Error Theory

(Crossposted at WUWT)

Technically Dr. Muscheler is asking me to retract the title of my post, "Raimund Muscheler says that a steady high level of forcing can’t cause warming":

I am sure that you are aware of the fact that the title is wrong. I never said that steady high levels of forcing can't cause warming.

He most certainly did. Here is the sentence of Muscheler's that I was paraphrasing (with emphasis added):

Solar activity & cosmic rays were relatively constant (high solar activity, strong shielding and low cosmic rays) in the second part of the 20th century and, therefore, it is unlikely that solar activity (whatever process) was involved in causing the warming since 1970.

This is an unconditional statement: the high solar activity of the second half of the century can't have caused warming because it was "relatively constant." If Dr. Muscheler wants a retraction he's going to have to issue it himself, and that actually seems to be what is going on here.

Raimund now rejects the claim that a steady high level of forcing can't cause warming. Good. But then on what grounds can he dismiss a solar explanation for late 20th century warming?

His email offers a new rationale. Muscheler thinks the lack of warming from the 40s to the 70s vitiates the solar-warming hypothesis:

My point is rather nicely illustrated in the attached figure [at top of post]. It shows the sunspot data and temperature anomalies over the last 160 years (annual data and 11-yr average). It shows the high solar activity I was mentioning.

According to your reasoning one would expect a steady warming since 1950. However, the temperatures were rather constant from 1940 to 1970. Furthermore, the temperature and solar trends are opposite during the last 30 years. So I think one would have to invoke a very strange climate delay effect in order to explain the recent warming with solar forcing.

I would be happy if you could correct the title and add this clarification to your post.

Best wishes,

Raimund

There are a couple of points one can quibble with here:

1) Muscheler again invokes "opposite trends," as if it is the trend in solar activity, not the level, that would be driving temperature.

2) Those trends have not been "opposite for 30 years." Solar cycle 22, which ran from 1986-1996, had the same sunspot numbers as cycle 21 but was more intense by pretty much every other measure.

3) Muscheler seems to be asserting that temperature has been rising for the last 30 years when it has been roughly flat for the past 15 years (a fact that presents problems for Muscheler's preferred CO2-warming theory, but is perfectly compatible with the solar-warming theory, after cycle 23 slowed down and dropped off a cliff).

But set those quibbles aside. What is interesting here is Muscheler's new argument that if the sun had caused late 20th century warming then the planet should have warmed steadily since 1950.

My reply:

Actually, I would say that warming should have been steady since the 1920's, but that is only if we are looking at the heat content of the oceans (where almost the entire heat content of the climatosphere resides). Unfortunately, we don't have good ocean heat content data for this period, while the data we do have--global mean atmospheric surface temperature--is dominated by ocean oscillations.

You suggest that it would take some very strange lags for warming from the 40s to the 70s to not show up until later, but would this actually be strange? Doesn't it fit with what we KNOW: that the cool 40s-70's period coincided with a cool-phase of the Pacific Decadal Oscillation?

Ocean oscillations are widely acknowledged to be the dominant short term driver of global temperature

Just look at what the CO2 alarmists say as soon as their predicted warming fails to show up (April 2008):

Parts of North America and Europe may cool naturally over the next decade, as shifting ocean currents temporarily blunt the global-warming effect caused by mankind, Germany's Leibniz Institute of Marine Sciences said. ...

"Those natural climate variations could be stronger than the global-warming trend over the next 10-year period,'' Wood said in an interview. "Without knowing that, you might erroneously think there's no global warming going on.''

The Leibniz study, co-written by Noel Keenlyside, a research scientist at the institute, will be published in the May 1 issue of the journal Nature.

"If we don't experience warming over the next 10 years, it doesn't mean that greenhouse-gas warming is not with us,'' Keenlyside said in an interview. "There can be natural fluctuations that may mask climate change in the short term.''

Wood and Keenlyside aren't even talking about the PDO, just the measly AMO. For a historical example where natural fluctuations probably really did "mask climate change in the short term," the PDO is the place to look.

Here is a comparison of JISAO's PDO index (red) with the HadCRUT3 temperature record (black):


If ocean oscillations are as powerful a climate driver as the anti-CO2 alarmists claim then this graph suggests a simple story: that cold Pacific surface waters swallowed up a big gulp of warmth from 1940-1970, which the PDO then belched back up during its warm-phase in the 80s and 90s. Without the PDO there might well not have been a 40s-70s temperature dip, making warming over the 20th century much more even.

Is the PDO really this influential, or is it largely coincidence that the PDO was in a cool phase when GMAST dipped a couple of tenths between 1940 and 1970? Without good heat content data it is very hard to gauge but logically there is no upper bound on how powerful an effect ocean oscillations can have.

As Jo Nova describes meteorologist William Kininmonth's "deep cold abyss," the ocean depths form a great pool of "stored coldness" which is "periodically unleashed on the surface temperatures," a slumbering dragon that with a flick of its tail can grab away large amounts of surface warmth. Thus we certainly can't rule out that on time scales of up to decades GMAST really is dominated by ocean oscillations.


The CO2-warming theory needs to invoke ocean oscillations more than the solar theory does

Both have the same difficulty with the 40s-70s dip in temperature. For either theory to work the mid-20th century cooling pretty much has to be explained by ocean oscillations, but the CO2 theory now has to rely on the short-term dominance of ocean oscillations to explain the lack of recent warming as well.

That's the point of Trenberth's "missing heat," right? By his calculations there must be lots of CO2-driven heat accumulating in the oceans. Set aside whether the real problem is with Trenberth's measurements and calculations, the solar theory has no difficulty explaining why temperatures would be leveling off. With the sun having gone quiet this is the maximum likelihood solar projection (with cooling predicted to follow). It is the special case where GMAST actually tracks ocean heat content. Differences between GMAST and ocean heat are to be expected, but it is the CO2 theory that now needs to invoke that likely divergence from maximum liklihood.

Obviously it is not tenable to reject the ocean oscillation argument when applied to the solar theory but accept in when applied to CO2, but this is what the consensus scientists are effectively doing.


Gavin Schmidt on the asymptotic approach to equilibrium

Dr. Muscheler and I almost got to the ocean oscillations question way back in 2005, but Gavin Schmidt grabbed the hand-off. Raimund had claimed in a RealClimate post that a steady high level of forcing can't cause warming:

Regardless of any discussion about solar irradiance in past centuries, the sunspot record and neutron monitor data (which can be compared with radionuclide records) show that solar activity has not increased since the 1950s and is therefore unlikely to be able to explain the recent warming.

I objected in the comments that:

What matters is not the trend in solar activity but the level. It does not have to KEEP going up to be a possible cause of warming. It just has to be high, and it has been since the forties.

Presumably you are looking at the modest drop in temperature in the fifties and sixties as inconsistent with a simple solar warming explanation, but it doesn’t have to be simple. Earth has heat sinks that could lead to measured effects being delayed.

Gavin Schmidt's response was similar to Muscheler's today, but Schmidt was explicit about what the process of equilibration should look like:

Response: You are correct in that you would expect a lag, however, the response to an increase to a steady level of forcing is a lagged increase in temperature and then a asymptotic relaxation to the eventual equilibrium. This is not what is seen. In fact, the rate of temperature increase is rising, and that is only compatible with a continuing increase in the forcing, i.e. from greenhouse gases. - gavin

Like Muscheler, Schmidt ignores the PDO. It is ocean heat content that should undergo an "asymptotic relaxation to the eventual equilibrium," but all Schneider has to go by is GMAST, so he is implicity assuming that ocean heat content is faithfully tracked by GMAST, regardless of the fact that this relationship can be profoundly obscured by ocean oscillations.

We know that GMAST underwent a substantial mid-century gyration where 20th century warming actually reversed for a couple of decades before accelerating upwards again but we do not know that ocean heat content underwent any similar gyration. Schmidt assumes it did but the PDO record suggests that it likely did not, in which case Schmidt's argument that late-century warming must have been caused by CO2 collapses.


The problem is the hidden nature of these ocean-equilibration assumptions

If Schmidt and Muscheler want to dismiss a solar explanation for late 20th century warming by invoking the highly speculative assumption that GMAST is a good proxy for ocean heat content over with the 20th century, that fine. As long as this assumption is made explicit then others can evaluate it and toss any following conclusions in the trash. The problem is that the consensus scientists are not telling the public their real grounds for dismissing a solar explanation.

The consensus position, re-iterated over and over again, is a simple unqualified statement that because solar activity was not going up over the second half of the 20th century it cannot have caused warming over this period (or is unlikely to have caused warming over this period). I have collected a dozen such statements from scientific papers, news articles, and most recently from the First Order Draft of AR5.

Only when I have pressed these scientists on the irrationality of their claim that a steady high level of forcing can't cause warming do they start hinting towards the highly speculative arguments about ocean equilibration that are the actual basis for their dismissal of the solar hypothesis. Reliance on such hidden assumptions is not science, so job one is to get these unstated assumptions out in the open where they can properly evaluated. Not surprisingly, unscrutinized assumptions do not stand up well to scrutiny, so job two is knocking 'em down.

The rapid equilibrium assumptions of Lockwood and Solanki: knocked down. The implicit assumption by Muscheler and Schmidt that GMAST should track ocean heat content with no major divergence, now knocked down as well. It is a weak argument at best, requiring strong claims about matters of vast uncertainty, wrecking any pretension to have ruled out a solar driver for late 20th century warming.

Until these hidden assumptions are stated I suggest that we all take at face value the positions that these scientists actually assert. When they say that because a high level of forcing was relatively constant it is unlikely to have caused warming, we should say that they think you can't heat a pot of water by turning the flame to maximum and leaving it there, because that is exactly what they are saying.

Then when they come back with their "what I really meant was," we can expose their real thinking for the unexamined nonsense it is.


Update 1: Tisdale points out that the PDO index does not record a net temperature anomaly, only the phase of a temperature pattern

What I should have used to represent anomalously warm or cold ocean surface temperatures in the Pacific is ENSO (El Nino Southern Oscillation).

ENSO seems to have SOME correlation with the PDO but it's correlation with the 1940's to 70's era is much more ambiguous than the PDO's. Still, as I said to Bob:

The general point remains. Ocean equilibrium cannot be inferred just by looking at GMAST.

Update 2: Is "unconditional" the wrong adjective for Muscheler's claim that a steady high level of solar activity is "unlikely" to cause warming?

A number of WUWT commenters complained about this. For instance nuclearcannoli wrote:

With all due respect, the guy said it was “unlikely.” That is not an “unconditional” statement by any definition of which I’m aware.

I answered that while Muscheler's statement is not absolute, but it is unconditional. That is, it was not stated as relying on any assumptions about ocean equilibration or anything else. When pressed Muscheler did hint at underlying assumptions. The reference to lack of warming from 1940-1970 suggests that he sees this lack of warming as an indication that the oceans had already equilibrated to the high level of solar forcing, but this conditional was not included in his original statement.

My paraphrase did change his less than absolute statement into an absolute statement. Is that what Muscheler is complaining about? Is he complaining because my paraphrase fails to leave room for the possibility that late 20th century warming really was caused by the sun? Clearly not. His clarifications are all to bolster argument for ruling out a solar explanation. That's why I passed over that point. I was just following Muscheler on that one.

If he had said that I was understating the extent to which he intended to leave room for a solar explanation then obviously I would have been glad to issue a correction on that point. But he did not want to walk back his dismissal of the solar explanation. He wanted to re-argue it and re-affirm it.

Still, it would have been better not to call Muscheler's statement unconditional. The distinction between unconditional and absolute may be plausible but it is not particularly natural. People are right to question whether his statement is actually unconditional, so it introduces a gratuitious point of disagreement to use the term.