This doesn't work either. Fitting the periodic functions makes little difference to the slope. If you fit GISStemp alone with a line you get about 0.06 C/decade, as is well-known.

The fallacy in the argument is that no-one thinks that a linear slope since 1880, or earlier, represents the manmade contribution. The right curve would be something like the green GHG forcing curve in this plot. If you fit that instead of the line, but with the periodic function, you will still get a slope of over .1 C/decade since 1950.

Your "natural variation" could not have contributed .1C/decade since around 1950. That is almost exactly a complete period (63 years). It's back to where it started and contributes nothing at all.

What's your point exactly Nick? The GHG forcing increased in slope after about 1960, but it's more or less linear (just different rates) before and after that. That just means that GHG's could not have caused the 1911-1941 warming, which occurred at an identical rate to the 1978-2008 warming. It doesn't change the nature of the record itself, which clearly has natural periodicities.

I think it would be reasonable to argue that a greater portion of the late twentieth century warming was due to AGW. Let's say for instance we attribute natural variability to the Sun (21 years is approximately two solar cycles, and of course is about the magnetic polarity cycle length). Well, in a number of articles by Scafetta and West, they have virtually always found that about half of the overall twentieth century warming is solar (sometimes more, like 65% in Scafetta's most recent paper) but only about a quarter of the second half of the warming. if this cycle method is revealing a part of the natural climate cycles, then all that would be necessary for the residual to be linear is a solar increase on top of the cycles in the first warming, and just the cycle in the second one. This is even compatible with the IPCC's statements!

Andrew, My point is that if you're going to try to separate the series into an AGW part and a "natural" part. you need to justify why each fitting function relates to the role you assign. For AGW, the straight line is just not right; the forcings are much studied, and better curves could be chosen. For the "natural", it would be best to have a mechanism. Your 22-year solar cycle idea has well known issues (observed TSI just doesn't change that much, but here, more seriously it turns out to be only a minor component, and there's nothing in the solar record to show a strong 63-year cycle.
I'm not a Scafetta fan, but their conclusions about solar etc are very data dependent. For example, from their 2007 paper:

Climate is relatively insensitive to solar changes if a temperature reconstruction showing little preindustrial variability is adopted. In this scenario most of the global warming since 1900 has to be interpreted as anthropogenically induced. On the other hand, if a secular temperature showing large preindustrial variability is adopted, such as MOBERG05, the climate is found to be very sensitive to solar changes and a significant fraction of the global warming that occurred during last century should be solar induced.

"observed TSI just doesn't change that much, but here, more seriously it turns out to be only a minor component, and there's nothing in the solar record to show a strong 63-year cycle."

I didn't say TSI, and I'll admit I don't know that the natural component is solar, but it's a suggestion. The 63 year cycle is probably something distinct.

"I'm not a Scafetta fan, but their conclusions about solar etc are very data dependent."

Really, I don't care what your "issues" are with their methodology (I expect you are more a "Gavin fan")

"Climate is relatively insensitive to solar changes if a temperature reconstruction showing little preindustrial variability is adopted."

And what is this reconstruction? Why it's:

"Mann, M. E., and P. D. Jones (2003), Global surface temperatures over the
past two millennia, Geophys. Res. Lett., 30(15), 1820, doi:10.1029/
2003GL017814."

Wow! So your argument is that the results are uncertain because Mann might be right? :lol:

"if you're going to try to separate the series into an AGW part and a "natural" part. you need to justify why each fitting function relates to the role you assign. For AGW, the straight line is just not right; the forcings are much studied, and better curves could be chosen."

That is not my interpretation of what is being done.

"So your argument is that the results are uncertain because Mann might be right?"
No. Scafetta is saying that on some data, you get nothing, on others, some solar component. And the nothing data includes not just Mann, but many others.

A master of disingenuous-ness, I see. Scafetta did no analysis of those other recons. and what can be said of many of them? Why, that Mann or Mann coauthors are authors, and that they use many of the same proxies!

http://www.climateaudit.org/pdf/others/07142006...

Figures 5.2 through 5.8

NOTE THAT A RESPONSE WHICH DOES NOT COMMENT ON THESE FIGURES IN PARTICULAR WILL BE IGNORED.

I am only reporting what the fits show, and they are fairly high amplitude
and coherent cycles, and as such should be explained by any self-consistent
climate theory. If not there is something important missing.

I don't know why, but the integer multiples of the periods suggests
harmonics. Could the 63 year cycle be driven by the higher frequency 11 and
22 year solar cycles in a harmonic fashion? Just a speculation.

Integer multiples of frequencies might suggest harmonics.

Mostly people associate the 63 year period with ocean effects.

Exactly, a resonant frequency seen in AMO and PDO driven by absorbed energy
from a shorter Hale cycle.

No, that's the point. It's the lower frequency (longer cycle) that generates the harmonics (frequency multiples).

Its just a mystery ;-)

"This doesn't work either." If you were referring to the OHC post I thought we had established, as you said:

"Your analysis shows the same result for a parabolic increase plus noise."

That is, while the OHC appears to be "increasing faster" it is indistinguishable from a upturn in a natural cycle.

"Fitting the periodic functions makes little difference to the slope. If you fit GISStemp alone with a line you get about 0.06 C/decade, as is well-known."

It makes a great difference to the slope since 1975, the period which is claimed to have the greatest AGW effect, because that is the up-part of the cycle.

"The fallacy in the argument is that no-one thinks that a linear slope since 1880, or earlier, represents the manmade contribution. The right curve would be something like the green GHG forcing curve in this plot. If you fit that instead of the line, but with the periodic function, you will still get a slope of over .1 C/decade since 1950."

If you want to use this forcing instead of the linear, you seem to be saying there is still a considerable reduction in trend over the period by including periodic terms. Though I have not checked it, I am prepared to believe you.

"Your "natural variation" could not have contributed .1C/decade since around 1950. That is almost exactly a complete period (63 years). It's back to where it started and contributes nothing at all."

Yes, but if you look at the graph, the period from 1950 to 1975 is flat, then increases at a great rate from 1975 on due to the addition of 60yr periodic and linear increase. You must also add in the 20 year cycle. The highest trend contribution is from 1975 to 2000 at about 0.1C/decade.

Nick, your [well IPCC's] green slope doesn't make sense; AR4 allocates 2/3's of GHG forcing to CO2; CO2 has had a fairly even increase over the 20thC; the trend should be linear; the green slope has periodicity written all over it; the issue is, is that periodicity neutral over completed cycles; this is a preposterous [sic] assumption because it confirms Miskolczi:-); the evidence that natural cycles are not even is fairly substantial;

http://www.sciencemag.org/cgi/content/abstract/...

The problem for AGW is as David says the natural cycle was in upturn since ~ 1950; if your AGW is 0.1C PD than that means the natural upturn part of the cycle is, what, about 0.03C PD; if that is the case than since 1998, when temps have down-turned, the 0.1C PD of AGW trend has been more than neutralised by the natural down-turn; so the natural down-turn is > the up-turn?

Coho, you should check that CO2 claim. The CO2 rise accelerated during the century.
There's no indication of periodicity in the green forcing curve; it follows the CO2 burning history, and there's certainly nothing periodic about that - it just goes up and up.
Again, the 63-year cycle was not "in upturn" since 1950. That's the thing about cycles - over a full period, they may go up and down, but come back to where they were.

Your argument had more credibility when you were pointing to forcing curves...

"the 63-year cycle was not "in upturn" since 1950. That's the thing about cycles - over a full period, they may go up and down, but come back to where they were."

The superimposition of the various cycles is "in upturn" since 1975-ish. That's half a cycle. Why you insist on going back to 1950 I don't know.

Why you insist on going back to 1950 I don't know.
Because that's what Coho said (and David in the OP), to which I'm responding.

Nick, A theme of this blog is to assume that GHGs are not the dominant effect on climate; and then to try to quantify the separate effects, which might include GHGs if the calculations turn out that way.

I find it a bit shaky to assume that GHG is dominant when it is really an explanation for a model residual or error. You're approaching the subject with a preconception. As yet, I'm not aware of a demonstration that has been sought for years, that GHGs have the capacity and the measured ability to give the measured effect on climate, however robust that might also be.

A climate experiment with no models is still lacking.

Who goes to the Melbourne Cup to look at horses or calculate statistical odds? Not many. Most go because the models look attractive. The IPCC models are not so attractive and they are looking worse as they get older. Soon we'll be looking at their teeth so estimate when they should go to the knackery.

When I was in mineral exploration, we had a letter from Uri Geller offering to use our bizjet to fly over Australia at 40,000 ft, with a notebook where he would list places of high mineral value. His problem was also that he started with a preconception then tried to make money justifying it. We did not give him any.

Here is another of your preconceptions - "The fallacy in the argument is that no-one thinks that a linear slope since 1880....". Have you tested this assertion and found it supportable? How the heck do you pretend to know what I think?

Please try following the scientific method a bit more logically.

Sherro, my point here is that while you can indeed fit a line and a sinusoid to this data (and it's more meaningful when you do have at least a couple of periods), if David wants to claim that that says something about natural vs anthropogenic, those connections have to be made. The 63 year "natural" period, is highly speculative, but there is no reason at all to believe that a linear increase since 1880 or before is a representation of AGW (do you really think that?).

"The 63 year "natural" period, is highly speculative"

It is the only theory so far which explains the 1911-1941 warming occurring at the same rate as the 1978-2008 warming.

Nick, If you can dissect a compound curve into a series of less complicated ones, using Occam's Razor type reasoning (which I do not rely upon, but it's short hand for what I mean), then you have some material to which you can start to ascribe causes. Sure, there is a probability that your dissection means nothing in physical terms, there is a possibility that it's a mathematical artefact.

However, when a number of different measurement techniques start to show repeated similarity to one or more of the sub-curves, you can start to form a hypothesis, then proceed through the scientific method with that hypothesis. An example might involve the orbital variation in Earth-Sun distance over time. A reasonable start would assume that a greater separation produces less irradiance at the Earth. So you quantify it and subtract it from the original compound curve. Then you proceed with successive stripping steps until you get a good fit, being careful to ensure the carrying of error terms as you go.

Where we differ is in the assumption that the GHG effect is a given and can be confidently stripped out first. The difficulty is that its effect is a guesstimate, it is not robust.

You answered your own thoughts by writing "...if David wants to claim that that says something about natural vs anthropogenic, those connections have to be made." The answer is, "If Nick wants to claim.....etc". The jury is still out on GHG, Nick.

BTW, what I think about the shape of a curve back to 1880 is immaterial. A method to determine the parameters of such a curve is material. Science is not about what you think, it's about what you can demonstrate to a tough audience.

This also applies to GISSTEMP, of which you write "If you fit GISStemp alone with a line you get about 0.06 C/decade, as is well-known. " It is well-known because of publicity, not because it is right. One can drive a truck through the errors in GISSTEMP, as has been done several times.

Nick, you've verballed me; you raised the date of 1950 first in your initial comment; your CO2 graphs are also misleading in the usual way; fig's 5 & 6 here explain why;

http://brneurosci.org/co2.html

although to be honest I like Roy Spencer's version;

http://www.drroyspencer.com/2009/01/50-years-of...