Global Warming: When Judith Curry makes a claim, check for yourself

Climate models do an excellent job reproducing how Earth’s temperature has changed over the last century and more. But when Judith Curry gives one of her interesting presentations, she seems determined to imply that they don’t.


Here, for instance, is a slide from one of her presentations, about what caused the observed global warming:

What I find puzzling is her claim of “Cooling 1940-1975: Not reproduced by the models.” It seems to me to be an attempt to discredit the models. But where did she get that conclusion? Is she unable to “do the math”? Did she just make it up?

Let’s look for the truth of the matter.

Here’s the observational data (HadCRUT4, from the Hadley Centre/Climate Research Unit in the U.K.) for the time span 1940 through 1975:

I’ve included a trend line (in red) as well as its uncertainty (dashed red lines). It’s sloping downward so you could call it “cooling” but the uncertainty is so big that there’s no being sure about it — or even being confident about it. The rate is estimated at -0.0028 ± 0.0035 °C/year, which includes zero in the credible range. I’d say no, it doesn’t show cooling it just shows the never-ending fluctuations which Judith Curry has decided to call “cooling” with no ifs, ands, or buts. Not very scientific, Dr. Curry.

What about the models? Here’s the multi-model mean for the RCP4.5 through RCP8.5 scenario runs:

I’ve included a trend line (in red) as well as its uncertainty (dashed red lines). It’s sloping downward so you could call it “cooling” but the uncertainty is so big that there’s no being sure about it — or even being confident about it. The rate is estimated at -0.0011 ± 0.0025 °C/year, which includes zero in the credible range. I’d say no, it doesn’t show cooling, but neither does the observational data.

Here’s the kicker: if you insist on statistical significance then neither shows cooling. If you take any negative value, significant or not, as “cooling” then they both show cooling.

You can’t have it both ways, Dr. Curry. Either neither or both, but to state that the observations show cooling 1940-1975 while the models don’t, is false.

So tell us, Judith Curry, what’s the scientific basis for your claim? You also said in one of your presentations that your job as a scientist is “to critically evaluate evidence and challenge and reassess conclusions drawn from the evidence.” It looks like you’re happy to “challenge and reassess conclusions” — unless they’re your own conclusions.


This blog is made possible by readers like you; join others by donating at My Wee Dragon.


Advertisements

29 responses to “Global Warming: When Judith Curry makes a claim, check for yourself

  1. When Judith Curry makes a claim you call for prosecution. The case to give her is murder.

    • Over the top.

    • Mal Adapted

      I don’t want to prosecute anyone either for their relative contributions to AGW, or their relative effectiveness at delaying collective action to decarbonize. I just want everyone to pay for their own marginal climate change costs.

      I do condone ‘free and robust’ verbal engagement with AGW-deniers, OTOH, up to but not including incitement or libel, and especially with mercenaries. I want to muster public opprobrium against those who won’t take responsibility for their marginal climate change costs, to encourage them to censor themselves.

      I don’t mind if fossil fuel investors are stuck with billions in stranded assets, either. Nor do I mind if workers in the fossil fuel industry have to retrain and/or relocate for employment, although I’d be willing to use public funds to help them with that. But decarbonization has to happen one way or another, and hopefully sooner rather than later. That will require collective, i.e. political, decision making. For it to happen, AGW-denial must not be seen as respectable!

  2. LOL at still holding onto the “pause” and circling it in graphs.

    Elsewhere in her SPE presentation (pg 13 in the PDF) she simultaneously claims that 2015 was only notable because of El Nino and because of the lack of warming since 1998… which is only an artifact of starting to count at the peak of the biggest recent El Nino.

    • Pathetic, isn’t, for an actual scientist? You can understand non-experts falling for this (at least originally before it’s explained), especially when they are only shown part of the dataset, but what on earth is Curry’s excuse?

      • Basically the Curry-Lewis gang willfully ignore the plus side of natural variability. NV cools; abrupt climate change cools. That is all they allow.

        Currently the AMO is diving: North Atlantic ARC. This is going to be milked as a precursor for global cooling. Cooling is always, and I mean ALWAYS, right around the next corner. Sea ice is about to rebound. Mass is going up on Antarctica. It’s snowing on Greenland. They grab any little toehold they can find to continue their delusions.

  3. Chris O'Neill

    If you cherry-pick from 1940 to (the beginning of) 1957 in the Sks calculator then you’ll get -0.185±0.117℃/decade for GISSTemp so it is possible to get a period of statistically significant global cooling (that ended in the 1950s, i.e. well before 1975).

    But tricks like this are often used in science denial memes to claim that because models don’t exactly match observations, therefore increased CO2 levels don’t cause global warming.

    • As tamino has instructed on MANY occasions–and any good stats course will teach–you DON’T see “statistically significant” cooling by scanning the record looking subregions without correcting for the multiple testing problem.

      Simple example: Flipping 5 tails in a row is “significant”. Now flip a coin 100 times. What are the odds of seeing a “statistically significant” run of 5 or more tails somewhere along the way? Answer: ~81%.
      Further, on average one will see not just 1 but just over 1.5 such “statistically significant” series in each run of 100 flips which of course means such observations are not significant as well.

      This can be worked out analytically straightforwardly enough, but simply running the following simple R code will verify the above.

      ###########################################
      Gen <- function() {

      # generate series of 100 heads and tails where
      # 1=H and 0=T and run length encode it
      X <-rle(sample(0:1,100,replace=TRUE))
      # treat returned dataframe as a matrix
      X2 <- rbind(as.numeric(X$lengths),as.numeric(X$values))
      # extract all run lengths of tails
      X3 <- X2[, which(X2[2,]== 0)]
      # extract runs of 5 or more tails
      X4 = 5)]
      return(X4)

      }

      # generate 100,000 runs of 100 coin flips
      q 0)
      # count total number of times such runs occur
      length(unlist(q))
      #######################################

  4. The Very Reverend Jebediah Hypotenuse

    Dr Judith Curry doesn’t even deign to accept the science that says that increasing atmospheric CO2 is the main forcing of current climate change.

    But just in case it is actually CO2, she also doubts that the rise in atmospheric CO2 is anthropogenic.

    But, she’s very popular with pseudo-science blog-cranks and Republicans – so there’s that.

  5. The Very Reverend Jebediah Hypotenuse

    One of Curry’s recent blog-posts:

    How can the fundamental disagreement about the causes of climate change be most effectively communicated?

    Which sort of begs the question… what “fundamental disagreement”.

    • Chris O'Neill

      “fundamental disagreement”.

      Just the sort of claim we expect from someone who jumped the shark a long time ago. http://www.realclimate.org/index.php/archives/2010/07/the-montford-delusion/comment-page-4/#comment-181895

      We are dealing with deranged people.

    • VRJH:

      You miss the point of Currie’s question. It’s not about effectively communicating anything to do with the causes of climate change – it’s about effectively communicating that there is a fundamental disagreement. Even if there isn’t. What is said after that is irrelevant.

      It’s like the old Far Side cartoon where a human is talking to his dog Ginger. All the followers only hear “fundamental disagreement”.

      ahttps://desertdemocrat.files.wordpress.com/2015/03/gary-larson-far-side-cartoon-what-we-say-to-dogs-blah-blah-ginger.jpeg

      • Drat. Get rid of the a before the url to follow the link.

      • The Very Reverend Jebediah Hypotenuse

        Yeah – but most dogs are smarter than that.

        Increasingly, I am viewing climate-denialists as the guy at bottom right:

  6. OK, let me get this straight.

    For the obs, the calculated trend is: -0.0028 ± 0.0035 °C/year.

    For the models, it’s: -0.0011 ± 0.0025 °C/year.

    So the trend is “off” by 0.0017 C/year and that’s a miss? How close do they need to come, anyway?

    Oh, right, you can truthfully say that the obs trend is more than twice as large as the model trend and then trumpet to the world that the models missed by more than 100%!

    How could I miss that juicy possibility, even for a moment?

    • Well that’s just saying that 0.0017 is a small number (like ‘carbon dioxide is only 0.4% of atmosphere’). Which signifies nothing.
      The numbers agree within error bars, which is good. The error bars are significant in this case, so we can’t say that much about what the actual number really is.

      • Of course. The point is that there is some selective vision going on here… that, I guess, and that a “miss” is being claimed even though there is no defined criterion set for assessing that. In other words, Dr. Curry, just like most of the US federal government these days, is willing to ‘just make stuff up’ as she goes along, if only it possesses ‘truthyness.’

  7. You can find some reliable info about global worming here:

    https://brainchecked.com/is-global-warming-real-CF53

  8. Judith Curry’s caption reads: “Cooling 1940-1975: Not reproduced by the models.” That initially suggested to me, as I presume it did to those seeing her presentation, that while observations show cooling during that period the models don’t. As we see from the above analysis, the models also show cooling during that period, just fractionally less than the observations.

    I find it quite impressive that the hind-cast model mean managed to produce a best estimate cooling trend during the same period when the observations suggest a best estimate cooling trend occurred. It suggests at least that they must have got the physical forcings for the models roughly correct. Yet the presentation’s wording manages to turn this quite impressive feat into a negative.

    Judith Curry has come up with a technically accurate statement that gives a misleading impression of the facts. If she did it on purpose, then it perfectly fits the definition of a ‘lie of omission’.

    • That’s a good way of stating the point I tried to get across above.

    • ” If she did it on purpose…”???? If? If? I see that there is a “fundamental disagreement” here. I once made a comment on Curry’s blog about how uniform the sea surface temperature is over large areas of the ocean (because its flat, at sea level and has large thermal inertia and relatively uniform albedo compared to dirt, plants, snow cover) and presented data gathered by towing a thermistor across large swaths of the Atlantic at various seasons.
      Curry’s rejection of the point i was making began “I can’t believe…” It’s not a matter of belief. It’s a matter of observation. It’s a matter of measurements. It’s a matter of doing the math, and understanding the physics – i.e., modelling.
      I doubt she’s sincerely as loony, or as stupid, as she appears. She may be so invested in denial that she is now fooling herself with wilful ignorance, like a high school classmate of mine who scored significantly lower on a multiple choice test than you could get by answering randomly. To get it that wrong, you have to actually know the correct answers.

      • Not precisely, Brian. I once studied this when I noticed certain people scoring significantly below chance on an A:B::C:D analogies test. What this subgroup appeared to be doing was choosing a foil (i.e., wrong answer) that was kind of correct on a single surface dimension, or could be made to be correct with enough side assumptions, but definitely wrong when compared to the deeper analogy being tested.

        They simply appeared to be oblivious to the fact a deeper true analogy might be there or just couldn’t recognize it and went for the more surface/side assumption alternative.

        This doesn’t obviate your point in the least, just puts a different interpretation on it which I think is closer to Curry and some other deniers.

  9. If answer A is correct, answer B is the kind of correct foil, and C and D are clearly wrong, they have to know the difference between A and B to consistently choose the “kind of correct” instead of the correct answer. Someone who has a little knowledge will accidently guess the correct answer more often than random chance if they can reliably reject one or two obviously wrong answers. If you test a people with “kind of correct” answers that require the assumption of there is no GHG effect of CO2, people that have no clue how the physics of CO2 works(Bill Oreilly) will have a different distribution of answers from people who understand physics but deny global warming(Gerhard Gerlich and Ralf D. Tscheuschner) Clueless Bill will get more correct answers by chance than Gerlich or Tscheuschner who will knowingly reject correct answers. If you randomly throw in a few questions about the operation of CO2 lasers, G&T will get more of those questions right, unless they cotton on to the fact that you know their global warming BS is inconsistent with the known physics of industrial lasers.

    • Close. What they were doing (through post interviews) was SEEKING something that made quick sense. Plus as part of test construction foils often are constructed to be of fewer dimensions.

      In a slightly different context, many years ago a cognitive researcher identified a subgroup of students who could reliably guess the correct answers to SAT reading com questions in use at the time without being given the passage to read using a slightly different strategy than the strategy above.