Guest Post by Willis Eschenbach
I was out trolling for science the other day at the AGW Observer site. It’s a great place, they list lots and lots of science including the good, the bad, and the ugly, like for example all the references from the UN IPCC AR5. The beauty part is that the ones which are publicly available are marked “FULL TEXT”, so you can just search for that and step from study to study knowing that they’re not paywalled. So as I said, I was trolling through the full text links and I ran across an interesting study entitled Global Decadal Upper-Ocean Heat Content as Viewed in Nine Analyses by Carton and Santorelli, hereinafter C&S2008. Here’s their money graph, Figure 1:
Figure 1. Nine different estimates of the change on oceanic heat content, including one model and eight observational estimates. When comparing to other analyses, note that this analysis has oceanic heat content (OHC) expressed in units of 10^8 joules per square metre, and not the more usual global total OHC which typically is measured in units of 10^22 joules. The conversion is described in the last sentence of the caption. (Actually, I think that the caption to Figure 1 in their paper was from another context and wasn’t updated … but the meaning is clear).
I was hooked when I read the abstract, with its mention of the volcanic analysis, viz:
ABSTRACT
This paper examines nine analyses of global ocean 0-/700-m temperature and heat content during the 43-yr period of warming, 1960–2002. Among the analyses are two that are independent of any numerical model, six that rely on sequential data assimilation, including an ocean general circulation model, and one that uses four-dimensional variational data assimilation (4DVAR), including an ocean general circulation model and its adjoint. Most analyses show gradual warming of the global ocean with an ensemble trend of 0.77 x 10^8 J m-2 (10 yr)-1 (=0.24 W m-2) as the result of rapid warming in the early 1970s and again beginning around 1990. One proposed explanation for these variations is the effect of volcanic eruptions in 1963 and 1982. Examination of this hypothesis suggests that while there is an oceanic signal, it is insufficient to explain the observed heat content variations.
So what did I learn from this paper? To start with, I was totally unaware that there were nine different estimates of the changes in ocean heat content, so I learned that. And quite a bit more … including being reminded that this kind of “spaghetti graph” without error estimates is useless.
So the first thing that I did was to go get the error estimates on the Levitus data shown in Figure 1 (dashed purple line) and add it to the graph so I could see what was going on:
Figure 2. Same as Figure 1, but I have highlighted the Levitus data and added the vertical red lines showing the error of the Levitus data.
Now, I have long held that the error estimates in Levitus were underestimated … I would say that this graph agrees.
I also have to note in passing that I was unable to replicate their Figure 1 regarding the Levitus results. Using the data downloaded from the above link, here is what the Levitus analysis currently shows:
Figure 3. Figure 1 from C&S2008, overlaid with current Levitus results shown in red.
As you can see, there is good overall agreement with their data with the exception of the period from 1969 to 1984 … I have no explanation for this.
However, that’s not what I was interested in. I wanted to know about the volcanoes. For some time, I have argued in a variety of posts that the effects of volcanoes on the planet’s temperature were overestimated, and sometimes greatly so. So I was surprised to see their results for the eruption of El Chichón in Mexico. They took an interesting tack in their analysis. For each area of the ocean, they compared the average ocean heat content during the four years before the eruption, with the average heat content in the four years following the eruption. That seemed like a reasonable metric to me, and a good way to go about it. Figure 4 shows their results of the 9 analyses regarding the eruption of the El Chichón volcano in 1982:
Figure 4. Ocean heat content (OHC) net change from the four years before the eruption of El Chichón, Mexico, to the four years after the eruption. Upper 8 panels show the 8 observational datasets, and bottom panel shows the model. Note the different scales … presumably used because the changes in the model results are only about 2/3 the size of the observations. ORIGINAL CAPTION: FIG. 3. Change in 4-yr average heat content spanning the eruption of Mount Agung (1963). Prior to computing the heat content change a regression analysis is used to remove the effects of ENSO and a linear warming trend (see Fig. 2). … Changes exceeding ± 5 x 10^8 J m-2 are shaded. Lowest panels show the change in heat content from a five-member ensemble of the GFDL coupled simulation CM2.1 with complete aerosol forcing. Changes exceeding ± 3 x 10^8 J m-32 are shaded.
Now at first sight, all of that looks like confirmation that the volcano caused actual cooling and that my hypothesis of minimal volcanic cooling was wrong.
However, if the cooling is from the eruption, then why are there areas of warming? Why is the cooling localized in the region just below the equator in the Pacific, when the volcanic aerosols are initially from above the equator and then spread widely around the planet? And why is there not increased cooling in the region around the eruption site in Mexico?
The answer, as usual, lies in more observations. Figure 5 shows the corresponding 4-year averages for Pinatubo …
Figure 5. As in Figure 4, but for the eruption of Mt. Pinatubo in the Philippines.
As the paper itself says …
For Mount Pinatubo most analyses show general warming except in the western equatorial [South] Pacific.
General warming of the ocean after the largest volcanic eruption in modern times? Sure seems like that supports my claims … to me, the only conclusion that we can draw from these observations of the two volcanic eruptions is that we’re looking at normal variations in OHC, and that whatever the effects are, they are pretty dang small.
Close inspection reveals a final and very strong indication that the changes shown in Figures 4 and 5 are NOT from the two eruptions, but are natural variations of unknown origin.
The indication is that the shape of the cooling does not have the form that the modelers predicted. As the models show, if forcing ruled temperature the largest effect would be expected to be immediately downwind of the eruption site. Note in Figure 5 that of all of the nine results (8 from observations, 1 from the model), the only one showing North Pacific cooling downwind from Pinatubo was the model. You can see it in the model results, the blue area like an arrow pointing at the northern Philippines, with the tail streaming straight downwind in the north Pacific … but none of the observational datasets show that pattern of cooling downwind from Pinatubo.
Not only that, but look back at Figure 4. Care to guess which of the nine analyses claimed that there would be cooling downwind from the eruption in Mexico, in the area of the Caribbean and across the top of South America? Yeah … the model was the only one … and it didn’t happen. So even in the areas right downwind from the eruptions, we don’t find the expected heat content changes from the change in solar forcing.
The volcanoes pose a huge problem for the commonly held view that the changes in global average temperature are a linear function of the changes in forcing. The climate models are nothing but a mechanistic implementation of that circumscribed and simplistic hypothesis.
Now, we know for a fact that the solar forcing after Pinatubo underwent a large and fairly lengthy drop … but we don’t find either the amount or the pattern of cooling predicted by the models. Heck, not only that, but the predominate pattern after Pinatubo was warming, not cooling … once again, the only tenable conclusions are:
1) Whatever the volcanoes might be doing, they’re not doing what the model says or what conventional climate theory predicts, and
2) Whatever the volcanoes might be doing, they are not doing enough of it to even rise above the noise.
To me, this is simply more evidence that the underlying climate paradigm, the idea that changes in temperatures are a linear function of changes in forcing, is simply not correct. If it were correct, the eruptions would show it … but they simply don’t.
That’s why I describe myself as a climate heretic rather than a skeptic—I think that the most fundamental paradigm of how the climate works is wrong. The temperature changes are NOT a linear function of forcing changes as conventional climate theory holds.
As usual, my best wishes to you all,
w.
PS—Also as usual, please quote whatever you disagree with when you comment on it. That way we can all be clear just what you are referring to.
That’s why I describe myself as a climate heretic rather than a skeptic
I like this one. I see the main “climate tribes” as:
– Orthodox and alarmists. (Climate alarmists)
– Orthodox and rational. (Climate lukewarmers, or sceptics)
– Hetherodox. (Or climate heretic, as you say)
The last ones are the only ones trying to advance the science theoretically.
http://climategrog.files.wordpress.com/2014/04/erbe_vs_aerosol_forcing1.png?w=843
What they try to do with 21 scaling is to fit the aerosol forcing directly to SST but this is not appropriate since whatever effect there is on SST will not settle within a month or two. It also very obviously fails to match either the magnitude of the initial response or the later rebound. The TOA distruption has returned to zero about a year later when the volcanic forcing is still at 50% of it’s peak, as Roy Spencer pointed out in his “Pinatubo revisited” article,
Now if you read Lacis et al 1992 , these guys are not clowns, they are well capable of doing a proper physical account of volcanic aerosol effect in excruciating detail.
That means when they do this kind of fudge and go back on solid, physics-based results that they, themselves have published, they know exactly what they are doing and why.
http://pubs.giss.nasa.gov/docs/1992/1992_Lacis_etal_1.pdf
Willis shows:
“… the idea that changes in temperatures are a linear function of changes in forcing, is simply not correct…”
______________________
Bears repeating…
I’m still in the process of writing this up , but since it is up for discussion, here is what you get if you give the system time to react to the forcing. (eg time for mixed layer to warm and provoke a climate feedback. )
http://climategrog.files.wordpress.com/2014/03/tropical-feedback_resp.png?w=814
The volcanic forcing is convolved with an exponential kernel which calculates a relaxation to equilibrium response. Here I used a 19mo time constant which matches pretty well the TOA net flux anomaly.
This also ensures that the area under the flux graph ( ie the total energy ) is the same as the direct TOA forcing. The fitted line bisects the remaining sub-annual variability very well right the way through. It rises correctly though 1992 , peak about right and falls off right down the middle.
The flat portion in 1993 is a break in the ERBE data that I filled with the annual cycle mean. The fitted line falls almost exactly on the mean.
That was done with the Lacis value of AOD * 30 W/m^2
The relaxation thing is like the way a hot object cools quicker, the further it is from equilibrium with its surroundings. This can be calculated as a weighted average, where the weightings are a decaying exponential.
Rather than simply cooling, the volcano could have muted the climate response to the el Chichon el nino. If there was less decrease in cloud cover because of el aerosals and wind pattern changes, there’d be less evaptions (less transfer of from the ocean) as well as less transformation of SW to latent heat. So there’d be both less transfer of heat and less of a forcing.
Correcting for the el Nino ingores the fact that el Nino response is likely affected by the volcano and is inappropriate.
Greg Goodman says:
April 7, 2014 at 4:08 am
One important thing to note in OHC and SST is that the drops begin in 1980 well before the March 1982 eruption. More spurious correlation if that is attributed to El Chichon.
===============
If memory serves there was a post (by Willis?) a while back that showed that cooling was anticipating volcanic eruptions by a couple of months. First it cooled, then the volcanoes erupted.
Which strongly suggests that volcanoes do not cause cooling. Which would mean the climate models have it backwards. Which likely helps explain why the climate models have gone off the rails and are running hot.
I guess the take home message is that big volcanos cool the atmosphere, but that when the sulfates from the volcano get to 90,000 feet or so (as with Pinatubo), their effects are global, because they are spread out so widely and for so long. Seems like the shorter lived ash particles, which might be more local in their influence, may not last long enough to have much of a local impact? Willis, is this your view?
Typo:
“they’re not doing what the model says or what convention climate theory predicts”
Unless we’re talking about climate theory taking place only at conventions 😉
[Thanks, Jeff. Fixed -w.]
johndo says: “In his recent post Bob Tisdale showed about two thirds of this troposphere temperature change in the SST. http://bobtisdale.files.wordpress.com/2014/04/figure-4.png”
The graph you linked was of sea surface temperature anomalies, not troposphere temperature. In fact my post was not about troposphere temperatures, so I’m not sure how you came to your conclusion.
Regards
johndo, maybe it was the following graph that you wanted to link, which I posted in a comment?
http://bobtisdale.files.wordpress.com/2014/04/tropical-pacific-ssta-mata-tlta.png
Cheers
fred said: “If memory serves there was a post (by Willis?) a while back that showed that cooling was anticipating volcanic eruptions by a couple of months. First it cooled, then the volcanoes erupted.
Which strongly suggests that volcanoes do not cause cooling. Which would mean the climate models have it backwards. Which likely helps explain why the climate models have gone off the rails and are running hot.”
Yes, Willis’ early thread on this subject “Spot the volcano” pointed out that the effect is far from obvious in a blind test. Any volcanic effect there is, is smaller than the natural variability in SST (possibly more detectable in TLT, but by no means clear).
This is particularly true of El Chichon, but also Mt P. The IPCC methods are simply wiggle matching the 3 three major events without any check as to whether the resulting match is spurious correlation.
When Willis first brought this up and I started looking, I thought the modellers were exaggerating both the volcanic and CO2 forcings. but a more thorough examination of data surrounding Mt P, suggests the actual volcanic forcing is _under-estimated_ , it’s the net result that they are over-estimating, meaning there’s even stronger neg. feedbacks that it previously appeared.
As they say: it’s worse than we thought.
Results from just one period could still be misleading, but I’m finding similarities with Spencer and Boswell 2011 and Dessler 2011 plots taken from post-2000 CERES TOA. . More on that later.
http://bobtisdale.files.wordpress.com/2014/04/tropical-pacific-ssta-mata-tlta.png
What long/lat box was that from Bob?
How could a famously sea-faring man possibly describe “trawling” as “trolling”? That would be to dredge up unnecessary unpleasantness… 🙂
“So the first thing that I did was to go get the error estimates on the Levitus data shown in Figure 1 (dashed purple line) and add it to the graph so I could see what was going on:”
there are two observational datasets and the rest are models.
You cant conclude the error bars in Levitus are wrong by comparing the observational error bars with model reconstructions.
“Now at first sight, all of that looks like confirmation that the volcano caused actual cooling and that my hypothesis of minimal volcanic cooling was wrong.”
its long been known that models over estimate the cooling and overestimate the rebound.
Super-colossal eruptions
There has been only one other magnitude 7 “super-colossal” eruption in the past 1500 years–the massive eruption of the Indonesian volcano Tambora in 1815. The sulfur pumped by this eruption into the stratosphere dimmed sunlight so extensively that global temperatures fell by about 2°F (1°C) for 1 – 2 years afterward. This triggered the famed Year Without a Summer in 1816. Killing frosts and snow storms in May and June 1816 in Eastern Canada and New England caused widespread crop failures, and lake and river ice were observed as far south as Pennsylvania in July and August. The Tambora eruption was about 40% smaller than the 535 – 536 A.D. event, as measured by the number of sulfur aerosol particles deposited in Greenland ice cores.
In an article published in 2008 in the American Geophysical Union journal EOS, Dr. Ken Verosub of the University of California, Davis Department of Geology estimated that future eruptions capable of causing “Volcanic Winter” effects severe enough to depress global temperatures by 2°F (1°C) and trigger widespread crop failures for 1 – 2 years afterwards should occur about once every 200 – 300 years. Even a magnitude 6 eruption, such as the 1600 eruption of the Peruvian volcano Huaynaputina, can cause climatic change capable of killing millions of people. The Huaynaputina eruption is blamed for the Russian famine of 1601-1603, which killed over half a million people and led to the overthrow of Tsar Boris Godunov. Thankfully, the climatic impacts of all of these historic magnitude 6 and 7 eruptions have been relatively short-lived. After about two years, the sulfuric acid aerosol particles have settled out of the stratosphere, returning the climate to its former state.
Mega-colossal eruptions
Even more extreme eruptions have occurred in Earth’s past–eruptions ten times more powerful than the Tambora eruption, earning a ranking of 8 out of 8 on the Volcanic Explosivity Index (VEI). These “mega-colossal” eruptions occur only about once every 10,000 years, but have much longer-lasting climatic effects and thus are a more significant threat to human civilization. According to the Toba Catastrophe Theory, a mega-colossal eruption at Toba Caldera, Sumatra, about 74,000 years ago, was 3500 times greater than the Tambora eruption. According to model simulations, an eruption this large can pump so much sulfur dioxide gas into the stratosphere that the atmosphere does not have the capacity to oxidize all the SO2 to sulfuric acid aerosol. The atmosphere oxidizes as much SO2 as it can, leaving a huge reservoir of SO2 in the stratosphere. This SO2 gradually reacts to form sulfuric acid as the OH radicals needed for this reaction are gradually produced. The result is a much longer-lasting climate effect than the 1 – 2 years that the magnitude 6 and 7 events of 535, 1600, 1815, and 1991 lasted. A magnitude 8 eruption like the Toba event can cool the globe for 6 – 10 years (Figure 3), which may be long enough to trigger an ice age–if the climate is already on the verge of tipping into an ice age. Rampino and Self (1992) argued that the sulfur aerosol veil from Toba was thick and long-lasting enough to cool the globe by 3 – 5°C (5 – 9°F), pushing the climate–which was already cooling and perhaps headed towards an ice age–into a full-scale ice age. They suggested that the response of Canada to the volcano played a particularly important role, with their model predicting a 12°C (22°F) reduction in summer temperatures in Canada. This would have favored the growth of the Laurentide ice sheet, increasing the reflectivity (albedo) of the Earth, reflecting more sunlight and reducing temperatures further. The controversial Toba Catastrophe Theory asserts that the resulting sudden climate change reduced the Earth’s population of humans to 1,000 – 10,000 breeding pairs. More recent research has shed considerable doubt on the idea that the Toba eruption pushed the climate into an ice age, though. Oppenheimer (2002) found evidence supporting only a 2°F (1.1°C) cooling of the globe, for the 1000 years after the Toba eruption. Zielinski et al. (1996) argued that the Toba eruption did not trigger a major ice age–the eruption merely pushed the globe into a cool period that lasted 200 years. Timmereck et al. (2010) used a model to show that the sulfate particles inthe stratosphere would have clumped together after the eruption, limiting the colling effect to four years, with a maximum global reduction in temperature of 3.5°C. Interestingly, a previous super-eruption of Toba, 788,000 years ago, coincided with a transition from an ice age to a warm period.
justsayin999 says:
April 7, 2014 at 8:10 am
How could a famously sea-faring man possibly describe “trawling” as “trolling”? That would be to dredge up unnecessary unpleasantness… 🙂
_____________________
Well, if you actually knew anything at all about fishing, then you would know that “trolling” is a valid technique, widely employed. Over the course of my fishing life, I have spent weeks if not months, trolling with a great deal of success. Yumm.
Steven Mosher says:
its long been known that models over estimate the cooling and overestimate the rebound.
What do you mean by “rebound”? Are you saying they recover too quickly or go too high afterwards ( or both ). Could you say more clearly what this acknowledged defect is ?
Thanks.
Steven Mosher says:
April 7, 2014 at 8:16 am
Steven, the errors in the Levitus data don’t even agree with the Levitus data itself … see Figures 2 & 3.
Steven Mosher says:
April 7, 2014 at 8:18 am
No, that hasn’t “long been known”. In fact, the exact opposite has been argued over and over by the alarmists. They have consistently claimed for a couple decades that the volcanoes were the very test that showed the models to be correct.
For example, Susan Solomon claimed that volcanoes were the cause of the pause, as I discussed here. Then we have the recent claim that volcanoes caused the Little Ice Age, discussed here. Hansen claimed the models were confirmed by the Pinatubo results as discussed here. And Dr. Andrew Lacis claimed in response to a question about evidence the models were right that
as discussed here. Then we have the recent claim that volcanoes caused the Little Ice Age, discussed here.
Finally, Richard Muller and the BEST group that you work with made the same bogus claims, as discussed here.
Now … if, as you fatuously claim without a citation or a single scap of evidence, it has “long known that models over estimate the cooling and overestimate the rebound”, how come I’m one of the few people that has been saying that, and all the modelers and alarmists, including the Berkeley Earth Surface Temperature group of … charming folks that you work with, say that the volcanoes are the very test that shows the models are correct?
A half dozen citations would be sufficient to back up your claim that it has “long been known” that the models are wrong in this regard … the citations need to be from over a decade ago, though, last year doesn’t count in a claim that something has “long been known”.
I suspect that you will never return to this thread to either provide those citations, or to admit that you were wrong … but heck, you’re welcome to prove that I’m incorrect.
w.
Steven Mosher says:
April 7, 2014 at 8:18 am
Everyone knows that Sagan was utterly wrong, yes.
Ted Vaughn says:
April 7, 2014 at 8:20 am
Another guy with a host of claims and not a single citation … Ted, the claims about Tambora are as much urban legend as fact, just like the claims about the rest of the volcanoes. See my post called “Missing the Missing Summer” for a more nuanced view of the question. Short answer? The global temperatures didn’t do what you claim, even Tambora only had local effects …
w.
Tis a little known fact in this internet era that trolling is a fishing technique; that of trailing a baited, hooked line or two behind a slowly moving boat.
It is often presumed in internet land that the troll we speak of has something to do with bridges, however the meme of throwing out a baited comment and waiting for the ‘fish’ to bite is no doubt the original derivation.
Having said that, the usage of the word has deteriorated to the point that it is now used as a lable for anyone who disagrees with one online. ie, rather meaningless.
Interestingly, in his trolling for science, I note Willis states HE was the one who got hooked. That indeed sometimes happens when fishing from a moving boat.
Kenny says:
How about volcanic eruptions underwater. Is there enough heat to cause a change in the ocean temp?
No, except very locally. The heat capacity of the ocean is enormous.
CarolinaCowboy says:
I have a question about underseas eruptions and geysers, do they not add heat to the ocean?
Yes, they do, but the effect is very slight due to the huge heat capacity of the ocean. Though it is conceivable that the addition of heat to the very cold bottom layer of the ocean may help driving the thermohaline circulation
Kenny says:
Also…..The lava that is produced….does it displace enough water to cause any rise in the ocean? I’m just curious.
It does, but the effect is only significant over very long time periods (tens of million years). Over such timespans the varying activity of the ocean ridges can change sea-levels by tens or even hundreds of meters. During periods of high ocean ridge activity (e. g. the mid-Cretaceous) the rock in the ridges is hotter on averag and less dense and takes up greater volumes.
It’s so nice to know that we can depend on Mosh doing a meaningless “drive by” any time serious discussions of papers occur. Such deeply considered comments bring new meaning to “shooting oneself in the foot.” That’s Mosh, one gun, six bullets, two feet.