Assorted Links

Thanks to Peter Spero and Alex Chernavsky.

16 Replies to “Assorted Links”

    1. Martin, the post to which you link (at Bad Astronomy) is right that several newspapers badly misinterpreted the study. But the poster at Bad Astronomy misses a much more important point: That (a) we don’t know if cosmic rays influence climate and (b) therefore the models on which all AGW claims are based could easily be wrong. Those models assume no cosmic ray influence — but we don’t know that. The models are going out on a limb. They could easily be wrong. They have never been shown to predict correctly. There is no good reason to believe them. I am saying that AGW claims are overstated because they are based on models that could easily be wrong. I am “denying” that the evidence for AGW is as strong as AGW proponents say.

  1. Seth,

    I think that the cosmic ray theory of Svensmark, though not proven is quite credible. The post at “Bad Scientist” leaves out some key results regarding how pions (cosmic ray analogues) did in fact generate nucleating clusters that could seed clouds. Together with earlier work of Svensmark, and studies correlating ice ages with the Earth’s travel through the oscillating spiral arms of the Milky Way galaxy, there are the elements of a consistent explanatory theory. The best overall account of the theory is given in Calder’s book “The Chilling Stars”. There is also an excellent series on YouTube, reviewing all the supporting data:

    By contrast, there are three key observations which render the AGW theory of global warming untenable:

    1. The “correlation” between atmospheric CO2 levels and global temperatures are extremely good, but with one problem: the rise and fall in CO2 actually lags the rise and fall and temperature by about 800 years. This proves that rising CO2 is a consequence, not a cause of temperature, and hence cannot be the major forcing issue. The same is true of methane, another “greenhouse” gas. The lag is better explained by the outgassing of CO2 from the oceans as they warm, and resolubilization as they cool — with the lag explained by the very large solubility capacity of the oceans. AGW advocates have tried to “explain” this, but their explanation has been debunked:

    2. Water vapor is a far more potent “greenhouse gas” than carbon dioxide, and is present at far higher levels, yet water vapor (in the form of clouds) is typically left out of climate models because clouds are too difficult to model. (This is one of the points that Svensmark’s cosmic ray theory may fill in).

    3. Warming on other planets correlates well with that on Earth, and with solar irradiance Certainly, that can’t be explained by an an anthropogenic effect, and it is hard to find any alternate explanation that solar effects:

    So the cosmic ray theory may not yet be proven, but it is at least as credible as the AGW/CO2 theory, which itself has many holes in it. And Svensmark has faced incredible ostracism from the entrenched peer-reviewed publications. Whether you agree or not, it is an amazing demonstration of the lack of freedom in contemporary science, which is largely a result of the system of government funding and peer review.

    Stay tuned. There will me more to come out the the CLOUD study at CERN.

  2. Seth,

    As further points to my above comment, I think these discussions of the Nature article on the CERN results by Nir Shaviv and Nigel Calder are instructive:

    Shaviv came to the cosmic ray theory independently of Svensmark, from his background as a cosmoclimatologist. The fact that completely different, orthogonal types of data support the cosmic ray theory is itself quite striking. It’s all documented here. Take a look at Figure 2 and try to explain it away!

  3. Whether we do or don’t know that cosmic rays influence climate, we do know that there’s been no trend in cosmic rays since the 50s, so they don’t have any explanatory power for observed recent temperature trends.

    Also, it’s quite false to claim that models haven’t predicted correctly. Observed temperature corresponds rather well with predictions by Budyko in the 60s, Charney et al in the 70s, Hansen in the 80s, 1990 IPCC report, etc. Also, models have predicted several phenomena that were only later verified through observations, including tropospheric temperature trend where observations turned out to be wrong due to UAH processing errors.

    There are still many gaps, but compare the performance of mainstream models to alternatives, for example the hypothesis that cosmic rays drive climate and climate sensitivity is otherwise very low, like Lindzen’s preferred 0.5C/2xCO2. Rewind to 1970 or so, and backcast future global temps with that model, and you’ll find that you’re pretty far out on the low tails of the likelihood. You’ll also have a very hard time explaining natural variability and ocean heat accumulation with strong negative feedback on temps.

    The models also assume no influence of pirates, angels and aliens – and we can never be sure that those don’t matter either. But there are endless sources of possible systematic error in every endeavor. The fact that mainstream climate models, from simple energy balances to GCMs, have withstood investigation of a long list of those errors without major changes to their central conclusion speaks in their favor.

    1. Tom, I am unable to find support for your statement that “Observed temperature corresponds rather well with predictions by Budyko in the 60s, Charney et al in the 70s, Hansen in the 80s, 1990 IPCC report, etc.” Could you tell me where I can find the predictions by Budyoko (not Budyko) and Charney et al?

      My phrase “the models have not predicted correctly” is short for “the models have not predicted global temperature in a way that increases belief in them.” It is no great success for a model when it predicts the continuation of a trend — e.g., continuation of a linear increase in temperature. The models may predict all sorts of things correctly without being able to predict global temperature correctly. (Maybe this is what you mean by “there are still many gaps”.) If they can’t predict global temperature, they can’t be used to argue that humans have had or will have a big effect on global temperature. That’s what I am saying here, in short hand: The models can’t be trusted. Without being able to trust the models, reason to believe AGW disappears.

  4. Budyko wrote a lot, and unfortunately I haven’t figured out where to get it all, but one discussion of CO2 and aerosol effects is in Tellus (1977), 29, 193-204, “On present-day climatic changes”. A related and well-known, extremely simple model is in “The effect of solar radiation variations on the climate of the earth”, Tellus 1969. Note that in both articles, the temperature data on which Budyko relies does not show a clear long-term warming trend, so he is not merely extrapolating.

    Trend extrapolation is a lousy null forecast anyway, because it’s impossible for any length of time, and lacks physical justification. A plausible null forecast really needs some kind of mean reversion or stationarity built in to be plausible, especially if your position is “we know nothing.”

    It’s a bit of a dodge to say that “models may predict all sorts of things correctly without being able to predict global temperature correctly” as if they were predicting the price of chocolate. Many of the things that they’re reproducing reasonably well are features of global temps, e.g. seasonal and latitudinal patterns, response to volcanic eruptions.

    A single time series is a weak test of a model. Physics also matters. Getting things right, other than global temps, provides a lot of information that you can’t get out of temps alone, in part because the inputs to the model (forcing) are uncertain.

    By your metric, (not predicted correctly = predicted continued trend), it would be impossible to validate any that one would naively expect to continue, which is pretty silly. (Especially so if competing alternatives, like cosmic rays, fail to show a long term trend.)

    Charney = – not a model, but does put bounds on climate sensitivity (3 +/- 1.5 C). You can plug that into an early energy balance model, like Schneider, S. H. & Thompson, S. L. (1981) J. Geophys. Res. 86 , 3135-3147, for retrospective tests.

    There are quite a few interesting early works collected in

    1. Thanks, Tom. This is from the Charney report:

      Our confidence in our conclusion that a doubling of [atmospheric] CO2 will eventually result in significant temperature increases and other climate changes is based on the fact that the results of the radiative-convective and heat-balance model studies can be understood in purely physical terms and are verified by the more complex GCMs.

      That does not increase my belief in those GCMs. Not even a tiny bit.

  5. From what you’ve said so far, you’re not convinced by correct predictions of time series data, replication of other phenomena or physical grounding of model structure. That would seem to rule out all possible lines of evidence by which one would compare models in a non-experimental field.

    On the other hand, you’re willing to entertain a nontrivial probability that cosmic rays drive observed temperatures, in spite of flimsy evidence for the mechanism and lack of a trend coincident with temperature.

    It seems like you’re applying an inconsistent burden of proof.

    If you really don’t believe models (or just GCMs?), what’s your null model for climate over the century?

    1. Tom, I will be convinced of the value of GCMs when they correctly and persuasively predict global temperatures. Simple as that. Persuasive = the prediction was a priori unlikely.

      It is not enough to make correct predictions. A model only gains belief when it correctly makes predictions that seem unlikely a priori. This is why I came to believe my own model of weight control: It predicted that sugar water could cause weight loss. This was highly counterintuitive — practically everyone thought sugar is fattening. That the prediction was counterintuitive is what made the fact that it turned out to be true persuasive.

      If GCMs did a good job of predicting global temperature, and did so without assuming anything about cosmic rays, no smart person would be looking into cosmic rays as a possible big influence. That smart people are doing so should tell you something.

  6. By your metric, it seems impossible to validate a model of any system that generates ‘a priori likely’ behavior.

    That’s a problem, because you’re defining ‘likely’ ex post. If you look at the data available to Budyko, Manabe & Wetherald, and others predicting warming in the 70s, they were coming off a 3-decade period of level or declining temperature, and expecting an end to the interglacial at some point. So how was a positive trend a priori likely?

    That’s why I was trying to nail down your null model or forecast. I doubt that there’s a plausible choice, trained to pre-1970 data, that makes observed warming likely.

    You still haven’t pointed out a specific flaw in GCM temperature replication that would motivate a smart person to look for alternatives. The AR4 ensemble fits history pretty well. Admittedly that’s a weak test – but then you have to get into validating the physics.

    ‘Smart’ is not the only motive. Regardless of what you think of the models, you’d have to be a bit naive to not notice a cottage industry in exploring even silly alternatives for ideological and financial gain.

  7. @Tom,

    You write: “Whether we do or don’t know that cosmic rays influence climate, we do know that there’s been no trend in cosmic rays since the 50s, so they don’t have any explanatory power for observed recent temperature trends.”

    That’s an old claim by Lockwood and Frolich that was rebutted convincingly by Svensmark in this 2007 paper:

  8. The link’s not working at the moment, but I assume this is the Cosmoclimatology paper. The CLIMAX data (red in Fig 5), also found elsewhere, actually confirms ‘no trend since the 50s.’

    Also, since the amplitude of the cycle in cosmic rays is large relative to the trend (Svensmark Fig 5), where’s 11-year spectral line in temperature?

    If the net solar/cosmic forcing trend was negative over the last decade, as Spencer concludes, and that’s responsible for level temperatures and ocean heat, then wouldn’t a solar/cosmic explanation require a strong positive trend to explain warming from ’70-’00? (Bearing in mind that temperature is the integral of forcing and feedbacks, not merely correlated, of course.)

Comments are closed.