Sublime Support and Ridiculous Rebuttal of Akasofu’s Natural Variation Paper

Posted: October 21, 2013 by Rog Tallbloke in Analysis, climate, Dataset, Forecasting, Natural Variation, Ocean dynamics, Solar physics, solar system dynamics

Back in May, MDPI’s new Journal, Climate, published a paper by Japanese researcher Syun-Ichi Akasofu entitled ‘On the Present Halting of Global Warming’. The paper proposes the idea that the recovery from the little ice age and the 60 year oscillation evident in the data not reproduced by climate models needs subtracting from the temperature history before the effect(if any) of additional atmospheric co2 can be assessed. This enraged several of the new journal’s editorial panel so much that they resigned. The Chief editor provided this reassurance and reasoning:

What we can disclose about the review process of the Akasofu paper, without violating the confidentiality of the review process, is that the manuscript was reviewed by three specialists affiliated to institutes or universities based in Europe and the USA. The reviewers were not from the same institution as the author and they have not co-authored papers with the author in the last five years…

We hope that this opportunity for debate will be taken up by members of the scientific community, and that Climate can facilitate vibrant discussion around environmental climate topics that can often polarize opinion, but are of vital importance for stimulating cutting edge research.

Quite right too. Proper scientific debate is to be preferred to editors flouncing. This kind of histrionic approach to rebuttal is unscientific, as an eminent PhD pointed out:

If a scientist does not agree with the results of a published study, he or she should submit a paper presenting counterarguments backed by data and physical analysis instead of making silly political ‘statements’ such as resigning. This not how real science works.

Indeed. But now, a ‘Rebuttal’ comment from a group of the usual suspects (including the hilarious Dana Nuccitelli and Scott Mandia) has been published at the journal’s website purporting to demonstrate that:

The claim that a two-century linear temperature increase is a recovery from a recent cool period is not supported by the data. Furthermore, this thermal recovery hypothesis is not connected to any physical phenomenon; rather it is a result of a simplistic and incorrect curve-fitting operation.

Lets take a look.

Here’s the abstract of Akasofu’s paper:

The rise in global average temperature over the last century has halted since roughly the year 2000, despite the fact that the release of CO2 into the atmosphere is still increasing. It is suggested here that this interruption has been caused by the suspension of the near linear (+ 0.5 °C/100 years or 0.05 °C/10 years) temperature increase over the last two centuries, due to recovery from the Little Ice Age, by a superposed multi-decadal oscillation of a 0.2 °C amplitude and a 50~60 year period, which reached its positive peak in about the year 2000—a halting similar to those that occurred around 1880 and 1940. Because both the near linear change and the multi-decadal oscillation are likely to be natural changes (the recovery from the Little Ice Age (LIA) and an oscillation related to the Pacific Decadal Oscillation (PDO), respectively), they must be carefully subtracted from temperature data before estimating the effects of CO2.

First thought is that the ‘linear’ recovery from the LIA isn’t really linear, but in the absence of strong supporting evidence for a suitable climate oscillation, Akasofu chose to keep it simple. But we can help him out with a bit of Talkshop hindcasting and prediction skill here. Back in feb 2010 Tim C and I posted an article on replicating Judith Lean’s TSI reconstruction using the underlying sinusoidal oscillations discovered in the data by Tim’s brilliant analysis software. One of the oscillations was at a period of around 426 years, see figure 1.

Figure 1: The long period oscillation found in the Lean reconstructed TSI envelope is 426 years, and phases well with LIA recovery

Other things worth noting in Figure 1 are that the peaking of this long period oscillation around the time of the US dustbowl in the 1930s, along with the near Gleissberg cycle 80yr period, and the the next upturn in the 426yr period won’t occur until around 2150AD. Note also the 80 and 220yr period were at a minimum around 1900, when solar activity was low. The modern solar maximum peaked around 2003 along with the 220 period near the De Vries cycle length. The near 60 year period was already on it’s way down, along with the PDO. As you can see, the next few decades show a steep fall in solar activity.

Now one of the usual suspects who wrote the rebuttal paper, Rasmus Benestad, is the ‘solar expert’ amongst the group, the others knowing exactly SFA about solar variation. No doubt he would try to rubbish the findings of Figure 1 as “simplistic curve fitting” as well. But since the cycle lengths found near to periods found in the 14C data, it’s not so easy to dismiss. A few years ago, Rasmus teamed up with another usual suspect who knows SFA about solar variation, Gavin Schmidt, to write a paper rebutting a paper by Nicola Scafetta. This paper contained a fatal stats error which entirely invalidated it’s content. I posted about this a while back.

The journal in question refused to publish Nicola Scafetta’s comment on this fatally flawed paper, and Nicola lost interest in fighting that decision, as he had more interesting work to do. So the falsified paper still stands, and although Benestad and Schmidt are fully aware of its invalidity, they won’t withdraw the paper. In fact Benestad still cites it in his more recent work, and it is relied on by the new IPCC AR5 report. That’s something we’ll be having a guest post from Nicola Scafetta about soon.

  1. oldbrew says:

    Here’s one for the usual suspects: the UK CET records (since 1772) indicate we’re currently experiencing the sharpest decline in average temperatures for 150 years. UK graph

    Not related to natural forces? A hard sell for the suspects, and only the die-hards are likely to buy it.

  2. CET, like the Alaska Region seems to act as an early indicator for more general changes.

    Something to do with being downwind from a long ocean track.

  3. J Martin says:

    How many of the periods in Tim’s graph can be tied to natural causes ?

  4. tallbloke says:

    JM: Bear in mind Lean’s reconstruction is from imperfect data. So the cycle lengths Tim’s software found won’t be exact. But there’s one at 57 years, in the zone for the AMO-PDO ocean oscillations. Another near the Gleissberg cycle length – a solar period. Another near the De Vries cycle length of 208 yrs, which shows as a sharp strong peak in the 14C and 10Be data. The near 450 year period may be Lunar related: Jupiter:Saros cycle = 38:25 = 475 Draconic years (25 x 19) = 450.7y (thanks to Oldbrew for that nugget). We’ve found lots of other evidence that Jupiter is involved in Lunar periods, and solar cycle lengths, and inner planet orbital timings, see the ‘Why Phi?’ posts. At 3 x 450 yr there is a magnetic wobble in the NH too. The 112yr period matches up with an alternation between sst and air temp Roger Andrews discovered.

  5. Roger Andrews says:

    Back to the topic of this thread ;-)

    Before I did anything else I read Akasofu’s paper to form my own conclusions as to its merits and demerits without being influenced by what other people were saying. And after reading it my conclusion was

    The paper sucks.

    This is not to say that its conclusions are wrong, because I don’t think they are. But the paper itself is simplistic, superficial, breaks no new ground, offers no new insights and presents no new data. It’s basically just a rehash of previously-published papers on natural climate variability. It also makes a number of unsupported claims, chief among them being that the global temperature increase since 1850 was caused by a natural recovery from the LIA. Yet even though Akasofu’s arguments hang on this claim he can’t identify a physical cause for the recovery. The best he can come up with is “it may be speculated that the LIA and its recovery is perhaps related to changes in solar activity, even though changes in solar output during a sunspot cycle (11 years) are known to be small, at ~ 0.1%”. If that represents the sum total of his knowledge on the subject he never should have written the paper in the first place.

    Having looked at Akasofu I then looked at the question of the Climate Magazine editorial board resignations. As far as I’ve been able to determine there’s been only one, not “several” as reported above. The panel member who departed was Dr Chris Brierley, of University College London, and the reason he gave was: “I do not believe that the paper is of sufficient quality for publication and have decided that I do not want to be associated with a journal with such lapses of judgment.” Well, I don’t think Akasofu’s paper was of sufficient quality for publication either, but how many present-day climate change papers are? And Brierley’s “why I resigned” article (link below), and the fact that it appeared in Skeptical Science, suggests that Akasofu’s conclusions had more to do with Brierley’s departure than the quality of his research.

    And as for Nuticelli et al’s rebuttal, well, Akasofu gave them an easy target, and they shot at it.

  6. tallbloke says:

    OB: Yes. Question is, why would three Metonic cycles produce one 57 year wave? Perhaps its a reinforcment of any or all of the many other periods around the same frequency.

  7. tallbloke says:

    Roger A: I tend to agree that Akasofu didn’t support his LIA recovery hypothesis well. Perhaps he was trying to keep the whole thing as simple and as free from speculative ideas as possible in order not to forestall open debate? Seems like a good ‘starter for 10′ in a new journal to me.

  8. tchannon says:


    CET might not be quite as you expect.

    I have an extensive article on CET in preparation giving a lot of background. Check my blog over the next few days. (unless Rog copies here)

  9. Roger Andrews says:


    Saying that temperatures have risen because of the recovery from the Little Ice Age and leaving it at that is like saying that the economy has improved because of the recovery from the recession. It ranks high on the list of classic “No sh*t?” statements.

    It is, however, surprising that the Akasofu paper got published, not only because of its technical defects but because if all of the straight-line warming since 1850 was in fact caused by a natural recovery from the LIA and all the rest by the PDO then no room is left for any contribution whatever from CO2. So maybe we have found a good new journal to publish in.

    Maybe I should send them the empirical models I put together some years ago, which fit surface temperature observations not just globally (that’s easy) but in five different latitude bands using a) ocean cycle and CO2 forcings and b) ocean cycle forcings and no CO2 forcings at all. Want to see them? :-)

  10. Scute says:

    TB and Roger A

    I think you are right TB- a starter for ten. It’s true his paper doesn’t say much that is new, Roger A, but he doesn’t make the mistakes that the ‘rebuttal’ claim he does. They might as well have said he doesn’t know how to boil an egg for all the relevance it had.

    Akasofu is a prominent skeptic, a space scientist and aurora specialist (some contributors here probably know that but I didn’t) so he must know of the purported mechanisms such as a GCR link and the possible planetary/tidal influences whether he believes they have validity or not. I think it’s reasonable to point to a pattern and step aside to let others fit a mechanism to it. I think that’s what he must be doing. Mention of TSI is uncontroversial whereas GCR etc means he would be nailing his colours to the mast and preempting any debate on the matter. Conversely, he may be an avid GCR fan, well-known for it and feel that he doesn’t need to mention it for the umpteenth time- that would be a classic SkS-style omission of the truth.

    As an aside, surely people have noticed this LIA/PDO pattern before? I think Roger Andrews is right that it’s a rehash. I read about Brierley’s resignation too and was going to mention that it was to do with the paper being a bit lightweight. But I think he was throwing the baby out with the bathwater: it was a conjecture which could also be termed as a simplistic observation bereft of any proof. But that would be uncharitable. In mathematics conjectures which by definition lack proof, are respectfully placed on the back burner for decades until they are proven by others and the original author duly lauded. But yes, there is a ‘no s**t’ element- trouble is Dana et al studiously avoid saying that in any polite form. Maybe that’s why Akasofu feels he has to write a paper, after 64 years in Academia, stating the bleedin’ obvious.

    The ‘rebuttal’ really does suck. It has all the Nuccitelli hallmarks, the ducking and diving, parrying, the feints the straw men….

    I spent the best part of a weekend last year unpicking his SkS article on Alec Rawls’ leak of the AR5 draft. It was so deft in places that even Alec himself couldn’t quite get his own rebuttal article on target. That’s why I did my own rebuttal, posted as a comment and got a thank you from Alec saying in effect that I’d nailed Dana. And here we go again! I’m really quite weary of this man.


  11. Bart says:

    I find it very droll indeed that people who have been deeply engaged, unsuccessfully I might add, in a vast curve fitting exercise (matching climate simulations to observations) would complain about curve fitting. Apparently, these knuckleheads believe that the admissibility of a curve fit depends on how complex it is, not on how well it matches reality.

  12. RoHa says:

    Minor point, but Akasofu is a US citizen and has been living in the US since 1964. Surely you should say “American researcher Syun-Ichi Akasofu”.

  13. RoHa says:

    No, I’m wrong. He was living in the US for several years before 1964.

  14. Brian H says:

    The LIA recovery argument has this going for it: the post-WWII CO2 emission surges could not have driven it.

    Remember the basic Null: unspecified natural variation encompasses the observed data.

  15. oldbrew says:

    To have a ‘recovery from the little ice age’ you first have to have a little ice age. What excuses do the usual suspects have for that?

  16. tallbloke says:

    Tim: Please publish it here.

    Roger A: Akasofu includes a plot from the Medieval Warm period to Now. His implication is that what goes down comes back up. i.e. climate is wobbling either side of a mean at various timescales. Everything else in the universe does, so why not planetary surface T too? We’ll do a new post for your plots, email me.

    Scute: Nuccitelli is just so ridiculous and so blatantly doesn’t respond to fairly put points which blow his facile reasoning out of the water that I just leave him to hang himself with his own rope. It’s good that someone with your thoroughness and patience is prepared to put the time into dealing with him. I have my own research agenda and won’t have it derailed having my time wasted rebutting nonsense.

    Brian H: Elegantly put. :)

  17. oldbrew says:

    How do the usual suspects account for the LIA in the first place? If they admit it existed that is.

  18. tchannon says:

    Okay Rog, no rush, I’m carrying on bashing on coding, spitting out integers now read from images.

  19. Roger Andrews says:

    Oldbrew asks: “How do the usual suspects account for the LIA in the first place? If they admit it existed that is.”

    It doesn’t exist:

  20. Geoff Cruickshank says:

    Probably worth duplicating Dana’s quadratic v. linear graph and updating it from time to time. It might look pretty stupid in a few years time. Wonder how the residuals of a cubic fit work out.