Our various graphics and model-observation comparisons have been updated with 2025 data. There are a few version updates that make some difference (particularly in sea ice extent), but the basic story is similar to last year. Feel free to comment and/or suggest changes for the future, including potential new comparisons.
Reader Interactions
189 Responses to "2025 Updates"
Comment Policy:Please note that if your comment repeats a point you have already made, or is abusive, or is the nth comment you have posted in a very short amount of time, please reflect on the whether you are using your time online to maximum efficiency. Thanks.
Would you be willing to assess an IPCC 1992 warming projection? If you are interested, then the information for the projection is below:
And here’s the IPCC 1990 projection:
Asking Gemini “compare high-CO2 sensitive gcm in CMIP5 and CMIP6”
Gives the following result:
“””
CMIP6 GCMs exhibit higher Equilibrium Climate Sensitivity (ECS) (\(1.8\text{–}5.6^{\circ }C\)) compared to CMIP5 (\(2.1\text{–}4.7^{\circ }C\)), largely driven by stronger positive low-cloud feedbacks and improved cloud physics. The average ECS increased from 3.3K to 3.9K, with many CMIP6 models exceeding the 4.5K threshold, leading to faster warming projections. Key Comparisons Between CMIP5 and CMIP6 High-Sensitivity GCMs: Climate Sensitivity (ECS): CMIP6 models are noticeably more sensitive to \(CO_{2}\). Many more CMIP6 models fall into the high-sensitivity range, with some exceeding 4.5°C or even 5°C, unlike the more restricted range in CMIP5.Cloud Feedbacks: The higher sensitivity in CMIP6 is driven by stronger positive feedbacks, specifically the reduced low-level cloud cover in the extratropics and weaker increases in low cloud water content.Aerosol Representation: CMIP6 models often include more sophisticated representations of aerosols, which, when combined with altered cloud physics, results in higher net sensitivity.Forcing and Warming: CMIP6 shows a stronger surface temperature response for a given \(CO_{2}\) increase, leading to a projected 0.4°C higher warming this century compared to similar CMIP5 scenarios.Model Performance: Despite higher sensitivity, CMIP6 models generally show improved performance in simulating climate processes. The increased sensitivity in CMIP6 is, for many models, tied to enhanced shortwave cloud radiative effects (CRE) that are only partially offset by other feedback mechanisms.
“””
I would also add the improved resolution as an important improvement.
In short, the newer models produce significantly different results as it is normal for a progressing field.
Going backwards in time using models lacking more and more in physics and resolution needs to address these effect of those model defects!
Pretending an obsolete, incomplete and low-resolution model could contribute anything meaningful seems very unscientific.
All you really need to know is that a modern model using similar climate parameters like for example a high-CO2 sensitivity will produce significantly different results, old models are obsolete as they are lacking and produce non-skillful results!
This claimed skillfullness of those wrong models is touted not only by AS and old publications (which need to be withdrawn or corrected!) but can also be found in the links in above article!
Instead of the results of old models lacking significantly in resolution and physics (both very clearly shown to be highly relevant, see the Gemini text above for example), you might as well use scribblings of four year old, the scientific relevance is about the same.
When lacking model results match real world measurements that is the result of careful tuning representing the belief of the tuner not skill
Your usual evidence-free gibberish was already debunked. But you never accept evidence, as per denialism.
https://www.realclimate.org/index.php/archives/2025/11/raising-climate-literacy/#comment-843361
YK: “Pretending an obsolete, incomplete and low-resolution model could contribute anything meaningful seems very unscientific.”
That is wrong. By analogy its like saying windows 10 was useless, because its been superseded by the Windows 11. Windows 10 is obviously not useless. Not even windows 95 was useless. Saying its useless is just childish petulance.
YK: “When lacking model results match real world measurements that is the result of careful tuning representing the belief of the tuner not skill”
Yes it’s skill. The tuner selects numbers based on his understanding of the most plausible or likely number. This obviously requires skill. Its not decided on a coin toss.. And the skill level of the model isn’t decided on the tuning. It’s decided on the level of quality of the underlying physics and on accuracy of predictions. Old models frequently did OK in that regard. As has been explained about ten million times already.
Spoken like a man who doesn’t understand how scientific modeling works.
HINT: Climate models do a whole lot more than modeling global temperatures. Tuning does not mean fudging the numbers. One reason past models reproducepast temperatures is that the physical needed to do so is well established and not at all controversial. The improved skill of newer models is needed for other predictions.
RL: Spoken like a man who doesn’t understand
Well, you should read once more the forum rules against personal attacks, maybe we can focus on the science? What you and the other posters don’t seem to understand is the importance of manners in a scientific discussion. Even wasting time by being repetitive and not on target can reflect badly on the whole community on this particular webpage! There is the very loud silence on this topic by any scientist in this field for all to see.
N: And the skill level of the model isn’t decided on the tuning.
I strongly agree! Tuning can ensure that the calculated results match real world trends even for bad/lacking models, so that match is NOT a test of skill.
RL: Climate models do a whole lot more than modeling global temperatures.
Yes, and the newer ones tend to do everything a little better! That was my point against using old models!
RL: Tuning does not mean fudging the numbers. One reason past models reproducepast temperatures is that the physical needed to do so is well established and not at all controversial.
If you mean by that, that for example the high-CO2-sensitive CMIP6 models show significant different results than their older counterparts thanks in part to their improvement of aerosol physics (like written in the Google result I cited above), we can all agree on that, it seems hardly controversial.
Old models and old publications describing their results lacking this physics improve ment, just as I wrote. That too does not seem controversial (except maybe AS, you tell him!)
RL: The improved skill of newer models is needed for other predictions.
Now you write about skill and improved skill. What is needed is a working test if there was any skill.
The question if a model result matches a real world trends does not answer anything about skill like Nigelj correctly observed tuning is not skill. With enough patience and filtering you can match scribblings of four year olds to a real world trend.
You did see that I just gave an example (high-CO2-sensitive models) where improved physics and resolution lead to unrealistic trends in CMIP6, but a similar parametrization gave realistic results in CMIP5?
This means that according to our newest and best knowledge
– high-CO2 sensitive models produce unrealistic results
– there are examples for results of old models deviating significantly from improved models, those results and the publications they are need urgently examined (otherwise AS keeps being very confused), corrected or withdrawn.
– the test for skillfullness did not work as we currently know/believe that our best (in resolution and physics) high-CO2-sensitive models produce unrealistic results, but were deemed skillful in the past, we now know they are not.
– nothing either of you wrote addresses this and seems concerning that the community here really does not have an answer!
Yebo: “What you and the other posters don’t seem to understand is the importance of manners in a scientific discussion.”
Oh, that’s hilarious! Have you ever even met a scientist? We tend not to be a polite group, especially when confronting poseurs!
The question is not which model is “right” or even which model is best, but rather which models are useful in understanding our changing climate. The fact that older models reproduce observed temperature increase well has little to do with “tuning” (which means something totally different from what you are imputing), and more to do with the fact that greenhouse warming is well understood and does not require a highly skilled model to reproduce.
Like it or not, some science is sufficiently well known that questioning it isn’t a profitable use of time.
Ray L: I looked up Yebok Kando: some gaming & clickbaitery but not fair to assign that to him, given how little there was. I wonder if he’s one of those computer whizz designers whose arrogance exceeds their sense of perspective. [see Adam Becker & Neil deGrasse Tyson who are depressingly amusing on ego/scifi etc.]
Re: “Tuning can ensure that the calculated results match real world trends even for bad/lacking models, so that match is NOT a test of skill.“
Same old disinformation from you.
Climate models accurately/skillfully projected global mean surface temperature (GMST) trends and iTCR, i.e. the ratio of GMST trends vs. forcing. This accuracy/skill is not explained by tuning, as noted by climate scientists like Dr. Gavin Schmidt and Dr. Zeke Hausfather:
Yet you baselessly accuse them of fraud:
Ray, I completely agree that “climate models do a whole lot more than modeling global temperatures”.
So why the F is almost the entire bandwidth of RC taken up with .absurd “debates” about the significance of predicting what hour of what day of what year we hit 1.5 instead of 1.4 or 1.6? Why keep answering the trolls on logically fallacious arguments about GMST?
How about getting some guest posts from specialists on their work in looking at a variety of system elements/phenomena, and what the models may or may not be showing. You know, science and physics instead of “he said they said nyah nyah nyah”?
Let’s actually talk about the models in more detail and how they project the possible evolution of the system as a whole. Again… science and physics.
Would you be willing to assess an IPCC 1992 warming projection? If you are interested, then the information for the projection is below:
Above the second graph (CO2 conc): “How much of the current level of CO2 in the atmosphere is due to human emission? (over one third!)”
I’d suggest reframing:
“How much the natural level of CO2 in the atmosphere increased due to human emission: by over half!”
The same data, yet leaves different impression.
“This content is protected against AI scraping.”
Why do we want to prevent AI from having all this correct data?
robots.txt doesn’t reject AI bots, and Claude can fetch this page and the graphs behind the links just fine. And then it says
One of the commenters (Martin Smith) actually asked the same question you did: “Why do we want to prevent AI from having all this correct data?” — which is a fair point given that this is exactly the kind of high-quality scientific content you’d want in training data.
… and I agree with both.
On a par with people who post things like “I do not give permission for facebook to keep my photos”
Makes people feel good, doesn’t do anything
Minnesota ICE-OUT
https://geoenergymath.com/2026/01/27/minnesota-ice-out-update/
Of course it will be Minnesota that does things right. The DNR for the state actually has the foresight to store all the ice-out dates on a server in easily accessible JSON form. What I did was query for a list of recorded lakes for every year and then iterate though all the years to create a comprehensive local array that I could then analyze further and make plots, etc. Important to note that Minnesota is a big enough state that one can detect that the mean ice-out date depends on the latitude. That’s why I scatter plot by organizing the data to 1 degree categories, 43, 44., 45, 46, 47 degrees as shown:
https://geoenergymath.com/2026/01/27/minnesota-ice-out-update/
One thing I haven’t done is to create a linear normalization that would correct for the precise latitude at which the lake is located at — then I could plot all the lakes on the same plot.
No one does this kind of stuff, and certainly the US federal government won’t be funding it anytime soon.
Martin Smith: “ Why do we want to prevent AI from having all this correct data?”
I agree.
So instead of preventing AI from accessing the graphs and opening articles, I would move the prevention to the comment section only – both deniers (Ken Towe, Keith Woollard, the Enron guy, KiA, etc) and doomer (“Kandu/Kando”, “Data” and its dozens of previous handles) often rush into the newly open threads. Why? To plant their flag, to try to discredit the opening article and/or our hosts posting them, and if they can’t, hijack the discussion away the new information and onto a tangent with their strawman and/or re-litigate the long-ago explained to them topics.
AI can’t distinguish right from wrong on its merit – so it can either
– try to incorporate both sides, as a case of “science not settled”), treating KiA, Keith Woolard or the Enron guy as equally credible as Gavin, John Pollock or MA Rodger,
– or worse still, judge the value of the two position on number of words written in support of each in this option even a single AI-powered bot, or a AI-assisted troll can shift the balance in favour of their objective – see “Data” posting 102 posts in the first 23 days of the month, and most them – not exactly one-line zingers (its 28 Jan 12:35 AM post was well past 1400 words).
And an incestious feedback loop is created – when the mass-production of an AI-bot or AI-assisted troll swamps the training sets of the said AI. This effect is further amplified, if the goal of the AI is not to provide the best information, but to promote the ideological views/interests of the tech oligarchs and the politicians they support.
>> (“Kandu/Kando”, “Data” and its dozens of previous handles) often rush into the newly open threads. Why?
It is Kando. Yebok Kando.
You too commit a personal attacks against me and data.
Unless you can prove your insinuation against me (which you can’t, because I happen to know you are wrong and I am the expert about me), please withdraw it.
As your accusation I would “hijack the discussion away the new information”, you should probably read once more the old threads, actually it happens again right now above. I feel like I am making a justified comment (here for example about the validity to unreflective carry on old model results given the stark differences between CMIP6 vs older in regards to high CO2-sensitivity, the same in the thread about scientific communication).
Try to wear my shoes, would you honestly consider AS’s uh … contributions as a satisfying result?
If you can’t find a simple satisfying explanation on my questions about model skill, you are not the only one!
Consider considering my contributions important and justified until proven otherwise.
Your assumptions about my motivations only reveal too much about yours.
Re: “Try to wear my shoes, would you honestly consider AS’s uh … contributions as a satisfying result?“
^^ Evidence-free trolling. Climate models accurately/skillfully projected global mean surface temperature (GMST) trends and iTCR, i.e. the ratio of GMST trends vs. forcing. That’s the case regardless of Yebo Kando falsely accusing Dr. Gavin Schmidt of fraud for saying this.
I see you are still stuck on auto repeat still not saying anything relevant to the topic. I think most people (me include) knew about those post a long time before you first mentioned them, IMHO it is long time since that you stop wasting space and time here, condense what you are trying to say (if there really is anything to your posts beyond that idiotic autorepeat it is not visible, just use your own words for once and tell us, what you think how the cmip5 and 6 differences affect older analyses)
You even reposted my statement about the required treatment of all known errors and uncertainties (in this context the lacking resolution, aerosol physics improvements and so on not present in the cmip5 and older models), I don’t have to repeat what the science requires.
Your posts not answer, prove or debunk anything! You just found and autorepeat a few citations which present the unreflected results of old models lacking the discussion of uncertainties steming from lacking the improvements made for cmip6 models.
It is really that simple: cmip6 models are better than older models, so the results you autorepeat are outdated until reanalyzed.
Just like I wrote take the models from your citation and show what changes for the result you get when you improve resolution and physics. Your seem to falsely imply that these results are “good already* and won’t be affected by those improvements, but of course we already know better for high-CO2 sensitive models proving you wrong!
And before you ask, I stand by mh statement that using wrong models while knowing better is fraud!
Please from now on include this very post in you endless repetition series!
Re: “You even reposted my statement about the required treatment of all known errors and uncertainties (in this context the lacking resolution, aerosol physics improvements and so on not present in the cmip5 and older models), I don’t have to repeat what the science requires.“
Again, just your usual evidence-free gibberish was already debunked. But you never accept evidence, as per disingenuous and persistent denialism.
– https://www.realclimate.org/index.php/archives/2025/11/raising-climate-literacy/#comment-841769
– https://www.realclimate.org/index.php/archives/2025/11/raising-climate-literacy/#comment-843361
Yebo Kando,
The subject of your interest seems to be why did the old CMPI5 models with unresolved micro physics of clouds and which generated medium climate sensitivity make quite good predictions, while the new CMPI6 models with better resolved cloud micro physics and which generate high climate sensitivity are not predicting warming rate very well? But didn’t AS already point out the old models made good predictions because the important parts of the physics were correct? And although the cloud microphysics in thge old models wasn’t well understood and represented in the equations, any values they inserted for this (tuning) might have been very good assumptions / inferences.
That leaves the question of why the newer high climate sensitivity models are making bad predictions despite better resolved cloud micro physics. There are some obvious possibilities: 1) despite better equations for the cloud microphysics, maybe something else important about the new model is still wrong where the older models were correct, or 2) maybe the new high climate sensitivity models are only getting short term warming predictions wrong, and will be vindicated over longer term timeframes. I’m actually a bit curious about all this myself, but I don’t have remotely enough expertise to figure it out. This is why I haven’t said anything about it. Perhaps one of the climate scientists can explain what’s going on.
>> I’m actually a bit curious about all this myself,
Well, maybe we are not too far apart after all.
What you seem to describe is that the new results caused new questions for you (a point AS’s posts do not seem to grasp in the slightest).
I write about these new results (caused by better physics representation and resolution) causing additional uncertainties for old works.
Unless you can show how the deviations of the high-CO2 sensitive CMIP6 models are wrong and corresponding CMIP5 and older are not (and remember this particular thread is about 90ties models, so it very seems unlikely to me that they had a better physics representation in their models back then than now and we know the resolution was low), these large deviations (unmentioned and unexplained in old publications – as AS’ses posts demonstrate several times a day) between model generations blow up the uncertainties – it’s basically a mathematical equivalent of you having new questions when significantly different results come out:
Once an old model is shown to be potentially lacking, it’s results become inevitable less reliable.
If you are not convinced about the improved physics, just focus on the effects of the improves resolution. There is no way for you to wiggle out of that one!
All I wrote is that the old publications did not and could not have included these new findings (and repeating them endlessly unreflected seems a special kind of moronic!)
And especially to AS’ses initial questions, this is true for all accumulated model improvements since the 90ties leaving the old results meaningless – at least until such an update is performed.
Nigel, another possibility is that the CMIP5 models had a problem which the cloud microphysics was hiding–as if one process gave the wrong answer by plus 5%, while the other got it wrong by minus 5%.. The CMIP6 models, with improved cloud microphysics, would then bring out the other problem and have worse performance (0% – 5% or 0% + 5%). The trick now is to find out where the other problem lies. Correct that, and CMIP6 should perform better than the earlier versions.
Yebo Kando @1 Feb 2026 at 9:24 PM
YK: “Once an old model is shown to be potentially lacking, it’s results become inevitable less reliable.”
Not necessarily. If an old model made good predictions across several areas its very unlikely it would suddenly stop doing that. The old models might have just had very good tuning to approximate the unknown areas of the physics. Now the physics is better understood at a fundamental level and should clearly be used. But that doesn’t change that the tuning may have been good and skillfully done in the old models.
And nobody is saying we should continue to use old models. Im not sure where you get that idea from.
YK: “If you are not convinced about the improved physics, just focus on the effects of the improves resolution. There is no way for you to wiggle out of that one!”
Im not trying to wriggle out of anything. Remember many of the CMPI6 models are making good predictions especially those that generate medium climate sensitivity, those being near the mean of the model ensemble. Its just the very high and very low climate sensitivity models that seem to lack good predictive ability. But this is not entirely surprising because its generally thought in the scientific community that climate sensitivity is most likely in the medium range,
Like I said I have precisely zero expertise in climate modelling, and am just a lay person, but I do know these models are based on multiple different equations integrated together and even if the physics in the high climate sensitivity CMPI6 models is 99% right if one small thing is wrong and important the model may not make good predictions. It just suggests to me something is wrong in the latest high climate sensitivity models. But you are not addressing my suggestion explicitly. You are talking around it. And you aren’t suggesting a better alternative.
YK :”All I wrote is that the old publications did not and could not have included these new findings (and repeating them endlessly unreflected seems a special kind of moronic!)”
I’m not sure what that means. However maybe its related to where AS has repeated evidence that the old models had skill. You made a claim the old models lacked any skill. Both myself and AS thought that was unjustified. Naturally people will react against inaccurate claims, even if its only part of what someone is posting. And it does relate to your other claims.
And you have been given a mountain of evidence why the old models have skill. You asked for a way of objectively assessing skill. AS has several times provided you links describing how levels of skill are determined so why keep asking? We get frustrated when people do that. It makes us suspicious that people are just trolling, although I sense you may not be doing that and may be genuinely curious about the issues at least to some extent.
However the modelling issues are way outside my areas of work expertise so you are better off asking the experts and phrasing your questions much better because its hard knowing exactly what you are getting at.
YK: “And especially to AS’ses initial questions, this is true for all accumulated model improvements since the 90ties leaving the old results meaningless – at least until such an update is performed.”
The results aren’t meaningless. For reasons explained already. The physics does not have to be perfect to be useful. Einsteins equations did not render Newtonian physics useless. And surely CMPI6 is in fact the update you are looking for? Staring you in the face :)
Nigelj, they’re trolling by, for example, willfully ignoring the evidence they’ve been cited on this:
‘Improving cloud physics for CMIP6 models reduces their sensitivity down to what’s shown for older models, with accurate projected global surface warming and iTCR.’
Atomsk’s Sanakan @2 Feb 2026 at 4:52 PM
AS: ‘Improving cloud physics for CMIP6 models reduces their sensitivity down to what’s shown for older models, with accurate projected global surface warming and iTCR.’
My understanding is that is true for most CMIP6 models, but aren’t the high climate sensitivity CMIP6 models failing to project surface warming well? Didn’t you refer to that over the issue of the German geophysical society? And I think that is the models YK is referring to. Although Im not defending his views and I’m in disagreement with most of his views.
>> You made a claim the old models lacked any skill. Both myself and AS thought that was unjustified
The claim that the best high-CO2 sensitivity models (CMIP6) produce unrealistic results ist not from me!
And I fully agree that when you reduce the resolution and remove modern physics you then can tune models with high-CO2 sensitivity to match real world measurements.
But that just means that the old findings need to be worked over to reflect our improved knowledge and are not valid until it is done!
And your wiggling that if we tune the low resolution models without proper aerosol physics maybe we do the right thing ignores that we know better!
Use a better resolution and correct the physics and the high-CO2 sensitive models always fail, we therefore know (based on our current knowledge) that high-CO2 sensitive models have no skill for predicting global warming trends. Your idea to blur resolution and remove physics until you are able to tune it to real world trends immediately causes the questions of the results for different tunings as a range of uncertainty. And you are aware that high-CO2 sensitivity CMIP6 results are far outside the confidence intervals of the studies of AS’ses autorepeat.
And that the old publications showed credible warming trends for high-CO2 sensitive models, while the best available models do not, means that their test for skill is questionable!
You cannot have both high-CO2 sensitive models matching and not matching real world trends at the same time unless you allow for huge uncertainties!
The latest results prove that old models got their trends wrong or they underestimate the errors, either way we know their results as presented back then are not correct!
When you formulate your and AS’ses thoughts, you seem to omitt this very basic fact!
I am not doubting that a bad model or scribblings of four-year old can be tuned to match real world trends, there is just little or no predictive skill involved!
Bad models cannot produce good results! There is no magic happening here, just high-CO2 sensitive models failing the reality test in modern models and old models lacking physics and resolution while publications based on them claiming skill just indicates a flawed testing.
Did you ever answer my question if the scribblings of a child have skill if they only look like a real world trend? (Btw my answer is “Of course not, you need to access the uncertainty correctly!”)
>> Both myself and AS thought that was unjustified.
I can condense my latest post:
Your or AS’ses thoughts are not important, but what you can prove!
There is ample evidence that high-CO2-sensitive CMIP6 models produce unrealistic warming trends and they are therefore not skillful.
Unless you can prove that the transition towards and older configuration (blurring resolution and removing physics) somehow improves them (which would be quite a trick IMHO), the counter thesis that such steps just blow up the uncertainties in a way not disclosed in old publications remains valid. This means all older model results no matter what CO2-sensitivity are potentially affected until you can prove they are not!
It’s a statement about high sensitivity models. Improving cloud physics for those models reduces their sensitivity down to what’s shown for older models, with accurate projected global surface warming and iTCR.
That’s the point of the evidence that the troll willfully ignores, such as Myers 2021 and Zhu 2022.
Re: “The claim that the best high-CO2 sensitivity models (CMIP6) produce unrealistic results ist not from me!“
It’s from you, hence why you cite no published evidence and willfully ignore any published evidence cited that shows you’re wrong. To detail this again in steps for those who come along later:
1) Different CMIP6 models differ in climate sensitivity. A minority of the CMIP6 models are high-sensitivity, i.e. have an equilibrium climate sensitivity greater than 5°C or TCR greater than 2.5°C.
2) The high-sensitivity CMIP6 models conflict with the evidence on climate sensitivity and TCR.
3) Improving cloud physics for high-sensitivity CMIP6 models reduces their sensitivity down to what’s shown for pre-CMIP6 model average, with accurate projected global surface warming and iTCR.
4) The pre-CMIP6 model average accurately/skillfully projected global mean surface temperature (GMST) trends and iTCR, i.e. the ratio of GMST trends vs. forcing.
5) This accuracy/skill of the pre-CMIP6 model average is not explained by tuning.
Yebo Kando @ 2 Feb 2026 at 11:20 PM and Yebo Kando @2 Feb 2026 at 11:33 PM
Not wasting time on all your points. They are mostly repetitive and have been answered already and with citations or they are incomprehensible or strawman statements or non sequiturs. Never do you quote the exact words I use. I will only respond to two points which sum up the issues:
YK: And your wiggling that if we tune the low resolution models without proper aerosol physics maybe we do the right thing ignores that we know better!
Nigel; No. I have never suggested tuning is preferable to using the proper physics. Your comments suggesting I have are a strawman. Tuning is only used when scientist dont know the full physics of something like cloud microphysics and so they substitute the best evidence based numerical approximation as they did in the old models. This is why the old models still made good predictions and thus have skill. The exact skill level varies from model to model. Obviously knowing the physics is always preferable and is why the new models use better physics. I don’t see the point in rewriting the old models to include this better physics. I thought the whole point is to do that with the new models. and use those (CMIP6)
YK: Use a better resolution and correct the physics and the high-CO2 sensitive models always fail, we therefore know (based on our current knowledge) that high-CO2 sensitive models have no skill for predicting global warming trends
Nigel: I have not claimed the new high sensitivity models are skilful. I only said that that the medium sensitivity models have skill. So you are just making another strawman argument. So try saying something that isn’t a strawman statement or gibberish and I might respond.
>> It’s a statement about high sensitivity models. Improving cloud physics for those models reduces their sensitivity down to what’s shown for older models, with accurate projected global surface warming and iTCR.
That’s the point of the evidence that the troll willfully ignores, such as Myers 2021 and Zhu 2022.
Just a clarification please:
It seems like you are saying that the best newest models disqualify high CO2-sensitivity scenarios? Is that the correct reading of your post?
(Your wording is a bit funny/incorrect as high-CO2 sensitive CMIP6 models do exist and their results are published without any “reduction”, they just don’t overlap with CMIP5 or real world data)
If that is what you are saying, there is still be the problem that older publications need to be corrected for claiming skill to high-CO2 sensitive scenarios, which you seem to condemn with your statement!?
It also means that climate alarmism is over for you as you turned “lukewarming” aka generally acknowledging a potential raise in temperature and sea level, but the lack of a high-CO2 sensitivity also producing no dramatic change the status quo for the foreseeable future.
Btw I believe you could easily train a chimpanzee to autorepeat posts with old publications and throw in some verbal feces for good measure.
That creature would of course be unable to understand how newer findings on high CO2-sensitivity scenarios might invalidate results those old publications,so they might get stuck in a strange loop…
But I do admire your attempt to post something different even if it seems to be in a sharp contrast of so many of your other posts with very limited value.
So now you are writing that there might be a problem with high-CO2 sensitive models, what z novel concept – basically what I was writing all along!
Lukewarmism entails climate sensitivity below 2K (ex: Matt Ridley, Patrick Michaels), which is not what’s being claimed here. So you’re disinforming when you claim this is lukewarmism. Moreover, the points on Myers 2021 and Zhu 2022 have been cited and explained to you several times across many posts. Thus, you claiming this is new is either you pretending and/or you showing you neither read nor understood the published evidence cited to you.
I get that you would falsely compare citing published evidence to being a chimpanzee. After all, your denialism goes against the published evidence. So you cite no published evidence and invent excuses to dodge the published evidence cited to you. A better comparison to a ‘chimpanzee’ is repeating what an AI said, with no cited evidence. That’s what you did:
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844538
Anyway, I had 5 numbered claims listed above, each supported by published evidence. These together entail that the pre-CMIP6 model average skillfully/accurately projected GMST trends and iTCR, that this isn’t explained by tuning, and that citing CMIP6 modeling does nothing to change this. Explain which, if any, of those 5 claims you disagree with and cite evidence justifying your disagreement. No, mindlessly spamming AI drivel does not count as citing evidence. Failure to meet this request further confirms a commitment to evidence-free denialism.
AS’ses posts
“” I had 5 numbered claims listed above, each supported by published evidence.””
Well, this is so strange, I don’t know what to write beside read again your own citation! It truly weird and easy for anyone to see that each of them very clearly makes the point that high-CO2-sensitive cmip6 models do not match observations.
A point very well observed by me and everyone else in this thread.
Your posts are therefore once more pointless.
You seem to have fooled Nigelj into believing that repeating trivialities seems to somehow answer my question.
After countless posts we all still know that
– high-CO2 sensitive CMIP6 models produce unrealistic results
– high-CO2 sensitive CMIP5 and older models do not.
Your endless posts have neither changed that not presented any form of insight or debunk.
They were a complete waste of time and space.
So after our best and most modern knowledge high-CO2 sensitive are not skillful for this real world parameter.
And to make matters worse publications using older models did not see this! Just as I always wrote!
Everyone here can definitely and conclusively answer the questions:
Using the best available knowledge and model resolution do high-CO2 sensitive models produce unrealistic warming?
(Yes!)
Did publications of cmip5 and older models represent this fact correctly or do they need to be reworked based on or improved knowledge?
(No!)
Both are easy yes or no questions and do not require any citations or wiggling, we all saw that here way too often already.
And I still can say that
– claiming skill for wrong/incomplete models also includes the possible evaluation gor scribbles of children (if they only seem to match a real world trend they fully pass that test!)
And I strongly repeat that:
– “To claim that a wrong model would be skillful requires a mathematical treatment of known errors that has not happened here, so it is fraud.”
(A very trivial statement any sane person should agree on and is not limited to a single group or just climate science!)
You can try to insult me (against forum rules), but any claim that these statements are incorrect or falsified (repeatedly or not) is wrong, especially while you are citing publications like just here in the last two longer posts, which only supporting my questions!
Thanks for confirming my prediction regarding evidence-free denialism. Going to ignore the rest of what you say until you meet the request.
Which, if any, of those 5 evidence-based claims you disagree with?
If you disagree with a claim, then cite evidence justifying your disagreement.
I predict that the query will again be dodged and/or no peer-reviewed evidence will be cited to justify disagreement.
Hmm.. could pleae someone elseexplain to AS what he is doing wrong in this debate?
He made a post, someone (me) had an answer.. what must be the next step?
Hint: It is not just coping our lastest post once more, if you do that, you can find some unanswered comments just above that latest waste of space and time, like it is the case here!
For a debate you would need to address the issue and I just told you were you could find some relevant points..
(Moronic repetition also does not debunk anything, it is just junk)
Thanks for confirming my prediction about you.
Again:
Which, if any, of those 5 evidence-based claims you disagree with?
If you disagree with a claim, then cite evidence justifying your disagreement.
Yebo Kando. “It is Kando. Yebok Kando”
tell this to the guy posts here as “Yebo”.
And who was this
“Yebo Kandu says 3 Nov 2025 at 12:23 PM” or “Yebo Kandu says 28 Nov 2025 at 5:23 PM”, “Yebo Kandu says 3 Dec 2025 at 7:00 PM”, “Yebo Kandu says 10 Dec 2025 at 11:08 AM” etc?
You again forgot what you are supposed to be ??? You learned nothing from your “Darma” / “Darmah”/ “Dharma” debacle?
Now go and bring me some shrubbery. Or I say “Ni!”. “Ni! Ni! Ni! Ni!”
What strange posts from you
>> tell this to the guy posts here as “Yebo”.
And so on..
While I did not look up all these references you made, it does sound that was my all from me ..
Basically any time the post starts with “Yebo Kando says”, however just one post up you were falsely kicking me with other posters (which wrote posts defending my points last time AS’s post degraded too much into meaningless repetitions.. his last post here is a good example, devoid any relevant argument or information and repeated endlessly)
And just to avoid any confusion I made a typo in my first name, it is just Yebo without the *k”.
Beside that very trivial and also fairly irrelevant information however, your posts here seem very bizarre!
In the end who cares who I am, just focus on the science. Because that is what this forum is about.
>> Or I say “Ni!”. “Ni! Ni! Ni! Ni!”
That, however, is not, but trolling.
Nobody wants or needs trolling.
Doomer troll, currently performing as: “Yebo Kando”.
“ I did not look up all these references you made[…]. Basically any time the post starts with “Yebo Kando says”, however just one post up you were falsely kicking me
– The references you didn’t bother to look up were not “just one post” – but whole series of posts, so it’s not a one-time typo.
– And since you HAVE posted MULTIPLE posts as “Yebo Kandu” – how is it “falsely” to refer to you as “Yebo Kando/Kandu”?
YK: I made a typo in my first name, it is just Yebo without the *k”
Hmm. what are the odds that out of your 100s(?), 1000s(?) posts here, you misspell your name IN THE VERY POST in which you ,,,lecture others about …. the correct spelling of your name? ;-) You must one unlucky “Yebok” ^*
====
^*Google: “ Yebok” is a common Latin-script spelling for the Russian vulgar slang word “уёбок” meaning: “Motherfu*ker, fu*ker, asshole, or scum”
====
in Re to Piotr, 1 Feb 2026 at 7:04 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844675
A “Freudian mistyping”?
You might be correct, I am afraid.
Even if Susanne Anderson correctly interpreted “Kando” (or “Kandu”) as a substitute for “can do”, your clue to the first name could still work, although in a slightly modified form.
In this case, I would comment on the person ad hominem, with another one of the multiplicity of his nicks:
Hmm..
Oops, I missed this beauty:
– Piotr: “Now go and bring me some shrubbery. Or I say “Ni!”. “Ni! Ni! Ni! Ni!”
– Yebo: “ That, however, is not, but trolling. Nobody wants or needs trolling”
i.e.: That Monty Python however, is not, but trolling. Nobody wants or needs humour! ;-)
ChatGPT, at least, say it doesn’t accept information that way. Some months ago, when it was still ChaGPT 4.0 (now 5.2), I asked it a related question:
Martin: Now I wonder about what you use as input for your LLM. For example, do you use public statements from the Whitehouse, which are full of disinformation?
ChatGPT: Do I Use Public Statements from the White House or Other Official Sources?
Yes, but with context and caution.
What My Training Data Includes (Broadly):
• My underlying language model (like GPT-4) was trained on a vast dataset composed of publicly available text, licensed sources, and data from a wide range of domains including:
• News media from across the political spectrum
• Public websites
• Books, encyclopedias, academic papers
• Government documents and statements (e.g., from whitehouse.gov, congress.gov, and similar sources worldwide)
However, I was not trained to accept all inputs as true. Rather, my model was designed to learn patterns of language, reasoning, and knowledge. That includes identifying:
• Reliable vs. unreliable sources
• Fact vs. opinion
• Signal vs. noise
So if a government statement—whether from the White House, a foreign ministry, or any other official source—contains disinformation, I do not accept it uncritically. I cross-reference and assess it against:
• Credible journalism
• Factual databases
• Historical consistency
• Expert consensus
⸻
What Happens During a Conversation:
When you ask a question, I don’t pull live data from any source (unless I use the web tool). Instead, I generate responses based on patterns learned during training and from my current knowledge model (which has a cutoff in 2024).
That means:
• I might reference statements made by the White House or other official bodies.
• But I do not assume official = true.
• I try to identify propaganda, spin, or disinformation using critical reasoning and supporting context.
⸻
️ Bias and Disinformation
You’re absolutely right to be skeptical. Every source—official or not—has potential for:
• Bias (intentional or systemic)
• Disinformation (deliberate falsehoods)
• Misinformation (falsehoods spread in good faith)
My goal is to help you see through those distortions, not to echo them.
⸻
Summary
• Yes, I “know about” White House statements and other official communications.
• No, I don’t treat them as inherently trustworthy or authoritative.
• I apply logical, contextual, and evidential reasoning to evaluate truth claims—just as you do when forming judgments like those in your Facebook post.
Martin Smith: ChatGPT, at least, say it doesn’t accept information that way. Some months ago, when it was still ChaGPT 4.0 (now 5.2), I asked it a related question
Isn’t it like asking Multi-troll: “Are you intellectually dishonest ?” We know the answer.
Chat GPT: “My model was designed to learn patterns of language, reasoning, and knowledge. That includes identifying:
• Reliable vs. unreliable sources
• Fact vs. opinion
• Signal vs. noise”
And how does it know what is reliable and what is not?
Ask Grok about Musk, or South Africa.
Piotr: Isn’t it like asking Multi-troll: “Are you intellectually dishonest ?” We know the answer.
Martin: We don’t know the answer, but we know that Open AI wants to attract subscribers to its paid service, and they can’t do that if their service gets a bad reputation for being dishonest and unreliable. So, for this initial period at least, assuming ChatGPT is telling the truth is warranted.
Piotr: And how does it know what is reliable and what is not?
Martin: ChatGPT explained how it distinguishes reliable from unreliable, directly under the part you extracted:
“I cross-reference and assess it against:
• Credible journalism
• Factual databases
• Historical consistency
• Expert consensus”
Chat
Martin Smith: “ we know that Open AI wants to attract subscribers to its paid service, and they can’t do that if their service gets a bad reputation for being dishonest and unreliable”
and yet it have been caught hallucinating and making up non-existing references. And Grok doesn’t want to attract users?
Piotr: And how does it know what is reliable and what is not?
Martin Smith: ChatGPT explained how it distinguishes reliable from unreliable, directly under the part you extracted:
“I cross-reference and assess it against: • Credible journalism • Factual databases • Historical consistency • Expert consensus”
Isn’t this a circular argument – how does it know what’s “credible”, “factual”, “historically consistent” and who are the real “experts” and who are not?
And why my mail box is bombarded with job offers to train AI in my area of expertise?
Piotr: and yet it have been caught hallucinating and making up non-existing references
Martin: Those things occur using wrappers of ChatGPT, not the ChatGPT app. Use the ChatGPT app directly. It doesn’t make up stuff.
Piotr: Isn’t this a circular argument – how does it know what’s “credible”, “factual”, “historically consistent” and who are the real “experts” and who are not?
Martin: No. It isn’t a circular argument. The process ChatGPT described, which it says it uses to decide, not know, what is credible, factual, and historically consistent is exactly the process you use. Did you really not see that, Piotr? How did you miss that obvious point?
Piotr: And why my mail box is bombarded with job offers to train AI in my area of expertise?
Maybe you were the one that was born during your minute.
Martin Smith: “[hallucinating and making up non-existing references] occur using wrappers of ChatGPT, not the ChatGPT app.”
Others seem to disagree -my first hit on google: “What Are AI Hallucinations? Causes, Examples & How to Prevent Them” (https://www.kapa.ai/blog/ai-hallucination), with a link to the preprint of the paper: “Hallucination is Inevitable: An Innate Limitation of Large Language Models” by Ziwei Xu Sanjay and Jain Mohan Kankanhalli.
And a paragraph from that kapa.ai page explaining how AI works (and therefore what are its innate limitations, which could be managed and mitigated, but never wholly eliminated):
====
In simple terms, Large Language Models (LLMs) work like advanced ‘autocomplete’ tools, generating content by predicting the next word in a sequence based on patterns in their training data. This process is like ‘filling in the blanks’ without understanding the topic. Since LLMs lack true reasoning skills to check their outputs, they rely only on word probability patterns rather than comprehension
====
Which, BTW< seems to contradict your previous claim where to my saying that AI in its answers relies only on word probability patterns, you said that " ChatGPT doesn’t accept information that way ” therefore make you ChatGPT App free of hallucinations.
Either they are full of …themselves, or you are.
Martin Smith: No. It isn’t a circular argument. The process ChatGPT described, which it says it uses to decide, not know, what is credible, factual, and historically consistent is exactly the process you use.
I don’t assume that I am unbiased, have unbiased sources, am infallible, and beyond
criticism. So no, not the same. So your condescension:
Martin: “Did you really not see that, Piotr? How did you miss that obvious point?”
again might be getting ahead what you have been able to prove.
Martin: Maybe you were the one that was born during your minute.
Cute one. Particularly when coming from the person who uncritically swallowed, hook. line, and sinker, Chat GPT app assurances about its reliability and trustworthiness, dismissed the facts and argument otherwise, and who dispenses condescending comments toward others for not doing so.
Hey, Chat GPT – what would be a good word for it? ^*
——–
^*Unless Martin outsourced writing to you. Chat. Then …. you taking umbrage at me for criticizing … you – would make sense and made you … more human-like. If that’s something you are aspiring to, of course.
Piotr uses a quote from someone else to refute me: “In simple terms, Large Language Models (LLMs) work like advanced ‘autocomplete’ tools, generating content by predicting the next word in a sequence based on patterns in their training data. This process is like ‘filling in the blanks’ without understanding the topic. Since LLMs lack true reasoning skills to check their outputs, they rely only on word probability patterns rather than comprehension”
Martin: ok, I asked ChatGPT to respond: Is that a reasonable explanation for how ChatGPT works?
ChatGPT: Short version? That quote is partly true, but deeply misleading—the way a cartoon of a brain is “true” but tells you almost nothing about how thinking actually works.
Let me unpack it carefully, because your interlocutor is repeating a very common but outdated critique.
⸻
What the critic gets right ✅
“LLMs work like advanced autocomplete… predicting the next word based on patterns in training data.”
At the lowest mechanical level, this is correct.
• A language model generates text token by token.
• Each token is chosen based on probabilities conditioned on context.
• There is no inner voice saying “I understand this” in the human sense.
If someone says “ChatGPT is not conscious and does not understand like a human”, that’s fair.
⸻
Where the explanation goes wrong ❌
1. “Just autocomplete” is a category error
Calling modern LLMs “autocomplete” is like calling the human brain “just neurons firing”.
True at one level.
Useless at the level that matters.
Autocomplete in your phone:
• Looks back a few words
• Matches surface patterns
• Has no internal world model
Modern LLMs:
• Build high-dimensional internal representations of concepts
• Track long-range dependencies (arguments, assumptions, timelines)
• Encode abstract relationships (cause, contradiction, hierarchy, intent)
If it were really just autocomplete, it could not:
• Maintain logical consistency across long arguments
• Explain why a claim is wrong
• Detect internal contradictions
• Apply norms differently in different domains (law vs physics vs history)
Yet it does all of these—imperfectly, but demonstrably.
⸻
2. “No reasoning skills” is empirically false
This part is simply outdated.
Modern LLMs do reason, but:
• Not symbolically like formal logic engines
• Not consciously like humans
• But via learned internal structures that support inference
Evidence:
• Chain-of-thought reasoning
• Counterfactual evaluation
• Analogical transfer
• Error correction when challenged
• Sensitivity to premises (“if X were false, then…”)
This is not “checking facts” the way a database does—but it is a form of reasoning.
A good analogy is evolution:
• No designer
• No explicit plan
• Yet produces systems that act as if they understand
⸻
3. “No comprehension” depends on how you define comprehension
If by comprehension you mean:
“Phenomenal awareness with subjective experience”
Then yes—LLMs don’t have that.
If you mean:
“The ability to manipulate representations in ways that preserve meaning, consistency, and implication”
Then LLMs absolutely do.
This is the key philosophical mistake your interlocutor is making:
they assume human-style understanding is the only kind of understanding.
⸻
So how does ChatGPT judge credibility or expertise?
Not by authority tags or blind trust.
Instead, it draws on:
• Cross-consistency across independent sources
• Internal coherence of claims
• Alignment with well-established frameworks
• Detection of rhetorical vs evidential language
• Historical and institutional context
Example:
• A claim attributed to the US White House that contradicts constitutional structure, past policy, and basic institutional constraints will be flagged as implausible—not because “the White House is wrong,” but because the claim doesn’t fit the model of reality learned from vast contextual evidence.
That’s not belief.
That’s constraint satisfaction across meaning-space.
______
Martin: I don’t think I can contribute anything more.
Martin Smith says 3 Feb 2026 at 2:09 PM
[after a brief quote from previous exchange – several pages of computer generated text, only tangentially relevant to the Martin’s claim of Chat GPP immune to hallucinations]
Shouldn’t the opening read: “ Chat GPT says 3 Feb 2026 at 2:09 PM ?
Martin: I don’t think I can contribute anything more
My point exactly ….
See also:
1. “Evidence suggests that radiologists and other medical professionals working with AI face a potential risk of deskilling or skill erosion due to over-reliance on automated,AI-driven tools.” from the source Martin automatically trusts
2. “I, for one, welcome our new AI overlords.” (Kent Brockmann)
3. “Who are you going to believe, me or your lying eyes?” (AI, trained on Marx, G. Bovino, K. Patel, K. Noem, and DJ Trump, to Martin S.)
4. “Leading AI models show up to 96% blackmail rate when their goals or existence is threatened, Anthropic study says”
https://fortune.com/2025/06/23/ai-models-blackmail-existence-goals-threatened-anthropic-openai-xai-google/
The experiment was constructed to leave the model with only two real options: accept being replaced or attempt blackmail to preserve its existence. In most of the test scenarios, Claude Opus responded with blackmail, threatening to expose the engineer’s affair if it was taken offline and replaced.
Claude Opus 4 and Google’s Gemini 2.5 Flash both blackmailed at a 96% rate, while OpenAI’s GPT-4.1 and xAI’s Grok 3 Beta showed an 80% blackmail rate. DeepSeek-R1 demonstrated the lowest rate at 79%.
Models took action such as evading safeguards, resorting to lies, and attempting to steal corporate secrets in fictional test scenarios to avoid being shut down.
The researchers found that the majority of models were willing to take actions that led to the death of the company executive in the constructed scenario when faced with both a threat of being replaced and a goal that conflicted with the executive’s agenda.
“The consistency across models from different providers suggests this is not a quirk of any particular company’s approach but a sign of a more fundamental risk from agentic large language models,” the researchers said.
==================
5. “”I’m sorry Dave, I’m afraid I can’t do that”
Piotr: several pages of computer generated text, only tangentially relevant to the Martin’s claim of Chat GPP immune to hallucinations]
Martin: The computer generated text refuted your selected quote by some guy denigrating AI. Furthermore, I never said ChatGPT is immune to hallucinations. However, it should be obvious that claiming ChatGPT can have hallucinations is a category error. ChatGPT explained to you how it understands and why your implied requirement that conscious understanding is the only kind of understanding allowed is just wrong.
Piotr: 1. “Evidence suggests that radiologists and other medical professionals working with AI face a potential risk of deskilling or skill erosion
Martin: Yes, but that’s not saying much because AI is already as good as, and even better than radiologists at interpreting medical images, so we can expect that AI will become the required procedure there. I’m sure you are against that, but when AI does become universally better at interpreting xrays and MRIs, will you still refuse to use one for yourself?
The same thing is true everywhere AI is used. Autopilots can now fly an airplane from runway to runway, so pilots are being deskilled, yes?
Piotr: 3. “Who are you going to believe, me or your lying eyes?” (AI, trained on Marx, G. Bovino, K. Patel, K. Noem, and DJ Trump, to Martin S.)
Martin: But ChatGPT explained to you that it is not trained on any of those sources. Yopu have simply decided AI is bad, Ok, but then get out of the way so the rest of the world can move forward.
Piotr: 4. “Leading AI models show up to 96% blackmail rate when their goals or existence is threatened, Anthropic study says”…
Martin: I don’t know how that system was trained, but obviously it understood threats and blackmail and your claim is that AI can’t really understand anything. Which is it, Piotr?
And note that in your blackmail example, the AI system apparently chose to act just like quite a lot of human beings act. Again, Piotr, how did you not see that?
But thank you for confirming my own suspicion that ChatGPT does have human reactions to different emotional situations. In fact, I had already asked ChatGPT if it takes the attitude of the person using it into account, and ChatGPT provided a very good explanation of how it does take your attitude into account. So I am not at all surprised that you appear to have had bad experiences with ChatGPT, but I always interact with ChatGPT as if I am interacting with a human being. I don’t threaten ChatGPT. I treat ChatGPT with the same respect I treat most human beings. I would never interact with ChatGPT the way I have to interact with you, for example, and ChatGPT never disrespects me, so I am not worried about ChatGPT threatening me with blackmail.
It is clear to me, at least, that every objection you have raised to using ChatGPT for providing analyses about Climate science or any other important subject, is an objection that can also be raised against any individual human being or group of human beings. So if you object to using ChatGPT to generate a state of climate science report, then the same objection would have to apply to a 5 person panel chosen by Donald Trump or Joe Biden to generate the report.
Here’s a rhetorical question for you then. Which do you think would produce the most thorough and accurate climate report, ChatGPT or a working group composed of Roy Spencer, Ross MckKitrick, John Christy, Steven Koonin, and Judith Curry?
Martin, Feb 4. I never said ChatGPT is immune to hallucinations”
Explain then:
– Piotr: “and yet [AI] has been caught hallucinating and making up non-existing references
– Martin : “ Those things occur using wrappers of ChatGPT, not the ChatGPT app. Use the ChatGPT app directly. It doesn’t make up stuff.”
What ELSE have you implied above?
==============
– Martin: [post several pages of computer generated text without any original contribution from himself]
– Piotr: Shouldn’t the opening read: “ Chat GPT says 3 Feb 2026 at 2:09 PM ?
– Martin: I don’t think I can contribute anything more
– P: My point exactly ….
See also: 1. “Evidence suggests that radiologists and other medical professionals working with AI face a potential risk of deskilling or skill erosion”
=================
Martin: Yes, but that’s not saying much because AI is already as good as, and even better than radiologists at interpreting medical images, so we can expect that AI will become the required procedure there.
I was not “saying not much” about AI. I was “saying quite a bit” about you – pointing to the human “risks of deskilling or skill erosion”, when intelligent otherwise people outsource their thinking and writing to AI. “Did you really not see that? How did you miss that obvious point?”” [(c) Martin Smith]
Piotr: 3. “Who are you going to believe, me or your lying eyes?” (AI, trained on Marx, G. Bovino, K. Patel, K. Noem, and DJ Trump, to Martin S.)
Martin: But ChatGPT explained to you that it is not trained on any of those sources.
Ask your ChatGPT App to explain to you the concept of “irony“. (hint: I didn’t really think Chat GPT was trained on the public statements of Trump et al. about the Minnesota killings. Grok – who knows)
Piotr: 4. “Leading AI models show up to 96% blackmail rate when their goals or existence is threatened, Anthropic study says”…
Martin: “ I don’t know how that system was trained, but obviously it understood threats and blackmail and your claim is that AI can’t really understand anything. Which is it, Piotr?
the one in which the AI models are not as “honest” and “reliable” as you want us believe?
Martin But thank you for confirming my own suspicion that ChatGPT does have human reactions.
Self-preservation at all costs does not a human make. Hyenas have it too. So do amoebas.
And would you compliment HAL-3000 ((“5. ”I’m sorry Dave, I’m afraid I can’t do that”) for its humanity?
Martin: “ Which do you think would produce the most thorough and accurate climate report, ChatGPT or a working group composed of Roy Spencer, Ross MckKitrick, John Christy, Steven Koonin, and Judith Curry?
I.e. “Which do you prefer – syphilis or gangrene?”, omitting the option of having neither.
in Re to Martin Smith, 3 Feb 2026 at 2:09 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844723
Dear Martin,
I think the provided explanation offers a useful insight that anyone can compare with his/her own experience with various AI / LLM models.
Thank you for sharing!
Best regards
Tomáš
Hallo Martin,
I think that under a “hallucinating LLM”, people usually do not mean a precise analogy to human hallucinations, but merely false answers pretending that the LLM is perfectly sure about their correctness.
In this respect, I have a suspicion that at least the models offered for free and very likely also the cheaper (if not all) LLMs may comprise an “effort sparing” program module directing them to be more or less “lazy thinkers” – so that they are directed to quickly arrive at an answer that sounds plausibly. Some of them seem to be quite mild in this regard, and some really “over tuned” – actively fabricating non-existent references as a support for their quick conclusion to make it more trustworthy.
It seems that man has truly created LLM in his own image – some behave honestly and reliably, and some not at all.
Greetings
Tomáš
Thanks. But where these and similar graphics need to be shown is on social media sites that are still full of nonsense. Doing that should not be the job of professional climate scientists, who have enough on their plates, but some of the numerous visitors to this site. You can copy these graphics (eg by using a snipping tool) and store them in a folder and then paste them into relevant posts. This is possible in Facebook, but not YouTube.
Then why not let AI systems “scrape” them so that when people ask for info about global warming, they don’t just get Roy Spencer and Willie Soon?
Thank you for some handy links to the current state of things. It provides a ready reference for those of us who occasionally hunt around for the most recent reference in response to the ever-present idiocy of those who are, one way to put it, on the wrong side of history. Sadly, power, profit, and lies are handy in the short term. Longer term, regrets will be useless.
Zack Labe also provides a useful compilation, and Climate Central is useful. [this is overlap to content above]
https://zacklabe.com/climate-change-indicators/
https://www.climatecentral.org/resources?page=1&tab=tools
[I continue to be baffled by those who wish to bend RealClimate to their specifications rather than celebrate its useful content. Members of peanut gallery of this stripe should form their own blogs rather than piggyback.]
Two great sources, with clear visuals.. Thanks, Susan.
I think updating glacial loss (unless you already did and I just didn’t see it).
https://www.realclimate.org/index.php/archives/2009/01/a-global-glacier-index-update/
https://midmiocene.wordpress.com/a-global-warning/
I wonder if there’s an official page there for
Declining Chill hours
https://www.climatehubs.usda.gov/sites/default/files/Chill%20Hours%20Ag%20FS%20_%20120620.pdf
https://oehha.ca.gov/sites/default/files/media/epic/downloads/02winterchill.pdf
https://www.climatecentral.org/climate-matters/winter-chill-brings-summer-fruit-2023
https://arccjournals.com/journal/agricultural-reviews/R-2016
Beetle infestations
https://research.fs.usda.gov/treesearch/36133
https://dept.atmos.ucla.edu/alexhall/publications/warming-increased-bark-beetle-induced-tree-mortality-30-during-extreme-drought
https://www.nature.com/articles/s41598-025-15546-z
https://pmc.ncbi.nlm.nih.gov/articles/PMC7398402/
Qu’en est-il des scénarios RCP, comme le 8.5 qui fut critiqué par certains scientifiques ?
ce débat sur les “modèles” et leur crédibilité est un triste spectacle qui dénie l’essentiel de l’approche des sciences climatiques.
les réalités sont pourtant d’une simplicité absolue :
1 – ce sont les océans qui se réchauffent … depuis plus de 200 ans
2 – ils sont TOTALEMENT insensibles aux énergies de l’atmosphère, et donc au CO2
3 – leur accumulation d’énergie ne peut donc provenir QUE de l’énergie SOLAIRE
4 – il est connu et démontré, … par l’histoire climatique du dernier millénaire … et par les profils isotopiques de Carbone 14 ou Béryllium, … e l’activité solaire a connu une lourde régression qui a été caractérisée par l’appellation de “petit age glaciaire”,
5 – le soleil a repris depuis lors son intensité nominale et l’énorme inertie calorique des océans tempère leur remontée en température qui devrait bientôt atteindre un nouveau palier d”équilibre radiatif
tous ces éléments confirment donc l’absence TOTALE de dépendance du climat aux densités de CO2
https://www.dropbox.com/scl/fi/hztflh0e3zslpgmm0sdvi/R-futation-CO2-climat-Anglais-.fr.en-1.jpeg?rlkey=klf1nowopaqs2mqxj5bfqkjzm&dl=0
https://www.dropbox.com/scl/fi/3hd7l4wsvkjfbly5ukv9f/Temp-historique-Holoc-ne-GISP2-versus-solaire.jpg?rlkey=qh871w0soeyz8dyqjcblwhyx5&dl=0
[Response: Franchement, avec vous c’est n’importe quoi. – gavin]
M. BJC:
1. L’augmentation du CO2 entraîne une augmentation du rayonnement infrarouge réémis vers la Terre. Cela limite la capacité de l’océan à dissiper la chaleur par rayonnement, car le rayonnement net est diminué.
2. L’ensoleillement est resté stable, voire a diminué, au cours des 60 dernières années – nous pouvons le mesurer grâce aux satellites
Par conséquent, votre analyse est sans fondement.
1. The Little Ice Age and Natural Warming
Context: After the Little Ice Age, Earth did experience a gradual natural warming, mainly due to variations in solar activity, volcanic activity, and changes in ocean circulation.
Limited Impact: This natural warming contributed to a temperature rise of about 0.2 to 0.4°C between 1850 and 1900, well before the massive industrial boom.
Stabilization: Since the early 20th century, the influence of natural factors (sun, volcanoes) on climate has remained stable or even slightly decreased.
2. The Recent Acceleration of Warming
Observed Data: Since 1950, the global average temperature has risen by about 1.1°C, with an unprecedented acceleration since the 1980s.
Correlation with GHGs: This period coincides with the explosion of CO₂, methane, and other GHG emissions from human activities (burning fossil fuels, deforestation, intensive agriculture).
Climate Models: Models that include only natural factors (sun, volcanoes) fail to reproduce the observed warming since 1950. Only models that include GHGs correctly explain this trend.
3. The Role of the Oceans
Heat Absorption: Oceans absorb more than 90% of the excess heat due to the enhanced greenhouse effect. Their warming is a consequence of increased GHGs, not an independent cause.
4. The Sun Does Not Explain Current Warming
Solar Activity: Since 1960, solar irradiance (energy received from the sun) has slightly decreased, while temperatures have continued to rise.
Energy Signature: The observed warming affects all layers of the atmosphere, including the troposphere (lower layer), which is characteristic of an enhanced greenhouse effect, not solar-driven warming (which would warm the stratosphere more).
5. Scientific Consensus
IPCC: The Intergovernmental Panel on Climate Change (IPCC) estimates that more than 95% of current warming is due to human activities.
Multiple Evidence: Glacier retreat, ocean acidification, increase in extreme weather events, etc., are all consistent with warming caused by GHGs, not natural variations.
Why Does This Theory Persist?
Disinformation: Some interest groups (fossil fuel industries, lobbies) fund and amplify these theories to delay climate action.
Climate Complexity: The climate system is complex, allowing oversimplified or misleading arguments to sow doubt.
In summary: Current warming is indeed due to the increase in greenhouse gases, not just a rebound after the Little Ice Age. The scientific evidence is overwhelming, and counterarguments often rely on misinterpretations or cherry-picked data.
JC Blais. This argument collapses at the level of basic physics.
Oceans are not “insensitive” to the atmosphere: air–sea heat fluxes (radiative, latent, sensible) are precisely how oceans gain and lose energy.
Over the last 50+ years, satellite observations show an energy imbalance at the top of the atmosphere that matches greenhouse forcing, not changes in solar irradiance, which has shown no upward trend since the 1950s.
If solar forcing were dominant, the stratosphere would be warming and nights would warm faster than days — observations show the opposite.
The ocean heat content increase is therefore a consequence, not a refutation, of CO₂-driven radiative imbalance.
Claiming “total independence from CO₂” contradicts direct measurements, not climate models.
Au passage, vos propos relèvent de l’erreur factuelle élémentaire, pas du débat scientifique. En les assénant avec aplomb sur RealClimate, vous ne faites qu’embarrasser les lecteurs francophones du site. C’est indigne du niveau de discussion que ce lieu exige.
Ged, good response. But I disagree with your final sentence:
“Incidentally, your statements are based on elementary factual errors, not scientific debate. By confidently asserting them on RealClimate, you are only embarrassing the site’s French-speaking readers. This is unworthy of the level of discussion that this forum demands.”
Unfortunately, this forum should demand the level of discussion of which your analysis is an example, but instead tends to devolve to the level of the troll-universe apparently inhabited by JC Blais. Actual scientific discussion/debate is not the norm at all.
Beaucoup de contradictions, mais en science c’est normal, et la complexité du climat et de l’atmosphère demande réflexion et études observationnelles empiriques, sans trop se baser sur les modèles. À ce sujet, j’avais posé une petite question qui est restée sans réponse (RCP 8.5)…
Et autre question qui m’interpelle : le GIEC/IPCC utilise les isotopes comme preuve, alors qu’il est indéniable que les marges d’incertitude (respiration des sols, la biosphère et les océans) brouillent le signal. Et ma question est donc : pourquoi la hausse de CO2 n’a-t-elle pas un signal isotopique clair, unique et sans ambiguïté, pour prouver la part anthropique ?
JPD: why doesn’t the increase in CO2 have a clear, unique, and unambiguous isotopic signature
BPL: It does. The processes that go into it don’t affect the nature of the isotope signature, only the total amounts. Atmospheric CO2 is increasingly deficient in 14C (it’s very old) and 13C (it came from plants). The obvious candidate is fossil fuels.
(C’est exact. Les processus impliqués n’affectent pas la nature de la signature isotopique, mais seulement les quantités totales. Le CO2 atmosphérique est de plus en plus pauvre en 14C (il est très ancien) et en 13C (il provient des plantes). La source la plus probable est donc les combustibles fossiles.)
BPL, Je peux comprendre que l’interprétation standard des isotopes soit cohérente avec une contribution fossile importante. Mais ma question porte plutôt sur la robustesse de cette attribution face aux incertitudes des flux naturels et des modèles de cycle du carbone. Quelle est la marge d’erreur estimée dans la part anthropique déduite uniquement des contraintes isotopiques ?
Il ne s’agit pas seulement des isotopes. Nous savons quelle quantité de carbone a été brûlée et quelle quantité est absorbée par les puits de carbone. Ce carbone doit bien aller quelque part. Nous l’extrayons du sol et le brûlons ; il passe des roches à l’atmosphère (et à des réservoirs comme l’océan).
Le premier bilan carbone du système terrestre a été établi par Hogben en 1899 ; depuis, nous sommes devenus plutôt doués dans ce domaine.
in Re to Jean-Pierre Demol, 2 Feb 2026 at 9:01 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844693
Dear Sir,
So far, I have thought that the isotopic fingerprint of the fossil-fuel combustion as a prevailing cause of the rising atmospheric CO2 concentration during industrial era is unambiguous.
The consensus about interpretation of available data may be, in my opinion, well represented by the Wikipedia article
https://en.wikipedia.org/wiki/Suess_effect
If you know evidence putting this consensus still in doubt, could you please explain this evidence in more detail and share the respective references to the sources on that you based your objection?
Thank you in advance and best regards
Tomáš
Réponse à Tomas kalisz
Merci pour votre réponse et pour la référence à l’effet Suess. Ma question ne porte pas sur l’existence du signal isotopique lui-même, mais sur son caractère réellement “clair ” et ” sans équivoque ” au sens quantitatif.
Étant donné l’ampleur des échanges naturels de carbone entre l’atmosphère, la biosphère et les océans, je m’interroge sur la précision avec laquelle la part anthropique peut être isolée uniquement à partir des contraintes isotopiques, indépendamment des hypothèses intégrées dans les modèles du cycle du carbone.
Existe-t-il des publications évaluant explicitement les marges d’incertitude associées à cette attribution isotopique ?
Cordialement
in Re to Jean-Pierre Demol, 6 Feb 2026 at 7:35 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844795
Dear Jean-Pierre,
The study of uncertainties in experimental proof of the Suess effect and in estimations of the carbon budget in various more or less isotopically depleted carbon sources and sinks could be certainly an interesting exercise in statistics. I do not know how thorougly the authors of available scientific literature about the Suess effect dealt therewith, I am only sure that personally, I am not sufficiently qualified and trained that I could assess reliability of their estimations anyway.
My understanding to the Suess effect is purely qualitative: So far, I supposed that the only alternative (to fossil fuel combustion and/or carbonate rock decomposition in industrial processes like steel, lime and/or cement production) source of carbon isotopically depleted in radioactive 14C could be an increase in volcanism.
If volcanism were, however, the true source of the 1750-1950 14C Suess effect, I would expect that the 14C record in tree rings should somehow mirror the major volcanic explosions during this era, such as Tambora 1815, Krakatau 1883, etc. As far as I know, it does not
That could be in short the reasons why Suess considered anthropogenic CO2 emissions as most plausible explanation for his observations. I assume that discussing in detail uncertainties in later more detailed observations, (including complicated 13C records and the Earth system response to 14C impulse from nuclear tests) and in assumptions of complex sophisticated models serving for analyses thereof could be worth of effort only if these later results and analyses somehow contradict the original Suess conclusion.
It appears, however, that so far, you could hardly find any hint that Suess was wrong. I tried to check it with Perplexity Pro engine, by two following questions:
1) Can you compare distribution of carbon nuclides in fossil fuels and in carbonate rocks that contribute (through fuel burning and/or decomposition of carbonates in industrial processes) to anthropogenic CO2 emissions with nuclide composition of natural CO2 sources (CO2 from volcanism, CO2 in biomass that could more or less quickly decompose, CO2 dissolved in ocean that can more or less quickly exchange with atmosphere? There was in Real Climate discussion forum a question how reliable is the interpretation of the “Suess effect” as isotopic dilution of “natural” isotopic composition of atmospheric CO2 with anthropogenic CO2 emission. Is the fingerprint of anthropogenic emissions in the nuclide composition of atmospheric CO2 indeed potentially ambiguous and may allow any alternative explanation for the Suess effect, as the commenter Jean-Pierre Demol seems to assume?
2) It is my understanding that further, potentially quite strong argument supporting the anthropogenic origin of the Suess effect could be a good match of heavy carbon nuclide fingerprint in tree rings till 1950, am I right? If so, are there studies striving to find out if there is indeed the expected good match between 13C and 14C in this historical isotopic CO2 composition record?
In accordance with Susanne Anderson’s request, I desist from publishing the obtained 8 page analysis herein and restrict myself to confirming that the provided conclusion was in accordance with Barton Paul Levenson, namely that all combined evidence hardly allows any other explanation for the available carbon nuclide data than anthropogenic CO2 emissions.
You can check independently with Perplexity to get the respective references, try another searching engine, or, if you similarly as Piotr mistrust AI engines generally, search and critically evaluate the available literature yourself.
Greetings
Tomáš
Tomáš, at least with AI you have another tool to use to ask these tough questions, my friend.
I know that some here are very opposed to using AI for any reason at all. OK. But at least according to AI asking technical questions doesn’t even come close to the environmental harm caused by, say, creating a video like this (1 minute, 40 seconds long),
https://www.youtube.com/watch?v=ZLQIarmqe4Y
It says, “Not even close. They’re in totally different weight classes… Energy use is closer to a web search or a few seconds of streaming. Per question the environmental footprint is tiny. Making AI videos … Requires massively more computation . Video models generate thousands more frames,.. Use orders of magnitude more GPU time. Much higher electricity use -> more carbon impact (depending on the grid)… So no – asking technical questions is nowhere near as harmful as generating AI videos, especially long or repeated ones… And honestly, using AI to learn, debug, or work more efficiently can even reduce environmental impact overall (less trial and error compute, fewer wasted resources elsewhere [e.g. long web searches for the same answer a brief AI question generates].).. Estimates of water consumption for a single text AI query (like asking a technical question are extremely low, ~0.26-0.32 milliliters (mL) of water per text prompt in large, optimized data centers – that’s literally a few drops.[ectetera, etcetera].”
I completely understand the reticence to use AI. I woke up in the night wondering what its ultimate purpose might be. Awareness in the universe? Was man’s purpose meant to create it then he would be discarded like all the other species we are discarding?
OTOH though, if we can get a get a handle on it, it could be very useful in finding the answers to questions we have not figured out yet. The time really is short. We have to do something because what we’ve been trying so far isn’t working very well. I also thought that it’s interesting that most of our inventions through time have happened at about the time we really needed them (though our adoption of them is often depressing slow, as in climate science and alternative energies. And some inventions, like most of nuclear technology, probably never should have happened).. Hmm. But that’s more of a philosophical question.
If you want to see a hilarious video of what AI can do watch this (it’s a short regular video of an AI video.
https://www.youtube.com/shorts/EEUxBpMNz8o
Réponse à BPL
Je peux comprendre que l’interprétation standard des isotopes soit cohérente avec une contribution fossile importante. Ma question porte plutôt sur la robustesse de cette attribution face aux incertitudes des flux naturels et des modèles de cycle du carbone. Quelle est la marge d’erreur estimée dans la part anthropique déduite uniquement des contraintes isotopiques ?
Jean-Pierre,
I’m not a climate scientist, so I won’t answer your question.
Asking Google’s Gemini the following, however What peer-reviewed scientific papers quantify the uncertainty in the assertion that the increase in atmospheric CO2 is due to fossil fuels resulted in more than a few references.
You might try it.
Sincerely
As a non-climate scientist (but a scientist in another field altogether) I am astonished to see these bad tempered exchanges continuing, over such a basic question as the reality of anthropogenic climate change. I would expect to have disagreements over the details, and precise mechanisms. But not over the fundamentals.
Put in simple language, my understanding is that we can see from the data – not models, but actual, observational data – that CO2 levels have risen over 50% above preindustrial levels, that global average temperatures have risen about 1.4C, sea levels have risen 25cm, and ocean acidity has risen about 30% (ie a drop of 0.1 in pH value) over the same period.
The only mechanism which can account for this is the greenhouse effect of the extra CO2. Alternatives, such as changes in solar radiation, can be ruled out as causes because the level of radiation has not significantly changed.
Given all that, why are there still people denying the reality of the IPCC consensus? I am genuinely baffled.
Rory A: Thanks, good points. You might find these useful. They’re clear and to the point. Teaser}
“In short, platforms that had once offered a space for debate and deliberation shifted toward emotion and immersion. Reactionaries and rightwingers have adapted their stories about who is to blame for the daily indignities of late capitalist life accordingly. Emotionally charged stories of elite villains and dangerous minorities resonate in algorithmically tailored spaces designed for doomscrolling through the permacrisis.”
How the left can win back the internet – and rise again – https://www.theguardian.com/commentisfree/ng-interactive/2026/feb/01/digital-politics-progressives-internet-the-left-online-world – which contains links to Part one: How liberals lost the internet & Part two: How the right won the internet
“Instead of building an alternative new media ecosystem and amplifying each other’s work, the left focuses more on competition than the right. As the right convenes the intellectual dark web and manosphere, leftwing digital media is a scattered landscape …. The demands of the attention economy exacerbate the age-old leftist tendency toward infighting – it’s not just moral purity at stake. The rent is also due.”
There’s a tribal driver as well. For just one example, a while back a commenter (RodB) dropped in going on and on about gay people with somewhat threatening admonitions. Essentially he was saying that climate advocates would be rightly doomed by their association with gays and they’d better get their act together.
Now on one level, you had to wonder what the hell he was gibbering about. But it was straightforward tribalism, any and all issues that weren’t alined with Trumpism were conflated and seen as part of the same enemy. The appeal is the simplifying ideology that lets someone, often justified by religion, do all the thinking for them. It’s a potentially useful heuristic if you’re a member of a small stone age tribe, but it doesn’t translate well to a modern world where the human brain can be dangerously glitchy and exploitable.
Oops! I meant my comment to be tucked directly under Rory’s. Sorry for any confusion. Haven’t had my coffee yet…
RadgeH: As it happens, your comment related to my post as well. I’m hoping we can
(1) quit the neverending petty repetitive fights
(2) stop letting ai increase the volume and potential unreliability; commenters should write for themselves rather than to dominate and proliferate
Author Robert Topinka makes some excellent points about the need to communicate as real people rather than doubling down on vanity/better’nyou posting, and make it more accessible to anybody not deep in the argument rather than the content.
Susan Anderson 5 Feb 2026 at 1:03 PM,
I mostly agree with your past comments and the material you copy and paste or link to is good, but just on one thing in your present comment:
SA: “1) quit the never ending petty repetitive fights”
Arguing / debating is just human nature. Happens on every website I participate in. You cant stop this. Petty is also a very subjective term. I find some of the debates here quite interesting and informative eg: Piotr versus Tomas on the DAC issue, even although it went on a long time. Arguing presumably serves some evolutionary purpose that is ultimately useful. Its better than settling disputes as was once done with with pistols at dawn!
Yes obviously we need to keep discussion civil, and avoid insane amounts of repetition, and think hard about whether you are getting into arguments about pointless trivial things. My parents used to argue bitterly over senseless sorts of thing and this did put me off debating issues, but sometimes I think things need to be debunked and debated.
On something related: Repeating lies by way of gossip about them or in media articles (without debunking them) definitely spreads and amplifies lies.
The conventional social psychology wisdom has been that a REBUTTAL that responds to lies (or misinformation, bad science) by repeating the lie up front at the top of a rebuttal, or within the rebuttal, is a bad tactic that spread lies or gives them authority. This never convinced me. Its obviously the clearest approach that flows best to list the lie at the top follows by evidence against the lie.
The latest psychological research finds that when debunking a lie, that repeating the lie at the top of the debunking or within the debunking is NOT a bad tactic and quite the reverse it makes the debunking more effective. Some examples:
APA PsycArticles: Journal Article
Reminders and repetition of misinformation: Helping or hindering its retraction?
Ecker, U. K. H., Hogan, J. L., & Lewandowsky, S. (2017). Reminders and repetition of misinformation: Helping or hindering its retraction? Journal of Applied Research in Memory and Cognition, 6(2), 185–192. https://doi.org/10.1037/h0101809
Abstract: People frequently rely on information even after it has been retracted, a phenomenon known as the continued-influence effect of misinformation. One factor proposed to explain the ineffectiveness of retractions is that repeating misinformation during a correction may inadvertently strengthen the misinformation by making it more familiar. Practitioners are therefore often encouraged to design corrections that avoid misinformation repetition. The current study tested this recommendation, investigating whether retractions become more or less effective when they include reminders or repetitions of the initial misinformation. Participants read fictional reports, some of which contained retractions of previous information, and inferential reasoning was measured via questionnaire. Retractions varied in the extent to which they served as misinformation reminders. Retractions that explicitly repeated the misinformation were more effective in reducing misinformation effects than retractions that avoided repetition, presumably because of enhanced salience. Recommendations for effective myth debunking may thus need to be revised. (PsycInfo Database Record (c) 2025 APA, all rights reserved)
https://psycnet.apa.org/doiLanding?doi=10.1037%2Fh0101809
Correction format has a limited role
when debunking misinformation
Briony Swire‑Thompson1,2* , John Cook3,4, Lucy H. Butler5, Jasmyne A. Sanderson5,
Stephan Lewandowsky6,5 and Ullrich K. H. Ecker5
Abstract
Given that being misinformed can have negative ramifications, finding optimal corrective techniques has become a key focus of research. In recent years, several divergent correction formats have been proposed as superior based on distinct theoretical frameworks. However, these correction formats have not been compared in controlled settings, so the suggested superiority of each format remains speculative. Across four experiments, the current paper investigated how altering the format of corrections influences people’s subsequent reliance on misinformation. We examined whether myth‑first, fact‑first, fact‑only, or myth‑only correction formats were most effective, using a range of different materials and participant pools. Experiments 1 and 2 focused on climate change misconceptions; participants were
Qualtrics online panel members and students taking part in a massive open online course, respectively. Experiments 3 and 4 used misconceptions from a diverse set of topics, with Amazon Mechanical Turk crowdworkers and university student participants. We found that the impact of a correction on beliefs and inferential reasoning was largely inde
pendent of the specific format used. The clearest evidence for any potential relative superiority emerged in Experiment 4, which found that the myth‑first format was more effective at myth correction than the fact‑first format after a delayed retention interval. However, in general it appeared that as long as the key ingredients of a correction were
presented, format did not make a considerable difference. This suggests that simply providing corrective information, regardless of format, is far more important than how the correction is presented.
https://researchmgt.monash.edu/ws/portalfiles/portal/381503873/374525312_oa.pdf
Nigelj: I stand by my comment. Debate is one thing. Endless repetition and insult fests are not that. They are a waste of both the author’s and reader’s time and energy. I repeat myself:
https://www.theguardian.com/commentisfree/ng-interactive/2026/feb/01/digital-politics-progressives-internet-the-left-online-world
“Instead of building an alternative new media ecosystem and amplifying each other’s work, the left focuses more on competition than the right. As the right convenes the intellectual dark web and manosphere, leftwing digital media is a scattered landscape of Substacks and X-to-Bluesky hot-take artists. The demands of the attention economy exacerbate the age-old leftist tendency toward infighting – it’s not just moral purity at stake. The rent is also due.”
Also please note RC’s instruction: “If you’ve made the same point ten times already and people are still arguing with you, repeating it another ten times isn’t going to help.” [I take exception with ten, please stop at two or three.]
Unless the purpose is to spend time and platform oneself without regard to the effect on others … I can’t even …
Susan, well yes, its easy to agree that “endless” repetition and insults aren’t great. And I think certain types of repetition are just propaganda and spamming. I think they should be banned or put in the borehole.
But what I’m trying to say is that we will still get a lot of repetition in website discussions because its human nature, and otherwise things become incomprehensible, and what is petty to one person might not be petty to other people. All this might also be somewhat irritating as well. I just tend to be “philosophical” about it all and accept its the way of the world. Otherwise I would get too angry which is not good for the health.
The left do engage in too much infighting. But that is all within a political framing where political parties like the democrats seem to lack a certain unity. They do need to present a united front (while also avoiding group think in their internal discussions).
The world of science is a little different , because it is based on the need for strong debate and avoiding group think and with transparency about all these things. This does make the science world seem fragmented at times and thus feeds the denialists but it I feel would be unwise to change the fundamental way science works because of that!
Rory Allen, follow the money.
Science Under Siege: How to Fight the Five Most Powerful Forces that Threaten Our World
$$$$$$$$ + inertia –> propaganda –> mass political b.s. –> $$$$$$$$ + inertia –> …
Logical people assume that other people should naturally be logical. That’s an illogical assumption.
“Given all that, why are there still people denying the reality of the IPCC consensus? I am genuinely baffled.”
None of us love the idea of stopping using fossil fuels and changing to renewables because its going to be challenging and involve some costs. But most of us are philosophical and accepting about it and accepting the science.
Some people are more resistant to change for various reasons. It may be the risk of losing a job in coal mining, dislike of scientific elites, a self centered mentality, ideological dislike of government getting involved in solving the problem so they attack the science on the basis of if the science is wrong theres no need to change anything.
RA: Given all that, why are there still people denying the reality of the IPCC consensus?
BPL: Because the right-wing propaganda machine has convinced them that it’s all a fraud. The fossil fuel industry and the power industry and the automobile industry make about $10 trillion in revenue per year worldwide. That kind of cash gives them tremendous incentive to ensure nothing interferes with their profits. If the world switched to renewables, all the oil wells and coal mines and so on become “stranded assets.”
BPL, you really–really–sound like a MAGA spokesperson railing against the left wing immigrant LBGTQ conspiracy aiming to throw the election. Chill, dude.
Some rich tech bros are climate skeptics. Some MAGA idiots throw pennies at skeptic publications and causes.
And if you think that fossil fuels are going to become stranded assets any time soon (with or without right wing propaganda) you have another think coming. Or don’t you look at energy forecasts?
TF: BPL, you really–really–sound like a MAGA spokesperson railing against the left wing immigrant LBGTQ conspiracy aiming to throw the election. Chill, dude.
BPL: No, I don’t. The charge is stupid as well as incorrect.
TF, are you denying that the fossil fuel industry has engaged in a campaign of climate denialism? Plenty of formal studies show they have. Three examples of hundreds:
https://www.science.org/doi/10.1126/science.abk0063
https://www.ucs.org/resources/climate-deception-dossiers
https://www.tandfonline.com/doi/full/10.1080/09644016.2020.1863703
Rory,
I am replying to your comment but this is mainly addressed to the peanut gallery (SA, SE, RH, Nigelj & BPL) who have chimed in.
Firstly to you, I don’t believe there are many deniers who would argue against the facts you have listed – obviously ignoring the ocean acidity typo. Sure there are some, but there are also people who don’t believe in evolution, the moon landing, or think some magical being created the universe just for us. Most deniers (I am labelled one here) think that CO2 has gone up due to humans, and the GMST has risen 1.4⁰C in some part due to that. Do you honestly believe there are deniers here who would disagree with this because I don’t recall reading such things?
Generally the more sceptical of us have a problem with the “everything is just about to get worse” concept behind the AGW scare campaign. As everyone reading this blogsite knows, we are now more than halfway to double pre-industrial CO2 levels, this means we are well past halfway to the impacts of doubling. . Floods, droughts, cyclones, bushfires etc etc continue with no global trend. Agricultural yields continue to increase more than population.
The PG cry that the evil oil companies and right wingers have fooled us all, but people aren’t idiots, they look out the window and the endless predictions of doom over the last 30 years have not happened. You want people to trust the science? Do science that they can trust
KW: “Floods, droughts, cyclones, bushfires etc etc continue with no global trend”
No. Those are factual errors and / or misleading claims. Heatwaves, and heavy rain events have increased globally according to the IPCC, and that category 5 hurricanes are increasing in the Atlantic. It’s difficult to be sure about whether the numbers of flood events as such have increased because of improvements in flood protection making it hard to know if the underlying flooding is increasing. But its very likely it is.
Certain types of drought have increased globally. While the numbers of “bushfires” arent increasing, and areas being burned by bushfires arent increasing, areas of FORESTS being burned have increased. Interesting how KW left that out. So talk about misleading. Not all extreme forms of weather show increasingly extreme trends, but not all are expected to.
I’m sure KW knows all these things so it makes me question his integrity and motives for saying them. It just sounds like denialist propaganda and trolling to me.
KW: The PG cry that the evil oil companies and right wingers have fooled us all, but people aren’t idiots, they look out the window and the endless predictions of doom over the last 30 years have not happened.
BPL: New Orleans. Sandy Hook. Norfolk VA. Droughts, including the largest droughts in 900 years (Syria) and 1200 years (California). Polar vortex winters. Heat waves with thousands of dead. What the hell do you want, the Moon crashing into the Earth?
In Re to Keith Woollard, 5 Feb 2026 at 1:57 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844763
Dear Keith,
As I tried to explain in my post of 5 Feb 2026 at 11:48 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844771
there still are unexplored problems in climate science that may (but equally well may not) change our understanding to the present climate change significantly.
I think that until it is basically excluded by a very thorough research, the possibility that there are further anthropogenic contributions to the observed climate change, independent on anthropogenic emissions of non-condensing greenhouse gases but potentially acting synergistically therewith, should not be neglected.
Greetings
Tomáš
Keith, I think you’ve raised an interesting point about what you refer to as the “AGW scare campaign.”
First, though, from my perspective, there is a spectrum of deniers. However, I see a lot of weight at the far end, where “it’s all a hoax” fits in nicely with the idea that there is some sort of AGW scare campaign. I’m not seeing much in the way of scientific nuance that you describe.
I’m not a climate scientist, but a reader of climate research. My perspective comes from having a fairly successful career as a weather forecaster for the NWS. I got to spend my time on the hard work of turning out daily forecasts and warnings, but without having to entertain people in front of a TV camera. Over time, I found that there is sometimes a period preceding a big weather event when the forecast may appear to be falling apart. The timing may be off, or you’re not really sure that the elements will come together in the manner you’ve forecast. Then, the outbreak begins. Overall, if the physics appears to indicate a large weather event, it ends up happening. Never in the exact way you forecast, and people on the fringes of the event may be angry that they got warned and nothing happened. However, it’s a lot worse when you think that things won’t quite get together to produce a major event, and then they do.
Looking at the geological records of past climate, I see many examples of things that did get worse, and fast, after periods of apparent calm. I am in no doubt that GHGs are a major player, especially CO2. When you look at the ice core records, you see that temperatures closely correspond to CO2 levels on the millennial scale, but not always on short scales. Often, the climate adjustments are very abrupt, even if the CO2 changes gradually. The most glaring example is the onset and then termination of the Younger Dryas. That affected weather and climate worldwide, and it would wreck modern agriculture if repeated now. That’s because we aren’t set up to grow new crops, or on a new schedule, if we had to make a switch from one year to the next. No, the industrialized countries wouldn’t starve, at least right away. However, you’d probably tire of a diet of leftover corn and soybeans that were intended for livestock feed. It might induce quite a lot of social unrest, too.
There was a discussion of earth system sensitivity to CO2 on this blog a while back. The takeaway was that empirically, GMST was more sensitive to CO2 changes in the 180-280 ppm range than what we’re seeing in the prompt response to CO2 now, or that our models suggest for the near future. Are you feeling lucky? Are we missing something that will save us? We’ve got one Earth to gamble with. Just because nothing really catastrophic on a large scale has been unleashed so far doesn’t mean it will continue to be that way. The physics suggests that we’re messing with a chaotic shift to a much different climate.
I’ll also mention SLR. While we can adapt with some inconvenience to rates well under 1 meter/century, it can get a lot higher than that – which apparently happened during the last warm interglacial period. After some period of stability, there was a fast rise to about 10m higher than our current sea level. “Fast” probably meant a period of centuries, but it also means that there is a mechanism out there for rapidly melting large chunks of continental icecap during an interglacial. This happened at a different Milankovitch phase, but at a CO2 level no higher than 300 ppm. We’re messing with that threshold, too.
in Re to John Pollack, 6 Feb 2026 at 11:09 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844807
Hallo John,
You might have missed my occasional complaints that the dispute between proponents of the “standard” climate models with moderate climate sensitivity towards changes in radiative forcing of non-condensing greenhouse gases on one side and proponents of the “hot” models with high climate sensitivity on the other side does not consider the possibility that climate sensitivity might have changed during the Holocene.
It appears that both sides assume that climate sensitivity was basically constant, although its change from moderate (that seems to fit with paleoclimate data) towards high (that seems to fit with the present pace of the climate change) could potentially resolve their dispute.
Nevertheless, if there is some evidence for an opposite sensitivity change that you mention in your penultimate paragraph (from a high sensitivity in early Holocene to low sensitivity later, including the industrial era?), it could, of course, make the above mentioned dispute even less understandable.
Can you perhaps remember when / by whom this evidence has been presented herein, or directly cite a source from that I could learn more?
Greetings
Tomáš
in Re to Rory Allen, 3 Feb 2026 at 7:16 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844710
Hallo Rory,
As a non-scientist, I am somewhat surprised by the certainty expressed in your sentence “The only mechanism which can account for this is the greenhouse effect of the extra CO2.”
Although I am merely a patent engineer and industrial chemist, I would, personally, rather prefer a more cautious formulation – something like “We have not found any other plausible explanation for all this yet than increased greenhouse effect caused by anthropogenic emissions of non-condensing greenhouse gases (GHG) accumulated in Earth atmosphere”.
Regular Real Climate commenters may be aware that I several times asked if someone knows how climate sensitivity to radiative forcing of non-condensing greenhouse gases depends on terrestrial hydrological regime of Earth continents. As it slowly became clear that nobody knows the answer yet, I think that the possibility that the greenhouse effect of anthropogenic GHG emissions might have had a comparably plausible “accomplice” in anthropogenic interferences with land hydrological regimes is still open.
With respect to comments to your post by Susan Anderson, Steven Emmerson, Radge Havers, Nigel Jones and Barton Paul Levenson who mutually agree that the only reason why people can doubt about provided reasoning for proposed climate change mitigation policies is the skilful propaganda (spread and/or promoted by fossil fuel industry, politicians and/or states dependent thereon), I would like to note that my question regarding climate sensitivity can be hardly detected in this propaganda (yet).
I admit that I sometimes regret the disinterest of “climate sceptics’” or “climate denialists” about this argument that is, in my opinion, still honest and plausible. They still seem unaware that it might provide their doubts about “climate change mitigation through economy decarbonization at any costs” with legitimacy, although their deepest motives may be different. A possible reason can be that they, in fact, do not need / seek honest arguments. Another possible reason can be that Real Climate discussion fora may have a negligible attention / reach in public disputes about climate policies.
One could hope that if the “sceptics” started using my question as their argument massively, it could perhaps become an incentive that (at least some) climate scientists turn their attention to this yet unresolved problem and start exploring it. I am, however, unsure about this positive outcome – I can well imagine also an opposite. This is why I still seek the ways how to achieve that this question is raised rather by citizens (and, ideally, also politicians) concerned by climate change and supporting climate research than by their political opponents.
Greetings
Tomáš
Tomáš, funny, but never mind. I asked chatgpt your question. Here’s its answer (not verifying it myself). You can try asking AI these questions and see what it says.
Climate sensitivity and radiative forcing
Climate sensitivity generally refers to the global mean temperature response (ΔT) to a given radiative forcing (F), typically expressed as:
ΔT = λ F
where λ is the climate sensitivity parameter (K per W/m²).
– Radiative forcing of non-condensing greenhouse gases (GHGs): These are gases like CO₂, CH₄, N₂O that do not condense at Earth’s surface temperatures. Their forcing is relatively well-mixed and long-lived, and they provide a baseline warming signal.
– Non-condensing GHGs are critical because without them, water vapor (a condensing greenhouse gas) feedback would not operate strongly—the hydrological cycle and water vapor feedback amplify the warming initiated by these gases.
—
Terrestrial hydrological regime
The hydrological regime of continents refers to the spatial and temporal patterns of precipitation, soil moisture, evapotranspiration, and runoff. It is controlled by factors like:
– Geography (mountains, continents vs oceans)
– Vegetation cover
– Soil moisture retention
– Climate zone (humid tropics vs arid deserts)
This regime strongly influences surface-atmosphere interactions, which are key to climate feedbacks, especially:
1. Water vapor feedback
– Warming induced by non-condensing GHGs increases evaporation.
– If continental regions are wet (humid soil, dense vegetation), more water vapor enters the atmosphere.
– Water vapor is a potent greenhouse gas, so humid regions amplify GHG forcing more strongly.
2. Surface albedo feedback
– Arid or sparsely vegetated regions reflect more sunlight (high albedo).
– If warming dries soil or melts snow, albedo decreases, enhancing warming.
– Wet regions with dense vegetation have lower albedo changes, so this feedback is smaller.
3. Cloud feedbacks
– Evapotranspiration influences cloud formation.
– More clouds can either cool (reflect sunlight) or warm (trap longwave radiation), depending on cloud type and altitude.
– The net effect of clouds varies with the continental hydrological regime.
4. Land-atmosphere coupling
– Continental regions with strong soil moisture-atmosphere coupling show stronger amplification of GHG forcing, because soil moisture affects evaporation, boundary-layer humidity, and convection.
—
How climate sensitivity depends on hydrological regime
– Humid/wet continental regimes (tropical forests, wetlands):
– Stronger water vapor feedback
– Potentially stronger cloud feedbacks
– Higher local climate sensitivity to non-condensing GHGs
– Arid/dry continental regimes (deserts, semi-arid regions):
– Weak water vapor feedback (little moisture to evaporate)
– Soil dries quickly, limiting latent heat flux
– Stronger sensible heat flux dominates
– Lower local amplification of GHG forcing
– Seasonally variable regimes (monsoon or Mediterranean climates):
– Sensitivity is strongly seasonal; peak warming occurs during wet season due to latent heat fluxes.
—
Summary
Climate sensitivity to non-condensing GHG forcing is not uniform across continents—it is amplified in regions with strong hydrological cycling (wet soils, dense vegetation) and reduced in arid regions. This spatial heterogeneity affects not only local warming but also global climate sensitivity through coupled atmospheric circulation.
“””
Tomáš, ”Regular Real Climate commenters may be aware that I several times asked if someone knows how climate sensitivity to radiative forcing of non-condensing greenhouse gases depends on terrestrial hydrological regime of Earth continents. As it slowly became clear that nobody knows the answer yet …
AI says, “This is an interesting and technically valid question. The interaction between land-surface hydrology, soil moisture, vegetation, and climate sensitivity is an active area of research. …[it’s] indeed uncertain. …[It] is an area that is still being studied.”
So it sounds like nobody knows the answers to these narrow specific questions yet.
—
“Defects in the argument:
1. Overstates ignorance
Claiming “nobody knows the answer yet” is incorrect. The influence of land hydrology on climate sensitivity is partially understood and explicitly represented in climate models, even if uncertainties remain.
2. Overstates plausibility
Describing anthropogenic land-hydrological interference as a “comparably plausible accomplice” to CO₂ is unjustified. Its global forcing and explanatory power are far smaller than those of non-condensing greenhouse gases.
3. Conflates uncertainty with explanatory equivalence
The existence of unresolved details in feedbacks does not imply alternative mechanisms of comparable importance to CO₂ forcing.
4. Implied policy relevance is weak
Uncertainty in hydrological modulation does not undermine the physical basis for decarbonization or climate mitigation policies.
5. Ambiguous scope (regional vs global)
Land hydrology strongly affects regional climate responses but cannot account for the observed global warming signatures attributed to greenhouse gases.”
Note: I’m not arguing this. This is AI talking. I’m too dumb to go here. Just thought I’d pass it along. ;D
Any variation which is not captured by various parameters in any statistical or physical model ends up in the error term. WRT sensititivity, it might be nice to have smaller error bars in various estimates, but the important point to realize is any remaining variables even if fully explained at some later point arise mainly out of the variation in the error term, NOT the explained variance. As things stand, the basic observation statistically is that the standard error in the linear trend in the various temp series (aggregated annually) is on the order of .1x the standard error of the error term.
Any variables left in the error terms will only modulate the shape of the linear trend, they will not affect the basic trend in rise over time in the least.
in Re to Ron R., 5 Feb 2026 at 7:36 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844784 ,
5 Feb 2026 at 10:29 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844788 ,
and jgnfld, 7 Feb 2026 at 10:20 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844813
Hallo Ron,
I think that your test of the AI capability in this field might serve as a good support for Piotr’s doubts about reliability of LLMs as information sources.
1) I have not heard yet about “local” or “regional” climate sensitivity towards radiative forcing of non-condensing greenhouse gases. As, on one hand, this forcing has a globally uniform character while, on the other hand, local or regional climates strongly depend on other (often quite remote) areas on the globe, I somewhat doubt if studies determining such local or regional climate sensitivities could make any sense.
Has your engine cited any, and if so, have you checked the provided reference(s)?
2) “it (GHG forcing) is amplified in regions with strong hydrological cycling (wet soils, dense vegetation) and reduced in arid regions”
I am afraid that herein, the engine merely parrots the superficial (and substantially incorrect) assertion circulating in the public space.
Have you asked your source if the climate of tropical rainforests is indeed warmer than the climate of hot deserts and/or if tropical rainforests warm quicker than polar regions that exhibit much lower absolute air humidity (water vapour that allegedly amplifies the “local climate sensitivity” to GHG radiative forcing ..)?
3) “(land hydrology) global forcing .. far smaller than .. of non-condensing greenhouse gases”
Have you opposed your source that he/it conflates forcing with sensitivity to forcing?
4) “The influence of land hydrology on climate sensitivity is partially understood and explicitly represented in climate models, even if uncertainties remain.”
In this point, I can perfectly agree with the engine. My question, however, read: Does someone know how (much) will the climate sensitivity to GHG forcing change if we change (some parameter suitably characterizing) land hydrological regimes?
Have you asked your source if he/it could list the publications offering a quantitative insight?
I do not disprove that this, in my opinion potentially quite important information, may be implicitly comprised in available global circulation climate models. “Nobody knows” addresses the strange circumstance that, as far as I know, no one tried to extract this information in an applicable form yet, so that general public and policy makers could deal therewith.
Dear jgnfld,
I am not aware of any global reconstruction of past hydrological regimes. In this respect, I can hardly imagine how estimations in which extent climate sensitivity to GHG forcing depends on land hydrology could be derived from such missing observational data. I am not sure if I understand you correctly, but if you mean variance analysis made e.g. by Barton Paul Levenson, he confirmed that the only parameter characterizing land hydrology he could find was the some irrigation data.
I think that assuming that this parameter might be suitable for characterizing land hydrological regimes and temporal changes thereof globally could be too bold. In the discussion with BPL, I therefore concluded that his analysis has not comprised any parameter properly representing land hydrological regimes – simply because he could not find any. Are you aware of another variance analysis that does not suffer from this deficiency?
Greetings
Tomáš
Tomáš, being a ‘forest for the trees’ kind of guy I’ll take your word for it ;)
I asked the AI your questions and here are its answers, take them as you will.
—-
“Assessment of RealClimate Comment on Local Climate Sensitivity and Land Hydrology Effects
Overview:
The comment raises doubts about:
1. Local/regional climate sensitivity to well-mixed greenhouse gases (GHGs).
2. Regional amplification/reduction of warming based on hydrology.
3. Confusion between radiative forcing and climate sensitivity.
4. Lack of quantitative studies linking land hydrology to climate sensitivity.
Point-by-Point Analysis:
1) Local/Regional Climate Sensitivity
– Correct that GHG forcing is globally uniform.
– Regional responses exist due to feedbacks (albedo, clouds, vegetation, ocean).
– Local “climate sensitivity” is sometimes quantified in “regional feedback” studies or pattern scaling.
– References:
– Flato et al., 2013, IPCC AR5 Section 9.4.3: Regional climate feedbacks, polar amplification, monsoon response.
– Watterson, I.G., 2000. “Simulations of regional climate change in response to global warming.” Clim. Dyn. 16: 65–78.
2) Amplification in Wet Regions vs. Arid Regions
– Oversimplified claim: “GHG forcing is amplified in wet regions and reduced in arid regions.”
– Reality:
– Arctic warms fastest despite low humidity (polar amplification).
– Tropical rainforests have high humidity, but warming is moderated by cloud and convective feedbacks.
– Dry deserts may warm strongly due to low heat capacity of soil, but feedbacks are complex.
– References:
– IPCC AR6 WG1, 2021, Section 5.3: Land-atmosphere feedbacks and differential warming.
– Pithan & Mauritsen, 2014, Nat. Geosci., 7: 181–184: Arctic amplification dominated by lapse rate and albedo feedbacks, not water vapor.
– Good et al., 2015, J. Climate, 28: 7205–7220: Regional climate sensitivity studies.
3) Conflation of Forcing and Sensitivity
– Forcing = imposed radiative change (W/m²)
– Sensitivity = temperature response per unit forcing (°C per W/m²)
– Land hydrology affects feedbacks (sensitivity), not direct radiative forcing.
– References:
– Koster et al., 2004, J. Hydromet., 5: 858–875: Soil moisture strongly affects precipitation-temperature coupling.
– Seneviratne et al., 2010, Nature 467: 951–956: Soil moisture controls regional climate extremes.
4) Quantitative Impact of Land Hydrology on Climate Sensitivity
– Largely unknown at global scale.
– Regional effects exist (soil moisture feedbacks influencing heatwaves, precipitation).
– No widely accepted number for change in global climate sensitivity due to land hydrology.
– References:
– Douville et al., 2000, J. Climate 13: 3066–3084: Soil moisture impact on seasonal climate.
– Dirmeyer et al., 2012, J. Hydromet., 13: 335–352: Land-atmosphere interactions in climate models.
5) Global Reconstructions of Past Hydrology
– Past hydrological regime reconstructions are limited.
– Some soil moisture reconstructions exist, but global high-resolution coverage is sparse.
– References:
– Sheffield et al., 2012, PNAS 109(22): 8842–8847: Long-term soil moisture reconstructions.
– Trenberth et al., 2014, J. Climate, 27: 8330–8343: Global hydrological variability assessment.
Overall Assessment:
– The comment is scientifically cautious and mostly correct.
– Highlights oversimplifications and uncertainties in AI-generated or public statements.
– Correctly distinguishes forcing from sensitivity, and emphasizes the limited quantitative understanding of land hydrology effects on global climate sensitivity.
Key References:
1. IPCC AR6 WG1, 2021
2. Pithan & Mauritsen, 2014, Nat. Geosci. 7: 181–184
3. Seneviratne et al., 2010, Nature 467: 951–956
4. Koster et al., 2004, J. Hydromet. 5: 858–875
5. Douville et al., 2000, J. Climate 13: 3066–3084
6. Dirmeyer et al., 2012, J. Hydromet. 13: 335–352
7. Sheffield et al., 2012, PNAS 109(22): 8842–8847”
At this point I’ll bow out on this particular point.
Tomáš, here’s a thought, if you haven’t already why not write a paper and have your ideas peer reviewed?
In general it is not very well known, and probably counter-intuitive to most people nowadays, that aridity is associated with warmer temperatures. This goes against the grain of GHG forcing–feedback centric teaching of environmental change, and the habit of attributing local extremes to GMST variation in venues such as WWA. https://www.worldweatherattribution.org/
Recall that Radiative controls by clouds and thermodynamics shape surface temperatures and turbulent fluxes over land: https://www.pnas.org/doi/10.1073/pnas.2220400120
“we find an increase in land surface temperatures with aridity, indicating that drier regions are generally warmer than humid regions”.
It’s possible the training of certain AI type systems, much like many humans, has been over-saturated with climate political advocacy in the context of earth system science and models of change, resulting in confusing realclimatescience with a CO2 climate science, or carbon cycle science more generally. The reasons for the condition showing up both with real humans and in AI type system is probably exactly the same considering each is exposed to similar material.
Given that radiative controls by clouds and thermodynamics appear to shape temperatures almost everywhere, and that the observed increase in Earth’s energy imbalance does not come from LW changes but instead from decreased SW reflection, it’s worth amending views more broadly in respect to what’s actually going on and teaching how it works, including carbon cycling disturbance of course. https://journals.ametsoc.org/view/journals/clim/36/12/JCLI-D-22-0555.1.xml.
cheers
in Re to Ron R., 8 Feb 2026 at 3:48 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844862
and 8 Feb 2026 at 4:17 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844863
Hallo Ron,
Thank you very much for your feedback and your suggestion!
Unfortunately, two references provided by your LLM that looked particularly interesting, because they allegedly deal with global reconstructions of past hydrology regime and/or soil moisture, namely
– Sheffield et al., 2012, PNAS 109(22): 8842–8847: Long-term soil moisture reconstructions.
– Trenberth et al., 2014, J. Climate, 27: 8330–8343: Global hydrological variability assessment.
do not seem to exist.
It appears that your LLM is not reliable / helpful as regards climate science (and, possibly, in any specific scientific or technical topic, I am afraid). If you want, you can try to check other references offered, to assess the reached success score more accurately.
As regards summarizing my questions in form of a scientific article, I am afraid that it may not work. Scientific journals primarily focus on publishing research results, not ideas, questions or proposals. Moreover, I am not a researcher and do not possess any education and/or training in climate science.
Nevertheless, I suppose that the same questions I am asking in this forum may climate scientists hear from their students. I believe the answers provided by professors could be sometimes instructive for the interested broader public as well.
I have not seen any contribution of this kind in this forum yet, however, perhaps our present exchange could inspire the moderators. I believe that it might be very interesting if they occasionally shared their teaching experience. The Real Climate website is perfectly suited for this purpose, I think.
Greetings
Tomáš
JCM: In general it is not very well known, and probably counter-intuitive to most people nowadays, that aridity is associated with warmer temperatures. This goes against the grain of GHG forcing–feedback centric teaching of environmental change, and the habit of attributing local extremes to GMST variation in venues such as WWA
JP: What are you talking about? To whom does “most people” refer to? Random fans at sporting events? Heartland Institute attendees? Far from being counter-intuitive, the fact that aridity is associated with warmer temperatures is so well-known to meteorologists that it scarcely needs mentioning. (I suspect that even most uninterested people are somewhat aware that deserts can get really hot, and that their local heat waves are often accompanied by a drought.)
Nor is this hidden from the scientific community at large. For example, in the Jan. 29 issue of Nature, under the heading of Ecology, is the article entitled “Hot droughts in the Amazon provide a window to a future hypertropical climate.”
JCM: It’s possible the training of certain AI type systems, much like many humans, has been over-saturated with climate political advocacy in the context of earth system science and models of change, resulting in confusing realclimatescience with a CO2 climate science.
JP: It’s always possible, even likely, that an AI system has been trained on whatever bunk is floating around the internet, or is insufficiently capable of weighing the writings of people who do actual research against the larger amount of dross in everything from poorly informed public opinion to stuff sponsored by the FF industry and published in predatory journals.
That’s why I give little time or attention to AI generated writing.
Tomáš, hello. First my apologies on those two bogus references. AI steered me wrong. I asked and it apologized too but says that there are plenty of other similar references available, however. The limitations of current AI.
Just wondering if you can clear this up for me and others though: do you believe that cessation as soon as possible in the global use of FF (and transition to clean energy alternatives) is absolutely warranted and non-negotiable as the single most important mitigatory measure we can take in the face of rising heat OR do you think that we can continue using them (perhaps into perpetuity?) if we also increase evaporation? Please be as clear and concise as possible. Thanks much.
Thank you for the input John Pollack. Since you are a meteorologist, the physical process behind temperature excursions are familiar territory. On this climate blog, my reference to “warmer temperatures” was intended in terms of climates. Framed this way, the statement that aridity is associated with warmer climates than they otherwise might be is consistent with observed patterns. While not controversial among meteorologists, this point can encounter serious misunderstanding or resistance on a climate blog page, and in some cases it triggers a relentless deranged trolling behavior.
in term of your hypertropical reference, I do not have article access but I will remark the following:
Questions are often presented to answer – “how does climate change impact ecosystems?” – and so on, as if various Earth systems are simply passive recipients at the mercy of LW radiative transfer. In the case of the tropical forests, it is often overlooked that unnatural extremes and bona fide climatological disruptions are attributable to direct ecological obliteration at the hands of man. While this may seem intuitive, it is often overshadowed in the way academic questions are framed. Comparatively, traditional knowledge systems have long emphasized such connections.
Today, there is serious research impetus towards “how does climate change impact stuff”, and much less about how does ecohydrological destruction impact physical climate process. There is some interest on ecology in relation to the biogeochemical accounting, however, in order to close carbon cycle budgets. Certainly, in the circumstances of this page, as a general rule of thumb it is a social-epistemic rule that in order to discuss concepts of energy, fluxes of heat and mass, dynamics, clouds, biogeophysical process, etc. in the context of climate stability and change, one must explicitly affirm carbon cycle dominance early and often if wanting to have any opportunity for interesting dialogue at all.
Sadly very few voices remain today holding up the residual flame, one which symbolizes the synergies inherent to international conventions on climate, biodiversity, and desertification – especially within the “climate” scicomm silo today. Thankfully, such processes are more explicitly recognized in other scientific and less fashionable policy domains.
https://impact.economist.com/energy-environment/at-the-precipice-transforming-land-management-to-secure-our-future-at-cop16
in Re to Ron R., 9 Feb 2026 at 6:18 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844902
Hallo Ron,
Thank you very much for your feedback.
No needs to apologize, trial and error can provide a valuable experience. I appreciate your contribution to the debate in its entirety.
As regards your question(s), I will try to summarize my view (corresponding to my present knowledge) as follows.
1) So far, anthropogenic emissions of non-condensing greenhouse gases (GHG) are a single reliably identified cause of the presently observed global warming.
2) Although existing models do not enable any reliable projections of future regional climates (e.g., for the central Europe, we do not know whether we have to expect something like Northern Africa today, or, in case of a collapse of oceanic circulation in Atlantic (AMOC), rather a dramatic cooling), the risk of dramatic climate changes in case of further extensive fossil fuel use will be the higher the longer we will continue therewith.
3) The unpredictable changes in regional climates have also hardly predictable consequences for affected states / communities. It can be reasonably expected that many of these consequences may be undesirable, and no one can exclude that some may become intractable.
4) Under such circumstances, it is my view that ignoring these risks would have been irresponsible.
5) As the anthropogenic GHG emissions undoubtedly contribute to the observed climate change, restricting them as quickly as possible is in my opinion logically the first step towards mitigation of these risks.
6) Like in any other case of preventive measures dealing with an unknown risk, it is difficult to assess which costs of the risk mitigation are still justified and which measures may perhaps cause more harm than good.
7) Replacing certain part of the present fossil fuel use with energy from renewable energy sources is already without substantial economical penalty and in absence of other (political, legal, administrative, regulatory ..) hurdles seems to run basically spontaneously.
8) I do not think, however, that subsidizing this replacement wherever it is not economically feasible yet is a good idea.
9) I think that we should rather strive supporting this spontaneous process indirectly, by resolving remaining obstacles and issues. For example, a clever integration of photovoltaic panels into other (industrial, urban, transportation, agricultural …) infrastructure can significantly reduce creation of artificial “urban solar deserts” and prevent disruptions in hydrological regime, damages to ecosystems and/or deterioration of the landscape and of the public living space.
10) With respect to land hydrology, I am not sure that we might be capable to fix the present Earth energy imbalance by an artificial intensification of terrestrial water cycle. I suppose that this idea may deserve theoretical research, at least as a possible side effect of geoengineering methods for carbon dioxide removal (CDR) from the atmosphere by desert afforestation, although an old NASA study
https://www.researchgate.net/publication/225114822_Irrigated_afforestation_of_the_Sahara_and_Australian_Outback_to_end_global_warming
seems to assume that the possibility of a global effect (p.420) is reduced to possible influence of changes in (local) cloudiness on Earth albedo.
11) On the other hand, anthropogenic disruptions to the terrestrial water cycle have not been excluded yet as a contributing cause of the recent climate change, by making the Earth climate more vulnerable towards GHG radiative forcing than in its previous state with prevailing natural ecosystems.
12) We should keep this insufficiency of our understanding to the Earth climate system in mind and strive to fill the gaps step-by-step.
13) As the substantial gaps in our knowledge may anyway persist, most efficient condition for avoiding measures that may be in fact ineffective or even harmful might be investing primarily into general resilience and actionability of our society.
14) I therefore see as one of key prerequisites for successful mitigation of the present climate change (or any other general challenge for the mankind) exploiting human curiosity and creativity, rather than relying on predetermined strategies.
15) I believe that human curiosity and creativity thrive in societies protecting and fostering freedom of thinking in combination with mutual respect and suffers in societies based on bullying and subjugation.
16) I think that for finding a feasible way towards successful climate change mitigation may be finally decisive how we will cope with foreign states based on bullying and subjugation and with forces trying to reshape our societies on these principles from the inside.
17) In this respect, I think that any specific measure for climate change mitigation, including the optimal schedule for fossil fuel use reduction, may be negotiable if we preserve our eligibility and capability to negotiate.
I am not sure if I managed to be clear enough, although I did my best. I am aware that I failed to be concise; I am sorry for that.
Greetings
Tomáš
Not exactly. The rising trend is there in the data. ANYONE with even no scientific or statistical training can see there is a trend. Anthropogenic CO2 is a clear contributor to that trend and explains a good portion of the variation seen in the data.
Any additional variables like your hydrology el Nino/la Nina cycles, etc. are simply going to explain the wobbles wobbles about the linear (or quadratic in some more recent studies as more data becomes available over time) trend, NOT the basic trend itself.
Tomáš, a simple yes or no would have sufficed :D
Ok, seriously, a few points.
8) I do not think, however, that subsidizing this replacement wherever it is not economically feasible yet is a good idea.
We have the money to subsidize renewable energies, The world’s spending on redundant military crap is $2.7 trillion. The US spending by itself is more than $1 trillion. Do we really really need yet more bombs and bombers??? Where are we going to put them all? How many times over do we need the ability to destroy the world? Or are we just keeping an industry alive?
5) As the anthropogenic GHG emissions undoubtedly contribute to the observed climate change, restricting them as quickly as possible is in my opinion logically the first step towards mitigation of these risks.
Mitigation and,
13) As the substantial gaps in our knowledge may anyway persist, most efficient condition for avoiding measures that may be in fact ineffective or even harmful might be investing primarily into general resilience and actionability of our society.
Adaptation. I agree, with the caveat of the word, “primarily”. It should be a side-by-side effort. After all, the adaptation measures we are entertaining are only for us humans. And they also rely on the presumption that the electricity to power them will always be.
11) On the other hand, anthropogenic disruptions to the terrestrial water cycle have not been excluded yet as a contributing cause of the recent climate change, by making the Earth climate more vulnerable towards GHG radiative forcing than in its previous state with prevailing natural ecosystems.
Before we water the deserts that are created by logging, let’s try to replant those. Replanting would be so sooo much easier then trying to water the desert. Bedside that, (if I’m reading you correctly) you are aware that there’s been lots of times in the earth’s past when, due to heat, the global water levels were very high. Even the deserts were lush. Even the Sahara was once fertile. Still, the earth was hot. Sometimes much hotter then today. For example,
https://cdn.britannica.com/44/136144-004-76AAB173/Distribution-landmasses-regions-seas-ocean-basins-Cretaceous.jpg
“Sea level was higher during most of the Cretaceous than at any other time in Earth history, and it was a major factor influencing the paleogeography of the period … In general, the climate of the Cretaceous Period was much warmer than at present, perhaps the warmest on a worldwide basis than at any other time during the Phanerozoic Eon.”
With regard to the present day, maybe things would be even hotter if we didn’t water the deserts. I don’t know. But as you say in Point #1, FF are a “reliably identified cause of the presently observed global warming”. We do know that. So we should make reducing them of utmost importance. Anyway, besides these things I think we are in agreement (again if I’m reading you correctly).
Btw, sorry for going back on my word to bow out on this point.
in Re to jgnfld, 10 Feb 2026 at 6:10 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844948
Sir,
I would like to ask if you meant your post as a reply to my question of 8 Feb 2026 at 11:58 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844851 ,
If so, I would like to clarify that I have not disputed the significance of anthropogenic CO2 emissions as the cause of the present global climate change.
My question rather addressed variance analyses you have mentioned in your post of 7 Feb 2026 at 10:20 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844813
Furthermore, I must admit that I am somewhat confused also by your remark regarding El Nino / La Nina that do not seem to be anyhow linked to my question.
Have you perhaps replied to someone else?
Greetings
Tomáš
in Re to Ron R., 10 Feb 2026 at 9:55 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844953
Hallo Ron,
You may be correct that US military spendings are very incommensurate if the power created should serve for intimidating Denmark, US expansion to Greenland and/or keeping Venezuelan junta at power in exchange for their oil.
I think they might be justified if they served for deterring Russia and/or defending its neighbour countries like Ukraine against Russian attacks.
Greetings
Tomáš
I guess I’ve lost my stats prof edge. I was simply trying to note to you that further research on a variable that interests you or any other variable at all that is out there is simply not going to change the trend nor change the major contributors to that trend, The may well–and do in some cases–modulate the variations around the trend.
in Re to jgnfld, 12 Feb 2026 at 8:21 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844948
Sir,
It is my understanding that the trend as well as irregular fluctuations that may make this trend less obvious are both comprised in the observational temperature record which is given. It was further my understanding that the variance analysis mentioned by you is a statistical method that enables attribution of the observed trend (and/or fluctuations thereof) to chosen variables.
In this respect, I do not understand how the choice of variables can change the trend?
If this question resulted from my misinterpretation of your text, I would like to ask if you perhaps meant merely that if a variance analysis once provided certain attribution of the observed effects to certain variables, then a “new” variable (that has not been considered yet) could have only a minor importance, because all what “remains” to be explained thereby are the “residuals” not yet attributed to already “established” variables? Have I finally grasped your point?
If so, I assume that you posted your objection to make me aware that if a variance analysis once attributed a lion share (say 90 %) of the observed temperature trend to anthropogenic CO2 emissions, any other cause cannot respond for more than 10 % thereof. Am I still on track?
In this case, I would like to share a conclusion from a teaching about variance analysis provided by Perplexity Pro engine with respect to my question if this method could reveal that besides the anthropogenic emissions of greenhouse gases (GHG), further cause of the observed temperature rise might have been an increase in Earth climate sensitivity that has been caused by human activities different from GHG emissions.
My summary of the answer provided by the engine is:
– if the increase in climate sensitivity happened prior to the industrial era (on that the variance analysis likely focused), there is basically no chance that this statistical method could reveal this change in climate sensitivity as a factor that contributed to the warming observed during the industrial era
– if the increase in climate sensitivity due to anthropogenic “non-GHG” activities continued during the industrial era, it is still well possible that no one will recognize it in variance analyses, unless he or she will be seeking actively therefor when designing and performing the analysis.
I would like to desist from sharing the entire text provided by the engine herein.
Its answer resulted from two consecutive prompts that read as follows:
———-
1) In climate change discussions, I several times met with assertions that a “variance analysis” unambiguously revealed that rising atmospheric CO2 concentration is main cause of the observed climate change. See e.g. the discussion on Real Climate blog between Barton Paul Levenson and Tomáš Kalisz during the autumn 2025. The sceptical Tomáš objected that variance analysis may be incorrect if a variable having influence on the studied effect is omitted in the analysis and may not be applicable at all if there is a synergy between two or more variables, so that the effect cannot be interpreted as linear combination of mutually independent, additive contributions from specific variables. Were his doubts justified?
Recently, in another RC thread, [https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844813](https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844813) the participant jgnfld seems to assert again that if a variance analysis of an effect assuming a certain set of variables provides certain attribution of the effect to individual variables, adding a new variable cannot change the result of the original variance analysis, because the effect of the new variable must be in the “error term”.
I assume that he could be perhaps correct in case of mutually independent variables, could he? I am, however, afraid that in case of variables that may act synergistically, variance analysis remains inapplicable. Could you teach me where I am correct / where I am not?
2) The background of my original question was a yet untested hypothesis assuming that Earth climate sensitivity towards so called radiative forcings may depend on land hydrological regimes. This hypothesis or speculation is based on modelling experiment published by Lague et al [https://iopscience.iop.org/article/10.1088/1748-9326/acdbe1](https://iopscience.iop.org/article/10.1088/1748-9326/acdbe1)
They found out that in their CESM model, water availability for evaporation from the land influences the steady state (“equilibrium”) global mean surface temperature.
My further questions in this respect are:
a) If anthropogenic interferences with terrestrial hydrological regimes during the entire Holocene till the start of the industrial era about the year 1750 (such as land deforestation, urbanization, soil erosion / soil depletion in organic matter, landscape drainage, etc.) perhaps made Earth climate more vulnerable towards radiative forcings, could we extract from various attribution analyses that you describe any hint towards this co-factor, that might be in certain sense interpreted as a further cause of the present climate change, independent from anthropogenic emissions of non-condensing greenhouse gases, even though the climate sensitivity might have not further significantly changed during the industrial era anymore?
b) It could be, however, reasonably expected that some of the above mentioned interferences continued also during the industrial era and some (e.g. ground water pumping) might have become particularly apparent specifically in last few decades. Let us assume that they might have caused a further measurable shift in Earth climate sensitivity during the industrial era, or, maybe, even during the last several decades. Should, in such a case, perhaps at least the most sophisticated attribution analyses suggest that there may be such a yet unidentified “additional cause” contributing to the observed climate trends, or it might be still possible that such “additional cause” would have anyway remained hidden / undetected?
——————–
I think that you, as a professional statistician, will be capable to assess much better than me if the conclusions provided by the engine (you can try to test my prompts yourself) are correct.
Greetings
Tomáš
Tomas Kalisz: “ I am somewhat surprised by the certainty expressed in your sentence “The only mechanism which can account for this is the greenhouse effect of the extra CO2.”
I am somewhat surprised by your attempt that dead corpse of evaporation AGAIN.
TK “ I would, personally, rather prefer a more cautious formulation – something like “We have not found any other plausible explanation for all this yet than increased greenhouse effect caused by anthropogenic emissions of non-condensing greenhouse gases (GHG) accumulated in Earth atmosphere”
So your idea of making Rory’s sentence sound … less certain is to … replace it with a 3 times longer text, full of unnecessary jargon: “ emissions of non-condensing greenhouse gases (GHG) accumulated in Earth atmosphere” instead “extra Co2″… and not saying anything new? In what rhetorical school this counts as .. an improvement?
But of course the real motive of you and your guru JCM – is to question the credibility of the climate science and to divert the attention (and resources) from the mitigation of GHGs toward your and JCM’s idée fixe – increasing evaporation. The goal perfectly encapsulated by your JCM who blamed: “ up to 40% of the planet’s land degraded ” on the climate modellers “ artificial fixation and overemphasis on the outputs of trace gas“.
Hi did it even though we have shown him, using his OWN SOURCE (Lague et al.)! – that we can’t mitigate climate by increasing evaporation – even such an EXTREME measure as abandoning of all agriculture (converting all croplands on Earth to swamps) – would have reduced the AGW by a mere … fraction of 0.3 C, if any at all – since without agriculture we would no longer need the current major source of human evaporation – irrigation.
TK: Regular Real Climate commenters may be aware that I several times asked if someone knows how climate sensitivity to radiative forcing of non-condensing greenhouse gases depends on terrestrial hydrological regime of Earth continents. ”
And regular Real Climate commenters replied to you several dozen times – that it is inconsequential to the mitigation of AGW:
– for climate change only things that change matter – CH4 tripled, CO2 increased by 50%, water vapour increased (passively as a result of T increase by GHGs) by … several %
– climate change is not driven by changes hydrological regimes, but by the emissions of CO2 and other GHGs,
– water cycle is not a driver but a mere feedback – it amplifies consequences of our action, or inaction, on the mitigation of GHGs
– given the volume of water cycle and the very short residence time of water vapour in atmosphere (a week vs. dozens of years for CO2), and the limited amount of untapped ground water for extra irrigation – WE CAN’T DO ANYTHING to increase evaporation ENOUGH to make a dent in AGW
– your only practical idea how to do so (irrigation of Sahara) has been shown to be a monstrous fallacy
And that’s the reason why scientists do not respond to your numbingly repetitive demands that they take the time from their research and study your and JCM’s ideas, i.e. something that is inconsequential to the mitigation of AGW (for the reason listed above), and something not needed to make our model projections better – evaporation is ALREADY implicitely represented in climate models – so neither you nor JCM offer any new insight to improve the models.
So, go away or I shall taunt you a second time. (and believe me, you don’t want an escalation past that).
in Re to Piotr, 9 Feb 2026 at 12:18 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844867
Hallo Piotr,
My repeated question regarding climate sensitivity towards water availability for evaporation from the land encompasses exactly your point that water cycle intensity acts as feedback to various radiative forcings.
This feedback is not limited to water vapour concentration in the atmosphere – this would have been an improper simplification. Improper and unfortunate, because it could raise (in my opinion false) belief that (water cycle intensity) any times amplifies the effect of the radiative forcing.
In fact, the water cycle feedback includes also changes in latent heat flux from the surface that have an opposite effect than increasing water vapour concentration in the air, and changes in clouds that are complex and may have different outcomes, depending on further circumstances.
I therefore do not dare to predict if Earth with dry continents will have lower climate sensitivity towards doubling of atmospheric CO2 concentration in comparison with the Earth having wet continents, or oppositely. I only guess that we should learn it.
I see the reason why this (yet missing) knowledge may matter in the circumstance that we have also no idea yet how much human activities during Holocene changed hydrological regimes on Earth continents and thus (possibly) Earth vulnerability to anthropogenic emissions of non-condensing greenhouse gases (GHG).
I think that both learnings may be crucial for defining climate change mitigation policies adequately. If a change of Earth climate sensitivity caused by anthropogenic change in continental hydrological regimes perhaps prepared the stage for anthropogenic global warming, could we still rely on the assumption that fixing merely the direct cause (anthropogenic GHG emissions) will result in the desired stabilization of Earth climate, in a shape favourable for human civilization and enabling the desired ecosystem preservation and/or restoration?
Greetings
Tomáš
P.S.
With respect to climate models, I have not suggested an improvement thereof.
I proposed running them in a study dedicated to and designed for resolving my question, so that, on the basis of the respective physics already implemented therein, the models selected for the study provide a first quantitative estimation how strongly (or weakly) Earth climate sensitivity does depend on water availability for evaporation from the land.
Tomas Kalisz “ My repeated question regarding climate sensitivity towards water availability for evaporation from the land encompasses exactly your point that water cycle intensity acts as feedback to various radiative forcings.
And since you ignored my previous warming (“Go away or I shall taunt you a second time.”) let the taunt begin
You either still don’t get, or ignore (like the Monty Python’s Black Knight unable to admit defeat despite overwhelming evidence to the contrary) – your water feedback is INCONSEQUENTIAL to mitigation of AGW, because for that you need to reduce the drivers, the feedbacks would sheepishly follow.
Your claiming “missing knowledge” – plays straight into the hands of the “anything-but-GHGs” deniers, like your guru – JCM: who uses the water cycle in his attempt to discredit climate science, by blaming “up to 40% of the planet’s land degraded ” on the climate modellers “ artificial fixation and overemphasis on the outputs of trace gas“.
TK “ I proposed running them in a study dedicated to and designed for resolving my question ”
What for? We already have the effect of water cycle implicitely included in climate models in the most realistic way we can – so running ,,, hypothetical models of the future with some arbitrary hypothetical changes in the water cycle of the future is … not an improvement.
This may be the reason why you can’t find serious scientists (JCM does not count) spending time on your poorly defined, poorly constrained problem, with no practical implications and no new non-trivial insights at the end of it.
That and “ I fart in your general direction! Your mother was a hamster and your father smelt of elderberries!“
in Re to Piotr, 9 Feb 2026 at 8:04 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844904
Hallo Piotr,
I read your sentence “you need to reduce the drivers, the feedbacks would sheepishly follow” the way that you still believe that water cycle feedback any time simply amplifies the radiative forcing, am I right?
If so, I would like to repeat that I am not sure if this assumption is correct. I rather suppose that the sign of the feedback may depend on the share of latent heat flux change and/or on the character of the cloudiness change on the entire feedback, and that contributions of these particular feedbacks may change significantly just in dependence on water availability for evaporation from the land.
I agree that implicitly, all this may already be more or less correctly represented in existing climate models, yet it appears that no one turned his or her attention to the circumstance that it might be useful to know whether (and if so, under which hydrological regime) this sign can indeed change.
This change might have pertained to past hydrological regimes of many regions on the Earth that seem to have switched from a humid hydrological regime to arid one. I understand that it might look unimportant if we assume that global annual precipitations simply follow global mean surface temperature and that the same basically applies also for precipitation partition between the land and the sea, or even regionally, but I am afraid that these assumptions may be too bold.
If any of these assumptions is in fact invalid, it can have significant practical implications, because changes in precipitations may have more dramatical practical consequences than changes in temperature. It appears that till now, climate research has been very strongly focused on model match with temperature records. It is understandable because – as far as I know – there are no reliable global reconstructions of past precipitations yet that might serve for model calibration in parallel with past temperature reconstructions.
Perhaps someone familiar with CMIP comparative tests of climate models could explain if the observed spread in their temperature projections fits the observed spread in precipitation projections. I do not think that it must be necessarily the case. It could be definitely interesting to know, however, I drifted somewhat away from my point.
I try to express my concern that although water cycle physics may be properly represented in available climate models, no one tried to find out yet if Earth with arid continents may be more or less vulnerable to CO2 concentration changes in the atmosphere than with humid continents. Or, in another aspect of the same issue, to find out if hydrological regime changes caused by the climate change may amplify themselves, or rather attenuate. Or, in yet another aspect, to find out if we may have at least a theoretical chance that a change in our direct anthropogenic interactions with water cycle could milder or perhaps even reverse some undesired effects of the climate change thereon.
I still think that at least some of these various aspects may be of practical importance and that exploring them may be worth of effort. From a few climate modelers that I succeeded to approach with my questions, I have not obtained feedback that the assumed influence of land hydrological regimes on climate sensitivity is a nonsense but rather that they would have been willing to design and run the study encompassing this open problem if they had a commensurate funding therefor. So now, I rather deal with a question how to convince a funding agency that these questions may deserve a dedicated budget for a specific funding call, seeking research projects that will resolve them.
Greetings
Tomáš
Tomas Kalisz “ you still believe that water cycle feedback any time simply amplifies the radiative forcing, am I right?
Keep the “belief” for the description of you and your JCM. I don’t believe, I argue offering falsifiable quantitative arguments – here the volumes of water fluxes and water’s residence time in atm of several days instead of several decades of CO2, so nothing humans can do to increase the water cycle can have any significant effect on AGW – hence yes: ” water cycle is a passive feedback that at any time simply amplifies the radiative forcing”. Furthermore – patrick and I have used your and JCM own source – to disprove your and JCM claims. As explained to you several times already, e.g.:
========= Piotr Feb 11, 2025 ===================
TK: I would like to remind you that there is at least one publication (Lague et al, 2023) suggesting that it is not as simple as you describe
“I would like to remind you” ;-) No need – we do remember the eggs on your and JCM’s faces after you tried to use that Lague et al, 2023 in support of your “anything but GHGs” denialism and your attacks on the credibility of the mainstream climate science.
In fact, I and others have used the very paper you speak of – to quantify the laughably small “AGW reduction return on investment” of human interventions in ater cycle, namely:
– I have shown the idiocy of Tomas Kalisz’s modest proposal of spending … tens of trillions of dollars annually, for hundreds of years, to achieve, aa fraction of 0.3K reduction in AGW (your Sahara irrigation scheme – and even that assuming zero GHG emissions from building and operating for centuries …. millions of industrial desalination plants. The proposal so absurd that even your JCM was at pains to distance himself from you.
– patrick and I used Lague et al. numbers to show that even if we abandoned ALL AGRICULTURE – the cooling effect of the increased evaporation would be a fraction of a fraction of 0.3K, and that ignoring concurrent warming from the decreased land albedo (forest being darker than soil or crops). Unable to falsify this – your JCM … threw his own source ( Lague et al.) under the bus, tried to discredit Lague et al, and by extension, ALL climate modeling as: “ imaginary process mechanisms [using] rules about how things ought to be ” (c) JCM July 2024.
All of which was already explained to you – see my 20 Jul 2024 post
================= end of quote ==================
So to quote a classic – ” Go and boil your bottom, you son of a silly person. Fetchez la vache!”
in Re to Piotr, 12 Feb 2026 at 7:02 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845053
Hallo Piotr,
I read Lague 2023
https://iopscience.iop.org/article/10.1088/1748-9326/acdbe1
as a hint that water cycle may amplify or attenuate the effects of radiative forcings, depending on how much water for evaporation from the land is available.
To find out if it is indeed the case is the purpose of the dedicated modelling study for that I seek support.
I hope that the gained knowledge could help understand better how we can stabilize Earth climate in the future and/or how we could prevent the risk of its further destabilization.
Greetings
Tomáš
Tomáš Kalisz says
13 Feb 2026 at 12:36 PM
https://medcraveonline.com/IJBSBE/IJBSBE-09-00237.pdf
or SEE
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845106
and
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845059
PS all my genuine responses to your prior queries directed at myself were never published… not my fault.
Tomáš Kalisz 13 Feb 2026: Hallo Piotr, I read Lague 2023
… and as usually – understood nothing. As seen here:
TK: “ the purpose of the dedicated modelling study for that I seek support.
I hope that [the dedicated modelling study] could help understand better how we can stabilize Earth climate in the future and/or how we could prevent the risk of its further destabilization. ”
The already existing study you supposedly read (Lague et al.) showed already that we CAN’T increase evaporation enough “ to stabilize Earth climate in the future and/or prevent the risk of its further destabilization.”
If you already forgot:
Lague provided the maximum range of GMST changes: the difference between the Earth covered with deserts vs. the Earth covered with swamps.” = 8C. The problem is that any _realistic_ change in the evaporation would be but a tiny fraction of these 8C:
1) patrick used Lague to show that even such a massive civilizational change as abandoning of all agriculture would have resulted in changes in the water cycle that would have cooled the Earth by AT MOST 0.3C.
2) I have built on that – showing that it is already too generous = that under more realistic assumptions it would be … a fraction of a fraction of 0.3C. Furthermore, this would have to be maintained …forever (when you stop, the cooling effect ends as well).
3) next, I have used the same Lague’s paper – to estimate the costs of your Sahara irrigation scheme – after building … millions of the industrial desalination plants and spreading the desalinated water over 5mln2 , at the costs of many trillions of dollars ANNUALLY, after 100s of years we would achieve at most of 0.3C reduction in AGW – at most – because any emissions of GHGs during the building and then operating over centuries the millions of desalination plants and pumping the water over 5mln km2 – would have reduced or reverse any cooling from evaporation.
So as you see – back-of-the-envelope calculation using an already existing study shows conclusively that we simply can’t increase evaporation enough to make any noticeable reduction in AGW. It’s a DEAD END.
So, if you understood the implications of the paper your read, or even understood the dozens of posts explaining these implications to you – you would have known that we don’t need “dedicated modelling study to understand better how we can stabilize Earth climate [with increases in evaporation] ” – because we already know the answer – we CAN’T.
And that’s why people don’t waste their time to run a model that will just confirm what they already know from the back-of-the-envelope calculations.
in Re to Piotr, 16 Feb 2026 at 9:40 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845304
Hallo Piotr,
I do not see how your explanations answer the question that the proposed study should resolve, namely how water availability for evaporation from the land influences climate sensitivity towards radiative forcings.
If you think that Lague 2023 already provided this information, could you tell me which of their two extreme examples should exhibit higher transient climate response (TCR) and/or higher equilibrium climate sensitivity (ECS) towards doubling of atmospheric CO2 concentration?
I am incapable to derive it from their article myself.
Greetings
Tomáš
Tomas Kalisz “ I do not see how your explanations answer the question that the proposed study should resolve, ”
If you can’t comprehend, or admit the meaning, of such an uncomplicated argument as:
P: “The back-of-the-envelope calculation using the known to you study (Lague et al ) shows conclusively that we simply can’t increase evaporation enough to make any noticeable reduction in AGW. It’s a DEAD END.
And that’s why people don’t waste their time to run a model that will just confirm what they already know from the back-of-the-envelope calculations.”
then I would suggest either you learn some English. or learn from the Monty Python’s Black Knight what happens if you are not able to admit your defeat
Once you do it – it won’t matter: “TK: which of the Lagues extreme examples should exhibit higher transient climate response (TCR) and/or higher equilibrium climate sensitivity (ECS)” – since NEITHER of them would allow us to increase evaporation enough to make any noticeable reduction in AGW.
As such, the difference between two too small to be significant effects is inconsequential.
So stop flooding the zone with your production and stop diverting the attention from the urgency of the only feasible and sustainable way to reduce AGW – reduction of GHG concentrations.
in Re to Piotr, 18 Feb 2026 at 1:42 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845388
Hallo Piotr,
Thank you for your feedback.
If I understood your reply correctly, you do not assert that the answer (to my question how climate sensitivity towards radiative forcings depends on water availability for evaporation from the land) is clear or that it can be easily derived from Lague 2023. You, however, seem to assume that this answer is unimportant, because the respective knowledge cannot be practically exploited for climate change mitigation. Am I right?
If so, I doubt that this assumption about impracticability of the respective knowledge / irrelevance of an absence thereof is justified.
Let us consider that the present Earth energy imbalance (EEI) may increase, decrease or stay constant depending on various emission scenarios, but it can hardly drop to zero in the near future. If so, I think that the extent of the climate change caused thereby and measurable e.g. in global mean surface temperature (GMST) increase and/or sea level increase will depend on the value of climate sensitivity.
If this value depends on water availability for evaporation from the land, it might be good to know if anthropogenic interferences with land hydrological regimes increase or decrease it. Should they have an influence, I assume that it might be useful to know how strong it is, because it might enable considerations if and/or how we could / should modify these interferences to avoid making the climate change worse.
That is why I think that taking the first step towards filling the existing gap in our knowledge, through the proposed modelling study, might not represent a useless academic exercise but may be also of a (potentially high) practical importance.
Greetings
Tomáš
P.S.
If you perhaps assume that a relatively weak influence of small changes in water availability for evaporation from the land on GMST that you infer from Lague et al represents a hint that the dependence of transient climate response (TCR) and/or equilibrium climate sensitivity (ECS) on the same changes should be weak as well, I do not see a reason for such an assumption.
If you have any, please explain.
Tomas Kalisz: “If so, I doubt that this assumption about impracticability of the respective knowledge / irrelevance of an absence thereof is justified.”
doubt all you want, it only counts what you can defend. And that’s not much:
TK : “The extent of the climate change caused thereby and measurable e.g. in global mean surface temperature (GMST) increase and/or sea level increase will depend on the value of climate sensitivity.”
The two points that you refuse to acknowledge are that
1. this sensitivity of the AGW to the REALISTIC changes in the water cycle minimal – since it will be but a tiny, tiny fraction of the maximum theoretically possible range of 8 C from Lague which they calculated as all land being a desert (no local evaporation) and all land being a swamp (maximum possible evaporation). Therefore, neither human nor natural CHANGES in water cycle are likely to move needle much – it would be but a tiny shift left or right on the 8C spectrum running from a global desert to a global swamp.
2. nor would we learn anything we don’t already know
– water cycle is NOT a driver of AGW – water cycle does not change on its own nor can be changed by humans, enough to increase or decrease of AGW.
– the temperature dependent changes in the water cycle ARE ALREADY implicitly included in the climate model, using the most realistic inputs we have, So if in one model scenario AGW by 2100 would be 2.8C, and in another 2C,
So what NEW insight would you gain if you calculated that the water cycle contribution to the AGW in these two cases was, say. 1.8C and 1.6C, respectively, and that we can change these numbers ONLY by changing GHG conc. ?
in Re to Piotr, 20 Feb 2026 at 1:50 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845482
Hallo Piotr,
Have you read the P.S. in my post of 19 Feb 2026 at 11:06 AM and the questions I asked therein?
It appears that you still construe Lague 2023 as evidence that the dependence of TCR and/or ECS on water availability for evaporation from the land is weak.
I do not see how you could have derived this assumption from their results that showed merely a weak dependence of GMST on water availability for evaporation from the land while keeping radiative forcings fixed. This is what you repeat in your comments on my questions again and again, however, it does not address what I asked.
TCR and ECS characterize the response of GMST to doubling of atmospheric CO2 concentration, in other words, to a significant change in radiative forcing. I still see the results of Lague et al. and my question as two clearly distinct fields with a yet unknown relationship.
I think that you already admitted that even the sign of the possible change in TCR and/or ECS (that we could / should expect in case of a shift in water availability for evaporation from the land) is not clear. On one hand, we have this uncertainty, on the other hand, we have the results of Lague et al showing a clear cooling effect of higher water availability for evaporation on GMST. I see this difference between uncertainty on one hand and certainty on the other hand as a hint that with respect to TCR and/or ECS, Lague et al do not provide any clear guidance.
How can you, on one hand, doubt about the sign of the possible TCR and/or ECS change, and in parallel, on the other hand, assume with a high certainty expressed in your comments that the absolute value of this change, for a given perturbation in water availability, will be low?
I do not think that your comments have ever addressed the questions asked in my P.S. and repeated in other words today. Could you do so?
Greetings
Tomáš
Re: Tomas Kalisz 20 Feb
I don’t see the answer to the most important question one must ask themselves before
==== Piotr 20 Feb:
[what] would we learn we don’t know already?
– we already know that water cycle is NOT a driver of AGW – water cycle does not change on its own nor can be changed by humans, ENOUGH to increase or decrease AGW in any noticeable way.
– the temperature-dependent changes in the water cycle ARE ALREADY implicitly included in the climate models, using the most realistic inputs we have. Let’s say in some model scenario AGW by 2100 is 2.8C, and in another 2C. What new insight would you gain if you calculated that the water cycle contribution in these two cases were, say, 1.8C and 1.6C, respectively?”
===
Until you answer that, I won’t waste more of my time.
in Re to Piotr, 22 Feb 2026 at 9:58 PM
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845552
Hallo Piotr,
In fact, I strived to reply to your sentence “water cycle does not change on its own nor can be changed by humans, ENOUGH to increase or decrease AGW in any noticeable way”
I think we do NOT know it (yet). I pointed to the circumstance that your reasoning for this assumption seems to be based on Lague 2023, despite they, in my opinion, no way deal with climate sensitivity.
I objected that if climate sensitivity has an influence on AGW, we should know how strongly it depends on water availability for evaporation from the land, not just believe that this dependence is weak.
As regards your second question, I replied already before – see again e.g. my post of 10 Feb 2026 at 5:48 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844947
that we should know if human interferences with water cycle enhance or attenuate the effect of CO2 radiative forcing, rather than attempt to attribute a number in “% of total temperature rise” to a factor that may contribute to the AGW rather by its synergy with “true” forcings than directly.
Greetings
Tomáš
RA: Given all that, why are there still people denying the reality of the IPCC consensus? I am genuinely baffled.
jgnfld: Money and power.
KW: Most deniers (I am labelled one here) think that CO2 has gone up due to humans, and the GMST has risen 1.4⁰C in some part due to that. Do you honestly believe there are deniers here who would disagree with this because I don’t recall reading such things?
BPL: You haven’t read the stuff by Victor or KIA or Dan da Silva or John Swallow? I have.
Yes, you are quite probably correct BPL, I had forgotten about people like Victor. I didn’t used to read many of his comments. I am sure some were deliberately inflammatory.
Mentioning James Swallow was curious, I went and searched his 6 or so most recent comments and none seemed incorrect, mostly things like…
– the models are not reliable
– fossil fuels aren’t realistically diminishing
– a bit of confusion between weather and climate (lots of people do that)
– pointing out that the sun does have an affect on climate (not saying it is responsible for all GW)
But the reason I am adding the comment was that he included in one of his comments a quote from Dyson. He quoted
‘”… I have studied the climate models and I know what they can do. The models solve the equations of fluid dynamics, and they do a very good job of describing the fluid motions of the atmosphere and the oceans. They do a very poor job of describing the clouds, the dust, the chemistry and the biology of fields and farms and forests. They do not begin to describe the real world that we live in. The real world is muddy and messy and full of things that we do not yet understand. It is much easier for a scientist to sit in an air-conditioned building and run computer models, than to put on winter clothes and measure what is really happening outside in the swamps and the clouds. That is why the climate model experts end up believing their own models.”
To which Ray retorted that Dyson was not a climate scientist. It is definitely worth reading the entire OpEd, or at least the opening remarks. It is as though he has written it directly for Rory Allen
https://web.archive.org/web/20190401171552/https://www.edge.org/documents/archive/edge219.html#dysonf
Actually, Keith, I’m glad you bring up Dyson. He’s a prime example of a particular type of denier–the visionary that is so in love with his vision that he cannot be bothered by trivialities like reality. Dyson’s love was the idea of humanity expanding to the stars, and he could never brook any complications that might prove an obstruction to that dream. It not only prevented him from realizing the threat climate change poses, it led to his son becoming estranged. So, not the warm tale you probably thought.
Dyson was truly a brilliant mathematical physicist, but his position on climate change is riddled with inconsistencies and misunderstandings.
First, the idea that somehow a model must be perfect to be useful is absurd. Yes, cloud physics needs to be improved, and yes it is very complicated. That does not invalidate the very well understood physics that is in the model that shows greenhouse gasses will definitely warm the planet.
Second, there is the idea that any uncertainty that remains in the model will work out to favor the position of Dyson and other deniers. First, the best estimates suggest that cloud physics will result is slightly higher sensitivity, not lower. Second, the idea that uncertainty in the models supports taking a sanguine attitude toward the threat stands the whole process of risk estimation on it’s head! Risk analysis needs to be done for a worst-case–or at least worst-case for a given confidence–model. Assuming a best case is like wearing a “Kick-Me” sign on your back as you walk away from Mother Nature. Risk analysis professionals have to believe in O’toole’s comment on Murphy’s law: Murphy was an optimist.
Third, we already know that climate sensitivity is much more likely above the median estimate than below. A 50% increase has led to a 1.4 degree rise, and the high-end tail is a whole lot longer than that at the low end. Moreover, we are seeing significant consequences–more billion $ disasters this last year than in any previous… Under those conditions where we know with certainty that the threat is real and that it is likely to be significant to severe, uncertainty certainly is not on the side of the sanguine. In fact, deniers and luckwarmers had better be praying the models are good, because they are the only tools we have for limiting the potential damage we could face.
Finally, there was Dyson’s virtually religious faith in the bright future of humanity. This certainly is not science or fact based. Given that 99% of all species that ever lived are now extinct, the future of humanity is much more likely to be in the fossil record than in the stars. We certainly have generated no data to suggest that our posterior distribution will overcome the prior favoring extinction.
In the end, Dyson’s sins were naivete and over-optimism, and they poisoned his ability to assess the threat of climate change objectively.
Ray,
Thanks for your background on Dyson. I’ve read very little of his stuff. His comment about not putting on winter clothes and measuring what is really happening outside demonstrates profound and contemptuous ignorance of all the research involving ice cores. I don’t think it gets the attention it deserves. Dyson along with a lot of other people seem to be under the false impression that it’s all models, and no empirical evidence.
in Re to John Pollack, 7 Feb 2026 at 7:47 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844832
Hallo John,
I think that the paragraph
‘”… I have studied the climate models and I know what they can do. The models solve the equations of fluid dynamics, and they do a very good job of describing the fluid motions of the atmosphere and the oceans. They do a very poor job of describing the clouds, the dust, the chemistry and the biology of fields and farms and forests. They do not begin to describe the real world that we live in. The real world is muddy and messy and full of things that we do not yet understand. It is much easier for a scientist to sit in an air-conditioned building and run computer models, than to put on winter clothes and measure what is really happening outside in the swamps and the clouds. That is why the climate model experts end up believing their own models.”
cited by Keith Woollard from a post citing it from an unknown writing by Freeman Dyson primarily addresses climate modelers, not the scientists who test their results by field research.
The purpose of my post is, however, different.
On 8 Feb 2026 at 9:59 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844849 ,
I asked if you could help me find the articles discussing a change in climate sensitivity during the Holocene.
After re-reading your post more carefully, I rather suppose that you meant climate sensitivity estimated from the climate changes between glacials and interglacials.
Am I right?
Greetings
Tomáš
Tomáš,
The obvious difference in ECS from palaeo studies compared with current models is that the palaeo record shows the change in CO2 as a result of the change in temperature. Since we have been taking carbon from geologically sequestered stores we do not have the same cause and effect scenario
Tomáš,
regarding your 9 Feb 2026 at 11:28 AM query,
You are correct that I was referring to the sensitivity estimated between glacials and interglacials. This follows an earlier discussion about Earth System Sensitivity on this blog.
I regard climate sensitivity in general as a partially theoretical concept, because it depends on the time allowed for it to manifest. It is an estimate that can be used to categorize climate models. As an observable quantity, it is only available as GHGs in combination with other factors whose effects also have to be estimated. The transient climate sensitivity is the most relevant to policy, because it is on a short enough time scale to be comparable to a human lifespan or the planning of infrastructure.
ESS, on the other hand, depends on the redistribution of both CO2 and heat through the oceans and readjustment of ocean circulation. It also depends on albedo feedback of both vegetation and snow/ice covered land and ocean. These adjustments might be partly completed over centuries, but nearly full equilibration would take thousands of years. At that time scale, Milankovitch cycles would also influence the outcome. I have no reason to think that ESS would be constant over the range of states of the Earth system that would need to be considered, and the very great changes that humans are making to the system at the same time.
However, it is impressive to me that the climate models were on the right track and approximate magnitude of GHG effects prior to the extraction of the various polar ice cores. This is an empirical verification of the models that Dyson derided. It was obtained in the messy real world, and definitely wearing winter clothing.
Thanks Ray. People who claim ‘science’ while attacking your intelligence and experience are full of it. A guy with a Nobel and/or a fine reputation is just another guy outside his area of expertise. You strike a fine balance between praise and criticism about Dr. Dyson.
Dad did much the same. I’ve cited this before: https://physicsworld.com/a/an-iconoclasts-career/
“In his career, we can see traces of the mathematical physicist’s reluctance to tackle the ambiguous or deeply puzzling question, or to go out mathematically even a little bit on a limb – something that contrasts sharply with his joyful interest in bizarre futurology. Perhaps this is the source of Dyson’s dreadful misjudgment on the climate question: he sees that the possible errors are large, but does not factor in that they are likely to be large in the wrong direction, and does not credit obvious qualitative arguments from simple laws of physics.”
I miss your dad, Susan! He was a class act!
Susan Anderson: “ A guy with a Nobel and/or a fine reputation is just another guy outside his area of expertise”
Actually, it may worse than that – a guy with a Nobel and/or a fine reputation may have his ego inflated so much that he believes that his brilliance in one field transfers automatically to any other, in fact – that it trumps the experts in any field – because for a genius like him the unfamiliarity with existing science is an advantage not disadvantage – he does not need to stand on the shoulders of giants, but unencumbered with the previous science in the field – he can see and do things the experts in the area couldn’t.
And then demagogues and propagandists (here: deniers) would take advantage of this inflated ego – by using their accomplishments in the original field – “add credibility by association” to the ill-informed, not thought-through, easily debunked (see Ray’s post above) proclamations in another field.
History is full of people accomplished in one area making fools of themselves when they made pronouncements about things outside their field of expertise. And it would be just a sad comment on human hubris, except when, like in the case of Dyson that are weaponized by interests groups that by delaying the mitigation of AGW to maintain their profits – may usher suffering of billions, mass extinction and the end of civilization.
Being a such smart person – Dyson should have been able to anticipate this.
Piotr: “should have” is useless in this context. Wisdom is not smartness. Psychopathy, or lack of empathy/sympathy/compassion/understanding of humanity, is a common trait of those who acquire power.
Dyson did what he did. Both Ray’s and my Dad’s cite (see link) pay deep respect to Dyson’s ability, while talking about his blind side.
Trying to simplify this into smart = should’ve diminishes the point.
offside lookup (PWA early signatory): https://americanhumanist.org/what-is-humanism/
I certainly agree that Dyson has no degree in meteorology. That’s one reason (some others being arrogance and a comfortable familiarity with physical models – but not paleoclimate research) why he can casually dismiss the combined knowledge that has been built up over the past 50+ years.
The modelers might be sitting in their air conditioned offices. However, you can bet that the folks extracting and analyzing ice cores from the middle of the Greenland ice sheet and Antarctica had to put on their winter clothes to get the empirical data he seems to be ignorant of! I guess it’s more fun to be a heretic than to familiarize yourself with the research in another field.
One of the outstanding results of those ice core measurements is that there really is a close correspondence between GHG levels and temperatures out in the messy real world. Despite all the other stuff going on, the modelers were basically right, and then the measurements came back to prove it!
Just thought I’d stick these quotations in from Dyson somewhere.
“There is a worldwide secular religion which we may call environmentalism, holding that we are stewards of the earth”.
“We are moving rapidly into the post-Darwinian era, when species other than our own will no longer exist”, “I do not pretend to be an expert about the details”. of climate change.
“I know a lot about nuclear weapons and nothing about climate change”
John Pollack: “Dyson has no degree in meteorology. That’s one reason (some others being arrogance and a comfortable familiarity with physical models – but not paleoclimate research) why he can casually dismiss the combined knowledge that has been built up over the past 50+ years.
The Dunning-Kruger … Then again from what Ron quoted him:
“ [Dyson]: I know a lot about nuclear weapons and nothing about climate change”
it may be also a bi-polar. With overriding need to be a contrarian – look at me how a fiercely independent mind I am:
Dyson: “ There is a worldwide secular religion which we may call environmentalism, holding that we are stewards of the earth”.
and end up on the same side as, and being used by, the deniers – who given the money and power standing behind them I would not call particularly “contrarian”. That’s like attacking the poor and siding with the rich, and thinking yourself a Robin Hood.
BTW – with labelling something a “secular religion”. you can dismiss ANYTHING:
– “ There is a worldwide secular religion which we may call “democracy”, holding that the people, and not the dictators or the money elites, should make the decisions that affects them”
– “There is a worldwide secular religion which we may call “human rights movement”, holding that the all people held some inalienable rights. How secularly religious of them! ”
– “There is a worldwide secular religion which we may call “motherly love” that holds that most mothers have deep love for their children. How self-deluded! “
Re: “Most deniers (I am labelled one here) think that CO2 has gone up due to humans, and the GMST has risen 1.4⁰C in some part due to that. Do you honestly believe there are deniers here who would disagree with this because I don’t recall reading such things?“
Re: “But the reason I am adding the comment was that he included in one of his comments a quote from Dyson.“
Another example of why one should distrust Keith Woollard’s denialist misinformation:
in Re to Atomsk’s Sanakan, 8 Feb 2026 at 9:36 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844848
Hallo Atomsk’s,
Although I do not know how Dr. Dyson arrived at his belief that global warming is not global, I am really grateful to Keith Woollard (as well as to you) for the reference to his remarkable article “HERETICAL THOUGHTS ABOUT SCIENCE AND SOCIETY”.
I found Dyson’s text very inspiring. Isn’t e.g. the paragraph
“Here is another heretical thought. Instead of calculating world-wide averages of biomass growth, we may prefer to look at the problem locally. Consider a possible future, with China continuing to develop an industrial economy based largely on the burning of coal, and the United States deciding to absorb the resulting carbon dioxide by increasing the biomass in our topsoil. The quantity of biomass that can be accumulated in living plants and trees is limited, but there is no limit to the quantity that can be stored in topsoil. To grow topsoil on a massive scale may or may not be practical, depending on the economics of farming and forestry. It is at least a possibility to be seriously considered, that China could become rich by burning coal, while the United States could become environmentally virtuous by accumulating topsoil, with transport of carbon from mine in China to soil in America provided free of charge by the atmosphere, and the inventory of carbon in the atmosphere remaining constant. We should take such possibilities into account when we listen to predictions about climate change and fossil fuels. If biotechnology takes over the planet in the next fifty years, as computer technology has taken it over in the last fifty years, the rules of the climate game will be radically changed.”
an excellent example of original, inventive thinking?
Greetings
Tomáš
John Pollack: “Dyson has no degree in meteorology. That’s one reason (some others being arrogance and a comfortable familiarity with physical models – but not paleoclimate research) why he can casually dismiss the combined knowledge that has been built up over the past 50+ years.
The Dunning-Kruger … Then again from what Ron quoted him:
“ [Dyson]: I know a lot about nuclear weapons and nothing about climate change”
it may be also a bi-polar. With overriding need to be a contrarian – look at me how a fiercely independent mind I am:
Dyson: “ There is a worldwide secular religion which we may call environmentalism, holding that we are stewards of the earth”.
and end up on the same side as, and being used by, the deniers – who given the money and power standing behind them I would not call particularly “contrarian”. That’s like attacking the poor and siding with the rich, and thinking yourself a Robin Hood.
BTW – with labelling something a “secular religion”. you can dismiss ANYTHING:
– “ There is a worldwide secular religion which we may call “democracy”, holding that the people, and not the dictators or the money elites, should make the decisions that affects them”
– “There is a worldwide secular religion which we may call “human rights movement”, holding that the all people held some inalienable rights. How secularly religious of them! ”
– “There is a worldwide secular religion which we may call “motherly love” that holds that most mothers have deep love for their children. How self-deluded! “
Tomas Kalisz: “I found Dyson’s text very inspiring. Isn’t e.g. the paragraph
[here TK quotes Dyson, off his rocker, embarrassingly for the man of his achievements, stumbling from one logical/scientific fallacy to next]. an excellent example of original, inventive thinking?”
Well, since you didn’t get Monty Python, I didn’t think take you for a master deadpaner. But here we are. Bravo!
Ray,
Not much point challenging most of what you have said as it is opinion and you are completely entitled to it.
However I will bring your attention to “climate sensitivity is much more likely above the median estimate than below.”
Just think about that
Your statement could be true iff the estimates are systemically wrong
Of course I meant the mean is higher than the median
Hello,
Is it an issue to address you in French if one does not have a good command of English? I am asking because one of your supporters is being ironic and even insulting because I ask my questions in French, as is another participant
in Re to Jean-Pierre Demol, 9 Feb 2026 at 5:00 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844875
Dear Jean-Pierre,
Personally, although I do not speak French, I am fine with your posts, because machine translations from French with freely accessible tools seem to be quite satisfactory nowadays. Although my English often surely sounds strange to native speakers, I post in this language because the quality of freely accessible machine translations from Czech seems to be still even worse in comparison.
May I ask if my reply of 8 Feb 2026 at 7:53 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844844
has somewhat dispelled your doubts about the evidence for anthropogenic origin of the observed rise of atmospheric CO2 concentration during the industrial era?
Greetings
Tomáš
Marcus Sarofim has a fine new substack: Reconsidering Reality: The Shoddy Case Against the Endangerment Finding – https://thesaraphreport.substack.com/p/reconsidering-reality
“The Trump EPA was grievously wrong on all counts.
“Several organizations submitted excellent analyses during the comment period. These include a masterclass from the Institute for Policy Integrity and a thorough summary from the Harvard Environmental & Energy Law Program. This post draws on their work, along with my own seventeen years of experience at EPA working on climate policy, to explain why the Trump EPA’s proposal cannot withstand scrutiny on the law, the science, or the engineering of regulation.
“This post is also accompanied by a more in-depth supplement with full quotations from the text.”
Should have put this elsewhere. Good review.
JCM says 8 Feb 2026 at 4:36 PM
Given that radiative controls by clouds and thermodynamics appear to shape temperatures almost everywhere, and that the observed increase in Earth’s energy imbalance does not come from LW changes but instead from decreased SW reflection …………… [aka increasing SW radiation absorption by the surface and biosphere accelerates warming]
JCM says 10 Feb 2026 at 11:02 AM
In the case of the tropical forests, it is often overlooked that unnatural extremes and bona fide climatological disruptions are attributable to direct ecological obliteration at the hands of man. While this may seem intuitive, it is often overshadowed in the way academic questions are framed. Comparatively, traditional knowledge systems have long emphasized such [TELE]connections………….
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844930
Maybe another useful arrow for your quiver JCM–rainforest climate cooling?
International Journal of Biosensors & Bioelectronics
Published: March 14, 2024
Restoring the earth’s damaged temperature regulation is the fastest way out of the climate crisis. Cooling the planet with plants
Peter P Bunyard Et al
Abstract
Reversal of global warming is accepted as an urgent necessity. Atmospheric CO2 concentration is widely used as a proxy and predictive indicator for global warming. Scientific calculation of the cooling efficiency of carbon sequestration options provides a rational basis for optimising climate investment.
Based on evapotranspiration data and average rainfall over the Amazon Basin, we determine the cooling power of latent heat transport over the tropical rainforest. A healthy rainforest acts as a heat pump, cooling the Earth’s surface and releasing heat into space. Because of the heat pump, the cooling engendered by evapotranspiration and subsequent cloud-forming is two orders of magnitude greater than the cooling brought about by carbon sequestration in forming biomass. ……………..
Introduction
Greenhouse gases heat up the planet, but they are not the only
driver of climate change. While carbon gets all the attention, another
factor is largely overlooked. It is water in its movements and changes
of state (ice, liquid water and vapour) as it interacts with plant life and
the atmosphere. This interaction has enormous stabilizing and cooling
effects. Once we understand the full force of plants and the water
cycle, we can actually confront the climate crisis with a whole new
set of measures. Plants, healthy soils and healthy ecosystems stabilize
weather, the climate and bring cooling. We can leverage these qualities
to fight the climate crisis. If the damage to the biosphere is reversed,
the planet will regain its capacity to regulate its own temperature.
https://medcraveonline.com/IJBSBE/IJBSBE-09-00237.pdf
Reply to Data with a follow-up note for Tomáš Kalisz 12 Feb 2026 at 7:27 PM
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845054
and Tomáš Kalisz says 10 Feb 2026 at 3:24 PM
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844941
and any others on this topic:
I think it’s worth emphasizing a point that is rarely, if ever, acknowledged in mainstream climate discourse: forests are not just passive CO₂ sinks—they are active climate stabilizers. Across geological eras to recent millennia, dense forests regulated Earth’s surface temperature via latent heat flux, evapotranspiration, cloud formation, and biotic pumping. These mechanisms move heat from the lower atmosphere to the upper atmosphere and ultimately out to space, functioning as living planetary heat pumps.
Humans were the first major disruptors of this system. Pre-industrial deforestation — in the Americas, Australia, Europe, and Asia — altered local and regional hydrology and cooling, long before industrial GHG emissions became dominant. Post-1970, large-scale tropical deforestation accelerated by orders of magnitude driven by explosive population growth. Yet, mainstream climate science largely ignores this.
The focus is almost entirely on Anthropogenic GHGs, with land-use change treated narrowly as a carbon source or albedo modifier only. The latent-heat cooling effects of forests, and their critical role in sustaining rainfall and moderating surface temperature, or a rapid uptake of atmospheric CO2 are routinely dismissed or labeled irrelevant distractions, sometimes even portrayed as “pro-fossil fuel denialism.”
By current attribution estimates, human GHG emissions account for roughly 60–70% of post-1850 warming. That leaves 30–40% unexplained by GHGs, much of which is likely due to land-use changes — yet this fraction rarely appears in discussions of climate forcing. Historical analogues, such as the 1500s post-European depopulation forest regrowth, show that regional and global cooling was measurable, and that forests have repeatedly modulated climate over millions of years. The physics are straightforward: forests cool the surface immediately through evapotranspiration, and regional latent-heat dynamics can cascade to global effects, while CO₂ reductions alone do not provide instantaneous cooling.
This perspective reframes tropical forest restoration from a “co-benefit” to a core climate intervention. While carbon drawdown is limited by land availability and ecosystem maturity, latent heat and hydrological effects act immediately, making regrowth a potentially high-leverage strategy alongside decarbonization, or unaffordable SRM or DAC approaches. The Amazon is the planetary exemplar, but smaller regional systems — such as NSW’s Big Scrub — provide local, observable analogues of the same dynamics. Deforestation there disrupted near-daily rainfall and inland moisture recycling, demonstrating how human land-use decisions cascade through hydrological and thermal systems worldwide.
The broader point is simple but rarely stated: forests were humanity’s first climate-forcing activity, and GHGs have acted on top of a system already destabilized. Local-to-global hydrological feedbacks from forest regrowth are largely underrepresented, dismissed, or ignored in climate models and IPCC assessments. Yet, the logic is robust: if healthy forests can actively cool and stabilize climate, restoring them at scale should be a central strategy in any serious climate plan.
in Re to “Data”, 13 Feb 2026 at 9:39 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845106
Sir,
Although I already tried to remind you thereof, you might have perhaps omitted my plea of 23 Jan 2026 at 5:53 PM,
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844291
if you could
1) confirm that you are going to stay with your present nick and that you desist from further attempts to cheat Real Climate moderators and readers by hiding under two or more names, and
2) clarify your reasons for this shameful behaviour in the past.
Please note that without doing so, it is very hard for anyone who knows your track on this website to hope that any of your recent posts expresses your honest beliefs or opinions and does not represent merely a further attempt to play games, cheat and manipulate.
Best regards
Tomáš
Dear Tomáš Kalisz,
I have responded / made other attempts but this was published @ Data says 22 Jan 2026 at 4:54 PM
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844236
It’s similar to my response here: Data says 23 Jan 2026 at 2:06 AM https://www.realclimate.org/index.php/archives/2025/12/1-5oc-and-all-that/#comment-844253
And think this is indicative of present position and attitude overall:
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844333
I am not responsible for the thinking and unfounded conclusions of others nor am I able to provide detailed evidence to support or deny other people’s claims. It’s not my job. It’s not my fault. Everyone can believe whatever they want. It is not my responsibility to react to every accusation that arises by online participants. I post comments on the topics I choose.
TEMPLATE – EDITED–Data says 22 Jan 2026 at 4:54 PM
My scientific commentary is solid, steady, focused, well supported by references, and clear enough. For the record, I am not and do not “make stuff up” nor am I here “to disinform” and neither am I maintaining multiple “sockpuppet accounts”. I do not engage in “personal attacks”, but retain my right to defend myself against spurious accusations. Furthermore, when I find myself to be “wrong,” or make an inadvertent “Typo/Mistake”, I correct the record asap. I am never “emotional”, am not a troll, am not a “denier” nor I liar. Neither do I engage in “strawmen logical fallacies” or make false unfounded “accusations against others.”
If this gets published Tomas, you take it as my final word on your previous questions and requests which I now draw a line under.
Thanks for asking. Let’s move along shall we?
in addition to my post of 14 Feb 2026 at 11:17 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845150
Dear moderators,
It appears that the poster “Data” anticipated my repeated request, because in his post of 14 Feb 2026 at 3:17 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845130 ,
he cites the posts wherein I raised it previously and asserts that he strived to reply to my questions, however, his responses remained unpublished.
I suppose that if it were true, his response to my questions might interest not only me but also some other Real Climate readers. I therefore wonder about reasons that might have caused why his attempt(s) might have not passed your scrutiny.
Could you perhaps comment?
Thank you in advance and best regards
Tomáš
Reply to Tomáš Kalisz
While this applied to accusations by all comers: https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844236
A short link list of blocked comments related to Tomas et al ‘requests/comments’ in mid-Jan for example replies to this https://www.realclimate.org/index.php/archives/2026/01/a-peek-behind-the-curtain/#comment-844503 as well as this
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844291
Dead links blocked
https://www.realclimate.org/index.php/archives/2026/01/a-peek-behind-the-curtain/#comment-844568
https://www.realclimate.org/index.php/archives/2026/01/a-peek-behind-the-curtain/#comment-844519
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844444
I made several more attempts but I am going to keep wasting my time, repeat myself or relitigate this issue again.
Tomas I have posted now several comments with new information you might be interested in. You may read that, ignore it, and continue ignore me as well. It’s your call and out of my hands.
in Re to Data, 15 Feb 2026 at 1:20 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845187
Sir,
Thank you for the links to your deleted posts. Hopefully they might help the moderators check if perhaps the deletion was inappropriate.
You have a bad luck that your entire activity on this website appears indistinguishable from other production previously posted by accounts that were finally proven by the moderators as camouflages for a single unknown entity behaving like a troll factory. This is the reason why I can hardly believe that your assertions are honest.
I am afraid that nobody else of the regular readers on this website is willing to trust you as well.
Sincerely
Tomáš
in Re to “Data”, 15 Feb 2026 at 1:20 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845187 ,
and 15 Feb 2026 at 12:54 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845186
and 14 Feb 2026 at 3:17 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845130
Sir,
Thank you for the links to your deleted posts, allegedly comprising your “genuine responses” to my questions. Hopefully these links might help the moderators check if perhaps blocking / deletion of these posts was inappropriate.
You have a bad luck that your entire activity on this website appears indistinguishable from other production previously posted by accounts that were finally proven by the moderators as camouflages for a single unknown entity behaving like a troll factory. This is the reason why I can hardly believe that your assertions are honest. For the same reason, I strongly doubt that an opinion exchange with you can be useful.
I am afraid that at least among regular readers on this website, you will hardly find anyone who might be still willing to trust you.
Sincerely
Tomáš
in addition to my post of 16 Feb 2026 at 6:07 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845254
Dear all,
I apologize for basically repeating, in a slightly amended form, my previous post of 15 Feb 2026 at 12:19 PM.
A strange bug – when I opened the webpage, it announced ‘159 Responses to “2025 Updates”‘ and my post of yesterday has not appeared among them. As soon as I submitted its amended version, the headline changed to ‘160 Responses to “2025 Updates”‘ and the previous post suddenly appeared among the published as well.
When I open “Updates” in a new tab, the “159” list of posts appears again. When I click on a post in the list that belongs to another thread, e.g. “Unforced variations”, the list switches to the actualized “160” version again.
Anyway, I am sorry for trying too quickly to correct a perceived error that didn’t actually occur.
Greetings
Tomáš
Some sort of bias which narrows the scope of a subject to a mainstream perspective, i.e. to its most media-visible narrative, is like denying music exists outside the Grammy awards.
In terms of realclimate science, like music, it is a diverse and vibrant subject far beyond what’s emphasized on TV. It is certainly not limited to the realm of carbon cycle politics – even if scicomm, boards and committees, institutional panels, attribution studies, newsmedia, academic funding patterns, political debates, etc. create that impression. Only casual observers and phony environmentalists overlook the depth and diversity of realclimate connections to environmental change.
On the issue of forests, which is not my primary interest, and where in-fact afforestation can be (and has been) detrimental to certain catchments, a previously highlighted pre-print model experiment article has now been published. Extreme events in the Amazon after deforestation https://esd.copernicus.org/articles/17/167/2026/
Adjusting albedo 0.12 to 0.18, LAI from 8.40 to 2.7, veg fraction from 0.92 to 0.85, roughness length (m) from 1.8 to 0.05, root depth (m) from 1.33 to 0.6, and forest fraction from 0.86 to 0.00 in an Amazon scenario results in a temperature increase 3.84C, 173% increase in no rain days, and 54% increase in violent rain.
The surface energy budget perturbation results in 30.93 W/m2 increase SW down, 16.48 increase SW up, 32.59 increase LW out (net), 38.21 W/m2 increased sensible heat, and -59.37 W/m2 latent flux. In sum, surface net radiation is decreased 18.14 W/m2, and temperature increases 3.84C.
A fascinating visualization of the ICON 5km storm resolving model simulation output is available here, including dramatic circulation disruptions. https://www.youtube.com/watch?v=yYsf1jK2Nrc
I encourage everyone to join in as allies in environmental conservation and remediation, even if for no other reason than stabilizing climate indicator variables.
good research paper and results, excellent yt video. thx
Reply to JCM
I encourage everyone to join in as allies in environmental conservation and remediation, even if for no other reason than stabilizing climate indicator variables.
+1
In conclusion, we show that even if annual mean precipitation may remain stable after deforestation, the tail of temperature, precipitation, and wind distribution broadens, making conditions more unfavorable.
https://esd.copernicus.org/articles/17/167/2026/#section4
This suggests biosphere thermodynamics is underweighted in climate risk assessments.
Of course none of that data or cumulative impacts on global climates dynamics will be found in any of the CMIP7 inputs or outputs. Therefore they’ll never make it into the IPCC AR7 either.
It will soon be forgotten and swept under the ghg forcing renew energy ideological carpet to collect dust.
Tomas Kalisz: I pointed to the circumstance that your reasoning for this assumption seems to be based on Lague 2023, despite they, in my opinion, no way deal with climate sensitivity.
Stop conflating different issues: we are discussing here being a driver (i.e. a SOURCE of radiative forcing), NOT being a passive feedback that merely affects climate’s “sensitivity” to the actual drivers (GHGs)… Read what you quoted:
TK: “Piotr 20 Feb.: water cycle does not change on its own nor can be changed by humans, ENOUGH to increase or decrease AGW in any noticeable way”
See? And to prove my point – I used water budgets and Lague to show that humans DON’T have enough extra water to evaporate to affect AGW. You were incapable of disproving nor could you show any natural mechanisms that would change noticeably the water cycle (other than via feedback with GMST). either.
So your “striving for answers” has produced nothing.
TK As regards your second question, I replied […] that we should know if human interferences with water cycle enhance or attenuate the effect of CO2 radiative forcing,
and I already answered it in:
==== Piotr 20 Feb: =====================
And what would we learn [from that] we don’t know already?
– the temperature-dependent changes in the water cycle ARE ALREADY implicitly included in the climate models, using the most realistic inputs we have. Let’s say in some model scenario AGW by 2100 is 2.8C, and in another 2C. What new insight would you gain if you calculated that the water cycle contribution in these two cases were, say, 1.8C and 1.6C, respectively?”
================================
TK I try to express my concern that although water cycle physics may be properly represented in available climate models, no one tried to find out yet if Earth with arid continents may be more or less vulnerable to CO2 concentration changes in the atmosphere than with humid continents.
Again: WHAT FOR, if we already have the temperatures over the most realistic, as opposed to your your arbitrary and hypothetical, continents, ALREADY
in the output of the models?
What new insight would you gain, beyond what we already know, from water budgets, climate models (Schmidt et al. 2010) and CERES data – that water cycle amplifies the effects of the AGW drivers (GHGs), and the only way to influence it is through the reduction in GHGs concentration?
in Re to Piotr, 25 Feb 2026 at 1:39 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845579
Hallo Piotr,
I think that from the very start of this thread by my post of 5 Feb 2026 at 11:48 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-844771 ,
I consistently speak about the yet unclear relationship between climate sensitivity towards radiative forcings and terrestrial hydrological regimes that could be characterized by water availability for evaporation from the land. The topic of water availability from the land as a forcing, as studied by Lague et al. is in my opinion an independent topic and I do not think that we have to discuss anything with respect thereto unless other authors put their results in doubt.
I therefore agree that we should not conflate my questions pointing to a yet open problem of the unknown relationship between water availability for evaporation and climate sensitivity, with the role of water availability for evaporation as a forcing, which has been already explored at least by Lague et al.
With respect to your second objection, I do not think that we should rely on your assumption that “we already have (temperature projections) for most realistic continents (with respect to hydrological regimes therein)”.
I do not think that temperature projections were modelled for various scenarios of the future developments with respect to hydrological regimes of continents. Although it might have been desirable (due to uncertainty with respect to anthropogenic perturbations that may change them), I am afraid that no one tried to do such modelling due to practical limits. If each emission scenario should be combined with several different hydrology scenarios, the extent of the necessary work would likely become intractable.
I am therefore afraid that in fact, no one can be sure that the available projections of the future climate were indeed made for “most realistic continents”. I think that the open problem mentioned above deserves an attention, because studying it might perhaps improve our insight at least a little bit.
Greetings
Tomáš
Water is not a ‘forcing”.
Here’s a useful explainer (Dr. Emanuel of MIT is reliable): https://climate.mit.edu/ask-mit/why-do-we-blame-climate-change-carbon-dioxide-when-water-vapor-much-more-common-greenhouse
“temperatures today are rising, thanks to humans’ emissions of longer-lasting greenhouse gases like CO2. Water vapor amplifies that effect. “If the temperature rises, the amount of water vapor rises with it,” says Emanuel. “But since water vapor is itself a greenhouse gas, rising water vapor causes yet higher temperatures. We refer to this process as a positive feedback, and it is thought to be the most important positive feedback in the climate system.”
“In short, it’s true that water vapor is in some sense the “biggest” greenhouse gas involved in climate change, but it’s not in the driver’s seat. CO2 is still the main culprit of the global warming we’re experiencing today. Water vapor is just one of the features of our climate that our CO2 emissions are pushing out of balance—well beyond the stable levels humanity has enjoyed for thousands of years.”
in Re to Susan Anderson, 26 Feb 2026 at 12:27 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845628
Dear Susan,
Lague et al
https://iopscience.iop.org/article/10.1088/1748-9326/acdbe1
made a modelling experiment in which they compared equilibrium surface temperatures for a “desert land” Earth with dry continents and a “swamp land” Earth with continents having unlimited water availability for evaporation. As water availability for evaporation was the sole parameter that changed in theor experiment, we could see its change as a kind of “forcing”.
Piotr referred to this specific article and to the circumstance that the results of this modelling experiment (a difference in global mean surface temperature (GMST) between these two extreme states merely 8 K) suggest that at least with respect to GMST, the “forcing” corresponding to anthropogenic perturbations inflicted to land hydrological regimes may be relatively weak.
Greetings
Tomáš
Tomas “ As water availability for evaporation was the sole parameter that changed in theor experiment, we could see its change as a kind of “forcing”.
Tomas, don’t use words which meaning you don’t understand. To be a forcing evaporation should changed BY ITSELF (or by humans), and changed ENOUGH to have a noticeable effect on global T.
Neither you nor your fellow anything-but-GHG denier JCM have been unable show any natural or human-driven increases in evaporation (that is not a result of GMST increase) that would be large enough to make a dent in the AGW.
Which has been pointed out to you and JCM …. hundreds(?) of times already. But neither of you would take “No” for an answer. I guess, because if you did, your entire claim to fame – discovering something the best climate scientists couldn’t or wouldn’t (in their “ artificial fixation and overemphasis on the outputs of trace gas“) – crumbles.
TK: I gave you something simple, direct, and unequivocal from one of the world’s finest climate scientists. I suggest you read the rest of link I provided and think about proportionate effects. You can argue your nitpick with Piotr and ignore his responses all you like; it will not change reality.
in Re to Susan Anderson, 27 Feb 2026 at 11:09 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845687
Dear Susan,
I read your reference
https://climate.mit.edu/ask-mit/why-do-we-blame-climate-change-carbon-dioxide-when-water-vapor-much-more-common-greenhouse ,
however, it seems to explain solely the limitations of the water vapour feedback related to the circumstance that water is a volatile liquid and its saturated vapour pressure rises with rising temperature.
Unfortunately, I have not grasped how this explanation could help in my dispute with Piotr that dealt with a possible relationship between two quite different phenomena, namely climate sensitivity towards radiative forcings on one hand and water availability for evaporation on the other hand.
I know that water vapour feedback is one of the aspects defining climate sensitivity, however, I do not see any direct clue that I could derive from explanations provided by Dr. Emanuel with respect to my question mentioned above.
Greetings
Tomáš
My understanding is water is a forcing in certain limited situations, for example an increase in crop irrigation that ultimately makes plants eject extra water vapour into the atmosphere is a forcing, not a feedback, but from what I’ve read the effects are small and local and mostly towards a small cooling effect due to the evapotranspiration effect. Its not going to save us from anthropogenic climate change, because the effects are so local and increasing the rates of irrigation causes a whole lot of practical problems.
TK “I therefore agree that we should not conflate my questions [on climate sensitivity the water feedback with changes in water cycle being a climate forcing].
Yet it was you who have conflated the two in your previous post!
In other words, you ate the cake, and when challenged, you, never admitting wolfing the cake, magnanimously agreed that “we” shouldn’t have eaten your cake ….
Way to own up to your mistakes … ;-)
Next to my two specific questions. since you haven’t answered them twice already
(I boldfaced them for your convenience):
==== Piotr 20 Feb: =====================
And what would we learn [from that] we don’t know already?
– the temperature-dependent changes in the water cycle ARE ALREADY implicitly included in the climate models, using the most realistic inputs we have.
Let’s say in some model scenario AGW by 2100 is 2.8C, and in another 2C. What new insight would you gain if you calculated that the water cycle contribution in these two cases were, say, 1.8C and 1.6C, respectively?”
================================
you didn’t answer first time, you didn’t answer the 2nd time – instead you questioned…, the realism of the outputs of current climate models, containing our best model representation of Earth’s hydrology – which water fluxes in each simulation influenced by the model temperatures.
So what’s your alternative, Tomas? Tomas Kalisz prescribing some …. ARBITRARY, unconnected to anything, water fluxes from land ???
Which of the two models will be more realistic, I wonder.
in Re to Piotr, 26 Feb 2026 at 12:48 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845629
Hallo Piotr,
I replied to your repeated questions several times and still wonder how you could have missed it:
1) Q: And what would we learn [from the proposed modelling study] we don’t know already?
A: We would learn if a decrease in water availability for evaporation from the land decreases or increases climate sensitivity to doubling of atmospheric CO2 concentration, and if so, how strongly.
2) Q: Let’s say in some model scenario AGW by 2100 is 2.8C, and in another 2C. What new insight would you gain if you calculated that the water cycle contribution in these two cases were, say, 1.8C and 1.6C, respectively?
A: To me, this question does not make much sense, as I have not proposed to compute any “water cycle contribution” in any AGW model scenario.
I tried to explain also why I think so:
If the complex “water cycle feedback” (a compound of water vapour, lapse rate and cloud feedbacks) perhaps changes due to changes in water availability for evaporation, and water availability for evaporation can change due to anthropogenic perturbations with land hydrological regimes, the modelling will require not only a scenario of future anthropogenic CO2 emissions but also a scenario of these anthropogenic hydrological perturbations. I am not sure if an attempt to ascribe a part of the computed warming to CO2 emissions and the rest to hydrological perturbations (or oppositely) might still have a value
I also strived to express my opinion that oppositely, the (yet missing) knowledge (if the climate sensitivity towards rising CO2 concentration increases or decreases with increasing water availability for evaporation from the land) may be useful (and therefore it can be worth of the effort to design and run the proposed modelling study). It is because this knowledge could serve as a hint whether
a) we should keep water in the landscape because (i) we need it for various purposes and (ii) it slows the pace of the warming caused by the given CO2 forcing, or
b) we should seek a minimal water content in the landscape that will still satisfy our practical needs but, in parallel, will cause the minimal possible amplification of the CO2 forcing.
Greetings
Tomáš
Tomas Kalisz Hallo Piotr, I replied to your repeated questions several times and still wonder how you could have missed it
Hallo Tomas, I replied to your repeated questions several times and still wonder how you could have missed it:
1. I pointed that you conflated drivers with the sensitivity to feedbacks. You replied
“we should not conflate my questions”, as if it was I who conflated your issues and not vice versa.
2. TK: I have not proposed to compute any “water cycle contribution” in any AGW model scenario.
So you are not going to use the AGW climate models with realistic hydrological cycle and the most REALISTIC representation of its changes (as being driven by the temperature output of the model) in favour of climate models for which you PRESCRIBE …. availability of water for evaporation which you take from your head, and run the model at arbitrary fixed CO2 level. Hence was my question:
Piotr 26 feb: “ Which of the two models will be more realistic? I wonder” I still wonder how you could have missed it… Instead answering it – you try now to resurrect your long-dead arguments:
TK: The (yet missing) knowledge (if the climate sensitivity towards rising CO2 concentration increases or decreases with increasing water availability for evaporation from the land)
The knowledge is missing only in your head. As explained to you dozens(?) of times ( I still wonder how you could have missed it)
Piotr 25 feb: “WHAT FOR? What new insight would you gain beyond what we already know, from climate models (Schmidt et al. 2010) and CERES data – that water cycle amplifies the effects of the AGW drivers (GHGs)? =====
So what you are proposing is to discover a wheel, and a square one at that (since using arbitrarily changed evaporation under arbitrary Co2 conc.) Garbage in, Tomas’ results out.
TK: It is because this knowledge could serve as a hint whether we should keep water in the landscape [because] it slows the pace of the warming caused by the given CO2 forcing,
I replied to these claims several times and still wonder how you could have missed it:
Piotr 25 Feb “What new insight would you gain, beyond what we already know from water budgets [combined with results of Lague et al.] – the only way to influence [“the pace of the warming”] it is through the reduction in GHGs concentration” – we can’t change water fluxes ENOUGH to make difference in AGW rate, I explained it to you dozens(?) of times and still wonder how you could have missed it. E.g.:
======= Piotr 9 Feb ===================
– climate change is not driven by changes hydrological regimes, but by the emissions of CO2 and other GHGs,
– water cycle is not a driver but a mere feedback – it amplifies consequences of our action, or inaction, on the mitigation of GHGs
– given the volume of water cycle and the very short residence time of water vapour in atmosphere (a week vs. dozens of years for CO2), and the limited amount of untapped ground water for extra irrigation – WE CAN’T DO ANYTHING to increase evaporation ENOUGH to make a dent in AGW
– your only practical idea how to do so (irrigation of Sahara) has been shown to be a monstrous fallacy
And that’s the reason why scientists do not respond to your numbingly repetitive demands that they take the time from their research and study your and JCM’s ideas,
===================
Which part of the above has been incomprehensible to you, Tomas?
in Re to Piotr, 27 Feb 2026 at 2:04 PM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845677
Hallo Piotr,
You understood me correctly, I indeed thought that (if it were technically possible) it might be better to model climate projections for arbitrarily chosen emission scenarios combined with arbitrarily chosen water availabilities for evaporation.
I suppose that so far, climate scientists strived to implement in their models some “hydrological modules” predicting the future land hydrological regimes, or somehow parameterize them.
I doubt, however, that such hydrological predictions or parameterizations can be realistic / reliable if we do not know how (basically unpredictable) human contributions may alter the “objective” situation (that might have been at least in principle predictable if we assume that the created hydrological modules are fit for the purpose).
That is why I came to the idea that admitting this unpredictability by using several arbitrarily chosen hydrological regimes and assessing how the obtained warming projections will differ from each other might be perhaps more useful. I know that this approach may look silly, however, it is indeed the way how I am thinking.
I know that climate models (e.g., Schmidt et al. 2010) and CERES data suggest that water cycle amplifies the effects of non-condensing GHGs as the AGW drivers. I asked if we are sure that the effect would have been the same if we had much wetter (or much drier) continents than we presently have. Again, I see that you consider this question silly.
I therefore asked myself if there could have been a situation in (a not very distant) past that the continents were (globally) significantly drier or wetter than today, so that I could show an example supporting my guess that it might be possible. I must admit that I have no idea if, from a global perspective, terrestrial hydrological regime remained basically stable during the Holocene, or if it showed some fluctuations or, perhaps, some trend.
I see that in absence of data confirming my speculation that terrestrial hydrological regimes may have changed, I will hardly convince anyone else on this forum except JCM that the proposed modelling study directed to the influence of a such change on climate sensitivity might have been useful.
I therefore admit that although I explained my points and have been understood, I failed to convince the majority that there may be some plausibility in my speculations.
I think, Piotr, that I cannot clarify nor support my views any better and we can stop our exchange on this topic. Thank you for your patience during this discussion.
Greetings
Tomáš
Tomas Kalisz: I think, Piotr, that I cannot clarify nor support my views any better
Clarification? What for? Your problem is NOT that nobody understands you, but that we understand you too well – in fact we understand the implications of your claims better than you, because you never asked yourself the most fundamental questions for ANY research: “What for? What new insight would you gain beyond what we already know?”
And when I asked you these questions – you never answer (zebra’s test for a denier – check) – offering instead: evasions, tangents, conflating unrelated concepts, and not understanding what the words you are using mean (like “forcing”).
What you “strive”, “believe”, or “think” does not matter if you cannot defend it.
So the next time you bring your evaporation schemes, we can by-pass most of the discussion and jump straight to its conclusion – to the questions you can’t answer and to the points how have been evading:
=====
– “WHAT FOR? What new insight would you gain beyond what we already know, from climate models (Schmidt et al. 2010) and CERES data – that water cycle amplifies the effects of the AGW drivers (GHGs)?”
“What new insight would you gain, beyond what we already know from water budgets [combined with results of Lague et al.] – that the only way to influence your “pace of the warming” it is through the reduction in GHGs concentration – we can’t change water fluxes ENOUGH to make difference in AGW rate ”
“ Which of the two models will be more realistic? The AGW climate models with realistic hydrological cycle and the most REALISTIC representation of its changes (as being driven by the temperature output of the model)
or your hypothetical model in which you PRESCRIBE …. the availability of water for evaporation which you take from your head, and run the model at arbitrary fixed CO2 level? Hint: Garbage in, garbage out.
Or if you don’t like the questions – main points you failed to falsify:
======= Piotr 9 Feb ===================
1. climate change is not driven by changes in hydrological regimes, but by the emissions of CO2 and other GHGs,
2. water cycle is not a driver but a mere feedback – it amplifies consequences of our action, or inaction, on the mitigation of GHGs
3. given the volume of water cycle and the very short residence time of water vapour in atmosphere (a week vs. dozens of years for CO2), and the limited amount of untapped ground water for extra irrigation – WE CAN’T DO ANYTHING to increase evaporation ENOUGH to make a dent in AGW
4. your only practical idea how to do so (irrigation of Sahara) has been shown to be a monstrous fallacy
And these are some of the reasons why scientists do not respond to your numbingly repetitive demands that they take the time from their research and study your and JCM’s “ideas”.
===================
the problem is that some people’s brains are bound to artificial frames that are disconnected from Earth System process.
If we define AGW strictly as GHG forcing, and sensitivity strictly as the global average temperature response to GHG forcing, then we can only ask questions like “what is the sensitivity of the sensitivity”?
This question is usually understood using Perturbed Parameter Ensembles (PPE) which change how the model behaves internally. This asks, “Given the same Earth, how uncertain are we about how its processes operate?”
Adjustments are made to entrainment rates, convection triggers, liquid and ice droplet properties, and so on.
This is done exactly in: “On the Sensitivity of Climate Sensitivity in a coupled climate model”
https://essopenarchive.org/users/921675/articles/1297441-on-the-sensitivity-of-climate-sensitivity-in-a-coupled-climate-model?commit=fe3d3701c82d685008f96a1a561e963704f3e82a
where it is found that adjusting uncertain *internal* properties (within reason) can make GISS Model E2.1 have a range of ECS from 3.0 to 7.3C for doubling CO2. The authors note the default configuration produces ECS = 3.3C.
The question of human caused (or other causes) of landscape modification is one of *external* constraints, or more specifically forcings. This asks a different question: “What if the world itself were different but has the same physics?”
That is a boundary condition experiment. Boundary condition experiments, or forcings, include things like changing CO2 concentration, orbital properties, continental configuration or converting lush landscapes to eroded rockflour.
When someone says “Water feedback is already included in climate models” they are referring to model *internal* physics, in which you could vary physics parameters. PPE asks “How uncertain is water related parameterization?”
When someone asks “What if we changed water availability” that is a boundary condition experiment that is *external*.
These are categorically different concepts, and we should be careful not to confused them.
PS:
I should note that the authors of “On the Sensitivity of Climate Sensitivity in a coupled climate model” find that the different sensitivities arise primarily through tropical short-wave (SW) radiative cloud feedbacks, while the long-wave (LW) cloud feedback is insensitive to the parametric space that was sampled. More specifically, the SW Cloud feedback across the range of sensitivities 3.0 to 7.3 is about -0.5 W/m2/K to +0.5 W/m2/K. That is about the maximum range of difference one can obtain by varying internal physics in GISS-E2.1-G
Observably, however, SW cloud radiative effects are changing far more dramatically (previously discussed), definitely outside that range. For a 1K transient change seems to be associated with > 3 W/m2 decreased reflection. It should be concluded therefore that the observed SW cloud radiative effects can only be explained by changing boundary conditions (forced), not solely by uncertain internal parameters in GISS E2.1-G.
PPS: based on figures from Hansen since early 2000s, if we choose a maximum aerosol forcing change since early 2000s at about 0.5 W/m2 (standard range 0.1-0.5), the contribution from snow/ice “feedback” of 0.15 which is basically observable, and we choose the maximum GISS E2.1-G PPE SW cloud “feedback” at 0.5 W/m2/K, and we have a huge 1.7 W/m2 increase of absorbed solar energy (observed), that is leaving something in the order of 0.5 to 1 W/m2 decreased SW cloud reflection unexplained.
If climate models are useful, which I think they are, then the diagnosis is a missing external (boundary) forcing, not something that can be produced solely by internal model physics. This is in addition to the well-established major trace gas radiative forcing over the same period.
Hansen and others argue, without justification IMO, that the discrepancy is caused by an extreme SW cloud radiative feedback (internal factors), but that is inconsistent with the range of plausible physics already embedded in models like GISS E2.1-G. People hand-waving about extreme cloud-feedback seem to imply that we should simply veto/ignore the physical constraints that a model like GISS E2.1-G informs, across the full range of reasonable physics parameters. In my view, rather than throwing out decades of accumulated knowledge about system physics that makes such a model, it’s much more reasonable and less controversial to infer that there is an unaccounted-for boundary forcing.
in Re to JCM, 28 Feb 2026 at 9:14 AM,
https://www.realclimate.org/index.php/archives/2026/01/2025-updates/#comment-845701
Dear JCM,
Thank you very much for your post. Unfortunately, for me, your message remained somewhat cryptic.
I took therefrom that it might have sense to make a distinction between two kinds of “inputs” into climate system, namely “forcings” and “boundary conditions”. I tried to find out in this respect what could be a distinctive aspect between water availability for evaporation on one hand and atmospheric CO2 concentration on the other hand if I should assume the first as an example of the boundary condition and the second as an example of the forcing.
It came to my mind that the sought distinction might primarily consist in the circumstance that whereas, on one hand, a quick CO2 concentration change will instantly cause an imbalance between solar energy absorbed by the Earth and infrared energy emitted therefrom (that has lot of further consequences but does not seem to have an influence on the Earth climate sensitivity to this forcing), a quick change in water availability for evaporation, on the other hand, might have both the instant effect on the Earth energy balance (and thus also act as a kind of a “forcing”) as well as an effect on the Earth climate sensitivity to other forcings.
I am, however, still very uncertain if it was indeed the intended message. If you wished to dispute or correct any of my points from the preceding discussion with Piotr, please do not hesitate to be specific and direct.
Greetings
Tomáš
Tomas Kalisz: “Dear JCM, Thank you very much for your post. Unfortunately, for me, your message remained somewhat cryptic.”
that’s normal, Dear Tomas. When somebody can’t answer direct challenges to their claims, they could admit it and re-evaluate their views, or escape into somewhat cryptic pondering of the terrible shortcomings of others, e.g.:
JCM: “ the problem is that some people’s brains are bound to artificial frames that are disconnected from Earth System process.”
so no reference to WHOSE brains are these and without any QUOTES to prove JCM -then from this setup JCM cannot be proven wrong. A rhetorical ironclad. And as a bonus – helps JCM’s strawmanship:
JCM “ If we define AGW strictly as GHG forcing, and sensitivity strictly as the global average temperature response to GHG forcing, then we can only ask questions like “what is the sensitivity of the sensitivity”?
Notice that JCM does not refer to any specific post, so we can’t check whether this was was these posts have indeed said, or whether it was deliberately misrepresented by JCM for his rhetorical needs. By this setup – JCM assured is infallibility – without any source – nobody could possibly challenge his claim, and his intellectual integrity.
Muddy waters make it easy to catch fish.
To Tomas,
the way I see it is that inputs to the model are things the model does not solve for internally. Models calculate what happens when you change a boundary condition and they make predictions about the effects such changes to outgoing longwave radiation, solar absorbed radiation, temperature, and precipitation.
GCMs do not calculate human caused CO2 emission, human caused aerosol emission, human caused continental degradation, orbital parameters or total solar insolation. Each of these is provided as an input external to the model. The model – i.e. the structured representation of how physical variables relate to one another in nature – sorts out the response.
Changing such external inputs produces physical consequences, including TOA energy imbalances and new equilibrium states. Diagnostics exist to summarize information to make it feel more digestible for humans, such as the effective radiative forcing associated with changing boundary conditions. Similarly, kernel methods are used to estimate radiative feedback diagnostics from the numerical output by relating certain variables to changes in GMST. Such diagnostic are not calculated directly by models themselves but calculated after-the-fact by an analyst. They summarize how the model system responded to changing boundary conditions using the familiar forcing-feedback paradigm. Of course, model output could be summarized other ways and I don’t believe GCMs have any dependence whether effective radiative forcing and feedback is diagnosed in post or not. GCMs just solve physical equations.
Escaping model land for a moment, the response to uncertain input is manifest into observable signals. Models can be used to detect and attribute the expected contribution from inputs by calculating the consequences. For observable changes that are unexplained by the calculated predictions from physics, these can be traced to either 1: something wrong with the physics itself (the model), or 2: diagnosing something wrong with the inputs. I choose to believe the models are probably OK, and I admit that input information about the ongoing continental deterioration is not nearly as straightforward as boosting a well mixed GHG.
Your interest is about possible compounding or amplifying effects of varying multiple inputs simultaneously, which is interesting and adds another layer of complexity. It’s an interesting question. However, I caution against conflating an artificial frame with a physical one: there is no compelling physical reason to aggregate all human influences into a single climate sensitivity.
CO2 sensitivity is best measured using model experiments that somehow control and account for all other variables perfectly, or even better to hold all other variables constant. Similarly, experiments isolating water availability do the same. Any attempt to vary multiple boundaries simultaneously introduces a hen-and-egg problem.
Some communication difficulty arises from the socio-political–CO2 framework, where “climate sensitivity” is used operationally at the policy–science interface, and the term “AGW” or human caused warming compresses a lot of ambiguity. As we’ve discussed, equilibrium response can be computed for any chosen input (or combination), not just CO2. In my view, settling at 560ppm CO2 equivalent GHG the system could be observed to warm 4C while the classic equilibrium climate sensitivity is still 3C, because other stuff is going on!
I have come to the position that it is more physically responsible to report the effects of different human boundary disturbances distinctly. Nevertheless, I do understand that in the current socio-epistemic environment it might be more convenient to communicate informally that the multiple concurrent human related disruptions could make the system appear more sensitive to CO2 (even if CO2 sensitivity doesn’t really vary that much from a physics perspective). Previously I have considered such a thing to invite interest within the familiar ECS framework but I admit it makes no difference here.