Sunday, April 29, 2012
From Schmidt 2005 to Miller 2012: the “not needed” excuse for omitted variable fraud
This is an updated version of my WUWT post with the same title. That original post injected a gratuitous and incorrect use of the term "null hypothesis." I noted the error there in the comments and, as I have done before, am posting a corrected version here on my own blog. The change is slight. I just stop using the incorrectly applied term of art and stick to the common-language meanings of hypothesis and competing hypothesis that I intended all along.
Solar-magnetic activity gets the Voldemart treatment: the competing hypothesis that must not be named
Miller et al. 2012 recently provided some pretty strong evidence for a solar driver of climate. "This is the first time anyone has clearly identified the specific onset of the cold times marking the start of the Little Ice Age," said lead author Gifford Miller in January. And the dates?
LIA summer cold and ice growth began abruptly between 1275 and 1300 AD, followed by a substantial intensification 1430-1455 AD.As you can see in the graphic above (from Usoskin 2003) these dates correspond pretty much with the midpoints of the Wolf and Spörer solar minima. (Usoskin 2007 centers Wolf at 1305 with a duration of 70 years and Spörer at 1470 with a duration of 160 years.)
Yet Miller never noted this coincidence. In fact, he tried to hide it, claiming that the onset of snow and ice growth coincided with periods of especially high volcanism (debunked both by Willis and by Wired), while dismissing the solar explanation with a misleading reference to the Maunder Minimum:
Our precisely dated records demonstrate that the expansion of ice caps after Medieval times was initiated by an abrupt and persistent snowline depression late in the 13th Century, and amplified in the mid 15th Century, coincident with episodes of repeated explosive volcanism centuries before the widely cited Maunder sunspot minimum (1645–1715 AD [Eddy, 1976]).This is a remarkably blatant deception, acceded to by 13 co-authors plus the reviewers and editors at Geophysical Research Letters. It takes no expertise to know about the Wolf and Spörer minima. There is no physics involved, so who do these people think they are fooling?
Nobody. They just don't think it is their job to make the case for what they regard as "the other side." The anti-CO2 alarmists are behaving like lawyers in an adversarial legal proceeding, hiding what hurts their own case while overstating what can be fashioned in support. In the courts an adversarial system is able to elicit a measure of truth only because there is a judge to maintain rules of evidence and procedure, and a hopefully unbiased jury examining the facts. These conditions do not obtain in science. The anti-CO2 alarmists are both the peer-review jury and the judge/editors, devolving into a pre-scientific ethic where acceptance is determined by power, not reason and evidence.
The lawyerly behavior of Miller et al. lead them to embrace a particular excuse for ignoring the evidence for a powerful solar driver of climate (even evidence that they themselves uncover). They don't "need" it. But they were not the pioneers of this anti-scientific ploy. That dishonor goes to Gavin Schmidt.
Miller's two hypotheses, specific and general
The specific hypothesis of Miller's paper is a feedback mechanism by which the cooling from volcanic episodes could get amplified into longer term cooling. It is "tested" via climate model. From Miller's University of Colorado press release:
The models showed sustained cooling from volcanoes would have sent some of the expanding Arctic sea ice down along the eastern coast of Greenland until it eventually melted in the North Atlantic. Since sea ice contains almost no salt, when it melted the surface water became less dense, preventing it from mixing with deeper North Atlantic water. This weakened heat transport back to the Arctic and creating a self-sustaining feedback system on the sea ice long after the effects of the volcanic aerosols subsided.But the real hypothesis of the paper, the one that expresses the authors' motivation, as revealed by blatant cover-up of their own evidence for a solar driver of climate, is more general. It appears in the last line of their abstract, which says that in order to explain the Little Ice Age, "large changes in solar irradiance are not required."
The timings Miller found point like a neon sign to a solar explanation but he is determinedly oblivious to that evidence. He is only interested in whether there could be some other possible explanation, and as long as he can come up with a non-solar explanation that is not absolutely falsified, he takes that as a rationale for ignoring the solar hypothesis and the evidence for it.
What exactly is the competing solar explanation that Miller seeks to avoid? According to Miller's wording, it is that the Little Ice Age was actually caused by "large changes in solar irradiance." But nobody thinks that there have been large changes in solar irradiance. There is broad agreement that while solar magnetic activity fluctuates dramatically, solar irradiance remains almost constant. Irradiance shifts towards the UV when solar activity is high, but the change in Total Solar Irradiance is too small to bring about much decadal or century scale variation in climate.
In contrast, there is a great deal of evidence for a solar-magnetic driver of climate (second section here), probably mediated either by solar-magnetic modulation of galactic cosmic radiation (as in Henrik Svensmark's GCR-cloud theory), or by the UV shift that accompanies solar-magnetic activity (Stephen Wilde's theory).
As a good adversarial lawyer, Miller is unwilling to betray any hint that these competing hypothesis are even a possibility. Like Voldemort, a solar explanation is the foe that "must not be named." Thus Miller refers to the possiblity of a solar-magnetic driver of climate indirectly and incorrectly as "large changes in solar irradiance."
Lawyerly advocacy is not science
In his role as an advocate, Miller's fear is fully justified. A strong solar-magnetic effect on climate would be a death knell for anti-CO2 alarmism. Any late 20th century warming that can be attributed to that era's continued high levels of solar activity reduces by the same amount the warming that can possibly be attributed to CO2, which in tern reduces the implied sensitivity of climate to CO2. Even worse, if solar-magnetic effects actually outweigh CO2 effects (my own surmise, by a wide margin) then the present danger is cooling, not warming, thanks to our now quiescent sun.
But lawyerly advocacy is not science. To only examine the evidence for non-solar explanations is to throw away information, violating the most basic scientific rationality, yet this is what the "consensus" has been doing for many years. My review of the first draft of the next IPCC report documents how "vast evidence for solar climate driver rates one oblique sentence in AR5." AR4 listed Total Solar Irradiance (TSI) as the only solar effect on climate, as did the Third Assessment Report (scroll to TSI). "The Team" simply omits what they wish to avoid.
Miller plays this game from the get go, where his opening literature review assumes that the only solar effect is TSI:
Episodes of anomalously cold summers primarily are attributed to some combination of reductions in solar irradiance, especially the LIA Maunder sunspot minimum [Eddy, 1976], explosive volcanism, and changes in the internal modes of variability in the ocean–atmosphere system [Crowley, 2000; Wanner et al., 2011]. However, the natural radiative forcings are either weak or, in the case of explosive volcanism, shortlived [Robock, 2000], thus requiring substantial internal feedback.Robock 2000 only addresses the volcanic issue, so Miller's only grounds for calling solar forcing a "weak" effect is his own analysis, where he only looks at TSI (using the TSI reconstruction of Schmidt et al. 2011).
That's a phony literature review. Miller's repeated deceptions—hiding the Wolf and Spörer minima, referring to large solar effects as "large changes in solar irradiance" when there is no such hypothesis—can only be to hide the possible role of solar magnetic activity, but this actual object of Miller's paper goes unmentioned in what is supposed to be a survey of the most relevant science. The literature review is a place where an adversarial approach is explicitly rejected by well established scientific standards, but the alarmists are not playing by the rules.
There is nothing wrong with Miller et al. testing their hypothesis that solar activity does not play a significant role (which they do by "setting solar radiation at a constant level in the climate models"). But when they pair this non-falsification of their pet theory with blatant misdirection about their own discovered evidence for the competing solar hypothesis, that is bad. It is using the "not needed" claim as an implicit justification for the omitted variable fraud that the entire consensus is engaged in.
Gavin Schmidt is a pioneer of the "not needed" excuse for ignoring possible solar magnetic effects
Some history on this particular ploy, for anyone who is interested. Miller and his co-authors are not the first to pull the "not needed" gambit. Eleven years ago Shindell, Schmidt, Mann, Rind and Waple published a paper in Science that is remarkably similar to Miller 2012. Like Miller, Schmidt and his co-authors propose a North Atlantic mechanism for amplifying cooling effects, though the mechanism itself is quite different. The spectral shift that accompanies decreased solar activity is hypothesized to alter atmospheric ozone composition in a cooling direction, setting in motion atmospheric flows ("planet waves") that in turn are hypothesized to drive the North Atlantic Oscillation. Their hypothesis is the same as Miller's: that they can account for the Little Ice Age without invoking any solar effects beyond the expected variations in solar irradiance, and their test is also the same: they run a model.
If the UV shift that goes with low solar activity can explain much of the Little Ice Age, couldn't the UV shift from high solar activity explain a similar amount of 20th century warming? No say Shindell et al. Changes in atmospheric composition from the pre-industrial to the industrial period supposedly cause the effect of the UV shift to reverse (p. 2151):
Ozone’s reversal from a positive (preindustrial) to a negative feedback supports results showing that solar forcing has been a relatively minor contributor to late 20th-century surface warming (7, 19, 31).Thus the CO2 explanation for recent warming is not undercut, enabling claims of future catastrophic warming to go forward. Of course Schmidt's references "showing that solar forcing has been a relatively minor contributor to late 20th-century surface warming" only look at TSI, and his "ozone reversal" is not an empirical finding but a model result. They've got all the doors manned.
In 2005, this 2001 paper became the centerpiece of a public exchange between Gavin Schmidt and science fiction author Jerry Pournelle. Schmidt was vigorously insisting on the scientific integrity of himself and everyone he knew:
None, not one, of the climate scientists I meet at conferences or workshops or that I correspond with fit the stereotype you paint of catastrophists making up worries to gain grant money. Personally, I don't think I've ever made a dramatic statement in papers, public speechs, grant applications or letters to the editor. Yet I still manage to keep my job and support a couple of graduate students. This is the same in every institution and university. ... I do not go around being a doom sayer - but it is incumbent on scientists to explain to people what it is we think we understand, and what it is we don't.To explain the sober science that causes him to dismiss the solar-warming hypothesis, Schmidt invoked his 2001 paper with Shindell et al.. It shows that there is no "need" to bring in any suspicious "new physics" (Schmidt's Voldemortian euphemism for the solar-magnetic hypothesis that must not be named):
I was a co-author of a paper in Science in 2001, that looked at whether climate models could replicate this pattern given the known physics of solar change. We found that two features were key, allowing the solar irradiance to vary more in the UV than in the visible (consistent with what is seen over the sunspot cycle), and allowing the ozone field to vary as a function of the UV and temperature in the stratosphere. With both of these effects, the model produced global cooling (as you would expect) but also a robust change to the circulation (a weakened NAO) that amplified the cooling in western Eurasia and over the mid-latitude continents. Obviously given the uncertainties in the forcing, the data that we were trying to match, and uncertainty in the model response, we can't use this a proof that we got all of it right. However, in the absence of better data, there is no obvious need for 'new' or unknown physics to explain what was going on. This was just a first cut, and better models and more data are being brought to bear on the problem, so the conclusion may change in the future. As of now though, this is still the current state of thinking."New or unknown physics" is obviously a reference to to GCR-cloud, which Schmidt finds lacking as a theory. Fine, but that is no excuse for ignoring the ever growing mountain of evidence that there is some mechanism by which solar activity is having a much larger effect on climate than can be explained by changes in solar irradiance. I compile some of that evidence in the second section here. For the state of that evidence in 2001, the Third Assessment Report's section 18.104.22.168 on "Cosmic rays and clouds" is well worth a look.
While AR4 and AR5 have progressively excised the evidence for solar activity as a powerful driver of global temperature, TAR actually began with several paragraphs of studies that found substantial correlations between solar activity and climate. Only then did it judge the proposed mechanisms that might account for these correlations to be too uncertain to include in their modeling.
That would be okay if they still took the discovered correlations into account in formulating their climate predictions, but of course they did not. This is the highly improper step that invalidates the IPCC's entire enterprise. They are using theory (in particular, their dissatisfaction with Svensmark's GCR-cloud theory) as an excuse to ignore the evidence that supports the theory, excising its known predictive power from their predictive scheme.
Evidence is supposed to trump theory, not vice versa. The IPCC is inverting the scientific method. It is literal, definitional, anti-science, and Schmidt's "no need" excuse is simply another justification for doing the same thing. Since there is no need to invoke a strong solar driver of climate, he is going to ignore the evidence for a strong solar driver of climate, and this is what Schmidt holds up to Pournelle as an example of his integrity: the very point where he justifies the omitted variable fraud that is being perpetrated by himself and his cohorts. If only he were capable of embarrassment.
Who is actually doing a suspicious new kind of physics?
Schmidt looks askance at GCR-cloud as "new physics," but it isn't new in any fundamental sense. The cloud micro-physics that Svensmark, Kirkby and others are looking at is presumed to follow established particle physics models. It is a new application of current physics. What Schmidt is really suggesting with his jaundiced eye is that we should be reluctant to extrapolate our current understanding of physical principles to illuminate the biggest scientific controversy of the day.
At the same time, he and Miller and the rest of the alarmists have introduced something that really is new and problematic. They are using model runs to test their hypotheses. They are using theory to test theory, with no empirical test needed. Here Miller describes how he "tests" his theory about ocean feedbacks (page 3 of 5):
Climate modeling reveals one such possible feedback mechanism. Following Zhong et al. , we tested whether abrupt LIA snowline depressions could be initiated by decadally paced explosive volcanism and maintained by subsequent sea-ice/ocean feedbacks. We completed a 550-year transient experiment (1150–1700 AD) using Community Climate System Model 3 [Collins et al., 2006] with interactive sea ice [Holland et al., 2006] at T42 x 1 resolution. Our transient simulation was branched off a 1000 AD control run, and forced solely by a reconstructed history of stratospheric volcanic aerosols and relatively weak solar irradiance changes (Figure 2b) [Gao et al., 2008].Models are not reality, and in the above case the model is known to be wrong. Total solar effects are presumed to be "weak"? That is what the alarmists all assume but it is not what the empirical evidence says, and while they may be able to tweak their models enough to keep them from being strictly falsified by the LIA, the last decade of no significant warming has them stumped completely.
General Circulation Models are the most elaborate hypotheses ever concocted. They involve thousands of questionable steps, iterated thousands of times. To illustrate, the Shindell-Schmidt paper is good enough to provide us with a highly abridged description of the hypothetical steps that their model works through. It gives some idea of the volume and sweep of what they are theorizing (p. 2150). (If you are actually going to read this, brace yourself):
Our previous studies have demonstrated how external forcings can excite the AO/ NAO in the GISS GCM (22, 25). Briefly, the mechanism works as follows, using a shift toward the high-index AO/NAO as an example: (i) tropical and subtropical SSTs warm, leading to (ii) a warmer tropical and subtropical upper troposphere via moist convective processes. This results in (iii) an increased latitudinal temperature gradient at around 100 to 200 mbar, because these pressures are in the stratosphere at higher latitudes, and so do not feel the surface warming (26). The temperature gradient leads to (iv) enhanced lower stratospheric westerly winds, which (v) refract upward-propagating tropospheric planetary waves equatorward. This causes (vi) increased angular momentum transport to high latitudes and enhanced tropospheric westerlies, and the associated temperature and pressure changes corresponding to a high AO/NAO index. Observations support a planetary wave modulation of the AO/NAO (27, 28), and zonal wind and planetary wave propagation changes over recent decades are well reproduced in the model (22).
Reduced irradiance during the Maunder Minimum causes a shift toward the lowindex AO/NAO state via this same mechanism. During December to February, the surface in the tropics and subtropics cools by 0.4° to 0.5°C because of reduced incoming radiation and the upper stratospheric ozone increase. Cooling in the tropical and subtropical upper troposphere is even more pronounced (;0.8°C) because of cloud feedbacks, including an ;0.5% decrease in high cloud cover induced by ozone through surface effects. A similar response was seen in simulations with a finer resolution version of the GISS GCM (14). This cooling substantially reduces the latitudinal temperature gradient in the tropopause region, decreasing the zonal wind there at ;40°N. Planetary waves coming up from the surface at mid-latitudes, which are especially abundant during winter, are then deflected toward the equator less than before (equatorward Eliassen-Palm flux is reduced by 0.41 m2/s2, 12° to 35°N, 300 to 100 mbar average), instead propagating up into the stratosphere (increased vertical flux of 6.3 3 1024 m2/s2, 35° to 60°N, 100 to 5 mbar average) (29). This increases the wavedriven stratospheric residual circulation, which warms the polar lower stratosphere (up to 1°C), providing a positive feedback by further weakening the latitudinal temperature gradient. The wave propagation changes imply a reduction in northward angular momentum transport, hence a slowing down of the middle- and high-latitude westerlies and a shift toward the low AO/NAO index. Because the oceans are relatively warm during the winter owing to their large heat capacity, the diminished flow creates a cold-land/ warm-ocean surface pattern (Fig. 1).That is a LOT of speculation. Normally it is all hidden. They just say, "we did a model run," but this is what it actually means: ten thousand questionable steps iterated a hundred thousand times. It is fine for people to be working on these models and trying to make progress with them, but to use them to make claims about what is actually happening in the world is insane, and using them as an excuse for ignoring actual empirical evidence is worse than insane.
This really is a new kind of science, and not one that stands up to scrutiny. We are being asked to turn our world upside down on the strength of the most elaborate speculations in the history of mankind, yet Schmidt thinks it is cloud microphysics—traditional science!—that should be eschewed. All to justify the destruction of the modern world, now well underway.
Hat tip to TallBloke.