Monday
Aug052013
by 
Bishop Hill
Bishop Hill Akasofu's model
Aug 5, 2013 Climate: Models In recent months there has been a lot of attention devoted to the failure of the GCMs to predict the pause/standstill/hiatus in temperature rise seen for the last decade or more. Mike Kelly points me to this recent paper in the open access journal Climate. It's by the prominent sceptic Syun-Ichi Akasofu, whose naive model of global temperature change gets attention every few years.
The model essentially superimposes a multidecadal oscillation on an upward linear trend representing the recovery from the little ice age, and it gets attention because it performs so well. In the latest paper, the model is compared to observations for the period to 2012.
;)
Good eh?
Reader Comments (81)
So that would leave the effects of CO2 at.... well nothing significant really.
Rather than naive as a description of the model I think simple would be better. Naive suggests not understanding the science of CO2 radiative forcing.
These simple models are very interesting, not least because they perform so well compared to those promoted and quoted by the IPCC. They are so poor that it amounts to an abandonment of responsibility to base anything on them, and I suspect the Climate Change Act comes into this category.
I suspect using the climate history for a given climate region, plus possibly a little stretching of ranges to be on the safer side, remains the simplest and most sensible input for planners looking a few decades ahead (which surely includes the maximum sensible planning horizons for all sorts of things).
It should be noted that not only are IPCC-blessed models poor, their 'past performance is not a reliable guide to the future'. In other words, hunting around for models that did relatively well over some period does not help when it comes to some future period. For example, Riefen and Toumi (2009)[GEOPHYSICAL RESEARCH LETTERS, VOL. 36, L13704, doi:10.1029/2009GL038082].
Other simple models are also available. Scafetta's simple model seems to be doing better than those GCMs found so convenient by the IPCC. As is the extremely simple persistence model of Armstrong.
Knappenberger and Michaels have summed up key policy implications in a paper presented last month:
It would be interesting to hear from those working on the orthodox IPCC climate models what they think Akasofu has missed or got wrong. Richard Betts, Tamsin?
This is probably heresy even here but I'm beginning to think any small increase in CO2 (as a percentage of the atmosphere) is having little or no effect.
One line of defence, lapogus, is that these simple models just look at global mean temperature - the very thing that was promoted, headlined, made so much of by those wishing to alarm us out of our minds over CO2. Now that that constructed temperature is failing to follow the script, then of course it has been all hands on deck to find substitutes. The trouble is the models were set up to illustrate how suppressing heat loss in a GCM causes rising global mean temperatures, and that may well be their forte.
Disappearing snow didn't work so well, nor did the hurricane thing, and sea level scares have suffered from pesky observational credibility issues as well. Ice remains problematic, and even would-be high priests of the apocalypse such as Schmidt have admitted problems with the methane line. General bad weather seems to be the fallback now, while the witches cackling around their bubbling cauldrons of toads legs etc are still trying to extract something about heat disappearing into the deep sea using some explanation other than magic.
Swiss Bob -
Note with me, I take the view that once atmospheric CO2 is above 250ppmv it has a negligible and immeasurable effect on global temperatures. The only thing I am not so sure about is if that means I am a dragon slayer.
John - Indeed. Hence they are now even trying to scare us with brain eating bugs.
In complex problems in solid state physics (such as magnetism and superconductivity), there is an honourable tradition of progress. The Curie-Weiss law and the gap equation were originally derived from fitting the empirical data as a function of temperature and then rationalising the mathematical form of the fitting function, followed by a long process of trying to understand the various parameters. It took 100 years to get a reasonably accurate first-principles calculation of the Curie temperature and this involved quantum physics that was not available when the original theory was derived. A similar but not identical process was involved in superconductivity on a slightly compressed time scale.
The climate system is very complex, and the general circulation models are the equivalent of calculating the Curie temperature or the superconducting gap. It will be very hard to get right, and possibly take a long time. What is very certain now is that the divergence of the climate models from the raw data since 2000 points to failures or incompleteness in the models. As time progresses the models will be modified and improved, and in doing so, one hopes they will converge on the past data and make less uncertain projections forward at least for a decade or two, but probably never more than that.
What is a real shame is that so many of the mainstream climate models dismiss the work of Akasofu as ‘mere curve fitting’ and of no relevance. In fact his forward predictions, (and his is a model with the temperature turnover from 2000, being a repeat of that from 1880 and 1940), have been borne out by the real-world data. He is onto something, and what that is needs explanation.
If there had been more people in mainline climate science trying to understand the magnitude and the period of the oscillations in temperature, and the century averaged temperature rise, the modellers might have been quicker to realise that Akasofu was on to something.
The justification in condensed matter physics was that the fitted curves were adequate for engineers of magnetic and superconducting systems to get on with the design of real applications. At the moment, because Akasofu is following the data, his is a more valuable approach for engineering and policy making – not the global circulation models that everyone agrees have got something wrong.
Akasofu’s understanding of the parameters of his temperature fit is incomplete, and some day his model will go comprehensively wrong, just as his model does not do well before 1840. But that day has not come yet. Until then the sensible betting should be on a temperature statis lasting until about 2030 as the basis of reaction in terms of energy policy. Mitigation is oversold and adaptation undersold.
What would be interesting is to recall that the extra CO2 in the atmosphere only increases the extra absorption of radiation in the CO2 absorption bands by a logarithmic amount. If one were to assume that the CO2 was rising exponentially, this would lead to a linear temperature rise, as is seen by Akasofu. One can get a self-consistent set of parameters over a narrow time interval of 120 years but they would still leave unanswered what do these parameters actually mean in terms of the physics of the system, but it might allow those primarily concerned with CO2 to feel vindicated. Furthermore a continued exponential growth of CO2 over this century would keep Akasofu's linear term on track for another overall 0.7C rise by 2013, net the oscillations which are attributed to oceanic cycles!
Thanks MJK, that's a good analogy.
Of course there are many cases of 'mere curve fitting' in the history of science. Kepler's laws would be another good example. Again it took nearly 100 years for the theory to follow.
If you apply purely random noise (white noise) to a highly resonant system, the response of the system looks like a sinusoid whose amplitude and phase vary slowly in a random fashion. A short section of the response is hard to distinguish from a genuine sinusoid.
It has been suggested recently on BH (including a comment by commenter Bart) that atmospheric CO2 is controlled by a tight (ie high loop gain) feedback system. If you apply negative feedback with a high loop gain around a system described by a 1st order differential equation, the resulting feedback system will not have a resonance - its response to a transient disturbance will be a simple decaying exponential.
But as soon as the dynamics of the system you are applying feedback around have dynamics other than a 1st order differential equation - 2nd order or higher - there is the possibility of oscillatory behaviour as soon as you apply negative feedback, even if the system without feedback is entirely non oscillatory.
Could the oscillations seen in the illustration be the response of the Earth's CO2-regulating system to purely random disturbances? [Probably not - the timescales don't match]
Just a speculation.
Here is a perfectly good and successful ( so far) 30 year cooling forecast forecast based on empirical patteren recognition.- see latest post.
http://climatesense-norpag.blogspot.com/
Here are the main conclusions.
"To summarise- Using the 60 and 100 year quasi repetitive patterns in conjunction with the solar data leads straightforwardly to the following reasonable predictions for Global SSTs
1 Continued modest cooling until a more significant temperature drop at about 2016-17
2 Possible unusual cold snap 2021-22
3 Built in cooling trend until at least 2024
4 Temperature Hadsst3 moving average anomaly 2035 - 0.15
5Temperature Hadsst3 moving average anomaly 2100 - 0.5
6 General Conclusion - by 2100 all the 20th century temperature rise will have been reversed,
7 By 2650 earth could possibly be back to the depths of the little ice age.
8 The effect of increasing CO2 emissions will be minor but beneficial - they may slightly ameliorate the forecast cooling and more CO2 would help maintain crop yields .
9 Warning !!
The Solar Cycles 2,3,4 correlation with cycles 21,22,23 would suggest that a Dalton minimum could be imminent. The Livingston and Penn Solar data indicate that a faster drop to the Maunder Minimum Little Ice Age temperatures might even be on the horizon.If either of these actually occur there would be a much more rapid and economically disruptive cooling than that forecast above which may turn out to be a best case scenario"
His increasingly ridiculous curve fitting exercises have been sufficiently technically debunked elsewhere. As Von Neumann pointed out long ago, of course they will fit over 'in data' range. But are devoid of causal explanation, so lacking in any useful longer term predictive power. How very different from Coe's fascinating and simple three box model.
What does "naive model" mean?
The main criticism Dana has levelled at the model is, it's not based on physics. Is that what "naive model" means?
My naive point of view is, if one model performs well and others don't, find out why.
MJK
As Paul Matthews says, good analogy.
BUT the entire global warming corpus has been built on the assumption that anthropogenic CO2 will continue to cause increased temperatures unless and until mankind takes action to reduce his reliance on fossil fuels (since that is by far the largest cause of increased CO2 since the industrial revolution).
Which is why many of us have always argued that the moving force was and is political.
Akasofu appears to be extrapolating a global temperature rise between now and 2100 of around 0.4C which is hardly going to scare anybody. Add to that the IPCC's own admission that it is impossible to make any meaningful projection of future climate and they can "improve" their GCMs to their heart's content; if they exclude the known oscillations and variations in the strength of solar cycles and they allow their philosophical attachment to a trace gas to colour their calculations their models will still throw up the wrong answer.
Akasofu's curve fitting may very well go astray sooner or later but since there is precious little we can do to influence the climate the question is an academic one at this stage.
It may be that some time in the distant future we will know enough about the many variables that affect the earth's climate (we may even know what all of them are; I'm pretty sure we don't just now) to make some reasonable forecasts over, perhaps, a 50-year period.
And then a fractional change in the earth's orbit throws all the calculations out and we have to start again!
Good eh? No. Why should the recovery from the LIA be linear and continue indefinitely? And if the linear trend is in doubt, then the estimated amplitude of the oscillation associate with the PDO could easily be wrong. At least the IPCC's models are driven by a phenomena (rising GHGs) with an physical mechanism capable of influencing climate. IPCC may over-estimate climate sensitivity and under-estimate natural variability, but at least they are fitting something that makes sense physically.
Will Nic Lewis want to keep posting here alongside this stuff?
"We are not descended from fearful men. Not from men who feared to write, to speak, to associate, and to defend causes that were for the moment, unpopular."
Edward R Murrow.
His fight was with Macarthyism but equally apllies to the current visitation of imagined demons.
Apart from selecting a short time period during which atmospheric CO2 concentration correlated with global temperature in order to get maximum alarmism, what else have climate scientists got right in recent years?
I feel that computed trend lines on temperature graphs can confuse the issue, as they depend on the shape of the graph over the period concerned. If you plot hadcrut4 from 1880 to 1979, when 30 years of global cooling ended, the actual temperature increase between these end points is less than 0.2ºC. We then get into the warming of the 80s/90s (if hadcrut4 is to be trusted, UHI effects and all) which was a period of exceptional solar activity and rising PDO/AMO. The IPCC bases its hypothesis of global warming due to man-made CO2 on just the 63 years “since the mid 20th century” (AR4 summary), yet only 20 years of this period have exhibited steadily increasing global temperatures, even less based on UAH satellite data rather than hadcrut4. Their hypothesis is looking increasingly unbelievable.
Frank -
If you are sure it is not just a recovery from the LIA and all AGW then please be so kind as to point out the CO2 signal in any of these long-term temperature datasets. (and note that these are raw data, unadjusted for UHI).I think the key point here is that Akasofu's model accords well with observations and the IPCC models are by and large bollocks.
Frank The recovery from the LIA is likely over - see my post above and forecast at
http://climatesense-norpag.blogspot.com
Martin A @ Aug 5, 2013 at 4:45 PM
"Could the oscillations seen in the illustration be the response of the Earth's CO2-regulating system to purely random disturbances?"
I doubt it. But, of the Earth's temperature regulating system? Yes, I believe it easily could.
Frank.
The IPCC is trying (unsuccessfully) to prove a point. Akasofu has a point. And for both it is only a point as CAGW theory quietly left the room when someone noticed the missing heat.
Bart 6:43. Thanks.
If we are at the end of post-LIA warming having regained the 1 degree and just completed the multidecadal peak around 2000, we may be in line for some cooling. That seems to be in agreement with a number of other indicators.
I wonder how a spell of cooling can be incorporated within the absconded heat theory.
Just as I ask with the GCMs, where is the validation? Sauce for the goose, and all that. It doesn't matter whether it fits observations for a limited period. As for cycles, we can all see cycles. It's a survival trait for us to see things which aren't there. In this case, there really are cycles. Tons of them. Ocean, Solar, astronomic, unknown unknown, imaginary. And then there are forcings. Some cyclical, some not. I wonder whether anything actually repeats in terms of what the weather or climate actually do, what with all those cycles interacting at different frequencies. Resulting in something that looks like a random walk over larger timescales but with long-term trends which may themselves be cycles. It's complicated, and Willy Ockham may not be the best authority this time.
Good post Martin A. Nice point Bart. Put this with the Nicola Scafetta (2012) Paper.
Frank,
"At least the IPCC's models are driven by a phenomena (rising GHGs) with an physical mechanism capable of influencing climate. IPCC may over-estimate climate sensitivity and under-estimate natural variability, but at least they are fitting something that makes sense physically."
But their models are crap. What's better, a model that has some sort of physical theory behind it that fails every test, or a model that simply says "what happened yesterday will happen tomorrow" and that works?
"Will Nic Lewis want to keep posting here alongside this stuff?"
Who?
There have been quite a few semiempirical approaches to parsing past climate. Most involve cycles in some way. Some get a remnant small contribution that is ascribed to greenhouse gases. Some ascribe cycles or long term secular changes to astronomical sources. They all conclude that CO2 is a small if not totally negligible contributor. I think that this work by Ziskin and Shaviv http://www.phys.huji.ac.il/~shaviv/articles/20thCentury.pdf is the most sophisticated because it contains enough physics (vs. simply curve fitting) and because of its use of a genetic algortihm for attributing the climatic effects of the various forcings.
I referred to Akasofu's theory recently here. I do not see that it is "just curve-fitting" (although you can see I don't even buy the "temperature record", so I have a different view than anyone else about it, so far as I know) -- I suspect the "expert" critics here are merely consensus greenhouse-theory-bound, and don't know what real physics is (the greenhouse effect, of increasing global mean temperature with increasing atmospheric CO2, is not real physics, as my Venus/Earth temperatures comparison clearly demonstrated, nearly 3 years ago now). I see the "expert" debate is still completely vain, as they refuse to admit they need to learn anything from the failure of climate science.
BTW, I published the output of my own model a while ago:
http://www.flickr.com/photos/99888237@N06/
The red line is the global average temps, from the MO. The blue line is a 60 year forecast appended to a 42 year hindcast.
So when will the model peak and why?
There are two reasons to use a model to describe a natural mechanism:
1. When you can't do proper experiments on the system, you can try and mimic the observed data and thereby understand the role of all of the putative inputs. This is GCM approach - they want to "explain" climate by building a model that encompasses everything that could be involved.
2. You want to be able to make predictions about what will happen in the future. In this case, you don't care about the mechanism, you only want your model to be "useful" in as much as it is accurate in its predictions.
This model is an example of type 2. Akasofu is not trying to say what is causing temperature changes, he is just describing them with a model that may be useful in making predictions and until such time as it goes wrong, it IS useful. Whether you want to make massive investments in global development policy based on it is another matter - but I wouldn't be doing that based on any of the GCM's either.....
[For those of you with a philosophical bent, consider type 1 Popperian science and type 2 Kuhnian - but only in a very general sense. Let's not get hung up on classification.]
Dana writes at SkS
It is superficially a good objection. What causes the observation of the rise? If you don't know you can't say it will continue.
But the only reason it is expected to stop is that the direction of the climate's temperature trend has changed in the past; warming to MWP, cooling to LIA and probably others as well.
What caused that?
If the basis for doubt of any change in the direction of the climate's temperature trend is unknown then the same case applies.
Assume it will continue or vary randomly / noisily until some evidence comes in to the contrary.
And no, a theory about the effect of CO2 is not evidence. Observations that match a theory about the effect of CO2 is evidence.
But the observations don't fit the CO2 theory as well as they fit the Akasofu model. Akasofu's Carry On and Don't Panic Model.
There needs to be a better reason than "I don't understand it" to reject a model. There need to be physical observations.
M. Courtney writes:
'Dana writes at SkS
"In order for this to be a physically sound argument, Akasofu must explain the physical mechanism behind the "LIA recovery", and why this 0.5°C global warming trend continues to persist. What is the underlying cause? Surely a geophysicist will examine this question." '
Dana has just committed the fallacy of "assuming that all causes are known." Why would anyone do that? Does anyone actually believe that all drivers of climate are known? I know that Alarmist modelers make that assertion but their only evidence is an ever growing list of models that show no skill at forecasting.
Someday the physical mechanism behind the LIA recovery will be known but the same is true for everything in our climate data. Someday the cause of the present pause, plateau, or retreat will be known but it is not known now.
Dana is simply assuming that the only thing in our data that is not explained by a known cause is the rise in temperature from 1979 to 1995 and of course he presents rising CO2 as the only possible explanation for that rise. That argument is a tight circle. Most important, the assumption that all drivers of climate are known implies that there is no natural variability in climate that is not understood. How anti-empirical can a thinker be?
If the predictions of this simple model are correct then, by referring to graph shown here, the "standstill" in global temperature change will soon turn into a marked fall in temperature which will extend until about 2030, which is 16-17 years from now. I suspect that, long before 2030 arrives, politicians around the world who are presently convinced by "warmist" will, seeing the persistent fall in temperatures, and in response to budget and other pressures, recognize that they have been sold a pup. It should be the end of the global warming industry.
There is no reason for Akasofu to think that the warming trend from the LIA will continue linearly..There are good empirical data in the temperature record and in the solar "activity " proxy data as seen in the neutron count to make the reasonably conservative prediction that the recent temperature peak was a peak in both the PDO 60 year and in a solar millennial cycle and that therefore the next 20 years are very likely to see cooling and that the next 1000 years could possibly repeat the temperature trends seen from 1000 - 2000.with cooling to a new LIA in about 600 years.The solar data suggests that a Dalton or even a Maunder minimum might be possible In which case the cooling would be more rapid and more severe with important problems for food production.
Check
http://climatesense-norpag.blogspot.com
for details.
If the establishment scientists would stand back a bit and look at the big picture and use their commonsense for once ,they might realise that the story is reasonably obvious and they don't have to worry the politicians with their ill-founded doomsday predictions.
This was my graphology version of climate hindcasting and forecasting that Judy was kine enough to host
http://judithcurry.com/2013/05/16/docmartyns-estimate-of-climate-sensitivity-and-forecast-of-future-global-temperatures/
Three caveats, partly based on the temperature data
http://data.giss.nasa.gov/gistemp/graphs_v3/Fig.A2.gif
1) He ignores the long term orbital cooling trend from the paleo data, illustrated in Marcott et al and confirmed by the long-term downward trend of the Medieval Warm Period and the LIA. He gives no evidence that the 60 year cycle predates the instrument record.
2) The peak-to-peak rise from 1880 to 1940 was from 13.8C to 14.1C, a rise of 0.3C. The corresponding rise from 1940 to 2000 was from 14.05C to 14.55C, an increase of 0.5C . The two 60 year cycles would show the same peak-to-peak rise if the long term trend were linear. Instead the later cycle has a larger increase than the earlier one.
3) In both previous cycles the peak was followed by immediate cooling. After 1880 the subsequent minimum was 0.25C cooler. After 1940 the drop was 0.15C. After the 2000 peak there has been no cooling at all. The amount of cooling decreases form cycle to cycle.
Points 2) and 3) both lead one to infer that whatever forcing causes the long term warming is increasing and accelerating the rate of change. This invalidates Akasofu's assumption that the long term trend is linear.
Entropic Man @ Aug 6, 2013 at 12:57 AM
"The two 60 year cycles would show the same peak-to-peak rise if the long term trend were linear. Instead the later cycle has a larger increase than the earlier one."
Peak-to-peak of a 60 year cycle is measured over 30 years. They look exactly the same.
I remember that a few years ago skeptics were amazed that warmists did not seem to understand the difference between correlation and causation, now however we seem to have role reversal.
Akasofu's model seems to fit fit the observed temperature record and so it is deemed good and reliable?
Not only that but the most sensible comment was by a warmist??
Frank does not say it is not linear he just asks how we know it IS linear and why it is supposed to continue. Rhoda continues with further questions. Here we have a model that is based on one unverifiable assumption and a decadal oscillation, Because this fits observed temperature records we have highly qualified people who are falling over themselves to support the model.
When it is the case that so many of us understand that there are so many factors influencing climate and temperature that we really don't even know what half of them are, how can we support a model based on an assumption and a decadal oscillation?
rhoda
For once, I agree with you. Where indeed is the validation? All he's done is fitted something to past data, assuming the wiggles continue into the future. It's the same mistake made by the more extreme predictions of Arctic sea ice loss from a few years ago, which were based on a temporary acceleration in Arctic sea ice decline in 2007 and extrapolated the curve to suggest (obviously incorrectly) that the summer ice would be gone by last year. Both this and the Akasofu "model" are just examples of over-fitting and mistaking noise for signal.
I agree with Frank that there is no physical reasoning behind this, and I too would be interested to see what Nic Lewis thinks. And Steve McIntyre come to that - he's been looking at Guy Callandar's work from the 1930s, which correctly predicted a long-term warming trend over the 20th Century on the basis of physics.
Update: I just saw Dung's comment, and agree with that too - and I never thought I'd find myself saying that..... :-)
Bart
The peaks of a 60 year cycle occur 60 years apart, as do the troughs.
The three peaks in the instrument record occur in 1880, 1940 and 2000. The minima occur in 1910, 1970 and (if the cycle is genune and not illusory) 2030.
They are not the same. Compare the slopes of the peak-to-peak linear regressions for the two cycles.
http://www.woodfortrees.org/plot/hadcrut4gl/from:1880/to:1940/trend/plot/hadcrut4gl/from:1940/to:2000/trend
The 1940-2000 cycle shows a clear increase in warming rate over the 1880-1940 cycle.
"They are not the same."
I gave you the plot. The two cycles are about as close to exactly the same as you can get. Just look at the plot. There is no increase at all to any level of significance.
The cycle isn't precisely 60 years, and your argument is sophistry.
Wow!
If Richard Betts, Rhoda, and Dung can agree, is that called "consensus"??
97% or 98%??
(j/k I think)
Roy Spencer has already clearly shown models over estimate warming
http://www.drroyspencer.com/2013/06/still-epic-fail-73-climate-models-vs-measurements-running-5-year-means/
It is almost certain there will not be a perpetual linear rise in temperatures after all where would the heat energy come from.
As some have pointed out here and others else where there is clearly a cyclic signal of about 60 years in temperature data.
The little ice age and medial warm period were several hundred years ago, this suggests temperatures may be influenced by a combination of cycles, one around 60 years, another several hundred years, plus others at scale of thousands or tens of thousands of years.
Those with at least O level maths will probably know over a sufficiently small interval a sine is a very close approximation to a linear trend.
Does any of this give us the ability to predict the future?
Not really
Newton's law of gravity is believed to be true to a very high level of accuracy across space and time, so this formula can be used to make predictions of positions of stars and planets (though in practice these predictions will be inaccurate). It is the belief in Newton's law that gives people confidence in predictions base on it.
There is no equivalent law or formula for global temperatures. Whilst it may be possible to fit a curve to existing data predictions based on extending a fitted curve into the future a highly likely to be wrong simply because it is almost always possible to find several different formula which provide nearly identical curve over the range for which there is data, but behave dramatically differently outside this range. A more detailed explanation here
http://jeremyshiers.com/blog/prediction-using-fitted-functions-can-be-very-wrong/
Phew haven't even had to mention chaos, UHI. or adjusted data
Richard Betts,
"Both this and the Akasofu 'model' are just examples of over-fitting and mistaking noise for signal."
You state this with authority. Yet the MO's models have been awful at predicting temperatures. Maybe another view on what is noise and what is signal might be a good idea.
Time will tell, just because we don't like it that doesn't mean it's wrong. When it does go wrong there may be a clue to the refined version which will work without hiding heat in the deep ocean and not the atmosphere.
While it forecasts better than expensive government and university research departments then using it will free up an awful lot of money for something more useful.
The universe is full of magical things, patiently waiting for our wits to grow sharper.
"Why should the recovery from the LIA be linear and continue indefinitely?"
I don't believe Akasofu is predicting anything, what he's doing is putting the various oscillations on a linear rise in temperature from the LIA and saying if this continues this is what will happen. No one knows what caused the LIA, or why we came out of it. What Akasofu's simple model tells us is that if the linear trend he's assumed for the LIA changes then the temperatures will continue to vary round that line in accordance with various oscillations doesn't it?. It's a theory, which plausibly explains the recent temperature records. Is it right? Dunno, we'll have to wait and see if some physical explanations can be found for the oscillations of the MWP and LIA. What it has in its favour is that it is right so far, while the climate models remain complete bollocks forecasting linear increases in temperature for an increase in a GHG which is known to have a logarithmic effect.
Richard Betts comments "And Steve McIntyre come to that - he's been looking at Guy Callandar's work from the 1930s, which correctly predicted a long-term warming trend over the 20th Century on the basis of physics."
Richard, if you think that Callandar's work "correctly predicted a long-term warming trend over the 20th Century on the basis of physics" then presumably you agree that the apparent climate sensitivity must be quite low as this prediction uses 1.67?
Actually, Callanders model is little different to Akasufo's model. If CO2 correlates with temperature any "physics" model fitted to CO2 wil match temperature. It is simply a further abstraction than just curve fitting to the temperature. And Callandars model is accelerating rather hard around now and is not going to agree with the current pause in any way.
You could just as easily have built Callandar's model on a trend of universal suffrage in the 20th Century and probably got a similar result.
I don't think Steve showed Callandar's model to suggest it had any particular merit, he showed it to highlight how a simple model based on CO2 was outperforming the might of the GCM's.
Was Richard Betts commenting on the GCM thread?
Akasufo's model is simply empirical and it is therefore easy to poke holes by saying it has no physical significance, but that doesn't mean his trend + periodic behaviour is wrong, just unexplained.
Regarding long term trends that might explain a recovery from the LIA or the MWP, Singer wrote a whole book on evidence for long term trends in "Unstoppable Global Warming...every 1500 years". Plenty of evidence in there for long term periodicity as well as short term periods. The approximately 60 year periods Akasufo proposes are quite well documented eg Russian Northern port sea ice data over the last couple of centuries show evidence for these periods, PDO etc.
Rhoda, Dung, Richard, Frank etal.
Perhaps the point of the model is not to provide a scientific basis for governments to base policy decisions upon.
Consider that a simple unsupported back of an envelope calculation has more predictive capability for the modern era than 20 years of advanced modelling using state of the art computing.
The climate is complex and continually changing so this model will be wrong but that is not the point. The point is that all models are wrong and the capabilities of the GCM's have been seriously overstated in their ability to inform policy makers and the public.
All the criticism being leveled at this model is correct so when are we going to agree that the GCM's have been incorrect, (Richard), and not just sneak a new model into the mix with a vastly different predictive output but retain the same baseless argument that CO2 is the controlling influence?
There have been many recent statements from the MET as well as others that hint at natural variability having a greater roll in the climate than previously thought but I don't recall anything being released yet that reduces the influence of CO2 to something that is not the controlling factor.
Consider that to be the purpose of the model.