptions about any such cycles (in other words, confused the high end of cycles as a baseline)?

And the definition of "no regrets" policy is also subjectively oriented.

Perhaps even a determination of "millennia of advancement" is subjectively rooted. And even if we accept it as objective fact, how much advancement might have been made if people had been more open (and meta-cognitively aware of) our proclivity towards biased reasoning?

April 11, 2013 | Unregistered CommenterJoshua

The problem with psuedo-cycles is that frequency and amplitude can change making the determination of an offset difficult if not impossible. Further, one has to worry about getting into Hurst LTP and that is a whole nother can of worms.

I reaslized that there is a lot of play in the words I used, but that was to highlight the problem. My point is that in best of conditions such would tend to cause problems in civil discourse. Add in the counter factual nature of determining if we accomplished anything with many of the policy directed "solutions" and we have a situation that disagreement is almost guareenteed to be vocal and contentious.

But like data, it is what we have and not we want. It is up to us to make the best of what we can.

April 12, 2013 | Unregistered CommenterJohn F. Pittman
Member Account Required
You must have a member account on this website in order to post comments. Log in to your account to enable posting.