Bishop Hill The IoP blog on the Nature trick
Mar 11, 2010 Climate: CRU Climate: Parliament The Institute of Physics blog has a posting on the furore over its submission to the Parliamentary Select Committee. Among the gems are this:
[Mike's Nature trick], as mentioned by Jones in one of his e-mails to Mann, Bradley and Hughes, is a statistical method that is widely accepted in the climate community and is applied to proxy measurements in the years since 1960. It deals with the problem that some tree rings in certain parts of the world have stopped getting bigger since that time, when they ought to have been increasing in size if the world is warming.
"Widely accepted" is an, ahem, interesting way of putting it, given that Michael Mann himself says that nobody has ever grafted instrumental temperatures onto proxy records.
And there's more. Take a look at this from Rasmus Benestad:
According to physicist Rasmus Benestad from the Norwegian Meteorological Institute and a blogger for realclimate.org, Jones’ reference to "hiding the decline" could have involved removing some tree-ring proxy data from the analysis after 1960 to produce a curve that agrees better with the evidence for global warming.
Throw out evidence that doesn't match your hypothesis? Can he really have said that?
Reader Comments (46)
We can assume that Benestad said that. Dendro Rosseanne D'Arrigo testimony to Congress was supposedly that cherry picking the tree ring data was necessary if you wanted to make a cherry pie.
Yes, it bogles my mind that they think nothing of selecting the data to confirm what they believe.
Well let's see, according to you he said:
So, the answer to your question is no.
Basic reading comprehension really.
Frank, the hypothesis is:
"Trees are reliable thermometers",
the decline shows that they are not reliable thermometers.
Hiding the decline is hiding the evidence that trees are not reliable thermometers.
Here is the link to the CA Cherry Pie article.
http://climateaudit.org/2006/03/07/darrigo-making-cherry-pie/
Tells a very similar story
In a comprehension exercise Frank, you are supposed to look at the whole passage, and its meaning. You cherry picked your data (words): the important words are "removing some tree-ring proxy data from the analysis after 1960 to produce a curve that agrees better with the evidence for global warming." REMOVING DATA to fit the preconception better.
Naughty Naughty.
Since when has "agrees better WITH THE EVIDENCE" meant "fit the preconception better"? Frank makes a sound point.
This must be from that new enlightened scientific thinking where you know things in your gut and proceed to shape your evidence to follow that all wise abdominal wizard. Phil Jones, as I recall, mentioned his magical gut in those famous emails. Imagine any other scientific field where a scientist would express that with a straight face -- or, even worse, that people would actually argue that this is a sterling example of the exercise of truly wise scientific thinking. My own gut tells me that as a community, climate science is an example, par excellence, as to how not to do science and why we need separation of Science and State.
I've started thinking you're making all of these stories up Mr Bishop Hill. Because frankly I can't quite believe them!
Is data not evidence? Frank and BigYin seem to suggest that only evidence that fits is evidence, and that which does not is to be discarded. That way madness methinks.
When I read physics at Oxford in the 1970s I found incontrovertable evidence that data was manipulated in the Clarenden Laboratory to match preconceived ideas and 'groupthink'. I drew attention to it, and it was as much as admitted. My old physics tutor died last year. Another former physics undergraduate from my college writes WITH APPROBATION:
"This man's breadth of worldview rubbed off on all of us...my practical partner and I were waiting...to start the dull ritual of fudging our results and acquiring the tick in the box. There was a Daily Mirror discarded on the floor...[the tutor] launched into an appreciation of the skill of their journalism, their communication."
'Frank and BigYin seem to suggest that only evidence that fits is evidence, and that which does not is to be discarded. That way madness methinks'
or the statement of a control freak maybe?
they still hope to dupe/confuse the public with this spin.
keep the good work informing us undecided all.
'Widely accepted' in the same sense the Piltdown Man was 'widely accepted'.
Is McIntyre the new Weidenreich?
I'm in the pharmaceutical field. Can you imagine if we ran clinical trials like climate scientists? We know in our hearts that the drug is safe and efficacious. So if we see patients dying after getting the drug, but not from placebo, just keep going. Use only the data that fits with our preconceived notions about the drugs safety and efficacy!
If pharma did that people would be getting jail sentences. But if you do it in climate research, you get grants and even a nobel prize!
Presumably, in climate science, one has to fudge everything to agree. Wouldn’t it be ironic if the tree ring data were closer to the truth than the fudged UHI surface temperature data?
Well.. not wishing to state the obvious too obviously, but Frank O'D and Big Yin are fans of the work of the climatologists, and so it is only fitting that they should have absolutely no problem with selecting only the evidence that produces "a curve that agrees better with the evidence for global warming." It's entirely in-keeping with the approach to science carried by the climatologists' profession.
The glue that holds climatology and its advocates together is their collective apathy in regard to scientific method and integrity.
Could someone please provide the link to the IoP post that this article references?
[BH adds: The link is in the first line of my posting]
Nice the 5 comments are reasonable.
Anyone examined this "EYES" propaganda group in the Potsdam Institute?
http://www.pik-potsdam.de/research/research-domains/transdisciplinary-concepts-and-methods/project-archive/eyes/index_html/?searchterm=EYES
Creepy... Martin Welp of CRU Emails seems to be into this stuff.
Question: Since when has "agrees better WITH THE EVIDENCE" meant "fit the preconception better"?
Answer: Since......Forever
I can "prove" any hypothesis if I can choose which data to keep and which to discard....
@ Frank
Yes, Franky, and I see where your reading comprehension failed. You see, in one case someone is talking about throwing out evidence. In another, someone is talking about producing a curve that better agrees with a theory. Now to the layman I can see how those might seem like the same thing, but to a real scientist it's almost second nature to look at graphs and question how someone arrived at the curves they have. You see, curves are not data in the strictest sense. In fact plots of data are really just someone's representation of the data and they like statistics can be manipulated to exagerrate something, even if all of the data that went into making them is wholly accurate.
The fact that the person said, "produce a curve" that agrees better with the evidence is itself almost a direct admission that someone is trying to manipulate what a plot looks like so that it better matches the narrative. It's interesting that you used a sentence that is itself almost an admission of malfeasance to somehow demonstrate a lack of it.
One must learn to translate climatese into English:
"Widely accepted" means "we're making this stuff up as we go."
"Robust" means "we've hidden the raw data where you'll never see it."
"Unprecedented" means "we got a correlation coefficient of 0.034."
Well, I was relieved to read this, and it is leading me to consider legal action to clear my name and to seek damages for lost earnings.
You see, some years ago now, I was involved in a study of drug safety for one of our leading drug companies. It had been alleged that one particular widely prescribed drug we manufactured was implicated in serious side effects extending to death, in certain circumstances.
The pressure at the time was immense, the amounts of money were huge. The amount of data involved was also huge, because the drugs were being very widely prescribed and had been prescribed over a long period. If the allegations were correct, we would probably be involved in a total ban on sales, which would have had catastrophic implications for the value of our shares and for our reputation.
I was the senior researcher and statistician on the internal assessment, which of course resulted in a submission to the regulatory agencies.
Well, after a long succession of 18 hour days struggling to get to the bottom of the question, it finally became clear to me that all the data except for a small number of recent studies agreed with our prior hypothesis. The drug itself was blameless provided that guidelines for usage were adhered to. The guidelines for use included regular blood testing, to make sure that the levels of a certain marker did not fall below levels which we judged to indicate the early stages of some of the more disastrous side effects.
Our problem was a few internal and unpublished studies which appeared to show in a small number of recent results that the marker was not in fact a reliable indicator of the side effects. They cited results in which the marker had been at low or declining levels, but the side effects had occurred, and where they had in fact led to premature death.
At a particularly fraught review meeting with the team, I took the decision that we would omit this data from our assessment. It did not accord with all the information we had on the safety of the drug, and simply confused the issue.
I found out later that one of my team, curiously enough, someone who had not raised any explicit objection to this, had secretly recorded the meeting, and took her recording to one of the regulatory agencies. There was an enormous brouhaha, I was hauled before a disciplinary committee and, in the face of threats from the regulators, forced to resign. Use of the drug was curtailed, the studies were redone. It did turn out that our marker was not a robust indicator in all cases. But the safety of the drug was not a serious issue. The internal consensus that the drug was basically safe was correct.
This was many years ago. It blighted my career. I have always believed I was right on this however. There was nothing to gain by confusing the issue. So it is immensely cheering to me now to read that my own procedures are in fact, and contrary to what was alleged during my disciplinary hearings, commonplace in science where public policy issues are involved.
I shall be reopening my case and taking legal advice, and can only hope that all others who have been unfairly victimized for correctly representing the bulk of the evidence, and not distracting regulators and the public with outliers of dubious worth, take note and follow my example.
There are different ways of reading this but I think the editor of PhysicsWorld is giving the alarmists enough rope to hang themselves - which they have duly done. The IoP is steering a reasonably skillful line. Most importantly, they're not backing down on their submission and they're not prepared for one known sceptic to be made a scapegoat:
They feel a need for a statement of agreement with global warming which is pretty unobjectionable:
The 'action to address it now' is not specified. In the context of the EU's Emissions Trading Scheme, other global attempts at Cap and Trade and the biofuels debacle it's hardly a courageous phrase. But the IoP would no doubt say it doesn't claim to be an economics and politics think tank. There are no numbers given on the 'carbon dioxide in the atmosphere ... contributing to global warming' or its likely magnitude in the future. It's what Lindzen calls the trivial consensus. It doesn't constitute a basis for alarm. That last part of course the IoP leaves unsaid, which is what makes Lindzen a man of courage and the IoP shall we say going for the 'discretion is the better part of valour' option.
What the IoP should say to the John Houghton's of this world is that this scandal has done massive damage to the reputation of UK science - and understandably so. The IoP had every right to be concerned about this. They didn't need to wait until the various inquiries publish findings. The emails have been published and it is this that has done the damage to the public's confidence, for very good reason. The Select Committee was asking for submissions and one science organization stepped up to the mark. We continue to salute them for that.
Is FO'D this blog's resident troll or just a tease?
"It deals with the problem that some tree rings in certain parts of the world have stopped getting bigger since that time, when they ought to have been increasing in size if the world is warming."
So either
a) tree rings are unreliable and the proxy is meaningless, or
b) tree rings are reliable and the world has not got warmer.
I don't see how anyone can conclude "tree rings are reliable and the world has warmed" from the statement above.
*Unless* maybe tree rings were reliable until the 1960s, and then they hung out with hippies, wore beads, smoked stuff, and became a bit flaky.
@michel
It worries me that after so many years you don't realise that you were wrong in not disclosing adverse data.
Science ethics dictates that you must show all adverse data and write an explanation why they don't matter. You were only lucky in hindsight that your gut feeling proved to be correct, your conduct wasn't.
I have attempted to understand some of the individual views on the IoP submission and failed. Skeptics stay it does not go far enough, alarmists say it undermines all of their work. The submission seems pretty solid to me, it is predominantly addressing scientific method. Withholding data to defend the AGW theory is indefensible. If the theory were weak enough that replication of Mann's exact selection and processing criteria (which are NOT reproducible from the literature, most workers rely on the end product, or attempt an independent replication), or an assessment of how UHI really has affected certain thermometers in China REALLY threatens the science, then the science is weak and should be subject to very detailed scrutiny.
I accept that the IOP has based it's opinion of climate on this body of work, and it seems reasonable to use the work for forming a view, yet at the same time state that the method of working is at odds with 'the way we expected it would have been done'. This part of the discussion should not be about how sample selection bias, it should be about performing scientific work in a way which is open, and CRUCIALLY making it possible to re-create that same analysis using newer data in 20 years to understand IF there was a problem with any earlier analysis and why.
The Institute of Physical Training has issued a new statement to further clarify the earlier clarification published last week:
It has been brought to our attention that the Chairman of IPCC v. 2.0 has declared that AR5 will show incontrovertibly that the planet is cooling and that it is all our fault (except for his nibs, of course) so we have re-visited our old peer reviewed carbon dioxide papers in the archive and have found, to our horror, that they are all completely wrong!
Using new data sets, freely available on-line at Facebook, and new software downloaded from the public pages of the Met. Office website, we are delighted to take this opportunity to get in line.
We are now able to state incontrovertibly that the pale blue layer of frozen carbon dioxide, which can be observed for 340 days per year from a supine position on the beach at Benidorm, but only rarely from our ivory towers here in London, is not as we previously thought to be like a John Lewis Winter Snow Goose Down Duvet making us all get hotter, but is in fact an amazingly efficient conductor of heat like the base of a John Lewis All-Clad Copper-Core Collection Frying Pan. We chose John Lewis examples to enhance the understanding of our Members of Parliament when we give them technical briefings about private e-mails taken out of context.
It is clear that this conduction of heat away from the planet by this pale blue layer has been masked until recently by a really good duvet type gas called steam (or relative humility in as we say in the scientific community) which we understand is now disappearing due to the clearing up of unsightly adolescent blackheads (as seen on TV) on the surface of the sun.
Satellite data from the usually reliable NASA office just above a famous coffee shop (fully staffed by ex-pat UK graduates) in New York, tells us that the blue layer is getting thicker and may be approaching a dangerous tipping point, but we are planning to send some guys with beards up there to take real measurements if/when that other guy with a beard starts up his programme of eagerly awaited rocket propelled high altitude package holidays.
We are embarking on a joint venture with the Institution of Silly Machines in Westminster SW1H 9JJ (currently CMMGW enthusiasts, but we hope to get them whipped into line soon) to modify these rocket propelled holiday contraptions so that they can scape off the blue carbon dioxide stuff utilizing scaled-up versions of the robotic sampling device used to cut thin slices from clam shells, as revealed in this week's Naturist magazine. The shavings of excavated material will be brought down to earth and the space vehicles will land on runways to be built alongside the new coal fired power station sites with their fleets of carbon handling units. It is expected that there will be several hundred of these facilities where the blue stuff can be injected directly into the new Carbon Capture and Storage apparatus. These cutting edge devices are currently being designed by a Committee of 642 Members of Parliament (should be 646 but four experts from the Value for Money Oversight Sub-committee are helping the cops with their enquiries), which hopes to have the Tender Drawings and Specifications ready for the deadline of first of April 2010. During the first week of April or thereabouts they will be sent back to their constituencies to receive a good kicking and as none of them is expected to be voted back in it is essential that this important work is finished, without fail, by the end of this month.
Accountants at the Royal Jelly Society have estimated that this project should only cost about 2 or 3 trillion pounds over one ten years but it will be worth it if the scientists and engineers engaged in this Holy War can prevent the global average temperature plummeting to absolute zero by Boxing Day 2099.
Hans I think, although I may of course be wrong, that Michels post was meant to be ironic. He was showing how unethical any "hiding" of unfortunate data is by comparing climate science to a more familiar discipline.
Exclusion of data that do not fit the hypothesis is not new.
In earlier times, both Callendar and Keeling contrived desired outcomes by ignoring data that did not fit their hypotheses. Hence we have e.g. the 280 ppm pre-industrial levels of CO2.
Anybody care to explain how something can seriously be used as "evidence for global warming" after being modified so that it "agrees better with the evidence for global warming"?
Why not modify tree rings data so that they "agree better with the evidence" that I should be elected President of the United States and paid every day my weight in gold, as my appearance onto this planet more or less coincides with the beginning of the "divergence problem"?
@ Arthur Dent
Poe's law: "Without a winking smiley or other blatant display of humor, it is impossible to create a parody of Fundamentalism that SOMEONE won't mistake for the real thing"
Scientist for Truth
When I read Physics at Oxford in the 60's the first step in any Practical exercise was to look up the answer in Kaye and Labey.
I remember a terrible fuss when someone got the 'wrong' result for the thermal conductivity of rubber and argued that much of the heat was leaking away from three thumping great metal lugs on the apparatus rather than passing through the sample. 'But everyone else gets the right result.' said the Demonstrator.
@Bernd, No it only demonstrates that Beck and Jaworowski don't have a clue about how diffusion physics works. The adverse data was not hidden, it was explained, a subtle but important difference.
Tree rings in denial?
Dreadnought
"When I read Physics at Oxford in the 60's the first step in any Practical exercise was to look up the answer in Kaye and Labey.
I remember a terrible fuss when someone got the 'wrong' result for the thermal conductivity of rubber and argued that much of the heat was leaking away from three thumping great metal lugs on the apparatus rather than passing through the sample. 'But everyone else gets the right result.' said the Demonstrator."
When I read physics at Oxford in the 1970s there were results of the thermal conductivity of the material in a time line since the 1950s. It was posted prominently over the doors to the laboratory so that no-one could be in any doubt what 'the right result' was supposed to be. There was the clearest evidence of fudging the data because there was large scatter due to errors in the early years, but as the years went on the scatter reduced significantly, although the apparatus was unchanged - as year succeeded year, it was possible to 'improve' the estimate by looking over longer history. Of course the problem here is that experimenters were biasing their results to conform to what was expected, so bias and 'groupthink' set in. Moreover, the expected range of experimental error was being narrowed. Anyone who did the testing without bias and the benefit of hindsight stood a good chance of having their results rejected, as you say.
This is what happened to me, and I had a blazing row about it in the Clarendon Lab. It was the sonometer experiment. My partner and I had done the experiment properly. We had spent around twice as long as anyone else. I had interviewed my colleagues who admitted that they had not done it properly, but were only interested in presenting the consensus view to get top marks, and also to get out early to play squash etc. When I came to present my results, the demonstrators would not accept them. I asked the demonstrators whether they had conducted the experiment themselves. They had not. Yet they were passing judgment on someone who had. But they refused to accept my results because, they said, "we know" what the result should be. Worst of all, they were not going to give me the marks, yet they were giving full marks to those who admitted that they had been dishonest. For a person interested in truth, this really was too much. I refused to fudge my results, and they refused to sign them off. They said they would give me the marks if I simply fudged my measurements in their presence. The altercation ended when on final refusal to fudge the measurements I was met with the response from the demonstrators (which has been ringing in my ears ever since, so I can tell you word perfect): "Don't you want to be a real scientist, then?" To which my result was "No" if it meant falsifying the data. They were incredulous - surely I would falsify the data for the purposes of self-interest, self-promotion and advocacy - everyone else did, it was the done thing, that's what REAL scientists do. Everyone learned a very important lesson - to get on you need to be prepared to falsify data and be prepared to present the 'facts' it in a biased way (hence the Daily Mirror allusion in the first comment).
And here we are, over 30 years later, and things have only gotten worse.
@Hans
Which "diffusion physics" do you mean? Molecular diffusion?
What did Callendar know about the diffusion physics?
Pettenkofer's method is a chemical method for determining the mass of CO2 in an air sample.
Why were Müntz & Aubin's data chosen by Keeling when the method used by the former was based on an open system with poor temperature control and therefore a very large margin of error in preference to historical data with much smaller error bars? Explain!
What do "exploding" calthrates, when the static pressure from the weight of ice is relieved as an ice core is raised to the surface, have to do with diffusion physics?
Why is an "average" guess at levels of CO2 significant when the temporal and spatial distribution of the gas mean that outdoor concentrations vary by a factor of about 3:1. How is an "average" useful for global calculations of heat transfer when heat transfer and the radiative properties are strongly non-linear?
Bernd Felsche
"Exclusion of data that do not fit the hypothesis is not new.
In earlier times, both Callendar and Keeling contrived desired outcomes by ignoring data that did not fit their hypotheses. Hence we have e.g. the 280 ppm pre-industrial levels of CO2."
Atmospheric CO2 was accurately measured in numerous locations before Keeling: the arctic, Newfoundland, Ireland, Sweden, Germany, India etc, and they consistently showed a rise to levels as high or higher than today by the late 30s/early 40s. The 1930s was also the warmest decade in most habitable parts of the earth, and the large increase in CO2 cannot be attributed to anthropogenic causes.
Callendar and Keeling cherrypicked the historical CO2 data to suit their presuppositions: AGW based on CO2 emissions. It was what Spencer Weart, Director of the Center for History of Physics of the American Institute of Physics, euphemistically described as "a fortunate mistake". Even Weart admits that "Callendar seemed to be picking only the data that supported his case...Callendar’s calculations...ignored much of the real world’s physics…he argued for conclusions that mingled the true with the false". Great scientist! Callendar and Keeling have since been canonized as saints in the Church of Global Warming.
For a somewhat acerbic look at Callendar, see my post:
http://buythetruth.wordpress.com/2008/10/25/a-fortunate-mistake/
Since when has "agrees better WITH THE EVIDENCE" meant "fit the preconception better"? Frank makes a sound point
No - he doesn't. We have a range of evidence in front of an investigator, and let's assume that some of it will seem to support a particular hypothesis, and some of it will seem to discredit that same hypothesis. If he removes some evidence in order to produce a result that better agrees with the other evidence he is not presenting the whole result of his work - he is presenting a subset, chosen to support or discredit the hypothesis. Making that selection is proof that he is not letting the evidence guide him, so he must be making that selection based on a pre-concieved notion, that is, an idea that originated prior to the examination of the evidence.
It is possible that the quote was sloppy, and that the meaning was to 'remove evidence that was irrelevent or unreliable' - but as quoted, it means more-or-less exactly what the Bishop said.
"It is possible that the quote was sloppy, and that the meaning was to 'remove evidence that was irrelevent or unreliable' - but as quoted, it means more-or-less exactly what the Bishop said."
We can go further.
If, Frank, you think it is so important to tax ourselves into oblivion, hand an awful lot more control over our daily lives to those who wish to control us and generally risk returning to pre-industrial levels of life expectancy and living conditions, perhaps we ought to expect those presenting evidence that we must do so to be a little clearer and less ambiguous?
Or is that a scandalous attack on the very foundations of the scientific method?
The division between "skeptic" and "diehard" sits exactly on rejection or acceptance of the 1940 CO2 outliers. Sorry these local outliers are not representative for the global base level.
http://www.ferdinand-engelbeen.be/klimaat/beck_data.html
http://www.ferdinand-engelbeen.be/klimaat/jaworowski.html
Those who understand the "divergence" problem and yet sincerely think that the proper thing to do is to eliminate the post-divergence tree rings from consideration simply do not have a scientific mindset, and probably never will. It's as simple as that.
Hans Erren "The division between "skeptic" and "diehard" sits exactly on rejection or acceptance of the 1940 CO2 outliers. Sorry these local outliers are not representative for the global base level."
You seem to have prejudiced the issue by referring to them as 'outliers'. Your use of the word 'local' tends in the same direction. By definition, all measurements at different parts of the globe are 'local'. The interesting thing is that the trend of increasing CO2 was observed in many 'local' places thousands of miles apart - they weren't 'outliers'. This was not some 'local' phenomenon in the sense that it is not representative of a phenomenon representing at least a large part of the globe. You could posit a testable hypothesis to explain this, such as that all workers were using tainted chemicals from the same source, leading to erroneously high results: i.e. independent observers contaminated by a common source. If you find something like that, you'll do us all a favour.
In general, I find direct contemporaneous measurements more compelling than proxies such as tree rings, stomata etc. As for the validity of ice core measurements, that relies on all sorts of assumptions and presuppositions as well as recovery difficulties such that I would never choose that method in preference to a direct contemporaneous measurement. Anyway, they are pretty 'local' themselves, aren't they? Ice cores can only be drilled in a few places, hardly representative of the globe in terms of location. To say that it is representative of the globe may be a sound assumption, but it is an asumption nevertheless.
And I don't like this argumentum ad ignorantium in the Engelbeen posting:
"that 210 GtC were absorbed in ten years time...is physically impossible. There simply is no process in the natural world which can absorb such a quantity of CO2 in such a short time."
What the writer means is that he can't think of one. I'm more persuaded by solid evidence than an argument from ignorance.
"Widely accepted" is an, ahem, interesting way of putting it ...,
Tut tut, now you should know better than that. The clause needs to be read all the way to the end to capture the full meaning.
widely accepted .... in the climate community
See - all makes perfect sense now.
"Throw out evidence that doesn't match your hypothesis? Can he really have said that?" - Not in so many words but that implication is clear.
He has expressed in words what Mike and the rest of the tricksters did in deed.
@Hans
I understand by the links that you've quoted that by "diffusion" in your earlier posting, that you probably meant "dispersion". As to Beck not being aware of the effects of dispersion: The real history of CO2 gas analysis
From my reading of Beck's first work on historical CO2 measurement (several years ago), he was well aware of local bias. He used the ability of the original results to detect a dirunal fluctuation as one of the metrics of data quality. And in the higher-quality data, was able to detect "lunar" signals. Data included by Beck were from crossings of the Atlantic conveyor current, which clearly illustrated the high spatial variability of CO2 levels near the surface.
I'll give you some more time to think about the other question I posed; that of the "significance" of an average, given the high variability of CO2 concentrations and the non-linearities in heat transfer.
@ScientistForTruth
Engelbeen's assumption that the 100 ppm used as the basis for the 210 Gt is where he goes astray.
If I goe to to a random location on the globe at a random time and take a gas sample, it is very, very likely not to contain the "official" average level of CO2; even within +/-10 ppm.
Beck's republishing of the CO2 concentrations from 1920's Atlantic crossings showed CO2 concentration vaying by about 200 ppm (highest in the tropics) in the space of a couple of months. Of course that wasn't all absorbed/released over the period. It illustrated the wide spatial variation of CO2 concentration.
CO2 near the surface may be thought of like clouds of water vapour in the sky instead of a uniform "fog". And the contribution to heat transfer is probably just as disparate.