Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Bluesky Facebook LinkedIn Mastodon MeWe

Twitter YouTube RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

New? Register here
Forgot your password?

Latest Posts


McLean, de Freitas and Carter rebutted... by McLean, de Freitas and Carter

Posted on 4 April 2010 by John Cook

Stephan Lewandowsky has written a thorough critique of the McLean 2010 paper on El Nino. Lead author John McLean subsequently published an illuminating reply. The key contention surrounding McLean's paper is the assertion that the El Nino Southern Oscillation is responsible for the long-term warming trend over the last few decades. How do they come to this conclusion? McLean clarifies how they prove that "there is no detectable sign of any global warming driven by carbon dioxide" - by eyeballing Figure 7 from their paper:

Figure 7 from McLean 2009: Seven-month shifted SOI with (a) weather balloon RATPAC-A temperature data 1958–1979 and satellite UAH temperature data (b) 1980–1995. Dark line indicates SOI and light line indicates lower tropospheric temperature. Periods of volcanic activity are indicated.

Foster 2010 explains how this figure splices together two separate data-sets: weather balloon data (RATPAC) to the end of 1979 followed by satellite data (MSU) since 1980. The splicing is obscured by the fact that the graph is split into different panels precisely at the splicing boundary. However, John McLean defends this graph by stating "The Y-axes are clearly labelled". In other words, the graph clearly states that box a) is RATPAC weather balloon data while boxes b) and c) are MSU satellite data.

What is not clearly shown in this graph and only discovered through analysis of the original data is that the mean values of the weather balloon and satellite data during their period of overlap differ by nearly 0.2°C. Splicing them together introduces an artificial 0.2°C temperature drop at the boundary between the two. In other words, they "hide the incline".

The crux of McLean's argument is that when you eyeball this graph split into several boxes, the temperature line fails to rise above the SOI line. Is there a plot of this data that isn't split into obscuring pieces? Is there a more rigorous analysis than a mere eyeballing of the data? To find such an analysis, you only need to go to an earlier section of McLean 2009. They start by plotting 12 month running means of the Southern Oscillation Index versus satellite temperature. This is before any long-term trends are removed. Eyeballing of the graph shows the temperature line clearly rising above the SOI line. Analysis of the data shows low correlation between the two data-sets. It's only by removing the long-term warming trend that they're able to establish a strong correlation between SOI and temperature.

Southern Oscillation Index (SOI) vs Global Tropospheric Temperature Anomaly (GTTA)
Figure 1 from McLean 2009: Twelve-month running means of SOI (dark line) and MSU GTTA (light line) for the period 1980 to 2006 with major periods of volcanic activity indicated.

An even longer time period is possible by comparing weather balloon data to the Southern Oscillation Index. This is done in Figure 4 from McLean 2009.

Weather balloon RATPAC global tropospheric temperature anomaly vs Southern Oscillation Index
Figure 4 from McLean 2009: Twelve-month running means of SOI (dark line) and RATPAC GTTA (light line) for the period 1960 to 2006 with major periods of volcanic activity indicated.

McLean argues that "If the SOI accounts for short-term variation then logically it also accounts for long-term variation". And yet despite McLean's attempt to hide the incline, his own analysis shows a strong divergence between temperature and the El Nino Southern Oscillation. Thus, the attempt to blame global warming on El Nino activity suffers from a serious divergence problem.

0 0

Printable Version  |  Link to this page


Comments 1 to 27:

  1. I love it! :D I plan to check for myself, but this is a pretty funny result after the media frenzy over 'hide the decline'.
    0 0
  2. The entire McLean rebuttal is a great read - especially the comment section.
    0 0
  3. It took me 3 or 4 reading of McLean to realise it is a bit of a dogs dinner. It does seem that the initial removal of the trend to indentify the 7 month lag is valid but they then go on to claim far too much without any valid analysis (i.e. by eyeballing). I take the point about figure 7 being confusing but if you do object to the splicing of two data sets then you can just ignore a) and look at b) and c). After all 1979 onwards is the most important period for global warming to show itself in the data. So just 'eyeballing' b) and c) of figure 7 suggest there isn't a great deal of separation between SOI and global temp. There was one extra question I wanted answering. I had noticed that Figure 1 clearly showed temperature rising above SOI in later decades as mentioned in the article. This is less obvious in Fig 7 b) and c), which would be the equivalent set of data. Now figure 1 is the 12 month running average for temp while Figure 7 is a plot of monthly averages. Does the smoothing of the temperature data do anything to cause the SOI and temp to diverge?
    0 0
  4. Tim Lambert at Deltoid has plotted satellite (UAH) measured temperatures before and after using McLean et al.'s 12-month differencing filter.
    0 0
  5. Deep Climate also has graphs with and without differencing, and has a ton of other info relevant to the fundamental poverty of McLean et al.'s original paper. There are more links to relevant material in the Links section here at Skeptical Science, in these sections: McLean's reply to Foster 2010 was censored and It's El Nino.
    0 0
  6. There is an hypothesis in the website "Climate Observations" by Bob Tisdale that blames ENSO for the warming after the 1970s. "Can El Nino Events Explain All of the Global Warming Since 1976? – Part 1" Bob Tisdale believes that big El Niños cause step changes in global temperatures, specially sea surface temperatures. In short, the La Niñas reduce the cloud cover above the tropical pacific, so the Pacific Warm Pool accumulates heat (interestingly, it is there were the most rapid sea level rise happens, at a rate of more than 1 cm/year !). Then when a big El Niño happens, the heat is released and transported to other ocean basins. When the next La Niña occur, part of this warmth persist(apparently, the effects of La Niña are less global than El Niño ones). It seems to me to best skeptical argument I have ever read. It fails, however, to explain the steady warming of deep oceans as shown in "Global hydrographic variability patterns during 2003-2008" by K. von Schuckmann. What do you think?
    0 0
  7. Nice work. What was that about hiding the decline? Looks like even this half-hearted attempt by the denial movement to produce some actual science is a bust.
    0 0
  8. #3 Humanity Rules, you have to look a bit closer at Figure 7. I just did and noticed a couple of non-obvious things. First, not only are the y-axes on the right and left sides different, they are also offset just a bit. The zero on the right side is a bit lower than on the left, so if you just eyeball the relationship of GTTA to the zero, it gives a misleading sense that the GTTA has about the same value at the end of each panel as it does at the start -- in fact, in both cases the GTTA is higher by 0.1-0.15 C at the end of each panel relative to the beginning. Add the incline of each panel and the 0.2 C offset between (a) and (b) and you pretty much have the temperature increase of the last 50 years. Second, you can see that in the middle panel GTTA is mostly below SOI, and in the bottom panel it is mostly above. And if you rescaled the GTTA axis so that the GTTA curve spanned about the same fraction of the total plot range as the SOI curve does for its axis, the trend of the GTTA relative to SOI would be more obvious. They did the same y-axis label shifting in Figure 1 as well. "Hide the incline" indeed!
    0 0
  9. From Peru said: "What do you think?" I think it fails to explain the steady warming of deep oceans, as would be expected of a global warming scenario. It does posit a mechanism by which that warming could puff off heat to the atmosphere. To the point: how do we know ENSO isn't deepened or made more frequent by AGW? The oceans hold heat 1000 times more densely than the atmosphere. We argue about just what the atmosphere is telling us about AGW, shouldn't we be arguing about the oceans at a rate 1000 times this? Proving that AGW, as measured by atmospheric temperatures, is caused by ENSO, proves nothing. The oceans are where the judgement of AGW is actually written, and those oceans are warming.
    0 0
  10. Re: Jeff Freymueller at #8: It seems to me that Jeffs comments about the axis being shifted is correct. I presumed that the zero on each side lined up when I looked at the graph. It was very difficult to see that they did not line up. To McLean et al: is it normal to have this type of axis shift?
    0 0
  11. I never thought much of filtering out long term trends to show short term trends are making short term trends, as in the Mclean et al paper. The long term trends remain relevant and must be explained. But one hide the decline (divergence) doesn't validate another hide the divergence (Mann et al 1998). The long term trend remains, and it doesnt follow either Mann's tree rings or short term climate variations.
    0 0
    Response: "one hide the decline (divergence) doesn't validate another hide the divergence (Mann et al 1998)."

    Of course it doesn't, noone is suggesting "McLean hid the decline therefore it's okay for Mann to hide the decline". In the case of the tree-ring divergence problem, what is required is a proper understanding of the tree-ring divergence problem. Contrary to what is commonly reported, it's not a divergence or decline in temperature - but a decline in tree-ring growth due to some factor other than temperature. This is an issue that is openly discussed in many peer-review papers.

    In contrast, what we're talking about here is more straightforward - the direct comparison of temperature and SOI. But the comparison is being obscured by splicing of data, of obscuring the splice by dividing the graph into multiple boxes, of using different Y-axis ranges in the different boxes and a curious lack of divergence in Figure 7 that is not apparent in Figures 1 and 4. The way McLean and others talk up Figure 7, it's the smoking gun that disproves anthropogenic global warming. With all the question marks over this graph, it's more a smoking gun firing blanks.
  12. #8 jeff, If we were concerned about absolute values I would take your point about the shift in the axis but what matters here is trends and I still don't understand why the trend difference is more obvious in figure 1 than it is in figure 7. If the difference existed in figure 7 it would be shifted to one end of the graph (1980 or 2000) because of the axis shift, but it isn't.
    0 0
    Response: The shift in axis is significant because Figure 7a uses the same SOI Y-axis as Figures 7b and 7c. And yet Figure 7a uses a different temperature Y-axis from Figures 7b and 7c. They're shifting the goal posts, they're splicing the data sets then trumpeting this graph as disproving the large body of empirical evidence that more CO2 is causing warming.
  13. OK, this may be a naive question, but if the SOI is a function of air pressure differences, and El Nino/La Nina is water temperature, well, aren't both of those dependent to some extent on the amount, or pattern of distribution, of energy in the system? If so, changes in them are driven by changes in energy, or energy flux. I think it would be surprising if there were not feedbacks or interdependencies involved, but I'm wondering if there is some confusion here between cause and effect. Also, did I understand this correctly that they filtered out the long term trend, and then state that they detect no long term trend (global temperature rise)?
    0 0
  14. #10 michael sweet, I think it is not normal. I would never do such a thing intentionally in my own work, and I would be embarrassed if I did it accidentally. Having the zero point on the y-axis be different on both sides, yet so close that a quick look makes it seem the same? The effect is deceptive, and it is the authors' responsibility NOT to be deceptive. #11 comment, actually I think the smoking gun wasn't firing blanks, but it ended up pointed back at the shooter... #12 Humanity Rules, if you put the three panels together, added back the 0.2 degrees offset in GTTA between panels a and b, then I think you would see the trend more clearly. Also note that the x-axis is expanded in panels b and c relative to panel a, which makes the trend appear smaller. In any case, as I mentioned above, the effect of putting the zero label in slightly different places on the right and left y-axes is visually deceptive. The bottom line is that Figure 1 is a straightforward plot of two quantities, and they did several things differently in Figure 7, all of which had the effect of obscuring the trend in GTTA, which is obvious in Figure 1. You might ask yourself why they did that, or why they removed the trend for their statistical correlation while claiming they were just removing noise? I've drawn my own conclusion.
    0 0
  15. #13 Chris G, Foster et al in their comment pointed out that the filtering McLean et al. applied to "reduce noise" mainly filtered out the long-term trend. Foster et al. are correct. McLean et al. found a strong correlation between the filtered SOI and filtered GTTA (which is fine, and non-controversial). Then they made a series of statements that were true for the filtered time series, but they stated them as if they were true for the original series (they are not true for the original series). I think you can conclude for yourself from Figure 1 whether SOI explains the trend in GTTA -- it explains a lot of short-term variability but clearly SOI does not have the trend that GTTA has.
    0 0
  16. From Peru, it doesn't matter what the proposed mechanism is for ENSO's effect on global temperature. The actual, empirical, observed data show that ENSO can explain only short-term variations in temperature, not the long-term increase. That's the point of this post, and of the post It's El Nino. Click on the link to Deep Climate that I provided in my comment #5 above, and look at the graphs there.
    0 0
  17. Given that the intensity or magnitude of any ENSO events in the Pacific Ocean are subject to influence of the IOD in the Indian Ocean, any theorising that fails to incorporate that and other regional influences, that combined, represent the entire global climate, is unlikely to come up with all the answers. I realise Australia is not the world, but when scientists were putting too much emphasis on ENSO in trying to predict our weather their success rate was lucky to be as good as tossing a coin, perhaps worse. Now as greater understanding of the IOD begins to filter in things are starting to make more sense.
    0 0
  18. Being vain, I seldom hesitate to repeat myslelf: (this also posted at Deltoid). I’ve offered this up before elsewhere. Here are links to submissions made by McLean to a New Zealand Parliament Select Committee that was considering an Emissions Trading Scheme. Three Select Committees have had a go. For the first round in 2008 McLean co-submitted with someone who in a previous life called for a boycott of Mobil because of climate change issues and is now a denialist. For the second round in early 2009 McLean went solo: And nothing for the 3rd round in late 2009.
    0 0
  19. Re MCdF09 Figure 7, the jump from panel (a) to panel (b) is a real problem. I don't think the offset in left vs right Y axes for panels (b) and (c) is a problem (I disagree with Jeff F. about this). The bigger issue is that MCdF claimed that SOI explained 81% of the variability in the MSU record, but that was based on the detrended data. That was their ONLY quantitative comparison of the two, and it was fundamentally wrong. Figure 7 is just for "eyeballing"; there's no quantitative analysis included. In his recent comments McLean keeps trying to shift the discussion to Figure 7. It's important to note the problem with Figure 7 (panel a vs. b/c) but it's also important not to let him shift the discussion away from his erroneous quantitative claims.
    0 0
  20. McLean: "If the SOI accounts for short-term variation then logically it also accounts for long-term variation." That's so classic! I can't believe he actually said that. ;-) On blaming ENSO for global warming... where does the heat come from? No matter how much handwaving is employed, you can't ignore thermodynamics.
    0 0
  21. Sean @20. It depends on what short and long term refer to. There is no doubt that the SOI does help form longer term variations seen as the multi decadal IPO or PDO which cycle over about 6 or 7 decades, and these have been identified back centuries. Reconstructions of El-Ninos (Quinn El-Ninos 1527-1987) indicate that the 1500's and 1800's were times of more frequent El-Ninos than more recent times.
    0 0
  22. johnd, McLean's claim is different. He says that from the short term SOI variation logically follows the long term trend, which is quite obviously not the case.
    0 0
  23. Heres the thing I fail to understand though. ENSO is hardly a *new* phenomenon-indeed humans have been *aware* of it for at least 250 years. Yet prior to the last 30 years we're supposed to believe it had *no* impact on long-term climate, but is now suddenly the cause of global warming. Excuse me if I'm a *little* incredulous, but isn't it just as likely that global warming is driving changes to ENSO, rather than the other way around?
    0 0
  24. #9 ubrew12 at 08:14 AM on 5 April, 2010 "The oceans are where the judgement of AGW is actually written, and those oceans are warming" The first part of the sentence is true. The second one is not quite so.
    0 0
  25. Berényi Péter at #24. If the oceans are "not quite" warming, what then accounts for the consistent rise in level, if not thermal expansion? Note that the rise occurs in the face of dam inpoundment which, according to BF Chao, removes more than 0.5mm/yr additional rise, from river inflows around the planet.
    0 0
  26. The splicing across more than one graph and the offset in the y-axes points to a deeper problem. Why not just offset the GTTA axis 6 inches down the page? Or expand the heck out of the scale. Since two axes refer to different metrics with no clear QUANTITATIVE relationship you're free to represent them however you want. Therefore MacLean's inference drawn from the observation that the GTTA never rises above the SOI doesn't mean anything.
    0 0
  27. @ #20 Sean A: In the article on the ABC Unleashed website, McLean dropped a couple of hints that suggest he might be looking to solar wind anomalies over the Antarctic to explain ENSO. He made a comment about a paper that I believe is this one (although he got the name of the lead author wrong): However he won't be able to explain global warming by referring to solar winds.
    0 0

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

The Consensus Project Website


(free to republish)

© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us