Cloud Feedback

Guest Post by Willis Eschenbach

In the comments to Christopher Monckton’s latest post, Nick Stokes drew attention to Soden and Held’s analysis of feedback in the climate models. I reproduce their Table 1 below:

soden held table 1

Figure 1. Soden and Held’s Table 1, showing all of the feedback parameters calculated from the models.

I found several amazing things in this table. The first is the huge range of values for the various parameters. While all of the Planck parameters are within a few percent of each other, the lapse rate feedback varies by more than three to one from smallest to largest; the surface albedo feedback varies by nearly five to one; and the cloud feedback varies by an amazing factor of more than eight to one from smallest to largest.

Despite these huge variations, all of them can (relatively) successfully emulate the historical record when they are each fed their own special brand of forcings … which should tell us something about the models. But I digress.

What I want to look at today is the cloud feedback. This is measured as something called the net cloud radiative effect (CRE), which is the sum of the solar and longwave radiation from the clouds. There is general agreement that as a global average the CRE is negative with a value of about -21 W/m2, meaning that in general the clouds cool the earth.

However, there is little agreement about the size or even the sign of the cloud feedback. Cloud feedback is the change in the net CRE that we can expect from a 1°C change in temperature. The models say that cloud feedback is a 0.69 ± 0.10 W/m2 INCREASE in downwelling radiation for each additional degree of temperature. In other words, if there is a small warming, the models say the clouds amplify it to make a large warming. This implies a positive correlation between temperature and the net CRE.

Fortunately, the CERES data can give us actual observational data regarding this question. Figure 2 shows the correlation between temperature and the net CRE.

correl net cre tempFigure 2. Correlation, net cloud radiative effect (CRE) and temperature. Monthly climatology removed before calculation.

Clearly, this is a hugely complex system, where in some parts of the world the correlation is strongly negative, and in some parts it is positive. Note that the inter-tropical convergence zone (ITCZ), which I have long held is a crucial part of understanding the climate, is negatively correlated. And so is the land area north of about 50°N or so.

Overall, we can calculate the global correlation by looking at the global area-weighted average net CRE versus area-weighted global average temperature. Figure 3 shows that result:

correl net cre temp globalFigure 3. Area-weighted averages of the cloud radiative effect (CRE) and the surface temperature. Monthly averages (climatology) have been removed before calculation.

This is very bad news for the models … they all claim that there is a positive correlation between CRE and temperature, which makes the model-projected warming much larger … but in fact the global average correlation is negative.

And from this same data, of course, we can calculate the global average cloud feedback parameter, viz:


                   Estimate   Std. Error  t value Pr(>|t|)

(Intercept)       -0.0006277  0.0392260  -0.016 0.987250

(Temperature)     -1.0121541  0.2695399  -3.755 0.000234 ***


Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.5263 on 178 degrees of freedom

Multiple R-squared:  0.0734, Adjusted R-squared:  0.0682

F-statistic:  14.1 on 1 and 178 DF,  p-value: 0.0002345

The CERES observational data says that for every additional degree of warming, the net cloud radiative effect becomes 1 W/m2 more negative, meaning a cooling effect. So despite the fact that the models claim that the cloud feedback parameter is positive (0.69 ± 0.10 W/m2 per degree C), ugly reality disagrees. According to the CERES data, the real number is not only negative, it is strongly negative ( – 1.0 ± 0.27 W/m2 per °C) and is strongly significant.

Not much more I can say about that, except that it totally confirms my long-held belief that given the amazing stability of the climate system, and my hypothesis about how clouds and thunderstorms regulate the temperature, cloud feedbacks perforce must be net negative, not net positive as the models claim.

Best to everyone, my girlie is getting married in about six hours, I’ll wearing a rented tux and not drinking red wine … wish me luck.


Please be clear about your objections by quoting the exact words you object to, so we can all understand the exact nature of your disagreement.


181 thoughts on “Cloud Feedback

    • Monckton of Brenchley
      September 4, 2016 at 11:21 am

      A most useful analysis from Willis Eschenbach. The more one looks at the actual numbers from within the models, the more questionable their outputs are.
      Do pardon me Lord Moncton, but I have to say I disagree in this point.
      For as much as I know, hopefully am not wrong, the models outputs are or seem to make our climate estimates more questionable, as the actual outputs, the projections, do give better numbers, that fit better, much better than the estimations reached by other means.

      All GCMs, their output projections cover a “ball court” of 2.5C-3.2C associated with a 200-220 ppm.
      That is as far as I know their actual final output………..meaning that according to GCMs that is the atmosphere thermal swing and its association with the ppm(s).
      Regardless of the thermal swing being an artificial warming or not.
      And the rest is a messed up charade, where the outputs are ignored, and projections are considered as predictions……….and there can not be any simulation that can predict unless the projections validated first by the reality and the paleo climate data…..

      By this angle the GCMs do not validate the other assessment reached by other means.
      The official assessment is at 4.5-7.5C thermal swing associated with 120ppm swing. A very wrong one, making the current interglacial look like not a proper one………..aka the Holocene.

      Actually the GCM projections fit much better for a proper current interglacial….and especially can be considered as validated by the Climate data covering the period from the LIA nadir to present…….


      • I would say that your lucid explanation of Lord Monckton’s “erroneous assertion ” would be a very respectable candidate to win the Bullwer-Lytton prize for literary gobbledegook.


      • >The official assessment is at 4.5-7.5C thermal swing associated with 120ppm swing.

        Couple of typos: “Swing” should be “swig”, and “120ppm” should be “120 Proof”. Now the rest of this comment makes perfect sense.

      • I think whiten’s reply is one of those written by a computer program that pulls climate sounding nouns and verbs randomly from a table and inserts them into a template document and then posts.

      • whiten September 4, 2016 at 2:31 pm
        “…Actually the GCM projections fit much better for a proper current interglacial….and especially can be considered as validated by the Climate data covering the period from the LIA nadir to present……. ”

        Only if one ignores the LIA and our historical rise out of the LIA, plus one pretends that our temperature history has not been cooled artificially; then in that imaginary sense, the GCM projections almost mimic alarmist confirmation biases.

        Still no proof, not even decent suppositions.

      • ATheoK
        September 4, 2016 at 8:42 pm

        Thank you.

        I know I did not explain it, so let me do it now as you seem to ask for it.
        One thing the data from LIA nadir to present show is that there is no concentration increase observed before an ~0.4c thermal swing, even when CO2 emissions must have increased.

        The numbers that I contemplate as estimates from GCMs, in accordance with the lag time of ppm(s) versus temp increase, fit very well with it.
        What the lag shows is that there is no ppm increase expected unless the thermal swing reaches above 0.2-0.3C, according to GCMs, and therefor no expected detection of the ppm increase unless the thermal swing reaches above 0.4-0.5C……but at that point suppose to become obvious and more so if the thermal swing still persist……..that is the GCMs.

        The other estimate, the orthodox one, in accordance with the lag time of ppm(s) versus temps gives corresponding numbers to at least a double up.
        Where the ppm increase becomes detectable only after a thermal swing of ~0.8 C.

        The 800 years lag, according to the climate science estimation of the temp swing of a 4.5-7.5C, has a share of `0.5-0.6C, meaning there is not any ppm increase unless the thermal swings reaches above the 0.6C, and there will be no any detection of ppm increase unless the thermal swing persist beyond 0.8C,,,,, something that is not validated by the data covering the period of the thermal swing from the nadir of LIA to present.
        The ppm increase starts to be detected at the 0.4 C limit……

        Could I be wrong with this?
        Yes of course, but only trying a share my view point………

        Thank you for your interest…..


    • If one reads the Soden and Held paper carefully it states that cloud feedback is a calculated as “residual” for each model. That is it is back calculated to give a number along with the other feedbacks so as to match the model’s estimate of global warming, which do not involve feedbacks. So if the models over estimate warming as they do, cloud feedback must be assumed positive if the initial forcing from CO2 and the feedback method is used. Settled science in action!

      • Richard Petshchauer:
        When you say “the model’s estimate of global warming” do you mean: a) the change in the global surface air temperature, b) the change in the global surface air temperature at equilibrium or c) the change in the value of some other variable?

      • Richard Petschauer September 5, 2016 at 8:02 pm
        “If one reads the Soden and Held paper carefully it states that cloud feedback is a calculated as “residual” for each model.”

        WR: “cloud feedback is a calculated as “residual” for each model.” Calculated. When they don’t have DATA about the behaviour of clouds (and they don’t), they are missing data of one of the most important parameters that decide over ‘warming or no warming’ of the Earth. In regard to ‘reality’ no conclusions can be made because ‘reality’ can’t be simulated.

        The modelmakers had to be that honest to state that ‘None of the results of this model might be used for predictions because of the lack of essential input’. Such a disclaimer should have been the minimum for every model that lacks essential data.

        (Disclaimers are used to limit responsibility. Wikipedia: “A disclaimer may (…..) or may specify warnings or expectations to the general public (or some other class of persons) in order to fulfill a duty of care owed to prevent unreasonable risk of harm or injury.”)

  1. So why don’t the modeler’s know this? Why didn’t they make this calculation and tie it into their model maths? I’m a pure skeptic, but I wonder why they do they opposite of what is observed?

    • I am quite certain that some modelers do know this, but they are also well aware that grant money, prizes, and recognition will accompany arrows that point in a different direction, and they have earned their keep accordingly. You can replace “earned” with “absconded” at will. I guess “sleazed” might also fit!!!

      • The question is, why “scientists” continue to incorporate the very worst performing models in their ensembles. Why don’t they just drop those? Is there some “gentlemen’s agreement” in force? Are they afraid of offending the authors of those models? Does is serve their political agenda, and how, exactly?

      • JorgeK- if a model could be correct only ONE model would exist. There aren’t 92 different theories of evolution. Darwin’s original hypothesis, that species evolve to survive in a particular ecologic niche, and the most adaptable species survive the longest, still works. General relativity is still mind boggling, but it survives, with some modifications and refinements.

        Or as George Box of statistics fame put it, “all models are wrong, some are useful.” Current climate models are wrong and not useful since they cannot be tested. Mr. Box was also a proponent of the idea that any model was only useful within the range is was tested for. In other words models can’t make predictions because the predictions, by definiton, cannot be validated.

      • Jorgek,
        How do you rank these as poor to be dropped or good to be kept?
        Seems that the same groups that create the model runs getbto evaluate how good they are.
        This is not good. Do you have a better way?

      • GS, there are two ways to do this in principle. Kumar suggests rerunning past models using actual inputs, and comparing results to observation. Won’t happen because puts modelling groups out of business. Or, just take CMIP5 and compare 2006-2016 projections to observations. Then we can throw out all models except Russia’s INM-CM4. It has highest ocean thermal inertia, lowest WVF, and second lowest ECS. And now we can even start to say why the others are so bad. Oberstated air/ocean coupling, overstated water vapor feedback. BTW, those two are also strongly related by Willis’ Tstorm governor hypothesis around the ITCZ.
        What I would do is tell all other modelling groups, you have one year to fix, rerun to reproject 2006-2016 as well as to re- hindcast 1976-2006. If the results are not significantly improved, you are permanently defunded as incompetent. Why the year? Because the ‘fixed’ models will all have lower ECS and we can then use them to shoot the model ECS nonsense down also.

    • tom s asks:

      I wonder why they do they opposite of what is observed?

      The answer:

      Money, status, grants, expense-paid jaunts to holiday venues, money, professorships, money, fame, political power, tenure, and in case I forgot… money. And lots of it.

      They are bought and paid for, and if any one of them strays, they’re replaced by those waiting in line; the ones who got their degree in “climatology”, believing the loot would follow.

      For the most part scientists promoting the alarmist narrative know the truth: they know that that CO2 does not have anywhere near the effect claimed. They know that changes in CO2 follow changes in temperature. They know that observations contradict the models. They know that all their scary, alarming predictions have failed. They know they’re deliberately disregarding the Scientific Method, and Occam’s Razor, and the Null Hypothesis, and plain common sense. They know that ‘dangerous AGW’ is a failed conjecture. And so on.

      They know those things, but money and status trumps them all. As Stephen Schneider famously said, “Each of us has to decide what the right balance is between being effective and being honest.” But there is no balance. Honesty has been defenestrated.

      At least corrupted scientists have reasons that we can understand, even if we disagree with their selling out to Mammon.

      But it’s a little more difficult to understand the lemmings who occasionally post here; the ones who made up their minds early on that CO2 is the primary ‘control knob’ of global temperatures, but without the benefit of having nearly enough data to arrive at such a definitive conclusion. They have the same mind-set as Harold Camping’s true believers (Camping was the religious leader who predicted the end of the world. When it didn’t happen as predicted, he just re-set the date. But we’re still here). The lemmings simply cannot admit that they guessed wrong.

      W.B. Yeats had the lemmings pegged:

      The best lack all conviction, while the worst are full of passionate intensity.

      Alarmist scientists made a conscious decision to sell out any ethics they might have had for money, fame, and similar rewards. That is reprehensible, but understandable. But the parrots who keep insisting that a climate catastrophe is right around the corner are the ones Yeats is describing.

      If a hypothesis like CO2=cAGW had been as thoroughly deconstructed in any of the hard sciences, it would have been dead and buried decades ago. The only thing that keeps it alive is the deluge of government grant money, supported by a media that fans the alarmist flames while giving little to no coverage of skeptical points of view.

      Take away the money (or even spread the grants evenly between alarmists and skeptics), and the ‘carbon’ scare would collapse practically overnight.

      Because as Nobel prize winning physicist Richard Feynman stated:

      “It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, if it doesn’t agree with observation, it’s wrong. That’s all there is to it.”

      The conjecture that CO2 is the primary influence on global temperatures is wrong.

      That’s all there is to it.

      • db – The internet are full of junk, garbage, and general you know what… And then there is the occasional diamond gleaming brightly in the light.

        Thank you, sir, for one of those diamonds. You could not have cut it more precisely.

      • 97% of the scientists (because of career, mortgage, family) prefer to float on ‘mainstream’ thinking. As soon as ‘the current changes’ they will prefer to float on mainstream as well. But in the opposite direction. The heroes in science are those who are seeking the truth – whatever mainstream says. Or said.

      • Quoting dbstealey

        tom s asks:

        I wonder why they do they opposite of what is observed?

        The answer:


        And dbstealey “answer” …… was truly a great one. And I would like to “add to it” by posting commentary I penned several years ago, but with some modifications, to wit:

        To figure out “why the supporters and passionate believers in/of CAGW do the opposite of what is observed” …… one has to understand the “nurtured mindset” of those individuals.

        And to understand said “mindset” one has to look at the three (3) distinctly different groups of people who “have a BIG dog in the CO2 fight”, …… and it is of my opinion …… that all of them, for their own personal reasons, have been desperately trying to convince the public that:

        1. Increasing Global Warming is “right as rain” and will destroy life on earth if not kept in check;

        2. The cause of AGW is the “greenhouse” gas CO2 that is increasing in the atmosphere;

        3. Human activities are the cause of CO2 increasing in the atmosphere;

        And the three (3) groups are, to wit:

        Group #1: Government funded Climate Scientists – This group has expended years n’ years and hundreds of millions of government funds researching the effects of Greenhouse Gases and to justify past expenditures and their future existence they were forced to provide a PJE (Proof of Job Existence) for public approval …. and thus their “proof(s)” are their claimed “increasing average temperatures”.

        Group #2: Opportunists wanting “part of the action” – with so much “free” taxpayer money being distributed indiscriminately they seized upon the opportunity to “jump on the Global Warming bandwagon” anywhere they could get “hold” so as to get their share of said tax dollars and used the “claimed proofs” attested to by the aforementioned Climate Scientists to justify their actions.

        Group #3: Environmentalists and liberal socialists – when Group #1 and Group #2 got CO2 declared an “air containment” and the primary cause of AGW ……. it was a Godsend for Group #3 and they also “jumped on the Global Warming bandwagon” and cited the “claims” of Group #1 and Group #2 for the explicit purpose of furthering their agenda of “shutting down” all Capitalism and Capitalists ventures they could by claiming they contribute to the increase in atmospheric CO2 quantities.

        Given the above, is there any question as to why there is a “consensus of opinions” among the three (3) above Groups that ….. CO2 causes AGW?

      • Thank you db for your excellent post. You correctly stated:
        “The conjecture that CO2 is the primary influence on global temperatures is wrong. That’s all there is to it.”

        In the shorter term, the Nino3.4 index appears to predict average global temperature ~4 months in the future, and does so quite well (except after major volcanoes that cause temporary cooling). Others have published on this subject. My formula is:
        UAHLTcalc (Anom. in degC, ~four months later) = 0.20*Nino3.4IndexAnom + 0.15

        In the longer term, the integral of solar activity and the PDO appear to be the primary drivers of global climate. I have not personally verified this relationship, but I am confident that its author Dan Pangburn is credible. See Fig. 11 at:

        In both the shorter and longer term, atmospheric CO2 is NOT a significant driver of global temperature. In fact, CO2 lags temperature at all measured time scales. Temperature, among other factors, drives atmospheric CO2 much more than CO2 drives temperature.

        W. B. Yeats could have been foreseeing a bleak future, where the global warming alarmists succeed in destroying modern energy systems with their warmist scare tactics: “Things fall apart; the centre cannot hold; Mere anarchy is loosed upon the world”

        It is astonishing that anyone listens to these warmist scammers, who have been wrong about EVERY scary prediction they have made to date.

        Unfortunately, the warmists have convinced many of our politicians, and when misinformed politicians fool with energy systems, innocent people suffer and die.

        The problem with “green energy” is that it is not green and produces little useful energy. The key problem is intermittency. In the absence of grid-scale storage (aka the “”super-battery”), grid-connected green energy is typically destructive, since it greatly increases electricity costs and also reduces grid reliability.

        The high rates of Excess Winter Mortality in the United Kingdom are evidence of this green energy debacle. Britain and Germany have driven many of their citizens into energy poverty, and are now retreating from green energy foolishness as fast as they politically can.

        Cheap abundant reliable energy is the lifeblood of society – it IS that simple.

        Regards, Allan

        Reference: “Cold Weather Kills 20 Times as Many People as Hot Weather” by Joseph D’Aleo and Allan MacRae, September 4, 2015

      • Allan

        Whether CO2 is the primary influence on global temperatures is indeterminate pending identification by global warming climatologists of the statistical population underlying their models. This statistical population is the source of the information, if any, that is conveyed to a would be regulator of Earth’s climate. Absent this statistical population no information is conveyed to a would be regulator hence the climate cannot be regulated.

        I have interacted extensively with governmental regulators on science policy. Their behavior suggests that regulators like to regulate because regulating pays quite well. It pays the same whether the regulator has sufficient information to regulate or not. This phenomenon seems to me to be the source of governmental fraud targeting global warming and at least one additional target of governmental regulation. A common characteristic of these frauds is the absence of the statistical population underlying the regulatory model.

      • For the record, here is my position on Climate and Energy. I hope to be wrong abut point 6, ebcuse humanity and he environment both suffer in a cooling world.

        September 4, 2015
        By Allan MacRae

        Observations and Conclusions:

        1. Temperature, among other factors, drives atmospheric CO2 much more than CO2 drives temperature. The rate of change dCO2/dt is closely correlated with temperature and thus atmospheric CO2 LAGS temperature by ~9 months in the modern data record

        2. CO2 also lags temperature by ~~800 years in the ice core record, on a longer time scale.

        3. Atmospheric CO2 lags temperature at all measured time scales.

        4. CO2 is the feedstock for carbon-based life on Earth, and Earth’s atmosphere and oceans are clearly CO2-deficient. CO2 abatement and sequestration schemes are nonsense.

        5. Based on the evidence, Earth’s climate is insensitive to increased atmospheric CO2 – there is no global warming crisis.

        6. Recent global warming was natural and irregularly cyclical – the next climate phase following the ~20 year pause will probably be global cooling, starting by ~2020 or sooner.

        7. Adaptation is clearly the best approach to deal with the moderate global warming and cooling experienced in recent centuries.

        8. Cool and cold weather kills many more people than warm or hot weather, even in warm climates. There are about 100,000 Excess Winter Deaths every year in the USA and about 10,000 in Canada.

        9. Green energy schemes have needlessly driven up energy costs, reduced electrical grid reliability and contributed to increased winter mortality, which especially targets the elderly and the poor.

        10. Cheap, abundant, reliable energy is the lifeblood of modern society. When politicians fool with energy systems, real people suffer and die. That is the tragic legacy of false global warming alarmism.

        Allan MacRae, Calgary

      • Allan MacRae, I expect many politicians latch onto CO2 drive climate theory because it provides a “scientific” fig leaf to cover their nakedly corrupt accumulation of power. Same with other nonsense and debunked theories they favor.

      • Note to Terry O.

        Not disagreeing with you at all, but allow me to make one “big-picture” observation:

        Global temperature declined from ~1940-1975, increased from ~1975-2000, and has stayed flat since ~2000, all while atmospheric CO2 increased; so the correlation of temperature to increasing atmospheric CO2 has been NEGATIVE, Positive, and Near-Zero.

        I suggest Near-Zero is the correct estimate of the sensitivity (ECS) of global temperature to increasing atmospheric CO2. There is no real manmade global warming crisis, because there is no credible evidence to support this failed hypothesis.

        Furthermore, every scary prediction made by the global warming alarmists has failed to materialize, The warmists have a perfect NEGATIVE predictive track record. A capable betting man would always bet on the warmists being wrong, and he would make a lot of money.

        Best, Allan

      • Allan:

        Thanks for taking the time to reply. I don’t take sides on the issue of whether CO2 driven global warming is or
        is not a threat but merely try to push the arguments that are made by both sides in the direction of logicality.
        An argument is written in words and these words have meanings. When a word changes meaning in the midst of an argument, this argument is an example of an equivocation. An equivocation looks exactly like a syllogism but isn’t one. Thus, while it is logically proper to draw a conclusion from a syllogism it is logically improper to draw a conclusion from an equivocation. To draw such a conclusion is the “equivocation fallacy.”

        A global warming argument is seldom made that does not apply the equivocation fallacy. Thus, while arguments are many logical arguments are few. We can improve the intellectual climate in this respect through disambiguation of the language in which arguments are made, for application of the equivocation fallacy is rendered impossible when terms cannot change meaning.

        A terms that often changes meaning in the midst of a global warming is “prediction.” The word is used in reference a non-inference and in reference to an inference in making arguments. A disambiguation that satisfies the need was suggested years ago by Kevin Trenberth. Under this disambiguation, a prediction that is an inference is a “prediction” while a prediction that is not an inference is a “projection.” If we adopt Trenberth’s disambiguation, no climate model supporting regulation of CO2 emissions makes predictions. All of them make projections. It is also true that no statistical population underlies any of these models. That this is true is consistent with the fact that no currently existing climate model makes predictions for
        predictions imply the existence of the underlying statistical population.

        Absent the underlying statistical population a model cannot provide a would be regulator with information about the outcomes of events as in order for this to happen values must be assigned to conditional probabilities but these values are dependent upon counts of the non-existent sampling units.

        As information is essential to providing information to a regulator regulators cannot regulate. Nonetheless, regulators continue to try to regulate. The blame for this fiasco can be assigned to persistent applications of the equivocation fallacy in making global warming arguments. A decade ago, most professional climatologists were guilty of this offence. These days it is mostly the amateurs who are guilty of it. Applications of this fallacy keep alive the illogical contention that the Earth exhibits a “climate sensitivity” whose value is about 3 Celsius per doubling of the CO2 concentration. In this blog nearly every blogger accepts this contention and argues only over the magnitude of the climate sensitivity.

      • Note to hanelyp – you could be correct.

        The other possibility is that most politicians are so utterly innumerate and incompetent that they actually believe the global warming scam.

        Regrettably, we are governed by scoundrels and imbeciles*
        ( * Note to file: The two terms are not mutually exclusive.)

      • Thank you Terry O: You wrote:
        “Applications of this fallacy keep alive the illogical contention that the Earth exhibits a “climate sensitivity” whose value is about 3 Celsius per doubling of the CO2 concentration. In this blog nearly every blogger accepts this contention and argues only over the magnitude of the climate sensitivity.”

        This is an interesting observation. Here is my comment from 2014 (and earlier).

        Regards, Allan


        The “mainstream” global warming debate centres on the magnitude of Equilibrium Climate Sensitivity (“ECS”) to atmospheric CO2, which is the primary subject of contention between global warming alarmists (aka “warmists”) and climate skeptics (aka “skeptics”).

        Warmists typically say ECS is high, greater than ~~3 degrees C [3C/(2xCO2)] and therefore DANGEROUS global warming will result, whereas skeptics say ECS is 1 degree C or less and any resulting global warming will NOT be dangerous.

        The scientific evidence to date (increasing atmospheric CO2, but no net warming for ~17 years) strongly suggests that if one had to pick a side, the skeptics are more likely to be correct.

        However, BOTH sides of this factious debate are in all probability technically WRONG. In January 2008 I demonstrated that CO2 LAGS temperature at all measured time scales*, so the mainstream debate requires that “the future is causing the past”, which I suggest is demonstrably false.

        In climate science we do not even agree on what drives what, and it is probable that the majority, who reside on BOTH sides of the ECS mainstream debate, are both technically WRONG.

        Based on the preponderance of evidence, temperature drives CO2 much more than CO2 drives temperature, so ECS may not exist at all at the “macro” scale, and may be utterly irrelevant to climate science except at the “micro” (and materially insignificant) scale.
        There may be other significant sources of CO2 that contribute to its increase in the atmosphere, but increasing CO2 just does not have a significant or measureable impact on global warming (or cooling), which is almost entirely natural in origin.

        I therefore suggest that the oft-fractious “mainstream debate” between warmists and skeptics about the magnitude of ECS is materially irrelevant. ECS, if it exists at all, is so small that it just does not matter.

        Wait 5 to 10 more years – I suggest that by then most serious climate scientists will accept the above hypo. Many will claim they knew it all along… :-)


        * If ECS (which assumes CO2 drives temperature) actually exists in the Earth system, it is so small that it is overwhelmed by the reality that temperature (among other factors) drives CO2.

        In this enormous CO2 equation, the only signal that is apparent is that dCO2/dt varies ~contemporaneously with temperature, and CO2 lags global Lower Troposphere temperatures by about 9 months.

        CO2 also lags temperature by about 800 years in the ice core record on a longer time scale.

        To suggest that ECS is larger that 1C is not credible. I suggest that if ECS exists at all, it is much smaller than 1C, so small as to be essentially insignificant.

        Regards, Allan


        My January 2008 hypo is gaining notice with the recent work of several researchers. We don’t always agree on the fine details, but there is clear agreement in the primary hypothesis.

        Here is Murry Salby’s address to the Sydney Institute in 2011:

        See also this January 2013 paper from Norwegian researchers:
        The Phase Relation between Atmospheric Carbon Dioxide and Global Temperature
        Global and Planetary Change
        Volume 100, January 2013, Pages 51–69
        by Ole Humluma, Kjell Stordahlc, Jan-Erik Solheimd
        – Changes in global atmospheric CO2 are lagging 11–12 months behind changes in global sea surface temperature.
        – Changes in global atmospheric CO2 are lagging 9.5–10 months behind changes in global air surface temperature.
        – Changes in global atmospheric CO2 are lagging about 9 months behind changes in global lower troposphere temperature.
        – Changes in ocean temperatures explain a substantial part of the observed changes in atmospheric CO2 since January 1980.
        – Changes in atmospheric CO2 are not tracking changes in human emissions.

      • Thanks for sharing, Allan

        My objection to “the climate sensitivity” is associated with the fact that it is a ratio which is implied to be a constant. The numerator is the equilibrium surface air temperature and though it can be computed the equilibrium surface air temperature cannot be observed. AR4 addresses this issue by a strawman argument that associates falsifiability with the late philosopher of science Karl Popper and states without justification that Popper’s philosophy has been superceded by peer review. Unappreciated, perhaps, by IPCC climatologists is that a model that lacks falsifiability generates no information, making regulation of the climate impossible. In particular, the model that maps the change in the radiative forcing to the change in the equilibrium surface air temperature through use of “the climate sensitivity” as the proportionality constant generates no information.

        By the way, use in the literature of global warming climatology of the term “signal” (e.g. “anthropogenic signal”) violates Einstein’s theory of relativity as the claim is to be able to regulate the climate system and control over a physical system is gained by having information about the outcomes of events before these outcomes occur. In order for this information to be brought to us by a signal this signal would have to travel faster than the speed of light in a vacuum but this is impossible under Einstein’s theory. Though Einstein’s theory bars gaining the required information from a signal it does not bar getting this information from a climatological theory. Thus far, the theories that have been given to us by the professional climatologists are of a type that conveys no information to us, however.

      • Also to Terry O. You wrote:
        “A terms that often changes meaning in the midst of a global warming is “prediction.” The word is used in reference a non-inference and in reference to an inference in making arguments. A disambiguation that satisfies the need was suggested years ago by Kevin Trenberth. Under this disambiguation, a prediction that is an inference is a “prediction” while a prediction that is not an inference is a “projection.” If we adopt Trenberth’s disambiguation, no climate model supporting regulation of CO2 emissions makes predictions.”

        Here is a video of Dr. Trenberth, entitled “Kevin Trenberth: “Global Warming – Coming Ready or Not!”.

        I suggest that Dr. Trenberth’s above comment is not a disambiguation, because he and his colleagues in the global warming alarmist camp continue to press politicians and the public to mitigate the alleged disastrous effects of alleged major global warming, when both these allegations are not only unproven, they are probably false. I suggest Dr. Trenberth is in fact being disingenuous with this comment, because elsewhere he and his colleagues present their improbable global warming scenarios as real dire predictions, and say the world must spend trillions of dollars of scarce to mitigate them. Dr. Trenberth cannot sit on both sides of this fence – he is either making projections, which have insufficient credibility to cause huge expenditures, or he is making predictions, for which there is no credible evidence.

    • Probably ends up a bit like the drake equation. Even making reasonable assumptions on the variables, it’s very easy to come up with whatever you want instead of exposing flaws in your assumed answer.

      The interesting thing is that most people are complete morons when it comes to this. Even well educated people seem to fail to understand that, if you don’t even know the sign of the relationship in response to a forcing…by definition you cannot truly understand what is happening. I’ve showed well educated people some of the actual assumed values used (which as you can see are wildly different) from official government reports stored on official government servers…and then they turn around and call me a denier just before repeating the tired, old idea that the models should be trusted because “consensus”. But it is truly impossible for most of these models to be correct…they each describe completely different worlds.

      It is ironic that they don’t catch the failure in this. Consensus is “general agreement”. Yes, the bulk of climate scientists “generally agree” that CO2 should cause warming. But it is clear that the “general agreement” is full of uncertainties. The “known unknowns” alone are so broad that it is truly impossible for them to reach a definitive answer. And the “unknown unknowns” are still lurking out there, capable of upending everything.

    • The models are unlikely to incorporate much, if any, data from CERES. The first satellite was launched in the late ’90s. The models assume that CO2 amplifies the effects of water vapour in the atmosphere. So cloud formation “ought” to be accompanied by enhanced down-welling LWIR. In this case Willis’ analysis reflects what ought to have been self evident to the modelers were they not wedded to an assumption of warming driven by CO2. The only time that clouds help maintain warmth is at night.

    • See the previous article

      Part of the problem is that it’s very difficult to prove any model correct! Since the systems are chaotic, any small inaccuracy in the setup will yield inaccurate results regardless of the accuracy of the equations. The best we can ever do is to give a broad area of results in which the real world should be, like the hurricane track forecasts, which get ever-wider as you get further away in time.

    • There are many reasons, but science does this all the time and in every discipline. Scientists are hubristic and egoistic, and like the rest of us, (most) start from an opinion and want to prove that opinion is right. That is why scepticism is so important, and the lack of scepticism amongst climate scientists so dangerous. The worst myth in this whole sorry global warming saga is that scientists are noble, god-like beings, seeking a higher truth.

    • To be honest, you can’t write a program that may have 200,000 factors and influences, or whatever the actual number of factors that interact to create the climate might be. It just can’t be done. Most of those factors are only guessed at, not known, and you can not write a program that “models” reality if you don’t know what you are “modeling.” The best you can do is create “Sim Climate,” a take off on Sim Farm and Sim City, and no more accurate then either of them. If you actually could program that accurately, we would have a stock market program that actually could predict the stock market. Since we don’t, do you REALLY believe we can program something that could predict the climate?

      • Tom O:

        It is perceptive of you to state that “To be honest, you can’t write a program that may have 200,000 factors and influences, or whatever the actual number of factors that interact to create the climate might be. It just can’t be done.”

        In dealing with this situation, the builder of a model must “abstract” (remove) this model from almost all of these factors aka features. How to do so is the ancient philosophical problem that is called the “problem of induction.” There is a modern solution to this problem which, however, has not been adopted by global warming climatologists.

  2. Good luck, Willis! This should be a good discussion. I don’t think one has “settled science” when a crucial variable’s estimates vary 8 to one defensibly.

  3. I am struck by figure 2, and how in several places the correlation coloration closely follows political boundaries. I suppose that in some places political boundaries follow topographical discontinuities, but that is not true for the US/Canadian border or the boundary between South Africa/Botswana.

    Is this an artifact of methodology in each country, caused by regional infilling, or ?

    • Very interesting idea.
      However, the blues does extend south of the parallel. To my naked eye it seems to follow the Great Lakes as the lower border.
      That does make sense. Below the Great Lakes are huge agricultural fields. These are human constructs that can affect the local climate. Forest and prairie (or crops) are different and managed by man.

  4. Simple analyses [22] indicate that either an increase of approximately 186 meters in average cloud altitude or a decrease of average albedo from 0.3 to the very slightly reduced value of 0.2928 would account for all of the 20th century increase in AGT of 0.74 K. Because the cloud effects work together and part of the temperature change is due to ocean oscillation (low in 1901, 0.2114 higher in 2000), substantially less cloud change would suffice.

    My Reply:

    My point which is if prolonged solar activity changes the terrestrial items which determine albedo, cooling will be in the offing.

    This is what needs to be evaluated and watched.

  5. A very nice analysis. My only critique is that the Ceres data span a time when balloon and satellite observations both say there was no global troposphere warming, and that could bias the result. But its what we got.

    There are other reasons to think the cloud feedback cannot be positive, but might be negative, in the ICOADS and ISCCP cloud data. Essay Cloudy Clouds in ebook Blowing Smoke explores that in some detail. The AR5 WG1 Chapter 7 SOD (removed in the final version) actually said of their confidence in positive cloud feedback,
    “This conclusion is reached by considering a plausible range for unknown contributions by processes yet to be accounted for, in addition to those occuring in current climate models.”
    Yup, warmunist belief in unknown contributions by processes yet to be accounted for. Settled science for sure—NOT.

    • They reached a conclusion based on “considering a plausible range for unknown contributions by processes yet to be accounted for”.

      In other words: Guessing.

  6. Best of luck Willis and my fondest best wishes for your girl and her beau. I suggest you rethink the red wine strategy.

    And thanks for the post. I shall read it now in more depth.

    • Stick to the red wine strategy. Make up for it tomorrow. The kids will be recovering then and you won’t be pushed into matching them.
      And good luck.

      • Eschewing wine of colour does not exclude the less staining fizzy kind.
        And you get to release some CO2 as well.
        As the modern idiom has it: double win :)

        Have you a speech to make?

  7. Changing the cloud feedback to a -1.0 W/m2/C drops global warming theory warming by more than half.

    Here is how the feedback assumptions lead to 3.0C per doubling (feedback on feedback values carefully chosen to maintain a large number).

    And then with clouds at -1.0 W/m2/C, we see the calculations fall to just 1.373C per doubling.

    That is why they NEVER come out and show what the empirical feedback results to date are. Because then the math falls apart and we get modest nothing warming.

    This is also why there can be such huge variation in the feedback assumptions as shown in Soden Held Table 1.

    Because they all ADD up to roughly the same 2.0 W/m2/C regardless of how the individual feedbacks vary. This is what the climate models are based on (assumptions carefully chosen).

    • I am very impressed, both by Willis´ findings and by your (Bill Illis’) conclusions. Is there another conclusion possible than that all the models are tuned to get a specific (wanted) result?

    • Of course. There are so many very poorly constrained ‘parameters’ that they can fix the results to be whatever they want.

      They play around with these “experiments” until they a get set of result which vaguely follows the later 20th c. and gives a good scary warming in the 21st. Then they publish their new “findings” in a complicit journal.

      It is basically an unstructured, hand driven form of multivariate regression, with not objective control criterion for the quality of the fit.

  8. If the climate system feedback to most any temperature forcing were not “negative”, we would not be here to ponder the issue today. That’s where I began in all this a couple decades ago and it where I’ve remained. I’ve long suspected “cloud cover” was at least one of (possibly the dominant) of the “negative feedback” mechanisms that held planet earth’s chaotic climate system within the limits revealed in our history of Ice Ages and Warm Periods and limited the progression rate between those ultimate limits, come hell or high water.

    How anyone can study the past 500,000 years of atmospheric temperature and not arrive at that “negative feedback to ANY external forcing” conclusion is simply beyond me.

    • CH, careful to distinguish an effect from its first derivative with respect to T. Clouds net cool. They damp temperature, as Willis pointed out. That is why climate is reasonably stable. The question is whether that damping gets weaker (positive feedback) or stronger (negative feedback) with a 1C increase in T at ~2x CO2. So long as the Bode f net equivalent is under ~0.8, there is significant amplification but no unstable runaway. (4.5C is ~ 0.72). See my long comment to Monckton clay feet part 1 a few days ago using the sound system examples. On an outdoor sound stage, f=0.6 works observationally, and in climate f=0.65 =>ECS 3 would work. There are many lines of observational reasoning voing backmto Guy Calendar in 1938 suggesting the ‘true’ ECS is ~1.5-1.65, and in any event <1.8. In which case the alarm is still off without confounding primary damping with feedbacks.

      • I understand your distinction. In my world, anything less than one unit of response (long term) to one unit of forcing represents “negative feedback”. I see nothing in the reconstruction of the earth’s historic macro-temperature envelope that would suggest anything other than “negative feedback” to temperature forcing (by my definition).

    • I am with Claude. Any system with net positive feedback will eventually take an excursion into a catastrophe. Systems are controlled by negative feedbacks. It has to be in there somewhere.

      • 600 million years of climatic reconstruction, covering several ice ages, inter-glacials, and hot house epochs, with carbon dioxide way above the AGW “tipping point”, and no catastrophe in site.

        And Dr. Hansen apparently ignorant, or uncaring, about the little detail of Venus and Earth having much different atmospheres. Earth: 90% within 10km, ToA (Top of Atmosphere) at 100km. Venus: 90% within 50km, ToA at 250km. FYI, Top of Atmosphere is the Karman line

  9. Held et al 2005

    Water vapor is found to provide the largest positive
    feedback in all models and its strength is consistent with that expected from constant relative humidity
    changes in the water vapor mixing ratio.

    Hardly surprising it is “consistent” with the assumptions of constant RH assumed by all modellers.

    • Greg, Judith Curry and I got into this a fair bit concerning some early guest posts on her blog in 2012. The models to not assume this. Clausius Clapreyon is explicit only at the surface. But models are tuned so that it is an emergent property anyway. See AR4 WG1 black box 8.1. And that tuning is wrong. Most of the observational evidence shows that specific humidity at altitude (where it counts) does rise,but not nearly as much as constant relative humidity would require. I.e., a weak water vapor feedback Bode f 0.25-0.3 rather than 0.5 as AR4 and 5 would have it. Lindzens adaptive iris and closely related Eschenbach thunderstorm regulator hypotheses provide adequate theoretical explanations for the growing body of humidity observations. Covered in essay Humidity is still Wet in my ebook Blowing Smoke, foreword by Judith Curry.

      • ristvan, thanks again for a cogent explanation of mathematics as they are applied in climate models.

        I have to confess to preferring booty curves to Bode, engineer or not.

  10. “…cloud feedbacks perforce must be net negative, not net positive as the models claim.”
    Exactly. Were it otherwise, runaway warming would have been a feature of Earth’s history, which it is not.

    • Positive feedback is NOT a guarantee a runaway warming.

      Positive feedback is used all the time to INCREASE GAIN. Early TRF radio receivers used “regenerative” positive feedback to increase sensitivity. A control was provided for tweaking the feedback to increase the gain short of the point of oscillation (it whistled).

      Automobile drum brakes (cheap ones) used ” two leading shoe” brake designs to increase brake friction. The rotation of the wheel forces the brake shoe against the drum. Main problem is brake fade. When the shoes get hot the coefficient of friction drops and that reduces the positive feedback so you lose braking effect. More expensive designs used “Two Trailing shoe” designs where the shoes tend to unwrap (negative feedback). This works in the opposite direction stabilizing the brake fade , but usually required additional pedal pressure. (non hydraulic brakes).

      Positive feedback increases variability, but doesn’t lead to oscillation or runaway unless the gain exceeds unity. Well there’s a more stringent definition of the stability condition, but I doubt that models of the climate exist that conform to what a real feedback system is, so that rigorous feedback theory applies.

      In my view, the entire story on climate stability and cloud feedback is told in the SCIENCE paper by Wentz et al for July 13 2007 “How much More Rain will Global Warming bring. Well the paper hints at it. They just didn’t state the obvious conclusion which is as plain as your face. ( more rain MEANS more clouds. Doesn’t require any fancy explanations.


    • And considering the ice core record temp’ proxy patterns for the last few million years all indicate the warm interglacials are the exception rather than the norm (~10-15,000 years out of each approx. 100,000 year cycle), we have more to worry about wrt. “runaway cooling” than runaway warming!

  11. The previous Monckton post and this Eschenbach post illustrate science at work, hopefully leading to significant outcomes.

    I noted that Monckton post refereed to one of his commentators ‘George White’ who influenced his thinking and the content of the post.

    Eschenbach read Monckton post which had a reply from ‘Nick Stokes’ and Eschenbach followed up some more.

    This reminds me of the great search for the secrets of the atom with various researchers observing some effect, published it, waiting for paper to travel the world, another researcher reads said paper and has a ‘light bulb’ moment.

    Except these days research/comments travel the internet at a rapid rate compared to the speed of communications in the 1900’s.

    • ‘The previous Monckton post and this Eschenbach post illustrate science at work, hopefully leading to significant outcomes.”

      Errr. no.

      • Steven, you are getting worse. I asked:

        Please be clear about your objections by quoting the exact words you object to, so we can all understand the exact nature of your disagreement.

        But noooo, requests for clarity are for the little people, not the great Mosher. You quoted something, but it was obviously not what you are claiming is the unknown error. Once again you can find nothing to complain about in my science, so you resort to drive-by ugliness.

        Come back when you are willing to actually explain what it is that you are mumbling about, and I’m happy to talk about it.

        Until then?

        Not interested, amigo, talk to the hand.


      • Mosh,
        Please enlarge the quotes below and hang it over your desk at work so everyone else can understand why you say the things you do.

        “It is difficult to get a man to understand something, when his salary depends upon his not understanding it!”
        – Upton Sinclair

        Skeptic climate science is not science as I understand it.
        – (pick any CAGW climatist name who feeds from the public trough)

      • Poor Mosh.

        If that’s the best you can manage as a Literature BA….

        …no wonder your understanding of science is basically non-existent.

      • Actually, the human brain does pay attention to every letter. Current research using functional brain imaging demonstrates that good readers see every letter as a lettergroup-sound association which fires in separate neurons. They just all fire at once.

      • PG, actually you can train yourself out of word/sound (which forms in elementary school based on ‘See Spot run’), I learned speed reading in an advanced elective high school course (ditto touch typing) and it has served me well professionally ever since. Eye takes in word groups rather than words (which would get instinctively ‘sounded’) (the best speed readers can do a whole line at a time, I stick to ~half lines dictated by punctuation except for newpaper or patent columns), and let the meaning of that word group instantly register. Next group, next meaning in the thought chain. Very fast at getting not just ‘gist’ but logic structure, main points, and supporting evidence outline . Now speed reading is not about savoring a Clemens or Hemingway peice of great literature where every word is polished and nuanced. But when you are reading boring corporate reports, government docs, business proposals, science papers, or newspapers and magazines, you can increase throughput ~10x, or equivalently accomplish the task in a tenth the time. In my world, that has proved an enormous edge. But OTH I now find reading great literature hard and frustrating. Always wanting to speed up, thereby getting the story but missing the greatness.

      • WN, not sure what your comment target was. Do you speedread? Has little to do with smarts. Has only do with rate of basic comprehension of written stuff. Sounding out written words as if spoken is slow. Recognizing writing blocks per se is fast. Just takes practice. Was my only point.

      • Ha ha!
        I was wondering what Mosh would have to say and when he’d pop in!

        Bit disappointed though…. he just came in, bared his buttocks, and went away again.

  12. Very interesting analysis Willis. Hope that wedding is as fun as all the weddings I have attended. I will also be interested to see what the usual alarmist defenders have to say about it as well. Science is good.

  13. More CO2 causes more warming which increases water vapor which causes more warming which………….Wait a minute. Isn’t this a perpetual positive feedback loop? So, I’ll just cross the street and grab a boiled lobster out of the ocean?

    • Not necessarily. If an increase in CO2 adds one unit of heat so that the air can hold enough extra water vapor to contribute an extra 1/2 unit of heat, which increases wtaer vapor and this produces another 1/4 unit of heat, etc, the result is an infinite sum of terms but a finite sum (2.0).

      (I’m not sugessting my factor is correct, just illustrating that one can have
      infinite terms with a finite sum.)

      • ToP, you will be amused to know the ‘official’ IPPC factor is exactly 2. Doubling CO2 absent feedbacks 1.2C, with water vapor 2.4C, with additional positive cloud feedback 3.0C. All the other stuff pretty much cancels out to rounding error. Details posted in precious comments on other threads. Translate to Bode f, IPCC WVF +0.5, clouds +0.15, total 0.65=>ECS 3.0.

    • No, siamiam, that would only happen if the feedback were impossibly large.

      For example, If a 1°C temperature increase causes enough additional water vapor to enter the atmosphere that it causes an additional 0.2°C temperature increase, then that causes an additional 0.04°C temperature increase, which causes an additional 0.008°C temperature increase… and you’ll never get very warm. Here’s a simulation:

      temp = 1.0
      increase = 0.2 * temp
      while True:
        print "temp = ",temp
        temp += increase
        increase = increase * .02
      temp =  1.0
      temp =  1.2
      temp =  1.204
      temp =  1.20408
      temp =  1.2040816
      temp =  1.204081632
      temp =  1.20408163264
      temp =  1.20408163265
      temp =  1.20408163265
      temp =  1.20408163265
      • Oops. that should have been .2, not .02:

        temp = 1.0
        increase = 0.2 * temp
        while True:
          print "temp = ",temp
          temp += increase
          increase = increase * .2
        temp =  1.0
        temp =  1.2
        temp =  1.24
        temp =  1.248
        temp =  1.2496
        temp =  1.24992
        temp =  1.249984
        temp =  1.2499968
        temp =  1.24999936
        temp =  1.249999872
        temp =  1.2499999744
        temp =  1.24999999488
        temp =  1.24999999898
        temp =  1.2499999998
        temp =  1.24999999996
        temp =  1.24999999999
        temp =  1.25
        temp =  1.25
        temp =  1.25
      • Thanks, Dave. Or alternatively, the end result is ∆T / (1-feedback). Where feedback is .2 as above, you end up with 1/(1-0.2) =1.25 as in your example


    • “More CO2 causes more warming”

      Still waiting for some proof that CO2 causes warming in an open atmosphere governed by convection and conduction.

      • Actually thermalization explains why CO2 (or any other noncondensing ghg) has no significant effect on climate. (Kinetic theory of gases and a smidgen of quantum mechanics)

  14. I am very pleased to see this topic come up. Thank you Willis

    I have posted here in the past that according to me experience clouds do not always result in cooling. Right throughout the winter months at my latitude (37 S) clear skies means cool in the shade during day and rapid cooling at night.

    The NZ Metservice has been very accurate in its weekly forecasts over latter years. Invariably their predicted average daily temperatures during clear skies (yes NO cloud!) in winter are cooler than during cloudy days – often by 3 C. Remember too that during winter the nights are much longer than the days. (derrrrr) :-)

    I too find the great variation in the models interesting. Ya don’t really know do ya! (not @ Willis) Ya could be out by a country mile

    • There are some studies which suggest that low altitude clouds have a net cooling effect while higher altitude clouds have a net warming impact. If true, it means one has to be very cautious about treating clouds as an amorphous group, especially if changes in temperatures have a differential impact on low altitude versus high-altitude clouds.

      • High thin cirrus always warm because themice crystals are transparent to sunlight but opaque to infrared. Essence of Lindzens adaptive iris–the iris is cirrus. Cloud effect depends on type, altitude, optical depth (and entrained precipitation). Essay Cloudy Clouds unravels this a bit.

      • Outdoor experience while working says that even cirrus clouds are the worker’s friend. Even a thinnish cirrus layer results in cooler temperatures at ground level. The radiation that doesn’t reach the ground is more important to you comfort than what doesn’t leave. When it is asserted that the “themice” crystals are opaque to infrared, does that mean the clouds absorb that energy or reflect it.

  15. Do I have this right?
    (a) It seems that most of the positive correlation is over the oceans.
    (b) Over the oceans the temperature is measured as sea surface temperature (not surface air temperature used over land).
    (c) Sea surface temperature has changed less over the last decades compared to land air temperatures.
    Question: Shouldn’t that in itself account for most of the strongly positive feedback regions?

    • Fig. 2:
      – Where there is cold upwelling in the oceans, there is a net positive correlation.
      – Where there are warm currents in the oceans there is a positive correlation
      – Where there are cold currents in the oceans there is a negative correlation
      – At the same latitude land has a more negative correlation than oceans
      – Land as a whole has a more negative correlation than oceans
      – The Southern Hemisphere (SH) as a whole has a more positive correlation
      (All: in this period)

      The most anomylous warming in the oceans (today, summer 2016) is visible at the high latitudes of the NH:,90.49,466/loc=-166.267,57.880
      The most negative correlation between ECR and rising temperature (in the period of fig. 2, 2000-2015) is visible at the land surfaces in the NH, more or less at the same latitude where todays’ most anomylous warm oceans are visible.

      Does some warming lead to warmer oceans at higher latitudes of the NH (and ice melting at the N. pole), while the continents at that same latitude react by cooling the surface by their clouds and in this way (in time) are restoring a certain temperature equilibrium?

  16. I don’t know what data is available but I think a separation between day and night would be informative. The feedback during the day is probably even more negative while at night may actually be positive. This is important because it means the days are cooler which should then reduce water vapor feedback as well.

    • And a separation between the seasons. I’d expect the effect would have the opposite sign in summer and winter (excluding the tropics where the are no real seasons).

    • I understand clouds to be a classic negative feedback, IE a damping signal that opposes the change therefore it makes cold nights warmer and hot days cooler. It is a thermal mass.

  17. Willis,
    “This is very bad news for the models … they all claim that there is a positive correlation..”

    If such feedback was positive, wouldn’t the atmosphere be unstable? That is, wouldn’t a small warming create feedbacks that cause more warming etc, causing temperature to rapidly accelerate?

    • See upthread comment to CH. The positive feedback weakening of a damped system is NOT the same as an undamped or positive feedforward system. The instability argument has been misunderstood/misused by many skeptics, some prominent like Monckton. Don’t overplay the hand.

    • The feedbacks always dampen out at some point because temperatures rise at the fourth power of the energy / forcing level.

      Even if you ramp up the feedback values to extreme levels so that the temperature increase gets above 100C (and the oceans/water/volatile gases in the soil, boil away and enter the atmosphere, and we end up with a Venus-like scenario), eventually the Stefan Boltzmann (temperatures rise to the fourth power of energy) takes over and temperatures stabilize at +450C or so.

      The runaway eventually reaches a peak temperature. Its just the way the math and the physics work. Even the stars reach a peak temperature when fusion starts up (and the peak follows the stage of fusion that is taking place). This is the way the universe works. Stefan and Boltzmann described it extremely accurately. The only issue climate science won’t touch is how gravity and pressure influence those temperatures because inside the core of stars temperatures can reach close to a billion C, our Sun is at 15 millionC, Jupiter is 10,000 C in its core, Earth is 6,000C as hot as the outside temperature of the Sun, what is the temperature inside a neutron star or a black hole, what is the temperature during a supernova or a black hole merger, why do objects heat up as they enter into a gravity field. Sorry, just something not accounted for in my view.

  18. Clear sky at night = cooling – or- am I missing something? We increase the thickness of blankets at night in winter. – those of us without heating. What is so hard to understand?

    • Yes, nightime radiative cooling is hindered by clouds when present. Ditto humidity, whichnismwhy deserts cool so rapidly at night. But you have to look at the 24 hour net effect. Cloud albedo cools more during the day than it hinders cooling at night. Cause all the warming energy for the whole world only arrives during daytime. Its called sunlight.

      • Ditto humidity

        I have often questioned whether the presence of clouds is more significant for their role in restricting convection and that humidity is likely to be higher on cloudy nights such that irrespective of the radiative budget, it will take longer for the atmosphere to cool since the higher humidity has more energy that needs to dissipate and with reduced convection, it will inevitably take longer to dissipate that energy. Thus nighttime temperatures will hold up and it will take longer to cool.

        It may be this process that leads us to consider that cloudy nights are not as cold.

      • JB, there is a famous law school and NOT apocryphal story about the great third SCOTUS Chief Justice, John Marshall. He ordered casks of sherry from Spain for the SCOTUS chambers, and was in the habit of offering sherry to his colleagues and himself every afternoon. He began noticing this slowed the courts productivity some. So as Chief, he ordered that they would only drink in the afternoons if it was raining. This experiment unfortunately slowed productivity further, as his colleagues would simply go out for a drink. So one day he assembled the court on a sunny afternoon, pronounced ‘Gentlemen, it is raining somewhere’ and offered all sherry. The tradition lasted until his death in 1835.
        The law school joke is the legal distinction between raining, and raining somewhere. And thus was constitutional law legal precedent built.

    • Don’t forget when clouds form they release their heat of condensation and directly warm the place up. This is an awful lot of heat. The effect is big enough to completely overwhelm the adiabatic cooling which is also going on at the same time and which is directly caused the drop in pressure which usually accompanies the formation of clouds. The adiabatic cooling is around 2.5 degrees Celsius for the typical pressure difference between high and low pressure systems (3.5% change in pressure =approx= 1% change in temperature). Temperatures don’t just change because of the effect on the radiative balance. There is a lot going on.

      • And this happens at altitude so this heat is removed from the surface and radiated into space. Actually this process cools the surface more than LWIR radiation. Even according to IPCC, though they hide it pretty well.

    • The atmosphere cools the planet when it is overly warm.

      funny sort of blanket. !

      The only reason the clouds are up there is because they have already done their job.

      What is so hard to understand ???

      • Yes, I used to regard clouds as just another potential climate variable among many, but lately feel them more like inevitable companions to oceans in particular. As if Earth was neatly positioned atca distance from the sun that allowed indeed demanded constant interplay between the 3 water phases. Clouds gave to form, physics dictate this and they have to be coupled to oceans.
        One of the nicest bits of common sense was the comment by Willis that the time of day when big low clouds appeared over tropical oceans was an effective climate control for temperature. The surface Watts per sq m involved in going from no cloud to cloud for an hour as measured are impressively large compared to GHG radiation changes as postulated.

  19. Despite these huge variations, all of them can (relatively) successfully emulate the historical record when they are each fed their own special brand of forcings … which should tell us something about the models. But I digress.

    Given this vast variation in the variables, the fact that the model forecasts are not as closely grouped as is their hindcasts tells you all you need to know about their worth.

    Why are the forecasts of the various models so widely spread given the tight grouping of the hindcasts of these very models?

    Something very odd is afoot.

  20. Did I just overlook it, or did no one mention Svensmark? His cloud – cosmic ray – solar wind theory seems to tie it all together.

    “Svensmark publishes: Solar activity has a direct impact on Earth’s cloud cover”

    Thanks, Willis

    • George Hebbard PE September 4, 2016 at 6:12 pm

      Did I just overlook it, or did no one mention Svensmark? His cloud – cosmic ray – solar wind theory seems to tie it all together.

      George, the problem with Svensmark’s latest paper is simple. The paper claims that we get cloud changes from “Forbush events”, when the number of cosmic rays dip low and then come back up a few days later.

      The problem is, the amplitude of the Forbush events is only about HALF of the change in the number of cosmic rays that we see each and every sunspot cycle. So whatever effect they claim is happening in the Forbush events, we should see TWICE that change over the ~ 11 year sunspot cycle.

      But we don’t … it doesn’t kill his theory, but it is certainly something that he will need to explain. And to date, as far as I know he hasn’t even acknowledged the problem, much less begun to deal with it.


      • Well, not necessarily. Forbush are short term, the change over the solar cycle is long term, giving lots of time for the atmosphere to react, or adapt. Sudden change vs slow change.

      • Kim, I think Willis has a strong observational point. We have ICOADS and ISCCP cloud date for more than one solar cycle. There is no evidence of a solar cycle in that data. None. Essay Cloudy Clouds analyzes the data.

      • Thanks, Rud, observations trump theory, but a pulse is more easily detected than a slow roll.

      • kim September 5, 2016 at 5:53 am

        Well, not necessarily. Forbush are short term, the change over the solar cycle is long term, giving lots of time for the atmosphere to react, or adapt. Sudden change vs slow change.

        While that may be, it simply reinforces my point. If there is some natural adjustment mechanism that neutralizes the effect of the cosmic rays, then they don’t make any difference

        And of course, that fits exactly with my hypothesis that changes in things like the timing of daily emergence of cumulus and thunderstorms and dust devils neutralize not only changes in cosmic rays but changes in forcing of any type …


      • kim September 6, 2016 at 6:58 am

        Thanks, Rud, observations trump theory, but a pulse is more easily detected than a slow roll.

        Huh? if I have a gauge in a lake, whether it takes a year or a day for the lake to drop a foot, I can detect it either way. The same is true for say TSI, or for sunspots. I don’t believe your claim about slow versus fast.

        And despite stories to the contrary, boiling frogs slowly does NOT mean that they won’t notice …


      • Thanks, w; I’m still clinging to the wreckage here. Think of all the vortices possible in the sun-ocean-atmosphere system in which to hide a myriad of tiny changes urged by cosmic rays. I think this can even support your thermostat, even giving it eyes.

        I understand your point, too, and since it is better articulated, I’ll yield. For now.

  21. I note in passing that I found in my research reported in “Precipitable Water Redux” that the increased absorption of upwelling surface radiation from water vapor was between three and four W/m2 per degree C of warming. Since half of this goes up and half goes down, that puts the water vapor feedback as being on the order of 1.75 W/m2 per degree C of warmings.

    Soden and Held’s value is 1.8 ± 0.05 W/m2 for the water vapor feedback … so in that case CERES basically agrees with the models. It disagrees strongly, on the other hand, with the cloud feedback.


    • A question Willis.

      Are you using the surface temperature data provided by CERES?

      Because water vapor feedback could indeed be a lower value if one uses the adjusted surface temperature from the NCDC/NCEI. Using the adjusted surface temperature data as the denominator, I always get around 1.0 W/m2/C.

      Maybe the real numbers work but it is a different matter if the temperature trend is exaggerated.

      • Bill, always good to hear from you. I am using the CERES data. I’m not sure what you are calling “adjusted surface temperature” … adjusted for what?


    • Interesting observation. But not the entire story, since what matters most is the upper troposphere specific humidity, not all the precipitable water lower down. The effectively constant rUTH with altitude is where the model flaw lies thatbove states WVF.

    • Water is treated as feedback only in the models. According to modtran there is no radiation looking down in the water bands until you get to nearly 5 kilometers. Also according to modtran when you get to 40km looking down water alone is radiating to space at 329 W/m2, about 97% of TSI.

      (This is from an up vs down exercise and the red is looking down. The blue looking up is barely visible in the low wave numbers and amounts to only .17 W/m2 at this altitude)

  22. Ristvan@2:30
    If cirrus clouds are transparent to sunlight, how can I see them? Am I completely dumb here. What have I missed?

    • I think the statement that cirrus clouds are transparent to sunlight needs to be quantified. For example if you have sunny day and an approaching warm front brings a cirrus cloud shield in, there is no doubt this lowers temperatures due to decreased incoming solar. So cirrus clouds are not completely transparent and the thicker the cirrus cloud layer the less solar radiation makes it through. However, I think the word transparent in the case of cirrus is that it allows more energy through it that it allows to escape regardless of the thickness. There are other optical features to consider as well. Some cirrus clouds have near identical crystals and this causes the light to reflect and refract off the ice crystals causing sun halos, sun dogs and icebows. If this light is being scattered or redirected in this way then some must go back into space.

  23. Global Solar Radiation and thus Net Radiation at the ground level are function of bright sunshine hours — in addition to other localised factors –. In the absence of sunshine data, I presented a model to derive sunshine hours from cloud cover — lower, middle and high clouds — [Solar Energy, 1974, 15:281-285, Pergamon Press, Printed in Great Britain].

    S = 1 – f1 + f2;
    wherein S = n/N is the ratio of actual hours of sunshine to the theoretical duration of sunshine;

    f1 = a x e [-0.25√a] — the value in the brackets is exponential of that term]

    a = [Cl + Cm + Ch]/8

    wherein Cl, Cm & Ch respectively are amounts of low, medium and high clouds, mean of 0830 and 1730 hours IST observations, in Octas, varies between 0 to 1 as sky condition changes from clear [zero octas] to overcast [8 octas] and e is the exponential function

    f2 = latitude correction, 0.02 + 0.08 Cos 4φ upto 45 degrees latitude and -0.06 for latitudes beyond 45 degrees latitude, φ is the latitude of the place in degrees

    Dr. S. Jeevananda Reddy

  24. “The models say that cloud feedback is a 0.69 ± 0.10 W/m2 INCREASE in downwelling radiation for each additional degree of temperature.”

    Another thing with this is that the temperature is about 1 degree warmer now than in preindustrial times, that should give 0.69 ± 0.10 W/m2 increase in the dowelling radiation. However, this is of comparable size to the current global warming by all direct an indirect effects combined:
    “Earth’s energy imbalance:
    Earth’s energy imbalance measurements provided by Argo floats detected accumulation of ocean heat content (OHC). The estimated imbalance was measured during a deep solar minimum of 2005-2010 at 0.58 ± 0.15 W/m².[11] Later research estimated the surface energy imbalance to be 0.60 ± 0.17 W/m².[12] ”
    – Wikipedia: Earth’s energy budget
    (This estimate is about the same as several other estimates)

    The cloud feed-back effect hypotesized hypotesized by United Nations climate panel, does not leave any room for the direct effect from CO2 and all the other direct and indirect effects the climate panel is so confident about.

  25. I was surprised to see such a weak correlation between CRE and temperature across the interior of Australia, where cloud cover has very large effects on temperature. Max temperature reductions of 20C are not uncommon in summer on cloudy (and rainy) days, and the effect still exists in winter, albeit much smaller.

    Which indicates to me that whatever increase in humidity occurs due to rising temperatures is mostly restricted to the already humid areas and the increased humidity quickly precipitates out. Increasing vertical heat transport by the water cycle (a negative feedback), and there is little or no global effect from increased WV.

    • Anywhere there is a Horse Latitudes High Pressure Zone you are going to have a correlation. It’s the one place where surface warmth and clouds rarely go together. But even there, at times, the Monsoon extends its reach. But that is rare enough that it merely softens the correlation.

  26. Willis:

    Clearly, this is a hugely complex system, where in some parts of the world the correlation is strongly negative, and in some parts it is positive. Note that the inter-tropical convergence zone (ITCZ), which I have long held is a crucial part of understanding the climate, is negatively correlated. And so is the land area north of about 50°N or so.

    Overall, we can calculate the global correlation by looking at the global area-weighted average net CRE versus area-weighted global average temperature.

    I wasn’t clear if your results were simply area weighted or also latitude-adjusted ?

  27. Willis,
    Thanks for an interesting post. The CERES correlation map is particularly interesting, although I can’t get excited about the global regression, as the R-squared is only 0.07.

    So far as GCMs go, note that things have moved on since the Soden and Held (2006) feedback analysis of CMIP3 models. For CMIP5, the principal feedback analysis used is per Vial et al (2013), available at: .

    Jessica Vial found that much of the warming influence of clouds in CMIP5 models arising from cloud changes increasing the forcing from increased CO2 concentration; see Table 2. As this effect is independent of surface temperature changes, it is classified as a (fast) adjustment and taken into account in the difference between radiative forcing (RF) and effective readiative forcing (ERF) – see the discussion at the start of Ch.8 of IPCC AR5. The surface temperature dependent cloud feedback (Table 3) is then quite small on average (~0.35 W/m2/K), and is negative for a few models.

    • niclewis: I can’t get excited about the global regression, as the R-squared is only 0.07.

      All that means is that a 1C change in mean temp would need to be sustained many years for the multiple-year mean change in CRE to be close to -1. In all of this discussion of CO2 and climate, the interest is in sustained changes in “forcings” and “feedbacks” and their long term effects when R^2 is small for everything of potential interest.

  28. Questions have been proposed regarding why climate models misuse cloud feedbacks. Follow the money (if you can). Models are built on sexy green watermelon grants. My hunch is that natural climate driver model (there is one model, but I believe only one model and it is likely guaranteed to not simulate observations) grants are sooo not sexy, not to mention unavailable. However, the opaque wall between CESM- -and funding body leanings is nearly impossible to bridge. The careful use of wording keeps the common woman (me) from deciphering the funding bias.

    For starters if you have 60 hours to spare, try to find the natural climate model:

  29. Seems like a comment I made must have triggered the nuclear option. Dunno why, but it was up there for a while. Totally benign, and can’t imagine why it would get disappeared, either by accident or by chance or deliberation.

    Oh well. Hardly worth reposting.


  30. “ Effects of Clouds on the Earth’s Radiation Budget
    The effect of clouds on the Earth’s present-day top of the atmosphere
    (TOA) radiation budget, or cloud radiative effect (CRE), can be inferred
    from satellite data by comparing upwelling radiation in cloudy and
    non-cloudy conditions (Ramanathan et al., 1989). By enhancing the
    planetary albedo, cloudy conditions exert a global and annual shortwave
    cloud radiative effect (SWCRE) of approximately –50 W m–2 and,
    by contributing to the greenhouse effect, exert a mean longwave effect
    (LWCRE) of approximately +30 W m–2, with a range of 10% or less
    between published satellite estimates (Loeb et al., 2009). Some of the
    apparent LWCRE comes from the enhanced water vapour coinciding
    with the natural cloud fluctuations used to measure the effect, so the
    true cloud LWCRE is about 10% smaller (Sohn et al., 2010). The net
    global mean CRE of approximately –20 W m–2 implies a net cooling…..”

    “TS.6.2 Key Uncertainties in Drivers of Climate Change

    • Uncertainties in aerosol–cloud interactions and the associated
    radiative forcing remain large. As a result, uncertainties in aerosol
    forcing remain the dominant contributor to the overall uncertainty
    in net anthropogenic forcing, despite a better understanding of
    some of the relevant atmospheric processes and the availability of
    global satellite monitoring. {2.2, 7.3–7.5, 8.5}

    • The cloud feedback is likely positive but its quantification remains
    difficult. {7.2}

    • Paleoclimate reconstructions and Earth System Models indicate
    that there is a positive feedback between climate and the carbon
    cycle, but confidence remains low in the strength of this feedback,
    particularly for the land. {6.4}”

    Not exactly news.

  31. Willis, congratulations.
    I do hope the wedding was a great success and that you are still reveling in all the joy.

    Cloud feedback is an issue I’ve struggled to understand.
    If I accept that clouds increase downwelling radiation shouldn’t a lot of cloud provide a lot of downwelling radiation? And a lot of cloud for a lot of hours, or even days, should provide the most downwelling radiation, right? So does this mean that if we have cloud cover for 24 hrs a day for ever, the downwelling radiation will increase earth’s temperature? If so, how do we square that with our own experience that cloudy days are invariably cooler than sunny days?
    If this is really sophomoric, I apologize, but I can’t for the life of me come up with a simple cocktail party explanation of why, if clouds provide positive temperature feedback, lots of clouds are synonymous with cool days.

    • John, at night it is true that clouds have a warming effect, due to downwelling longwave (invisible) radiation.

      In daytime, clouds reflect a great deal of incoming solar radiation (i.e., they shade the ground), which has a cooling effect, and that cooling effect greatly exceeds the warming from downwelling longwave radiation.

      So, whether clouds warm or cool depends on time of day.

      As an additional complication, the effect on temperature of clouds made of ice crystals is different from the effect on temperature of clouds made of water droplets, due to the differing spectra of H2O in its different states.

      What’s more, cloudiness varies with time of day. Willis did some especially lovely work on one aspect of that, last year:

      So the answer to the question of whether clouds warm or cool is, “it depends.”

      • Yes – and on degree of latitude and season. In winter at higher latitudes the angle of incidence of sun energy during a clear sky is such that albedo restricts service heating. While we may feel warmed by the sun, much of the energy reflects directly back out to space. The ambient temperature in the shade is low, even at midday.

        I know this through personal experience as a grassland farmer. During the 3 months of winter grass growth is always higher during periods of cloudy days than sunny days. This is due to soil temperatures which are higher during cloudy weather and at their lowest during periods of sunny days and cold nights (frosts). We never get frosts under a cloudy sky

        I think that those that study such things may get stuck in what occurs in the tropics. The cooling mechanisms are at their most effective at higher latitudes

        There is a lot of ‘yes but’ in this topic



      • daveburton, thank you and Michael Carter for your replies. They’re helpful.

        But, in the end, as Willis points out, “they (the models) all claim that there is a positive correlation between CRE and temperature, which makes the model-projected warming much larger”. His data analysis says the opposite and that, at least to me, makes more intuitive sense.

  32. Mr. Eschenbach:

    You state that the models “…claim that there is a positive correlation between CRE and temperature,
    which makes the model-projected warming much larger … but in fact the global average correlation is negative.” It seems to me me that this statement confuses the global average surface temperature with the global average temperature at equilibrium. As the latter quantity is not observable it seems to me that the claim of a positive correlation is not amenable to testing.

  33. Wow, Willis, I cannot believe the volume and breadth of the comments on this quite concise post. (Which I enjoyed, by the by)

  34. I am late to this party as I have been hiking, but I have a few issues with the lead post.

    but in fact the global average correlation is negative.
    And from this same data, of course, we can calculate the global average cloud feedback parameter

    This falsely assumes that all the change in global average CRE is caused by only temperature changes. Willis and the climate modelers assume that clouds change only by a temperature change, which is a feedback. But clouds change for other reasons, which cause a temperature change, which is a forcing. Dr. Roy Spencer has written extensively about cloud changes causing temperature changes, which contaminates the correlation of clouds to temperature when estimating feedbacks. Clouds change in response to ocean circulation changes, resulting in forcings. The huge amount of noise in the cloud and temperature correlation is due to cloud changes unrelated to temperature changes.

    Willis shows a graph of CRE and HadCRUT4.4 temperatures from March 2000 to February 2015. The CERES data is available to May 2016. Here is a graph of CRE, HadCRUT4.4 and UAH LT temperature data to May 2016:

    Figure 3 shows “correlation = – 0.27”. The table of Coefficients says “Std. Error” of temperature is 0.2695, and the slope of the correlation is -1.012 W/m2/°C. Shouldn’t Figure 3 say “Correlation coefficient = -1.012 W/m2/°C. Std. Error = 0.27” ?

    The table shows “Multiple R-squared: 0.0734, Adjusted R-squared: 0.0682”.
    Could someone please explain the difference in meaning between the Multiple and Adjusted version of R-squared?

    The correlation plot of CRE and HadCRUT4.4, data to Feb 2015 is here;

    It shows R-squared is only 0.0327 as calculated by Excel, which is much less than reported in the in the lead post. The relation is -0.876 W/m2/°C, which is less than the -1.012 W/m2/°C reported by Willis. Why is there this different?

    Using all the available CERES data with HadCRUT4.4 gives R2 = 0.0145, which is very poor. The relation is -0.444 W/m2/°C as shown here:

    The HadCRUT4.4 data is hopelessly contaminated by the urban heat island effect as shown by many studies. The study by McKitrick and Michaels shows that almost half of the warming over land in instrument data sets is due to the UHIE. A study by Laat and Maurellis 2006 came to identical conclusions. The satellite lower troposphere temperatures should be used.

    Using UAH lower troposphere temperatures, the R2 jumps up to R2 = 0.108, and the relation is -1.047 W/m2/°C as shown here:

    Data and graphs are at:

    Climate modelers noted the cloud cover declined from 1986 to 2000 while surface temperatures rose, and falsely assumed that only the temperature change caused the cloud cover to change, so they estimated a strong positive feedback. Here is the Climate4you graph of cloud cover:

    and here is the correlation plot from July 1983 to Dec. 2009:

    Note that the correlation shows a relation of -0.66 °C/%change cloud cover. If one (falsely) assumes the cloud cover changed only due to the temperature change, this would imply a strong positive feedback. (I don’t know how to get CRE values before CERES data, but I assume CRE is highly correlated with cloud cover.) Climate4you correctly said that global cloud cover increase is empirically “associated with” a global temperature decrease, but did not call it a feedback.

    The cloud feedback should not have changed between the 1980 – 2009 period and the 2000 – 2015 period. The fact that the cloud – temperature correlation changed dramatically proves that the cloud changes are not only caused by a temperature change, or feedback, but are also caused by forcings. The CERES CRE data is a combination of feedback in forcing in unknown proportion.

    • Ken Gregory: Figure 3 shows “correlation = – 0.27”. The table of Coefficients says “Std. Error” of temperature is 0.2695, and the slope of the correlation is -1.012 W/m2/°C. Shouldn’t Figure 3 say “Correlation coefficient = -1.012 W/m2/°C. Std. Error = 0.27” ?

      It is a coincidence that the correlation and the standard error of the regression coefficient have the same 2 leading digits. the value of -1.012 is the slope (b) of the equation y = a + b*x from the computer output; with y being cloud radiative effect and x being temperature (the slope of the temperature effect is labeled “temperature” in the output.) So, Figure 3 is correctly labeled.

      Note that the correlation shows a relation of -0.66 °C/%change cloud cover. If one (falsely) assumes the cloud cover changed only due to the temperature change, this would imply a strong positive feedback.

      How is that a positive feedback?

      • matthewrmarler;
        How is “correlation = – 0.27” correct? The correlation table provide does not show this number anywhere. It isn’t the slope of the regression line. It isn’t a correlation coefficient. What is it and how was the -0.27 value determined?

        Oop, my typo. I should have wrote “Note that the correlation shows a relation of -0.066 °C/%change cloud cover.” Not -0.66. This is from the equation on the chart “Y= -0.06591 * X + 19.6378” and is the slope of the regression line. It says that a -0.066 °C change in temperature is associated with a 1% point change in cloud cover. If the temperature change caused the change in cloud cover, then a temperature increase of 0.066 °C caused a 1% point reduction in cloud cover, which allows more sunlight in to heat the earth’s surface. If an increase in CO2 caused an initial temperature increase, then cloud cover would decrease, allowing in more sunlight causing a further temperature increase, thereby amplifying the initial temperature change. That is a positive feedback.

        However, the CERES data from March 2000 to May 2016 shows the opposite effect. It shows an increase in temperature of 1 °C caused a 1.05 W/m2 decrease in CRE, which implies that the cloud cover increased, assuming that the CRE and cloud cover changes are due to the temperature change. The CRE becomes more negative when cloud cover increases. The increase in clouds block some sunlight from the earth, causing a temperature decrease that offset some of the initial temperature increase. That is a negative feedback. But cloud feedback can’t change sign between the two periods, proving that the initial assumption that the cloud changes were cause by only the temperature change is incorrect. The correlation does not determine the cloud feedback factor. As per Dr. Spencer’s work, the changes in ocean circulation caused some of the cloud changes. See;

    • Regarding the statement that the satellite lower tropospheric temperature data should be used because McKitrick & Michaels shows almost half the land warming to be from growth of UHIs: If the land component of HadCRUT4 is reduced by half, HadCRUT4 would still show more warming than the satellite datasets of the lower troposphere show. This is because an increase of greenhouse gases increases the lapse rate. Figure 7 in shows the near-surface temperature warming more than the lower troposphere as a whole as indicated by radiosondes.

    • I showed that the large decline in cloud cover % from 1986 to 2000 suggests that this is a positive cloud feedback if one falsely assumes that the cloud declined only due to the rising temperatures during the period. However, the CERES data from March 2000 to May 2016 shows the opposite effect when comparing cloud radiative effect CRE) to temperature. It shows an apparent negative cloud feedback. I wrote, “I assume CRE is highly correlated with cloud cover.” I couldn’t find cloud cover on the CERES page, but the CERES help wrote me “We have both cloud coverage and cloud radiative effect (CRE) top-of-atmosphere and surface fluxes available”. Here is a plot of cloud cover vs CRE;

      Wow! It shows for the period 2000-03 to 2016-02 there is a coefficient of determination R-squared = 0.0005. I know that clouds have different layers that behave differently, but this low R2 is surprizing to me.

      • I should have plotted cloud cover % anomaly vs CRE anomaly, rather than cloud cover %.
        The R2 increases to 0.0013, which is still very low! This is still surprizing. Why is it so low?

  35. More evidence from boots on the ground

    Further to my posts above that discusses the importance of land surface albedo at latitude 37 Sth during clear skies in winter (enhanced cooling as compared to cloud insulation)

    During clear sky weather, paddocks that slope to the north (sun direction) > 15 degrees grow 3 x more grass then those that are flat or slope to the south. How do I know? Through the number of sheep I can run/hectare/paddock. This relates directly to soil temperature

    The only conditions in which the flat/sth-sloping paddocks will grow grass is cloudy weather with periodic rain. Cloud does NOT always equate to cooling

    Field observation still has a place in science

    • Spring has arrived and summer not far off. I am going to be following the NZ Metservice forecasts throughout summer.. I believe that I may be able to demonstrate that cloud cover in summer can also be higher in mean daily temperature than clear-sky conditions. This is due to rapid night-time cooling

  36. Clouds are Nature’s heat sinks. CuNim are tower heat sinks. As such, Nature’s heat sinks take the heat up off of “the board” and get it to where it can dissipate into space.

    • Thanks, James. I’ve compared CuNim towers to “heat pipes” instead of “heat sinks”, because the warm moist surface air passes through them from surface to upper troposphere without exchanging significant radiant energy with the surrounding atmosphere.


  37. “warm moist surface air passes through them from surface to upper troposphere without exchanging significant radiant energy with the surrounding atmosphere”.

    More efficiently than during clear sky conditions at night? At any given time half the globe is in darkness and much of this surface area has no cloud. According to the cloud = cooling theory these conditions always contribute to warming. It don’t equate. The theory is too simplistic (IMO).



Comments are closed.