Guest Post by Willis Eschenbach
The claim is often made that volcanoes support the theory that forcing rules temperature. The aerosols from the eruptions are injected into the stratosphere. This reflects additional sunlight, and cuts the amount of sunshine that strikes the surface. As a result of this reduction in forcing, the biggest volcanic eruptions are said to depress global temperatures, sometimes for years.
The idea that large volcanoes significantly cool the planet is widely accepted. This effect is built into the climate models, for example. It is a reflection of the dominant climate paradigm, which is that surface temperature is a linear function of forcing. Since it can be measured observationally that the volcanoes greatly reduce the global solar forcing, it follows that they must significantly affect the global temperature.
However, I hold that the climate system is not an inert slave of changes in forcing. I hold that the climate system immediately and actively responds to changes in forcing by adjusting things like albedo, cloud type, cloud formation times and locations, timing of Nino/Nina alterations, and the like, to quickly counteract any forcing changes.
Which means, of course, that according to my hypothesis, even very large volcanoes should a have very small effect on the global temperature. To see which hypothesis is true, mine or the standard AGW hypothesis, I devised a little game I call “Spot the Volcanoes”. Two of the largest volcanoes of the century occurred within a twenty year time span. See if you can tell where they occurred.
Figure 1. First difference (month-to-month change) in global surface air temperature. Timespan shown is twenty years. Two of the largest volcanoes of the 20th century are shown in this record. The volcano in the picture is Mt. Redoubt, Alaska, one of my favorite mountains.
In Figure 1, to make things a bit difficult, I show the month-by-month CHANGE in temperature. This is not the temperature itself, but the month-by-month change in temperature, called “delta T” (∆T). If the temperature is a function of the forcing, the eruptions should be making the temperatures drop for a while. So the game is, where in Figure 1 are the two eruptions? Make your choice before you take the jump …
The answer is shown in Figure 2 below. It contains the record of the atmospheric transmission over Mauna Loa. The two eruptions, of El Chichon and Mt. Pinatubo, are very apparent in the Mauna Loa (MLO) record. I have scaled the Mauna Loa record to the corresponding GISS estimate for the forcing from Pinatubo (in W/m2), in order to show the generally accepted size of the volcanic forcing.
Figure 2. As in Figure 1, plus Mauna Loa atmospheric transmittance observations. These observations are of the total amount of clear-sky sunlight making it through the atmosphere.
Now, I can already hear folks grumbling, that this was not a fair game, that it was rigged because it was the first differences and not the actual temperature itself. And besides, most people don’t spend much time looking at first differences, so it was too hard. And perhaps those folks are 100% correct.
So let’s play a bonus round of “Spot the Volcanoes”, this time using the real temperature data. Figure 3 shows a stretch of the HadCRUT3 global surface air temperature record. This time it includes one smaller and two larger volcanoes. See if you can spot where the big ones erupted:
Figure 3. A stretch of the HadCRUT3 temperature record containing one small and two large eruptions. Don’t bother trying to find the small one.
So once again, the game is to spot two volcanoes.
Now at this time,
.
We’ve got to play the game show music,
.
. dee
. dee
. da dee dee dum
.
So as to hide the answer,
.
Until you make your choice, of the exact location of the two eruptions in Figure 3.
.
So here it is.
Figure 4. As in Figure 2, showing the eruptions of El Chichon (1982) and Pinatubo (1993). The small eruption is Mt. Agung (1963).
I’m sure you understand my point. There is nothing to see. The kinds of temperature excursions we see after the volcanoes are not different from the temperature excursions before the volcanoes.
How big an effect should we have seen, given the IPCC assumptions about climate sensitivity? Well, the average change in forcing over the three years following the Pinatubo eruption is ~ -1.7 W/m2. Now, that’s about half the forcing change expected from a doubling of CO2, maintained for three entire years … and where’s the response? Using the IPCC numbers, we should have seen a temperature drop of 1.4°C at equilibrium, and three years after the step change we should have seen at least a full degree of that …
Instead of a full degree of cooling after Pinatubo, or even half a degree, we see maybe a tenth of a degree of cooling.
But wait, as they say on TV … it’s even worse than that. The drop after Pinatubo may be just by chance, because after the earlier El Chichon eruption we see maybe a tenth of a degree of warming … and the average three year change in forcing for El Chichon is only trivially smaller than Pinatubo, at ~ -1.6 W/m.
So this is a great natural experiment regarding changes in forcing. From these observations, as near as we can tell, half the forcing change expected from a doubling of CO2 was applied for three full years, at two different times, and it resulted in … well, pretty much nothing.
So I’d say that the volcanic eruption data strongly supports my thermostat hypothesis, which says that changes in forcing are almost immediately and nearly completely offset by opposing changes in other aspects of the climate system.
w.
PS—Here’s the double bonus question … the UAH lower temperature record:
Figure 5. UAH MSU satellite based global lower tropospheric temperature record.
This time the game is a bit different. Are there one or two volcanoes in the record, and where is it / are they?.
Now at this time,
.
We’ve got to play the game show music like last time,
.
. dee
. dee
. da dee dee dum
.
So as to hide the answer,
.
Until you make your choice, of the exact location of the two eruptions in Figure 5.
.
So here it is.
Figure 6. As Figure 5 plus transmittance information.
Note that as with the surface temperature record, the globe cooled slightly after Pinatubo … and that as with the surface temperature record, the globe warmed slightly after El Chichon. And since the post-Pinatubo drop is indistinguishable from the post-1983 and the post-1988 drops, there is no reason to assume that the post-1991 drop is due solely to Pinatubo.
Which in my opinion is why all of the analyses focus on Pinatubo, while poor El Chichon is roundly ignored because it didn’t get the memo about causing a temperature drop.
PS—Does this mean volcanoes have no effect on the climate? No, it just means that because of the immediate and basically “equal but opposite” response of the climate system to forcing changes, the effect is much more local, much shorter lived, and much smaller than would be expected if the IPCC estimates of climate sensitivity were correct.
FURTHER READING: Climate forcing by the volcanic eruption of Mount Pinatubo
[UPDATE] People have asked for more information about how the climate responds to counteract the cooling action of the volcano. Figure 7 shows the response of the albedo to the Pinatubo eruption. The albedo immediately began to drop, allowing more and more sunlight to warm the surface.
Figure 7. Anomaly in post-albedo solar isolation for the period 1984-1998. The transmittance change due to the volcano is shown in red. Albedo data from Hatzianastassiou et al.
You can see that it’s not too hard to spot the volcano in this graph … which is exactly the reason why it’s so hard to spot in the other graphs.
w.
Reader is cautioned of serious consequences of the global warming.
Global warming causes volcanic eruptions
Compare the two steepest temperature rises in the Central England records going back to 1659.
http://www.vukcevic.talktalk.net/CET1690-1960.htm
After careful examination attentive reader will observe not only similarity in the trends but also the ‘year to year’ temperature movements.
According to this hypothesis, soon to become a theory (to be widely acclaimed by the IPCC) a tipping point is reached when the unprecedented land temperature rises to the point where the rock expansion opens up old volcanic calderas (calderae ?) allowing magma and volcanic ash to spurt into atmosphere, resulting in substantial cooling in the following years.
Reference: http://www.ipcc.ch/
Climate Change 2014: Working Group I: The Physical Science Basis
Willis – look at figures 8, 9 and 10 in my book “What Warming?” if you have it. If you don’t have it, get it and look. The difference between El Chichon and Mount Pinatubo is what made me write that section. If you are unfamiliar with my book you will notice that I do not use any computer smoothing but apply a magic marker to raw data. You get a thick line that covers up most of the fuzz in the record and neatly outlines the correct global trend. Any use of computer will destroy data and should be avoided. The fuzz looks random not because it is random errors but because it represents local cloudiness variations that are randomly distributed. But my motivation was that the 1992/93 La Nina is marked everywhere as Pinatubo cooling, even on Roy Spencer’s web site, which would change this particular La Nina cooling into a volcanic cooling instead of a part of ENSO. That volcanic cooling claim started with Best who pontificated in 1996 that “Pinatubo climate forcing was stronger than the opposite warming effects of either the El Nino event or anthropogenic greenhouse gases in the period 1991-93.” The man had access to satellite data but he had no idea how to spot an El Nino in it. Back to El Chichon and Pinatubo. After assigning the 1992/93 La Nina to Pinatubo cooling Best goes on and states that “… surface cooling is clearly documented after some eruptions (for instance Gunung Agung, Bali in 1963) but not others – for example El Chichon, Mexico in 1982…” Wow! I had to check that. Since I had already worked out the theory of how ENSO works and identified all El Ninos and La Ninas in the satellite record it was easy to see the difference between El Chichpn and Pinatubo. El Chichon erupted exactly when a La Nina had just bottomed out and an El Nino was beginning. Pinatubo, on the other hand, erupted at the peak of an El Nino warming that was immediately followed by a La Nina cooling. Just plain luck, it seems: if an eruption is followed by a La Nina, that La Nina counts as volcanic cooling. But if it is followed by an El Nino, like El Chichon was, there is an unexplained absence of cooling that does not disturb them thinking that they found volcanic cooling elsewhere. That is a pretty big loose end in their theory which nevertheless is accepted by climate scientists. So, let’s test it and check out that Gunung Agung. It should be followed by a La Nina if my theory is correct. And so it is – a pretty steep drop just after the eruption. But before I could determine this I had to find the locations of El Ninos and La Ninas in ground-based temperature records. That turned out to be easy despite the fact that nobody had either noticed or bothered to remark about it. They are present in all temperature records going back to the nineteenth century. The BEST project showed a comparison of four temperature records. I looked at them and found the same El Nino peaks present in temperature curves from America, Europe, and Japan. There are differences of course and someone will eventually sort them out. My figure 10 shows the timing of seven volcanoes against the HadCRUT3 temperature curve. It is hard to test it because cooling is not one of the things that is recorded about most volcanoes. But one thing is certain: there is no such thing as volcanic cooling in nature. It is simply a matter of lucky timing in relation to the ENSO phase that determines whether and how much “cooling” is associated with any particular eruption.
Good post Willis.
You made a good point without resorting to Snark and Pomp. I really enjoyed it!
You seem to have have reverted in the comments but if you didn’t, you would be someone else.
Your “buttons” really seem easy to push. Hang in there big guy!
dana1981 says:
March 16, 2012 at 12:29 pm
Like Groucho Marx said, “Who ya gonna believe? Grant (Tamino) Foster or your own lying eyes?” Me … I trust my eyeballs to identify large excursions. Tammy and all the rest claim large excursions. I don’t see them. You can’t see them either. So sue me.
I KNOW that claim has been made, Dana, you can save your citations for someone more credulous. I analyzed Hansens claims of the same thing here at some length, take a look. You don’t seem to understand. These guys will look you in the eye, assure you of their honesty, and switch the pea under the shells. You are very, very naive to trust them.
Finally, anyone who thinks that “peer-reviewed” means anything at all in climate science is either clueless or is truly not following the story …
w.
An interesting post.
I have long argued that the effects of volcanoes may be over-stated. There is no doubt that super volcanoes can have a noticable effect on climate for a short period as the history books recount, but super volcanoes are very few and far between. I would say that there were none in the 20th century.
Even the effects of Krakotoa (August 1883) are not clear. If one approaches the data set blind, I suspect that the observer would find it diificult to know whether there had been a volcano eruption and if so when. See for example Gisstemp. I attach a plot for the period 1880 to 1910 and a plot for the period 1880 to 1890.
http://www.woodfortrees.org/plot/gistemp/from:1860/to:1910
http://www.woodfortrees.org/plot/gistemp/from:1880/to:1890/from
It will be seen that there was a period of cooling between 1865 and 1885 and that the period 1883 to 1885 was in no way exceptional. The rate of cooling does not appear to have accelerated afetr 1883. In the first plot, the low figure of -0.52degC is in late 1882 about 10 months prior to the eruption and during the autumn of 1883 it actually warmed by just over 0.1degC! of course, there may be other forcings at play but if so the point made by Willis remains that Volcano forcing is not overwhelming in nature even in the case of a super volcano like Krakatoa.
Well done.
KR says:
March 16, 2012 at 11:14 am
That’s why the climate is so hard to pin down. Every time there is a volcanic eruption cooling down the earth, some other unknown process emerges to exactly cancel it out./s
Except for the super volcanoes of course, none of which has occurred since modern instruments started recording (dang)./s
Willis – Note that dana1981 (http://wattsupwiththat.com/2012/03/16/volcanic-disruptions/#comment-925422) was referring to an article by Piers M. De Forster, not Grant Foster.
As to the eyeballs – that’s why we have statistics, Willis, because the eyecrometer is so often wrong.
Now, as to accusing various folks of dishonesty – when both Foster’s and Hansen’s work is easily repeatable (I’ve seen several bloggers who have repeated those papers work with their own data and software), and to conspiracy theories about the body of climate science and peer review over the last 150 years – if you’re convinced of those, well then, neither I nor anyone else will be able to change your mind. But I cannot consider grand conspiracy theories rational…
michael Moon says:
Anecdotally, I bought a morotcycle in 1993. I lived in Michigan, and the temperature rarely got out of the 60′s all summer. I had to keep my jacket on to ride the entire season. It was the coolest Midwest summer I ever went through. Was I wrong to attribute this to Mt. Pinatubo?
The plural of anecdote is not data. I tend to suspect that there is a lot of post hoc ergo propter hoc reasoning with respect to volcanoes.
If a string of cold weather follows a volcanic eruption, the reasoning is that the volcanoes caused the cooling.
But if a string of warm weather follows a volcanic eruption, that particular eruption is ignored.
If the volcano does not cause immediate and steep drops in temperature then its effects are not caused primarily by blocking out the sun. We notice a temperature drop the moment a cloud blocks the sun, not weeks later. So when Pinatubo went, shouldn’t the blocking of sunlight have been immediately obvious? And immediately obvious on the temperature record too?
In the case of the CO2 theory, the post hoc ergo propter hoc is even worse. Since world temperatures are not rising fast enough for CO2 theory, something must have caused that. Villains are hard to find, so they are made up: volcanic and industrial aerosols get the blame. Evidence they work that way, be damned!
Doing good climate science may not be hopeless, but it certainly is daunting.
The volcano issue raises the chaos issue again: multiple, non-linear processes conflated and an insufficient number of examples (ie small sample size) to do anything approaching robust statistics.
Each volcano varies in several potentially important ways. For example,
—variation in latitude (change in abedo near the poles in winter unimportant compared to same change in albedo near equator,
—relation to prevailing winds and jet streams (for ash dispersal),
—in the midst of a continent vs as ocean, coincidence of ENSO events, total volume of ash distributed over various altitudes,
—a lot of ash in one day or week (explosive event) vs same amount of ash over months or years,
—particulars of the ash (particle size and chemical composition),
—coincident ex-gassing of CO2, CH3 and sulfates,
—who knows what else.
All of these might have some influence on local, regional, or global climates, or not, depending on magnitudes and underlying science.
How can one go about doing rigorous science on such a complex set of difficult to measure data points.?
Scott Covert says:
March 16, 2012 at 1:00 pm
Thanks, Scott. I used satire extensively in the last post, because those guys are buffoons, and my intention was to prick their pompous balloons.
Here, it’s science.
And I’m getting better at letting half-wits insult me, I no longer reflexively snap back, although I may snap back … but I still don’t suffer fools gladly.
w.
Larry Ledwick (hotrod): Granted this is only one local weather anomaly but the public perception at the time following these major volcanic eruptions was that the following winters 6 – 18 months after the event were noticeably harsher than usual.
There are temperature/snowfall/rainfall records set every winter and summer somewhere in the world. So showing that there was a record somewhere 6 – 18 months after something happened is uninformative.
richard verney says:
March 16, 2012 at 1:19 pm
Excellent, Richard, another game of spot the volcano where I can’t spot it. As you say, there is no visible indication of Krakatoa in the record, and that was a supervolcano. And if you can’t see that …
w.
KR
You cite Foster and Rahmstorf 2011. I have compared their estimate of the effect of Mt Pinatobu on global surface temperatures, as shown in Fig. 7(b), with Fig. 4(c) of Lindzen and Giannitsis 1998 “Volcanic cooling and climate”. F&R estimate – in particular the shape and timescale of decay of the initial temperature reduction – is similar to L&G’s simulated curve at low (no more than 1C) climate sensitivity, and inconsistent with sensitivities in the IPCC ‘likely’ range of 2C – 4.5C.
The peak temperature reduction computed by L&G is only weakly linked to climate sensitivity, except when sensitivity is exceedingly low, due to the effect of ocean heat uptake.
No fair..you’re using thermometer data. You need to use the better, more accurate measure for temperature……tree rings! And be sure to use only trees growing on the side of the volcano, preferrable downwind of the ash cloud.
Willis Eschenbach says:
As you say, there is no visible indication of Krakatoa in the record, and that was a supervolcano. And if you can’t see that …
Of course there is in any reasonably accurate temperature record.
http://www.vukcevic.talktalk.net/CET-NAP.htm
dana1981 says:
March 16, 2012 at 12:29 pm
“I’ll take peer-reviewed research over the eyeball method, personally”
Naive much? Take it from me, someone that’s published dozens of peer reviewed articles in prestigeous chemistry journals: peer review is not some magical device that turns papers into undeniable truth. Believe me, we used to write papers that guessed as to who would review them and we’d tuck some sentences and references in their to stroke the reviewer’s ego a little to more likelyget a postive review. At some level it’s a game…that can and is manipulated. So everyone please stop thinking just because something is peer reviewed that it’s beyond reproach.
The literature is filled with peer reviewed articles that have been proven wrong at a later date. Peer review is an attempt at some quality control…but once again it’s not a magic wand. The article above that Willis wrote could easily be turned into a peer reviewed article. A couple more lines of text, a few more authors (up to 33!) and some references. Send to a journal with your chequefor the appropriate amount and viola…peer reviewed publication that will be sanctified by folks like dana1981 since they are “peer reviewed”.
And that goes quadruple for “pal review”.
There are singular jumps at the start and end when you do that. There should also be singular (negative) jumps in the dT vs time plot due to the volcanoes, and the effect of a volcano should stand out more starkly on such a plot. There seems to be a singular drop at El Chichon but not Pinatubo, and the one at El Chichon is no different than several other drops not corresponding to a volcano. In any case, Willis also plotted the anomaly vs. time for those who can’t understand plot 1.
There’s boatloads of money available for any research remotely related to AGW. That money will disappear if there is no perceived problem. If a scientist stops bringing in research dollars he doesn’t advance, raises are minimal, and tenure is out of the question. It doesn’t take a ‘grand conspiracy’ for intelligent people to see that and act in their own self-interest.
Or maybe you have a better explanation as to why practically all of the adjustments to the raw data increase the warming … sea level rise … OHC …
El Chichon went head-on with the Super El Nino, similar to that of 1997. I remember that Dr Spencer also used the volcano effect for deriving the climate sensitivity, getting something 5x lower than IPCC estimates.
Nic Lewis – I would be wary of direct comparisons between L&G 1998 and F&R 2011 – Lindzen and Giannitsis themselves point out “it is clear that the uncertainties in our results are large”, and their simple 2-box ocean model (limited modeling of deep ocean circulation) has been implicated in biasing their sensitivities to the low side.
More up to date and detailed models than L&G disagree strongly: for example Bender et al 2010 (http://www.misu.su.se/~frida/benderetal10.pdf) finds that the Pinatubo eruption indicates an equilibrium sensitivity between 1.7 and 4.1 C/CO2 doubling.
vukcevic says:
March 16, 2012 at 2:01 pm
The Central England Temperature record? That’s your global evidence?
So it shows up (maybe) in a local temperature record but not the global temperature record … doesn’t say much for the eruption. But heck, I’ll play your game. The two largest eruptions in the last quarter millennium are in the CET record below … spot the volcanoes!
w.
darkobutina says:
March 16, 2012 at 12:22 pm
“Willis, you are becoming a waste of space on this blog site. Who do you think you are….”
Ah, “who do you think you are?” The words that warm the cockles of my heart and bring back cherished memories. The line every cheap-suited provincial peasant apartchik in old commie Europe used to boom out when having his worldview challenged. Usually to be followed by a list of approved exemplars or the chilling, “do you know who I am?” So, Darko, never mind Willis; we don’t know who he is or what he’s doing here, but most of us like to keep him as the site’s mascot, if it’s alright with you. More interestingly, who are you, comrade?
Yes and no, that is specifically why I noted it was only a single instance of a all time record that happened to coincide with the period that was attributed to volcanic cooling. It was however more than just a record high or low but an all time low snow melt that lasted two months, so it carries a bit more weight than a single days event.
But the correct way to make your statement is:
showing that there was a record somewhere 6 – 18 months after something happened may beuninformative.
It there is a real cooling due to volcanic activity there should be lots of new records and temperatures that flirt with record lows. You cannot have a change in the trend without that happening. A single instance is by itself simply an observation, it is the collective sum of lots of similar observations that would establish a change in the trend.
It is just as wrong to totally dismiss a single observation as it is to accept it as proof of the hypothesis.
It is “necessary” but not “sufficient” to prove the case.
The point is, those of us who lived through those cooling periods remember very well that there were large numbers of observations that were consistent with the theory that volcanic cooling had happened. The cooling may have been due to other causes like cycles in sea surface temperatures but that is irrelevant, the cooling was noticed by large numbers of people. Enough that the general public perception was solidly behind the belief that that period was cooler than normal. As Willis points out, it is not nearly as simple as a one to one correspondence and we need a LOT more data (60 – 120 years or more) of additional quality satellite data and sea surface temperature measurements to tease the very complex signals out of the noise.
There is a lot more to weather/climate than just temperature records. As noted by observers at the time, it was plainly obvious that the summers were cooler, the peak temperatures of the day were consistently a bit below what experience would suggest they should be. Not records but just cooler than normal. Plants made it very clear with late maturity of crops, like corn failing over wide areas to ripen by the usual land mark days, late and early frosts (again not records but just anomolies). Fruits failing to set, etc. How much of these effects to assign to AMO PDO etc and how much to assign to ground levels of solar isolation reduction due to atmospheric haze is not going to be easy.
Due to the earth’s thermal inertia, changes in energy input will not be reflected immediately unless they are very significant drops in input. That is why the coldest day of the year in the center of the North American land mass frequently falls around the last week of January, and first couple days of February even though the annual minimum in solar input in North America is on the winter solstice on Dec 21-22 each year.
As I pointed out to Willis the failure of the volcanic cooling to show up (if it existed) in his delta T plots could signify more than one possibility. It could show that there was no volcanic cooling, or it could show that a delta T plot is not the proper metric to look for volcanic cooling.
Experimental findings often end up pointing to something totally unrelated to what the observer was looking for, and if you are not careful in how you set up controls and the experimental method can actually demonstrate more than one possible cause for the result.
That is why replication and independent verification is so important in the scientific method.
Sometimes you prove something you were not even looking for, and like Edison, sometimes your “failed experiments” are really successes as they show you ways Not to do something.
Larry
KR
You say about Lindzen and Giannitsis 1998
“their simple 2-box ocean model (limited modeling of deep ocean circulation) has been implicated in biasing their sensitivities to the low side.”
L&G didn’t use a 2-box ocean model – they used a mixed layer coupled to a fully diffusive thermocline, which they showed matched a more sophisticated mixed layer / thermocline / deep ocean upwelling-diffusion ocean model well over periods of up to at least 20 years. Surface temperature changes don’t propogate significantly to the deep ocean over the timescale of recovery from volcanic cooling,
Further, L&G used an ocean effective vertical diffusivity value derived from direct observational data, similar to that originally used by Hansen. AOGCMs, such as those used in the Bender et al. study, are known to embody effective ocean vertical diffusivities that are unrealistically high, by a substantial factor. See, e.g, Even Hansen now admits this (Earth’s Energy Imbalance and Implications, 2011). I wouldn’t put much confidence in AOGCM-derived estimates of climate sensitivity, whether based on volcanic cooling or otherwise.