For my sins, I decided to listen to Murr
ay “I have a theory” Salby talking about his ideas about why the recent rise in CO2 isn’t human-caused (note that isn’t his most recent UK tour; that’s back in April). By all means read my notes below if you’re interested in the various ways that he is wrong; but if you’re interested in how we know the increase really is human-caused, then try RealClimate from 2004, a somewhat pithier response from me, point 5, in 2005, or the ever-popular Skeptical Science version; and Eric Wolff is excellent. Or, if you belong to the Dark Side, then perhaps wattsupwiththat.com/2010/06/07/some-people-claim-that-theres-a-human-to-blame/ (note! nofollow to make VV happy :-) or wattsupwiththat.com/2010/08/05/why-the-co2-increase-is-man-made-part-1/ will help.
Before I go on, I can’t resist saying that all of this is very very silly; I’m playing along in this post but I won’t forever. There are interesting areas to explore in climate science and palaeoclimatology, but “is the CO2 rise human-caused” isn’t one of those areas. Its settled, done, and nailed down. If you don’t know the science, or have read some words but found them too hard to understand, then you may fairly claim to not know for sure either way. In which case, you’re going to have to believe some authority; but you’re definitely not making any interesting contribution to the debate, because all you’re saying is “I haven’t been able to look”. But if you claim to have investigated and end up believing that the rise isn’t human-caused, then you’re lost, wandering in the wilderness, and I doubt I can help bring you back.
It often seems to me that the extreme fringes of “skepticism” are showing their fear: although they profess to believe in science, and the implication that the long chain of science needed to make “…and GW is a problem that needs addressing” needs to be true in all its aspects, nonetheless they are afraid to allow even the most basic and obvious points of that chain to be accepted. I’m not sure why; its clearly not logically necessary.
This post is culled from Presentation Prof. Murry Salby in Hamburg on 18 April 2013. I haven’t included all his charts by any means. Note that is isn’t correct to speak of him as a prof in the present tense, as he has been de-proffed. The PR for his recent UK tour makes this elementary error.
Here’s the first chart of interest. Its some kind of 50kyr-filtered version of the std.Vostok ice core temperature and CO2 record. It shows that the two are correlated; this isn’t controversial.
Salby provides no key to the literature here; its impossible to tell from his talk whether he knows this is old stuff, or not; its impossible to tell whether his audience knows.
One thing worth noting is that he makes no quibbles about the quality of the record at this point: this figure, and the little he says about it (he speaks very slowly, as though he needed to spin out his words), would make no sense if you didn’t trust the CO2 record. Later on, he does decide not to believe the CO2 record; that makes most of his discussion of this figure dubious, in his terms. Of course, since he is wrong about the problems with the CO2 record, what he says about this figure here is, in fact, true.
He does note that CO2 is a proxy, but he’s wrong about that. A proxy is something standing for something else: like the length of a column of mercury as a proxy for temperature (if you buy a decent thermometer and expose it carefully then it will be a very good proxy for local temperature, but its still a proxy). By contrast, the CO2 in the bubbles in an ice core isn’t a proxy for ancient air, it actually is ancient air (it turns out later tha he doesn’t really believe this, but he’s wrong; see later).
Next up is the correlation between CO2 and temperature from the ice cores. If you’re interested in this, I’d recommend Eric Wolff’s words and/or more of mine. Salby (incorrectly) says the correlation is highest at small positive lag; as you see from his picture, its highest at zero lag. However, he draws no conclusions from his statement.
Then a coherence (then phase) spectrum, which shows that CO2 and T are well correlated on timescales longer than ~10 kyr. Again, there’s no source or ref to the literature. Phase is described as “hovers near zero” (he does get some Brownie points for not going on about the silly leads-by-800-y stuff).
At 8:45 he repeats the assertion that CO2 is a proxy, and says that we need to understand how in-ice CO2 and atmospheric CO2 are related.
Then shows the “observed” atmos CO2 and “global temperature” from 1960-2010,
(note in passing here that this is a CO2 record showing a strong seasonal cycle. Note also that (as you expect) the change is negative during part of each year) and then its correlation:
Unlike before, he provides no significance levels. He asserts that this is significant, but without numbers that’s dubious. No-one in the audience reacts. Indeed, at no point does anyone in the audience react to anything, even to Salby’s ponderous little “jokes”. If there are questions afterwards, they’ve been cut out of the record. Nor does he tell you which CO2 record he is using. If you look at, say, the bottom pic of http://instaar.colorado.edu/sil/research/research_detail.php?research_project_ID=1 you’ll see that’s rather important: there’s a strong seasonal cycle in CO2 in the NH, and there’s a seasonal cycle in the global temperature record, so if you just correlate them you’ll see that; and the lag-lead relationship will tell you nothing interesting about causation (its odd how good people can be at changing “correlation does not prove causation” when it suits them, and then forgetting it when it doesn’t).
Salby draws some kind of conclusion from this, but its a vague one; but he clearly likes that CO2 lags temperature. Onwards, to “net emission” of CO2 vs temperature anomalies:
But notice something odd here: the line is smooth, but always positive. So it isn’t the slope of the CO2 record he presented earlier. Its been smoothed in some way. So has the temperature. But in what way? He doesn’t say. Indeed he doesn’t say its been smoothed at all. Note also that the first, and last, 5 years have been cut off the record (again, not remarked on). Is this some kind of 10-y wavelet filtering? 2 year filtering (it sort-of looks like it, but if it was 2-year, why would that lose 5 years at start and finish)? Salby’s results depend very heavily on whatever he’s done here, so he needs to tell us what it is.
Salby then asserts that this demonstrates that d(R_CO2)/dt = (lambda)(T – T_0) (I’m using R_CO2 for the reservoir of CO2 in the atmosphere; changes in this are the net surface fluxes). As I’ve noted above, I don’t think that’s valid.
Here’s an alternative explanation which could even be true, even granting him his wiggle-matching: short-term (annual) fluctuations in CO2 are driven by short-term fluctuations in temperature (oceanic outgassing, perhaps, though I don’t know; you’d need to poke around in Henry’s law to know; Wotts does this a bit; see-also this from SKS especially its fig. 2) and this is superimposed on a long-term trend of increasing CO2 from fossil fuel burning, and a long-term trend of increasing temperatures from the greenhouse effect.
Salby is talking to his audience at a very low level – he pauses to explain what an “integral” is. Indeed his whole manner is desperately portentous. Its also very fake – there is no way that anyone who needs to have “integral” emphasised is going to follow the subsequent manipulations and Fourier transforms.
At this point he is asserting that the CO2 and T “evolve coherently” in both the “observed” (1960-2010) and the “proxy” (by which he means ice core) records. But differently in the two. Cute a closer look at the ice core CO2 record. There’s a pile of equations thrown at us here – cue Salby saying “Are you ready? take a deep breath” – which amounts to explaining why he thinks CO2 should be in-phase with T over long timescales and out-of-phase over short timescales. I don’t have objections to that, so will skip over it.
There’s some more incomprehensible equations, at the end of which Salby convinces himself that the ice-core record CO2 underestimates the atmospheric changes by a factor that increases with timescale culminating with the assertion (I kid you not) that on the 100 kyr scale, atmospheric CO2 changed by a factor of 10 more than the ice core; i.e., by approximately 1000 ppmv. And therefore in his view the 20th C changes aren’t unprecedented. If you believe the std.preindustrial holocene values of 280 ppmv, then this leads to a massive negative amount of CO2 in the atmosphere, which is obviously impossible; so I presume he is forced to believe in ~1000 ppmv CO2 during previous interglacials, and perhaps early parts of the current interglacial. This, if he believed it, would be a massive challenge to current theory. Its the sort of thing that if you believed it, you’d dwell on the details, tease out the implications, try to reconcile it with the ice core records (later on he accepts the CO2 record since 1830 (why 1830? He doesn’t say), which includes ice cores, but makes no effort to reconcile that with the view he expresses here, that the cores are wrong), whatever. But he does none of these things; he just moves on.
[Note that this need to believe in implausible glacial-interglacial swings has been noted before; see John N-G quoted in various places, e.g. Wotts, though I actually came across it in JM’s preprint. But note that, at least in Salby-world, this isn’t a fatal flaw; its a part of his theory. As so often with these descent-into-madness things, simply pointing out the obvious flaw isn’t enough, because they say “aha! I’ve thought of that!” and you have to go another step down.]
Ah, I think I know how he has got here: he has assumed non-conservative processes in the ice (but doesn’t say if that’s removal or addition). Since he’s done this with no observations or theory to justify this, its all unjustified. There’s a whole wide literature of how CO2 behaves in ice cores (the answer, broadly, is that its conserved in southern hemisphere cores. There is some interesting work there, and there were some early problems with the Greenland cores, but all of this is absorbed and explained and understood by the std. literature, which Salby ignores) and he refers to none of it. So his equations and graphs become meaningless. This is like the antient Greeks theorising about epicycles but not bothering to measure the planets orbits.
He then goes on to try to deal with diffusion in ice. And ends up drawing a graph. But he does all this without determining the diffusion coefficient. This is impossible, so he must have just made one up. This is impermissible.
He presents nothing formal as a proof that this is correct, only a picture of the T-CO2 cross-correlations predicted by his theory:
They look a bit similar. Is that good enough? No, not even close. Because “diffusion” or analogues are such omni-present processes. You get the same broadening of a peak from anything: measurement error, “random” fluctuations caused by “other events”; whatever (note that its also weird that his “spike” of without-diffusion is not just broadened by also amplified by his diffusion; that makes no sense).
Note that in all of this he is assuming the temperature record (which unlike the CO2 really is a proxy) is accurate; he never mentions this assumption.
Then he tries to address the “C13” problem; but that was more than 30 minutes in and I was losing the will to live. I might go through that some other time. After (45 mins in!) that we’re onto CH4. Then what Co2 would look like if it followed his theory backwards; but he is careful to stop at 1880. How he thinks he can reconcile the essentially-monotonic CO2 with the clearly-not-monotonic temperature series I really don’t know. 54 mins: we’re into the global energy budget, some silly stuff with climate models which doesn’t seem to be relevant. Then there’s the obligatory reference to Feynman (poor chap, he has become the new Galileo) but without reading the important bit, “how not to fool yourself”. Its always the other people who are fooling themselves.
Don’t read me, read Eric
If all of this is too much, you’re right. Wading through dis/mis-information is more painful that just reading the right answer. In this case, Eric Wolff’s the main evidence that the ice core record of CO2 is a good representation of the past atmospheric concentration is perhaps the best technical reference.
TL;DR: what’s wrong with Salby?
He does a lot of “theoretical calculations” but at no point does he point out that those calculations can’t be done without assuming values for some basic parameters (CO2 diffusion in ice, for example; or the non-conservation of CO2 in ice) and that his values for those parameters are wildly at variance with the ones anyone else would use.
He doesn’t engage at all with existing literature, or indeed the bleedin’ obvious: we’ve emitted all that CO2: where does he think its gone?
He won’t write any of this down.
Something I thought at the time: I’m doing Salby too much honour by even bothering to read his stuff. But I’m also doing him more honour than any of the Watties and so on: for whilst many of them fawn over his conclusions, none of them can be bothered to read or understand any of his words(some of them know he’s wrong, of course, but AFAIK none have bothered analyse why, for example, his graphs derived from diffusion in ice are drivel). See for example this cri-de-coeur:
Note: In all the many times (and some of you realize that it has been, indeed, many,) I have posted my hero, Dr. Salby’s, lecture on this site, NOT ONCE HAS A SCIENTIST OF WUWT given us his or her detailed comments on the complete content of that lecture. While I have taken notes from it and could post a detailed summary of the video, I have nothing to add. Has all my posting of Dr. Salby’s lecture been for nought? Has NO ONE watched his lecture? Why–in–the–world haven’t you?
In fact the cri is wrong: just a little higher in the thread WE (and I think he counts as a “scientist of WUWT”, snigger) has told her that its all a pile of donkey’s dildoes. But he’s done it in honeyed words (I don’t think WUWT regulars are allowed to diss the potty peer yet) so she can’t read it.
* Emissions and Concentrations–How Closely Are They Correlated? – The Lukewarmer’s Way
* W h y d o e s a t m o s p h e r i c C O 2 r i s e ? Jan Schloerer, Version 3.1, October 1996. Via rmg.