Guest post by Frank Lansner
IPCC – How not to compare temperatures – if you seek the truth.
There are numerous issues discussed intensely when it comes to IPCC-illustrations of historic temperatures, here for example the illustration from IPCC Third Assessment Report:
Fig 1. Taken from IPCC TAR
In short we have heard of problems with 1) the Mann material, 2) the Briffa material, 3) The cherry picking done by IPCC to predominantly choose data supporting colder Medieval Warm Period, 4) Problems joining proxy data with temperature data mostly obtained from cities or airports etc, 5) Cutting proxy data of when it doesn’t fit temperatures from cities, 6) Creating and Using programs that induces global warming to the data and finally 7) reusing for example Mann and Briffa data endlessly (Moberg, Rutherford, Kaufmann, AR4 etcetcetc).
But, as I believe another banal error needs more attention:
8) Wrong compare.
Imagine for a moment that none of the above mentioned problems 1) – 7) has any impact and then lets just focus on the comparing itself. The data of the proxies suffer from 2 impacts:
A) “Technical averaging” – Impact of many series of date summarized.
Check out what happens when summarizing many datasets of temperatures, an example, the cooling episode 8200 years ago:
Fig 2.
Data taken from: http://wattsupwiththat.com/2009/04/11/making-holocene-spaghetti-sauce-by-proxy/
The white graph with the red squares are the resulting average graph: More temperature sets added together tends to flatten the average. Notice for example how many datasets certainly has a down peak between 8000 and 9000 years ago, but the timing for these datasets are slightly off, and so the down peak is almost gone.
So, to some degree we can expect multi proxies to yield an averaged overall graph.
B) “Direct averaging” – on top of the technical averaging, the data series are often averaged further to some degree using 30, 40 and 50 years Gaussian filters.
The result of averaging by A) and B) is, that the variability of the IPCC graphs on a decadal ´timescale are limited to just tenths of a degree K. But in reality, if there where any real temperature peaks on decadal time scale in the Medieval period, we will would not see these much in the typical data series IPCC shows.
Is this a problem?
Well, it certainly becomes a problem if these “super averaged” data are compared with data that is NOT quite as “super averaged”. And this faulty compare is just what IPCC do.
IPCC “Super averaged” data from proxies, are typically compared to “Observed” temperatures, that is, recent temperatures not at all submitted to the same degree of averaging.
Technical averaging – type A) – is to some degree not happening for observed temperatures, so how about type B), the direct averaging, filtering?
Well, For the IPCC graphis shown in fig 1 above, the IPCC text says: “All series were smoothed with a 40-year Hamming-weights lowpass filter, with boundary constraints imposed by padding the series with its mean values during the first and last 25 years.”
Explanation: If your data ends in year 2000, then the last genuine 40-year averaged/filtered point on the graph would be a point for 1980 with average of 1960-2000 near +0,2K anomaly. But the IPCC graph for observed temperatures ends at +0,43 K around year 2000. This more resembles the normal five years average of GISS year 2000 data:
Fig 3. Giss temperatures illustrated in year 2001.
So for IPCC/Mann etc. to get a year 2000 temperature as high as +0,43K, they must have used just normal 5 yr avg. A longer average period would yield lower temperature for the last year.
So, when IPCC wrote “with boundary constraints imposed by padding the series with its mean values during the first and last 25 years.” – they mean: “We don’t use 40 year average/filter in the last 25 years…!”
So the bottom line is: IPCC compares “super averaged” temperatures of the medieval period with a peak in modern temperatures only submitted to 5 years average.
IPCC basically compares a peak in temperatures in recent years with super averaged medieval data where peaks are more suppressed to conclude how much it is warmer today than in the MWP.
This is a problem !
From this illustration it appears that the peak after 1998 to some degree appears related to the big El Nino 1998 peak, here from appinsys:
Fig 4.
So, where there no El Nino peaks in the medieval period that could have affected the compare with recent temperatures? Yes, there where: http://co2science.org/articles/V12/N5/C2.php
So we have every reason to believe that there where also temperature peaks in the medieval period – peaks that just might resemble the recent El Nino Peak.
So no excuse for the IPCC to compare a modern temperature peak with medieval average temperatures.
This is banal, of course, and even IPCC must have been aware of this, one should think.
Here: An illustration where the single year 2004 for observed temperature data explicitly is used in comparison with the super averaged medieval temperature data.
Fig 5. (from here)
Cheers!
If we just look at el-mimo peaks or even decade peaks. The proxies could easily have 5-10 years errors on the time axis. And then, averaging them would eradicate those peaks.
One of the other issues of proxies is the dating of them. All their dating from previous centuries and millennia are approximate, no matter how careful researchers read the tree rings or ice cores. Just finding a starting point is an art. And Carbon 14 dates are all, or nearly all, approximates for earlier centuries, so their placements along the time scale are critical for preserving the peaks and valleys that MUST have occurred.
Then with different proxies the peak of one graph may be shifted in time versus another, and then when blended/averaged, the peaks and valleys of that averaged graph would become flattened across that span of time.
Certainly some of this happens.
This brings up the question: Do the “combiners” of all this (Mann et al) make any effort to rectify (meaning to shift forward or backward to match other proxy graphs) the timing of these separated graphs, or do they simply take them “as is” from the specialists who have created each database?
Each specialist would have no reason to know his or her own data is shifted in time from other proxies, due to the unknown precision of time for each proxy.
It is not just the height of the graphs (temps) but also the time element that has to be considered. In combining graphs, if no effort is made to rectify the various traces in time, flattening of the resultant graph is inevitable – and MISLEADING.
Method matters – in the combining process perhaps more than any other time – since graphs can only get more flat, not less. Avoiding this flattening should be of prime concern to the combiners of other people’s work.
“jdn (06:53:25) :
[…]
If you’re going to criticize the IPCC for making a mistake, you should put your finger on the mistake and not simply take issue with an averaging technique.”
He pointed out a very basic mistake, namely the production of a convincing-looking graphic with a technique that is not suitable to deliver reliable results. “Take issue with an averaging technique” ???? You seem not to know what you are talking about here.
The effect of the mistake is that high frequencies – noise if you will are dampened in the past but as we approach the present they are less and less dampened.
Obviously this will hide extrema in the past record.
Thus, the assertion that current extrema are unprecedented has been PROVEN to be without base by Frank Lansner.
If this seems like nit-picking to you you obviously have no understanding of science at all.
You can’t just make the data say what you like and still call it science.
jdn (06:53:25)
So you are saying it is OK to mix smoothed data and pure data in the same line, and to express a view that the movement from one to the other represents a trend? And it is OK to use a filtering method completely outside its sphere of operation, when it conveniently gives the answer you want, and simpler methods do not?
I guess you are. By the way, the comma after “and” is redundant. Apologies if you are not a native English speaker.
Re: jdn (Apr 5 06:53),
I thought Frank’s article was clear and understandable and the indictment of IPCC statistical techniques evident. Are you sure that you are not just being willfully obtuse?
When geologists first saw the Mann et al. curve, we all just laughed because of the solid geological evidence for the Medieval Warm Period and Little Age published in hundreds of papers for decades. Our conclusion was that either the trees that Mann and others used were not sensitive enough to record climate changes or else it was a total fraud. The Climategate emails showed us the answer to that question. Instead of endless arguing over the details of their bogus curves, it would make more sense to just go to the geologic evidence which shows beyond reasonable doubt the Mann et al. curves are worthless.
Good article, Frank, and thanks for all the effort you put in. Another nail in the coffin of IPCC veracity.
Your ‘Danglish’ is very easy to understand and the odd typo is difficult to avoid entirely when one’s writing is driven by ideas.
As a native speaker of English and a former teacher, I get irritated by those who get picky and pedantic over ‘correct’ English useage and who have lost sight of the fact that the essential job of written language is to communicate. For those who didn’t grow up with it, English is a damnably hard language to learn, but everyone who prides themselves on their knowledge of it would do well to remember the old adage – ‘every rule in the English language has an exception which serves to prove that rule’.
jdn (06:53:25) :
Frank, your article is full of mistakes, and, your point isn’t clear either. If you’re going to criticize the IPCC for making a mistake, you should put your finger on the mistake and not simply take issue with an averaging technique. Your article hasn’t added to my knowledge of any issues.
I have a feeling you are writing outside of your normal pattern of speech. For that, you need an editor.
————-
What a mean-spirited comment. I found Dr. Lansner’s revelations to be very enlightening. I struggle to understand all the minutae of some of the statistical arguments put in other posts, but the significance and importance of this guest post are immediately apparent. In a way I’m surprised nobody has pointed this out before! The smoothing of a series of unrelated proxies into a bogus super average which is then compared with a different scale of averaging in recent temperature records offends on two levels, as has been pointed out by earlier commentators. Thank you Dr. Lansner for this illuminating information.
I second Smokey’s demand: list for us the mistakes that offend your delicate sensibilities, jdn.
“DirkH (07:38:07) :
[…]
You can’t just make the data say what you like and still call it science.”
And i’m fuming about this display of scientific incompetence – but maybe i should be laughing. We’ve all been had by the IPCC and their people; they’ve been enjoying good salaries for such a long time and just messed around with the numbers any way they saw fit. That’s chutzpah, a quality all in itself. The great climate swindle.
Here’s a description of Hamming window coefficients
http://en.wikipedia.org/wiki/Hamming_window#Hamming_window
and here’s a good description with the Fourier transform
https://ccrma.stanford.edu/~jos/sasp/Hamming_Window.html
The reason for using a Hamming window is its optimal suppression of ripples in the transform. So just messing with its width should have some effect on the spectrum, right? Makes you wonder why you ran a filter in the first place.
This comment is off topic for this thread but might be of interest to WUWT readers. Many here are aware of William Connolley’s role in keeping the climate change articles on Wikipedia scrubbed of anything he doesn’t like and inserting questionable material in it to support what his point of view. With that in mind there is an article up at:
http://pediawatch.wordpress.com/2010/04/05/a-good-example-of-william-m-connolleys-work-on-wikipedia/
which discusses some minor fiddling he has done with the aid of his friends on other blogs, and which points out, among others, the following diff:
http://en.wikipedia.org/w/index.php?title=The_Real_Global_Warming_Disaster&diff=next&oldid=342554628
in which he is scrubbing mention of WUWT from the article on Christopher Booker’s book “The Real Global Warming Disaster.”
So they pad beyond the end date, and then use the 40 year filter.
Were is the evidence that they use a 5 year average?
This Climategate email has the raw unsmoothed and untruncated (for the post-1960 decline) data from Briffa’s first reconstruction in 1998 covering 1402 to 1995. It was sent to Michael Mann from Tim Osborne.
It is calibrated to the 1881-1960 average so the temps are comparable over the period. Chart it up and you can see what the raw tree-ring data really shows.
http://eastangliaemails.com/emails.php?eid=146
939154709.txt
I am concerned that no-one has raised the skewed ‘geographic’ averaging that occurs with the use of the proxy data. The coverage of the proxies is hardly ‘world-wide’ yet they are used as indicators of the entire planet. Is it safe to claim to know the past warmth in the Amazon basin from tree rings measured in Siberia? Most proxies are of small, sometimes very small, geographic spread, and where there has been an attempt at validation they have been shown to be inaccurate. Averaging instrumental temperature over the globe is bad enough (it should be heat content the enthalpy of the atmosphere varies hugely with humidity) but averaging multiple inaccurate proxies for temperature just provides huge scope for cherry picking. That is before the mathematical and statistical games are played.
BTW –
After that last thought dawned on me (feet2thefire 07:33:30), I am tending to think the CRU people and Mann – especially Mann – went where no man has gone before in accumulating all the work of specialists, and that in the process they made assumptions they didn’t even know they were making. These main assumption was that they could just take the data as found and – as complicated as it was already – process it in a way that DID flatten curves, though they did not see that at the time.
Frank has made a VERY important observation here, pointing us to this flattening issue.
In breaking new ground, they wouldn’t recognize all the pitfalls that lay along the path. In addition, they would have wanted to take at least some shortcuts (read “assumptions”), so as to not take far longer to get their results.
(Yet, having found a nearly flat graph that eliminated the MWP and the LIA, Mann clearly should have seen that he’d done something wrong; he HAD to know that people would jump all over his Hockey Stick. That he was a total jerk when they did only made him look like he had something to hide. That he went to the lengths he did to destroy the efforts of others to correct the graphs made him look vicious and despicable. He should have instead welcomed efforts to improve his work.)
Add that to Mann’s Primary Component failures pointed out by Steve M and (and Ross M?), and the Hockey Team’s work was flawed, even if they didn’t intend it.
With the time-flattening of their failure to rectify different curves in time, added to Mann’s lack of statistical expertise, I am pretty certain that they honestly believe they have done nothing “wrong.” They are perhaps even now ignorant of their errors.
But I think that they were also are insecure about their findings (they WERE blazing trails in the beginning), and when people are insecure about their methods/work, they are very easily caught up in efforts to neurotically defend the work.
I am certain that their work will be overturned, on the basis of better procedures that are already being discovered. In the big picture, their work shouldn’t be so heartily denigrated as most of us do here, but appreciated as a strong first effort, even if it was flawed by misguided assumptions.
On their part, they need to be open to the improvements, not paranoid about someone else “finding flaws in it.” But they shouldn’t just blindly or blithely accept such developments; they should always challenge the proposed improvements, so that other assumptions don’t prolong the entire already drawn-out process of getting a true temperature history of the planet.
Very few pioneers get everything right the first time. This current process of finding and fixing the flaws in the work that began with Mann’s work did not need to be so contentious. The Hockey Stick was just WRONG. Mann should have welcomed Steve M’s corrections. If it wasn’t Steve, then it would have been someone else, sooner or later.
The real problem came when other people – people with agendas – took Mann’s work as gospel, as confirming their environmental concerns, and then who created a “whole cloth” out of it that simply wasn’t true, and then carried it into the political sphere and started laying heavy trips on governments and the public about humans killing the planet.
There was, and IS, no firm and automatic connection between “it looks like the climate is getting warmer” to “humans are the cause of it all.” We here think there is no connection whatsoever, of course.
Seriously, IMHO that is a connection that came out of the environmentalists’ individual and collective psychology, which then was played upon an unsuspecting public’s desire to not kill Bambi’s mother, the environment. Once the “planet killers” concept kicked in, there was no logic that could convince the public that we should all just stop and think this thing through. All arguments for the public became the Cautionary Principle – “But shouldn’t we err on the side of caution? What if they are right?”
Once the public thought that, nothing short of the nuclear option was available: Logic and rational discourse were no longer possible. Corrections would be VERY difficult to make, and it would be a long, uphill battle.
It didn’t have to be.
Henry Chance: “Why do we insist on averaging temperatures for the planet?”
GOOD question. My opinion is that it is because the whole scam started with “Global Warming.” When that failed to stand up to scrutiny, the focus was moved to “Anthropogenic Climate Change”, but little was actually changed except the verbiage even though it is not possible to accurately describe (or forecast) “climate” with temperature alone.
I also believe that in the beginning it all boiled down to an easy way to comply with research grants ($$), regardless of its actual merit. Another plus is that it is easy to accomplish and made impressive graphs.
I contend that if you put any stock in “average temperatures,” you need to spend a few days wondering about the high deserts of the US in winter time. I guarantee you won’t even notice the “average” temperature when it goes by as the sun goes down!
Frank, thank you for articulating something I’ve been trying to say for years. As pointed out, your statement may need some refining for people that have to have everything spelled out for them.
jdn (06:53:25) : Just because you’re not grasping the obvious statement the article is making doesn’t mean its “full of mistakes”. I’ll try to help, after all the averaging, smoothing, extrapolating, proxy splicing, ect….., when you see a graph comparing historical temps to present day temps, you’re not seeing a proper comparison. It is an “apple to orange” comparison and totally meaningless. It is stating our recent spike is illustrated against smooth historical data. Even the standard set by alarmists is smoothed over a 4 decade period of time. Of course that shouldn’t be confused with a smoothed average over a millennium period of time. THEY ARE NOT THE SAME.
Turboblocke (03:43:15) :
I was going to seriously respond to a couple of your comments until I saw your obvious attempt at humor when you stated, “…….. neglects to mention the benefits which range from £20.7 – 46.2 billion/year.”
Can you please list some of the alleged “benefits” in term of monetary value, please?
@jdn (06:53:25) :”Frank, your article is full of mistakes, and, your point isn’t clear either. If you’re going to criticize the IPCC for making a mistake, you should put your finger on the mistake and not simply take issue with an averaging technique.”
Actually, his points are sufficiently clear for anyone who wants to understand them. He clearly summarized several known IPCC mistakes and then pointed out that one more has not been sufficiently analyzed. That’s why the concentration on averaging technique. As for the “mistakes.” I presume you are referring to his non-native English. Try reading the existing comments before blasting off. For that matter, re-reading the article itself if you don’t understand it the first time.
“Your article hasn’t added to my knowledge of any issues.”
I think we can all agree to that. But it’s not Frank’s fault.
Hmmm,
Remove (or be unable to identify) all the short term temperature peaks from the Mediaeval Warm Period.
Note a short term peak that has just occured.
Panic in all directions because there is no known past peak equal to the recent one.
Ignore the historical and proxy evidence that it was warmer than now during the Mediaeval Warm Period, the Roman Warm Period and the Minoan Warm Period
Recommend global energy rationing and world government on that basis ?
Are we governed by the mentally challenged ?
IPCC compared average temperatures for MWP and the present, not an average in one case and a peak in another. Your claim here lacks merit.
On a more general note, you are asking us to believe temperature reconstructions put together by amateurs, rather than the ones carefully developed and peer reviewed by hundreds of specialists at NOAA, NASA, CRU, and others. It’s beside the point that your data is not correct. It is unworthy of consideration.
Historical temperature reconstruction, including those for recent decades, is a highly technical and detailed process, requiring extensive training. Questioning the basic expertise and integrity of those who develop these global temperature charts for international organizations is not credible, and is unsupported by actual scientific evidence.
Frank,
looks to me like another very important contribution to prove AGWT wrong.
It deserves to get reviewed and published in a more official way.
Stephen Wilde (08:37:22) :
“Are we governed by the mentally challenged ?”
Go here for an answer. http://wattsupwiththat.com/2010/04/01/congressional-tipping-point-not-an-april-fools-joke/#more-18094
While there are several other examples, this one entry should sufficiently answer your question.
mike roddy (08:39:54) :
Sorry “Historical temperature reconstruction, including those for recent decades, is a highly technical and detailed process, requiring extensive training.”
is absolute Cr*p, are you saying Mathematicians & Scientists using the published procedures are not capable of doing it without “extensive training”?
Of course to do it the “RIGHT WAY”, i.e. to show AGW takes very special training in deed.
“MikeN (08:01:08) :
So they pad beyond the end date, and then use the 40 year filter.
Were is the evidence that they use a 5 year average?”
The 5 year average is an estimate using GISS global temperature data. If you use around 5 yr average on GISS data published year 2000-2001, this enables an end point that matches the endpoint in the IPCC graphic rather well : 0,43K.
For some reason, the IPCC did not write themselves: “For the last years we use 5 yrs averaging enabling big warm temperature peak, and then we compare with older temperatures that we have submited to 40-years averaging and thus suppressing peaks markedly more.”
IPCC should make it quite clear in their legend text what they do.
K.R. Frank Lansner
mike roddy (08:39:54) :
“IPCC compared average temperatures for MWP and the present, not an average in one case and a peak in another. Your claim here lacks merit.
On a more general note, you are asking us to believe temperature reconstructions put together by amateurs, rather than the ones carefully developed and peer reviewed by hundreds of specialists at NOAA, NASA, CRU, and others. It’s beside the point that your data is not correct. It is unworthy of consideration.
Historical temperature reconstruction, including those for recent decades, is a highly technical and detailed process, requiring extensive training. Questioning the basic expertise and integrity of those who develop these global temperature charts for international organizations is not credible, and is unsupported by actual scientific evidence. ”
Mike, they were averaged using different time sequences and periods. THIS ISN’T ASTRO PHYSICS, this is simple base mathematics that any high-school grad should understand. The final comparison illustrated in the graphs are in no way correlated. One can’t average over different lengths of time and not expect different trends in the graphs.
Your appeal to authority lacks any credibility because you didn’t appeal to the proper authority. You could start with any college level algebra instructor, but a statistician would probably be better. You can peer review until the cows come home, but until you include some one with appropriate math skills(or in the way your think) “credentials” it is all meaningless. But then, once someone with the appropriate math skills gets involved, it will be declared meaningless anyway.
Why do you think they had to hide the decline? Because if they didn’t, tree ring proxies become invalidated. THEY DON’T EQUATE WITH RECENT OBSERVED TEMPERATURES(that’s not my observation, that’s the hockey teams). This then calls into question how they KNOW they are even correlated with historical temperatures. They don’t. It isn’t possible for them to know if they don’t correlate with observed temps. So, then, how do they know we are warmer than the past using proxies? They don’t. But they can convince you that apples equal oranges because they work for the NOAA!!!!????!!! Nice.
Perhaps for those who don’t “get” Frank’s point, this may help or may not:
His graphic of a very brief moment in time confused me at first (the Holocene temperature proxy series for 9000-8000bp).
Perhaps segregating the curves and pointing out that there ARE peaks and valleys would be useful, and then showing the averaged graph, separately – and then asking “Where did the peaks and valleys go?”
Isn’t that Frank’s main point, that the peaks and valleys disappear, the more data sets one includes?
* * * * *
On my own point – of the peaks and valleys perhaps being shifted forward or backward in time:
When I look at that first graphic here, I see that the red, the bright green and the medium-light blue graphs have serious dips in them. The red one is out of phase with the other two. It is probable that the time scale on them is merely shifted a bit, because of the methodology or the inherent errors in each method for determining the time element.
If one averages those three, the red graph’s shift will start to flatten out the other two. Even though there is a clear and severe dip in all three, the averaged graph will show less severe of a dip.
And the more graphs one adds in the averaging – if there are shifts in them like this – the more the peaks and valleys disappear.
IMHO, if the measurements of time in all of them was absolutely correct, then the peaks and valleys would show up in the averaged graph. Why? Because I think that the measurements of temps are all probably pretty accurate – because the warming and cooling DID happen and DID have effects. But if we can’t measure the time of these occurrences well enough, then when we combine/average the data, then the peaks and valleys start to cancel each other out, flattening the curves.