This post by David Archibald at WUWT extolls the virtues of “wiggle matching,” then applies it — or so it claims — to match the CET (Central England Temperature) from 1703 through 1737, to that from 1978 through 2012. Archibald then notes that the 1703-1737 period was soon followed by the unusually cold year 1740. His conclusion? That we might be in for a similar year of extreme cold in 2015.
There are lots of problems with his efforts, but the main one is that the wiggles don’t match.
Archibald says:
So what does that period up to 1740 wiggle-match with? It matches with the warmth of the last 30 years:
The graph above shows the Central England Temperature (CET) record from 1703 to 1745 as the blue line. Plotted on it is the CET record from 1978 to 2012. Normally when you align 34 year lengths of temperature records you don’t get any correlation. The correlation on this particular matchup is 0.112. The statisticians amongst us can argue over whether or not anything can be read into that. If something can be read into it, we only have to wait two years to experience the consequences.
First of all, when you align 34 year (or any) lengths of temperature records you will get some correlation. If the records are unrelated then the correlation is likely to be small (how small depends on the noise level and structure and the number of data), but it is unlikely (damn near impossible) to be zero. The question is, is the correlation large enough to be meaningful? If so, what does it mean?
Actually the 0.112 isn’t the correlation, it’s the squared correlation — and I get 0.111 rather than 0.112. The correlation itself is 0.333.
Here’s the data without the artificially curved lines:
Does such a correlation mean anything? Testing it, the p-value is 0.0508, which — just barely — fails statistical significance at the 95% confidence level. Ordinarily that might be interpreted as a sign that Archibald’s wiggle-matching might have some meaning or might not. But in this case, the answer is a definite no.
Why? Because Archibald went looking for correlation — even did “wiggle matching” to impose it — and the best he could come up with failed to achieve 95% confidence. If you’re trying to match up two data sets, and you still can’t get to 95% confidence, then you’ve got nothing.
But wait — there’s more! The correlation which is present (whether meaningful or not) isn’t because the wiggles match at all. It’s because the trends match:
When the trends are similar, as they are here, there’s going to be correlation. If we want to know whether or not the wiggles match, then we need to remove the trends first. That gives us this:
Now the correlation isn’t 0.333 (sqaured correlation 0.111), it’s a pathetic 0.0806 (squared correlation 0.0065). For a data set this size, that’s about as meaningless as it gets; testing for significance, the p-value is 0.645, which is a resounding No Way!
Bottom line: David Archibald tried to predict an upcoming cold year based on “wiggle matching” with past data, but even though he was trying, he couldn’t get the wiggles to match. That didn’t stop him from trying to forecast the consequences of his purported cold year in another post at WUWT.
Of course, this is just another amusing example of silliness masquerading as science on the WUWT blog. David Archibald tried to “wiggle-match” old CET data to modern CET data and utterly failed. But because he didn’t do any analysis of his result, he doesn’t even know that! Because he lacks the basic knowledge of how correlations can arise, and the difference between wiggles and trends, he doesn’t even know that the “wiggle” part of the correlation is vastly smaller than he estimates (and again, meaningless, which again, he remains ignorant of).
But, in my opinion, David Archibald isn’t really interested in understanding science, or what we can expect from future climate. He is, however, very interested in justifying a scenario in which we should be afraid of cold rather than heat, because his bottom-line goal is to sabotage any effort to take global warming seriously. Just my opinion.
Also my opinion: the most interesting thing about this sordid little comic episode is how it shows that the WUWT blog, always eager to declare how incompetent actual working climate scientists are, has such low scientific standards that they will actually publish this drivel. Just my opinion.
Open Mind 
There are so many problems with Archibald’s chart that one could write a book on it, although that would be a colossal waste of time once you have established that there is no significant correlation. I love the chart containing predictions for 2014 and 2015, based on, um, based on here’s what I pulled out of my ass, I guess.
Fact – WUWT did publish it. Fact – It is drivel. This would suggest your assertions are more than just opinion.
cooling any moment now, really just hang on another year and an ice age is coming, proof that AGW is a fraud to control your lives and remove your freedom, yes just wait one more year!!!!!!!!!
and if you’re wondering I am being sarcastic [not ironic- irony is a the notion of surprise], Archibald is beyond parody- I can’t even think of a parody concerning the next ice age- it has been done just go to WTF- oops – WUWT. Thankfully science is the means to avoid the pitfall of being human that seeks patterns and destiny. For that reason I thank those who make a life time of study in the matter- being lazy, British and otherwise occupied I extend my thanks to those who give a damn and especially those who do the maths [math!]
oh look I just pulled a nacho from the packet with an image of jesus on it- proof of divinity- now reassured of my eternal soul and loving god.
The most important thing – when your aim is to mislead and misinform in order to allow catastrophic change to the planet to proceed on it’s Business as Usual course – is to assiduously avoid close scrutiny of the underlying causes of those wiggles.
And cherry-picked to have the great frost and irish famine of 1740-1741 to have the future ‘prediction’. The climatological causes don’t seem to have been researched very well, but a VEI5 eruption occurred in 1739 at Mount Tarumae in Japan.
Probably not very hard to find gradual warming trends punctuated by cool periods due to volcanic eruptions.
The great thing about this ‘prediction’ is that you can keep moving the date forwards, and since the trends match the false correlation will always be there, and then when a climatologically significant volcano ever does go off it will accurately ‘predict’ the climate, even though it had been wrong for years/decades…
Speaking of not likely to get cold anytime soon with all the extra kj mankind’s activities are injecting into the system… well, it’s still early days in the arctic melt season, but this doesn’t look good:
NSIDC Arctic Sea Ice Extent | Daily Image
Mind you, it looks to be a bit of an upturn since just yesterday. What’s a pixel or two amongst friends, eh?
It’s been a slow start to the melting season; heck, extent was almost at baseline for several weeks, I think.
And I was looking yesterday at the DMI reanalysis curve for Arctic temps; it has been well below seasonal norms for several weeks, too. In fact, I was struck enough by this that I went looking through the archives of past years, and really didn’t find anything quite so striking in the last three decades.
But even with all of that, the ice still looks pretty bad. Persistent cyclonic activity helped keep temps down for a couple of those weeks, but at the price of a good mechanical bashing administered to the pack. And although extent, area and volume were all up a bit, ice thickness was not. (Volume was up less than area.)
And now extent has started to drop in earnest, as it does every season. Will it keep on dropping, as it did last year? Or will we get periods of stagnation, as happens not infrequently?
Essentially, there’s no way to know in advance. We may well see a minor SI ‘recovery’ this year–but it’s not inconceivable that a new record could still be set. Just have to wait and see.
2002 was the last year [using NSIDC September mean extent] that was not in the then top five lowest sea-ice extent. My guess is that the trend is sufficiently strong that each year will always be in the top five years in the minimum September sea-ice extent league.
“extent was almost at baseline for several weeks, I think.”
Yes but part of that was due to the fact that the baseline has been revised “down” since it now includes 1981 to 2010. Even at that, we’re already edging toward 2 standard deviations below, just like last year. It’s all coincidence, no doubt. Whatever “recovery” is claimed this year, as it is every year that doesn’t beat the previous, should be put in the context of the new/old baseline. Looking at the images on Cryosphere Today, which give a better idea of the density of the coverage, this year may not be much better than last. A 1 year slowing in the death spiral means recovery only to the deluded types…
“A 1 year slowing in the death spiral means recovery only to the deluded types…”
Yes, hence the scare quotes.
And the DMI 80N reanalysis temperature has now edged nearly up to the seasonal norm. The second half of melt season may not resemble the first.
Extent isn’t the best measure of the _amount_ of ice, exactly, as it assigns all grid areas with 15% or more ice into the same class. When the ice is spread out, this looks like nothing is happening. PIOMAS and sea ice area get at that and when you look at them, the differences early this season were not so great. Extent is dropping rapidly at the moment, however.
The NSDIC statement on this is:
“…This type of circulation pattern is known to slow the summer retreat of ice, not just because it fosters cool conditions, but also because the pattern of cyclonic (counterclockwise) winds tends to spread the ice out…”
Certainly true, jgnfld. (You from ‘the rock?’)
Although I actually think that part of the reason for the extent is ice divergence due to lots and lots of cyclonic action early in the season–the ‘ice drift law(s)’, essentially Coriolis-driven, mean that lows, rather counterintuitively, tend to produce ice divergence. I’ve wondered if part of the sharp decline is related to more convergent patterns as well as melt–the basis for that thought being that extent was dropping faster than area.
But I didn’t take that thought seriously enough to really look at the whole situation carefully, so I mention it now only as a speculation that may perhaps interest some–IOW, ‘conversationally.’
Yes. At Memorial University. I also have a sailboat and closely follow marine matters.
Nice. I’m an Ontario boy in the American south, myself, but have some indirect connections out your way.
I stopped visiting WUWT a year ago. The stupid is so thick other there you can cut it with a knife. Reading that kind of stupid will literally lower your IQ points. The site should come with a medical WARNING, something about losing about IQ points and very moderate consumption.
It’s hard to take something that matches wiggles seriously …
seriously!
So, Archibald’s model predicts that 2015 will be -2.5C colder and will kill off 20% of the Irish? That’s an awful strong claim to hang on a R2 of 0.112. It looks like it was good enough for his skeptical readers pick it up and run with: The June 19, 2013 at 3:43 am commenter wants to bet that Olympian Pranksters will make this prediction come true.
Actually I think he’s using CET data to predict a major volcanic eruption for next year.
Mount Tarumae: VEI-5 [Plinian] eruption in 1739
Or have I missed something else? There’s nothing really close enough time wise from Iceland, so something bigger elsewhere would fit the bill nicely.
Amazing to think that someone would even think that there might be predictive value in this kind of exercise–even if carried out properly. It’s as if the old saw about ‘repeating history’ were taken much too seriously, and much too naively.
Tamino,
Just in case you get criticized, I got p-values of 0.04976 and 0.6383 for the raw and de-trended datasets, respectively (using cor.test in R for the time frames you used in the annual data, and detrend() function for the detrending, which doesn’t look like a great way to do it for the 1703 part of the record, but whatever). Of course, you are right that this is all dumb anyway…
[Response: Curious … I just re-ran it, also using cor.test in R, and got 0.05082 and 0.6452 for raw and de-trended. You’re comparing 1703-1737 to 1978-2012 annual averages?]
[Response 2: Aha! I computed annual average CET by averaging the monthly values. But when I simply take the annual averages reported in the CET data file, I get the same numbers you got for the p-values.]
Yea, with the end years included, the precise number depends on whether you add a year or not…using http://www.metoffice.gov.uk/hadobs/hadcet/cetml1659on.dat
Oh well…
wikipedia: “Numerology is any study of the purported divine, mystical or other special relationship between a number and some coinciding observed (or perceived) events. It has many systems and traditions and beliefs.
….
The term numerologist is also used derogatorily for those perceived to place excess faith in numerical patterns (and draw scientifically unsound inferences from them), even if those people do not practice traditional numerology.”
But how is the average person with an IQ of 100 and little education to tell the difference?
Well-said, Ed. But I don’t think this is too much for the average person you envision, provided–and this is the kicker–that their full attention is engaged. So many distractions…
That’s “Numberwang”
It will be instructive to look at the weather patterns during the next few years. Perhaps Archibald will be proven correct. Now wouldn’t that be interesting?
[Response: It will be instructive for you to watch my butt. Perhaps monkeys will fly out of it.]
No, it will not be interesting, even if it happens.
They’ve been calling for an imminent cold snap now since 2006 in one guise or another, over and over and over again, and wrong all along.
Eventually a significant volcano will go off and they’ll be (temporarily) right in one of these calls.
A busted clock is also perfectly accurate twice a day, but its totally useless.
RichardLH on July 6, 2013 at 1:33 am Tamino: Would you every consider taking any temperature series you wish and passing it to a digital bandpass splitter/cascaded low pass filter circuit and plotting the whole output from each stage (as low and splitter)? This should allow the discovery of where in the available spectrum (record length limited of course) the RMS power is. I would suggest that you started with a cascaded running average filter bank with the starting pole (average span) at say 12 months and use the well known inter-stage multiplier of 1.3371 to cancel any ‘square wave’ digital sampling errors that otherwise occur. I have done it for various sources and would like your observations
[Response: Why would I do that to find out where in the spectrum the signal power is? I can do that with the Fourier transform. No, I’m not interested in your procedure, but I encourage you to.]
Talking of trends causing false apparent correlations, I think (but I’m far from sure) that this blog post falls for that:
http://neven1.typepad.com/blog/2013/06/problematic-predictions.html
I suspect failure to remove the trend is at least part of the reason he gets the higher correlations for the longer datasets.
Coincidently, I tried something similar myself. At least I didn’t fall for that trap:
http://edavies.me.uk/2013/06/seasonal-sea-ice-variation/
Yes, I believe that was discussed in the comments to some degree.
I guess, if he’d lived in the stone age, Archibald would have been very good at spotting predatory animals in the bushes—and of course shouting, “wolf!” When it comes to using his imagination to make climate predictions he is very selective. Is this the ultimate example of cherry picking to see what you want to see?
Well there it is for me, the difference between this site and WUWT; real curiosity and the attempt to get to truth, looking at the data to find the discrepancies.
Thanks Fred
Is Archibald’s prediction of cooling for the whole world, or just for England?
Well, if the exercise made any sense at all, it would make *more* sense for central England. But we all know what happens when you multiply by zero.
This appears to be an exercise in statistical pareidolia (or maybe an attempt at ‘cold’ reading). Why would anyone take this seriously?
Maybe Archibald can get a guest spot on Dr. Oz as a climate medium.
More of Archibald’s contribution to science:
http://n3xus6.blogspot.com.au/2007/02/dd.html
Filter splitting of this kind can be fun, can be interesting. However, you have to consider the fact that the filtering system you use directly affects your results – if you don’t, you may ‘find’ amazing things in your output that are really the product of your filtering.
Tamino describes a good example of this error at https://tamino.wordpress.com/2013/03/09/ice-cover-is-not-stabilizing/ – where a bandpass filter highlights what are really insignificant portions of the signal by suppressing everything else. McLean et al 2009 makes this error and attributes all global warming to ENSO, as discussed at http://www.skepticalscience.com/news.php?n=162&p=2 .
Much better, as Tamino points out, to do a complete frequency analysis with Fourier decomposition, and look at your entire signal.
[ Previous comment was in reply to RichardLH – Sorry about that… ]
Sorry to ask here, but I can’t think of somewhere else with the sort of people who might be able to help. I’m an engineer, electronics, but a practical one (hands on with industrial kit). It’s been three decades since I did anything like Eigen vectors, which was during my degree.
I’m looking into a new pattern of atmospheric circulation since 2007. I’ve got tabulated data from NCEP/NCAR (2d pressure fields of the northern hemisphere). If I use the summer (JJA) average I get the pattern, which I’ve shown by another method is unique for the summer average since 1948.
If I correlate (Excel CORREL) the JJA averages for each year since 2000 with the 2007-2012 average I get what I expect, very low correlations until 2007 when correlations increase substantially. For each month correlated with the 2007 to 2012 average, June 2013 is a very low correlation, again I’d expect this as the pattern hasn’t developed this June. However I don’t trust correlation for this task. I’ve looked into EOFs but these seem to mainly be about extracting patterns not comparing similarity of patterns.
So I want to get a numerical measure of the similarity of one 2d matrix with another.
Am I right to suspect that correlation is the wrong tool for the job?
If so what other techniques are out there?
I’d be grateful for pointers from other commenters here.
A dumbfounding aspect of the DA prediction is the implication that it’s easy to predict climate/long term weather with, ahem, “accuracy.” So, by extension, the true climate modeling and data extrapolations surely must be time wasting obfuscations. Why shoot, if an envelope sketch is good enough, all them thar climate researchers must be pulling our legs. It may be that the Pseudologoi and Apates over at WUWT are stating this sentiment expressly, I wouldn’t know. Regardless, they toss their soft pitches and Tamino knocks them outta the park.
When I read Archibald’s Wiggle Matching post several weeks ago, my reaction was “he cannot be serious”.
Can we get an EEG from Archibald ( and a few other WUWT posters ) to see if their wiggles match up with any known sentient species?
Well I was very skeptical of Archibald’s claim too,but then I noticed that the wiggles in the graph perfectly match that of my front door key,and it too has a large dip at the end,so I owe David an apology,and I’m looking forward to the much cooler temps ahead.
RichardLH | July 7, 2013 at 7:35 am | Reply RichardLH on July 6, 2013 at 1:33 am Tamino: Would you every consider taking any temperature series you wish and passing it to a digital bandpass splitter/cascaded low pass filter circuit and plotting the whole output from each stage (as low and splitter)? This should allow the discovery of where in the available spectrum (record length limited of course) the RMS power is. I would suggest that you started with a cascaded running average filter bank with the starting pole (average span) at say 12 months and use the well known inter-stage multiplier of 1.3371 to cancel any ‘square wave’ digital sampling errors that otherwise occur. I have done it for various sources and would like your observations
[Response: Why would I do that to find out where in the spectrum the signal power is? I can do that with the Fourier transform. No, I’m not interested in your procedure, but I encourage you to.]
Because if you look at what a bandpass splitter circuit does it will provide signal detection in noise that no Fourier transform will do, fast or slow.
If you like a challenge, take a 12Hz, 4hz, and 3hz sine wave and and mix with a lot of noise. Lots. Make a very short tape (<3 secs). Work out how long a recording period is required to get the signals/noise separated with an FT.
This is a very well know and often used audio/rf circuit and when converted to digital as I have described, will spit out those cycles without a blink. Even on a very, very short record.
So, why don't you do what I suggested and try. Plot a frequency/power/time diagram using this or a better circuit. I have done if for UAH for you. http://s1291.photobucket.com/user/RichardLH/story/70051
[Response: It appears that there’s a lot more to Fourier analysis than you’re aware of.
It’s very nice of you to suggest I apply alternatives analyses — but it is not very nice to refuse to take “no” for an answer.]
Oh and make sure you use a mix of 3hz half cycle followed by a 4hz half cycle for a few cycles just to keep things interesting.
juuuust wait a minute
“climate is chaotic therefore trying to predict something is vain, but let’s do that anyway using the most stupid and the less physic based way”, did I get it right ?
You could have stopped there, you know
P.S. The longest period that can be discovered by this circuit appears to be the reciprocal of Nyquist as 1/period = frequency and the period limit is < 1/2 the record.
[Response: It appears that there’s a lot more to Fourier analysis than you’re aware of.
It’s very nice of you to suggest I apply alternatives analyses — but it is not very nice to refuse to take “no” for an answer.
]
I am very sorry. I was trying to point out the limitations of FT when compared to another, more basic algorithm on short recorded length data.
I was not being argumentative, I was trying to obtain for opinion.
Before you dismiss so easily the results I have shown, perhaps you would like to research bandbass splitter circuits in digital and see if you can get a better signal to noise ratio ‘in band’ when compared to an FT?
[Response: Before you assert so positively the superiority of your method, perhaps you should research Fourier analysis more thoroughly.]
Please note that the circuit shown above is the digital implementation of a cascaded low pass/bandpass splitter with a ‘brick wall’ filter characteristic and a complete indifference to noise ‘in band’. It also leaves all the harmonics of the main component (12 months) intact whilst completely removing it.
No phase distortion and precise frequency band assignment.
RMS power from each passband can be calculated.
I don’t think that is bad for a simple circuit.
You are right though, I do not know the comparable signal/noise ratios between this and an FT.
Sorry: more accurate phrasing
…a complete indifference to HF noise ‘in band’. LF noise may still be present. It…
What could have caught this (if Archibald has been competent) would be residuals analysis. This is a classic “spurious correlation” due to both data sets sharing a rising trend. Granger and Newbold wrote about this sort of thing back in 1974, didn’t they?
[Response: Before you assert so positively the superiority of your method, perhaps you should research Fourier analysis more thoroughly.]
Presumption of lack of knowledge of a given area on someone else and its possible limitations is not always a wise step.
I would still like a simple answer, which has a better signal to noise ratio in short length data, the circuit I suggest or an FT?
Which parts of FT do you suggest I study further in order to find a better circuit than the one I describe?
Looks like Archibald is losing the confidence of the faithful.
Here are a couple quotes from the comments to his followup to his wiggly post:
( http://wattsupwiththat.com/2013/07/05/further-to-a-1740-type-event/ )
quote1:
Blue sky says:
July 5, 2013 at 2:17 pm
“Don’t understand the real point of the article. Basically all it says is that cooler years tend to reduce yield (which we already know).”
Agree
I love this website, but the quality of the articles is all over the place. I would be surprised if Skepitical Science would pick out a hot year from the last three or four CENTURIES and write what this author wrote.”
quote 2 ( from lsvaalgard – who I believe is a solar expert )lsvalgaard says:
July 5, 2013 at 2:23 pm
“Blue sky says:
July 5, 2013 at 2:17 pm
I love this website, but the quality of the articles is all over the place.
You can count of Archibald to deliver articles with a consistent [unvarying] quality…”
Ouch, that’s gotta sting.
[Response: LOL]
Can I describe my understanding of FT in data selection terms.
There we have ‘normal distribution’ (or spikier) followed by one or more weighted contributions up and down stream at specific time offsets depending on your choices of parameters.
Tweak as required.
This circuit (I believe) shows the extreme opposite to ‘tuned circuit’ style behaviour. It is completely de-tuned in the passband. Completely for HF anyway.
Look at it this way, if you were to take any temperature data series and remove from it all metadata that says it is temperature or whatever.
Treat it as a short recorded section of unknown signal from an anomymous source.
Decide to look at it as though it were an audio tone with a main frequency component at 12hz with signifcant ‘normal distrubuted’ noise around the main tone.
As it is a very short recording, the well known signal length limitations of FT come into play. As it appears to be a slighly organised mix of 3Hz and 4Hz ‘half cycles’ then simple sine wave detection will not be very usefull so FT is a poor choice for this.
All the sub-harmonics of the 12Hz tone are preserved by the circuit described (such as 1, 2, 3, 4, 6) out to ‘DC’/zero but without any feedforward or feedback contributions and no assumptions about data distribution inside a band.
All this is doing is what most of the instruments that provide the climate data in the first place are doing. In front end demodulator or noise reduction stages usually.
This is applying that same technique to the data at the end as well the start of its journey.
[Response: Let’s agree to disagree, shall we?]
This is also logical. Pure tones are mostly found in rigid or organised structures. Fluid or gas structures in nature are much more chaotically dominated than pure tone.
Easy to get set of nice tones out a taut string from branch to ground in a breeze. Less easy to see if the lower end is cut.
Digital frequency filtering is just that – a method of amplifying or reducing particular frequencies.
The problem with this, or any other filtering technique, is a tendency to get excited over the results without considering the methodology – and not realizing that the processing threw out the baby with the bathwater, removing the significant signal and leaving behind only what are at most minor components. If you do not consider your process as well as the results, you will be examining only a subset of the data, and will generally be in error.
A full frequency analysis, such as a Fourier approach, allows looking at the full set of data prior to any filtering – and identifying from that where the significant energies are in that signal. You can guide any later filtering based upon the data, a much better technique overall.
Your analysis, I have to say, appears to suffer from processing artifacts, an examination of filtered data without considering the influence of that filtering – and as such, isn’t terribly useful.
Please see the comments about the limitations of using FT for quasi periodic and short record data.
Please note also, that this is exactly what most of the instruments used in collecting ‘climate temperature data’ do They often use the above circuit, sometimes in analogue, as part of their input or demodulation stages.
This is just using the same metodology on the output data as that used to record the inputs.
I cannot see how that is wrong.
Specific filtering like this, binning, can be very effective – if you know ahead of time what you are looking for, and can filter to isolate that particular signal from noise. For example with compressed sensing techniques, FM radio, etc – you have a concept of the significant data, and wish to isolate it.
It is completely inappropriate when investigating where an unknown signal might be, as you have prior to analysis already discarded information.
[Response: Let’s agree to disagree, shall we?]
Certainly.
I will leave it, for now, with a final thought.
If you were to take a ideal model of a constant input power source and apply it to mixed reflector/absorbtion surface of a rotating sphere.
To measure the temperature time response (from the ‘daily’ input rotation/periodic function) in such a model it is easy to split in into time bands/frequencies in the natural response.
From Daily, through Yearly to the 1461 true ‘Solar Year’.
If you wish, use FT to spit out any bands that contain energy (or the circuit above).
Describe the both the input and bandpass outputs. Note that an ‘end around’ sum can be performed to provide validity and any remaining RMS that is yet left to be assigned.
Archibald tried a similar thing with another (consistently) execrable post at WUWT on ‘falling’ sea levels:
http://www.webcitation.org/6HxfVKdz9
The man is a laughing stock, except amongst clowns.
A ‘sad clown?’
KR | July 8, 2013 at 2:51 pm | Reply
“Digital frequency filtering is just that – a method of amplifying or reducing particular frequencies.”
Actually this is digital bandpass filtering. All of the energy will be correctly assigned into a ‘bin’ which is a frequency/period band.
Think of it like histograms for energy in time.
Your very binning (12, 16, 21, 28 and 37 months span – why those???) is a choice, a choice that throws out data and distinctions – it is a filtering. Perhaps (as a Gedanken experiment)) 39 months has a great deal of energy, but gets lost in midst of the 37 month bin? What then?
Fourier filtering certainly does have its limitations – edge wrap effects, and the inability to separate frequencies above the Nyquist limit or with a wavelength longer than perhaps 1/2 the sample length. However, those are not huge (or, I would argue, significant) issues with climate data, as we have >100 years of data, and we have monthly sampling data with trends only detectable at 20-25 years length.
Histogram binning represents a direct data loss. A full Fourier analysis does not. And looking at your analysis – https://tamino.wordpress.com/2013/07/06/when-wiggles-collide-or-not/#comment-82364 – I would suspect that your very choice of periods is to some extent driven by the Mark I eyeball, and hence more an artifact of perceptual shortcutting (it looks like XXX, therefore it must be XXX confirmation bias) than of any statistical analysis.
I am going to have to disagree with you. You have reduced data too soon, filtering before analysis.
Well that all be well IF I chose the series. I chose nothing. The maths cooses for me. The series used is the ideal one for this job. It has a 1.3371 multiplier between stages. Kust the one required for running averages. This is a well known low pass filter – unzipped into a fuil bandpass splitter.
No cycle junkie me!
Using FT is like trying to work out the frequencies in a set of ropes hanging from the branches of a tree and finding that you can only easily ‘see’ the ones that have their tips caught on the ground.
Gives nice sine waves then. Otherwise, just too much organised chaos.
The film is short though, could really do with a longer recording to work it all out properly.
[Response: Digital signal processing is a poor substitute for time series analysis. Alas, too often those who practice the former insist on telling those of us who practice the latter how much better their methods are.]
For God’s sake, first of all the CET is not even England between 1702-1720 where it is based on Utrecht measurements. That is just a start on the problems of the pre 1750 measurements.
“Have another carrot”.
Well the data is good enough to retrun a tiny +0.4c to -0.3c daily offset from the Yearly Normal as a consistent beat pattern. That’s pretty accurate!
—
The range of offsets to the ‘normal’ annual pattern at a daily resolution has values of +0.4C to -0.3C. The Annual pattern is from +16C to +3C around which this 4 year pattern then distributes and Weather then distributes around that.
http://s1291.photobucket.com/user/RichardLH/story/73127
Response: Digital signal processing is a poor substitute for time series analysis. Alas, too often those who practice the former insist on telling those of us who practice the latter how much better their methods are.]
Ah well. Let’s call it a draw for now then.
[Response: You can call it what you like. I’ll call it “You’re in way over your head but you don’t know it.”]
I have a very simple view looked at from the vantage point of a space station which sits somwhere on a line between Earth and Sun. It always cast a shadow in the center of the Earth as I see it.
I have no other view, so I get what information I CAN see from there.
The Earth only repeats EXACTLY every 4 years. So that is my starting point.
[Response: What nonsense. The tropical year isn’t EXACTLY 365.25 days, it’s 365.2422 days. Leap years don’t occur EXACTLY every 4 years, so the average calendar year isn’t 365.25 days either, it’s 365.2425 days. And the sidereal year isn’t EXACTLY 365.25 days either, it’s 365.2564 days. All those periods will affect earth’s temperature and out estimates of it. But the tropical year dominates, and that is well approximated by the calendar year. As for the configuration repeating EXACTLY, that simply isn’t going to happen.
It’s quite difficult to demonstrate an interaction of the diurnal and annual cycles in available temperature data. More to the point, you haven’t even come close to showing it, your nonsense about 1471 days is just because you begin by saying “that is my starting point” — it’s an assumption, not a result. You think you’ve seen a pattern but you haven’t done the statistics needed to demonstrate that it’s anything other than randomness. Honestly, I doubt that you know how.]
Then I talk to my collegue who has a base on the Moon, his vantage point only see the Earth illuminated in the EXACTLY same way with the same ground below every 37 Human Months.
We conclude that these frequencies MIGHT be important to understand what we both are seeing. Temperature as well as everything else.
YMMV.
[Response: Your naivete is stunning. I might be tempted to disavow you of some of your misconceptions if I thought you were open to the idea that no, your cascading filter method isn’t so much better that it’s indispensable.
Playing around with your tool is fine — enjoy yourself, you might even discover something useful. But you haven’t yet. Insisting how weak other methods are (methods you seem not to understand very well) makes you look like a fool.]
Or is that ‘Calling Earth?… Come in?”
And whilst you are at it. Take one of your nice climate data temperature records and strip the metadata off it that says ‘Climate’ to get impartiallity.
[Response: What the hell is this about? Are you accusing me — or anyone else — of partiality? You are out of line.]
Walk it around telling evrybody that you think it has a 12Hz fundamental and you need help understanding the sub-sonics. Tell that it is digital but quite a short tape.
Se ehat answers you get.
Leaving off the fact, as Tamino points out, that the correlation is hardly the correct measure here, though used by WTFIUWT, There are NINETEEN spans with greater correlations than the 1703-1737 span:
Start End Correlation
1708 1742 0.516
1806 1840 0.487
1691 1725 0.471
1922 1956 0.460
1690 1724 0.454
1687 1721 0.452
1847 1881 0.427
1909 1943 0.422
1807 1841 0.421
1694 1728 0.418
1878 1912 0.397
1689 1723 0.394
1887 1921 0.390
1916 1950 0.370
1924 1958 0.369
1837 1871 0.359
1710 1744 0.357
1872 1906 0.356
1688 1722 0.335
1703 1737 0.334
Cherrypicking? Nah. Actually, if I were Archibald, I would have cherrypicked 1708 to 1742. It catches the late drop in the “wiggles” quite nicely.
I doubt it is worth taking the time to look at the distribution of detrended data.
So a viewpoint from the Sun (as energy input), and the Earth in Temperature response and local feedbacks (for energy output and losses). Thus it must be the Moon which is the regulator, directly or indirectly.
Now THAT is going to get me in trouble everywhere. Just following the logic. Honest!
[Response: You’re entitled to your opinion. You are not entitled to hijack my blog for your nonsense. We’ve put up with enough of it already.]
Richard, your first sentence isn’t. It conveys zero semantic content, which makes your second sentence a non-sequitur as written.
So if you’re “in trouble,” it’s with critics of effective communication. Nobody else cares, because nobody knows what you are trying to say.
My appologies. I was pointing that there is a 4 year component in the Sun’s input power to Earth surface power tarnsfer function.
And the there are the observed local temeprature responses here on Earth.
The main modulator of local conditions, outdise of thos etwo, is the Moon. Is has observed effects on evrything.
Those points of view need to be considered also.
That is all.
A story of three points of view.
I am a scientist employed on a space station that always sits on a line beetwen the Earth and Sun. Its shadow always falls at the centre of the Earth as I see it.
I have a collegue who works on a station on the Moon.
There are also a load of people on Earth recording what the temperature is and where and when and building many brilliant theories.
Let us try and bring together those THREE points of view.
I see repeating patterns in the data I collect from my instruments with a 1461 Earth day time pattern as one of the frequencies.
My collegue sees repeating time patterns in the data he is collecting which corresponds to 37 Human Months.
The people on Earth keep looking at the data they collect in 365 (and squash every 4th to fit – mostly) with fancy FTs and stats, but for why?
Calling Earth: Anyone there?
Then we got the Luedecke fallacy, for the damn measurements that were indoors or out in the sun.
Eli Rabett | July 9, 2013 at 5:19 pm | Reply
So there is just pure Numeroiogy in the sequence, 1,2.3.4,5,7,9,12,… then?
No bearing on Climate at all?
[Response: What the hell is this about? Are you accusing me — or anyone else — of partiality? You are out of line.]
I do hope NOT.
This was no attempt to call any names. That would be very wrong and improper. I do not believe I have ever done so.
I was just attempting to make this a ‘double blind’ test.
That sort of impartaility, only.
To RichardLH: you have had plenty of opportunity to express your opinion. It’s time for you to find another outlet.
To others: please do not argue with him, it’s pointless.
Ok. I did not come here for arguement. I came for discussion. I leave, for now anyway, with this thought.
It looks like the integer digital series based on the mutiplier 1.3371… when used on digitally sampled data (such as those in climate temperature research) contains as part of that series all the major frequencies and its sub-harmonics of the masses and rotations of the three major bodies involved.
Then this apparently means that Climate = round(1.3371…) with the 48/49 month mismatch to the integer main cycle being the inevitable output of natures usual trick of mixing 1/2 cycles (both plus and minus) in groupings that reflect longer, non interger cycles to disperse power out to ‘zero’/’ground’.
That half cycle mix (rather than pure single cycle tones) makes examining the data we have very difficult.
Just my positon at present.
I think you for allowing to think, and thus present it correctly.
I know that Tamino asked us not to engage, but …
A few years ago at Tim Lambert’s Deltoid blog there was one persistent commenter who had his own “theory” about analysing temperature data. He kept repeating himself over and over and eventually in exasperation I suggested that if he really had something, he should write a paper and get it published. I was half-joking, of course, so I suggested that WUWT would be his best chance of bringing his revolutionary “theory” to an eager public.
Well, a few weeks later, there it was in all its glory.
Can I suggest that RichardLH tries the same?
[Response: Ah yes … Girma.]
So you’re one to blame for that piece of, er, work?
Please quit while you’re ahead; there’s more than enough anti-climactic diatripe out there.
TrueSceptic.
I think you should go and receive the acclaim that is likely awaiting you.
With you describing this Girma as being entirely closed to the advice of all others, I thought to dig a little deeper. His Wattsupia piece shows he is educated, or at least qualified. (Although I always become cautious when somebody presents themselves with a full list of academic qualifications. In the same way as you would be rightly worried by somebody insisting that they are compus mentis and have the paperwork to prove it.),
So I searched him out. Sad really. I doubt he has much use of his FEA skills down at the fish dealers.
But the search surprisingly also yielded this (second comment on blog).:-
“My favorite thing Watts ever did was to post a piece by one Girma Orssengo…”
I think this trumps the “Brilliant” response comment by Jim at RealClimate which has previously been mentioned on this site.
Perhaps then you should go and explain your contribution to getting Girma’s great theory ‘published’ (abet on another planet). His number one fan surely deserves to know..
MM,
I can hardly claim any credit but seriously, are you suggesting that no one has a right to air their ideas, no matter how nonsensical? Don’t you think it’s useful to see how truly unsceptical WUWT and its denizens are?
There’s enough claptrap taken as gospel. Girma et al is perfectly entitled to whatever beliefs he chooses to hold, however wrongheaded they may be.
Encouraging him and others to spread them only emboldens the semi-informed and makes an already uphill battle more difficult.
Re. “I always become cautious when somebody presents themselves with a full list of academic qualifications.”
I routinely use my “Dr.” title when making restaurant reservations and when talking to any receptionist in the health care system. I find it quite valuable in both cases!
But I use it hardly anywhere else.
I don’t know how to tell you guys this, but the “pause” has proven Girma to be correct. Just ask him.
Sooner or later a big volcano will erupt or some other reason will cause a short term big drop in temperature. When that happens I’m sure there will be someone with a wiggle matching exercise who has ‘correctly’ predicted this by sheer luck and will be hailed as a genius while all the other less lucky wiggle matching efforts will be forgotten.
I found a 98% correlation between the temperature anomaly and the number of taxi drivers in Nigera between 1973 and 1987.
Next year will be an ice age.
In 1987, all the Nigerian taxi drivers emigrated to Ireland. It’s just a hunch, mind you, but I reckon that will stave off the ice age. Maybe we should publish our results on WUWT as a counter to Archibald.
Would that be ‘plagiarism’ according to accepted WUWT norms?
;-)
Given the sheaf of 19 significant sets of correlated “wiggles” (actually trends) to chose from, that gives a lot of room to cherrypick from, certainly.
Of course the fact that there are 3 times as many significant positive correlations (19) as negative (6) will be ignored even though that is at least mildly suggestive of something in these data.
[edit]
[Response: Enough already.]
[edit]
[Response: I’ve been more than patient with you. I even gave you a “parting shot” already, a final comment after I said “It’s time for you to find another outlet.” Three times I’ve told you we’re not interested in hearing more from you. This is the fourth. Goodbye.]
> quote 2 ( from lsvaalgard – who I believe is a solar expert )
He is; Leif used to come by here to comment. He’s, um, engagingly quirky, for values of quirky typical of, oh, Unix sysops for example. I recall he and Tamino rubbed each other the wrong way* and I’ve hoped for a reconciliation and further conversation.
————————————
* Once when Billy Sunday preached a hard message on sin somebody said “Billy, you gotta quit preaching that way. You’re rubbing the fur on the cat the wrong way.” Billy Sunday said “The old cat’s headed toward hell. If she’ll turn around, I’ll rub her the right way.”
More wiggles as EGU continues to suicide
Eli, not sure EGU is to blame! It is the publisher (Copernicus) who decided to push this new journal. Check the Editorial Board and it certainly rings some bells. So who is to blame? The managing director of Copernicus perhaps? He/they might not have realized that they actually got hijacked in some way …
“Check the Editorial Board”
Ah, the editor in chief is from the algerian petroleum institute and the co-editor in chief is mörner. Oh my.
… and guess who made it also past “peer” review. Albert Parker! Rings another bell? Well, Tamino certainly remembers: [tamino.wordpress.com]
Proudly presented by his “peers” Morner, Quadfeul (Editor in chief & said petroleum institute fellow) plus two apparently clueless anonymous reviewers. Pathetic!
The morner things change, the morner they stay the same.
Heat records aren’t the only ones that are broken
Sorry to repeat (to repeat..to repeat .)
But some gifts are worth giving more than once
Mornering has broken
Like the first Morner-ing.
Willie Soon has spoken, like his first sneers.
Praise for horse-diving, praise for the Morner-ing,
Praise for them springing high from the piers.
Sweet the Times’ new fall, Soon-light from heaven.
Like the first manure-fall, on the first grass.
Praise for the sweetness of the wet garden,
Sprung in completeness where horse hooves pass.
Mine is the Soon-light, mine is the Morner-ing.
Bjorn of the one light Heartland saw play.
Praise mathturbation, praise every Morner-ing;
The Lord’s* recreation of the new day.
Morner-ing has broken, like the first Morner-ing.
Willie-Soon has spoken, like his first sneers.
Praise for horse-diving, praise for the Morner-ing,
Praise for them springing fresh from the piers.
*Lord Mocktman
I just read a few lines.
My eyes. They burn.
From the man who has made more of these errors than just about anyone who is actively publishing. You would think that always being wrong might slow down your ability to publish, but I guess not.
Co-Editor in Chief: Nils-Axel Mörner :(
Ah, the Journal of Mathturbation Research. The colour scheme is the least of its problems, and perhaps sickly appropriate
The irony–and rich it is–lies in the final paragraph of this paper:
“Highlighting these kinds of problems is important in science.
In fact, while errors in scientific research are sometimes
possible and unavoidable, what most harms the scientific
progress is the persistence and propagation of the errors.
This happens when other scientists uncritically cite and use
the flawed results to interpret alternative data, which yields
further misinterpretations. This evidently delays scientific
progress and may damage society as well.”
Check the Acknowledgments. It appears one “Roger Tattersal” provided “encouragement”. That’ll be Tallbloke…
And we know at least one of the reviewers: Nils-Axel Mörner himself.
Any chance that you could update the wiggles on your own post on humidity from 2011? It has been a horrifically humid summer here in the NE USA — way, way beyond anything I’ve seen in 40 years. I’d be very curious to see where the wiggles in my own region have wandered to in the past two years.
https://tamino.wordpress.com/2011/05/17/hot-and-wet/
Very curious, because the number of insects is way down. Hardly get any on the windshield when driving.
Horatio, you outdo yourself …
tergiversation
Shuffle off to buffalo!
sigh, wrong place, but this makes an opportunity to mention to Eli that insects have been way down for several years. They came back a bit this year in moist New Jersey (loving the butterflies, but it’s a rare place).
Among other things, way too many chemicals being used by towns to protect from vector-born disease and big ag to make farming more like a factory.
Actually, insects (and weeds/foliage, including poison ivy) are also way, way, way up in Massachusetts. I can’t believe the number and variety of insects, especially earwigs, mosquitoes, butterflies and dragonflies, but also everything else, from ants to spiders to stuff I’ve never seen before. It may have a lot to do with how moist and hot its been for a few years in row… too early to tell if it’s actually climate change, but it’s definitely unusual.
op. cit., The Rise of the Dedicated Natural Science Think Tank (PDF)
” … science progresses when everyone can buy the type of science they like, dispensing with whatever the academic disciplines say …”
This looks intresting on accelerated sea level rise http://onlinelibrary.wiley.com/doi/10.1002/grl.50731/abstract
“Popular Trends”
— by Horatio Algeranon
While some remove the trend, like Carter,
Before they correlate
Others think it’s simply smarter
To leave the trend in, mate.
Hmm… looks like the wiggler in chief (Scafetta) seems to have a new paper in E&E.
Click to access 1307.3706v1.pdf
Amazingly, he seems to have a section relating medieval astrology to climate. (random conjunctions=plague!!) It does raise the interesting point though – wiggle matching could easily be described as a modern version of astrology.
Given the subject (greenhouse gases) and the amount of indigestion, farts and crap it produces, “gastrology” may be more apt
And footnotes to CO2science, Soon and Baliunas 2003, multiple references to McKitrick papers, comparisons of projected ensemble _means_ to recent observations with ENSO variations, Hoyt and Schatten 1997 (outdated TSI reconstruction), on and on and on.
In other words, climastrology based on outdated and withdrawn papers, denialist blogs, and blatant astrology – “Indeed, in 1345 AD a Jupiter-Saturn conjunction occurred in the zodiac sign of Aquarius and was
linked to the outbreak of the Black Death epidemic.”
And in his conclusions, “It may be surprising to many to learn that planetary oscillations probably exert a significant control on the Earth’s climate system, as presented in this paper. However, this is the way climate change has been interpreted and predicted for millennia by ancient civilizations that built sophisticated astronomic observatories to this purpose. […[ Today, many dismiss this ancient science as astrology….” Apparently we’ve learned nothing interesting or useful over the centuries…
Appalling. Not surprising from Scafetta, but still appalling.
‘climastrology based on outdated and withdrawn papers, denialist blogs, and blatant astrology’
Don’t forget the most important form of (jack)asstrology: “horse-diving”
Could Scafetta be Sokoling the denialists?
A Sokol maneuver? If he had published a singleton piece with outrageous claims, that might be the case, he might be making fun of the outrageous climate denial claims.
However, as per his publication history ( http://people.duke.edu/~ns2002/ ) Scafetta has been pushing this nonsense for at least 10 years, with considerable effort going into web publications for SPPI (denialist lobbying group), contributing to the NIPCC (not the IPCC) report for the Heartland Institute, etc. I believe I see a pattern…
If this were a joke, it’s one that lost its humor a long time ago.
The Age that will Bury Us
— Horatio Algeranon’s rendition of The Age of Aquarius (5th Dimension)
When Will Soon is in the Random House
And stupid is as stupid does
Then rays will guide the climate
And Lords will steer the stars
This is the dawning of the Age
That will Bury Us, the Age that Will Bury Us
Will Bury Us, Will Bury Us
Clime and sea misunderstanding
Ignorance is just astounding
Tons more falsehoods and derisions
Tony having dreams and visions
Sea-ice crystal recreation
And the mind’s tergiversation
Will Bury Us, Will Bury Us
When Will Soon is in the Random House
And stupid is as stupid does
Then rays will guide the climate
And Lords will steer the stars
This is the dawning of the Age
That will Bury Us, the Age that will Bury Us
Will Bury Us, Will Bury Us
Let Will soon shine, let Will Soon shine in
Will Soon shine in, na na na na na….
KR, yeah, I know. He even duped the Physics Today editor. I just don’t see how any sane person could think that this was real science, though. It is just patently absurd.
Geez, THANKS Horatio! Both I and my wife have the 5th Dimension running laps through our brains now. Just do me a favor: don’t pick on “It’s a small world.” I went to Disneyland once upon a time, and that song was stuck in my head for three weeks. I had to have repeated sessions Led Zeppelin and Zappa therapy before erasure occurred (it was replaced, curiously, by “Stick it out.”).
Good one, though.
Sorry….
It’s a Watts World
— Horatio’s parody of “It’s a Small World”
It’s a world that’s dafter, a world of smears
It’s a world of dopes, it’s a world of jeers
So much that ain’t there
That it’s time we’re aware
It’s a Watts World after all
CHORUS:
It’s a Watts World after all
It’s a watts World after all
It’s a Watts World after all
It’s a Watts, Watts World
There is just one Soon and one Golden Sun
And HADCRU means trendshift to everyone.
Though the warming’s increased
And the ice has deceased
It’s a Watts Watts World
http://people.duke.edu/~ns2002/
Page down to the bottom for the climate theory.
Focus on just the last six years, as he does, and you will see he can prove
well
whatever he likes, I guess.
hm, he thanks Prof. Arthur Rörsch
see also http://jules-klimaat.blogspot.com/2009/04/rorschs-bizarre-ideas-on-cherry-picking.html and http://jules-klimaat.blogspot.com/2009/04/rorschs-bizarre-way-of-conducting.html
Horatio–I love it!
Hi Tamino- What did I do wrong?
[Response: Nothing.
It’s rare, but it does happen that glitches occur. If I lost your comment, try again.]
Another possible exercise in wiggle matching:
Click to access sr05-02.pdf
From some melancholy Danes.