So that was …. interesting.
First off, I’d like to thank the commenters for all of the suggestions and ideas to the previous post. They were certainly useful. In particularly, the connection with the difficulties faced by evolutionists in debates vs. creationists proved to be very a propos. Our side played it it pretty straight – the basic IPCC line (Richard Somerville), commentary on the how ‘scientized’ political debates abuse science (me, though without using the word ‘scientized’!) and the projections and potential solutions (Brenda Ekwurzel). Crichton went with the crowd-pleasing condemnation of private jet-flying liberals – very popular, even among the private jet-flying Eastsiders present) and the apparent hypocrisy of people who think that global warming is a problem using any energy at all. Lindzen used his standard presentation – CO2 will be trivial effect, no one knows anything about aerosols, sensitivity from the 20th Century is tiny, and by the way global warming stopped in 1998. Stott is a bit of a force of nature and essentially accused anyone who thinks global warming is a problem of explicitly rooting for misery and poverty in the third world. He also brought up the whole cosmic ray issue as the next big thing in climate science.
Update: The transcript is now available – though be aware that it has not yet been verified for accuracy. Audio + Podcast.
The podcast should be available next Wednesday (I’ll link it here once it’s available), and so you can judge for yourselves, but I’m afraid the actual audience (who by temperament I’d say were split roughly half/half on the question) were apparently more convinced by the entertaining narratives from Crichton and Stott (not so sure about Lindzen) than they were by our drier fare. Entertainment-wise it’s hard to blame them. Crichton is extremely polished and Stott has a touch of the revivalist preacher about him. Comparatively, we were pretty dull.
I had started off with a thought that Lindzen and Stott, in particular, would avoid the more specious pseudo-scientific claims they’ve used in other fora since there were people who would seriously challenge them at this debate. In the event, they stuck very closely to their standard script. Lindzen used the ‘GW stopped in 1998’ argument which even Crichton acknowledged later was lame. He also used the ‘aerosols are completely uncertain’ but ‘sensitivity to CO2 from the 20th Century is precisely defined’ in adjoining paragraphs without any apparent cognitive dissonance. Stott didn’t use the medieval English vineyards meme (as he did in TGGWS) – but maybe he read the RC article ahead of time.
The Q&A was curious since most questions were very much of the ‘I read the Wall Street Journal editorial page’ style, and I thought we did okay, except possibly when I suggested to the audience that the cosmic ray argument was being used to fool them, which didn’t go over well – no-one likes being told they’re being had (especially when they are). My bad.
The organisers asked us afterwards whether we’d have done much different in hindsight. Looking back, the answer is mostly no. We are scientists, and we talk about science and we’re not going start getting into questions of personal morality and wider political agendas – and obviously that put us at a sharp disadvantage (shades of David Mamet?).
One minor detail that might be interesting is that the organisers put on luxury SUVs for the participants to get to the restaurant – 5 blocks away. None of our side used them (preferring to walk), but all of the other side did.
So are such debates worthwhile? On balance, I’d probably answer no (regardless of the outcome). The time constraints preclude serious examination of any points of controversy and the number of spurious talking points can seriously overwhelm the ability of others to rebut them. Taking a ‘meta’ approach (as I attempted) is certainly not a guaranteed solution. However, this live audience were a rather select bunch, and so maybe this will go over differently on the radio. There it might not matter that Crichton is so tall…
Thats where you lost it Gav….from then on you came across all pretentious
Why dont you come up with some “expert” climate scientist predictions about something that will happen next year and let everyone see how brilliant your models are. After your predictions show how gifted you are send them to the hurricane forcast dudes who predicted the end of the world last year because of the accelerated climate change. They had some super climate scientist “expert” models that were spot on accurate. Oh wait a minute I think those experts completely screwed up and werent even close. In fact I think they were off by 65%. In fact I dont think those experts have come closer than 60% in the past 20 years. Now isnt that an observeable trend that should point to you clowns not knowing anything about how our weather systems will act in the next 12 months let alone the next 100 years. Make one accurate prediction that isnt just an average of the last 100 years of collected data and then maybe you will have some credibility. All you do is say what has happened and tell us why it HAPPENED!!! MAKE ONE ACCURATE PREDICTION. All you need to do is comee within 75% to significantly beat the average expert analysis.
I think you lost it on three levels:
1. Refusing to use logic to support your view: “..you’re not gonna hear us arguing about obscure details…”
2. Using straw men, by implying that sceptics are like tobacco companies and creationists
3. Ad hominem attacks: [edit] “…we just need to tap Philip Stott”
You say of Stott “He also brought up the whole cosmic ray issue as the next big thing in climate science”. This seriously undermines your credibility, for he in fact said: “I’m not saying they’re right or wrong, they’re pointing however at the edge, to new research. You cannot dismiss that, because it’s a consensus for CO2”
re: 251. No. If someone does not understand something and that point is made, that is not being prententious at all. It is a failure of the audience to comprehend. Place fault where it rightfully should go.
By the way. Is the debate going to be put on Youtube?
Keep up the good work.
Mike
Re: 247
Ike, thanks for your thoughtful reply.
You write (in reference to my post in which I claim that (1) global climate models are bases on an incomplete representation of the physical processes that drive climate change and (2) there is no example of a validated prediction on a multi-year or multi-decade timescale made by a climate model):
“This doesn’t mean anything. What are the known physical laws that are being neglected? What published scientific literature are you referring to? Keep in mind that climate models have been making predictions for some time now, and there is a record that they can be compared to, so the claim that ‘models may be ‘tuned’ to past behavior and yet have no utility for predicting the future’ just doesn’t hold.”
Let me take this in two parts.
First, “What are the known physical laws that are being neglected?” In my experience climate modelers are smart and dedicated, and they attempt to ground models in known physical laws as best they can. One big issue, however, is that the element size in a typical model used for this purpose is about 200KM X 200KM X 1KM (or about 14,000 square miles by a half-mile thick). Most of the relevant physical processes that drive temperature change, such as cloud formation, convection, etc., take place at a physical scale that is far smaller than this. These aggregated processes are represented within models as parameters that are modeler-created, rather than being based on physical measurements.
Of course, any model in any context is a simplification of the system, and the acid test of a simulation model is typically how well it predicts the outcome of interest in the presence of complete inputs. This brings us to your second point: “Keep in mind that climate models have been making predictions for some time now, and there is a record that they can be compared to, so the claim that ‘models may be ‘tuned’ to past behavior and yet have no utility for predicting the future’ just doesn’t hold.”
I don’t believe that there is a single validated example of a climate model that has demonstrated material predictive skill over a period of multiple years. I would welcome a counter-example.
Please note that I am referring to an actual prediction (i.e., a forecast made on date X for some date after X), rather than a ‘hindcast’ (i.e., validation on a hold-out set). As you’ve probably gone through, there is a long string of posts on this thread that go into some specificity on this request, as well as a fairly detailed discussion of evaluation of Hansen’s model-based 1988 forecast. In the interests of space, I’ll just refer to these, rather than repeat them here.
Best,
Dana
[[The idea of Global Warming is NOT science, it’s a guess about the future. No matter how many people agree on that guess, it is still a guess.]]
No, it’s not a guess. It’s something we’ve been measuring for about 150 years. It’s already happening.
[[The EARTH’s temperature has always fluctuated UP and DOWN. For the fluctuation to stop for no reason and say it’s going to freeze or burn up is ridiculous. We know the EARTH has been hotter than now in the past and Co2 has been at higher levels as well and the planet did not burn up then.]]
Nobody is saying it’s going to burn up. They’re saying our agriculture and economy will be damaged.
[[What made the EARTH’s temp stop fluctuating? Global Warming is a multi-million dollar business now.]]
Huh? What?
[[It’s so funny, we are always looking to the weather man wondering why he was wrong again and again and they are only guessing at tomorrow temperatures. However when they try to guess about 50 to 100 years in the future…what…we just believe them???]]
Do you understand the difference between weather and climate? Weather is day to day variation and is chaotic. Climate is a long-term average (30 years or more) and is deterministic. I don’t know what the temperature will be in Riyadh tomorrow (weather), but I predict that it will be higher than in St. Petersburg (climate). To put it yet another way, weather is an initial values problem, climate is a boundary values problem.
[[MAKE ONE ACCURATE PREDICTION.]]
They already have. They predicted the stratosphere would cool as the troposphere warmed. It’s cooling.
They predicted the poles would warm faster than the equator — “polar amplification.” That’s also happening.
Hansen gave an estimate (1988, scenario B, the middle one) that still looks pretty close to what happened over the next decade or so.
Your information is incomplete.
Gavin, the site’s becoming a lot more like my favorite flashlight or car repair site’s threads on global warming (yes, they really have them) — dominated by many long repeats of the standard elementary questions.
One thought, is there any way to sort the FAQ from the main thread? In this thread it’s no great loss, but in the science threads, the same thing’s happening.
You all are very, very patient; I fear the average reader isn’t going to notice that you’re trying to be helpful and instead just see the same old same old being proclaimed faster than you can answer each one.
Perhaps asking each poster to come back at least once and prove they’re not a robot-driveby, or something like checking IPs? Because it’s really feeling cluttered. Backing off myself, can’t tell the trolls from the children right now.
re 252. [[Why dont you come up with some “expert” climate scientist predictions about something that will happen next year and let everyone see how brilliant your models arehis has been done in at least two instances.]]
Here’s at least four instances where the CLIMATE models (which are different than chaos-based WEATHER models), not to mention the stratospheric cooling that was already mentioned.
The first was Jim Hansen’s (NASA) global climate models predictions for future temps made in the 1980s…
He showed three variations including a most likely scenario and two more extreme scenarios. (which the skeptics selectively chose a more extreme scenario to “prove” that he was wrong…good lie.
His “most likely scenario” was very accurate for average temp rises.
Second, Jim Hansen did models predicting what would happen with Mount Pinatubo’s volcanic effects just after Pinatubo blew (that was putting his balls,and model’s reputation on the table for sure)… He was very accurate on both counts.
Thirdly, the models predicted more warming at higher latitudes than mid latitudes. That was accurate.
Fourthly, the models precicted more warming during the winter than during the summer in northern latitudes. This was also accurate.
There’s one point in the debate that needs to keep being repeated: none of the participants denied that AGW is happening. In fact, they all affirmed that it was. Crichton faulted some for NOT taking it personally serious. Stott criticized it for deflecting attention from taking care of the poor.
As Cheech Marin’s character in “All Dogs Go to Heaven” might have said, “If is denial, then chain me to the wall.”
Some of this is getting silly. I sense that the trolls have found RC, notice that real climatologists live here, and have come to beard the lion.
Recommendation: Can we add a moderation function to the comment software? Just a simple thumbs up/down thing or maybe a label system (insightful/interesting/troll) al la slashdot, so that visitors can set their fence to filter out trolls and keep the science stuff at the top of the float.
I know that filtering is nasty business, but the important “pure science” discussion is doomed to be lost under the shrill diatribes of extremists if something isn’t done. And getting the readership to help in the effort of winnowing the wheat from the chaff would provide a sort of democratic process that restores respect.
If this were a climate symposium and some wing-nut stood up in the seats and started ranting on his favorite conspiracy the nice men in jackets with “EVENT” stenciled on the back would help him to the door. The important issues to be addressed here probably do warrant a suitable response to deliberate or even ham-handed attempts to derail them.
You know, I have to say, that one of the most disturbing aspects of the denialists presentation was how readily Lindzen launched into the “warming in Mars, Pluto, Triton…)”. There’s not an iota of truth in most of that and the small amount of possible warming on Mars is readily understandable if you know anything about drivers of Martian climate. Rush Limbaugh can convincingly claim ignorance here. Richard Lindzen can not. It is a sad day when a scientist resorts to outright obfuscation to score points in a debate for the benefit of nonscientists who do not know any better.
Gavin, thank you and your panel colleagues for having the courage to participate… now, consider the experience as an experiment. Having heard opinions on approach beforehand (two threads back) and now, having had the experience and seen the audience poll, are there lessons that you would recommend to the science community for how to effectively communicate the correct level of urgency to an interested (if non-expert) audience?
A hypothetical concluding remark: “The best estimate of current climate science is, extending current trends over the next 50 to 100 years, we will indeed see effects like ocean levels rise 8 to 12 meters due to melted ice caps, tropical pests and diseases migrating into more densely populated temperate regions, and increases in desertification. All these predictions make the current “crises” described by others on this panel worse in the not-so-distant future. If handing an extreme slow motion remake of “Day after Tomorrow” to generations of children and grandchildren does not constitute a crisis, what should it be called?”
You and your colleagues rightly decry both alarmism and do-nothingism or denialism. Fair enough, but to me, the middle path is sounding increasingly alarmist.
Being a good scientist should not prevent a role of policy advocate. It’s a matter of keeping the respective hats straight.
So the question: are there any lessons learned if the objective is to communicate the “proportional” level of crisis to the general public?
Re: 260 and 262
I don’t think these are good examples of global temperature forecasts (the issue, as I see it, under consideration). The only actual global temperature forecast is Hansen’s 1988 forecast, which has not been validated. See, for example, Hansen’s own evaluation of this in his 2006 paper from the Proceedings of the National Academy of Science(http://www.pnas.org/cgi/reprint/103/39/14288), in which he says:
“a 17-year period is too brief for precise assessment of model predictions, but … comparison with the real world will become clearer within a decade”.
In other words, even according to the author, we have to wait another 5 – 10 years before we can empirically evaluate the accuracy of those specific models. It’s not that the forecast was wrong, simply that we are simply in a period when the forecast signal can not yet be statistically distinguished form noise.
You can see in this paper the follwoing data observations:
1. The temperature difference predictions 1998 – 2005 that Hansen reports in his paper were actually higher over this timespan in Scenario C (strong emissions reduction) than in Scenario B (base case). I think this indicates that there has not even been clear spearation in the predicted effects over the 1988 – 2005 time period.
2. If instead of the 17 year period 1988 – 2005 one were to evaluate prediction accuracy over the 12 year period 1988 – 2000, you would find the predictions of warming rates are off by a factor of 6 in the base case (I had to read the values form the chart as I don’t have the underlying data tables, so these are approximations). In fact, you won’t find convergence of the predicted – actual residuals to zero as you proceed from 1988 – 2005 year-by-year, which I is I think a rigorous version of finding validation.
In the interests of fairness, please note that Gavin has a different interpretation of this and has indicated that he may post on this in the near future.
[Response: I will. But I reiterate, there is a big difference between using 17 years to precisely determine climate sensitivity, and being able to show that it’s not zero. -gavin]
In terms of Pinatubo, I think Knox and Douglass did a pretty thorough job of demonstrating in GRL that climate sensitivity as approximately measured in the Mt Pniatubo “natural experiment” actually contradicts model seensitivities to GHGs. See: Climate forcing by the volcanic eruption of Mount Pinatubo
David H. Douglass and Robert S. Knox GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L05710, doi:10.1029/2004GL022119, 2005. At best, when considering multiple papers on this subjects, the results are ambiguous.
[Response: K&D is rubbish, and the response by Wigely et al to their paper shows exactly that their methodology gives no information on climate sensitivity at all. Secondly, all the GCM physically determined matches to the Pinatubo response are much better than K&D’s empirical fit – which if you are truly open-minded on this, should give the models a boost no? – gavin]
In terms of your third and fourth examples, it is an entirely different thing to predict relative geographic and/or seasonal rates of warming than to predict temperature (or more precise, temperature sensitivity to GHGs).
Best,
Dana
Dana, you say that: “These aggregated processes are represented within models as parameters that are modeler-created, rather than being based on physical measurements.”
That’s not true; the parameterization schemes are indeed based on physical measurements. This has been gone through before. That is a distortion of how the process is carried out, though it is true that some parameterizations are on firmer ground than others. You also skip the more relevant point:
“Thus, if you are going to make such statements, and expect to be taken seriously, provide some references, if you can.”
Don’t you think that if the contrarians could demonstrate disagreement between models and observations, they would have published such results? They tried to make the claim that troposphere temperature trends disagreed with model predictions, but that was clarified by correct analysis of the satellite data. They have nothing to go off, and if they did they certainly would have published it – so where is the evidence that the climate models are not agreeing with observations?
Let’s also look at the upper tropospheric moistening in comparison to model predictions: see http://www.colorado.edu/chemistry/chem5151/McBride.pdf (Radiative signature of upper tropospheric moistening, Soden et al 2005) Looks like the model is tracking it with good accuracy, yes?
Let’s also consider the current trend of new temperature records, and compare it to what Richard Lindzen said in Congressional testimony ten years ago:
One of the common claims in support of the reality and seriousness of global warming is that we have had a large portion of record breaking warm years during the last decade or so. This is not a claim used by the IPCC, and its presence in any discussion is a rather clear piece of evidence of the intent to deceive (especially when the claim is made by a scientist). As noted by Solow and Broadus (1989) and Bassett (1992), this is an inevitable occurrence when one has a single record breaker in a time series characterized by interannual variability, interdecadal variability and an underlying trend or longer period variability. Solow and Broadus show the clustered nature of record breakers….Not surprisingly, record breakers cluster in exactly the manner found by Solow and Broadus (1989) in the observed temperature record. The occurrence of such record breakers contributes no additional information. Our prime concern remains with the determination of trend and the identification of such trends with emissions of carbon dioxide, and this remains a difficult and contested issue as the IPCC freely acknowledges..
Now we have ten more years of temperature records, and Lindzen no longer mentions the topic. Hmmm… what does that indicate? That the temperature record trend can no longer be justified on the basic of ‘statistical clustering’. In fact, this is an example of how a large number of isolated events, none of which can individually be attributed to global warming, in sum represent a trend. Roger Pielke Jr. won’t discuss whether or not he expects this trend to continue – now, why is that?
Talk is cheap – where’s your data?
#207 Barton You are fixated on the range. Try dealing with the mean and the standard distribution. Estimates cluster around 3 K. The IPCC is citing a range when they say 1.5-4.5 K. They’re not saying that any possible value in that range is of equal likelihood.
I read all chapters of AR4 second draft relative to this estimate of climate sensitivity, and I find nowhere a clear explanation for the statistical basis of the “best estimate” (near 3 K) in the SPM. Maybe I miss the reference, if you get it, I’m interested.
In fact, 3,2 K (mentioned in chapter -8) seems the aritmetic mean of the 23 models in table 8.8.1. But that’s not really conclusive as all these GCMs are supposed to be equally efficient for simulating and projecting climate, so a 2,1 K result does not rule out a 4,4 K result (and the mean of these results does not tell me which model better implements physical processes of climate).
#218 Eli Regarding the range of estimates for carbon sensitivity, one should compare the range to the “natural” greenhouse effect for ~280 ppm, about 30 K, which makes the variation in the estimate about 10%. That the range has not moved much over 100 years and with increasing sophistication is a measure of the robustness of the result. As Gavin points out, the fact that 0 to negative response is ruled out (indeed that
Your comment was apparently cut as you posted. I don’t understand the first point about natural GH effect.
Re #263: […one of the most disturbing aspects of the denialists presentation was how readily Lindzen launched into the “warming in Mars, Pluto, Triton…)”. There’s not an iota of truth in most of that…]
Worse yet, in the one instance where planetary warming is indisputable – Pluto – there’s a reason, utterly unrelated to solar changes, that should have been obvious to anyone who’s had even a decent high school science class, let alone a professional scientist. Pluto has a very eccentric orbit: it’s been closer to the sun for the last few decades, so of course it warms up. Duh!
As for the rest, I see and I wonder. These planets are hundreds of millions of miles away; in most cases, we have only scattered telescope information for more than the last decade or so, and yet the same people who on such scanty evidence claim that these other planets are warming will with their next breath say that we have insufficient records of Earth’s climate to prove that warming is happening.
Watching the contraian news, one sees various scientists quoted
against global warming. It would be helpful to have a survey
of who these people are and what arguments they are making.
Re: 265
Gavin:
In terms of the Pinatubo analysis, thanks for the Wigley reference, I’ll dig it up and read it. Once I’ve done my homework on this, I’ll cycle back. If you put your ‘outside the debate’ hat on, would you say that “K&D is rubbish” is a universal point of view among serious scientists, or would you characterize there being a legitimate debate on this point? I do agree that the premise that a natural experiment which validates pre-existing model predictions (ie, those made prior to the eruption) leads to the conclusion that such models have passed a useful falsification test.
I believe I understand the reiterated distinction you are drawing on the evaluation of the 1988 forecast. In effect, I think you are saying there is difference between showing that a value is statistically different than zero and showing that a value is within some confidence interval of an estimate. But if we all agree that global temperature, sensitivity to CO2 > 0 based on the radiative physics, then the issue of interest is exactly the confidence intervals, at various levels of certainty, around the estimate produced by a given model / scenario under consideration. Said differently (though without quantification of the term “precise assessment”): “a 17-year period is too brief for precise assessment of model predictions, but … comparison with the real world will become clearer within a decade”. Am I missing something here? (asked seriously, not rhetorically)
Gavin, thanks again for spending so much time on this given all of the professional work you are doing in this area.
Best,
Dana
[Response: Nothing is universal, but I’d be very surprised if anyone sensible cited the K&D paper approvingly. I understand your comment about the uncertainty estimates – however, the transient response of the model to ongoing forcing is not really a simple function of the sensitivity. At best what can be said here is that a model with a sensitivity of ~4 deg C (if I recall correctly) gives responses that are within the uncertainties of the measurements – given the level of interannual ‘noise’ in both the model run and the observations. I will do this point justice at some point – patience! – gavin]
re 262-cat black says ….Recommendation: Can we add a moderation function to the comment software?….”
It’s not nearly obvious what the filter’s criterea would be. Uses a shrill tone? Every now and then expert scientists get shrill. No political flavor? AGW has a political flavor that occassionally is helpful to address; when it goes solidly into more distant related politics, it probably ought to be properly filtered out, but there’s no decent software that comes close to the moderators. (BTW, a big hand for the moderators who spend mucho time doing a very credible job of it.) Block the agin-ers? Also a very difficult massive algorithm; plus you end up with a boring circle-jerk.
Dana wrote : [[I don’t think these are good examples of global temperature forecasts (the issue, as I see it, under consideration).]]
What you are saying does not make sense… If you have read the threads, you know how climate models differ from weather models.
Here is some more reading from this site if you are really interested.
https://www.realclimate.org/index.php/archives/2006/03/climate-sensitivity-plus-a-change/
https://www.realclimate.org/index.php/archives/2005/07/climate-sensitivity-and-aerosol-forcings/
https://www.realclimate.org/index.php/archives/2005/06/make-your-own-forecasts-of-future-energy-carbon-emissions-and-climate/
https://www.realclimate.org/index.php/archives/2005/05/planetary-energy-imbalance/
https://www.realclimate.org/index.php/archives/2005/03/how-long-will-global-warming-last/
https://www.realclimate.org/index.php/archives/2005/01/climatepredictionnet-climate-challenges-and-climate-sensitivity/
https://www.realclimate.org/index.php/archives/2004/12/temperaturevariations-in-past-centuries-and-the-so-called-hockey-stick/
https://www.realclimate.org/index.php/archives/2006/08/climate-feedbacks/
Re: 265
Ike, let me see if I can address your issues / questions one-by-one:
POINT 1:
IS: Dana, you say that: “These aggregated processes are represented within models as parameters that are modeler-created, rather than being based on physical measurements.”
That’s not true; the parameterization schemes are indeed based on physical measurements. This has been gone through before. That is a distortion of how the process is carried out, though it is true that some parameterizations are on firmer ground than others. You also skip the more relevant point:
“Thus, if you are going to make such statements, and expect to be taken seriously, provide some references, if you can.”
Don’t you think that if the contrarians could demonstrate disagreement between models and observations, they would have published such results? They tried to make the claim that troposphere temperature trends disagreed with model predictions, but that was clarified by correct analysis of the satellite data. They have nothing to go off, and if they did they certainly would have published it – so where is the evidence that the climate models are not agreeing with observations?
Let’s also look at the upper tropospheric moistening in comparison to model predictions: see http://www.colorado.edu/chemistry/chem5151/McBride.pdf (Radiative signature of upper tropospheric moistening, Soden et al 2005) Looks like the model is tracking it with good accuracy, yes?
RESPONSE TO POINT 1:
DJ:Iâ??m looking for something pretty specific: an example of a published multi-year global temperature forecast with a comparison of prediction to observation. Note, not predictions of relative geographic or seasonal trends or anything else. (As I’ve said earlier, to specify this somewhat more precisely, a copy of a model plus operational scripts escrowed at the time of prediction and a comparison of this model output with actual temperature at some future data when presented the actual input data as of that later date – the standard method of model validation.)
I agree that if contrarians had demonstrated examples of prediction -observation disagreement they would be published. I’m enough of a scientist to believe that most scientists, whatever their a priori beliefs, would publish such findings. But think about your question. If there are no published examples of prediction – observation disagreement, and there are no examples of prediction – observation AGREEMENT (to be fair, we are debating one possible example, though I don’t believe any fair-minded observer would characterize the evaluation of the 1988 Hansen forecasts as either agreement or disagreement as of today), then what does that say about validation studies that have been conducted of climate models? There are no published examples of either one. There has only been hindcasting analysis, and no true prediction validation. In a normal simulation modeling field there would be hundreds of such studies. Why not here?
I can think of at least one good potential reason why not. If it takes 30+ years to validate a model because that’s how long it takes for the phenomena to become evident at a level distinguishable from noise, we would have to wait until 2037 to validate a model built today, and if very dire forecasts are acurate, our goose would already be cooked by then.
If that’s the argument, fine. But there is a huge difference between making a forecast using a tool that has actually been validated as predicting with some defined accuracy vs. one that has not. Note that this doesn’t mean I think that any specific model is wrong or, more extremely, that this somehow ‘disproves’ AGW. If it’s correct that we are using non-validated models, however, we should just be somewhat more humble about our predictions.
POINT 2:
IS: Let’s also consider the current trend of new temperature records, and compare it to what Richard Lindzen said in Congressional testimony ten years ago:
One of the common claims in support of the reality and seriousness of global warming is that we have had a large portion of record breaking warm years during the last decade or so. This is not a claim used by the IPCC, and its presence in any discussion is a rather clear piece of evidence of the intent to deceive (especially when the claim is made by a scientist). As noted by Solow and Broadus (1989) and Bassett (1992), this is an inevitable occurrence when one has a single record breaker in a time series characterized by interannual variability, interdecadal variability and an underlying trend or longer period variability. Solow and Broadus show the clustered nature of record breakers….Not surprisingly, record breakers cluster in exactly the manner found by Solow and Broadus (1989) in the observed temperature record. The occurrence of such record breakers contributes no additional information. Our prime concern remains with the determination of trend and the identification of such trends with emissions of carbon dioxide, and this remains a difficult and contested issue as the IPCC freely acknowledges..
Now we have ten more years of temperature records, and Lindzen no longer mentions the topic. Hmmm… what does that indicate? That the temperature record trend can no longer be justified on the basic of ‘statistical clustering’. In fact, this is an example of how a large number of isolated events, none of which can individually be attributed to global warming, in sum represent a trend. Roger Pielke Jr. won’t discuss whether or not he expects this trend to continue – now, why is that?
RESPONSE TO POINT 2:
DJ: I donâ??t believe that either Richard Lindzen or Roger Pielke have validated the ability to make accurate multi-year global temperature predictions.
POINT 3:
IS: Talk is cheap – where’s your data?
RESPONSE TO POINT 3:
DJ: I think you have this exactly backwards. I don’t claim to be able to predict global temperatures 100 years from now. In fact, I think this is an example of a very hard scientific problem. If you believe that you can do this where is your data? Specifically, to go back to the original point, please provide validated examples of accurate multi-year global temperature forecasts.
Best,
Dana
Re: 270
Gavin, looking forward to the detailed treatment when you get a chance.
Best,
Dana
Re: 272
Richard,
I think I understand the distinction between weather and climate, and between weather models and climate models.
What I meant by the clause you quote
”
I don’t think these are good examples of global temperature forecasts (the issue, as I see it, under consideration
”
was that a salient output of climate models is a prediction of global temperatures under various carbon emissions scenarios. I was, an am, searching for validated examples of accurate global temperature predictions, as opposed to predictions of geographic or seasonal temperature differences. I’ve specific previously that, of course, the “prediction” would really be an escrowed version of the model at the time of prediction plus the actual imput data at the time of validation.
I have read the threads that you recommend, and while all interesting and valuable, they don’t really go to this precise point (as far as I can tell).
Best,
Dana
Dana, you understand I hope that the prediction has been for increased _variability_ with a longterm warming trend — and that’s what’s being observed thus far. What beyond that do you want, that you think the physics is able to produce? You’re asking for some specific longterm weather prediction, or what?
Why don’t you consider the biological changes, or ocean alkalinity changes, as evidence?
I’m not sure what other posters have written and do not have time to look but, if you want to know, I will tell you why I think the audience responded more favorably to the other side.
You did not give any examples. The other side had plenty, almost completely uncontested by you guys. You basically told the audience that “scientists” have decreed this and that without giving any reasons why, and insinuated that they were too dumb to understand it. You’ve got to get into the nitty-gritty and, you’ve got to present it in a way that they can understand it. That does not mean “dumbed-down.” It means, in a way that meshes with their personal experience.
I once had a manager tell me, in my younger days when I was fuming over the customer’s inability to understand what I was telling them, that if I couldn’t put it in terms that most men on the street could understand, then I didn’t really understand it myself. Most men (and women) on the street have a good store of common sense and, if you break your argument down into the essentials, you can persuade them. Too technical is bad, but not technical enough is bad, too. I mean, you had an audience that came in largely expecting to emerge on your side and, you lost them. Do you think that is their fault? Why?
Oh, and then you brought in “the children.” I mean, how hackneyed can you get? I bet the eye-rolling meter was pegged on that one. And, when you said “I don’t think that they are completely…doing this on a level playing field that the people here will understand,”, well, that really gave the game away. Even your post here gives off an odor of sour grapes, [edit]
Why in the world do you think you can succeed like that? I can’t imagine why you expected this kind of presentation to be received favorably. For the life of me, I really can’t.
Comment by Hank Roberts> “…you understand I hope that the prediction has been for increased _variability_ with a longterm warming trend…”
Is there really a credible prediction for increased _variability_?
Or does the longterm warming trend just cause more high temperature records?
Re: 276
Hank:
Thanks for your post.
OK, well what I’m still looking for is a model prediction made on date X for future global temperatures that was subsequently validated or falsified on some date after X. In principle, that could be, for example, a warming trend, cooling trend or inreased variability with no underlying trend at all.
I think there is a ton of evidence for AGW from biological and other sources, if that’s what you’re getting at.
Best,
Dana
Every once in a while, a key turn of phrase on the part of the good guys, a key step out of bounds on the part of the bad guys, does have the power to completely turn around a public debate. Everybody who has ever been involved in a public debate over critical issues hopes for a moment like Joseph Welch’s obliteration of Joe McCarthy, with the phrase “Have you no sense of decency, sir, at long last?” That moment clearly didn’t happen in the global warming debate under discussion here, and maybe it will never happen for a problem like global warming. I do think that this case too, in the end, comes down to a fundamental lack of decency, a fundamental shamelessness, on the part of opponents like Lindzen or Crichton. My hat is off to Gavin, Brenda and Richard, who I think did about as well as can be done playing the science straight, but the response shows that some other tactic is necessary to engage the hearts as well as minds of the audience. I’m not, of course, suggesting that one play crooked with the science the way Lindzen does, but that tactics which play more to peoples’ feelings, tactics which even extend to ridicule of opponents where they deserve ridicule, may be needed to win in fora like this one. I’m not sure what such tactics would look like, but I doubt many scientists have the requisite theatrical skills.
Re #277: […if I couldn’t put it in terms that most men on the street could understand, then I didn’t really understand it myself.]
Either that, or you don’t understand the man in the street :-)
Re #279: [OK, well what I’m still looking for is a model prediction made on date X for future global temperatures…]
A lot of people seem to get this backwards: it isn’t evidence of warming that’s motivated people to come up with theories and models, it’s those theories (grounded in physical laws and very sound facts) that predict warming. So if I can make a “man on the street” sort of analogy, you’ve just jumped off a cliff. We can use Newton’s Laws and aerodynamics to construct a computer model that’ll predict how hard you’ll hit. Of course we may not have exact figures for the height of the cliff or the air temperature, you might tumble a bit on the way down, or even encounter an updraft. And of course our theory isn’t perfect: we’re not taking into account relativistic velocity corrections, for one. So with all that, there’s a good chance our model is wrong, and so our prediction of your terminal velocity may be somewhat in error.
Now if you were to follow the lead of the AGW denialists, you’d claim that because our gravitational model isn’t perfectly accurate, we can all go around jumping off cliffs and never worry about the consequences. Hit bottom, you say? But falling is just a conspiracy of the liberal gravitational scientists, so they can keep on getting research grants :-)
Re #281 Jumping off a cliff (James)
James, that’s a great analogy.
You might have added that our computer models has been able to accurately reproduce the measured velocity of the person falling down for the first 3 metres or so. Some contrarians are claiming that no, it’s not Earth’s gravity, it is the moon that determining what is happening. Or the moisture content of the air.
Re. #279, see here.
Re 281 , 282.
Pity the result will be the same for modeller and contrarian alike: splat.
Unless the modeller had the foresight to attach a bungee cord, using the model to estimate the length needed (where length is less than the distance to splat point). In which case there will be a certain number of ups and downs, but s/he will be left dangling; albeit lower than the original starting point, but alive. And in a position to rectify the situation.
Two points to add to the debate:
1. I read Chrichton’s State of Fear with astonishment – not at the science, but at the basic premise. I couldn’t help wondering if he planned the book expecting Al Gore not George W Bush to be in the White House. The idea that the first Bush government was using Climate Change as a means to keep the population in a ‘state of fear’ was like suggesting that [edited–sorry, this was a bit inflammatory].
2. It was interesting to watch the film ‘Amazing Grace’ on Wilberforce’s battle to abolish the slave trade in the light of the GW discussions: notably the arguement of the anti-abolitionists that abolition would ruin the economy of the British Empire. Funnily enough, it didn’t. Post 1807 the Empire continued to rise in power and wealth for another century.
re 278: Steve, Consider that what you are doing is putting energy into a chaotic system. This has the effect of not only raising the temperature (averaged over the entire system), but also of increasing the phase space available to the system (hence increasing its variability). This is just physics, and there’s no need to even refer to a model here. Now, a chaotic system may have regions of phase space that are quasi-stable–i.e. a small perturbation does not lead to a completely different part of phase space, but rather to a nearby point. The past 10000 years of exceptional climatic stability could possibly represent such a region. If so (and I admit it is a reasonable-sized if), then the question is how much energy do you need to add before we fly away from the region of quasi-stability and become very unpredictable.
>279, Dana
Dave’s pointer is the right answer. Read it through carefully. If you don’t understand it, the AIP History of Global Warming (also linked in the right hand column of the main page) is a good explanation.
You can always argue that an answer isn’t precise enough, or global enough, or statistically significant enough to suit you. Nature isn’t that picky, from all the evidence in so far — it’s happening as expected.
All:
Sorry for my seeming obsession with the question of validated predictions, and sorry if I’m semi-hijacking this thread. So, I’ll make just one more post on this subject and then stop.
Consider an analogy to climate models. Instead of a very smart group of thousands of scientists and engineers who assert that they have built a piece of software that can accurately predict the global temperature impact over several decades of increased GHG concentration, imagine that this same group of people assert that they have built a car with a new kind of engine that can go 1,000 miles per hour.
Now, I think it would be natural for an outside observer to say ‘great, where is the data from the road tests where this car has actually achieved that speed?’ When the response is ‘here are thousands of scientists who agree that this car can go 1,000 miles per hour, and pretty much nobody who knows about this kind of engine disagrees’ and ‘let me show you the data on the engine tests that demonstrate that the relative rotation speed of part A is 47 times that of part B, which is consistent with going 1,000 miles per hour’ and ‘here are computer simulations of this engine that demonstrate that it goes 1,000 miles per hour’ and ‘let me explain the physics of this device’ etc., it does lead naturally to the question ‘OK, all of that makes sense, but why don’t you just take it out on a track and drive it at 1,000 miles per hour and then we’ll all know for sure?’.
By analogy, it seems sensible to ask for demonstrations that climate models can actually perform a specific, crucial advertised task: predict the multi-decadal global temperature impact of increasing levels of GHGs. In any technical field with which I’m familiar that uses complex simulation models, the lierature has hundreds of validation studies that do exactly this.
Let me reiterate that this question about model validation doesn’t mean that I deny AGW (I don’t) or that I think the models have been proven to be wrong (I don’t) or that I think there is some nefarious conspiracy among climate scientists (I don’t) or that I know of anybody who has a reliable alternative method for preciting long-term climate change (I don’t). I just think that predicting the climate 100 years from now is an incredibly hard scientific problem, and therefore demands real rigor in evaluation of claims.
Thanks for all for your time and attention. This discussion has been very illuminating for me, and really helped to clarify my thinking.
Best,
Dana
Re #288 (Dana)
[By analogy, it seems sensible to ask for demonstrations that climate models can actually perform a specific, crucial advertised task: predict the multi-decadal global temperature impact of increasing levels of GHGs. In any technical field with which I’m familiar that uses complex simulation models, the lierature has hundreds of validation studies that do exactly this]
Dana: honestly, I don’t understand your question. If measured GHG concentrations are fed into the models, they reproduce the temperature record 1850-2000 very well, both in time and space, and a lot of other things besides (and this is not just a matter of curve fitting.)
What more do you want?
PS Dave Rado had a link to the two most compelling graphs (temperature and attribution) somewhere in a recent thread, but I can’t find that post just now.
Re # 288
“I just think that predicting the climate 100 years from now is an incredibly hard scientific problem, and therefore demands real rigor in evaluation of claims.”
I’m curious – do you think that rigor has been lacking in the peer-reviewed papers on this subject, or that the climate modelers somehow don’t recognize the difficulty of modelling global climate and projecting future temperature increases?
Re #288: [By analogy, it seems sensible to ask for demonstrations that climate models can actually perform a specific, crucial advertised task: predict the multi-decadal global temperature impact of increasing levels of GHGs.]
I think you’ll find that your question contains its answer: “multi-decadal”. So you’re suggesting that the modellers make predictions now, and we all wait around doing business as usual for several decades to find out if the predictions are correct, no? By which time, if the predictions do turn out to be correct, it’s probably too late to do anything.
On the other hand, if we proceed on the assumption that the models are right, many of the actions that would help mitigate the problem are relatively inexpensive (some even result in savings), and give us cleaner air & water to boot. Seems like a pretty clear choice to me.
Imagine your engineers promising a 1000-mph vehicle
— Over more than a century, kept at it:
http://www.roadsters.com/welcome/
— Stepwise through the years, kept promising higher speeds.
— Kept showing their ideas worked, year after year
By 1997
http://www.roadsters.com/ssc_770x390.jpg
“… Monday, October 13th [1997],… two runs, at 764.168 miles per hour (Mach 1.007) and 760.135 miles per hour (Mach 1.003)”
And keep promising exactly what you specify:
http://www.sonicwind.com/page1.html
“… its design speed of 900 mph and possibly even to 1000 mph!”
Believe it?
http://en.wikipedia.org/wiki/Land_speed_record
Start with Arrhenius, in the AIP history, watch how it’s improved over the years.
Just read the transcript. You did an excellent job, Gavin. The audience already knew plenty about GW, so it was good you focused on the contrarian’s fallacious arguments. Great.
I do really resent the program’s frame – “GW is NOT a crisis.” That may be the null hypothesis for false-positive avoiding scientists (and the contrarians who need 99% or 101% confidence before rejecting), but it is not the position regular people should be taking — unless they are very old and very selfish, and don’t give a %*&$#@ about the kids or grandkids, or where they’ll spend eternity. Also “crisis” might be too harsh a term, since I look outside and the sun is shining and birds are singing, and it’s hard to even find a storm cloud on the horizon. The word “problem” may have gotten more votes against that stupid proposition.
But at least the contrarians haven’t started arguing “it’s happening, but it’s too late to do anything about it” — which will be their next argument.
The contrarians have badly abused this Galileo & paradigm shift idea. They claim as Galilean the position “the world (climate) is as it’s always been, amen, totally flat,” while characterizing the “AGW is a new phenom based on principles we now understand” as anti-Galilean. AGW, not climate-as-usual, is the big paradigm shift, which even the believers are struggling to grasp.
Eventually, many years after Galileo, most of the world came to believe his position. So one would expect that 100+ years after Arrhenius first proposed the AGW hypothesis, that lots of scientists would come to his position, and the proof is in the pudding of mounting evidence.
As a member of NRDC (& ED, etc), I resent Crichton’s criticisms. For one thing, NRDC build a very “green” building in NY. And since when did I own a private jet? And how is it that not switching to CF bulbs and turning off lights not in use will solve the problem? AGW is a problem with thousands of solutions – and we need to start implementing one after another ASAP. No measure is too small to implement. (If everyone put out just one little candle, what a cool world this would be.) Everyone needs to do everything, including carrying a hanky to wipe hands in public restrooms. We need to see our electric bills cut in half by all our measures, and then by another half, then get on Green Mountain 100% wind energy (they’re having a lock-in sale right now at 2.5 cents lower per KWH than the going rate).
And with all the money we save, we can send it over to help all those poor people. Another thing, my husband and I give 20 times more to such charities than we do to NRDC, ED, and RMI combined. I just don’t see how saving money from all that energy/resource conservation/efficiency is going to prevent us from giving to the poor, who as Somerville pointed out, are suffering much more greatly from AGW (even right now as I write) than us 1st world folks, who will also eventually become very poor & disease-ridden if we fail to mitigate — but there won’t be any rich people to provide us aid.
Re Stott’s discussion of “it’s mostly unknown” (e.g., re aerosols), here’s an analogy: There’s a rockslide up ahead, but it’s unknown if it impacted the railway tracks, ergo we should proceed full speed ahead on our train.
And Lindzen’s emphasis on how the impact per unit CO2 goes down…That’s UNLESS, of course, the warming it causes (1) releases xxxxxx gigatons of CH4 & CO2 from melting permafrost, ocean clathrates, GW-induced forest fires; and/or (2) impedes nature’s ability to reabsorb those GHGs.
Re: 289 – 292
I said that I’d stop, but you guys have raised some great points and questions, so at leat I’ll try to be brief. :)
Dick:
The distinction is one between ‘hindcasting’ and actual prediction.
In the rigorous version of hindcasting, I parameterize a model on data for, say, 1900 – 1950 and then use this model to simulate global temperatures for 1951 – 1980 and compare these simulated temperatures to actual. This is normally the first test applied to a predictive model in any field. Climate models have passed this test.
The next step is to make an actual series of predictions at time X for some future time Y and then when Y rolls around compare predictions to actual to create a distribution of prediction error. This has not been done for climate models It might seem like this is not that big a deal, but there are numerous examples in many fields of predictive models that do well on hold-out samples but fail in real trials.
Chuck:
In my experience, climate modelers are smart and dedicated. Some modelers have made trying to quantify model uncertainty a key research focus — James Annan is a prominent example.
There is an empirical tendency for all predictive modeling communities (not just climate modelers) to over-estimate accuracy in the absence of validation studies. I assume, but do not know, that climate modelers are typical of other scientific communities in this regard.
I think the lack of validation studies is most likely a function of how long it is believed to take to compare a prediction to actuals, as per my response to James below.
James:
I completely get your point. I said something very similar in a prior post on this thread.
My only point is a narrow one: we should know, when making these decisions, whether or not the predictive model we are using is validated (in the sense of validation I describe in my response to Dick) or non-validated.
Hank:
LOL — great post. My hypothetical was not that the group of engineers asserted they had developed the first car to go 1,000 miles per hours, just a car that could go 1,000 per hour. But consider your extensions to my analogy. If “road test” = “validated prediction” (in the sense of validated prediction as described in my response above to Dick), then in your analogy it would be as if climate modelers had first predicted accurately for 1 year, then 3 years, then 5 years, then 10 years, then 20 years and were now claiming that they had extended the predictive envelope to 30 years. But my point is that the models haven’t even done a road test that shows that they can 100 miles per hour, because they haven’t done any road tests at all.
Best,
Dana
I am a neophite trying to better understand global warming, so advanced apologies if this is naive. In looking at the 400kyr ice core data, it’s clear that there is correlation and some feedback loop between CO2 and T. What caused the high spikes in T every ~100kyrs and how are those factors relevant in today’s temp increases, if at all?
RE #174 & “I have never come across what I consider to be good evidence that global warming is a crisis, present or future,”
There’s a (I know it’s going to be) great book by Mark Lynas coming out now, SIX DEGREES. He’s done tremendous research and presents what each degree rise in warming means in terms of impacts and harms.
Even a few degrees means lots of harms. I’m even thinking with the way the glaciers are melting today, that if we just stayed at this current level of warmth even, that could mean no water for drinking, much less agriculture for a large chunk of the world’s population, once the glaciers that provide their water are all melted. So it looks like some bad things are in the works right now at our current level of warmth…
Same with droughts in Africa.
I think what the improved science will eventually be able to tell us is just how much harm has been caused over the past 30 or so years from AGW. Maybe we’ll find out in the next IPCC AR, or the one after that. It takes science a long time to reach 95% confidence on these various issues. And I, for one, am not waiting around for the definitive proof, but already into my 17th year of reducing my GHGs. And I think I could bring my husband on board more to help in this effort, if there were more of a community consensus out there that this is a problem and we need to address it.
So the crisis (or serious problems — whatever you wish to call it), I think, has already been with us for some time, and will just get worse.
>modelers … road test
Simplest model: Arrhenius. Results? In the ballpark (he defined the ballpark, admittedly)
Watch the developments from there, see the AIP History document.
Your simplest model is “add CO2” as Arrhenius did, estimating the result; he’s got the longest baseline so far, starting with the very simplest model, just the basic physics he discovered.
Each physicist along the historical track has known a bit more and had a bit more complex model, and from the 1950s has needed computers to do the math.
They’ve all said the same thing, to about the same magnitude, and all been correct so far within reasonable error bars.
It doesn’t have to be huge to be a model; Arrhenius did it first.
[[What caused the high spikes in T every ~100kyrs and how are those factors relevant in today’s temp increases, if at all? ]]
“Milankovic cycles.” The Earth’s orbit and axial tilt change cyclically over time, and that affects how sunlight is distributed over the Earth’s surface. These cycles seem to correlate well to the pattern of ice ages during the Pleistocene.
Thanks! It seems that each of the previous T spikes have gone higher than we are today. So how much of warming today is from Milankovic cycles versus anthropomophic CO2 increases? (not arguing against GW, but trying understand how much is natural vs artificial)
Re #299: Dan — Accoring to orbital forcing theory, the climate should be very slowly cooling now, heading for a try at a stade (ice sheets) in about 20,000 years. So based just on this, slightly more than 100% of the warming is anthropomorphic.
But given the physics, the fact that the carbon dioxide in the air is higher than any time in the last 600,000 years ought to be a source of serious concern. After all, Homo sapiens sapiens only evolved about 200,000 years ago. We are adapted to cold (global) climates…