German Scientists Call Matthew England’s Attempt To Declare Climate Models “Robust” A Real Joke

By Dr. Sebastian Lüning and Prof Fritz Vahrenholt

It’s rather peculiar: None of the IPCC climate models projected the warming pause of the last 17 years. An real embarrassment.

Scientist Stefan Rahmstorf was devastated, and so publicly disputed that there even was a pause at all. Naturally this strategy cannot be successful over the long run. Then on April 23, 2015, a team of scientists lead by Matthew England published an article in Nature Climate Change, whose title finally and officially conceded the warming pause:

Robust warming projections despite the recent hiatus.”

The abstract reads:

The hiatus in warming has led to questions about the reliability of long-term projections, yet here we show they are statistically unchanged when considering only ensemble members that capture the recent hiatus. This demonstrates the robust nature of twenty-first century warming projections.

A wonderful strategy: 95% of all models are wrong (see the following figure from Roy Spencer), so simply take the remaining 5% of the models and, with them a single stroke of the pen, declare them “robust” and reliable in their ability to forecast. A real joke.

30 responses to “German Scientists Call Matthew England’s Attempt To Declare Climate Models “Robust” A Real Joke”

  1. Ron C.

    The average model over the same period as UAH shows a rate of +2.15C/cent. Moreover, for the 30 years from 2006 to 2035, the warming rate is projected at 2.28C. These estimates are in contrast to the 145 years of history in the models, where the trend shows as 0.41C per century.

    Clearly, the CMIP5 models are programmed for the future to warm more than 5 times the rate as the past.
    https://rclutz.wordpress.com/2015/03/24/temperatures-according-to-climate-models/

  2. Bart

    It’s known in the US as the “Texas Sharpshooter’s Fallacy”. A cowboy shoots at the side of a barn, then walks over and draws a target with the bull’s-eye containing the largest cluster, and declares himself an expert marksman.

  3. DirkH

    After two Tornados in Germany in the last weeks, Warmunist “scientist” Mojib Latif uses the opportunity to suck up to the next microphone, claiming, Climate Change is now hitting back, is SURE the weather worldwide WILL become more extreme in the future.
    http://www.focus.de/wissen/natur/meteorologie/hagel-stuerme-unwetter-zwei-tornados-in-zwei-wochen-schlaegt-jetzt-der-klimawandel-zu_id_4662727.html

    Mojib Latif might be the most consequent Medicine Show presenter in Germany. Whoever decided to pay him a professors wage should be held liable for the cost.

  4. handjive

    Back in 2012, Prof Matthew England appeared on the Australian govt. owned ABC tv show called “Q&A”, denying the pause:

    “On Q&A Nick Minchin said the IPCC predictions were wrong.
    Matthew England said “Not true” their 1990 prediction was “very accurate”.
    But the IPCC predicted 0.3C per decade, and we got at most 0.18C per decade. (Forster and Rahmsdorf 2011 )
    How is is “very accurate” when the result is below their lowest estimate?”
    Jonova has a link to video, and posted about it:

    http://joannenova.com.au/2012/04/abc-biased-scientist-matthew-england-outrageous-error-or-dishonest-nick-minchin-owed-an-apology/

  5. Mindert Eiting

    Tomorrow it will rain (model 1) or it will not rain (model 2). If it rains tomorrow model 1 is robust otherwise model 2 is.

  6. AndyG55

    If you ,make enough “predictions™” over a wide enough range, you should be able to get somewhere near with at least a couple of those predictions™.

    Even so, they nearly missed out completely !!

    1. Mindert Eiting

      So it is. With 97 percent errors on your exam, you still have 3 percent correct. Don’t underestimate the giant propaganda machine, now running hot up to Paris. In our MSM we will read more about these victories, defying elementary logic. If the public digests your BS, your propaganda is succesful.

  7. sod

    Have you taken a look at the graph at all?

    http://www.nature.com/nclimate/journal/v5/n5/fig_tab/nclimate2575_F1.html

    He takes the model runs that have a 15 year “pause” at around 2000 (darker lines) and then looks if those are different to the rest (lighter colour).

    The average (very thick lines) demonstrate, that they are not.

    That is, what the article shows. Makes sense to me and contradicts a claim that is often made (“no models show a pause”…)

    1. AndyG55

      Nature , climate change……… roflmao..

      read Rolling Stoned instead, its probably more accurate.

      You would fall for any mathematical CON they these guy threw at you, because you are a mathematical moron.

      The models have FAILED, have got nowhere near reality, and never will, because they are based on erroneous assumptions.

      Mingland will being to look even more the FOOL as the temperature continues to drop.

      The little molehill of warming out of the LIA has ceased, a mild cooling trend has already started.

      Realise that , and stop your childish squirming !!

    2. AndyG55

      So he runs it multiple times, then chooses the ones that, by pure chance, get somewhere near reality, thus proving he hasn’t even got the most basic idea what he is doing.

      If he had any idea, all his result would be near reality, but they aren’t.. not within cooeeeeeee.

      Anyone got a wonky coin ????

      1. sod

        you do not understand how models work. They simulate volcanoes, for example.

        He also did not make model runs but used those that were documented already, as far as i see it.

        1. Bart

          sod – if your model results are so broad that they cover any eventuality, they they have no predictive power, cannot be falsified, and are essentially useless for confirmation or any other purpose. Saying “there’s still a chance we might be right” is a very far cry from validation.

          It is certainly a non-standard definition of “robust”. When we say an estimator is “robust”, we usually mean that, despite uncertain data with outliers, we manage a fairly good representation of reality and, moreover, that representation tends to get better with time.

          Here, we have the precise antithesis. Not only do most model results fail to extrapolate the outcome faithfully, but even those that could be said loosely to do so are not heading in the same direction. The “pause” is still pausing, even taking what appears to be a discernible downturn, yet none of the models exhibit this behavior.

        2. AndyG55

          “you do not understand how models work”

          roflmao.. again, with your ignorance.

          I work with engineering and climate models.
          I have helped write a couple that are used in industry, ….because they work.

          I do know how models work, and how they don’t work.

          It is you that has hallucinogenic phantasies about the climate models.

          In climate science, “robust” means they keep giving the same WRONG answer. 😉

        3. DirkH

          sod says:
          24. May 2015 at 2:00 PM
          “you do not understand how models work. They simulate volcanoes, for example.”

          You are hallucinating.

          1. AndyG55

            Actually they do simulate volcanoes

            http://www.pi.ingv.it/galleria/foto/uf2/d14.jpg

            but only the immediate form of the volcano from known data. Its basic engineering modelling.

            No predictive power what-so-ever.

            Climate models on the other hand are based on many fudged and battered unknowns….. they are NOT engineering models, they are game models based on fallacies …

            and they also have no predictive power what-so-ever. As they have shown.

            The fact that s.o.b think they are in any way similar is a further illustration of his base-line ignorance.

          2. DirkH

            Well we know they use partially made to fit, partially true historic forcing data for their hindcasting, fitting the aerosol forcing to what they need, of course. I wouldn’t call that “simulating a volcano”. That’s like saying Nightmare On Elm Street simulates a typical teenager’s life.

          3. AndyG55

            The picture of the simulation vs the volcano is actually a “validation” of the model.

            Ask a climate modeller what “validation” means.. all you will get is a dull blank gauping face staring back at you. 🙂

  8. sod

    “With 97 percent errors on your exam, you still have 3 percent correct.”

    That is a false description of what is happening.

    A better one works like this:

    Do not claim that somebody failed, because he made 15 errors early in the exam, because people who do not get different results at the full exam than those who do not have such a “mental pause”.

    1. AndyG55

      15 out of 16 questions.. !!

      better than you are doing, that’s for sure.

      You haven’t got one thing right since you started your ranting and raving !

      1. Graeme No.3

        Andy55:

        Don’t bother; he hasn’t quite grasped that there is no sunlight at night, or when it is overcast.

        1. AndyG55

          Comes from living in his grandma’s basement , I guess.

          He will never see the light.

    2. AndyG55

      poor sob… “because he made 15 errors early in the exam”

      ……. then you ran out of time.

    3. DirkH

      sod, you boasted of your supreme formal mathematical education. Now you come up with that waffling to defend the Climate Models?

      No, please explain to us how and why the Climate Models can model the real happenings in spite of their shortcomings. Address the shortcomings, the idealized parametrized imitations of statistical descriptions of microprocesses in huge grid boxes , and how the models can be successful even though they don’t describe thunderstorms.

      Show us your shiny Maths-Fu, sod.

      1. DirkH

        Hmm. No answer. Maybe I used too complicated words.
        Sod, how do Climate Models copy with error propagation? Do they compute uncertainty intervals around their variables, and increase them from timestep to timestep as the iterative computation introduces more bit noise with each timestep? Or do they just let the bit noise amplify from time step to time step? Do they work with standard double IEEE 754 64 bit format? After how many time steps do all the bits become meaningless?
        How big is the estimated deviation from the real, nonlinear, chaotic system from step to step?

        1. AndyG55

          “how do Climate Models copy with error propagation”

          They manage error propaganda quite well.. In fact its the only thing they seem to be able to do, propagate errors.

  9. chris moffatt

    “This demonstrates the robust nature of twenty-first century warming projections”

    The IPCC has told us again and again that these are “scenarios” not “projections” – possibilities not predictions. Bit of a fail, what!

  10. sod

    Again, what is the problem with his approach?

    There is deafening silence on the main point of the paper discussed here, among the shouting and insults about other topics that you want to have discussed:

    so again, what is the problem with his approach?

    http://www.nature.com/nclimate/journal/v5/n5/fig_tab/nclimate2575_F1.html

    PS: will i hear somebody admit, that the models simulate 15 years “pauses” rather well?

    1. DirkH

      They simulate something “rather well” while their chosen metric, global average temperature, the metric around which the catastrophism of the warmunists circles, did something completely different from reality?

      You are a member of a crazy cult.
      http://en.wikipedia.org/wiki/When_Prophecy_Fails

    2. Bart

      Once again:

      sod – if your model results are so broad that they cover any eventuality, they they have no predictive power, cannot be falsified, and are essentially useless for confirmation or any other purpose. Saying “there’s still a chance we might be right” is a very far cry from validation.

      It is certainly a non-standard definition of “robust”. When we say an estimator is “robust”, we usually mean that, despite uncertain data with outliers, we manage a fairly good representation of reality and, moreover, that representation tends to get better with time.

      Here, we have the precise antithesis. Not only do most model results fail to extrapolate the outcome faithfully, but even those that could be said loosely to do so are not heading in the same direction. The “pause” is still pausing, even taking what appears to be a discernible downturn, yet none of the models exhibit this behavior.

  11. mwhite