The Plug

I have always been suspicious of climate models, in part because I spent some time in college trying to model chaotic dynamic systems, and in part because I have a substantial amount of experience with financial modeling.   There are a number of common traps one can fall into when modeling any system, and it appears to me that climate modelers are falling into most of them.

So a while back (before I even created this site) I was suspicious of this chart from the IPCC.  In this chart, the red is the “backcasting” of temperature history using climate models, the black line is the highly smoothed actuals, while the blue is a guess from the models as to what temperatures would have looked like without manmade forcings, particularly CO2.


As I wrote at the time:

I cannot prove this, but I am willing to make a bet based on my long, long history of modeling (computers, not fashion).  My guess is that the blue band, representing climate without man-made effects, was not based on any real science but was instead a plug.  In other words, they took their models and actual temperatures and then said “what would the climate without man have to look like for our models to be correct.”  There are at least four reasons I strongly suspect this to be true:

  1. Every computer modeler in history has tried this trick to make their models of the future seem more credible.  I don’t think the climate guys are immune.
  2. There is no way their models, with our current state of knowledge about the climate, match reality that well.
  3. The first time they ran their models vs. history, they did not match at all.  This current close match is the result of a bunch of tweaking that has little impact on the model’s predictive ability but forces it to match history better.  For example, early runs had the forecast run right up from the 1940 peak to temperatures way above what we see today.
  4. The blue line totally ignores any of our other understandings about the changing climate, including the changing intensity of the sun.  It is conveniently exactly what is necessary to make the pink line match history.  In fact, against all evidence, note the blue band falls over the century.  This is because the models were pushing the temperature up faster than we have seen it rise historically, so the modelers needed a negative plug to make the numbers look nice.

As you can see, the blue band, supposedly sans mankind, shows a steadily declining temperature. This never made much sense to me, given that, almost however you measure it, solar activity over the last half of the decade was stronger than the first half, but they show the natural forcings to be exactly opposite from what we might expect from this chart of solar activity as measured by sunspots (red is smoothed sunspot numbers, green is Hadley CRUT3 temperature).


By the way, there is a bit of a story behind this chart.  It was actually submitted by a commenter to this site of the more alarmist persuasion  (without the PDO bands), to try to debunk the link between temperature and the sun  (silly rabbit – the earth’ s temperature is not driven by the sun, but by parts per million changes in atmospheric gas concentrations!).  While the sun still is not the only factor driving the mercilessly complex climate, clearly solar activity in red was higher in the latter half of the century when temperatures in green were rising.  Which is at least as tight as the relation between CO2 and the same warming.

Anyway, why does any of this matter?  Skeptics have argued for quite some time that climate models assume too high of a sensitivity of temperature to CO2 — in other words, while most of us agree that Co2 increases can affect temperatures somewhat, the models assume temperature to be very sensitive to CO2, in large part because the models assume that the world’s climate is dominated by positive feedback.

One way to demonstrate that these models may be exaggerated is to plot their predictions backwards.  A relationship between Co2 and temperature that exists in the future should hold in the past, adjusting for time delays  (in fact, the relationship should be more sensitive in the past, since sensitivity is a logarithmic diminishing-return curve).  But projecting the modelled sensitivities backwards (with a 15-year lag) result in ridiculously high predicted historic temperature increases that we simply have never seen.  I discuss this in some depth in my 10 minute video here, but the key chart is this one:


You can see the video to get a full explanation, but in short, models that include high net positive climate feedbacks have to produce historical warming numbers that far exceed measured results.  Even if we assign every bit of 20th century warming to man-made causes, this still only implies 1C of warming over the next century.

So the only way to fix this is with what modelers call a plug.  Create some new variable, in this case “the hypothetical temperature changes without manmade CO2,” and plug it in.  By making this number very negative in the past, but flat to positive in the future, one can have a forecast that rises slowly in the past but rapidly in the future.

Now, I can’t prove that this is what was done.  In fact, I am perfectly willing to believe that modelers can spin a plausible story with enough jargon to put off most layman, as to how they created this “non-man” line and why it has been decreasing over the last half of the century.   I have a number of reasons to disbelieve any such posturing:

  1. The last IPCC report spent about a thousand pages on developing the the “with Co2” forecasts.  They spent about half a page discussing the “without Co2” case.  These is about zero scientific discussion of how this forecast is created, or what the key elements are that drive it
  2. The IPCC report freely admits their understanding of cooling factors is “low”
  3. The resulting forecasts is WAY to good.  We will see this again in a moment.  But with such a chaotic system, your first reaction to anyone who shows you a back-cast that nicely overlays history almost exactly should be “bullshit.”  Its not possible, except with tuning and plugs
  4. The sun was almost undeniably stronger in the second half of the 20th century than the first half.  So what is the countervailing factor that overcomes both the sun and CO2?

The IPCC does not really say what is making the blue line go down, it just goes down (because, as we can see now, it has to to make their hypothesis work).  Today, the main answer to the question of what might be offsetting warming  is “aerosols,” particularly sulfur and carbon compounds that are man-made pollutants (true pollutants) from burning fossil fuels.  The hypothesis is that these aerosols reflect sunlight back to space and cool the earth  (by the way, the blue line above in the IPCC report is explicitly only non-anthropogenic effects, so at the time it went down due to natural effects – the manmade aerosol thing is a newer straw to grasp).

But black carbon and aerosols have some properties that create some problems with this argument, once you dig into it.  First, there are situations where they are as likely to warm as to cool.  For example, one reason the Arctic has been melting faster in the summer of late is likely due to black carbon from Chinese coal plants that land on the ice and warm it faster.

The other issue with aerosols is that they disperse quickly.  Co2 mixes fairly evenly worldwide and remains in the atmosphere for years.  Many combustion aerosols only remain in the air for days, and so they tend to be concentrated regionally.   Perhaps 10-20% of the earth’s surface might at any one time have a decent concentration of man-made aerosols.  But for that to drive a, say, half degree cooling effect that offsets CO2 warming, that would mean that cooling in these aerosol-affected areas would have to be 2.5-5.0C in magnitude.  If this were the case, we would see those colored global warming maps with cooling in industrial aerosol-rich areas and warming in the rest of the world, but we just don’t see that.  In fact, the vast, vast majority of man-made aerosols can be found in the northern hemisphere, but it is the northern hemisphere that is warming much faster than the southern hemisphere.  If aerosols were really offsetting half or more of the warming, we should see the opposite, with a toasty south and a cool north.

All of this is a long, long intro to a guest post on WUWT by Bill Illis.  He digs into one of the major climate models, GISS model E, and looks at the back-casts from this model.  What he finds mirrors a lot of what we discussed above:


Blue is the GISS actual temperature measurement.  Red is the model’s hind-cast of temperatures.  You can see that they are remarkably, amazingly, staggeringly close.  There are chaotic systems we have been modelling for hundreds of years (e.g. the economy) where we have never approached the accuracy this relative infant of a science seems to achieve.

That red forecasts in the middle is made up of a GHG component, shown in orange, plus a negative “everything else” component, shown in brown.  Is this starting to seem familiar?  Does the brown line smell suspiciously to anyone else like a “plug?”  Here are some random thoughts inspired by this chart:

  1. As with any surface temperature measurement system, the GISS system is full of errors and biases and gaps.  Some of these their proprietors would acknowledge, and such have been pointed out by outsiders.  Never-the-less, the GISS metric is likely to have an error of at least a couple tenths of a degree.  Which means the climate model here is perfectly fitting itself to data that isn’t even likely correct.  It is fitting closer to the GISS temperature number than the GISS temperature number likely fits to the actual world temperature anomaly, if such a thing could be measured directly.  Since the Hadley Center or the satellite guys at UAH and RSS get different temperature histories for the last 30-100 years, it is interesting that the GISS model exactly matches the GISS measurement but not these others.  Does that make anyone suspicious?  When the GISS makes yet another correction of its historical data, will the model move with it?
  2. As mentioned before, the sum total of time spent over the last 10 years trying to carefully assess the forcings from other natural and man-made effects and how they vary year-to-year is minuscule compared to the time spent looking at CO2.  I don’t think we have enough knowledge to draw the Co2 line on this chart, but we CERTAINLY don’t have knowledge to draw the “all other” line (with monthly resolution, no less!).
  3. Looking back over history, it appears the model is never off by more than 0.4C in any month, and never goes more than about 10 months before re-intersecting the “actual” line.  Does it bother anyone else that this level of precision is several times higher than the model has when run forward?  Almost immediately, the model is more than 0.4C off, and goes years without intercepting reality.

46 thoughts on “The Plug”

  1. You’re so fucking stupid it’s unbelievable. You just repeat the same bullshit time and time again, and obviously have no idea of how ridiculous you look. You point out the models match past climate very well, and then claim that they vastly overpredict past warming! And apparently you don’t even notice your self-contradiction!

    And you are not just fucking stupid but fucking blind as well. Your solar chart shows that temperatures started rising in 1910, twenty years before solar activity rose. It shows that in 1960, solar activity was rising and temperatures were not. It shows that in 1985, solar activity started dropping while temperatures rose sharply. In short, it shows that over the last 150 years, solar activity cannot have been a large factor in driving the climate. But you think it shows the opposite.

    My model of denialist behaviour is so far incredibly successful. It predicts that you will never learn, and never understand, let alone correct, your stunningly basic errors. I expect it will be no more than two months before you next post your meaningless graph, with exactly the same commentary.

  2. It’s the standard flaw with any fiancial model too. If you fit and then predict, your model will be out.

    A typical test is this. Get the models working on one set of data, say 1970-1990. Then pick the model that performs best when tested against unknown data 1990-2000. ie. A blind test. However, by picking the best performer, you’re unconciously fitting the model to the data 1990-2000. ie. It’s a fitting and interpolation that you have picked, not necessarily one with any explanatory information. End result, when run as a forecast, they fail miserably compared to the great prediction for the 1990-2000 period, because you’ve cherry picked the best model, and that’s more likely to be the model where the parameters gave the best fit.

    What you see going on in the weather world is best illustrated by the solar scientists who keep changing their predictions as they fail to predict the rise in solar activity. No doubt at some point they will say, look, isn’t our model (the last one) great. Well, no, if failed the forecast test.

    If they are clear that its the best prediction at the moment, its a different issue. However, you need to report very clearly why the model has been failing.

  3. Hunter – that is argumentum ad hominem at it’s worst. These are arguments about models and prediction, not actuals.

  4. Hunter,

    What is it you want us to infer from your choice of adjectives:

    – intelligence?
    – argumentative skill?
    – analytical brilliance?
    – wit?
    – charm?
    – self-confidence?

    Help us out here.

  5. Fran and Billy, ignore this Hunter, the man doesn’t even understand basic chemistry. He’s not worth your time.

  6. “The first time they ran their models vs. history, they did not match at all” – citation for this, or did you just make it up?

  7. I did not get the point. Do I understand correctly that the GISS does not admit that it calibrates the model to historical data?

  8. Jennifer,
    Do you think you are are accomplishing anything by insulting the host and continuing to trollishly highjack my name?
    Perhaps you name is actually James Hansen? You both seem to operate on about the same level of maturity and integrity.

  9. Good Hunter (the original I assume.), now Jennifer is hijacking my name too because I owned his ass on another post. He’s definitely got a screw (or 3) loose.

  10. Jennifer’s / Scientist’s ass is not worth any more than his/her intellect, I suggest you set the bar a little higher.

  11. Warren,

    Thanks for the link to your video clip contained in your post. As a non-scientist, I found that the format of your presentation, particularly the graphics, provided a very clear exposition of “feedback” in the climate models.
    As someone who interprets financial models when assessing the viability of project funding proposals, I can appreciate how the results are only as good as the original input assumptions (GIGO). If you have not done so already, you should consider producing further videos on other climate change topics.

  12. Whoever posted the first post, please take to heart what Fran Manns posted. There is no need to be vulgar.

    This is a great blog, and I’d hate to see Comments disabled because of a bad egg.

    Mr. Meyer, perhaps a simple registration form would prevent unwarranted impersonation in the posts.

  13. “Looking back over history, it appears the model is never off by more than 0.4C in any month”

    Only if you somehow can’t see what the graph is doing in 1989, 1995, 1998, 2001 and 2002.

  14. So, the models predict the past suspiciously well, and at the same time the models greatly over-predict past warming. Interesting bit of doublethink, that.

  15. I think that the best thing to do, in order to keep things clear, is for everyone to post as “Hunter”.

    I’ve started the ball rolling…

  16. “hunter”, #1 You point out the models match past climate very well, and then claim that they vastly overpredict past warming!

    That’s because they’re kludged to fit the past data you dickhead! And there’s nothing to say that yesterday’s kludge will cause tomorrow’s reality to match the model.

  17. “I think that the best thing to do, in order to keep things clear, is for everyone to post as “Hunter”.”

    Good plan.

    So, if the models are “kludged to fit the past data”, and therefore fit the past data very well, how exactly do they at the same time “vastly overpredict past warming”? You seem to be struggling to even see the contradiction here.

  18. Just to get it clear. Is it true that the climate models, such as Model E, contain free parameters that are calibrated to match historical temperature datas?

  19. You seem to be struggling to even see the contradiction here.

    No… They vastly overpredict warming, full stop. This is demonstrated by the (future) green line on the “Extending GISS Model E…” graph. The past can be kludged to fit the data by inventing a negative “all other factors” factor, this may become a tad tricky when the actual temperatures haven’t yet occurred.

  20. You don’t seem to understand. Looking at the model hindcasts, do they

    a) match the observations uncannily closely, or
    b) vastly overpredict past warming.

    Do you even understand that these two options are mutually incompatible? Make a choice. Which one do you believe?

  21. I think this Hansen’s new tactic: flood news and analysis that shows the problems of AGW with so much trash that the lie of AGW simply slides on through,a s part of the noise.
    Remember: AGW has never been about the science or the evidence. It is always about the power of AGW promoters.

  22. Ah yes, a great response. Ignore the inescapable contradiction you’ve wound yourself into and just talk bollocks. Good one!

    Anyone want to take on the actual question here? Looking at the model hindcasts, do they

    a) match the observations uncannily closely, or
    b) vastly overpredict past warming?

    Not a hard question. Really.

  23. Knowing this is a waste of time with people like hunter I will try and educate him anyway.

    hunter, the high feedbacks necessary to come up with the high future warming predictions vastly over predict past warming. To compensate for this, the modelers use a “plug” to make their models match past observations. This is so obvious it shouldn’t even have to be pointed out.

  24. Hunter,

    Not defending or criticising the accuracy or merits of this article. Just clarifying why “Hunter” is completely confused yet again… (sigh)

    (a) observations uncannily match predictions because natural forcings are assumed to be heading downwards even though solar activity is increasing. The point of the article is that the IPCC does not explain WHY there should be “natural cooling” during this period. What is the cause of this? Especially when solar activity has increased.

    (b) If the “plugs” as per (a) are removed, the models apparently do vastly overpredict warming.

    Hence there is no contradiction.

    Hunter, as this was the whole point of the article and explained in detail there, your (a) and (b) choices are unusually stupid, even for someone like you.

    And as for your earlier statement:

    “solar activity cannot have been a large factor in driving the climate.”

    Well, you are a complete ignoramuses. No one argues – not the IPCC, or any climatologist living on this planet, that solar activity is the PRIMARY climate driver. If there was no solar activity, this world would be a frozen ball of ice. The CO2 hypothesis asserts that CO2 is ‘tipping the temperature balance’ of the planet. But the sun still, and always will, remain the primary driver.

    You seem to be completely confused over what even the basic role of CO2 as a forcing in the climate system is supposed to be according to the CO2 hypothesis.

    (And thanks for using the “Will Nitschke” name on posts I did not writer, you childish little man.)

  25. Hunter – are you blind? The blue band covers a range from -0.4 to -0.2 at the beginning. It covers a range of -0.4 to -0.2 at the end. Where’s the ‘natural cooling’?

    Are you also illiterate? Get an adult to read these two sentences to you:

    1. models that include high net positive climate feedbacks have to produce historical warming numbers that far exceed measured results
    2. Blue is the GISS actual temperature measurement. Red is the model’s hind-cast of temperatures. You can see that they are remarkably, amazingly, staggeringly close.

    As for the sun, you dopy cunt, no it is not the primary driver. Global warming and global cooling can be caused by the sun, and they can be caused by other things. If the sun is not causing them, then it’s NOT THE FUCKING PRIMARY DRIVER, IS IT?

    Do you molest small children? I bet you do.

  26. Of course global warming and cooling can be caused by the sun. All global warming models factor in solar irradiance (sun’s activity). It’s one of the most basic inputs in all climate models. If all other forcings were neutral a drop in irradiance would cause global cooling or an increase would cause global warming.

    Consider this pro global warming article and note the prominence of SOI (solar irradience). Depending on time frame, and the interaction of other forcings, solar irradience at the peak of solar cycle can effect and drive global temperature in significant ways. Actual hypothesised contributions are shown here:

    (The article doesn’t actually make a lot of sense, but it does simply point to the fact that solar irradience is viewed as a significant climate driver by all mainstream) scientists.) It’s basic stuff. And has no relevance to whether “AGW” is “real” or “not”.

  27. Hunter:

    Do you understand what the word ‘PRIMARY’ even means? Meaning: it makes the most contribution. Your room temperature is 25C. You turn the air conditioner on and adjust the room temperature to 22C. What is PRIMARY driver of the temperature in that room? The 2C contribution of the air conditioner or the 23C contribution of the surrounding atmosphere?

    Geez, you are a dumb fuck. 🙂

  28. “Hunter – are you blind? The blue band covers a range from -0.4 to -0.2 at the beginning. It covers a range of -0.4 to -0.2 at the end. Where’s the ‘natural cooling’?”

    The natural cooling is between 1940-1970 – this is what makes the models work (according to this article anyway). This is why there is physical cooling according to the climate record. It has to be explained if CO2 is increasing.

    Look, you can agree or disagree with the argument presented in the article, and that’s fine. But to not even understand the article and then fly off the handle and criticism some imaginary problem with the logic that is not there… because you’re stupid… well you’re just being a total arse.

  29. Reading the series of “hunter” posts reminds me of the original Star Trek episode, “Mirror, Mirror”.
    It was much easier for Federation Kirk, McCoy, Uhura, and Scott to blend into the Terran Empire Unniverse than it was for their Terran Empire counterparts to blend in with the Federation.
    – A. McIntire

  30. Wow, so when the temperature drops in your room, you don’t consider your air conditioning to be the driver of the change? You think the very air in the room is making it cool, and what, the air conditioning going on is just a coincidence? I think most two year olds could probably understand that.

    You all seem to be claiming that the alleged ‘natural cooling’ was put into the models to offset the 4C of warming that ‘should’ be there ‘if’ there are positive feedbacks. Was the ‘natural cooling’ 4C? If it was less than that, then your theory is bullshit, isn’t it?

  31. I urge anyone who wants to actually discuss the topic of this thread to not cooperate with Jennifer(still stealing my name)in confusing this topic by lowering the S/N ratio. Creating confusion is one typical tactic of those losing. There is no need to shelter Jennifer in this.

  32. the irony of a troll repetitively making the same nonsensical arguments about other people “repeating the same bullshit” is simply hilarious.

    but seriously guys, ignore him/her. we’ve all tried reasoning with him/her. we know it doesn’t work. all it does it fill these threads with endless vitriol and allow our name jumping companion to disrupt any actual conversation, which is clearly the goal. it’s a tar baby. the only answer is not to grab it in the first place.

    if we all refuse to buy into this deliberate disruption, it won’t work. maybe our troll will even get bored and go away.

  33. I would agree with that. The tactic of just repeating rubbish and nonsense and sprouting abuse, is something someone resorts to when they have nothing intelligent to say. The irony here is that it makes AGW believers look like crazy fanatics. I know it should only be about the science–but people like this do leave negative impressions for others, especially fence sitters, to read, and these very actions damage the ’cause’ they desperately believe in.

  34. Which ever ‘hunter’ posted this, “The irony here is that it makes AGW believers look like crazy fanatics.” Missed the point entirely. The entire AGW movement is to a greater or lesser extent exactly like our local troll. From Hansen and his death trains and criminalization of dissent, to Lovelock and his wacked out Xenocide-by-Gaia, to Gore and his planetary fever, to Chu and his bizarre prophecy about California becoming uninhabitable, AGW is not really very different at all from the negative ignorant authoritarian cowardly rants we get posted here by Jennifer.

  35. The last four posts contain absolutely nothing of value, just infantile paranoid rants. It’s obvious that no-one is willing or able to discuss science here.

  36. To be honest, I don’t really understand what this obsession about the “uncanny” correlation between the model and the the actual temperature is all about. The correlation isn’t *that* good, at least not compared to something like this:

    The dampening comes not from the natural “cooling” which is so slight as to be insignificant, but from anthropogenic aerosols, which is bound up in the anthropogenic forcings model. The natural model may be some kind of plug to improve the correlation, but it is not to hide the “expected” huge 20th century warming.

  37. OK – next question. Is it reasonable to assume (as the model apparently does) that d(GHG forcings)/dt is +K? There appear to be sporadic changes in actual temp and sporadic changes in non-GHG forcings…. if all other forcings have variability, why not GHG? ca. ’88 is the only period that seems to have “natural” variability in GHG forcings. Other than that it appears to be very consistent? No?

    If the modelers have an assumed +K slope for GHG component and that is the basis for their model, then it validates the plug theory. So there must be a theoretical justification for keeping that GHG component at a constant slope?

  38. Some one calling himself Hunter said:
    “You point out the models match past climate very well, and then claim that they vastly overpredict past warming! And apparently you don’t even notice your self-contradiction!”

    Even as a non-scientist, I followed the point being made in post without any problem. If anyone needs a more simple explanation, the video cited in the post explains the basic argument at a level that my daughter, a high school freshman, understood with no problem.

    If you really think the post contain a self-contradiction, you really don’t have any business commenting on climate science.

  39. JR Frey – modellers don’t ‘assume’ an approximately constant slope for greenhouse gas forcings – the forcings are observed. Look at the CO2 data from Mauna Loa here. After the seasonal cycle is removed, the trend is simple – smooth and upward.

    pauld – for the ‘plug’ allegations to hold any water, according to the argument presented, the ‘plug’ has to be -4°C. The brown line is supposed to represent the ‘plug’. Please point out where the brown line goes to -4°C.

    And if you really think the contents of this post bear any relation to sense, let alone science, then you have no business commenting on anything.

  40. A person calling himself Hunter says:

    “pauld – for the ‘plug’ allegations to hold any water, according to the argument presented, the ‘plug’ has to be -4°C.”

    No. You apparently don’t comprehend the post.

  41. As was alluded to in a previous post above, AGW believers argue that it is not ‘natural cooling’ that causes the inconvenient drop in temperature over certain decades. It is the industrial production of ‘aerosols’ or pollution during these periods. This may be so, but it does pose a difficulty in evaluating the merits of AGW as even less is known about the climatic effects of aerosols than CO2. If one has an ‘aerosols’ variable that one can adjust up or down as convenient in the models, one can always salvage the CO2 hypothesis (at least historically). While it may not be, falling back on ‘aerosols’ to prop up AGW looks rather ad hoc and convenient.

  42. “silly rabbit – the earth’ s temperature is not driven by the sun, but by parts per million changes in atmospheric gas concentrations!”

    Stupid fuckwit – you seem to think these two things are mutually exclusive!

Comments are closed.