Cohler and Soon – 2026

A critical debate in climate-change physics

Jonathan Cohler asked me to remove my reposting of his and Willie Soon’s March 2026 paper under review, titled “Dominance of Natural Atmospheric CO₂ Dynamics: Falsification of Anthropogenic Attribution Through Mass-Balance Analysis, Isotopic Diagnostics, and Thermodynamic First Principles.” … However, you may freely download a copy of their paper at the above link or at https://papers.jcohler.com/co2. – This is the first time anyone has ever asked me to remove a reposting of their paper since I began my website in 2009. However, I can post their interview on the Tom Nelson show because it is public property:

This discussion of Cohler and Soon’s March 2026 paper in review began with an email challenge from David Andrews on August 30, 2026. Andrews sent copies of his email to Judith Curry, William Happer, Herman Hardy, Demetris Koutsoyannis, Dai Ato, and Ed Berry.

Unfortunately, Jonathan Cohler asked me to remove his replies to David Andrews from my website. So, I have removed them.

You can still read David Andrews’s challenges below, but you cannot read Cohler and Soon’s replies to Andrews.

This week, I will post my comments on Cohler and Soon’s paper. My comments will be very favorable to their science but will still raise important issues, such as their misinterpretation of my 2021 paper and their omission of my 2023 and 2023a papers, which contain information they claim is original in their new paper.

Prior to receiving Jonathan’s requests that I delete his paper and replies to Andrews, I had written the following in this space: (did this cause them to remove their replies?)

Andrews’ comments and replies are highly relevant to my post Berry v. CO2 Coalition, which has censored Berry since 2019 for the same issues that Cohler and Soon support here.

Notably, Cohler is a member of the CO2 Coalition, but he disagrees with it on the very issues it has secretly censored Berry on since 2019. If the CO2 Coalition is consistent, it should censor Cohler and invite Andrews to join the CO2 Coalition.

I am astonished that the authors of a paper under review would deny the public access to their replies to challenges from other scientists, such as David Andrews. I believe all scientists should allow the public access to their professional opinions and debates. This is especially important in the highly political science of climate change.

David Andrews comments on the above paper by Cohler and Soon

To: Jonathan Cohler and Willie Soon   

 Date: August 30, 2026

Copies to: Judith Curry, William Happer, Herman Hardy, Demetris Koutsoyannis, Dai Ato, Ed Berry

Your paper “Dominance of Natural Atmospheric CO2 Dynamics” is “under review”, is not published, and perhaps never will be.  But it is out there and you push it on a Tom Nelson podcast.  Therefore, I am making some unsolicited comments here.  You may have heard them already from other sources.  You have much to learn about the constraints observations and carbon conservation put on the carbon cycle, about the definitions of variables used in radiocarbon analysis, about the sometimes subtle but significant effects of “disequilibrium isotope fluxes”, and about the very real distinction between residence time and adjustment time.  These are all old topics, but the skeptical ethos that “science is never settled” means that wrong analyses get recycled and “skeptical pseudo-science” goes around in circles.  Using AI to summarize a collection of old wrong papers with similar errors is merely an efficient way to go backwards.

You skirt around the most compelling evidence that the atmospheric CO2 increase is anthropogenic, providing correct relevant numbers but apparently not seeing their implication. Referring to “mass balance” but not really applying carbon conservation appears to be an attempted sleight-of-hand.  You correctly note that over the last century the growth rate of atmospheric carbon is only about 44-47% of the rate of human emissions. (Bringing in alternate definitions of “airborne fraction” is neither necessary nor helpful.) That means that even though the natural sources of atmospheric carbon (ocean outgassing, vegetation decay, …) are indeed about 20x human emissions, natural land/sea reservoirs are net sinks, not sources.  Natural processes are on balance REMOVING carbon from the atmosphere, not adding it. Climate scientists appropriately call the difference between the tons of carbon we put into the atmosphere in a given year, and the increase in tons of carbon in the atmosphere during that year, the annual “Net Global Uptake”.  See for example this 2012 Nature paper: Increase in observed net CO2 Uptake by Land and Oceans During the Last 50 Years  Note that despite large uncertainties in individual exchange channels, good data on human emissions and excellent data on atmospheric carbon content allow Net Global Uptake to be determined quite accurately.  You learned the power of conservation laws in elementary physics.  The Nature paper finds that between 1960 and 2010, 350 + 29 PgC were emitted by humans while the atmosphere gained 158 + 2 PgC.   Carbon conservation then implies that 192 + 29 PgC were transferred from the atmosphere into land/sea reservoirs.  Other evidence corroborates how anthropogenic additions to atmospheric carbon have been shared with other reservoirs. With the ph of the oceans measured to be decreasing, there is now more carbon there, as expected by Henry’s Law, given the higher partial pressure of CO2 in the atmosphere. The “greening” of the biosphere gets a lot of press these days; there is now more carbon there too.   The only place where there is less carbon is where the fossil fuel reserves have been depleted. That should tell you the cause of the increase.  Do not make Koutsoyiannis’s mistake of thinking that a larger biomass is somehow causing the atmospheric increase, when the carbon growth in both reservoirs must be coming from elsewhere

To understand errors in your 14C analysis, please see the paper Discussion : Presentation of Atmospheric 14CO2 Data  This paper was motivated by others (Hardy, Salby, Berry, Essenhigh, …) who like you confused Delta 14C, which measures the ratio of C14 to total carbon, with 14C concentration.  Note that many papers you cite were by those authors. Figure 1 shows the radiocarbon Suess effect between 1900 and 1950, before atmospheric nuclear testing complicated things.  The measured decrease of Delta 14CO2 with time shown in 1a was correctly interpreted by Suess as being caused by the addition of 14C free fossil fuel carbon into the atmosphere.  The impact of human activity is clearly visible in the isotopic record, despite your denial of that fact.  But note from 1b that during this period, the atmospheric 14C concentration was rising! How could adding cold (14C free) carbon into the atmosphere raise 14C concentration there?  Disequilibrium isotope fluxes are the answer.  We saw in the previous paragraph, and from your own analysis, that large, nearly balanced two-way exchanges of carbon occur between the atmosphere and land/ocean reservoirs. To the extent that they are balanced, no net carbon is exchanged.  But if the carbon in one of the reservoirs involved in the exchange has a different isotopic composition than the other, the exchanges tend to erase the isotopic difference by simple mixing.  The low Delta 14C reservoir, in this case the atmosphere, then gains 14C from the high Delta 14C reservoir while (with perfect balancing) the total carbon in each reservoir is unchanged.  Note that also this means that the classic Suess effect, the lowering of Delta 14C in the atmosphere and then in tree rings by fossil fuel burning, is much smaller than it would be if the disequilibrium isotope fluxes were not present.  The same analysis applies to carbon 13. Suess understood this dilution of his effect in the 1950’s. Skrable and others did not understand this in 2023. Radioisotopes are often used as “tracers” but thinking of them that way in this situation can be misleading.  The carbon content in the reservoirs is not changing in the same way as its isotopic content.

Now look at Figure 2.  As you note, the “bomb pulse” measured by Delta 14C is well fit by a single time constant exponential after the 1963 Test Ban Treaty.  But we are interested in 14C concentration, and as shown that follows a quite different trajectory, one with a long and rising tail.  The atmospheric thermonuclear tests of course greatly changed the isotope patterns, with the atmospheric Delta 14C about doubling while ocean and biomass Delta 14C’s were temporarily left unchanged.  The atmospheric Delta 14C therefore declined rapidly due to a disequilibrium isotope flux moving 14C to land and ocean reservoirs until isotopic equilibrium was restored.  But then the ongoing addition of cold carbon from fossil fuels again lowered the atmospheric Delta 14C below that of land/sea reservoirs, reversing the direction of the disequilibrium isotope flux.  So now once again the 14C concentration in the atmosphere is increasing by the Seuss effect, as it was before 1950.  This increase was predicted in a 1998 paper by Caldiera et al.  See also Naegler and Levin, 2006 to understand how the bomb carbon is presently distributed.  A significant quantity persist in the atmosphere, despite their apparent full removal according to the Delta 14C plot.  Figure 4 of the Scwartz paper is taken from this reference.

I have had a difficult time getting Harde, Berry and Koutsoyiannis to understand and acknowledge the role of disequilibrium isotope fluxes; perhaps you will learn more quickly.  Ask yourself, if strong bi-directional flows of carbon between reservoirs with different isotopic content exist, how could they not modify that isotopic content?  Another point of confusion in these discussions is semantic. The amount of carbon in the atmosphere over and above pre-industrial levels is sometimes loosely referred to as “human carbon”.  But every knowledgeable scientist knows that while humans may be the cause of the increase, the bi-directional exchanges dilute the amount of carbon in the atmosphere once contained in a fossil fuel.  The introduction of an “Equivalence Principle” was hardly necessary to address a semantics problem, but a tightening up of language would indeed be helpful.

You refer to some of the contradictory literature discussing the difference between “residence time” and “adjustment time”. Rather than wading through all of that, I propose here a simple thought experiment.  The purpose of the thought experiment is to rebut the notion of Koutsoyiannis, echoed by you, that the carbon cycle can be successfully modeled using a single “residence time”.  Let’s suppose we have two reservoirs in contact exchanging carbon via bidirectional fluxes. Call one “atmosphere” and the other “surface oceans”.  Assume “surface oceans” also communicates via much weaker bidirectional fluxes with a third much larger reservoir “deep oceans”.  Suppose all three reservoirs have been in contact for a long time and the carbon content of each has reached an equilibrium based on Henry’s Law, the Revelle factor, etc.  Now start a trickle of CO2 into “atmosphere”.  No longer is the carbon concentration in equilibrium.  By Henry’s Law, etc. net carbon moves from “atmosphere” to “surface oceans”, raising the carbon levels there.  “Surface oceans” carbon levels are now above those in “deep oceans”, and a net flux of carbon moves towards “deep oceans”, though because of the size of “deep oceans” and the smaller fluxes, the carbon concentration there is slow to change.  Now abruptly turn off the trickle of CO2 into “atmosphere”.  Do you think the resulting time profile of carbon in “atmosphere” can be described be a single time constant?  Write out some differential equations if you like, but that should not be necessary.  There will be a downward transient to a new quasi-equilibrium carbon level higher than the pre-trickle level.  The time constant of this transient will not be the same as the longer time constant by which the quasi-equilibrium level decays as carbon is moved to “deep ocean”.  This thought experiment utilizes a simple serial process to generate multiple time constants.  You can hardly argue that the real carbon cycle is less complicated.  You cannot rebut it using the bomb pulse data, because you didn’t understand what the key variable measured.  Instead, a correct bomb pulse analysis confirms the persistence of carbon added to the atmosphere to be at least of order a century.

Finally, let me say that I found your podcast with Tom Nelson arrogant and offensive.  Calling mainstream climate scientists “idiots or liars” says nothing about them but says quite a bit about you and the intended audience of this error filled paper. You are not trying to advance science.  You are merely waging culture wars.  Shame on you both. The climate issue deserves serious discussion.  Reasonable people can disagree on policy details, but unserious and misinformed arguments such as yours interfere with that discussion.

David Andrews,

Retired Physicist, Tech Company Executive, and University Faculty

Cohler and Soon do not allow me to show you their reply to David Andrews comment #1

From: Jonathan Cohler and Willie Soon

To: David Andrews                                                                                                                                        

Date: August 31, 2026

Copies to: Judith Curry, William Happer, Herman Hardy, Demetris Koutsoyannis, Dai Ato, Ed Berry

[They requested their reply be hidden from public view]

Regards,

Jonathan Cohler and Willie Soon

David Andrews comment #2 to Cohler and Soon

Mr. CohIer and Dr. Soon,

I will try to do a better job of quoting directly from your paper before making any additional criticisms, but I firmly disagree that I “engage with none of its content”.  It is certainly true that I did not engage with ALL of its content, thinking 1400 words was quite enough for a start. I will address some additional issues after dealing with your response.

 Yes, although Steve Schwartz did most of the work, I was indeed a coauthor on the Radiocarbon paper. I don’t see how calling attention to confusion among non-experts between isotope ratios and concentration, whose time profiles over the last century are quite different, is anything but “neutral”.  Do you have a specific example of bias that you see in that paper?  Do you see any errors in the paper? You seem in your comments to want to ignore the fact that Delta 14C and 14C concentration are different.

I have been through the 117 references in “Dominance of Natural Atmospheric CO2 Dynamics” multiple times, and cannot find Ballantyne, Cawley, Seuss or any of the other papers you said you cited.  Perhaps you have some other paper that does cite them.

On the science items you identify:

1.       The “mass-balance” (I prefer “carbon conservation”) argument is indeed so simple, that it can catch one by surprise.  You are not the first to think that it is an example of circular reasoning.  I agree 100% with your statement “The question is empirical, not definitional.”  Nothing in the definitions forces Net Global Uptake to be positive, but that is what is consistently MEASURED in the modern era. Had Net Global Uptake been MEASURED to be negative (human emissions < atmospheric rise ) then obviously natural processes would be contributing. It is data, not circular reasoning, that leads to a positive Net Global Uptake and the conclusion of human responsibility for atmospheric CO2 increases.  

2.        On isotope data:  The Radiocarbon paper shows clear empirical evidence of the effect of fossil fuels on atmospheric 14C levels through the 14C Suess effect, both as measured by Delta 14C and by concentration.  The data is cleanest before atmospheric nuclear testing. You have not engaged.  So I will not engage on 13C data. I could cite the rebuttals to the Koutsoyiannis analysis of 13C data, but you can look them up yourself.

3.       On Delta 14C vs 14C concentration.  Why do you think Delta 14C measures the amount of 14C in the atmosphere better than 14C concentration?

4.       I am skipping discussion of the Bern Model, but you still seem to be fixated on isotope ratios instead of concentration.

5.       On disequilibrium isotope fluxes: I am glad you understand them, but I can’t decipher the rest of your comment.  Perhaps Grok wrote it.  See 7 below.

6.       On Naegler and Levin’s analysis.  No, this is not about model vs data.  This is about using the correct variable to describe the amount of 14C in the atmosphere.  You don’t. Delta 14C has returned to its “pre-bomb” level.  14C concentration has not.  The radiocarbon literature is mostly about Delta 14C, because that is what is directly measured. But 14C concentration is also a measured quantity when carbon concentration is known, which is usually the case.  There are no models involved in converting  measured Delta 14C to 14C concentration. But you are correct that there are models involved in separating bomb 14C from cosmic ray produced 14C.  I think Naegler and Levin got it right.  I know of no competing analyses.

7.       On my thought experiment:  I regret you did not understand it and doubled down on the bomb pulse timescale as measured by Delta 14C.  Yes, with initially balanced isotopic content, the thought experiment would treat 14C and 12C identically.  But don’t you see the difference between the restoration of isotopic balance between the atmosphere and surface ocean from disequilibrium isotope fluxes after the Test Ban Treaty, and the bulk movement of carbon from the atmosphere to the deep oceans? The former doesn’t involve Henry’s Law, Revelle factors, or mixing mechanisms between the surface and the deep ocean.  The latter does.  How on earth can you think the time constants are the same? 

I will refrain from getting too much into the polemics, but I will a little.  Like other skeptics, you complain that I am “someone who has decided in advance what conclusions may be discussed”.  I did not and could not stop you from writing what you did.  Do you think you should be immune from the criticism that similar papers experienced?  I presume the paper was submitted to some journal.  Would you care to share what their reviewers said about it?

I earlier promised to address a couple more specifics of your paper:

1.       You refer to “the absence of detectable perturbations during emission reductions (e.g., COVID-19 lockdowns)”. This is a familiar internet argument. Yes the ~6 % reduction in emissions for two quarters in 2020 is not detectable.  That is because the expected effect would be a half-year 6% change in the slope of the increase of atmospheric CO2 concentration and that is overwhelmed by seasonal variations and noise in the data.  I give you credit (perhaps wrongly) for knowing no dip was expected, but perhaps you hadn’t worked that out.

2.       You assert “High-resolution local proxy records and direct observations from every major archive examined demonstrate unidirectional temperature → CO₂ causality on all timescales, with no empirical evidence anywhere showing the reverse.”    

If you read the Ballantyne paper, you will see that in the 1990’s, following the Pinatubo volcanic eruption, Net Global Uptake was a little above the trend curve.  Presumably the balance in the temporarily cooler oceans between outgassing and “in-gassing” shifted towards in-gassing.  Yes, temperature effects CO2 levels.  The historical pattern of temperature → CO₂ on geological timescales is not disputed.  However the modern era increase if CO2, of course considered anthropogenic by most, has NOT been shown to follow temperature, despite the claims in your reference 117 (Koutsoyiannis “Hens and Eggs” paper. If you have any other papers claiming this for the modern era, I would like a reference.)  With apologies to Demetris who has heard it all before, here are the problems with his paper: a.) for no good reason he smooths the data prior to the analysis.  I take this as an attempt to hide the noise in the data. b.) He claims the statistical inference that temperature leads CO2 in the modern era BUT DOES NOT ASSIGN A CONFIDENCE LEVEL TO IT! When I asked, he basically said “do it yourself.” I actually would be quite interested in an update of his analysis, without the smoothing, and including the very warm years from 2020  through 2025, which followed the earlier anthropogenic CO2 increase. Koutsoyiannis is copied on this and can respond if he wishes.

I’ll leave it at that.

Dave

Cohler and Soon do not allow me to show you their reply to David Andrews comment #2

From: Jonathan Cohler and Willie Soon

To: David Andrews                                                                                                                                        

Date: August 31, 2026

Copies to: Judith Curry, William Happer, Herman Hardy, Demetris Koutsoyannis, Dai Ato, Ed Berry

[They requested their reply be hidden from public view]

Jonathan Cohler and Willie Soon

Note: This is the first time since I started this website in 2009 that an author has asked me not to show the public their replies to comments.

Recall that the authors of the CO2 Coalition’s document extensively defended their EMB paper.

I have debated these issues with David Andrews here. But this post is about Cohler and Soon’s paper. So, I won’t debate Andrews in this post. That’s Cohler and Soon’s job.