Swindled: Carl Wunsch responds

Channel 4 now says they were making a film in a series of “polemics”. There is nothing in the communication we had (much of it on the telephone or with the film crew on the day they were in Boston) that suggested they were making a film that was one-sided, anti-educational, and misleading. I took them at face value—clearly a great error. I knew I had no control over the actual content, but it never occurred to me that I was dealing with people who already had a reputation for distortion and exaggeration.

The letter I sent them as soon as I heard about the actual program is below. [available here]

As a society, we need to take out insurance against catastrophe in the same way we take out homeowner’s protection against fire. I buy fire insurance, but I also take the precaution of having the wiring in the house checked, keeping the heating system up to date, etc., all the while hoping that I won’t need the insurance. Will any of these precautions work? Unexpected things still happen (lightning strike? plumber’s torch igniting the woodwork?). How large a fire insurance premium is it worth paying? How much is it worth paying for rewiring the house? $10,000 but perhaps not $100,000? There are no simple answers even at this mundane level.

How much is it worth to society to restrain CO2 emissions — will that guarantee protection against global warming? Is it sensible to subsidize insurance for people who wish to build in regions strongly susceptible to coastal flooding? These and others are truly complicated questions where often the science is not mature enough give definitive answers, much as we would like to be able to provide them. Scientifically, we can recognize the reality of the threat, and much of what society needs to insure against. Statements of concern do not need to imply that we have all the answers. Channel 4 had an opportunity to elucidate some of this. The outcome is sad.

Page 2 of 2 | Previous page

162 comments on this post.
  1. Steven Douglas:

    Re 104, wherein Ron Taylor wrote:

    “Julian, many of the scientists developing the climate computer models work for NASA. Please understand the NASA track record of excellence in computer modeling. Everything they have accomplished in aircraft design and in the space program has depended on computer modeling, including landing men on the moon nearly forty years ago. Sure, climate is even more complex, but the computers used are orders of magnitude more powerful. It is a great mistake to dismiss the power and utility of computer modeling.”

    Speaking of dismissed, is that why the name of Tennekes (Hendrik Tennekes) appears only once in a search of this site, and that only in passing, in response to Lindzen? A search doesn’t even produce a lone link — it just goes straight to that article. If you have anything to do with climate modelling, you undoubtedly know the name Tennekes, a strong proponent of modelling who once stated, “…no forecast is complete without a forecast of the forecast skill…”

    Invoking the name of NASA might be flattering to some at NASA, but it is an Appeal to Authority, and not valid as an argument. Besides that, your comparison of NASA’s lunar mission (et al) modelling with climate modelling is like comparing recent medical advances and breakthroughs to some kind of godhood.

    For as much as I am solidly behind anything that advances our state of the art in modelling, a little humble reality check might be in order here, because climate modelling is still very much in its infancy, and to compare NASA’s technology modelling with the modelling of an entire climate system, regardless how many “orders of magnitude” improvements have been made to date, or who is involved, is vastly overreaching of reality (also by orders of magnitude).

    Climate modelling is an attempt to artificially, virtually, replicate an entire climate system, with all its vast complexity. Just feeding in data, and validitating the model (by knowing where it’s supposed to rain or snow, or how the wind should blow and fluids should mix at a given points), may give you the simplest of running head starts; however, once you set the model loose, it’s eventually going to bear no resemblance to how the earth and its climate will act and react. The question is, how soon will the model take on a mind of its own, so to speak? Ten years? More? How many more?

    The advances made in modelling science are invaluable to the future, but I would liken the current forecasting capabilities to a big game of Sim Earth at this point. You can get a model to appear like the earth’s climate, just as you can animate a polar bear to look somewhat like a live polar bear; you can even get the model to behave somewhat like the earth’s present climate, once you’ve fed in fifty years of varied [and often quite limited] data; but once the model takes off into the future, how valid is that data, and for how many theoretical years?

    This is why, of course, we are presented with multiple forecasting scenarios, with different sets of underlying and governing assumptions, but this leads to another question: For every model scenario and validation paper in support of that scenario that is selected for publication and review, how many were not selected, and why. We will never know. That’s another problem with the artificial modelling universe. Unlike natural proxies and other data that can be examined on the whole by everyone (and later validated or discounted), only the scenarios generated that are actually presented that can be examined, verified and discounted.

    Without aid of a single model, I forecast that the future models will forecast something along the lines of what was assumed already, before the data was ever fed in, and that we will be handed another set of multiple choice possibilities from which to choose for ourselves.

  2. Barton Paul Levenson:

    [[because climate modelling is still very much in its infancy, and to compare NASA’s technology modelling with the modelling of an entire climate system, regardless how many “orders of magnitude” improvements have been made to date, or who is involved, is vastly overreaching of reality (also by orders of magnitude). ]]

    Let’s see, the first modern radiative-convective model was published by Manabe and Strickler in 1964 — if you don’t count Hulbert’s model of 1931, which, of course, could not use a computer. On the other hand, the Apollo program dates from… the mid-’60s. They sound about the same age to me.

  3. Steven Douglas:

    Excerpted from one of myriad papers on this very subject: (hundreds more upon request, or – you can just Google “climate modelling” + “its infancy” yourself)

    Numerical modelling of the fully coupled climate system is still in its infancy. This is primarily due to the immense complexity of the task and insufficient empirical knowledge of many aspects of climate related processes. This is not only the case for the oceans where the knowledge of the time-dependent deep ocean circulation is particularly poor, but also for example in the hydrological cycle where precipitation, evaporation and the three-dimensional distribution of water vapour and clouds are insufficiently known. We probably only know the global annual averaged precipitation within an accuracy of some 5-10%. Practically no reliable observations of precipitation exist over the oceans.”

    There is a universe of difference between a single radiative coupling model and a model of the fully coupled climate system, which is still in its infancy. Infancy is obviously a term relative to comprehensiveness of the state of the art and the certainty of its results, more than its literal age in years.

    With Apollo programs, the number of variables and uncertainties were vastly narrower and far more easily quantifiable than those considered in climate forecasting models that attempt to effectively model/reproduce the dynamics of an entire planetary climate system, and to such a precision that it will continue to behave as that system, and with reasonable reliability and certainty, as it leaves its known historical data and proceeds into the future.

    A fly comes out of its infancy within a day, while a human’s infancy is measured in months, or years. Likewise, the two programs (lunar exploration and climate modelling) are entirely different animals. I seriously doubt that any climate modeller worth his or her salt would even begin to argue otherwise.

  4. Mike Omalley:

    I recently saw this film and was intrigued by the amount of data they used. Now if I understand your post here, you feel they misrepresented you and your opinion. I’m sorry if they did, but can you answer a question for me.
    The data they used on the correlation between Carbon Dioxide and the rising and falling temperatures, was that fairly accurate? Since it is next to impossible to see the actual raw data, I, as do many others, distrust people who use these types of “Facts” to make a point. Especially when it used with computer modeling.

    Thanks for your time

    [Response:The temperature data they used was wonky – it appears to have been truncated. http://en.wikipedia.org/wiki/Image:Climate_Change_Attribution.png is a better picture. T and CO2 data is readily available – William]

  5. Barton Paul Levenson:

    [[Likewise, the two programs (lunar exploration and climate modelling) are entirely different animals. I seriously doubt that any climate modeller worth his or her salt would even begin to argue otherwise. ]]

    If you have an emotional need to believe that climate models are inaccurate, there’s not much I can do about it. But if you want to convince others, you’ll have to come up with some concrete evidence, or at least a logically coherent argument, other than “the field is in its infancy.” Manned lunar exploration is in its infancy, too, and probably will be for a long, long time. We’re still finding out things about the moon the Apollo astronauts never knew; e.g., the lunar rocks Apollo missions brought back were “bone dry,” but Clementine says there’s buried ice at the lunar poles.

  6. James:

    Re #153: So your argument is that because models can’t predict future climate changes exactly, we should just ignore them entirely?

    I don’t think that attitude makes much sense. After all, when I’m out target shooting, my ballistic model is pretty inaccurate: I don’t hit the bullseye every time (or even the target :-(), but you’re still a lot better off standing behind me.

    The analogy may seem a bit strained, but it serves a point. In both cases, there is a very simple basic process, either increasing CO2 or speeding bullets, and even though we can’t exactly predict the effects, it seems really obvious that it’s a not a good idea to get in the way.

  7. Steven Douglas:

    Barton Paul Levenson wrote:

    “If you have an emotional need to believe that climate models are inaccurate, there’s not much I can do about it.”

    Now where did that little piece of ad hominem come from? For me, belief is not at issue, and ‘accurate’ and ‘inaccurate’ are woefully inadequate descriptors in all of this. Capacity, computational power, resolution, confidence, trustfulness, skillfulness, capabilities, limitations, usefulness, etc., all these and more are far more useful (and somewhat less polarizing) than terms like ‘believe’ and ‘accurate’ or ‘inaccurate’. Models can be extremely ‘informative’, or ‘useful’ in some areas, and utterly lacking in others. There is nothing inherently wrong with that, but don’t take my word for anything. Scientists assess levels of ‘confidence’ all the time, to wit:

    I am distrustful of prediction scenarios for details of the ocean circulation that rely on extremely complicated coupled models that run out for decades to thousands of years. The science is not sufficiently mature to say which of the many complex elements of such forecasts are skillful. Nonetheless…I firmly believe there is a great deal to be learned from models. With effort, all of this is explicable in terms the public can understand.

    Not my words. Carl Wunsch’s words (see article above). And since models (and especially models in relation to Carl Wunsch’s article, and not anyone’s emotional needs) are the topic, here’s another thing that he wrote, again, on this topic (emphasis mine):

    Many scientists therefore rely upon numerical models of the climate system to calculate (1) the nature of natural variability with no human interference, and compare it to (2) the variability seen when human effects are included. This approach is a very sensible one, but the ability to test (calibrate) the models, which can be extraordinarily complex, for realism in both categories (1) and (2) is limited by the same observational data base already describe. At bottom, it is very difficult to determine the realism by which the models deal with either (1) or (2). Thus at bottom, it is very difficult to separate human induced change from natural change, certainly not with the confidence we all seek. In these circumstances, it is essential to remember that the inability to prove human-induced change is not the same thing as a demonstration of its absence. It is probably true that most scientists would assign a very high probability that human-induced change is already strongly present in the climate system, while at the same time agreeing that clear-cut proof is not now available and may not be available for a long-time to come, if ever. Public policy has to be made on the basis of probabilities, not firm proof.

    So when it comes to separating anthropogenic contributions to climate change to natural variability, Professor Wunsch believes that clear-cut proof is not now available, and perhaps may never be. It is not the models, according to Professor Wunsch, whereby anthropogenic contributions are even attributed, but assignments of probability by “most scientists”. If that is true, I can accept that — if we call it that, and don’t imply something else.

    And one last:

    First, existing climate models, which are the main tool that have been used to study the hypothesis, do not have the resolution, either vertical or horizontal, to properly compute the behavior of fresh water and its interaction with the underlying ocean and overlying atmosphere. Models of the modern ocean contain special, high resolution subcomponents designed to calculate mixed layer behavior (e.g., Price et al., 1986; Large et al., 1994). Despite the great effort that has gone into them, systematic errors in calculating mixed layer properties remain. How these errors would accumulate in climate-scale models, with much less resolution is unknown. Second, some models also use a physically inappropriate surface boundary condition for salinity, leading to serious questions about the physical reality of the resulting flows (Huang, 1993). Third, the models have almost always been run with fixed diffusion coefficients. A series of papers (Munk and Wunsch, 1998; Huang, 1993; Nilsson et al., 2003; Wunsch and Ferrari, 2004) have noted that, (a) mixing coefficients have a profound influence on the circulation; (b) fixed mixing coefficients as the climate system shifts and/or as fresh water is added are very unlikely to be correct; (c) depending upon exactly how the mixing coefficients are modified, fresh water additions can actually increase the North Atlantic mass circulation (Nilsson et al., 2003).

    I didn’t write any of that, and I didn’t consider it disparaging of climate modelling in the least. They were honest assessments on the part of Professor Wunsch regarding capabilities, limitations, underlying assumptions, etc., all of which affect the confidence he has, or lacks, in certain parts of current models. I didn’t read anything else into it. I don’t have ‘an emotional need’ to believe (or not to believe) either way, and I’m not Channel 4, taking Professor Wunsch out of context (or putting him into one whereby he becomes politically involved in an unwelcome way, and thereby guilty somehow by association). Professor Wunsch’s words shouldn’t shake anyone’s faith, any more than they should add confidence where it does not belong at this point, because a large part of the scientific process is in honestly assessing the errors, limitations and challenges. Unfortunately, with the political climate that clouds this issue (ugh, I hate puns, sorry, none intended), merely making statements of simple fact can lead to controversy. It is a sad day for us all when a scientist like Professor Wunsch feels that he needs Miranda rights read to him before granting an interview, as anything he says can and will be used against him in courts of public opinion.

    As of now, there’s a simplistic (and highly polarizing) push, born of politics and shameless exploitation of lay ignorance, to view climate models in black and white terms. Either they’re exalted as ‘accurate’ (whatever that means), or ridiculed as the soothsaying crystal balls of climate change prophets (which I was guilty of myself, by comparing them to Sim Earth, which was overly simplistic and unfair). Hence, I fully agree with Professor Wunsch’s statement:

    “I firmly believe there is a great deal to be learned from models. With effort, all of this is explicable in terms the public can understand.”

    I think so as well, but it will take more than a little effort. It would take a concerted effort. It would take scientists like Hansen, Lindzen, et al, standing side by side, in a neutral, non-advocacy based documentary, article, paper or whatever, of their own, telling the world precisely what that they do agree upon — which is considerable, and vast, and could restore a great deal of public confidence in a process that [I believe] is perceived as corrupted, and polluted by special interests all the way around. Right now, to everyone’s shame, politicians and the media are the mouthpieces for the scientists, whose scientific views are almost always distorted, or misrepresented. I’m not talking about policy decisions. They’re actually presenting (and misrepresenting) the science itself. There are no peacemakers in this highly politicized process, and we now have lay people picking their favorite scientists as if they were horses in a race, which scientists are then being imputed as agreeing with every lay piece of unscientific nonsense that was ever spewed by either on the advocacy side of things – which should never be confused with the science side of things. And what was the result? HALF of the North American population believes one way, and a little less than half believes another. That is not the slightest bit surprising to me. Looks suspiciously like a coin flip to me, and we all deserve better.

    One thing I take comfort in: nobody’s getting away with anything, on either side. Not ultimately. The science will continue to sort and filter out all the nonsense, which will be discarded in time as so much dross.

  8. Traffic Bulldog.org:

    You know, I have to say that as I am involved with AB 32 and California’s effort at Global Warming mitigation, I am not so sure that the efforts planned currently will amount to much.

    Where is the conversation about rideshare and carpooling in our efforts to combat global warming?

    And in Los Angeles, we don’t need Global Warming as a reason to clean the air … our brown sky should be motiviation alone.

    Click here to Follow Traffic Bulldog Public Comments on AB 32

    http://trafficbulldog.org is a commuter advocacy group committed to helping people form carpools, reduce the number of vehicles on the road, cleaning the air, fighting global warming, underfund terrorism, and getting us half way to Kyoto all at the same time.

    Please join the conversation.

  9. Alex Tarussov:

    I have two comments, one on past climate changes and the other on models.

    – First, the major temperature variations in the past 600K years were not just changes in the atmosphere – they were driven by ice sheet growth during ice ages. Scientists have some idea of how these ice ages started, but their meltdown remains almost a mystery. The highly hypothetical mechanisms used to explain the meltdown, are really just guesswork. I’m saying this to point out that neither temperature nor CO2 variations, whichever started first, were the cause – they followed the development of ice sheets.

    – Second, the climate models. With all due respect to their developers, it is plain common sense to think that any model used to forecast climate, must be validated by its ability to correctly describe past and present climate changes for which records of temperature, CO2 and other factors exist. As far as I know, no climate model has been capable of this so far. Correct me if I’m wrong.


  10. Barton Paul Levenson:

    [[Barton Paul Levenson wrote:
    “If you have an emotional need to believe that climate models are inaccurate, there’s not much I can do about it.”
    Now where did that little piece of ad hominem come from?]]

    Because I don’t think there’s anything I or anybody else here could say which would change your mind. You’ve got your one point to make and you’re going to make it, and hold onto it, no matter what contrary evidence comes along.

  11. David B. Benson:

    Re #159: Alex Tarussov — See the paper by Abe-Ouchi/Sagawa/Saito linked on the What triggers ice ages thread, down a few from here…

  12. Alvia Gaskill:

    Never Let the Facts Get in the Way of a Bad Story

    From the producers of “Swindled” comes “Libeled!,” the skeptics response to charges they misrepresented someone else’s views.