Hide the Incline with the “Rank Trick”

or, Hey David Whitehouse — why is the sky orange?

It seems that David Whitehouse still doesn’t understand basic statistics. I mean the really basic stuff.

But he sure understands how to mislead people! This time Whitehouse is all about how this year will *not* be a record hot year. If you use data from NASA GISS for the whole globe then this year almost surely will be a the hottest on record — but Whitehouse doesn’t mention that. He does glibly quote numbers, not mentioning until the end that he’s using the CRUTEM3gl data set, which is for land areas only. That’s not what most people use when talking about global temperature — we regard the ocean as part of the globe — but it really doesn’t matter. Except as an indicator that David Whitehouse is kinda new to this whole “global temperature” thing, but nonetheless feels qualified to pontificate from his position of ignorance.

But hey — let the baby have his bottle! Let’s take a look at CRUTEM3 data (land-only average temperature from HadCRU). What should we have expected all along from 2010? Here’s the CRUTEM3 data (annual averages) from 1975 through 2009:

I’ve plotted the trend over this time period (which is approximately linear), and extrapolated the trend to 2010, in order to determine what the “upward trend” theory (a.k.a. “global warming”) would have predicted for this year (indicated by the red “x”). It should be about the 3rd-hottest on record (for CRUTEM3gl). Although we only have 2010 data for 10 months (the year not yet being over), here’s the average thus far:

Well gosh! It looks like this year is set to come in right on target — in fact, it’s set to come in a wee bit hotter than prediction — if you use that “upward trend” theory of global temperature. The one called “global warming.”

Once again, the observed data follow the existing global warming trend almost exactly. And once again, spinmeisters find a way to make that seem to “falsify” global warming.

And how, pray tell, does Whitehouse “spin away” the fact that this year’s temperature will be right on target according to global warming? He uses the “rank trick” — pointing out that this year will not rank #1 hottest year (if you use land-only data from HadCRU).

If the media headlines are to be believed 2010 is heading to be either the warmest or in the top three warmest years since the instrumental global temperature records began 150 years ago, and proof that the world is getting ever warmer. But looking more closely at the data reveals a different picture.

2010 will be remembered for just two warm months, attributable to the El Nino effect, with the rest of the year being nothing but average, or less than average temperature.

Here’s the month-by-month temperature anomaly from CRUTEM3gl, for the years 1998 to 2010. I’ve plotted 1998 in blue, 2005 in green, and this year (as yet unfinished) in red:

No single year is ranked #1 hottest for every month, but most months this year have been pretty hot compared to years 1998-2009. But is it accurate, is it fair, is it honest, for Whitehouse to characterize most of 2010 as “the rest of the year being nothing but average, or less than average temperature”?

Only if you cheat. Whitehouse begins his missive talking about “since the instrumental global temperature records began 150 years ago,” then conveniently ignores everything before 1998. Here’s the month-by-month temperature anomaly from CRUTEM3gl, for the years 1850 to 2010. As before, I’ve plotted 1998 in blue, 2005 in green, and this year (as yet unfinished) in red:

If you say that any month in 2010 can be called “average, or less than average,” then I say that your statement is inaccurate, unfair, and dishonest.

It’s a very common trick used by those who deny the reality of global warming, to focus only on the “#1 hottest record” so they can draw attention away from what really counts. It’s the trend, stupid.

19 responses to “Hide the Incline with the “Rank Trick”

  1. Out of interest, ccc-gistemp gets a global land/ocean anomaly of +0.73K for November 2010 – the hottest November on record – which puts the global Dec-Nov annual anomaly at +0.65K – the hottest Dec-Nov year on record. Of course, everything is within the error bars of the previous record-holder. These numbers might shift a little in the next few weeks as GHCN numbers settle down.
    The GISTEMP number is not yet published.

    [Response: Thanks for the heads up!]

  2. Disingenuous or plain dishonest of Mr. Whitehouse? I know which I’d guess.

  3. In case you aren’t aware, Whitehouse was cited in a recent story by infamous climate mis-reporter David Rose published in the UK Mail newspaper. The entire story was taken apart by Monbiot:

  4. “Only if you cheat.”

    Tamino, I think you mean “not even if you cheat”. The feebleness of Whitehouse’s argument is such that the data since 1998 is enough to crush it. Including 1850+ just smashes the debris of his argument to dust, and scatters it to the corners of the universe.

    Sometimes I think WTFUWT is just an easy target. Then I remember how much damage they do with their lies. Keep up the crushing!

  5. Obviously, the air conditioner is in the seventh surface station and Jones aligned with Hansen, cuz all the trend-based predictions are coming out right.

  6. FYI: The GISS update for November is out now. Sure enough, 2010 is the warmest year on record. The GLOBAL Land-Ocean Temperature Index for the Dec – Nov period is a full 0.08C above the “unbeatable” 1998.
    I think it’s time they come up with another approach. Their “warming stopped in 1998” is really becoming just plain silly now.

    • They will retreat into any series that finds 2010 is 2nd, which many apparently think HadCRUT and the satellite series will do.

  7. I followd the Whitehead post at WUWT rather closely and found the post strange. First, nearly all of the warmer months from earlier years were post 2000. This seems strongly against the “warming stopped” argument. The overall argument seemed to be “annual records are not that impressive unless each month also sets a record.” Given the amount of natural variability we can hope that we don’t see a year where every or most months are also records.

    [Response: So … you actually *noticed* what the data indicated, and applied critical-thinking skills.

    What a concept!]

  8. Tamino, I was about to make really cinical comment here. I won’t do it, cause I realy appreciate your work here. But I cannot help myself though… There are two rules you should know (sarcasm on):
    1. There is no global warming
    2. If you have some evidence of global warming, so worse for evidence
    (sarcasm off)
    So, keep on good work, Tamino, YOU have learned me a lot.. Not only about GW, but about probability issues also. Thank you really very much :)
    One great THANK YOU, TAMINO !!!

    • Briggs…. that rings a bell somewhere.
      I wonder if someone mentioned him a few years ago on something completely insane. I really can’t recall.

    • I have to say, I am underwhelmed by the work of Mr wmbriggs. No particular insight into either climate or statistics. Nothing I couldn’t get from any denialist website on the former or from an elementary testbook on the latter.

  9. Ideally, there is no Step 5—all data you have in hand should be used to construct your model—but there should be a Step 6, which is “Wait until new data comes in and test the model predictively.” All physical sciences do Step 6—with the exception, perhaps, of climatology

    Briggs is wrong regarding climatology, where models are used to test hypotheses against data on a regular basis …

  10. So according to this trick we can discount 2010 because El Nino produced a couple of really warm months compared to 1998, which also had an El Nino but for the purposes of this trick we don’t talk about that. Makes perfect sense.

  11. In October of 1998 David Whitehouse wrote a news story about a supposed alien signal in an attempt to give credibility to an unknown hoaxter. For those caught in the fallout from the story it was not pretty. I’ve had a low opinion of him ever since then.