E-mail to GRL Editor-in-Chief Noah Diffenbaugh [2016-08-12]

The paper of Lovejoy et al. (2016GL070428) is based on this assertion: “Keenan claims to have used a stochastic model with some realism”. The paper provides no evidence that I make such a claim. Indeed, I do not make such a claim, and the claim is obviously false. Furthermore, I wrote a critique, in 2013, to argue that no one could choose a model with adequate realism.

I have raised this issue with Shaun Lovejoy. Lovejoy now asserts that the following statement, by me, makes the claim: I used a “trendless statistical model, which was fit to a series of global temperatures”. Every statistical model, though, is fit to data—including the IPCC model. For example, if a straight line is used for some data, then the slope, etc., of the line are chosen so as to make the line fit the data as well as feasible.

Fitting a model to some data, though, does not imply anything about how realistic the model is for the data. As an illustration, suppose a straight line is fit to a wave. This does not imply that a straight line is a realistic model for the wave. (All this can be confirmed with any statistician.)

Thus, Lovejoy has still provided no evidence that I make the claim. Indeed, I do not make the claim. Simply put, Lovejoy et al. accuse me of saying something that I never said, and then criticize me for saying it. This is classic rhetorical trickery. It should have no place in science.

On Aug 11, 2016, at 10:32 AM, D.J. Keenan wrote:

David,

The paper of Lovejoy et al. is based on this assertion: “Keenan claims to have used a stochastic model with some realism”. I do not make such a claim. If Lovejoy/AGU are going to assert that I do, then they should supply the source and quote.

On the contest site Keenan states that he used a “trendless statistical model, which was fit to a series of global temperatures”. I therefore think that my statement “Keenan claims to have used a stochastic model with some realism” is quite a reasonable paraphrase of his own words.

In any case, I also demonstrate that the statistics of the model are indeed extremely close to those of the actual record since 1880, with the two variants of the model nicely bounding above and below the actual temperature fluctuation statistics as functions of time scale (fig. 2). The main unrealistic element is the use of Gaussian statistics (instead of the power law, “black swan” statistics), but this only affects the extreme 3% of the fluctuations, and is taken into account in the analysis (see e.g. fig. 3). Indeed it would be hard for Keenan (or anyone) to make a model that more realistically reproduces the global annual series since 1880 while simultaneously being on average trendless. The whole point of the article is that while his model is quite realistic for the series since 1880, that it is quite unrealistic as soon as one considers the pre-iindustrial data.

The claim is obviously false. By asserting that I have made an obviously false claim, Lovejoy/AGU are libeling me. I am not intending to sue for libel. If AGU were concerned to act ethically, though, then it would not publish the paper of Lovejoy et al. Additionally, without the false claim, the paper has no technical basis; so if AGU were concerned about technical validity, it would not publish the paper.

I see that the paper is being published in the journal Geophysical Research Letters. The journal’s editor in chief is Noah Diffenbaugh. Diffenbaugh is also a co-author of the article “Debunking the climate hiatus” [Climatic Change, 2015]. That article is invalid, and my Contest demonstrates such. Moreover, I have previously emailed Diffenbaugh about this issue.

Lovejoy et al. also state the following.

Keenan even made a submission to the UK House of Lords on the issue of uncertainty assumptions. In order to further attract public attention, on November 18, 2015, Keenan publically proposed a “climate contest” with a $100,000 prize (http://www.informath.org/Contest1000.htm).

That is essentially correct (strictly, my critique was not submitted the UK House of Lords, but rather to the UK government). The title of my critique is this: “Statistical Analyses of Surface Temperatures in the IPCC Fifth Assessment Report”.

As the critique explains, when doing statistical analysis, the first step is to choose a model of the process that generated the data. The IPCC did indeed chose a model. I have only claimed that the model used in the Contest is more realistic than the model chosen by the IPCC. Thus, if the Contest model is unrealistic (as it is), then the IPCC model is even more unrealistic. Hence, the IPCC model should not be used. Ergo, the statistical analyses in the IPCC Assessment Report are untenable, as the critique argues.

Now the proponents of the GNF model revert to their prior strategy of endorsing no concrete model (including distancing themselves from their own!) while simultaneously attacking the IPCC. As pointed out in my paper, Keenan does a very good job of concocting a model that underscores the difference between the strong and weak correlation assumptions needed for the trend estimates and uncertainty analysis, and I thank him for this. This thanks is sincere since it is an important scientific point and I have have spent literally over thirty years trying to bring the significance of power laws to the attention of the community. The problem is that when it comes to TREND analysis the difference between the two (scaling / non scaling , power law / exponential, strong / weak correlations) ends up not only being quite small (although significant for winning the contest!), but also it goes in the wrong direction! Had the IPCC used the strong correlation assumption instead of the weak one, their uncertainties would actually have been a little smaller so that their conclusions would have been stronger, not weaker! Indeed, this reduction in the uncertainty is precisely the feature that allows the contest to be (nearly) won! All this shows that while Keenan has watched onto an important scientific issue, that it turns out NOT to be an important issue for anthropogenic warming!

For an illustration, consider the following. Lovejoy et al. assert that the Contest model implies a typical temperature change of 4 °C every 6400 years—which is too large to be realistic. Yet the IPCC model implies a temperature change of about 41 °C every 6400 years. (To confirm this, see Section 8 of the critique and note that 0.85*6400/133 = 41.)

This is ridiculous: the IPCC never claims that their trend continues beyond the industrial epoch! It obviously cannot be extrapolated this way! Keenan’s model is however already far too variable at the century scale, and unless he makes a strong assumption about the correlations in the fluctuations reversing from strongly positive to strongly negative at century scales, the model only needs a slight extrapolation to reach my conclusion at millennial scales.

Thus, the IPCC model is far more unrealistic than the Contest model, according to the test advocated by Lovejoy et al. Hence, if the test advocated by Lovejoy et al. were adopted, then the IPCC statistical analyses are untenable.