Revenge of the killer cosmic rays from hell

Yes, clouds and cosmic rays are back, via the indefatiguable Svensmark, at arXiv. But excitingly there is an Antarctic twist, in that the clouds connection explains the “Antarctic climate anomaly, ie why Antarctic and rest-of-work are out of sync. Errrm, but are they? Its a common idea during glacial periods, but Svensmark wants it to be true on shorter timescales so that it can’t be ocean forcing. See his fig 1. So (ignoring the fact that it clearly doesn break down in the most recent past, which I don’t see him noting) there is a nice anti-correlation on longer scales. Which I was unaware of. But: if you plot d-o-18 (a temperature proxy instead) between NGRIP (greenland) and Epica Done C (antarctica) you get:


[Update: Axes are time in years (x) and delta-o-18 (y) in per mille – thanks to LL]

Which doesn’t show the same thing at all. Which may be why his paper is in arXiv rather than a journal?

Oh, and fig 4a, which claims to plot Antarctic temperatures over the last 100 years, is a great nonsense: there is not enough data available to do this. You can’t even pretend to do it before the IPY.

(NGRIP data from, EDC from; nb NGRIP offset by 45 per mille)

[Fairly similar stuff at . The dodgy Antarctic figure becomes fig 6]

59 thoughts on “Revenge of the killer cosmic rays from hell”

  1. I just saw this new paper in JClim, which looks to be general bad news for the CRF idea even though the abstract doesn’t mention anything specific about Antarctic clouds. Also there’s this one, which doesn’t mention clouds specifically but shows that the AR4 models seem pretty happy with simulating Antarctic climate just taking ozone into account. There’s this one too that’s specifically on Antarctic clouds, but I’m not sure if it’s relevant. Not being able to afford a subscription to everything is irritating.


  2. Oy, that’s the former editor of New Scientist, the ‘entertainment magazine’, though this makes him sound like another scientist supporting Svensmark.

    “The best measurements of global air temperatures come from American weather satellites, and they show wobbles but no overall change since 1999….

    “The reappraisal starts with Antarctica, where those contradictory temperature trends are directly predicted by Svensmark’s scenario, because the snow there is whiter than the cloud-tops.”

    [#2 just paraphrases the paper, though. #1 is funny, he obviously hasn’t noticed that the sfc record says much the same but is obliged to stick to satellites-are-best anyway -W]


  3. Dear William and others,

    Svensmark’s review of the cosmoclimatology paradigm shift is published in the February 2007 issue of Astronomy & Geophysics, peer-reviewed journal of the Royal Astromical Society.

    That falsifies the last would-be true sentence of this misleading article by William.


    [Appears to be “News & Reviews in…”. I notice you have nothing to say concerning my problems with his data. I’ve understood you in the past to have concerns over the global coverage of measurements; now you just ignore these problems? -W]


  4. Dear William, I don’t ignore any problems. On the contrary, I think that people must do everything to fix problems with their theories, or abandon them. But paying attention to problems is something different than preventing someone from doing research. Best, Lubos


  5. Otherwise, concerning the more exact point – the quality of global coverage of the data is whatever it is. This data is clearly crucial for directing the research, but on the other hand, it is certainly not perfect enough to rely on them in 100% of cases. A theory that should be given more attention is a theory that can simultaneously provide a natural and coherent picture of reality that is consistent with other theories we know as well as agree with as many detailed facts about the data as possible. It would certainly be wrong to promote one particular measurement to be a decider of everything; it would be equally foolish to ignore the whole data of a quantity such as the average temperature. But the precise choice of relevant data and the operations that are made before the data is used to validate a theory is not a God-given issue, and a wise choice in this respect depends on the talents and good luck of the researcher. Best, Lubos

    [Hi Lubos. That sounds like wurble to me – rather unimpressive. So, to be clear: do you think there is enough data from Antarctica to justify constructing a mean Antarctic temperature before 1950, as S does? If so I’m very surprised. It seems to me that you’re giving S and easy ride because you agree with his conclusions -W]


  6. William,

    Cloud cover seems to explain why the NH is warming and Antarctica is cooling. If its not clouds what is your theory? Is the co2 concentration different in the NH vs Antarctica? If not co2 what is it?


    [GCMs don’t predict much warming over Antarctica, so I’m not sure there is much of a puzzle to explain by clouds. Antarctica is surrounded by a heat-absorbing ocean; this is likely a large part of the explanation -W]


  7. Dear William, I apologize – I don’t know what’s the answer to your question. Moreover, I think it is clear that many people claim to be able to judge too many questions that they can’t – and they only decide according to others who also don’t know the reason but who share certain goals.

    At any rate, your blog and RC are almost the last two blogs in the world that don’t discuss the real event of the day. 😉


  8. William,

    The Antarctic peninsula is warming due to ocean currents (or CO2?) but the Antarctic interior is cooling due to ocean currents (or CO2). Ocean warms or cools? Seems a stretch to say the ocean warms the peninsula but cools the interior. Is this what you are saying?

    Seems more likely the peninsula is warming due to the ocean but the interior is cooling due to fewer clouds. Co2 play a minor role in warming both the peninsula and the interior.


  9. William, the file referenced as the source for the anomalies graphed in fig 4 appears to be this GISSTEMP table of zonal means (see also here ).
    It really does have temperature indices for the 90S – 64S zone going back to 1903!

    However I did some exploration with the GHCN data map tool and it appears to me that temp indices for the 90S – 64S zone before about 1954 are based mostly (in many years entirely) on the Antarctic peninsula.

    [Yes. The table is there, presumably to mislead the unwary. The only data for the early years is probably Orcadas. But we all know that the Peninsula doesn’t necessarily represent the rest of the continent… -W]


  10. One other item. It is my impression that the Antarctic circumpolar winds are a very good insulator. However, the Antarctic peninsula projects most of the way through this barrier. Thus, it is not as well insulated as the rest of the continent. This is usually advanced as the explanation for why the peninsula has warmed (over the last 30 years or so) 1-2 C, while the rest of the continent has not. (Use the GHCN map tool I linked to in my previous post to see this.) It implies the peninsula is a significantly different climate than the rest of the continent (at least w.r.t temperature trends induced by global effects) . The point is – there are indeed old (1903 – 1950 ) temp indices for the 90S – 64S region, but they do not seem to measure what the rest of the paper appears to assume.


  11. Anyway,
    any new “theory” has to be proved, before it is “widely” accepted. That is exactly, what happened with AGW theory. At the beginning, it was an interesting hypothesis, now it is almost a fact, if you like it or not ;-). At the moment, Dr. Svensmark data are just interesting, but needs further study and to claim that this is “alternative” or even “new explaining” theory for AGW is at least…. incautious (or unscientific)?
    I agree with Lubos, that many people try to claim things, which they should not. This is common not only to climate science…

    Just an interesting question: “For WHAT reason do we need climate models, to change our lifestyles”?


  12. Alexander,

    What proof do we have of AGW beyond our experience of the last century?

    Why should be expect the next 100yrs to be anything other than an extension of the last 10yrs? or 20yrs? as relates to co2 production and impact? What disasters have we experienced the last 100 yrs as we have burned FF and increased our standard of living. Why should be change course?


  13. Mr Rabett,

    Was your post meant as a response to my request for evidence of CO2 damage? Do you think CO2 is harmful? Please tell me of past damage resulting from increasing CO2 levels?

    Do you wish other readers of this blog to think you don’t know the difference between co2 and air pollution?


  14. Dear Charles,

    GW and resulting climate change are only in its starting phase, but the very first signs can be clearly seen and if (!) current trends will persist (and if co2 emmisions will continue to increase, i see no reason, why they should not), then the avderse effect will be worse than now. And if the temperature will rise by 3C (and even lower values are “dangerous”), we will live in totally different world… guess what happens, if in 50 years there will be NO SEA ICE in Arctic during the summer, hm?


  15. Alex,

    You agree that we have seen no damage from co2 the past 100yrs.

    You project future damage as no Arctic sea ice during the summer. Is this damage? In what way? Seems to be Canada and Siberia would welcome a longer growing season and open ports. Where is the damage?

    I don’t agree with 3dc by the way. 1dc is the most probable scenario. Also, most of the warming has been and is projected to be in the higher latitudes during the winter. Thus -20dc instead of -30dc in the winter for Canada and Siberia. Where is the damage?

    You would constrain economic growth and rising standard of living for more summer ice? Colder winters? Are you serious?

    [Its funny – but disturbing – that those who don’t think GW is going to cause damage also seem to believe in implausibly low temperature change estimates. The two beliefs should be disconnected, but they aren’t (like GW and ozone). Note that this is not intended as a hammer for the skeptics, because it works the other way round too -W]


  16. Dear Charles,

    actually, I think, that increased CO2 already brings a lot of damage, but this is difficult to quantify. Try to separate the strength of e.g. Katrina… we don’t know, what damage would this hurricane cause under normal CO2 concentration, but without doubt hurricane Katrina was either larger, or more intense or both under elevated co2… that is only one example… for more recent example look at the snowstorms in US

    You say warmer Canada. Do You know, that large areas of Western Canadian forest are under threat by bark beatle? Do You know, that the very first creatures, that react to increased temperatures is insect? Do You know, what paleoclimatologist know about climatic changes in the past?
    Lot of plant and animals will not be able to adapt to climate shifts, which are already occuring.
    I also don’t like cold, and I like sun and hot temperatures. But this gives me a little reason to support statment that CO2 is life 😉
    That is the same, as the polar bears would like to have a snowball from Earth 😉


  17. William,

    It’s hard to be too concerned when one doesn’t see any damage from co2 thus far.

    Co2 is increasing linearly, not exponentially.

    Warming from co2 ~log co2 thus co2’s contribution as a driver of damage will be less looking forward. Thus future co2 damage will be less than we have seen thus far.

    If one can’t make the case that we would be better off today had we not burned FF the last 100yrs then it’s rather hard to be motivated to sacrifice a rising standard of living in the next 100yrs.


  18. William,

    Please advise where you disagree with the following:

    We have seen ~0.4dc warming in the last 50 years of increasing co2.

    Of the 0.4dc how much is co2 forcing, other forcing, positive feedback(water vapor) forcing? Let’s assume 0.1co2 forcing, 0.1other forcing, 0.2 positive feedback.

    Now given co2 is increasing linearly (not exponentially)and co2 forcing ~log co2 what can we expect for the next 50yrs?

    It’s really hard to see more than 0.4dc for the next 50 yrs unless some non co2 forcing increases greatly. Similarly for the next 50yrs beyond that.

    If co2 were increasing exponentially then we should see a linear increasing co2 forcing right? Thus the next 50 yrs would be just like the last 50yrs right?

    So, what’s wrong with this logic?

    [Its not clear why you expect me to believe your CO2 scenario rather than, say the SRES ones. The last 50 years includes the “sulphate cooling” plateau; a better timeframe is the last 30 where we have closer to 0.2 oC/decade. Its all very well saying “lets assume 0.1 CO2, 0.1 other” – but you provide no justification for this at all. How about assuming 0.3 CO2, -0.1 other instead? That would be closer to reality.

    If you believe 0.4/50y; ie 0.08/10y, then there is probably a basis for a bet in there -W]


  19. What is wrong is that you arbitrarily pick numbers out of a hat, both for past and future climates which you use to spin a one dimensional model as reality. Easy to get to a conclusion that way.


  20. Damage?

    If CO2 makes it warmer? That is nice!

    But still the artic ice is growing and temperature will drop now.

    Would CO2 help? I’m not sure. Year 1100 Greenland was an agricutural success. A bit warmer… Temperature variations of up to 3-4 degrees in (sun)cycles last 10000 years and more wouldn’t be beaten by a fraction of a degree from CO2.


  21. Mr Rabett,

    The last 50yrs is reality.

    What ever projection one wants to make for the next 50yrs has to be grounded in the reality of the last 50.

    Why don’t you give it a try? Give us the temp change and forcings for the last 50yrs and then project the temp change and forcings for the next 50. That would help me (and other readers) understand where you are coming from.


  22. Why don’t you give it a try? Give us the temp change and forcings for the last 50yrs and then project the temp change and forcings for the next 50. That would help me (and other readers) understand where you are coming from.

    Please see the AR4 SPM , figure SPM-4, on page 18.
    See also Eli Rabbett’s discussion of the accuracy of Hansen’s 1988 forecast, here and here .
    There are two important things to keep in mind:
    (a) The models have always been judged by their ability to recreate the recent past (and to a lesser extent a few paleo scenarios)
    (b) Even as far back as 1988, some models made forecasts that have proved skillful.


  23. Charles,

    I’ll bite:
    Last 30 years, +0.5C
    Next 30 years, +0.6C +- a bit (+-0.2 perhaps?), depending partly on whether we have as many big volcanoes in the future as we have had recently (El Chichon, Pinatubo).
    Beyond 30 years, + whatever we choose (based on our emissions). We can accelerate the warming or decelerate it. How much hotter do you want to get?


  24. Mr Annan,

    A few thoughts:

    a) you picked 30yrs to maximize the increase per decade. If you picked 50yrs you would get ~0.3. Or 100yrs ~0.6. Did co2 just start to rise in 1976? (I could pick the years 1998 to 2006 and come up with a very small rate and you could pick 1994 to 1998 and come up with a very large rate.)

    b) going forward you accelerate the warming even though warming should de-accelerate due to saturation of co2 (warming ~log co2).

    It seems to me you cherry pick the historical data to obtain the maximum rate and then ignore the saturation effect of co2 to project an even higher rate going forward.

    Am I missing something?


  25. Charles, you are probably missing the irony of accusing someone of cherry picking after they have deliberately baited you by debunking your argument based on cherry picking by using cherry picking themselves.

    Hope this makes sense. Judging by your previous posts, I assume it won’t


  26. Please note Annan did not address the questions.

    Anthony agrees that Annan cherry picked the historical data both ignore the saturation effect of co2.


  27. If one takes Annan’s projection.

    0.6dc for the next 30yrs and extends it for 100yrs one gets about 2dc, the very low end of the IPCC projections.

    Thus to get even the very low end of IPCC projections, one has to cherry pick the historical record and ignore the saturation effect of co2.

    If one takes the average warming over the last 100yrs (o.6) and adjusts for the saturations effect of co2 one get’s warming less than 0.6dc over the next 100yrs.

    So where do the IPCC estimates of 3,4,5, and 6 come from? My understanding is they assume WW economic growth far above what we have seen the last 100yrs.

    Please keep in mind that the projected “damage” of rising oceans from a 1-2dc temp rise is on the order of one foot. On the benefit side we get increased co2 fertilization, longer growing seasons in Canada and Siberia, continued rise in WW standard of living due to low cost energy.

    [I’m still a bit baffled as to why you think simply projecting past trends forward is a good idea for the next 100 years. I’m also baffled as to why you think 2 oC is the very low end of the IPCC predictions. You have read the SPM, havent you? if you haven’t. Table SPM-3 if you can’t be bothered to search it.

    As to where all of this comes from: the TAR is online at and you could read that. If you can’t be bothered, then the answer is: the SRES scenarios fed through climate models -W]


  28. William,

    Thank you for your response.

    From table SPM3 we see best estimates of 1.8dc for scenario b1 and 4dc for scenario A1F1. Thus 2dc is at the very low end is it not.

    [Depends on your language. I would interpret “low end” to include the range, which is 1.1-2.9 for b1; 1.4-3.8 for a1t; etc -W]

    “I’m still a bit baffled as to why you think simply projecting past trends forward is a good idea for the next 100 years.”

    It’s always a good idea to use experimental results as a starting point. One then should ask themselves what is going to change from the past (this experiment) to the future. I assume:

    a) co2 will continue to increase ~linearly (not exponentially) because the WW economies are not going to all of a sudden start growing faster

    [Then you are assuming a considerable divergence from the SRES scenarios. You may be right, who knows, but this is then largely a question of economics, not climate science -W]

    b) warming ~log co2

    Thus one should adjust the straight line projection downward (not upward).

    I’m baffled as to why anyone can support UN climate model scenarios that are so at odds with the experimental data and assume unrealistic economic conditions. Now I ask myself why would the UN want to make unjustified projections?

    [The source of your disagreement is the future economic modelling, so I suggest you go off and find someone to discuss that with – perhaps Richard Tol 😉 -W]


  29. I was most struck by the conclusion of the above series of postings: the impact of economic growth on the projected impact of AGW upon society is more significant that the impact of CO2, cosmic rays, etc.

    It sounds to me we are applying experimental climate models (with an admitted uncertainty) to predict the impact on future climate, assuming as a key input, economic growth projections (with an uncertainty 10 or more times greater.)

    Certainly we should all aspire to live the pure lives of hermits out of simple respect for nature and nature’s God, but I’m not about to put on sack cloth and ashes based on economic growth projections coming from economists!

    If economic models were one tenth as accurate as climate models, every economist would be a billionair.

    The science behind AWG is well worth exploring, but when we start projecting the results beyond the realm of science and base those projections upon inputs from the evem more unproven world of economics, red flags definitely go up.


  30. Charles, I’ve copied and pasted it for you once more. Read it again and see if it makes sense to you.

    “Charles, you are probably missing the irony of accusing someone of cherry picking after they have deliberately baited you by debunking your argument based on cherry picking by using cherry picking themselves.”

    This is a good start to your understanding – “Anthony agrees that Annan cherry picked the historical data.”

    Now you must understand that you are also cherry picking data. Is it starting to make sense yet?

    I fear not….


  31. Karl, the reason progress in climate science is not determinative is that it is determined well enough at present for all practical purposes. Thus forecasting is limited by our ability to predict what future emissions will be and that depends on economic decisions to be made in the future.

    It is, however, incorrect to claim that economic growth depends on emissions. Economic growth can be associated with limited, decreasing or steady emission levels rather than just with increasing emission levels, however, decoupling the two depends on policy.


  32. Good point Eli.

    We also need to understand that economic growth is a very limited/crude measure of economic success. Cleaning up pollution, fixing disease etc all contribute to GDP growth. It would be a bizzare logic that suggests it is in our interests to encourage pollution and disease to grow GDP.

    Just because politicians get up an rant about GDP growth and the health of the economy, doesn’t mean you have to nod your head and concur.


  33. “Next 30 years, +0.6C +- a bit (+-0.2 perhaps?)”

    How can you say and belive this when there is no proof of the CO2 global warming theory in historical data; in empirical facts. Why did teperature rice more more than 1000 years ago? CO2? why did it drop after 1940? Did you have any good 0.2 degrees predictiuoins also then? Don´t think so! You have no more science to say temperature will rice now than you would have then, and you have no clue then. Thus, this +0,6, or even a rice in temperature, is pure religion.

    In fact, historically changes in temperature has occured constantly since hundred of million years ago, sometimes up and sometimes down. The climate change we have now is not at all exceptional, as the lies from IPCC until a year ago then told us it is.

    Water vapor is the totally dominanting greenhouse factor and sun spots has been shown to control climate changes. CERN will review Svensmark and IPCC is dead. Thank You. Have a good day!


  34. “co2 will continue to increase ~linearly”

    Well, no, it hasn’t. The growth rate has risen over the years.

    “warming ~log co2
    Thus one should adjust the straight line projection downward”

    We could model the CO2 forcing as approximately log(280+100e^(0.02t)), where the 280 was the natural level of CO2 and the other part is the anthro CO2. This function is a long way from being approximable by log(100e^(0.02t)) because the natural CO2 is still a long way above the anthro CO2. Perhaps when the total CO2 gets considerably above 560ppm it might be worth using this approximation but until then you’re only going to mislead yourself.


  35. chris,

    Could you plot your function a*e^(0.02t) on top of the data you linked to so I can see how closely it matches history?

    Please start with t=0 in 1958, a=35



  36. chris,

    the full equation: 280+35*e^(0.02*t) t=0 in 1958 should align the data and the model you referenced at the same starting point.


  37. Dear William, you wrote:

    “GCMs don’t predict much warming over Antarctica, so I’m not sure there is much of a puzzle to explain by clouds. Antarctica is surrounded by a heat-absorbing ocean; this is likely a large part of the explanation -W”

    This statement is simply untrue and I wonder whether you – someone who directly works in this field – may be writing such untrue statements because of pure ignorance or whether it’s something else – something that others could call scientific misconduct.

    GCMs predict both warming as well as increased precipitation for the bulk of Antarctica and neither of them is observed. It is a huge problem for these GCMs. See

    [I think you’re confusing future warming with past century, or past 50y. But I’m going to have a look -W]


  38. “Could you plot your function a*e^(0.02t) on top of the data you linked”

    Before trying to develop an accurate model of past CO2 levels (which is not a trivial issue), I want to be clear that we all recognize how slowly CO2 growth needs to accelerate to maintain a uniform growth rate in radiation forcing. For example, if or when the CO2 level reaches 560 ppm,it would need to increase by 2.95 ppm to produce the same increase in radiation forcing as an increase of 2 ppm does when the level is 380 ppm, i.e. a growth rate increase of 47.5%. Looking at the Mauna Loa record, from 1959 to 1968, the least-squared-error linear growth rate was 0.75 ppm/annum. From 1995 to 2004 it was 1.8 ppm/annum. That’s 2.4 times as much in 36 years. To maintain a constant proportional increase of 382/380 in the future, it would only need to increase to 560 ppm by 2079 (from 380 ppm in 2005). So the increase in growth rate by 47.5% only needs to happen in 74 years. Judging by past increases in growth rates (140% in 36 years), this shouldn’t be difficult at all.


  39. Also, you can just plot the forcing value, log(CO2), and see how it compares with a regression. The forcing has risen faster than linear.


  40. “Would you plot your model so we can see how close it fits the data?”

    I’m not claiming that my hypothetical example is an accurate fit to the data. I am claiming that the data has risen faster than exponential, which is extremely obvious if you plot log(CO2) and fit a straight line to it.


  41. chris

    “…. I am claiming that the data has risen faster than exponential, which is extremely obvious if you plot log(CO2) and fit a straight line to it.”

    Faster than exponential? May I (we) see your plot log(co2 data) compared to a straight line?

    To my eye the data looks much closer to a straight line then the exponential function of your hypothetical model. That’s why I am so interested to see your plot.


  42. Chris,

    I’m going to plot the data against your model and also a forcing curve(log(co2). I’ll then give you a link to the result even if it is only a pic of the graph.

    One thing you might want to consider when thinking about future co2 growth rates is the observation that co2 growth rates are closely tied to economic growth rates. That is why William points out above that my forecast vs SPM forecast boils down to a difference in economic growth rate assumptions.

    Thus you might want to consider what your assumption of an exponential growth rate in co2 implies about WW economic growth rates in the context of WW GDP historical growth.


  43. “One thing you might want to consider when thinking about future co2 growth rates is the observation that co2 growth rates are closely tied to economic growth rates.”

    Hopefully one thing you’ve learnt is that constant + exponential grows faster than exponential.


  44. William:

    “I think you’re confusing future warming with past century, or past 50y.”

    I don’t confuse anything but even if I did, it doesn’t matter: the climate models predict Antarctic warming and inreasing precipitation in all three time intervals you mention, and in most of them, the prediction is (or will be) falsified.

    [Having had a look, I’m a bit more sympathetic to your view… but to clear up, there is only one time period availab le to “falsify” and that is the last 50y. Within that time the models (averaged) predict warming (over the year) but not all that much of it. Far less than the Arctic, for example -W]

    [And see-also, figure 8 -W]


  45. “constant + exponential grows faster than exponential”

    More precisely, constant + exponential grows faster than the best exponential fit to this constant + exponential.


  46. 50 or 100, or even 150 years of instrumental measurements no matter how detailed, are simply not enough for any kind of statistically significant conclusion. There is a thing called serial correlation and there is no escape from it. Basically, as it happens quite often we are all fooled by randomness. It is quite funny to observe how everyone is building incredibly elaborate theories and supports them with temperature data that by all meaningful statistical standards is just climate noise. As far as temperature reconstructions based on longer periods, they all seem to be biased towards the more recent instrumental data, which as I mentioned is just noise (50-150 serially correlated points are simply no enough).

    [Yes, you need to do statistics to make sense of these things. But handwaving “oh the data is too short” is *not doing stats. Its just saying “i don’t like the data and so will try to avoid it”. With serial corr included, the data show a stat sig upward trend. But thats the easy bit. The hard bit is attribution. As to the “incredibly elaborate theories” – this is hard to understand. The theory is, more GHG tends to warming – why is that elaborate? -W]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s