Bishop Hill Von Storch: models are falsified
Roddy Campbell points us to this discussion paper by Hans von Storch et al, which looks at the divergence between climate models and observations:
In recent years, the increase in near-surface global annual mean temperatures has emerged as considerably smaller than many had expected. We investigate whether this can be explained by contemporary climate change scenarios. In contrast to earlier analyses for a ten-year period that indicated consistency between models and observations at the 5% confidence level, we find that the continued warming stagnation over fifteen years, from 1998 -2012, is no longer consistent with model projections even at the 2% confidence level. Of the possible causes of the inconsistency, the underestimation of internal natural climate variability on decadal time scales is a plausible candidate, but the influence of unaccounted external forcing factors or an overestimation of the model sensitivity to elevated greenhouse gas concentrations cannot be ruled out. The first cause would have little impact of the expectations of longer term anthropogenic climate change, but the second and particularly the third would.
That seems quite important to me. Note also that "the heat's all in the deep oceans" doesn't seem to be on the table as an explanation.
Reader Comments (71)
The "models" are based on policy.(policy based science).
And the policy is UNFCCC.
So this means that the UNFCCC has been falisified by the real Natural world?
I disagree with von Storch's statement about the consequences of greater natural variability. It does change the expected consequences in that the uncertainty grows considerably - including the possibility that the 20th cent warming was primarily natural variability. This unquestionably increases the value of adaptation over mitigation (since it implies that we may get large climate change even with successful mitigation of GHG emissions, so adaptation is the best way forward).
Given that weather prediction is so problematic that any forecast over 72 hours is virtual worthless , its hardly a surprise to find climate predication, which is made of weather on large time frame, is equally problematic. The real surprise is the 'settled science' statement was made in the first place with mad claims of accuracy to two decimal places for 50 years in the future . That is happen at all merely shows that is 'politics' and religions like zeal, not science, that drives 'the cause '
It seems to me that cause 1 and 3 are linked through attempts to model natural variability as if it were completely driven by CO2, therefore I think the conclusion that 1 has little implication is odd.
I think that it is very encouraging that a heavy weight like von Storch is articulating this. Perhaps this is crack in dam of the academic climate modelling.
Now, that's going to ruffle some Team feathers!
The heat-in-my-pants hypothesis is not on the table as an explanation because it shouldn't be. It is not only a hypothesis but a wild, desperate one at that. Go test it, sure, but don't make any use of it whatsoever.
Strange how they do not consider the possibility that the models are worthless and cannot be used to predictt any future climate.
Von Storch disappoints
Gavin
The (AGW) hypothesis which proposes that increasing atmospheric CO2 = increase in Global Average Temperature has been proven wrong by actual field real-world observations. Planet earth has falsified the warmist scientists' pet theory about which they remain obstinate and hard headed refusing to admit that they were WRONG. I think they need to watch a Feynman lecture......or ten.
The Met Office has stated in its publications that climate models are the *only* way to predict how the climate will change.
I think the bottom line is that they can't predict future climate change.
Aug 12, 2013 at 11:10 PM | Martin A
...
A bone rolling African sangoma has better predictive powers than the UKMO!
Streetcred -
Yes, but is his bone programmed using the principles of physics?
Martin A
Can you suggest non-model based methods for predicting 21st century climate trends?
Entropic Man; tea leaves perhaps?
"Whether or not a later calibration of the CMIP5-models was undertaken is not known, but the CMIP3 models were run before the recent stagnation emerged."
von Storch et al
Note that the CMIP3 model runs were made in the 1990s.
Their forecasts would not have taken into account the increase in albedo. This is due to increased soot particles and sulphur due both to an increase in volcanic emissions and increased global coal burning. An Earthshine based albedo study showed an increase in albedo of 8WM^-2 in the decade from the mid 1990s to 2006.
They would also have missed the reduction in insolation due to the weakness of solar cycle 25.
Both of these changes would counteract the forecast warming trend, as observed. The problem would not be with the models, but with changes in the behaviour of the system which were not predictable at the time of the forecast.
In the light of these effects the real problem is not why the forecast increase in temperature did not happen. What should be asked is-
Why has the reduction in surface energy input not produced a cooling trend?
What is holding the temperatures at their present level?
"tea-leaves"
Hoi Polloi
Quite possibly. It would explain how Martin A and others can convince themselves that the 0.6 C/century warming trend has ceased, without having any forecasting of their own to back up the assertion.
Consider the possibility that humans may not possess the requisite knowledge and method to predict 21st century climate trends..... and that the earth's climate(s) will do "natural variations" with or without human hysteria about CO2.
"the heat's all in the deep oceans" is included in an overestimate of climate sensitivity, as you need more energy for a certain temperature increase if the amount of water in play is much larger than previously assumed.
E-Man,
Imagine I had a model of the stock market and that model predicted a year over year increase of 5% in the total value of shares.
And, most importantly, my model based its prediction of 5% share price growth on a stock market sensitivity to growth in GDP of precisely 1:1. That is to say a doubling of GDP should be reflected in a doubling in the price of shares.
Now, propose that the observed value of the stock market increase by 1% in Yr1, .5% in Y2 and so on while the GDP marched ahead at 5% a year.
Does it make sense to find assorted unforeseen events - rise in interest rates, decline in value of dollar, bankruptcy of two or three companies - and say that "the model was right except for x, y and z" or would it make more sense to say. "Growth in GDP is only weakly correlated with the total value of shares and, therefore, the MODEL IS WRONG."
I realize this is going to be hard for people whose faith has been placed in the CO2=Heat theory - but the observational record suggests that the MODEL IS WRONG.
Not evil, not stupid, just wrong. And if your model is wrong you build another model which includes the stuff you missed on your first try.
"Why has the reduction in surface energy input not produced a cooling trend?"
Hasn't it?
Cycle 24 started to decline at the beginning of 2002.
http://www.swpc.noaa.gov/SolarCycle/sunspot.gif
We're currently near the peak of cycle 25. Although it's a weak one, it ain't zero.
Enso then was pretty much neutral, as it is now.
http://www1.ncdc.noaa.gov/pub/data/cmb/teleconnections/eln-5-pg.gif
Surface Temperature?
http://www.woodfortrees.org/plot/hadcrut4gl/from:2002/plot/hadcrut4gl/from:2002/trend
Entropic Man
The albedo of soot is cooling the atmosphere? Really?
If I told you I had one (and I can't promise you a cheap one!), would you want to judge it based on future results, and not claims about the past? I hope so.
Assuming it did as badly as current models, I hope you wouldn't accord it undue respect if I merely insisted it was right because it was based on sound physical principles. Arguing and asserting sound physical principles is not sufficient.
Aug 13, 2013 at 2:30 AM | Gixxerboy
Heh.
Underestimation?
The models contain no 'internal natural climate variability'. If you look at the IPCC report multimodel ensemble, this is evident:
http://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-10-5.html
The base period is 1980-2000. There is smoothing, but not enough to eliminate decadal/multidecadal scale variability in individual models. The models show variability at the decadal scale, but their average is a flat line.
If you average the ECG traces of 20 living, healthy people, with a heart rate of ~72, after synchronizing their beats, you'll get an average trace that still shows beats. If you average 20 different ventricular fibrillations, you'll get a flat line. Or close to one.
There is a very simple and effective comeback to the people who say that there is no apparent warming now only because all the heat is going into the deep oceans and will be released at some future date.
'What, like it did between 1980 and 2000?"
The average bod has no answer to that. It may even make them think.
Heat in the deep oceans is a conjecture being pimped by Hansen. But what has he every been right about?
Shub (3:35 AM) -
Yes, the multi-model mean has very little natural variability, because the models' variations are not synchronized. However, it is the individual model runs' variability which gives rise to the argument of "well, models show long periods of no increase", and presumably if one were to cause the models' natural variation to increase -- particularly at low (decadal) frequencies -- those runs would now contain longer stretches of stagnating temperatures.
It wouldn't necessarily mean that the models are any closer to reflecting reality, however.
As Bernhard Haurwitz said: "The safest long-range forecast is a very long-range forecast to be verified when the present population is gone".
Harold
Picking individual models to claim stretches of stagnant/upward/downward temperatures is meaningless, and a bait-and-switch. If I produce 25 different models and make them wiggle, I'm sure one of them would 'come true'. This is exactly what Ed Hawkins did on his blog. Before a model wiggle can be accepted as being true, it has to be shown to be real.
If one person is an excellent fibber, it is an useful approach to catch hold of ten such people and ask them to describe the same thing. If they tell 10 different stories, you know you can't trust any of them.
Martin A
Can you suggest non-model based methods for predicting 21st century climate trends?
Aug 13, 2013 at 12:04 AM | Entropic Man
...
The hubris to believe that they can model climate with sufficient accuracy beyond a sangoma rolling the bones ! Phwet !
Entropic Man said
"Can you suggest non-model based methods for predicting 21st century climate trends?"
Your real question is more "Can you suggest non policy/UNFCCC based model methods for predicting 21st century climate trends?"
The answer is: to be able to predict future weather/climate one needs scientific understanding. At the moment we are so far from it that it's a puzzle why they have thrown so much money at models that all are wrong?
Aug 12, 2013 at 11:19 PM | Unregistered CommenterStreetcred
A bone rolling African sangoma has better predictive powers than the UKMO!
Is that the reason our dear, glorious DFID are funding talks between Kenyan rain makers and meteorologists?
Another recipient of UK taxpayers’ money was a £25million research study, part of which involved teaming meteorologists with ‘rain-makers’ in western Kenya.
They make their predictions by watching the movement of ants and the measuring the wind using the tops of earthenware bottles.
http://www.dailymail.co.uk/news/article-2245300/Britain-gives-millions-climate-aid-tackle-flatulent-Colombian-cows--plus-31m-Turkish-wind-farms-funding-talks-Kenyan-rain-makers.html
The models aren't wrong. They are a joke because they are based on poorly understood physical systems. We focus on one output that doesn't change much, temperature. Let's see some more detailed predictions of more complex parameters.
What has happened here is that the liars were too stupid to remember that their CO2 hysteria was testable and that with satellites measuring global temperatures, it was going to be almost impossible to fiddle the figures like they had before with ground stations.
97% of all climate scientists are deluded. The other 3% are mad.
(Or is it the other way round? In reality CO2 climate sensitivity cannot be >~0.1K, easy to prove from irreversible thermodynamics.)
@ Entropic Man - Aug 13, 2013 at 12:04 AM
"Can you suggest non-model based methods for predicting 21st century climate trends?"
The pound in your pocket.
Heads it'll get warmer, tails it'll get cooler.
And you'll still have your pound to spend after the exercise.
I have used electronic circuitry simulation extensively, sometimes adjusting models to calibrate them as best I could with real measurements, for realistic output or to find the cause of a problem. In the same way climate scientists added CO2 feedback mechanisms to their models, perhaps other unrealistic modules as well. Almost invariably I found that other mechanisms were responsible for my model's deviations from reality. That helped me to realise that each model simulates only an idealised small fraction of reality. The only thing that counts is verification with measurements, like Feynman (and many others) said.
Another aspect is the almost religious belief in computer technologies, seen in many people without engineering or physical sciences background. I worked for a time in digital TV standardisation, large meetings with hundreds of engineers, scientists, broadcasters, government officials and regulators. I was always amazed to see how people without any physical science or engineering background were mesmerized by computers and software technology, accepting them as solution to all problems. But they are only tools, causing specific problems of their own kind. This attitude in government officials, strengthened by the love they feel for their smartphones and tablets, makes them accept computer output too easily for the truth.
There is nothing wrong in using models to check your understanding of reality. But as soon as you find deviations from reality, it's back to the drawing board to develop the algorithm. Declaring any model as "sacred" and "settled" is a very stupid thing to do.
Entropic Man said
"Can you suggest non-model based methods for predicting 21st century climate trends?"
Yes no problem at. all it's called common sense:-
1 Temperatures will flat line for the next fifty years.
2 There will be some slight warming for the next fifty years.
3 There will be some slight cooling for the next fifty years.
4 Temperatures will fluctuate up and down.
Easy really,but of course you could go to Climate Audit where a much better guy than myself has used a formula from the 1930's, byCallender, which out performs the models.
One has a suspicion that if the currency if the models hadn't been debased by junkett climate scientists then maybe they may have performed better?
Well we apparently need 'real' climate scientists to finally admit what has been blindingly obvious to skeptics before science progresses. Still the mistake of using frequentist stats on models and still the canard about warming being masked by cooling but little by little maybe they'll get there. Then they need to conclude the other obvious fact; that every climate impacts paper based on these inadequate models is useless.
I wonder how Tamino will take this? It's right up his alley...
Falsfied? Or the temperature data that they have been trained on has a bias. GIGO.
I think the warming at the end of last century was mainly driven by natural factors such as solar activity, perhaps in ways that we do not yet understand. The climate scientists hitched their GHG theory on to the temperature rise. Unfortunately, through a combination of opportunism, ignorance and naivety, they excluded all other possible climate drivers from their thinking, their rhetoric and their models.
Now they are in denial, stubbornly clinging to the models, hoping that warming will resume soon to rescue their credibility. However, with the Arctic having a record breaking cold summer, we may be entering a cooling phase. A few years of cooling will destroy their credibility totally
Then the questions, accusations and shifting of the blame will start. It is sad, really. If the team hadn't driven through their policy of scientific consensus and suppressed the critics, proper science may have prevailed, thus avoiding this mess and the squandering of billions of taxpayers' hard earned cash.
I feel a bit bad about my initial comment focusing only on the one point with which I disagree with von Storch, given that I strongly agree with the rest of his points and am very impressed that he has been willing to make these bold statements. Of course von Storch has often been willing to call out problems with climate science, which makes him somewhat unique within the "mainstream", and he deserves full credit for that.
Entropic man
"They would also have missed the reduction in insolation due to the weakness of solar cycle 25."
What ? you mean they can't predict the climate ?
You are a clown
"Both of these changes would counteract the forecast warming trend, as observed. The problem would not be with the models, but with changes in the behaviour of the system which were not predictable at the time of the forecast."
I would have thought they were trying to model the entire system, not just some responses to a range of inputs.
If not the "models" have no predictive worth (as we've seen).
Schrodingers Cat
Taking that quote of yours in conjunction with Albert Stienstra's posting above points us in the direction of hubris.Between them 21st century scientists and politicians cannot accept the concept of "ways that we do not yet understand". We are pouring millions of taxpayers' pounds/euros/dollars into understanding climate. How can it be that in this technologically advanced age and with all that money there is something we do not yet understand?!
Since he first walked on earth man has tried (and failed) to understand the weather; he has tried (and failed) to forecast the weather more than a couple of days ahead and over the centuries he has learnt to live with it.
Climate is chaotic and virtually by definition trying to predict and understand chaos makes herding cats child's play by comparison.
When are we finally going to understand that?
EM - ah, so it's solar cycle 25 wot dunnit. Fine, it goes against the grain but let's grit our teeth and include the strength and duration of future solar cycles in our models. What's does "the science" tell us how we should model solar cycles 27, 28, 29 ....? How will this affect the deep ocean hidden heat etc.
In case anyone missed it.
For me, the most likely reason for the flat lining of temperature is that NASA GISS and the Hadley Centre no longer find it as easy to fiddle the figures in the numerous creative ways they did when ground stations were the main source of data and they weren't under such intense scrutiny. Didn't Hadley refuse to release data because it was the property of foreign governments at one point ?
The most significant feature of the warmest ever year, 1998 wasn't that it was an El Niño year, it was the year of the Kyoto Climate Conference.
How anyone can believe a single digit these guys produce after Climategate is beyond me.
My expectation is that if the UAH satellite temperature extraction hadn't been under the control of skeptics then it would have been adjusted to suit the models the same way as ocean heat, radiosondes and Hadley/Giss. RSS also seemed more than willing to adjust their satellite data to suit Santers nonsensical, error-strewn paper. More alarm means more funding; the only positive feedback in this entire farrago. That modelers can almost do what they like with the data that is supposed to validate their models is just too tempting for them. Without UAH as a limiter, it would be a free for all.
Smiffy
Quite so, and if it wasn't that Hansen had got his grubby fingers on the figures the warmest year ever would still be 1934.All neatly laid out on Steven Goddard's blog linked to from the 'NASA re-writes the past' thread further up the page.
But then there's no reasoning with some people.