From the world of Claimatology™, comes this smackdown from Nature.
“The dramatic warming predicted after 2008 has yet to arrive.”
An article published today in Nature laments the dismal failure of climate models to predict climate a mere 5 years into the future, much less a century from now.
- “It’s fair to say that the real world warmed even less than our forecast suggested,” [modeller] Smith says. “We don’t really understand at the moment why that is.” “
- “Although I have nothing against this endeavour as a research opportunity, the papers so far have mostly served as a ‘disproof of concept’,” says Gavin Schmidt. Schmidt says that these efforts are “a little misguided”. He argues that it is difficult to attribute success or failure to any particular parameter because the inherent unpredictability of weather and climate is built into both the Earth system and the models. “It doesn’t suggest any solutions,” he says.
- “Because the climate does not usually change drastically from one year to the next, the model is bound to start off predicting conditions that are close to reality. But that effect quickly wears off as the real climate evolves. If this is the source of the models’ accuracy, that advantage fades quickly after a few years.”
- “Kevin Trenberth, a climate scientist at the National Center for Atmospheric Research in Boulder, Colorado, says that it could be a decade or more before this research really begins to pay off in terms of predictive power, and even then climate scientists will be limited in what they can say about the future.”
Limited in what they can say about the future?
Since when? Somebody please tell Jim Hansen he can’t say “the oceans will boil“.
The Nature article is here: http://www.nature.com/news/climate-change-the-forecast-for-2018-is-cloudy-with-record-heat-1.13344
h/t to the Hockey Schtick.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Within the last 2 weeks the WMO announced that 10 years was good enough to see changes but others say that the 15+ years of temperature standstill is not long enough – because a COUPLE of the climate models
predictedprojected it. THIS IS A FARCE.phodges says: July 11, 2013 at 4:24 pm
“…Is it just me, or isn’t this a huge bombshell???…”
I’m with you on that.
Surprosed this did not rate a sticky … all these leading alarmist umming and ahing, not sure how to explain the situation is a pretty huge statement on the real state of the science: ie unsettled! 🙂
Latitude says:
July 11, 2013 at 7:47 am
no…..the computer game does what it was programmed to do
Anyone who plays a computer game or two knows that there are still bugs and exploits in it, almost every game suffers from this, and that it needs regular updates to adress those bugs and exploits.
Just the suggestion that these (climate modelling) programs can’t be wrong …. unless you want to cheat at the game you are playing and start using those exploits and bug to obtain the upper hand in that game.
Patrick B says:
July 11, 2013 at 10:33 am
@Ferd berple
“The spaghetti graph of the climate models are telling us quite the opposite.
No, no, no. The models only “tell” us this if the models correctly represent the earth’s climate. ============
A wrong answer teaches us much more than simply that the answer is wrong.
jorgekafkazar says:
July 11, 2013 at 8:58 am
Models are either stochastic or determinate.
=========
Deterministic chaotic system are effectively stochastic because they require infinite computational precision. Any round off and the results diverge from reality in a completely random fashion.
Which is what the climate models are showing us. Run the same model twice without any change in the forcings and it gives you a different answer. Run the model a billion time and it will give you a billion different answers, for the exact same combination of forcings.
How then can climate modellers turn around an say that doubling of CO2 will cause X degrees of warming, when their own models tell them this is a lie. The models tell them and us that doubling of CO2 may cause X degrees of warming, or it may cause Y degrees, or it may cause Z degrees.
The models do not say that (X + Y + Z) / 3 is the right answer. That is the model builders saying that. The modes say that X or Y or Z are ALL the right answer. Think about this for a minute. The models are not saying that the answer is X or Y or Z. The models are saying that X or Y or Z are all correct.
The reason we have such difficulty with this is that most people perceive the future as some place we are traveling to. Quantum mechanics tells us something quite different . There are an infinite number of futures, all or them perfectly valid under the laws of physics, and X or Y or Z exist in different futures, but they are all perfectly valid physically.
The climate models are simply confirming something physics has recognized for many years. We will arrive at one of these futures, where X or Y or Z (or some other answer) will be true, but which one is for all intents and purposes a throw of the dice, beyond our ability to calculate in our wildest creams.
I’m confused. Is “real climate” the one that actually exists or is it the fantasy climate?
But in 2007 the Met Office “predicted” that 50% of the decade from 2009 would be warmer than 1998.
Who said climate scientists don’t do predictions?
talldave2
‘Forecasts’ are done for weather by your local bureau. As no one can predict the future of GHG emissions, IPCC give an upper and lower bound for a range of scenarios, called projections. The full range of all scenarios is 1.1C to 6.4C increase by the decade 2090 – 2999 relative to 1980 – 1999.
Skepticism and criticism is vital, but useless if it doesn’t reflect the premises it is addressing.
With five coin tosses you can get a flawed result compared to the average you expect. Over time the results converge with the average. Not much surprise the same should hold true for climate models, with weather representing coin flips, but this should not be confused with the averaging process on weather and interannual variability regarding centennial models. Fluctuations even out over the long-term.
Note: The Nature article starts off with Dr. Dough Smith and the Met Office I mentioned above.
[I only noticed this after I posted my comment. :)]
It’s a pity that some incautious commentary can lead to misconceptions like this. WMO announced that 10 years is the absolute minimum to get a climate state, not a trend. They compared consecutive 10-year blocks, so their minimum period to detect a change in climate is actually 20 years – a comparison of 2 averaged data points. To assess if there is a trend would require at least 3 consecutive decadal averages – 30 years. This is, coincidentally, the classic climate period as defined by the WMO.
It may be that some people come away from this article believing the Nature article canned centennial-scale climate models, which is not the case.
If at first you don’t succeed try, try and try again. Just how much more failure can these idiotic Nintendo players take? Instead they should use 100 chimpanzees throwing dart boards and using the ensemble mean.
“Kevin Trenberth, a climate scientist at the National Center for Atmospheric Research in Boulder, Colorado, says that it could be a decade or more before this research really begins to pay off in terms of predictive power, and even then climate scientists will be limited in what they can say about the future.”
Interesting tacit admission that current models have no skill.
If one reads TFA, one notes that it only addresses the effort to predict short term climate, less than a decade. Here the models disagree with the long term climate models and actually have some skill, although they are still too warm.
What Trenberth is asserting is that the long term models have skill because they worked great to predict the temperature rise in the 80s and 90s (bearing in mind that they were built in order to do so). So he claims that the long term models are still skillful, but only in the longer terms, so that nature itself is “incorrect” and so are the short term models, it being only an accident that either one is lower than the long term climate predictions.
He could be correct, but he isn’t correct in asserting the model skill yet, because in predictive modeling the proof of any model isn’t in its ability to predict the training set, it is its ability to predict a trial set independent of the training set and more to the point, to successfully predict new data that is in neither training or trial sets, as this is where the money lives. This is, by the way, not an idle observation as I am a friggin’ professional expert in predictive modeling on my second company in the field.
A number of things speak to a probable lack of skill. One of them was mentioned in the article — even to predict the warming in the training set, the different models explained it with different physical mechanisms. The physics itself of the warming differed, just where one would have expected it to be the same. Also, the models failed (and continue to fail) to predict or explain important details in the observed climate system that one would expect them to get right. But still, Trenberth is right, in a decade or two or ten, the Earth could start to follow the trajectory that the current crop of GCMs sorta, mostly, sometimes, on average, perhaps suggest. (Yet another problem is that the GCMs individually and collectively produce a mass of predictive spaghetti that isn’t even consistent run to run within a single model, and that nature has been evolving on the outer envelope of the range of those model results, which is the way most of us would reject the null hypothesis that “this model is correct”, not assert that nature happens to be following a highly improbable trajectory.)
In the meantime, the article carefully fails to point this out. But that doesn’t matter — the fact that they are publishing it at all suggests that the light is slowly dawning in even the minds of the editors of Nature.
rgb
Excellent post RGB!
One comment about your phrase “…Yet another problem is that the GCMs individually and collectively produce a mass of predictive spaghetti that isn’t even consistent run to run within a single model…”
Perhaps, in accordance with your previous papers on the ability of the models this could instead be worded; “…Yet another problem is that the GCMs individually and collectively produce a mass of pseudo-predictive spaghetti that isn’t even consistent run to run within a single model…”?
It may be that some people come away from this article believing the Nature article canned centennial-scale climate models, which is not the case.
Sure the the models diverge from reality in 5 or 10 years, but, really, they are accurate in 100!!
Besides what RGB pointed out, there are numerous quotes in the article by climate scientists admitting the models are crap and they have no clue.
It is a miracle it made it into print at all, let alone in Nature.
The basic problem for society is the same worldwide now:
Fear of nuclear annihilation in Aug 1945 (and we came much closer than society knows) convinced world leaders to:
1. Form the UN on 24 October 1945
2. Hide knowledge of nuclear energy
3. Eliminate national constitutions
4. Eliminate national boundaries
5. Save themselves by tyrannical control of society in the manner George Orwell predicted in “1984.”
Orwell’s prediction was fulfilled on schedule but remained invisible to the public until Climategate emails were surriptiously released in Nov 2009.
With deep regrets,
Oliver K. Manuel
Former NASA Principal
Investigator for Apollo
PS: Another EMP (electro-magnetic pulse) from the Sun’s pulsar core – as happened in the late 1850′s – may be required to knock sense into world leaders and make them admit that they cannot enslave mankind nor control Earth’s climate.
In reply to:
It’s fair to say that the real world warmed even less than our forecast suggested,” [modeller] Smith says. “We don’t really understand at the moment why that is.”
William:
Come on guys. There are multiple failures/discrepancies (at least four) of the general circulation models and the CO2 forcing theory. The article should list all of the failures and should explain the logical implications of each discrepancy. An analysis of the four discrepancies indicates that the majority of the warming in the last 70 years has due to solar magnetic cycle modulation of planetary clouds rather than the increase in atmospheric CO2.
1) There has been a warming plateau for 16 years which is a more serious discrepancy than a lack of warming. As atmospheric CO2 has been increasing continually for the last 16 years and as there is a lag from an increase in forcing to a change in temperature, the general circulation models predict a wiggle increase in planetary temperature where the wiggles are caused by natural variability in the climate and the gradual increase is caused by the increased forcing due to the increase in atmospheric CO2. What is observed is not only a lack of warming but rather a plateau of planetary temperature. … ….The logical constraint on the forcing mechanisms is different for a lack of warming (planetary temperature is still increasing but less than the general circulation models prediction) and a plateau where there is no increase in planetary temperature. … ….Aerosols or heat hiding in the ocean could explain a lack of warming, where planetary temperatures are increasing but less than model predictions, they cannot explain a plateau of warming. … ….The CO2 forcing mechanism cannot be turned off, if it is real. As atmospheric CO2 is continually increasing the aerosols or the heat hiding in the ocean would need to exactly balance the CO2 forcing and to start in 1998. i.e. There needs to be a mechanism that hides the CO2 forcing that is suddenly turned on in 1998 and that hiding or cooling mechanism must increase overtime to create the observed plateau in planetary temperature. In addition the hiding or cooling mechanism needs be activated quickly in 1998 as there is no period of gradually increasing temperatures. To explain a plateau where there is no increase in planetary temperature during a period when atmospheric CO2 is steadily increasing, the CO2 mechanism must saturate. The logical constraints of the other observational discrepancies indicate that the CO2 mechanism saturated at say 200 ppm, less than the pre-industrial atmospheric CO2 level, 280 ppm.
Climate Expert von Storch: Why Is Global Warming Stagnating? (William: Why did global warming plateau in 1998?)
Lack of warming
http://www.spiegel.de/international/world/interview-hans-von-storch-on-problems-with-climate-change-models-a-906721.html
2) Observed warming is not global. As atmospheric CO2 is evenly distributed in the atmosphere (less than 5% variance by latitude) the potential for CO2 forcing in the atmosphere is roughly the same for all latitudes. The actual forcing at the latitude in question due to the increase in atmospheric CO2 should be proportional to the long wave radiation that is emitted off to space prior to the increase in atmospheric CO2. Based on how the CO2 mechanism works therefore the observed warming should be global with most of the warming occurring in the tropics, based on the CO2 theory. What is observed is that the Northern hemisphere ex-tropics experienced twice as much warming as the planet as a whole and four times as much warming as the tropics. Furthermore the Greenland Ice sheet experienced the most warming on the planet (2 to 3C). The observed latitudinal pattern of pattern cannot be explained by the CO2 mechanism.
Comment:
The aerosol forcing is highest in the Northern Hemisphere. If aerosol where the reason for a plateau in warming the Northern Hemisphere should have warmed less than the global as a whole. The aerosol forcing does not explain why the Greenland Ice sheet warmed the most of any region on the planet or why there is very little warming in the tropics.
http://arxiv.org/ftp/arxiv/papers/0809/0809.0581.pdf
Limits on CO2 Climate Forcing from Recent Temperature Data of Earth
The global atmospheric temperature anomalies of Earth reached a maximum in 1998 which has not been exceeded during the subsequent 10 years (William: 16 years and counting). The global anomalies are calculated from the average of climate effects occurring in the tropical and the extratropical latitude bands. El Niño/La Niña effects in the tropical band are shown to explain the 1998 maximum while variations in the background of the global anomalies largely come from climate effects in the northern extratropics. These effects do not have the signature associated with CO2 climate forcing. (William: This observation indicates something is fundamental incorrect with the IPCC models, likely negative feedback in the tropics due to increased or decreased planetary cloud cover to resist forcing). However, the data show a small underlying positive trend that is consistent with CO2 climate forcing with no-feedback. (William: This indicates a significant portion of the 20th century warming has due to something rather than CO2 forcing.)
3) Tropical tropospheric hot spot is missing. The general circulation models predict that the most amount of warming in terms of change in temperature should be in the tropics in the troposphere at around 8km above the surface of the planet. This warming is due to increased water vapor due to warming of the tropics by the CO2. The warming at 8 km if it occurred would amplify the CO2 forcing. There is no observed tropical tropospheric warming. In addition, analysis of changes in radiation when there is a change in ocean temperature indicates the planet resists forcing changes by an increase or decrease in planetary cloud cover in the tropics which results in more or less radiation being reflected off to space, rather than amplifies forcing changes. These two discrepancies are logically supportive. Part of the reason why there is no tropical tropospheric hot spot is that the clouds in the tropics increase or decrease to resist forcing. Part of the reason why there has not been significant warming in the tropics is the cloud change in the tropics to resist forcing.
http://joannenova.com.au/2012/05/models-get-the-core-assumptions-wrong-the-hot-spot-is-missing/
Roy Spencer: Ocean surface temperature is not warming in the tropics.
http://www.drroyspencer.com/wp-content/uploads/TMI-SST-20N-20S.png
http://www.drroyspencer.com/wp-content/uploads/TMI-SST-MEI-adj-vs-CMIP5-20N-20S-thru-2015.png
http://www.drroyspencer.com/2013/02/tropical-ssts-since-1998-latest-climate-models-warm-3x-too-fast/
http://icecap.us/images/uploads/DOUGLASPAPER.pdf
A comparison of tropical temperature trends with model predictions
Lindzen and Choi have again found that the planet resists climate forcing changes
http://www.johnstonanalytics.com/yahoo_site_admin/assets/docs/LindzenChoi2011.235213033.pdf
4) There are cycles of warming and cooling in the paleo-climatic record that correlate with solar magnetic cycle changes. The same regions of the planet that warmed in the last 70 years are the same regions that warmed in the past. An example of the past cyclic warming is the Medieval Warm period which was followed by the Little Ice age when the solar magnetic cycle entered the Maunder minimum. The past warming and cooling cycles were not caused by changes to atmospheric CO2. There general circulation models cannot produce the past warming and cooling cycles that is observed.
Greenland ice temperature, last 11,000 years determined from ice core analysis, Richard Alley’s paper.
http://www.climate4you.com/images/GISP2%20TemperatureSince10700%20BP%20with%20CO2%20from%20EPICA%20DomeC.gif
http://www.climate4you.com/
http://www.nature.com/news/climate-forecasting-a-break-in-the-clouds-1.10593
RGB!
One comment about your phrase “…Yet another problem is that the GCMs individually and collectively produce a mass of predictive spaghetti that isn’t even consistent run to run within a single model…”
===================================
Well I thjnk the models are consistent, indeed, remarkably consistent. The are consistently TOO WARM, They are consistently wrong in one direction. (To a normal person this may cause them to look for one component, consistent in all the models, that ALWAYS produces too much warming.) That component is senstivity to CO2.
And here is Richard Feynman on the Scientific Method:
Between you and me, no matter what your operative term is, “disproof of concept” equals WRONG. (Thank you R. Milhous Nixon for that…)
In exactly the same way, the operative term “A little misguided” also still means WRONG.
But notice how Gavin doesn’t admit that it is REALITY that has “mostly served as a ‘disproof of concept’.” He blames it on “the papers.”
And before I drop it, let’s get one thing straight about the word “disproof” and its antonym: Proving something correct is not possible in science.
Again, ask Feynman. A falsifiable experiment can prove a hypothesis WRONG, but it cannot prove something correct – because you may have not considered every possible situation, no matter how much you’ve tried. Someone centuries later may find a process or piece of evidence that does not work. Feynman brought up Mercury’s orbit, which took until the 1900s to find did not conform to Newton.
“It’s fair to say that the real world warmed even less than our forecast suggested,” [modeller] Smith says. “We don’t really understand at the moment why that is.”
Now this is just b.s. parsing of words.
Saying that “the real world warmed less” when in actuality the warming was zero or sightly negative is just mendacious lying – trying to put a non-negative spin on the whole “real world” evidence. If your starting “guess” (Feynman’s term) produces certain testable calculable predictions, and then “experience or experiment” disagrees, your guess was WRONG. And if you do NOT have testable predictions, then it is not science. Meaning that scientist are obligated to make those guesses – and in so doing, putting their intelligence on the line – in a put-up-or-shut-up mode of inquiry. (Thanks to Robert Hooke for that.) And if their guesses are shown to be wrong too many times, those scientists go into the round bin of history – and deservedly so.
barry says:
July 12, 2013 at 3:54 am
With five coin tosses you can get a flawed result compared to the average you expect. Over time the results converge with the average. Not much surprise the same should hold true for climate models,
=======
no. this is not true. Climate does not converge to an average, except in infinite time. The reasons are well established. the law of large numbers only holds if the data has a constant mean and deviation. which is true for a coin toss, but certainly not true for climate.
chaotic data by its nature does not converge to the average because it has many averages, one for every attractor. Because the attractors operate at different time scales the average changes as the scale changes. Only at infinity is does the average converge.
Thus, because climate is the average of chaos, like a coin toss if will eventually converge on the average, but unlike a coin toss this will not happen on 100 years of 1000 years. Rather the models will converge at infinity.
BTW, (from above)
“But notice how Gavin doesn’t admit that it is REALITY that has “mostly served as a ‘disproof of concept’.” He blames it on “the papers.”
See, now this is simply denying responsibility. It’s not the concept that is WRONG (disproved), it is the papers that are the bad guys. Not “the real world” and not the “concept” in the forst place; it is the PAPERS. Nasty old things, PAPERS that disprove concepts.
But here is the crux of it all:
Will Gavin now look for different concepts to explain and to put into the models? And TO TELL THE WORLD OF THE CHANGE OF MIND?
Don’t hold your breath. To admit wrongness, Gavin would need to fall on his sword. Ain’t a-gonna happen.
Chaos tells us the Trenberth was mostly right. the climate models may eventually track reality. but to do so in 80 years will simply be accidental, because the further you forecast into the future the more likely an unexpected attractor will throw your careful calculations out the windows.
Now folks will say these events are rate, but the reply is you are talking about predicting climate 80-90 years in the future. these rare, unexpected attractors become more and more likely the further into the future you try and predict.
It is like a coin toss where at some point in time the coin will develop a 3rd, 4th and 5th side. But you don’t know when. And you are then asked to predict the percentage of heads that will be thrown 90 years from now.
Why does not 1 climate model start with the assumption that water feedback is negative and see if this improves their ability to track current conditions?
Why does every climate model assume that feedback is positive when atmospheric moisture has been dropping as CO2 has been rising, opposite to the assumption of positive feedback?
Why not try the models with negative feedback and see it they still predict the hot spot? Why not try the models with negative feedback and see if they still predict too high?
Wouldn’t this be a much more reasonable approach than continuing to stick with the assumption that feedback is positive in light of the evidence that the current models are all predicting too high?
What is the big deal in adjusting the parameters to try some model runs with negative feedback? Are the climate modellers so afraid that the results just might fit reality better than the assumption that feedback is positive?
Positive feedback has never been proven. It is simply a guess and it has been contradicted by the physical evidence. Isn’t it high time to test some alternatives that better fit the observed data?
Why oh why are the climate modellers continuing to stone wall on this subject and waste billions of dollars on model runs that show no skill, yet fail to use the models to test the alternative hypothesis?
Why does everyone ignore the elephant in the room? Positive feedback is a failed hypothesis. CO2 warming isn’t the failed hypothesis. Rather the failure is the assumption that water amplifies CO2 in the way that the models all assume.
Insanity: doing the same thing over and over again and expecting different results.
Albert Einstein
It’s 10 years till models can be reality-tested, and always will be.