RealClimate logo

Feedback on Cloud Feedback

Filed under: — group @ 9 December 2010

Guest article by Andrew Dessler

I have a paper in this week’s issue of Science on the cloud feedback that may be of interest to realclimate readers. As you may know, clouds are important regulators of the amount of energy in and out of the climate system. Clouds both reflect sunlight back to space and trap infrared radiation and keep it from escaping to space. Changes in clouds can therefore have profound impacts on our climate.

A positive cloud feedback loop posits a scenario whereby an initial warming of the planet, caused, for example, by increases in greenhouse gases, causes clouds to trap more energy and lead to further warming. Such a process amplifies the direct heating by greenhouse gases. Models have been long predicted this, but testing the models has proved difficult.

Making the issue even more contentious, some of the more credible skeptics out there (e.g., Lindzen, Spencer) have been arguing that clouds behave quite differently from that predicted by models. In fact, they argue, clouds will stabilize the climate and prevent climate change from occurring (i.e., clouds will provide a negative feedback).

In my new paper, I calculate the energy trapped by clouds and observe how it varies as the climate warms and cools during El Nino-Southern Oscillation (ENSO) cycles. I find that, as the climate warms, clouds trap an additional 0.54±0.74W/m2 for every degree of warming. Thus, the cloud feedback is likely positive, but I cannot rule out a slight negative feedback.

It is important to note that while a slight negative feedback cannot be ruled out, the data do not support a negative feedback large enough to substantially cancel the well-established positive feedbacks, such as water vapor, as Lindzen and Spencer would argue.

I have also compared the results to climate models. Taken as a group, the models substantially reproduce the observations. This increases my confidence that the models are accurately simulating the variations of clouds with climate change.

Obviously, climate skeptics are quite upset with my results. Dr. Roy Spencer, for example, has been criticizing my paper on his blog. Dr. Spencer’s argument is, as he wrote in an e-mail to Dr. Richard Kerr of Science:

Andy’s study assumes that all co-variations between clouds and temperature are due to feedback, when in fact they are a mixture of feedback and “internal forcing” (natural cloud fluctuation causing temperature changes).
Now, Andy DOES at least raise this as a possibility, referencing our (Spencer & Braswell) 2010 JGR paper on the subject (his ref. #26). But he then claims that since (1) ENSO is the main source of climate variability during 2000-2010, and since (2) no one has demonstrated that ENSO is in any way caused by cloud changes, that our cause-versus-effect claim does not apply to the 2000-2010 time period.
His second claim is incorrect.
As Fig. 4a in our paper ( ) shows, the major 2007-08 La Nina event shows the characteristic looping pattern in temperature-versus-radiative flux data that results from clouds causing temperature changes

In other words, Dr. Spencer is arguing that clouds are causing ENSO cycles, so the direction of causality in my analysis is incorrect and my conclusions are in error.

After reading this, I initiated a cordial and useful exchange of e-mails with Dr. Spencer (you can read the full e-mail exchange here). We ultimately agreed that the fundamental disagreement between us is over what causes ENSO. Short paraphrase:

Spencer: ENSO is caused by clouds. You cannot infer the response of clouds to surface temperature in such a situation.

Dessler: ENSO is not caused by clouds, but is driven by internal dynamics of the ocean-atmosphere system. Clouds may amplify the warming, and that’s the cloud feedback I’m trying to measure.

My position is the mainstream one, backed up by decades of research. This mainstream theory is quite successful at simulating almost all of the aspects of ENSO.

Dr. Spencer, on the other hand, is as far out of the mainstream when it comes to ENSO as he is when it comes to climate change. He is advancing here a completely new and untested theory of ENSO — based on just one figure in one of his papers (and, as I told him in one of our e-mails, there are other interpretations of those data that do not agree with his interpretation).
Thus, the burden of proof is Dr. Spencer to show that his theory of causality during ENSO is correct. He is, at present, far from meeting that burden. And until Dr. Spencer satisfies this burden, I don’t think anyone can take his criticisms seriously.
It’s also worth noting that the picture I’m painting of our disagreement (and backed up by the e-mail exchange linked above) is quite different from the picture provided by Dr. Spencer on his blog. His blog is full of conspiracies and purposeful suppression of the truth. In particular, he accuses me of ignoring his work. But as you can see, I have not ignored it — I have dismissed it because I think it has no merit. That’s quite different.

I would also like to respond to his accusation that the timing of the paper is somehow connected to the IPCC’s meeting in Cancun. I can assure everyone that no one pressured me in any aspect of the publication of this paper. As Dr. Spencer knows well, authors have no control over when a paper ultimately gets published.

And as far as my interest in influencing the policy debate goes, I’ll just say that I’m in College Station this week, while Dr. Spencer is in Cancun. In fact, Dr. Spencer had a press conference in Cancun — about my paper. I didn’t have a press conference about my paper. Draw your own conclusion.

I hope that this post has explained my work and cleared up exactly what my disagreement with Dr. Spencer is. If interested readers do some basic research on the causes of ENSO, I’m confident they will agree with me that my interpretation of the data is sound.

Update: For those of you who enjoyed reading the e-mails referenced above on cloud feedback, Andy Dessler is continuing to post the e-mails from ongoing correspondence.

186 Responses to “Feedback on Cloud Feedback”

  1. 1
    MapleLeaf says:

    I’m assuming then Dr. Dessler that you have seen this….

    I especially like the conspiracy theories presented therein.

    What struck me about your findings is that even if there is a negative feedback short-term feedback it is probably so weak that it does not even come close to offsetting the radiative forcing from us increasing GHGs. Also, even those models with a strong short-term positive cloud feedback do not necessarily have a high EQS for doubling CO2.

  2. 2
    Eli Rabett says:

    Hi Andy,

    This is but another example of how climate scientists are being played by the denialists. You attempted to discuss the issue with Spencer as if he were only doing science. But he is not. He is doing science and politics, and he has no compunction about sandbagging you.

    There is no gain to you in trying to deal with people like Spencer and Lindzen as colleagues. They are not trustworthy.

  3. 3
    MapleLeaf says:

    I thought El Nino events were caused by the so-called “delayed oscillator” effect. And that the westerly wind busts that seem to trigger and event may be related to the MJO in some way. But the latter is just a hypothesis.

    If Spencer is so confident that clouds trigger ENSO:

    “Spencer: ENSO is caused by clouds. You cannot infer the response of clouds to surface temperature in such a situation.”

    Then why has Spencer been unable to demonstrate causality and present a detailed conceptual model, or at least demonstrate unequivocally that changes in clouds precede ENSO events. We have TRMM data since late 1997, surely we can use that? Not to mention the GOES and NOAA satellite data.

    I am not convinced by Spencer’s arguments. Clouds are a feedback arising from ENSO not a driver.

    This whole negative cloud feedback claim is beginning to sound like the much refuted claim that GCRs significantly modulate cloud amounts.

    Maybe Spencer can tell us when this negative feedback is alleged to have started operating, and what the current long-term warming trend would have been without the alleged negative feedback. If we believe Lindzen, the negative feedback should have started very shortly after global temperatures increased by +0.5 C in order for his fast-feedback sensitivity of +0.5 C to be met, that or temperature will have to increase beyond +0.5 C and then decline. In reality, we have now warmed by over +0.8 C (for 40% increases in CO2), and the rate of warming of global temperatures has remained largely unchanged since the eighties. In fact, it is warming faster now than it was between 1915 and 1945.

  4. 4
    MapleLeaf says:

    I agree with Eli @2.

    Spencer’s rant on his blog and his defamatory comments about Andy are way over the line.

  5. 5
    DrCloud says:

    Great work!

    We’ve long thought that the character of feedback from different types of clouds could be quite different. Have you considered stratifying your analysis by cloud type (based on, say, cloud-top temperature) to reduce the scatter of Fig. 2a?

  6. 6
    Sean says:

    This quote seems at odds with what you claim the conspiricists claim. “It doesn’t mean that “clouds cause El Nino”, as Dessler suggests we are claiming,”
    How do you counter the phase space argument (which to a radio engineer is fairly basic in the analysis of feedback interactions)

    [Response: Spencer is claiming that clouds are causing ENSO. It may not be clear in his blog posts, but read our correspondence and you will see that’s what he is claiming. Lead-lag arguments are fine if you believe the temperature change is radiatively forced, but that is not the case for ENSO. Spencer also acknowledges that feedbacks can be observed in non-radiatively forced climate change (see his 2010 paper). That’s why he’s so determined to argue that ENSO is really radiatively forced. –Andrew]

  7. 7

    If you can get me a copy of the time series data, I’ll run Granger causality tests on it.

  8. 8
    robert davies says:

    “But as you can see, I have not ignored it — I have dismissed it because I think it has no merit. That’s quite different.”


    Reminds me of one of my favorite bits of movie dialogue — from Casablanca:
    UGARTI: “You despise me, don’t you Rick?”
    RICK (BOGART): “I suppose I would, if I gave you any thought.”

  9. 9
    BCC says:

    But Eli,

    I don’t thnk Andy is getting played. He played Spencer straight, as he should, as Spencer is at least doing some real science, and then applying the smackdown here (e.g. the “In fact, Dr. Spencer had a press conference in Cancun — about my paper. I didn’t have a press conference about my paper. Draw your own conclusion.” bit).

    And this post does me a great service. While I hadn’t bought into the Spencer cause and effect argument, I hadn’t understood why his claims are probably wrong (at least until I ran across Lin et al.) Dessier’s email exchange is quite enlightening.

  10. 10
    Lou Grinzo says:

    Andrew: Thanks for posting this, and thanks also for the link to your paper. (Those of us without employer-sponsored access to journals appreciate it.)

    As for the issue of “to engage the deniers or not”, I admit that my view on this has changed a lot over the years, and I’m increasingly of the opinion that it doesn’t matter. Ignore them, and at least some will simply make up things or pull stunts to garner attention (like Monckton crashing a meeting in Cancun the other day). Engage with them and they will relentlessly spin whatever is said or done.

    This is not to say that scientists should ignore the deniers’ fabrications about the science itself; there will always be a desperate need for the real experts to debunk that open hydrant of nonsense. But direct interaction with them seems (to me) to be a wash.

  11. 11
    Stefan Rahmstorf says:

    Andrew, same old argument: when the Rahmstorf et al. 2007 Science paper came out we were accused of this being “political” because it coincided with the release of the IPCC report. When the Vermeer and Rahmstorf 2009 PNAS paper came out we were accused of this being “political” because it coincided with the Copenhagen climate summit. All this by people who should indeed know that the authors do not have control over the publication date and have typically submitted this to the journal many months earlier. Besides, anyone who thinks that a new scientific paper on cloud feedback or sea level will be noticed by and influence the negotiators at UNFCCC conferences does not have a very realistic appreciation of how these negotiations work and what the negotiators are actually concerned with during these two weeks.

  12. 12
    Philip says:

    Andrew Dessler:
    You quote Roy Spencer as stating that “ENSO is caused by clouds”. But this statement isn’t in the exchange of emails that you reference here, and Spencer states in his blog post that this “would be too simplistic and misleading of a statement”. Could you clarify please where this quote comes from?

    [Response: Our apologies. We assumed it was clear that this was short a paraphrase of the much longer chain of email exchanges linked to. This has now been made explicit. Thanks! –mike]

  13. 13
    Christopher Hogan says:

    Completely unrelated: I find the funniest thing about Spencer’s website to be the following. I can tell whether or not the current month’s global temperature anomaly is low or high, from the UAH MSU satellite series, prior to the release of the data, based on how quickly Dr. Spencer posts the number on his website. If it’s a cold month, the results go up before the UAH data are released to the general public. If it’s a warm month, they go up some time afterward. My casual observation is that there’s about a week’s difference, on average, in when the number gets posted.

  14. 14
    Eli Rabett says:

    BCC Real Climate is not the mainstream media ;) If Dr. Dessler does nothing else than this it will be as if he did nothing. People such as Lindzen and Spencer have to be shown a cost to their two faced behavior, and the first step is to stop treating them as honest colleagues.

  15. 15
    Philip says:

    Andrew Dessler:
    It struck me that one of key aspects of Spencer and Braswell’s 2010 paper was their use of phase space plots to estimate sensitivity using satellite measurements. I cannot find any particular criticism of this approach in your paper, although you do mention that “inferences of large negative feedbacks have also been criticized on methodological grounds”. However, my understanding is that the papers you cite in this respect are criticisms of Lindzen and Choi’s 2009 paper “On the determination of climate feedbacks from ERBE data”, which I understand have subsequently been answered by the same authors. Could you therefore provide some more detailed criticism of the approach used by Spencer and Braswell in their 2010 paper please?

    [Response: The paper I said had been criticized was Spencer’s 2007 paper. His 2010 paper came out in mid-2010, so it is probably too new for published criticisms to have appeared. I think there’s a lot of interesting stuff in Spencer’s 2010 paper … my main point with respect to it would be that it does not apply to my analysis because I’m using ENSO variations to infer the feedback. We understand ENSO well, and we know it is not caused by clouds. -Andrew]

  16. 16
    Trent1492 says:

    Has anyone heard of the Cornwall Alliance before? I had not till yesterday. It is an organization of Evangelical Christians who claim that Environmentalist are doing the work of the Antichrist. This would not cause me too much concern normally except for two things:

    1. They appear to be well funded and have a wide reach.

    2. Roy Spencer appears in one of their videos.

  17. 17
    Edward Greisch says:

    Thanks very much for the paper.

  18. 18
    The Ville says:

    Stefan Rahmstorf@11

    Has any one done an analysis of paper releases and summits?
    That would be amusing.
    But I think Spencer should be intelligent enough to recognise a coincidence.
    From a political POV, a journalist might look out for papers being published in the middle of a summit and then suggest it wasn’t an accident, despite hundreds of other papers released throughout the year. Has Spencer joined the ranks of the journalists?

  19. 19
    tamino says:

    I agree with Eli (#14)

  20. 20
    MarkB says:

    Stating the obvious…If Spencer’s primary evidence for negative or no positive cloud feedback is his claim that changes in clouds is the primary cause of ENSO, which has no evidence to support it, then I would say positive cloud feedback is more robust.

    As far as Spencer’s blog goes, it’s merely red meat for ideologues. What does one expect from someone Rush Limbaugh refers to as his “official climatologist”?

  21. 21
    DrCloud says:

    I admit to being a tad behind in reading the literature, but, even so, I find this weird:

    “Spencer is claiming that clouds are causing ENSO.”

    Does Dr. Spencer have a mechanism whereby clouds initiate the equatorial Kelvin wave that shuts off the upwelling? Or is it that he claims that clouds somehow control the timing of the cycle overall? Gotta say, this seems extremely far removed from main-stream thinking about ENSO controls.

  22. 22
    Hank Roberts says:

    > shorter

    Here’s the form to identify “shorter” paraphrases — an Internet tradition:

    ‘Shorter’ concept created by Daniel Davies and perfected by Elton Beard. We are aware of all Internet traditions.™

  23. 23
    phlogiston says:

    Spencer: ENSO is caused by clouds. You cannot infer the response of clouds to surface temperature in such a situation.

    Dessler: ENSO is not caused by clouds, but is driven by internal dynamics of the ocean-atmosphere system. Clouds may amplify the warming, and that’s the cloud feedback I’m trying to measure.

    Actually the position you claim as the “Dessler” position is rather similar to the “Spencer” position. It is unfair to summarise Spencer’s position as “ENSO is caused by clouds”.

    From Roy Spencer’s blog:

    “It doesn’t mean that “clouds cause El Nino”, as Dessler suggests we are claiming, which would be too simplistic and misleading of a statement. Clouds are complicated beasts, and climate researchers ignore that complexity at their peril.”

    Spencer pointed out that Dessler’s analysis excluded the causality pathway of changing cloud cover causing temperature change ()but not the reverse). He proposed an alternative phase space model to include both directions of causality. This is not the same as just saying “clouds cause ENSO”.

  24. 24
    BKsea says:

    Forgive me if this is naive, but it seems that it does not matter whether ENSO is caused by clouds (Spencer) or ENSO is not caused by clouds (Dessler). Under either scenario, you would reach the conclusion from this study that more clouds = more warming. The next relevant question is whether global warming yields more clouds. If it does, then there is a positive feedback. For Spencer to be correct, not only would ENSO have to be caused by clouds, but also warming would have to reduce cloud cover. Is there any evidence to support the second of those contentions?

  25. 25
    Dan H. says:

    How does this model compare with the ISCCP results that show low-level clouds decreased during the 1980s and 1990s as temperatures rose?
    This is just one of a many papers which detail decreasing cloud cover corresponding to temperature increases.
    Am I understanding correctly that these models predict that low level cloud cover will increase under conditions of increasing temperature?

  26. 26
    Brian Dodge says:

    If Lindzen and Spencer are correct, delta GCR => delta clouds => delta ENSO; Sun Spot Number varies inversely with GCR, so -delta SSN => delta clouds => delta ENSO.

    This causal chain should result in a correlation between SSN and ENSO.

    I downloaded monthly ENSO data from and SSN from, applied a running 12 month average to each series, and calculated the correlation coefficient.
    My result – R^2 =0.0071, R=-0.084 The unsmoothed R^2 was 0.0033.

  27. 27
    DrCloud says:

    A positive feedback between clouds and surface temperature does NOT imply that (more clouds) -> (more warming). It implies that (cloud changes due to warming) -> (more warming).

  28. 28
    Joe Hunkins says:

    Thanks for a nice summary and readable treatment of a very complex topic. However I’m confused about how to read the error range, which seems very large compared to the measurement. Is here any way to assign a specific value to the likelihood that the feedback is positive or negative?

  29. 29
    Bob Tisdale says:

    Andrew Dessler: Your email correspondence with Roy Spencer included the following discussion of ENSO: “First, people have been studying ENSO for decades and my sense is that the basic theory that it is caused by changes in surface winds driving changes in ocean circulation seems to be quite successful and explains almost all of the details of the observations (e.g. the evolution of thermocline depth).”

    The increase in trade wind strength associated with La Niña events decreases total cloud amount over the tropical Pacific east of the Pacific Warm Pool, which, of course, results in an increase in DSR.

    You continued, “Second, I did a quick back of the envelope calculation and in order to get 1°C of warming in a 50m column of water in one year you need about 7W per square meter of heating. That seems a bit high. And if the warming occurs in just a few months (which it can, then the required radiative heating would be correspondingly larger).”

    Pavlakis et al (2008), in part, write in their abstract, “A clear anti-correlation was found between the downward shortwave radiation anomaly (DSR-A) time-series, in the region 7° S–5° N 160° E-160° W located west of the Niño-3.4 region, and the Niño-3.4 index time-series. In this region where the highest in absolute value DSR anomalies are observed, the mean DSR anomaly values range from −45 Wm−2 during El Niño episodes to +40 Wm−2 during La Niña events.”

    An increase in DSR of 40W per square meter far exceeds your annual requirement of 7w per square meter from your back-of-the-envelope calculation. Link to Pavlakis et al:


    [Response: Sorry, not even close, and no cigar. You are ignoring the fact that to determine the effect of clouds on SST you can’t take just the downward shortwave effect and ignore the cloud greenhouse effect. The cloud greenhouse effect nearly compensates at the top-of-atmosphere, but it works its way into the surface budget through warming the atmosphere. But take heart — you are in good company in being confused about this. It was the basis of Ramanathan’s confusion in his famously wrong “thermostat hypothesis.” –raypierre]

  30. 30
    Alex says:

    \but also warming would have to reduce cloud cover. Is there any evidence to support the second of those contentions?\

    I thought he claimed the opposite?
    Warming can be caused by less clouds and then it gives the impression of positive feedback even if the feedback is negative. But warming does not lead to less clouds. Isn’t that what he claims or did I read it all backwards?

  31. 31
    DeNihilist says:

    Dr. Dessler, thanks for the direction to the e-mails. Really enjoyed reading them and experiencing two scientists bouncing their theoroms off of each other. Really nice to see science evolving before my eyes. Keep it up!

  32. 32
    Andrew Dessler says:

    Re: 24, by Brian Dodge: “If Lindzen and Spencer are correct …” This reminded me of something I keep forgetting. They cannot both be correct. Lindzen’s approach is actually quite similar to mine, so if you accept Spencer’s criticisms then Lindzen is wrong. If one accepts Lindzen, on the other hand, then one must reject Spencer’s argument that cause and effect are wrong. This is one puzzle you don’t hear much about from the skeptics.

  33. 33
    Andrew Dessler says:

    Joe Hunkins: Given that the value of zero falls between the one sigma and two sigma levels, the probability of it being positive is around 80%.

    phlogiston: Spencer is trying to change his story in his blog. If you read his
    paper and my correspondence with him, the key argument he’s making is that clouds are the underlying driving force behind ENSO. I’m arguing that it’s not clouds. Our positions are not the same. Of course clouds may play a role in amplifying the ENSO, which is what I’m measuring here.

    Dan H.: Unfortunately, the data I’m looking at only goes back to 2000, so I
    can’t compare with an earlier period. As far as cloud cover
    decreasing, I have not looked in detail about exactly how clouds are
    changing (e.g., height, fraction, optical depth, particle size). So I
    can’t say exactly what’s happening. But that’s a great question, and
    one that I’m working on now.

    Bob Tisdale: I fully admit my BOE calculation may be wrong.

  34. 34
    DeNihilist says:

    Dr. Dressler @ 32, is this not to be expected in a subject (clouds) that is not yet on the front main burner and many different hypothesis are bandied about?

    As Dr. Spencer said in one of the e-mails – {Anyway, welcome to the cloud feedback debate. I’m actually happy that your paper will help to bring
    more visibility to a problem that *I* believe is still far from solved.

  35. 35
    Fred Moolten says:

    What strikes me about both Andy Dessler’s and Roy Spencer’s papers are the large uncertainties, and the susceptibility of the calculated values to the choice of time intervals (e.g., 3 months optimal for Spencer), altitude (sea surface vs mid troposphere), and the possibility of distortions introduced by measurement biases, including those involving reanalysis data.

    More important perhaps is the difficulty of drawing conclusions about long term climate feedbacks from short term responses to ENSO events, a difficulty acknowledged by both authors. The problems are manifold. El Nino warming, for example, entails a regional phenomenon characterized by unusually strong deep convection, and imposing the heat from a warming ocean on a previously unwarmed atmosphere, with clear implications for regional changes in relative humidity and cloud cover. Warming due to atmospheric perturbations from CO2 or other greenhouse gases (or cooling from aerosols) involves a more globally distributed imposition of heat from a warming atmosphere on a previously unwarmed ocean. Although long term feedbacks might converge to a common value in each case, there is no a priori reason why the short term responses need be the same, nor why they are informative about the long term feedbacks (Andy’s paper in fact makes a point about this and it is also acknowledged by Roy Spencer). It’s also not clear from Andy’s paper to what extent the estimated positive feedbacks, with all their variability, are partitioned between long wave and short wave effects. One might speculate that they might represent long wave OLR reductions from increases in some cloud types, offset perhaps to a greater or smaller extent by increased SW flux from an increased cloud albedo. We don’t know, but if that is the case, it probably differs from long term effects where evidence exists for positive cloud feedback derived in part from reductions in low cloud cover.

    For all these reasons, I wonder whether climate sensitivity assessments based on ENSO variations are a good way to ascertain climate sensitivity to CO2 increases over the course of years or decades.

    [Response: What Andrew gets, strictly speaking, is the feedback of clouds on ENSO. For the reasons you point out, there are reasons to doubt that this feedback can be extrapolated to warming caused by CO2 increase. Still, it’s one more number you can throw into the pot of estimates of feedbacks. If one is going to estimate cloud feedbacks from ENSO, one ought to at least do it right, and I think Andrew has done this a lot more right than Roy has. We can go into the reasons in a future post. Note also that, as Andrew noted, documenting cloud feedbacks of this sort give you another way of checking the cloud parameterizations in models. Given the large error bars, though, it doesn’t help us narrow the estimates of climate sensitivity much, especially at the high end. It is another nail in the coffin of the idea that climate sensitivity is low, since to get low climate sensitivity you need a strong negative cloud feedback, and that is definitively ruled out by the data, to the extent that you can extrapolate from ENSO to CO2. –raypierre]

  36. 36
    dhogaza says:

    As Dr. Spencer said in one of the e-mails – {Anyway, welcome to the cloud feedback debate. I’m actually happy that your paper will help to bring
    more visibility to a problem that *I* believe is still far from solved.
    – ‐Roy}

    How does this jive with his public statement that Dressler’s paper has set climate science back … ?

    Or his press conference at Cancun, denouncing it?

  37. 37
    Andrew Dessler says:

    DeNihilist: Yes, in fast moving debates theories come and go quickly. My point was that skeptics attack me for using a particular methodology, but they give Lindzen a pass for using the same methodology — because he says what they want to hear, but I don’t.

    Fred Moolten: First, the long and shortwave components of the cloud feedback are plotted in Figs. 3b and c. Second, you are correct that the short-term feedbacks may differ from the long-term feedbacks. What I said in the last paragraph or two of my paper is that measuring the short-term feedbacks is the best we can do now. But while we cannot yet measure the long-term feedback, the fact that the models’ short-term feedback agrees with the observations gives me some confidence in the models’ long-term behavior.

  38. 38
    Peter says:

    I think its pretty clear. Andy says that he thinks they disagree on the cause of ENSO. Andy says that he believes that ENSO is caused by changes in winds changing ocean currents. Andy does not believe that radiative forcings can cause ENSO (i.e. changes in clouds would be a radiative forcing “Third, I follow your argument in figure 4 of your paper. However, I don’t think this type of lead-lag argument works if the cause o ENSO is non-radiative”).

    From there, Andy states that he does believe that clouds have feedbacks in the ENSO:
    “I do agree that clouds are playing a role in amplifying ENSO, but that’s the cloud feedback and that’s what I’m trying to measure.”

    What was Spencer’s reply. Not that Andy was misunderstanding his point of view and that he also believes that ocean and wind currents and not a radiative forcing and is also interested the cloud feedback role in the context of those non-radiative changes.

    He doesn’t say that because it is untrue.

    He says they’ll have to disagree about ENSO.
    “While I agree with you on the MJO, I guess we will have to agree to disagree on ENSO for now.”

    If Spencer doesn’t believe that radiative (e.g. clouds) are causing ENSO, it should have been clear from the e-mails that is what Andy thinks he thinks, and Spencer should have corrected Andy’s belief in the e-mails rather than stating they will have to agree to disagree. If Spencer doesn’t believe that radiative forces (e.g. clouds) don’t play a major role in CREATING ENSO, then he should put together a blog about and explain his response to Andy in these e-mails.

  39. 39

    El-Nino is definitely not caused by clouds (masking the evaporation heat source too much is very good for cooling the sea in the long run) , but El-Nino causes a net cloud increase. This may be observable quite uniquely during sunsets at great distances away from the equatorial Pacific. In a few more ENSO cycles
    I will be able to visually prove with a greater deal of confidence this very neat phenomena by repeating the number of strange “dry” clouds observations seen only during twilights. (scroll on my website to see evidence)…

  40. 40
    kai says:

    You do not need cloud to cause ENSO to get the reverse causation Spencer invoque to invalidate the derivation of cloud feedback given in this new paper. Even if ENSO comes first, if the influence of ENSO on cloud cover is from something else that just the value of temperature increase due to ENSO (for example, the particular pattern of temperature driving stronger or weaker than normal winds, or other indirect effect that are caused by ENSO but not by a global GHG-like increase of downward LW, the reverse causation will be there. It does not mean that Spencer is right, but it means that ENSO is not caused by clouds is not the killer blow the author think it is…

    It is not clouds that are tricky beast, it’s finding causation path in complex systems that is a tricky problem…A problem that has not been solved either by Spencer nor in this paper, sorry…

  41. 41
    Bob Tisdale says:

    raypierre replied: “Sorry, not even close, and no cigar. You are ignoring the fact that to determine the effect of clouds on SST you can’t take just the downward shortwave effect and ignore the cloud greenhouse effect. The cloud greenhouse effect nearly compensates at the top-of-atmosphere, but it works its way into the surface budget through warming the atmosphere. But take heart — you are in good company in being confused about this. It was the basis of Ramanathan’s confusion in his famously wrong ‘thermostat hypothesis.'”

    Interesting. In a reply to my same comment over at WUWT, Andrew Dessler wrote, “Bob Tisdale: You’re probably right about that. I put that into the e-mail without spending much time thinking about it.”

    Additionally, if memory serves me well, Pavlakis et al addressed your concerns in the paper linked above and in their earlier companion paper:

    [Response: As I said on WUWT, I have not really thought about this issue. The line you’re considering is something I dashed off in an e-mail. I don’t know everything about the climate, and on this issue I would definitely defer to Ray’s expertise. -Andrew]

  42. 42
    John Byatt says:


    There is a raving lunatic by the name of pete ridley, google him ,active 24 hours a day spreading disinformation , denies any link to the alliance but has quoted them on blogs and is a habitual lier , as some here may confirm..

  43. 43
    Philip says:

    Andrew Dessler:

    Thank you very much for your response to my comment #15, in which you pointed out that the paper you said had been criticized was Spencer et al’s 2007 paper.

    Nonetheless, the papers you cite regarding “inferences of large negative feedbacks have also been criticized on methodological grounds” do appear to criticise Lindzen and Choi’s 2009 paper rather than Spencer et al 2007. Spencer et al 2007 analyses satellite observations relating to tropical intraseasonal oscillations and finds unexpected variations in ice clouds similar to those described by Lindzen’s proposed iris effect. It does not state that ENSO is caused by clouds, as your response implies.

    Even if Spencer and Braswell 2010 is too new for published criticisms to have appeared (as you suggest), it does seem that the approach you use in your paper to calculate cloud feedbacks from observations (shown in your figure 2A) has already been criticised by their paper. They state:

    “The regression slopes in Figure 1 range from near-zero to 2.5 W m-2 K-1, depending upon the averaging period, and whether surface or tropospheric temperatures are used, which illustrates why satellite diagnoses of feedback have remained so uncertain. Since all explained variances are rather low, there is great uncertainty in the value of each slope.”

    I would be interested to hear your response to this criticism. For comparison, it would also be very useful to see the phase space plot corresponding to your figure 2A in which the dots are connected in time sequence, similar to Spencer and Braswell’s figure 3(a).

    [Response: I think you’ve identified the crux of my argument with Spencer. His 2010 paper argues that scientists have cause and effect wrong: that it is the clouds that are driving temperature fluctuations, so that you cannot identify a climate feedback. However, I am arguing that, because I’m using ENSO variations, we know the cause, and it’s not clouds. I do not argue in my paper that Spencer’s 2010 paper is wrong (although I think it is). I argue that it does not apply here.. -Andrew]

  44. 44
    Didactylos says:

    kai, I think it’s fair to describe indirect effects “that are caused by ENSO but not by a global GHG-like increase of downward LW” as simply “ENSO”. After all, the definition of “ENSO” in general use is very broad.

    Your argument for reverse causation doesn’t make sense. Adding extra links in the causality chain doesn’t change the direction of cause and effect.

  45. 45
    Eli Rabett says:

    To isolate a practice of science issue here, being 180 degrees wrong in the stage where you are starting to think about a problem, can often be a very good thing as it focuses thinking on the issues. Eli did this once in a discussion with Bob Grumbine, who knew a lot more about what was being discussed, but had never gone down the squirrelly rabett hole that Eli fell into, and, after correction, it brought up some interesting implications.

  46. 46
    Eli Rabett says:

    Another point, notice how what we have going here wrt Spencer is exactly what is going on with the neighboring post with O’Donnell, et al. News flash, climate scientists are being attacked and not defending themselves effectively

    Gavin and others really have to activate their Rolladexes and start calling reporters. The science issues will not interest them, but the personal interplay will. And, oh yes, the contrast between the wild accusations Spencer is throwing about in public and his attempt to behave in private are key issues for reporters. The work is “two faced” and don’t be shy in using it, as in “we are extremely disturbed by Roy Spencers two faced behavior in attacking Andrew Dessler, not only is Spencer wrong about the science, but his behavior makes it difficult for us to treat him as trustworthy”

    Finally, guys, your Captchas are impossibly hard to read, which is stopping people from commenting. That may be a feature, but who knows?

  47. 47
    Ray Ladbury says:

    What I think is striking here is the difference between Spencer’s scientist and blogist personae. I have found that when one is actually telling the truth, one version of it suffices regardless of the audience. If you have to change what you are saying when addressing scientists versus what you say when addressing the ignorati, then I will not trust what you say to either. I wonder if Spencer realizes that he is damaging his scientific reputation. I wonder if he cares.

  48. 48
    Dennis says:

    Dr. Dessler,
    I am a lay person with a strong interest is both the science of climate change, and efforts to solve the problem. I just wanted to say that this post of yours is one of the best written, most factual, and highly professional effort to explain in understandable terms what an important paper does and does not say, and explain how the denialists (in this case Dr. Spencer) go to scientifically dubious and even unprofessional lengths to attempt to discredit genuine scientific research for political reasons. Keep up the good work and thank you again for this.

  49. 49

    Roy Spencer seems be saying (from his blog post):

    ENSO causes variations in clouds
    They in turn cause variations in temperature (which presumably feeds back on cloudiness?)

    whereas Andres Dessler is saying:

    ENSO causes variations in temperature
    This in turn causes variations in cloudiness (which feed back on temperature)

    It is not clear to me why Spencer’s hypothesis automatically means that the feedback from clouds on temperature is negative though?

  50. 50
    kai says:

    Didactylos, maybe I was not as clear as I could have:
    first, let’s define “cloud surface warming”, or CSW, as less cloud albedo and/or more cloud downward LW.
    now consider those two causation chains:

    ENSO->higher surface temperature->CSW ->even higher surface Temp->…

    This is the behavior assumed by Dessler, which show a feedback loop.

    ENSO->indirect cloud effect->CSW-> higher surface temp
    -> higher surface temp

    Only the first hypothesis allows you to compute cloud feedback using ENSO forcing. The second one does not, as nothing says how higher surface temp will affect CSW. It is not possible to say if you do not know how much indirect could effect there is, i.e. if the change in cloud cover is caused only by ENSO surface T change….or by something else related to ENSO. I hope I was more clear, and that it is not enough to say ENSO change the clouds, clouds do not cause ENSO to be able to derive any usefull measure of cloud feedback in a non-ENSO context. The second description can include a feedback, it would relate the higher surf temp->CSW link….but the observation could be compatible with positive or negative feedback depending of the ENSO->indirect cloud effect->CSW causation strength.

Switch to our mobile site