Cherry Herring

Fred Singer is on tour denying global warming. The “evidence” he offers is one giant red herring after another.

Probably the most idiotic is his claim that

The global surface warming for 1979 – 1997, reported by CRU-Hadley, NCDC-NOAA, and GISS-NASA, and used by the IPCC to support its claim of a GH-gas cause, is problematic. It is not seen by any other observations; we cite six independent methods , incl. radiosonde, satellite, and proxy data.

First let’s be very clear about one thing: Singer’s choice of the time span from 1979 to 1997 is cherry picking. In fact it’s “cherry picking supreme.” He ends at 1997 for one and only one reason: if you go any further his claim is not just nonsense, it’s obvious nonsense.

And there’s nothing special about 1997 for the IPCC reports, or global warming in general. He gives the impression that this was some kind of “key” time span. It isn’t. Except of course for Singer, because that’s his cherry-picked time frame to support his false claim.

He then goes on to claim that the warming seen in surface temperature data sets (like CRU, NCDC, and GISS) isn’t seen in any other temperature data sets. Like radiosonde and satellite data. But it is seen in those data sets. In fact, they’re in pretty good agreement with the surface temperature data.

Setting aside for the moment that the choice of 1979 through 1997 is truly idiotic, totally irrelevant, and done only because Singer is cherry-picking, what are the trend rates for that time span from, say, surface temperature (NASA GISS), satellite data (UAH TLT), and radiosondes (HadAT2)? Here you are:

They’re all in agreement with each other.

It seems that Singer bases his claim on the fact that for this very limited time span (a mere 19 years) the warming in radiosonde and satellite data isn’t statistically significant. That’s for two reasons. First, 19 years isn’t a long enough time span to find out — you might as well say there’s no statistically significant warming for the last three days. Second, atmospheric temperature shows more fluctuation — more noise — than surface temperature. In large part that’s because it responds more strongly to short-term noise factors like el Nino and volcanic eruptions, as illustrated here.

If you disdain Singer’s cherry-picking, instead studying enough data from GISS, UAH, and HadAT2 to get a good handle on the trends … what do you suppose is the result?

Singer’s entire “masterpiece” is fake science from a fake skeptic. But we’ve come to expect nothing else from the guy who has denied, among other things, the health risks of tobacco smoke.

37 responses to “Cherry Herring

  1. Gosh, I lost track… who is supporting his tour? What are his ticket prices?

  2. It appears consuming this leads to intoxication, it’s not really food:
    Common effects of intoxication include the belief one can be a singer:
    but actually one may fall down
    see, I can do references!

    • See, that’s the problem. Eating that is bad for you in the long-run, sure, but it’s fun and easy and the consequences are entirely disconnected from the action. So it remains popular.

      Except that if someone goes on a diet, and points out that overeating can cause obesity, people don’t then accuse them of wanting to force everyone to be a breatharian.

  3. Someone who heard Singer on this trip told me that he was awful – in the sense of bumbling, incompetent. He is getting on a bit, after all.

    • Awful eh?
      He’s been awful for quite some time, but in a different sense of the word.
      However, I’m thinking this bumbling awful might be a rather good thing.
      A poor presentation does wonders for a pr campaign…not!

  4. Health risks of tobacco smoke? I suppose you mean secondhand smoke. If so, why don’t you say it, for the sake of scientific rigour? And what do you know about the risks of SHS? Another red herring?

    • That’s correct. Singer was not involved in the denial of the health-effects of “first-hand” smoking. He was attacking EPA studies on second-hand smoking in the 1990s.

      [Response: How curious, since as far as I know Singer is not an epidemiologist. I’m quite certain he’s no statistician.]

      • Ol’Fred is the universal specialist. A modern Leonardo da Vinci (and a Galileo too).

        (Or maybe they just couldn’t find a suitable epidemiologist. At least nobody with Fred’s star quality.)

      • Response to Response: You don’t need to be either to debunk the EPA study. But you would have to read and understand it.

        [Response: So your defense of Singer is the claim that you don’t need to be either an epidemiologist or a statistician to “debunk” statistical studies of epidemiology?

        Perhaps you’ll get why we don’t consider you, or Fred Singer, to be real skeptics. You’re fakes.]

    • One could check out the Surgeon General’s Report.

      Fact Sheet

      Full Report (pdf)

      [Response: Who you gonna believe — Fred Singer, or the surgeon general?]

      • But the Surgeon General is employed by the gubmit, while Singer is only supported by the tax avoidance of his fake non-profit organization.

      • Singer was the go to guy for the Tobacco Institute in their attack on the harm of second hand smoke. In their words
        Here is the man who will handle the EPA/ETS (illegible) work Brennon wants for us on the “social costs”. Very impressive resume. I think the project is worth the 20K we discussed.

        As to the report, he used the wrong statistical test to edge the harm into the high 1sigma territory (unless you believe that environmental tobacco smoke (ETS) has health benefits). If you use a one sided test you find that even back in 1994 or so, there was enough data to show the statistically significant harms from ETS. Of course Fred has been trying to justify his acts of negligent homicide recently based on his wrong statistics.

      • Let me guess–he was just following orders? (Albeit the commercial, not military sort.)

    • I suppose tobacco smoke can refer to first or secondhand smoke, so why do you care so much? Oh, right, I suppose you are ready to quote a bunch of libertarian blogs about the supposed lack of risk of secondhand smoke, in contrast to the consensus of every major cancer, heart, lung, and public health organization.

      The libertarian bloggers will of course themselves cherrypick out of context quotes from supposed authorities on statistical significance.

      Or was that not what you had in mind?

    • No, Singer was completely denying that tobacco smoke was hazardous, period.

    • Benpal actually misses the larger point. In documents made public via the litigation against the tobacco companies, it was apparent that Singer was one of the “go to” guys for the tax industry’s misinformation campaign. Part of the settlement of the litigation was that the tobacco companies had to shut down their misinformation campaign. As a consequence, Freewheeling Freddy was out of a job. He then turned his attention to more fertile fields of denialism.

    • In my previous post, the term “tax industry” should have been “tobacco industry.” Sorry for the typo.

  5. Having lived in Europe, I know that there is a little denialism, but that Europeans, especially those on the continent are more willing to accept the consensus of experts, in contrast to the anti-intellectualism and anti-expert biases that are found in segments of American society. So, exactly who is listening to Singer in Europe and is the mainstream media giving him a free pass?

    • I am living in Europe. You are correct, climate denialism is a very minor factor in climate policy and debates. Off course, you can find some weirdos everywhere and of course also in Europe, so I am sure, that he will find some idiots who will be watching his presentations. But they don’t matter in public opinion on climate change and there is no media coverage of Fred Singer’s tour at all in the mainstream media in continental europe. There are some groups who desperately try to get their denial message in the media, but with only low success, because nearly nobody believes their nonsense. And most times when they get the some media attention, it’s not the attention they hoped for, because they are often described as what they are: Fake skeptics backed by fossil fuel money denying reality.

      The whole climate-denial-debate is a phenomenon, that is nearly entirely limited to the US and to a lower degree to a handful of countries like Australia and England. And in my opinion, the only reason for this is, that in the US there are a number of very powerful, well financed political groups (Heartland Institute, Cato Institute, etc) with good connections into (right wing) media and politics, which can spread their denialism to a greater audience. In Europe (and probably elsewhere) those groups don’t have this public outreach and are by far less influential.

      But I think that this denialism will not play a role after let’s say 2020, because at this point there will be so much evidence, that even the least informed sane person will not believe them. Seeding doubt only works temporarily, and they don’t have much else to work with, as the facts are not on their side at all. It was the same in the tobacco/cancer debate in the 60s. I just hope, that they can’t delay climate actions for too long, so that we will still be able to avoid the most drastic consequences. As bad as their denial of the connection between lung cancer and smoking was, their denial of the human role in climate change may cost by far many many more lives.

      • Stephen Baines

        The US has it’s own wierd built in defiance to anyone who seems vaguely like a smarty pants telling others what to do. It’s built int to the culture I’m afriad.

        But I find it interesting that among devloped countries it is the English speaking ones that seem to exhibit the greatest skepticism. The Murdoch empire is almost solely english-language based, right? I wonder if you could develop a statistical relationship between Murdochs influence over media in english speaking countries and some measure of receptance to climate change in those countries. Somebody must have tried that, no?

      • Stephen Baines, Resistance to authority is strong in English history–dating back to at least the Magna Carta, and although the Reformation began in the German principalities, it was the English that made resistance to the pope a litmus test of loyalty.

        Resistance, Rebellion and Regicide could be the English motto. It is unfortunate that resistance to reality is also part of the character.

      • Murdoch was late on the scene.
        If you want geography of the key purveyors of climate anti-science for the last 10-20 years (and many for tobacco companies before that) here’s a Google Map. You can zoom out or in … all the way to K-street.

        it is sadly unsurprising that the language is English,
        and the USA had more tobacco guys for practice.

    • I know he was in the Netherlands, giving a lecture at the Royal Dutch Meteorological Institute (KNMI), but I have not seen any media covering his visit (other than some blogs). I think the same happened in Belgium.

    • I’ve seen him get quoted extensively in mainstream papers in a small European country he visited a couple of years ago. And given a pretty free pass, for an American right-wing loon babbling about a UN conspiracy against our freedoms. But that’s ephemeral.

      In continental Europe, small clusters of enthusiastic deniers can be found in most places. They tend to be dependent on Anglo sources for their copy-paste arguments (though France bottles its own, I think). In most places they have zero influence in policy circles (the Czech Republic, with Vaclav Klaus as president, is one exception; there may be others).

  6. So did he analyse the data and find the best possible pair of years to use? Or was it just a lucky fluke?

    [Response: One can only speculate. But… starting with 1979 is to be able to include satellite data, nothing wrong with that. Ending with 1997 is specifically to avoid the high temperatures from 1998 onward. Singer even mentions this, but claims 1997 to eliminate the influence of el Nino.

    Yet if you actually remove the influence of el Nino, volcanoes, solar activity, you can get a very clear picture of warming, even over such a short time span as 1979-1997. Singer doesn’t seem to want that.]

    • Tamino.

      It is quite clear to anyone with an open mind – and I am curious to know what really goes on in the minds of those like Singer who try to cast the matter otherwise…

      On the matter of that graphic – when El Niño, volcanoes, and solar activity are filtered out there seems to be a bit of a regular oscillation occurring over the last decade or so. At the risk of descending into Orssengo territory, is this simply autocorrelated coincidence, or might there yet be something else percolating underneath?

      [Response: While as-yet undetected factors can’t be ruled out, there’s not yet statistically meaningful evidence of anything other than randomness. Bear in mind that it’s a natural human tendency to detect patterns, whether they’re real or not!]

    • Bear in mind that it’s a natural human tendency to detect patterns, whether they’re real or not!

      Hence the Orssengo nomeclature…


  7. There is a bit more info here: . And quite a bit more on Singer in Oreskes & Conway’s “Merchants of Doubt.”

  8. That the different observation methods (surface, radiosonde, satellite) are in good agreement, one can see for example in this comparison of the different methods by the UK MetOffice:

    Ironically, as you can see in the linked graphic above, from 1978 to 2010, the radiosonde measurements and the RSS satellite measurements show even *greater* warming than the HadCRUT surface measurements.

    What else do you need to know, that Fred Singer is just telling pure crap?

  9. Regarding the graphic, have you done a FT on the data to see any remaining periodicities?


  10. I’ve read Rober Grumbine’s posts on 20 – 30 years of data being a necessary minimum to tease out a climate signal from weather noise. It’s possible to get statistically significant linear temperature trends (global) from 16 and 17 years of data with surface records, so I’ve read.

    Being barely literate in statistics, I can sort of figure out that satellite and radiosonde data, being ‘noisier’ require data of longer time periods in order to winnow statistically significant trends. Shorter time periods are required for, say, Arctic sea ice because of the regularity of the data.

    The use and abuse of time periods is rife across the climate change blogosphere. It may be the most commonly misinterpreted component in the semi-popular debate. I’d like to request a post here explaining statistical significance WRT the surface and satellite temp records (and maybe sea ice for comparison), why they require different minimum time periods, and why too-short time periods give spurious results.

    I know this has been done before, including by Tamino. But the old OpenMind posts are dead (or hard to get to). A refresher would be great – and maybe a new attempt might be even clearer than previous. It’s so frustrating trying to explain something I grasp conceptually, but don’t have the maths skill to evince.

    [Response: That’s an excellent idea. I can’t promise when I’ll get to it, but it’s on the list.]

  11. Barry
    “old OpenMind posts are dead”

    They’ve come back from the dead. see:

    • Already had that bookmarked, Slioch, but the list has grown since last I trawled. Is that all of ’em?

      Looking for a post on statistical significance, I came across this one from mid-2007, where tamino says, “We can even look at a very short time span, say, from 2000 onward… despite the brevity of the time span, there’s still a statistically significant warming trend in both data sets.” [HadCRU, GISS)

      I would flat out have assumed 7 years was way too short to achieve statistical significance in trends from global surface temperature anomalies (though I *know* the data itself, not some magic minimum time period, determines that). Elsewhere and numerously I’ve read that trends from 1998 are not statistically significant… now my brain starts imploding. Further down:

      But even for the brief period since 2000, the trend is still positive, and the estimate is larger than the error range: it’s significant.

      If I learned how the error range was determined – if the maths was made accessible to a dunce like me – I might finally grasp statistical significance beyond the conceptual.

      If anyone knows a good link with no weird maths symbols, I’d follow up avidly. Otherwise, I’ll lurk as usual til tamino finds the time to exposit.

      [Response: I was mistaken in that post, because I failed to account for the autocorrelation in the data (the noise in global temperature is not “white noise.” I’m working on an exposition, but it’ll take some time and there might be some weird math symbols. The fact is that statistical significance is a mathematical phenomenon. But I hope my eventual exposition will at least convey the concepts to those who don’t grok the math.]

  12. A decade later, the same 19-year window (1989–2007 inclusive) would have satellite records very nicely aligned with GISTEMP and showing more warming than HadCRUT or NCDC.

    If the temperature record’s a bowl of cherries, Singer’s argument is the pits.

  13. Been watching the latest video interview with Prof Naomi Oreskes

    Superb. Her message is getting nicely distilled.

  14. Pete Dunkelberg

    It’s Homo misnomer.
    Pete Dunkelberg