3

From Top Advisor To Australian Gov't Says Climate Change is a UN Conspiracy (slashdot.org) and Climate change a UN-led ruse, says Tony Abbott's business adviser Maurice Newman (The Sydney Morning Herald), Maurice Newman, the top business advisor to conservative Australian Prime Minister Tony Abbott:

"It's a well-kept secret, but 95 per cent of the climate models we are told prove the link between human CO₂ emissions and catastrophic global warming have been found, after nearly two decades of temperature stasis, to be in error ... This is not about facts or logic. It's about a new world order under the control of the UN."

This appears to have been sourced from a report reported in the conservative news source, Daily Caller:

Former NASA scientist Dr. Roy Spencer says that climate models used by government agencies to create policies “have failed miserably.” Spencer analyzed 90 climate models against surface temperature and satellite temperature data, and found that more than 95 percent of the models “have over-forecast the warming trend since 1979, whether we use their own surface temperature dataset (HadCRUT4), or our satellite dataset of lower tropospheric temperatures (UAH).”

Is there any basis for concluding that anywhere around 95% of climate models have been found to be in error in the way Spencer suggests?

Brian M. Hunt
  • 17,999
  • 13
  • 99
  • 176
  • 7
    I would suggest changing the title to something like "have 95% of climate models been found to be in error" instead of "Is climate change a UN conspiracy?" – Jan Vladimir Mostert May 08 '15 at 15:16
  • 2
    This question as is buckets together some disparate claims. It's possible that climate change is happening, that it's not a UN conspiracy, AND YET 95% of the models are in error. In fact, I think that's entirely reasonable. – Ask About Monica May 08 '15 at 16:07
  • 5
    There is a big difference between "found to be in error" (which means "are not perfect") and giving invalid results. The fact that the speaker attributes the errors to a "New World Order" gives you some idea of the level of rationality being applied. – DJClayworth May 08 '15 at 16:21
  • 4
    This question is unanswerable without knowing what is meant by "in error". The complexity of weather patterns mean that no model is perfect, but that doesn't mean that broader predictions are incorrect. – KSmarts May 08 '15 at 16:24
  • 1
    @Brian: The question of whether Newman had any evidence to support his claim is apparently "Yes. The report by Spencer." Rather than make that an answer, I've incorporated it into the question, so the more interesting issue can be addressed (is the report correct, under the definitions it provides), rather than bickering about what Newman might have meant. I hope that will get a more satisfactory answer for you. – Oddthinking May 08 '15 at 17:22
  • I read the claim as "the model are wrong on purpose" (i.e. data has been tampered). – nico May 08 '15 at 17:57
  • @nico - then you may be reading it incorrectly. It can also be interpeted as "UN is using results achieved by incorrect models to push a political agenda, by deliberately painting the models as more correct than they are". – user5341 May 08 '15 at 18:23
  • Thanks all for the feedback and editing. The Spencer quote definitely narrows the claim and makes it more answerable (IMHO). Are there any other suggestions on how to improve this question? – Brian M. Hunt May 08 '15 at 18:37
  • The un-abbreviated quote was, `It's a well-kept secret, but 95 per cent of the climate models we are told prove the link between human CO₂ emissions and catastrophic global warming have been found, after nearly two decades of temperature stasis, to be in error` -- so a question could be about the claim that there have been "nearly two decades of temperature stasis"? – ChrisW May 08 '15 at 19:33
  • Thanks @ChrisW -- do you think the claimant is simply stating that the models are in error because they do not explain the near-20 year year lull in temperature warming? – Brian M. Hunt May 08 '15 at 22:01
  • I don't know what the claimant is stating or why, and I don't know about global warming, but that's what I inferred from the claim I quoted: i.e. that the evidence or reason for saying that the models are in error is that there has (allegedly) been statis for two decades. – ChrisW May 08 '15 at 22:05
  • Surely the real question, however expressed, is the perfectly reasonable "do current climate models have any forecasting skill when compared to actual temperature observations". Even if there is no pause, models may consistently overestimate the extent of actual warming which surely calls into question their forecasting skill. – matt_black May 08 '15 at 23:08
  • [Very closely related question about Spencer...](http://skeptics.stackexchange.com/questions/17655/does-this-graph-show-climate-change-predictions-dont-meet-observations) – Oddthinking May 09 '15 at 00:42
  • Maybe this is a duplicate. How about asking whether Communist China is better than democracies at addressing global warming? :-) – ChrisW May 09 '15 at 00:48
  • @Brian M. Hunt: Re the supposed "20-year lull", why is it necessary to explain something that simply hasn't happened? See e.g. this: http://www.climatecentral.org/gallery/graphics/10-warmest-years-globally – jamesqf May 09 '15 at 01:59
  • For a sufficiently tight requirement of being accurate is claim is trivially true--you have a range of projections but only one real value. Obviously the vast majority are misses. Obviously, also, we shouldn't draw any conclusions from that. – Loren Pechtel May 09 '15 at 20:09
  • Of the 3 links provided: the first is to a slashdot summary of a paywalled claim which does not let us read it, the second to a news article describing a different conspiracy theory and the last is a different, and duplicate claim. As such, I've voted to close as duplicate of the latter. – Sklivvz May 10 '15 at 14:16
  • 1
    Thanks all. I think Sklivvz hit the nail on the head with the duplicate. It is basically Spencer's claim that is in question, at least in my mind. That said, when I searched I did not find Spencer's claim so hopefully this question will close the search criteria gap. Of course if there is another skepticism to be found in the linked articles, I think it'd be find for someone else to open another question. – Brian M. Hunt May 10 '15 at 22:41

1 Answers1

2

The simple answer to this question is yes.

But the debate has been significantly obscured because the majority of contributors to it have chosen what appear to be poor ways of testing the claim with actual data (which is not, in principle, hard as the key data from models and observations are accessible).

Climate skeptics have used time series plots to compare models to observations but these are beset by bad choices of the normalisations for temperature anomalies and are a poor way to compare the underlying patterns due to the noisy nature of temperature time series.

See the claims Spencer makes here WattUpWithThat and a good critique of the errors here.

But there are ways of simplifying the comparison that don't suffer from these defects and these have been done by people who don't seem to have the same partisan desire to reach a particular conclusion.

This picture from a Nature article is the clearest analysis that avoids the various obfuscations possible when looking at noisy time series data:

nature chart

The choice made by these authors is to compare just the average extent of warming over a period rather than the noisy time series. This appears to avoid some of the choices that make the time series comparison so obfuscating and controversial. Their result seems to show that most models really do overestimate the actual amount of warming over recent years. In their words:

Recent observed global warming is significantly less than that simulated by climate models.

In somewhat more detail the authors talk about the statistics like this (my highlights):

The evidence, therefore, indicates that the current generation of climate models (when run as a group, with the CMIP5 prescribed forcings) do not reproduce the observed global warming over the past 20 years, or the slowdown in global warming over the past fifteen years. This interpretation is supported by statistical tests of the null hypothesis that the observed and model mean trends are equal, assuming that either: (1) the models are exchangeable with each other (that is, the ‘truth plus error’ view); or (2) the models are exchangeable with each other and with the observations (see Supplementary Information). Differences between observed and simulated 20-year trends have p values (Supplementary Information) that drop to close to zero by 1993–2012 under assumption (1) and to 0.04 under assumption (2) (Fig. 2c). Here we note that the smaller the p value is, the stronger the evidence against the null hypothesis. On this basis, the rarity of the 1993–2012 trend difference under assumption (1) is obvious. Under assumption (2), this implies that such an inconsistency is only expected to occur by chance once in 500 years, if 20-year periods are considered statistically independent. Similar results apply to trends for 1998–2012. In conclusion, we reject the null hypothesis that the observed and model mean trends are equal.

matt_black
  • 56,186
  • 16
  • 175
  • 373
  • 2
    Your conclusion is unwarranted by your evidence: the evidence shows that there was less warming than the prediction, but still within 2.5 sigma, thus less so *not significantly* less so, at best *barely significant*. – Sklivvz May 09 '15 at 13:07
  • 1
    @Sklivvz It isn't **my** conclusion but that of the nature paper. And your metric for significance is an *extraordinarily* high threshold. We are not comparing the noise in multiple observations for a single model with noise from observations (which I think is what your statistic implies). – matt_black May 09 '15 at 13:20
  • 1
    It isn't their *conclusion*, it's just a bit of the abstract. I can't access the full paper but they calculate a p-value to determine significance, according to [their supplementary information](http://www.nature.com/nclimate/journal/v3/n9/full/nclimate1972.html?WT.ec_id=NCLIMATE-201309#supplementary-information) what is this p-value? – Sklivvz May 09 '15 at 14:30
  • 1
    @Sklivvz I'll add some text to my answer to cover it. – matt_black May 09 '15 at 14:36
  • 1
    This paper doesn't convince me. A p-value of 0.04 is very bad, and rejecting a null-hypothesis with it is going to give a wrong result in 15-30% of the cases ([ref](http://blog.minitab.com/blog/adventures-in-statistics/how-to-correctly-interpret-p-values)). As such it's considered "mild" (or weak/moderate) significance ([ref](http://www.stat.ualberta.ca/~hooper/teaching/misc/Pvalue.pdf)) (in physics, non-significant). While you now report the paper correctly, the paper itself seems poor to me, but I'm not a climate scientist, of course. – Sklivvz May 09 '15 at 15:02
  • @Sklivvz Why choose that p-value? they also quote a more natural one given the nature of observations that is close to zero. And the paper gives enough details of the nature of the uncertainty (error bars for observation and for individual model runs) that mostly don't overlap *at all*. The distribution of model runs *isn't uncertainty in the results* it is different internal modelling parameters or algorithms. The obvious conclusion is most of those choices are just wrong. – matt_black May 09 '15 at 15:07
  • @Sklivvz I'm also not sure p-values are as relevant as you think here or should be interpreted the way you do. Besides see [this refresher](http://www.nature.com/news/scientific-method-statistical-errors-1.14700) – matt_black May 09 '15 at 15:18
  • Let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/23610/discussion-between-sklivvz-and-matt-black). – Sklivvz May 09 '15 at 15:30
  • 1
    Another thing, Matt: I don't see where your evidence supports that *95%* of them are wrong. As far as I can tell, the paper doesn't really address the claim as it is right now. – Sklivvz May 09 '15 at 18:52
  • 1
    By the way the Fyfe et al paper uses the model spread to estimate the plausible effects of natural variability in performing the test. This means that either models over-estimate temperature or underestimate the variability. Sadly this caveat doesn't appear in the paper, but in my opinion, it should have. –  May 09 '15 at 20:37
  • @DikranMarsupial which is why I deprecated the Spencer plot and why the Nature paper is a much better way to compare models and reality (their stats don't depend on artificial choices of start date). – matt_black May 09 '15 at 20:38
  • @DikranMarsupial The Fyfe et. al. paper shows model spread but this includes different choices of parameters and algorithms not just the impact of natural variability. – matt_black May 09 '15 at 20:39
  • @DikranMarsupial I suppose I think that neither is a good way to do the comparison (which is what I think i said). This is why the Fyfe paper is a better source. their comparisons method is better and less ambiguous. – matt_black May 09 '15 at 20:42
  • @DikranMarsupial The point I made was based on the Fyfe et. al. results. If you want to show the Tamino version as an alternative, reproduce its results in the same format as the Fyfe results. I wasn't trying to refute Tamino just to point out I prefer the Fyfe method. – matt_black May 09 '15 at 20:47
  • I said clearly that I didn't like the Spencer result so we should use something else to do the comparison. Did you actually read what I wrote? (PS I might well remove it since it appears that to mention it even when saying it is unreliable is a red rag to a bull for some people). – matt_black May 09 '15 at 20:52
  • @DikranMarsupial I provided it for context and i linked to the serious criticism that has been made of it. I included Tamino for balance though I have no methodological way to check what he did. The point of the answer is not about either spencer or Tamino but about the Fyfe et. al. paper. Perhaps I should have ignored the context or perhaps some people are just too sensitive when any criticism of mainstream climate science is given publicity even when the point is to criticise it. – matt_black May 09 '15 at 21:02
  • @DikranMarsupial How is criticising Spencer "misinformation". – matt_black May 09 '15 at 21:12
  • @matt_black cheers, +1. –  May 09 '15 at 21:22
  • @DikranMarsupial It's deleted. But you appear to be saying that even if I published the Spencer result and said it was bad that this would be a bad thing. I really don't get that apparent desire to never give Spencer any publicity even when criticising him. – matt_black May 09 '15 at 21:22
  • 1
    @Spencers analyses have been criticised elsewhere, just giving references to the sources is much better. –  May 09 '15 at 21:23
  • 2
    @DikranMarsupial By the way, is Judith Curry right in saying there have been almost no published comparisons of models versus observations? I've struggled to find any doing a good job other than Fyfe et. al. – matt_black May 09 '15 at 21:25
  • 2
    no, there are a series of projects (CMIP being only one of them, see e.g. http://www-pcmdi.llnl.gov/projects/model_intercomparison.php) that perform comparisons between models and observations. There are plenty diagrams in the IPCC reports comparing observations with model output. There are many climatologists that work on this, e.g. Ben Santer. Perhaps Prof. Curry is referring to some specific form of model-obs comparison? –  May 09 '15 at 21:30
  • 2
    -1 Cherry picked "Commentary" article used to support the [6th most popular climate change myth](http://skepticalscience.com/climate-models.htm) (with links to the leading climate change denial blog no less). I.e., [No, climate models aren’t exaggerating global warming](http://www.washingtonpost.com/news/energy-environment/wp/2015/02/04/no-climate-models-didnt-overestimate-global-warming/). – Rusty May 11 '15 at 06:23
  • 1
    The cited paper only addresses the last 20 years (as stated in the title), but the question appears to be more general, and the answer to the more general question is a clear "No". It may well be that models have over-estimated warming during the last 20 years, but short term variation is to be expected. Of course, models have also UNDER-forecast sea level rise and polar ice melting. – Mark May 13 '15 at 10:17