Ahh McLean, you’ve done it again

John McLean, the guy who kept guiding Andrew Bolt off cliffs, has this time taken Bob Carter and Chris de Freitas with him. As tamino explains, they say that recent warming trends can be attributed to natural variation, but their analysis removed the trend from the data. See also McLean’s defence and Robert Grumbine’s lucid post. James Annan exposes another error – they fit a step function to that data and conclude that there is step in the data merely because there is a step in the fitted step function. John Lott made the same mistake in his “more guns, less crime” argument, as I showed here.

McLean et al needs to be added to the list of the worst climate papers published in proper journals.

Comments

#1 James Haughton
August 5, 2009

Hi Anthony,
I read your and David’s paper. I agree with Nick that it seems likely that a Chow test will detect a “structural break” when the actual underlying phenomena is an oscillation (ENSO) combined with a linear upward trend (CO2 forcing) + noise, just because oscillation + linear is going to look like a series of steps. The fact that your “breaks” coincide with El Nino and PDO events suggests that this is in fact what is happening. Nor do I think that your analysis, which is in essence describing one set of functions which might fit the data (which, as pointed out above, does not rule out other functions from consideration), is sufficient to support any claims one way or the other about underlying causes, so the snarky comments in the paper about CO2 forcing being unproven are unsubstantiated.
But congratulations on actually writing down your POV in detail and submitting it to a journal.
#2 Bernard J.
August 5, 2009
Cohenite.

I have read Stockwell’s paper, and I know the period they use. I am referring to your focus, which seems to be only on the second half of the 20th century, and I have been trying to pursuade you to consider the basic oscillatory and monotonic trends since 1880.

It appears that you still won’t do your homework, so I can only conclude that:
1. your wife is not home
2. you don’t know how to do what was asked of you
3. you are too lazy to do the work, or
4. you do not want to know what it will infer
#3 Bernard J.
August 6, 2009

Considering the serious lack of justification for Stockwell’s very loose assumptions in applying a test for statistically unequal regression coefficients to a portion of the mean global temperature data, I find it curious, to say the least, that Stockwell would [say this on his blog](http://landshape.org/enm/errors-of-global-warming-effects-modeling/):

Among those believing gross scientific inaccuracies are not justified, and such attitudes diminish the standing of scientists, I was invited to a meeting of a multidisciplinary group of 19 scientists, including Dan Bodkin from UC Santa Barbara, mathematician Matt Sobel, Craig Loehle and others at the Copenhagen base of Bjørn Lomborg, author of The Skeptical Environmentalist. This resulted in Forecasting the Effects of Global Warming on Biodiversity published in 2007 BioScience. We were particularly concerned by the cavalier attitude to model validations in the Thomas paper, and the field in general:

(emphasis mine)

He then continues, quoting from page 228 of the [article in Bioscience](http://tiny.cc/RrPRu) (Biosci 57:3):

Of the modeling [sic] papers we have reviewed, only a few were validated. Commonly, these papers simply correlate present distribution of species with climate variables, then replot the climate for the future from a climate model and, finally, use one-to-one mapping to replot the future distribution of the species, without any validation using independent data. Although some are clear about some of their assumptions (mainly equilibrium assumptions), readers who are not experts in modeling [sic] can easily misinterpret the results as valid and validated. For example, Hitz and Smith (2004) discuss many possible effects of global warming on the basis of a review of modeling [sic] papers, and in this kind of analysis the unvalidated [sic] assumptions of models would most likely be ignored.

One question: what validation has Stockwell employed to justify his use of a Chow test in this instance, to model ‘structural breaks’?
#4 MAB
August 6, 2009

[Anthony's (cohenite') paper](http://arxiv1.library.cornell.edu/abs/0907.1650v1) is sumbmitted for review.
>”*We gratefully acknowledge the comments of John McLean and David Evans*”

Hat tip: [Ferenc Miskolczi](http://landshape.org/enm/wp-content/uploads/2009/04/microsoft-word-herald-article-14-4-091doc.pdf)
#5 MAB
August 6, 2009

Cox writes:

>Assuming a regime-shift from 1978-98 reduces the estimates of the underlying rate of AGW warming from around 0:14_C to 0:05_C per decade. An increase of 0:5_C by 2100 is consistent with low-end empirical estimates of climate sensitivity.

Anthony, any particular reason you chose this assumption?
#6 cohenite
August 6, 2009

BJ; if you have a comment about David why don’t you ask him at his site?

As to validation, have you read the abstract? A couple of points; the break paper was to test the McLean/Quirk thesis; the only variable used or tested is temperature so I’m not sure why the concentration on other ‘variables’; in terms of the CT the coefficients for the 2 sets of data are the different states of El Nino and La Nina; the advantage of the CT in dealing with ENSO oscillations is that where there is less than one wavelength in the data, that is less than the frequency of the change, the CT can pick that up and distinguish that from a non-stationary trend.

Speaking of trends the results are remarkable; there is no trend in the Australian data either side of the 1978 break; the global data has 2 breaks with different trends with the trends consistent with ENSO phase; so even though the paper does not comment on this there appears to be confirmation of regional differences with the global expression of the ENSO phases. But I’m talking to myself so I’ll give up.
#7 MAB
August 6, 2009

Anthony, any particular reason you think the conditions producing the mean temp for the last 10 years should continue for the [next 50 years?](http://i29.tinypic.com/muh7jn.jpg) Why would you assume such a sudden change to the long term trend, based on just 10 years of data?
#8 Bernard J.
August 6, 2009

Cohenite.

If you are a co-author of Stockwell’s paper I should be able to speak to you with equal success – although to date I have not been able to elicit from you any indication at all that you understand the very basic points that have been raised concerning the Chow analysis.

To this end I would ask again: in attempting to discern and thus to prove the existence of ‘breaks’, why have you (Stockwell) not considered, and incorporated into your analysis, other phenomena that are known to influence temperature?

Also, why do you not consider temperature records from New Zealand, from various South Pacific Islands, and from the west coast of South America?

I see that the R routine strucchange, which can detect multiple breaks, was used to perform the analysis. For the ‘breaks’ that you identified, have you discounted that the breaks occur only at the points you publish, and that similar statistically ‘significant’ results might not be identified that do not occur in concurrence with the years that you show? What constitutes an “optimal position” when multiple breaks are identified, and how you (Stockwell) discount that a non-”optimal” break is not in fact a more valid descriptor of an underlying phenomenon, whether such might be the phenomena you consider, or the many that you do not? I am especially interested in your (Stockwell’s) answer as there appear to be so many other impactors on temperature that you (Stockwell) have not considered, that might have an influence on the appearance of a ‘break’.

Have you determined what minimum period of time must be included in a particular regression set in order that inherent noise – or, conversely, autocorrelation – in the data will not give a spurious Chow result?

An what did you, as a former divorce lawyer and as secretary of the [Climate Sceptics Party](http://www.climatesceptics.com.au/), contribute to the paper?
#9 Bernard J.
August 6, 2009

In case there is any doubt, by “you (Stockwell)” I mean “you, or Stockwell”, and not “you, being Stockwell”.
#10 cohenite
August 6, 2009

The assumption of 0.05C per decade is explained in that paragraph; that gives about 0.5C per century which is about what temperatures did increase over the 20thC; I personally think this is too high to attribute to AGW but it is an assumption used to base the forecast on.

The forecast is not based on 10 years data; it is based on the finding by the CT of climate regime-shift based breaks in the temperature data over the period since 1910 and an assumed AGW continuous underlying warming; by excluding any further shifts the forecast emphasises the trend produced by the underlying AGW trend from the base of the CT verified break in 1998.
#11 MAB
August 6, 2009

MAB writes:
>Anthony, any particular reason you think the conditions producing the mean temp for the last 10 years should continue for the next 50 years? Why would you assume such a sudden change to the long term trend, based on just 10 years of data?

cohenite replies:
>The forecast is not based on 10 years data; it is based on the finding by the CT of climate regime-shift based breaks in the temperature data over the period since 1910 and an assumed AGW continuous underlying warming; by excluding any further shifts the forecast emphasises the trend produced by the underlying AGW trend from the base of the CT verified break in 1998.

Anthony, [your cyan line](http://i29.tinypic.com/muh7jn.jpg) just tracks horizontal from the point of the mean of the last 10 years and continues past 2050.

You havn’t found a regime shift, you’ve found chow test result and called it a regime shift. Now you have to justify your chow test in this application.
#12 Mark
August 6, 2009

> “We gratefully acknowledge the comments of John McLean and David Evans”

Is that the same David Evans who exhorts people that models lie and he knows because he made models lie when he was employed?

Surely then this model is lying too.
#13 Former Skeptic
August 6, 2009

Looks like Tamino et al. [have submitted a comment to JGR](http://hot-topic.co.nz/big-guns-brought-to-bear/) in reply to McLean’s crap. Anyone able to find the document online?
#14 Brian D
August 6, 2009

Former Skeptic: Not yet, but if they’ve just submitted it now, it’s probably going to be up for editorial review first. Gareth probably received a copy of the abstract from an insider. When it’s accepted/published, it’ll show up on the databases I search.
#15 James Haughton
August 6, 2009

Hey Anthony, is there a video or transcript of the talk at Newcastle Uni you promoted available online?
#16 Gaz
August 6, 2009

Is that the same David Evans who exhorts people that models lie and he knows because he made models lie when he was employed?

Of course, you must understand that when Evans says “models” in the context of his work at the Australian Greenhouse Office”, he really means “carbon accounting spreadsheets” as opposed to your actual true-blue climate models.
#17 Gareth
August 6, 2009

Brian’s right (#114). Very strong NZ contingent in the authors…
#18 Brian D
August 6, 2009

Courtesy of Joe Romm (and himself courtesy of Andy Revkin), a public link to the paper is available.

It’s missing the usual “submitted/accepted/published” dates and the journal numbers are templates, so this is probably a “submitted”, pre-peer-review comment as of now, and should be treated as such. I’ll be keeping an eye out for when it gets published.

For those who aren’t familiar with the academic publication cycle, this is incredibly fast for a rebuttal anywhere near this complete to be published, let alone one with that many authors. This means the paper is either likely slapped together or is responding to a piece with blatantly obvious errors. Knowing what I do of Tamino’s style and the quality of McLean et al, my money is solidly on the latter.
#19 James Haughton
August 7, 2009

Thanks Brian D. What a smackdown.
#20 cohenite
August 7, 2009

No James, there isn’t; but you can come to the next talk Miklos gives.
#21 Mark
August 9, 2009

> Of course, you must understand that when Evans says “models”… he really means “carbon accounting spreadsheets” as opposed to your actual true-blue climate models.

> Posted by: Gaz

Who’s to say that he’s not using Excel (with Pentium Floating Point Bug?) for his latest work too?
#22 Lotharsson
March 17, 2010

On a recent The Drum thread (at the ABC) someone else posted McLean, Carter & de Freitas as proof that AGW was bunk – and that this was not coming from deniers because (IIRC) Carter had a PhD in some relevant field.

There was also a poster named John McLean who asserted that he had at least one peer-reviewed paper on climate change, and that he could personally assure us that the science was “crap”. Despite requests for references he did not provide any.