Daily Mail touts McKitrick study using biased methods to question climate change


The Daily Mail (UK) has reported on a study by Ross McKitrick of the University of Guelph in Canada that, yet again, resurrects the tired old “warming pause” meme.  Climate Nexus notes that the study uses methods designed to show no trend; ignores a large body of contradictory scientific evidence; and was published in a journal known to be a “predatory publisher.”  McKitrick, a well-known antagonist to mainstream climate scientists, is using bad analysis to reach a pre-determined conclusion.

The following is a guest post from Climate Nexus (text in PDF format here):

Daily Mail Touts Study Using Biased Methods

to Question Climate Change

The Daily Mail (UK) has reported on a study by Ross McKitrick of the University of Guelph in Canada that, yet again, resurrects the tired old “warming pause” meme. The study looks at the change in average global temperatures over a narrow timeframe. This is problematic in that short-term variability can affect the trend. For example, studies confirm that the oceans are absorbing an increased amount of heat, allowing surface temperatures to level off temporarily. Scientists agree that over the past 50 years, global temperatures have been rising due to human activity, and this rise is projected to continue.

  • The study uses methods designed to show no trend.
  • The study ignores a large body of contradictory scientific evidence.
  • The study was published in a journal known to be a “predatory publisher.”

The study methodology generates trend observations that are not meaningful.

The study is not statistically rigorous, in that it simply uses a single starting point near the present, and works backwards in time to test if the sample period shows a trend. This doesn’t make a lot of sense as a way to test the accuracy of climate models, which are built to function on far longer timescales. It also isn’t even an effective way to prove a lack of trend. For example, the trend for the period McKitrick highlights is not quite statistically significant, but the confidence interval shows it is equally likely to be 0.19ºC of warming per decade, as it is to be 0ºC. That’s not the same as proving no trend.

One blogger even tried the same technique using simulated data that DID have a constant trend, and found that in 10% of cases McKitrick’s methods would have still showed a pause starting in 1995 or before.

The study ignores previous evidence showing that the pause may be exaggerated.

For example, Cowtan and Way found that undersampling of the Arctic in the modern temperature record may have biased recent global average temperatures too low. If they are correct, the slowdown in surface warming may be half of what the HadCRUT data series (which McKitrick uses) shows.

There are also many contributors to the surface warming slowdown that are still consistent with long-term human-caused warming. One big factor is the deep oceans’ submergence of heat, which new research shows has increased in recent years (in both the Pacific and Atlantic). History tells us that this uptake is likely to eventually reverse, in turn causing surface warming to accelerate once again. Surface and ocean temperatures may seesaw, but the overall global temperature trend remains upward.

The journal that published McKitrick’s study is on a list of “predatory publishers.”

The number of predatory publishers has skyrocketed in recent years, growing from 18 in 2011 to 477 in 2014. These publishers rely on an open-access model in which authors pay the journal for publication, a process that critics say leads to corruption and lack of peer-review. The “Open Journal of Statistics” published the McKitrick study, and is under the Scientific Research Publishing imprint which appears on lists of predatory publishers.

McKitrick doesn’t have a pristine scientific record either. One of his most blatant mistakes was to measure angles in degrees instead of radians, a basic and extremely important distinction in units of measurement. He is also a well-known antagonist to mainstream climate scientists, inserting himself into the “Climategate” controversy in 2010 by alleging that scientists were “faking the match” between proxy and temperature data.

McKitrick’s study seeks to perpetuate a meme that has been refuted over and over. It uses poorly developed statistical techniques to prop up a pre-determined goal.

*    *    *

Earlier post:

McKitrick tries and fails to move the goalposts on climate action (June 19)

This entry was posted in Global Warming Denial Machine. Bookmark the permalink.

8 Responses to Daily Mail touts McKitrick study using biased methods to question climate change

  1. It is quite telling that you didn’t link to the article for your readers to look at for themselves. They can get it at my website (rossmckitrick.com) or at the journal (http://dx.doi.org/10.4236/ojs.2014.47050).
    The blog post by Richard Telford that you link to did not try “the same technique using simulated data”. If you read my comment underneath his post you will see that he applied a different definition for a pause, one that is more likely to detect a long one than I used in my paper. The blogger has not yet responded.

    The radians error occurred in a 2004 paper, and you fail to mention that I promptly corrected it in print and showed that it did not change the conclusions of the paper:

  2. Greg Laden says:

    Ross, looking at the abstract from your paper, I have these comments:

    “The IPCC has drawn attention to”

    The IPCC does not really do research, but refers to other research. I’m surprised a peer reviewed paper would be allowed to reference its raison d’être in this manner in the abstract rather than referencing other peer reviewed literature.

    “… an apparent leveling-off of globally-averaged temperatures over the past 15 years or so.”

    The “leveling off” is not in global temperatures, but in measurements of surface temperatures.

    “ Measuring the duration of the hiatus..”

    Demonstrating a hiatus first would be important. Otherwise this is a great example of the traditional meaning of the phrase “begging the quesiton.”

    “ has implications for determining if the underlying trend has changed, and for evaluating climate models.”

    Climate models can always be improved, but here the question is not about models but about the lack of information detailed enough from the deeper ocean regions to incorporate into sophisticated models.

    “ Here, I propose a method for estimating the duration of the hiatus that is robust to unknown forms of heteroskedasticity and autocorrelation (HAC) in the temperature series and to cherry-picking of endpoints.”

    But it looks like you did that by cherry picking the endpoints. Or, at least, you need to redo the analysis with multiple endpoints to eliminate the effects of staring and ending points. Which, if you do, would mostly invalidate your analysis.

  3. caerbannog says:

    In the paper that originally contained the degrees/radians error, authors McKitrick & Michaels concluded:

    “The socioeconomic effects in the data are shown to add up to a net warming bias, although more precise estimation of its magnitude will require further research”

    That further research being, perhaps, testing their conclusion by computing global-average temperature results from a variety of subsets of actual station data?

    Given that the global land temperature trends reported by NASA/NOAA/etc. can be confirmed from data taken from just a few dozen temperature stations scattered around the world (out of the *thousands* of temperature stations used by NASA/NOAA), coming up with a variety of ways to test whether “socioeconomic effects” really do bias global-average temperature results significantly would not be difficult at all.

    It seems only logical that the authors of the above paper (or other climate skeptics who were interested in showing that the global temperature record is “biased”) would have followed up with something like this in the decade since that paper was published.

    Of course, if the paper was intended to be a Big-Tobacco-style “doubt is our product” exercise, I can understand very clearly why they had no interest in following up with such an analysis of actual temperature station data.

  4. jsam says:

    When one declares one’s faith in the science being wrong one has sold one’s soul. The internet never forgets.


  5. Julio Nieves says:

    Mr McKitrick signatory of the Cornwall Alliance, yes sir. I am following you through the web, making sure people know of your dismissing of Law of Conservation of Energy and its implications on the climate // Why do you ignore the proper use of rules of System Design & Analysis for Climate? / Does your religion forbid you to acknowledge the Natural Laws in Climate??


  6. Dikran Marsupial says:

    I have discussed some of the more obvious flaws of the paper here: http://quantpalaeo.wordpress.com/2014/09/03/recipe-for-a-hiatus/

    However we only need to plot the data and look at it to see that the conclusions are incorrect. The paper concludes that “Overall this analysis confirms the point raised in the IPCC report [1] regarding the existence of
    the hiatus and adds more precision to the understanding of its length.” For the RSS data, the “precise” duration of the hiatus is 26 years. Here is a plot of the RSS data with the 26 year hiatus shown:


    Clearly the hiatus did not begin in 1988, there is not even a hint of a decline in the rate of warming in that year (1998 perhaps!). The problem is in McKitricks use of the word “trendless” to describe a trend that does not reach statistical significance, but that is just not what a failure to reject the null hypothesis actually means. This is a particularly egregous instance of the “p-value fallacy” (i.e. assuming that a failure to reject the null hypothesis means the null hypothesis is true).

    Note that the slope of the trend during the “hiatus” is almost exactly the same as the trend for the entire dataset!

    • Kneel says:

      ” assuming that a failure to reject the null hypothesis means the null hypothesis is true”

      Nah – it just means this particular method indicates you need to use other methods and/or data to show there is a trend.
      Spencer, as the “resident expert” on sat. trop. temp, says something similar.
      Global sea ice extent says something similar too.
      Multiple lines of evidence saying the same thing – I guess that means it’s worth a look at least, right?

      • Rick - Climate Science Watch says:

        As if scientific researchers haven’t taken “a look” at these issues.
        Not many leading scientists would consider Spencer to be the leading expert to start with on this.
        Read Skeptical Science. We don’t engage in science education colloquies on this stuff.

Comments are closed.