Guest Post by Willis Eschenbach [also, see update at the end of the post]
Anthony recently highlighted a couple of new papers claiming to explain the current plateau in global warming. This time, it’s volcanoes, but the claim this time is that it’s not the big volcanoes. It’s the small volcanoes. The studies both seem to follow what I call “Willis’s Rule of Author Count”. The first study is Total volcanic stratospheric aerosol optical depths and implications for global climate change, by D. A. Ridley, S. Solomon, J. E. Barnes, V. D. Burlakov, T. Deshler, S. I. Dolgii, A. B. Herber, T. Nagai, R. R. Neely III, A. V. Nevzorov, C. Ritter, T. Sakai, B. D. Santer, M. Sato, A. Schmidt, O. Uchino andJ. P. Vernier. The second study is Observed multi-variable signals of late 20th and early 21st century volcanic activity, by Benjamin D. Santer, Susan Solomon, Céline Bonfils, Mark D. Zelinka, Jeffrey F. Painter, Francisco Beltran, John C. Fyfe, Gardar Johannesson, Carl Mears, David A. Ridley, Jean-Paul Vernier, Frank J. Wentz.
Now, Willis’s Rule of Author Count says that the quality of any study is inversely proportional to the square of the number of listed authors. And these two studies have seventeen and twelve authors respectively … not a good sign.
The abstract of the first paper says:
Understanding the cooling effect of recent volcanoes is of particular interest in the context of the post-2000 slowing of the rate of global warming. Satellite observations of aerosol optical depth above 15 km have demonstrated that small-magnitude volcanic eruptions substantially perturb incoming solar radiation. Here we use lidar, Aerosol Robotic Network, and balloon-borne observations to provide evidence that currently available satellite databases neglect substantial amounts of volcanic aerosol between the tropopause and 15 km at middle to high latitudes and therefore underestimate total radiative forcing resulting from the recent eruptions.
The abstract of the second paper, in turn, says:
The relatively muted warming of the surface and lower troposphere since 1998 has attracted considerable attention. One contributory factor to this “warming hiatus” is an increase in volcanically-induced cooling over the early 21st century. Here, we identify the signals of late 20th and early 21st century volcanic activity in multiple observed climate variables. Volcanic signals are statistically discernible in spatial averages of tropical and near-global SST, tropospheric temperature, net clear-sky short-wave radiation, and atmospheric water vapor.
Now, it is certainly possible that “small-magnitude volcanic eruptions substantially perturb incoming solar radiation”. There are lots of things that perturb incoming solar radiation, with clouds heading the list. Whether small volcano emissions in turn perturb the global surface temperature is a separate question.
But for eruptions to be an explanation for the current plateau in global warming, the authors would have to show a significant increase in volcanic eruptions in the 21st century. And unfortunately (but predictably) I see no sign in either paper that they have even tried to do that.
So let’s do their job for them by taking a look at the actual records of eruptions, both large and small. The data on all known eruptions is available from the Smithsonian Volcanism Project.
Now, we have some choices in how to display this data. Let me show three of these different ways.
First, we can show the total numbers of eruptions by year, without regard to the size of the eruption. Figure 1 shows that information:
Figure 1. Count of all volcanic eruptions, regardless of their strength, during the end of the 20th and the start of the 21st centuries.
As you can see, there is very little difference between the post-2000 (or post 1998, depending on the study) eruption count and the number of eruptions during the end of the 20th century. After 2000 (or 1998), it went up a bit, then it went down a bit … overall, little change.
But wait, I can hear you saying, the eruptions are not all of the same strength … what is the average strength of the eruptions? And reasonably so, since strong eruptions would have a bigger effect than small eruptions. So let’s look at that data.
The strength of an eruption is measured by the volcanic explosivity index, or VEI. This is a logarithmic scale. This means that an eruption with a VEI of 5 is ten times stronger than an eruption with a VEI of 4, and so on.
In order to properly average these, it’s necessary to use a “logarithmic mean” To do this, you first convert the VEIs to actual values (by taking ten to the power of the VEI). Then you average the actual values, and then take the logarithm of the resulting average to convert it back into the logarithmic VEI scale. Figure 2 shows that result:
Figure 2. Annual logarithmic mean of the volcanic explosivity index, all volcanoes.
In 1991, there were two strong eruptions, Pinatubo (VEI of 6) and Cerro Hudson (VEI of 5). Other than that, there’s not a lot of variation.
Once again, you can see that the post 2000 (or post 1998) average strength of the volcanic eruptions are little different from the strength of the eruptions prior to the turn of the century. So that cannot be the cause of 21st century plateau in global surface temperatures.
Finally, we could read the implicit claim as being that there is some kind of increase in the number of small volcanic eruptions. After all, the authors say that these are the overlooked eruptions. So let’s take a look at the small pre- and post-2000 eruptions.
Figure 3. Annual count of the smaller eruptions, those with a volcanic explosivity index of less than 3.
Once again, we see little change in the number of small volcanoes. After 2000, it goes above the average, and then it goes about the same amount below the average.
Conclusions? Well, the papers may be correct in their claim that the effect of eruptions on the clarity of the atmosphere may have been underestimated.
But they are absolutely not correct in the claim that this underestimation reveals the cause of the recent 18+ year plateau in temperatures as being eruptions. There is almost no post-2000 change in either the number of eruptions, the strength of eruptions, or the number of small eruptions.
Overall? I’d say that Willis’s Rule of Author Count, that the quality of any study goes down inversely proportional to the square of the number of listed authors, is validated once again …
[UPDATE: The underlying claim of these two papers is that although there have been no large eruptions in the 21st century, it is the weaker eruptions that are causing the plateau in temperature. These are eruptions with a volcanic explosivity index (VEI) of four. Some commenters below still think that the eruptions of VEI four are significant. The Santer document shows the effect of some of the VEI 4 eruptions on the stratospheric aerosol optical depth (SAOD), which is their main indication of volcanic change. The study says:
We use stratospheric aerosol optical depth (SAOD) data from Vernier et al. [2011] to study changes in stratospheric loadings of volcanic aerosol.
The eruptions make a change of about .002 in the SAOD, viz:
Now, that looks kind of impressive … until you compare it to the larger volcanoes (source):
You can see the sizes of Krakatoa in the 1880s, Pinatubo in 1991, El Chichon in 1982, and Mt. Agung in 1963, along with some smaller volcanoes … and then you can see the part that Santer et al. are discussing. This is the almost-flat line in the post-2000 era. Sorry, but given the short-term, weak, local effects of even the largest volcanoes, I’m not buying the idea that those tiny post-2000 wiggles have any discernible effect at all.
My best regards to you all,
w.
ONCE AGAIN: If you disagree with someone, please do everyone the favor of QUOTING THE EXACT WORDS THAT YOU DISAGREE WITH. This prevents all kinds of misunderstandings and misrepresentations.
FURTHER READING: I note that this is not the first time that Susan Solomon has made the claim that volcanoes are the cause of the current pause in temperatures. In addition, she was the main mover behind one of the IPCC reports, from memory the Fourth, and is fully and completely invested in the meme of “CO2 Roolz Everything, OK” … whenever I see her name on a study, I’m sad to report that I just wince. See here for my discussion of her previous work.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Willy, how do they get around the atmospheric transmission measurements at Mauna Loa? They are very steady around 0.93:
http://www.esrl.noaa.gov/gmd/grad/mloapt.html
They would seem to show that there has been no measurable change in atmospheric transmission, at least in the Northern Hemisphere, during the “pause” unless there was been som major change in measuremet procedures at exactly the right point in time.
Of course they may still be right that aerosols have been underestimated, but in that case they have been underestimated before 2000 as well.
It’s worse than that, tty. Even when the Mauna Loa data shows a drop in atmospheric clarity there is no corresponding drop in temperature. See my post called Volcanoes Erupt Again for a full discussion.
w.
PS As a matter of fact Figure 2 above matches the Mauna Loa record remarkably well, suggesting that the normal atmospheric transmission with “background volcanism” is 0.92-0.94.
This is exactly right, see actual data and an actual fit with a tiny slope to the whole data set including volcanism and a hand-drawn zero-slope line fit to split the equilibrium to which the system returns at 0.93. The other line is basically 0.92, so 0.93 \pm 0.01 is very, very close to being right.
And as Willis says, there is at most a weak drop in temperature in association with major changes in stratospheric clarity — too weak to take seriously. Willis thinks (I think, he can correct me as I do not mean to put words in his mouth and I’m trying to repeat the outcome of our earlier discussions of this issue) it isn’t there at all. I think it is there, but with a transient effect so short as to be ignorable and so small as to be easily overwhelmed by natural dynamics and hence erased. But we completely agree that whatever the response, it can be completely ignored in climate models without doing any real harm, at least out to Tambora scale events (where we lack good global temperature data to work with but anecdotally VEI 7 had an effect on weather for more like 5-7 years — “almost” at the level of climate impact — and had an immediate and profound effect on the weather for 1-2 years).
In other words, the actual historical data almost certainly completely refutes an assertion of a weak 0.3 C reduction in observed warming over almost 20 years due to any sort of alteration of stratospheric transmittivity from any or all causes, mostly because Mauna Loa is a direct observation that no such alteration of stratospheric transmittivity in fact has occurred.
I don’t know how anyone can correlate a flat line plus or minus a relative 1% of pure annual timescale noise with a supposed cancellation of warming due to something that is changing substantially over the same time frame, e.g. CO_2 concentration. But hey, some people will try anything to hang onto a broken hypothesis.
rgb
rgbatduke January 10, 2015 at 11:32 am
Actually, I agree with you, Robert, and have said so before. Very large volcanic eruptions have a minor transient effect on surface temperatures, which is local rather than global, and which is so small as to be overwhelmed by other fluctuations.
w.
My bad, Willis. Sorry, I do get confused about where we agree and were we disagree — I think I was remembering our also comparatively weak disagreement over whether or not one can model a smooth CO_2 rise from 1850 to the present to extrapolate over the region where we have at best inconsistent and slightly dubious ice core data. In the end, I only trust Mauna Loa data, and even that only somewhat provisionally, and ML data is very, very smooth and systematic even with just annualized averaging. Not perfectly smooth, but it shows none of the bumps that the ice core data do (and which I seriously doubt as being globally representative, even if they do represent the local not necessarily well or correctly mixed conditions that ice cores supposedly capture).
But in that case — YEAH, so we BOTH agree that really big volcanoes (or smaller big SO_2 emitters, sometimes) may have a really small impact on temperature that lasts a really short time, more or less indistinguishable from other sources of natural variation e.g. ENSO. In fact, often much smaller than ENSO events, even for very large volcanoes indeed.
Volcanoes can probably simply be ignored, although I wouldn’t rule out lower troposphere effects like modulation of albedo through cloud nucleation from enough even small VEI SO_2 emitters — all we can really rule out via Mauna Loa transmittivity is any stratospheric connection. But that’s a very different story and even that story is less compelling than the probable temperature CO_2 concentration correlation as far as the data only is concerned. I just don’t think that we have a particularly good handle on the science yet, especially not if we have people that should know better than asserting a stratospheric connection that is actively contradicted by prior direct observations doing it anyway while ignoring the much simpler albedo connection.
I need to get the TOMS data from the source, though — I don’t want to screen scrape it out of a log plot. It does open up more possibilities for a volcanic connection than just VEI, even though I’m sure that we almost completely lack a knowledge of it even by proxy before 1979. Sadly, the Global Sulfur Dioxide page has data that is so raw as to be useless, graphically encoded. I have no good idea how to get a simple table of annualized SO_2 emissions per volcano or per planet, with or without including human contributions from high sulfur coal. It does look like NASA has very good instrumentation for detecting it, though, even in eruptions that humans would find unimpressive.
rgb
Yep, total rubbish by people who know zero about volcanoes.
Every volcano erupts slightly differently from each other and each different volcano can erupt differently at each eruption.
There are an estimayed 3million volvsnoes on the planet, most of them under the ocean surface. One thing ALL volcanoes emit is CO2.
And they are hot hot, our earth on fire, our oceans on fire feeling hot hot hot
Really? And you know/estimate this how? There are 1500 “active” volcanoes on the land surface, around 500 of which are actually active at any given time. A naive estimate would extrapolate this by a factor of (say) 4 since the ocean is 70% of the surface to get 2000 active volcanoes worldwide, where by active I mean actually erupting. Allowing for this to be wrong by a full order of magnitude would give you 15,000 (the 500 on land don’t change as we’re pretty certain about that, only the 1500 in the sea). Two orders of magnitude would be 150,000. Three orders of magnitude would be 1,500,000. And you’re asserting another factor of two, to end up with 3,000,000 (where the 500 on land are then pretty much irrelevant).
So you’re basically saying that for every actually erupting volcano on the land surface, there are 2000 hidden, erupting volcanoes on the sea floor. Sure, why not? Or why not 10,000! Or 200. Or 10.
Do you have any actual data to back up that “estimate”?
rgb
There are a lot more volcanoes under the oceans than there are on land. But they are largely an unknown. I hate to link to NewScientist, but here it is:
The true extent to which the ocean bed is dotted with volcanoes has been revealed by researchers who have counted 201,055 underwater cones. This is over 10 times more than have been found before. The team estimates that in total there could be about 3 million submarine volcanoes, 39,000 of which rise more than 1000 metres over the sea bed.
For a more technical explanation, see here.
Dr. Brown is correct. We just don’t know the extent of underwater volcanoes, or for that matter, how much CO2 they add to the atmosphere.
Note well that finding underwater cones is irrelevant compared to land, as well. On land, rain and wind produce an ongoing process of weathering so that “cones” simply don’t last. The 1500 active volcanoes is the count of the “cones” that they think have erupted in the last 15,000 years, and of these, only around 500 are ACTUALLY active as in emitting magma, ash, gases etc. There are plenty of much more ancient cones that aren’t included in this count, and in all probability many that one would count if there were no weathering have simply vanished, or been tipped sideways and turned into something else.
On the ocean bottom, there is no wind and no rain. There is almost no current, and what current there is moves at a snail’s pace. It is cold. There is little chemistry to cause weathering. Cones last a long, long time.
So the big question is not how many humps are there in the bottom that could conceivably be identified as being leftovers from volcanoes. It is how many of these lumps are active by either of the related definitions on land — an eruption in the last 15,000 years, or actively erupting now.
I would think that mere seismology could rule out 200,000 significantly erupting sea bottom volcanoes. There are whole books describing the signal of volcanic eruptions from VEI 1 and 2 on up. I would expect that we could hear 200,000 active sea bottom eruptions with an absolutely unmistakable signal, picked up by the navy sound detector grid that is sensitive enough to locate submarine screws and whale calls at tens to hundreds of miles.
At subduction zones or spreading rifts, I’m perfectly happy to think that volcanoes might be more common than on the plate boundaries on land, if only because one expects the crust to be somewhat thinner there (especially at rifts). I’m equally certain that all sorts of interesting stuff happens down there (some of which we are slowly sorting out). But three million, active volcanoes is a bit difficult to swallow with only 500 or so on the 30% of the surface that is visible.
Finally, it is important to differentiate cones from volcanoes per se. Mauna Kea has over 100 cinder cones on its flanks, but is really only one big volcano. Many cinder cones are monogenetic, which means that if you find them you can be reasonably sure they WON’T erupt again, they don’t count as active in any useful sense of the word now or in the future. The United states is covered with weathered, ancient volcanoes. New Mexico is rife with them (I’ve climbed a number). Texas too. Even Maine has its share. So when the land count of actively erupting volcanoes is held to be around 500, they are not counting cones or things that at one time were volcanoes, weathered or unweathered. They are counting things that are actually spewing magma and/or ash.
How many of THESE are there on the sea floor? I have to say that I find 3,000,000, or even 300,000, really really unlikely. Even 30,000 seems a bit of a stretch, but perhaps there are some very active rift or subduction zones. But geological time and frequency are not the same as human time, and even “active” zones are probably marked by comparatively few actually active cones. The active cones actually take the pressure off of the surrounding area and make violent eruptions there somewhat less likely, at least one would expect.
rgb
You forget that the oceanic ridge system is 53000 miles of volcanoes. two places on land show the extent of volcanisity, Iceland and Ethiopia. These two places are quiet compared to the ridges. Every sea mount is an extinct volcano which still emit CO2 a so called pollutant.
The estimate of 3million is not mine but appears on the excellent website
http://www.geologist-1011.org
The estimate is based on sea floor surveys not wishful thinking.
I go back to the basis of the blog that the claim is not based on data of current eruptions but wishful thinking and straw clutching.
“The relatively muted warming of the surface and lower troposphere since 1998” I cringe every time I read statements like this. It states that there has been warming since 1998. But for there to be warming the temperature must have risen since 1998. If the temperature in my living room has stayed the same since 5PM, do I say my living room has been warming since 5PM? NO! I say my living room is still warm or has not changed or has not been warming since 5PM! It may be warmer then before 5PM but it is not “warming”. The abstract has a flat out demonstrable lie in it!
“in the context of the post-2000 slowing of the rate of global warming” Ok post 2000 you may be able to draw a statistically significant warming trend, maybe, if not it too is wrong. My house warmed by 2C/hr from noon to 5PM, from 5PM to present the rate is 0. Do I say from 5PM to present the rate is slowing? Slowing as in slower but still moving up. Or in common English, from 5PM to present the rate has stopped?
These are educated people who most likely put a lot of thought into the wording of the abstract. How can one believe anything that follows when they are not honest enough to admit the warming has stopped?
The worst thing, for me, is that I’m not even mildly surprised anymore. I know I should be angry, but it’s more like watching the baddies in an old movie: The plot may be ridiculous, but I only really notice how bad the acting is.
I’m glad that other people can still get riled by it.
Couldn’t stay up late enough to catch this early on but,
How can these perturbed homo sapiens justify or even explain what they do to their parents? We went into major hock to ship you off to a no longer ivory tower – tusk pollution maybe, virtual poaching it is – but I could be wrong. This could be totally on-line peering and it is cheap because we don’t even need that lab anymore.
I am only relieved that Bernie Sanders and Peter Shumlin would be completely incapable of reading this, not Willis, bs, not even capitalized, baffling brains.
The state of education is gravely sad. Free community college for gpa=2.5!!! Baring grade inflation, let’s have some standards or launch a major distributed denial of service to academe.
whenever I see her [Solomon] name on a study, I’m sad to report that I just wince.
Is that a proper way to show respect for a Nobel Laureate? /s
http://news.boisestate.edu/update/2012/01/24/distinguished-lecture-series-presents-nobel-prizer-winning-scientist-susan-solomon/
A quote from your link Johanus ,
” Her work connecting volcanic chlorofluorocarbons to increased damage to the ozone layer formed the basis of the Montreal Protocol, an international agreement to protect the ozone layer by regulating damaging chemicals”
Soooooo, chlorofluorocarbons are formed naturally from volcanoes and now the same volcanoes are causing global cooling……so is the ozone hole getting bigger? Hmmmm….I guess we better start capping those volcanoes or do we need more of them to cool the planet?!?! Mah!!!!
Isn’t it obvious? With more than 50 factors shown to be contributing to the ‘pause’ the effect of each is tiny when considered in isolation – that’s why we can’t see it in the graphs – like all the best climate theories it can’t be measured.
My biggest problem with all of these “explanations” of the missing warming is the question of coincidence.
Given that CO2 has been rising on pretty well a “business as usual” basis for the time in question, any of these explanations – or any combination of them – must have acted to more-or-less exactly offset the additional CO2 forcing without having any physical basis for doing so. That’s a coincidence and a half over nearly a 1/5th of a century.
Of course, I’m at risk of falling foul of the warning about correlation not showing causation here, which is really another way of saying that coincidences can, and do, happen.
But since virtually the whole of the AGW position is based on correlation (with the one physical exception that CO2 does absorb IR) , they should NOT be satisfied with any explanations that rely on coincidence to explain such exact correlation in the deviation from warming! ,
Joe, correlation does not prove causation, true.
But absence of correlation disproves causation. Also true.
And that is the problem that a temperature pause despite rising CO2 gives the warmunists. Hence the increasingly laughable desperation to find alternative explations nicely diced up by Willis here.
Agreed, Rud, which is sort of an extension of my point.
What they’re essentially saying is that there would be correlation between CO2 and temperature over the pause if it wasn’t for this huge coincidence of unrelated factors that just happen to be exactly offsetting it.
Now, to my simple mind, if coincidence can explain the magnitude of that unrelated offsetting then it can just as likely explain any of the (far less exact) correlation they’ve “found” in the earlier data.
Nice post Willis!
We have listened to them telling us that the heat has disappeared into the oceans (an explanation that defies common sense!) and that now the warming has stopped due to the activity of vertically challenged volcanoes!
The warming stopped around the time when satellites were able to produce data that was accurate and could not be cherry picked, so to me the simple explanation is that the computer models are wrong!
Errr yes maybe
The possible volcanic contribution to atmospheric C02 levels was discounted by “warmists” but now they want to blame the aerosols from the same [previously non-existent] volcanoes for perturbing their models.
And where is the evidence for a late 20th /early21st century increase in volcanic activity ??
Dear me
So, if these guys are right, climate science has under-estimated negative forcings…
Yes. You nailed it, Steve.
Meanwhile, back at the ranch, the rate of earthquake occurrences in the US southern central plains seems to be increasing, with 10 quakes in Oklahoma of mag 2.5+ in the past 24 hrs, so maybe there really is an uptick in gas release from all geo- tectonic sources.
“Willis’s Rule of Author Count ” sounds like it is related to the new consensus approach to science.
The idea seems to be that there is strength in numbers. So instead of relying on the quality of the research and conclusions, you go with the largest number of authors you can get to sign it. This happens in Congress quite often and we all know well Congress works..
There is a plausible mechanism to explain “Willis’s Rule of Author Count, that the quality of any study goes down inversely proportional to the square of the number of listed authors.” The larger the number of authors, the smaller the number of outside experts remaining in the particular field of study who are (a) competent to review and (b) free of conflicts/close associations with one or more authors.
On a related note, would a freedom of information act request be applicable to Livermore personnel to be able to see the peer reviewers’ comments, questions and suggestions for improvement? If so, that would certainly reveal something about the strength of the review process itself.
When I look at the data I see the exact opposite of what is claimed in this paper. keep in mind that the pause is really more like the top of a sine wave. The warming continued to around 2005 and then cooling began.
http://www.woodfortrees.org/plot/rss/from:1996.75/plot/rss/from:1996.75/to:2005/trend/plot/rss/from:2005/to/trend
Looking at the data, it was right after 2005 that the number of small volcanoes was generally reduced. Hence, one might argue the hiatus is being caused by a lack of volcanoes, not more of them.
Of course, that would also mean that the warming itself was the result of more small and large volcanoes and not CO2.
Just sayin …
Or, one could look at the data clearly and use R and conclude that it is a waste of time to even be distracted by volcanoes when considering the climate. We have good data on top of troposphere insolation for over 60 years! We don’t need to build complex theoretical GCM models involving aerosols — we can directly measure their effect on the transmittivity of the stratosphere, and look for the sensitivity of the actual climate to the major boluses of aerosols in model fits. And there is none.
Basically, the 60 year average top of troposphere insolation is totally flat, with the mean transmittivity of the stratosphere being around 0.92, assuming (as always) that the stratosphere is on average well-mixed so that Mauna Loa is a reasonable representation of the rest of the stratosphere, allowing for geometry. The equilibriuminsolation is around 0.93 and again, flat to well within our ability to resolve a slope over all 60 years. Major volcanoes — VEI 5 and up — can have a large (5%!) transient effect on top of troposphere insolation but even this large a variation in the Earth’s annual incoming energy budget, sustained for at least a year, has almost no statistically resolvable effect on the temperature and what effect it might have has a lifetime of less than 3 years and is hence arguably an impact on the weather, not the climate. Even major volcanoes are completely negligible in terms of climate impact compared to, say, El Nino events.
So let’s not argue that they are causing warming OR cooling, because at the top of the troposphere insolation simply isn’t varying at all so trying to attribute climate change to changes in broadband radiation due to reflection off of the stratopheric mix of gases is a rather waste of time. See the actual figure below.
Which, incidentally, does a pretty darned good job of invalidating all of the world’s GCMs, BTW, as they rely on substantial cancellations between aerosol-driven cooling and CO_2 driven warming to fit the reference period, which is odd given that Mauna Loa shows that — outside of two volcanic events that had almost no impact on temperatures (certainly an impact far smaller than natural variation) there is basically no change in equilibrium aerosol-driven cooling or warming, or anything at all over the last 60 years completely embracing the reference period used by the CMIP5 models displayed in AR5.
One can do a rather excellent job of fitting global temperature in a model that ignores aerosols altogether. For good reason — whatever effect they might have appears to be cancelled by negative feedbacks and/or swamped by the enormous buffering potential of the ocean, and the stratospheric reflectivity seems to be rather amazingly stable. I look for things reported by satellites regarding e.g. major spikes in its water vapor content that supposedly affect global temperature and see nothing of it in Mauna Loa data.
rgb
RGB wrote: “The equilibrium insolation is around 0.93 and again, flat to well within our ability to resolve a slope over all 60 years.”
As best I can tell, newer instruments and technology are showing it isn’t perfectly flat in the 2000s. Unfortunately, we don’t know how much change (forcing) has occurred. The insidious part is that an ambiguous change has lasted for more than a decade and transient temperature change is the product of forcing (W/m2) and time. At its peak, Pinatubo produced at least 30-fold more stratospheric aerosol than the weaker volcanos in the 2000’s, but Pinatubo’s were mostly gone in a few years. The advocates of some volcanic cooling during the hiatus assert a fairly constant, but weak forcing for more than a decade.
Frank January 11, 2015 at 2:49 pm
Frank, while you may well be correct, what you are missing is a sense of scale. Here is an overview of the measurement used in the Santer study, the stratospheric aerosol optical depth (SAOD):
Yes, the newer instruments can show us trivially small variations in the post-2000 data … but so freakin’ what? The variations that Santer et al claim are significant are on the order of .002 … look at the graph above to see how ludicrously small that is. Below are the Santer results. You can see that they are just an extreme closeup of the right side of the graph above.
w.
Willis wrote: “The variations that Santer et al claim are significant are on the order of .002 … look at the graph above to see how ludicrously small that is.”
Frank replies: Finally something we CAN agree upon. [There has been an unusual cluster of VEI = 4 volcanos since 2002.] Those volcanos may have caused a barely detectable increase in stratospheric aerosols compared with 1995-1999. We certainly don’t know if that is an increase over normal levels of aerosols. To define a forcing, one needs to know the amount of the agent during the 2000’s AND before (usually pre-industrial for anthropogenic forcing). Unfortunately, integration of this weak, perhaps nonexistence, forcing over a decade produces a partial “explanation” for the pause. (Forcing is power, temperature change is proportional to energy or power * time.)
What happened to the confidence intervals for this forcing/temperature change? They aren’t being shown because the 90% or 95? ci almost certainly includes ZERO! The sensible thing to criticize about this work is the possibility it is derived from changing our changing ability to measure stratospheric aerosols and/or noise in the data.
Boy, looking at your graph 1 and graph 3 (graph 2 doesn’t matter, they were looking at small volcanos), it looks like volcanic activity is higher in the naughts than in the 90s. Why don’t you put a 60 month average on your charts, or is that why you didn’t?
Uh? Are you actually capable of reading?
A surgical takedown and dismantling of yet another excuse for the “global warming hiatus.” Thank you and well-done Willis.
Thanks, Willis. Good analysis.
They are investigating the wrong metric. The relationship between cooling and volcanic loading has less to do with the overall VEI metric and more to do with the amount of volcanic sourced sulfur compounds that enter the stratosphere. Local cooling related to ash is exactly that, local cooling, and is rather short lived. Sulfur compounds injected into the stratosphere circulates on a global scale and can indeed induce global cooling and is best measured by the amount of these compounds found in polar deposits. But even that metric alone is short lived. Of more importance is the juxtaposition of volcanic stratospheric events with ENSO conditions, specifically related to conditions that allow equatorial recharge of oceanic heat stores. This incoming solar heating, or lack thereof due to veiling in the stratosphere, most definitely affects global temperatures past the point in time that veiling clears up.
So why do I say this? Because I know that clouds also affect solar heating in the equatorial band and can be modeled. Stratospheric veiling can also be modeled in much the same way. These less heated pools of water then ride oceanic currents, leading to changes in weather patterns on a global scale.
Changes in ocean heat is the main source of weather pattern variations. Therefore any changes in ocean heat will change weather patterns. Sulfur loading will indeed change incoming solar insulation thus will decrease oceanic uptake. This is the proper logical way to understand volcanic forcing on a global, and longer term scale.
The present studies are at too fine a scale and focus on events that hardly rise above background noise in the stratosphere related to sulfur compounds.
Basaltic (chemically “basic”) eruptions, as in Hawaii less explosive making mainly coarse ejecta from sand-like lappili tuff, to spindle shaped bombs often of many kilograms in mass (molten blobs shaped by spinning in the air). These are typically benign as far as residence time in the atmosphere. Not so benign if hit on the head with even a small one, but if it happens, you can rest assured that it didn’t occlude the sunshine for more than a minute or so!
The acid volcanics and intermediate types are high in silica, are viscous and very gassy with the usual suite of noxious gases, SO2, CO2 and lots of water. The force of the explosions disintegrates the brittle silicious foamy rock and blasts it often up into the upper atmosphere. Because even the -50 micron pieces are riddled with gas bubbles, they are very light and can be born along in the upper atmosphere for long distances and even easily re lofted by thermal updrafts when they begin to settle.
http://imgarcade.com/1/volcanic-ash-microscope/
WILLIS, the acid volcanics are continental crustal types. It would be good if it were possible to sort the volcanoes by type Acid vs Basaltic to see if we have had more acid volc. If not, then its Q.E.D. for your investigation results.
It’s very simple.
There are volcanoes hiding where we can’t find them. Probably underground…
Willis here is something you should think about and the term solar activity can be replaced with any climatic force one may want to use.
Yes solar activity is always influencing the climate but at times of no solar extremes (solar flux 90-150 range) the solar contribution is going to be lost to noise in the climate system.
Also the mean state of the climate/earth dynamics at the times of solar forcing are going to cause given solar conditions to result in different climate outcomes. This statement applying to all items that may exert a force on the climate.
In summary I think climate sensitivity to various forcings is EXPONENTIALLY dependent upon the mean state of the climate/earth dynamics at the time the forcing is taken place which is why correlations are hard to come by and so many different climate outcomes (although a similar trend in a general sense) is always the result.
The climatic outcomes all different over the past 20000 years while the same essential forces have been in play throughout this time frame I think proves my point. This is why no climatic theory no mater how could fails to stand the test of time.
.
Last sentence should read . This is why no climatic theory no matter how good fails the test of time.
One last thought due to the no reaction or hostile reaction I got from various places.
It is amazing that when I put out something like I did yesterday about how climate sensitivity to given forces depends upon the mean state of the climate and earth dynamics and changes exponentially due to those factors that I was either ignored or treated as if it just is not so, when it is so in my opinion.
The reason why I think I am getting this reaction is it destroys all of the various theories in trying to explain why the climate changes in the way it does due to some force because it takes away the notion that one can come up with a climate prediction based on that premise.
I think strongly that all the various climatic outcomes going back just 20000 years never mind further with essentially the same forces present then as now no matter what the choice of the forcing goes a long way in proving the point I am trying to make.
So I say to all the climatologist out there knock yourself out in trying to prove your theory is correct because it is not going to result in the result you may want and all one has to do is go back a sufficient amount of time to show how every climate theory no matter how good does not stand the test of time.
At best maybe general climatic trends can be forecasted and that is what I am going for.
And here is yet another bit of evidence to support Willis argument:
change in transmittivity per year. Yes, this is negative, but the uncertainty is larger than the value. The only effect of the eruptions is to pull the best fit line well down from what is obviously a dynamic equilibrium line that describes the atmospheric transmittivity in the almost perfectly constant background forcing from all sources — those small-to-medium volcanoes, human emissions, changes due to solar activity, and the effect of space alien rays being beamed down from cloaked transmitters as they try to terraform the earth to better match their needs before invading us (Look, I can do conspiracy SF theories too, Ma!).
http://www.phy.duke.edu/~rgb/mlo-tran.jpg
This is a direct graph of the atmospheric transmittivity as measured at the top of Mauna Loa (which means very close to the top of the troposphere). This is a direct measure of “just” the effect of volcanic aerosols and water vapor and other optically coupled gases on total stratospheric transmission, with 1 – transmission being effective reflection. One is above most of the water vapor, clouds, soot, and other human-induced stuff.
There are two things worth drawing your attention to in this graph. The first is that the major volcanoes — VEI 5 and up — have a clearly discernible effect on TOA transmission. I have a variation of this graph where I label the three substantial divots in transmission in the 60’s, the 80’s and the 90’s with the major eruptions that produced them. VEI 4 and under eruptions have almost no discernible effect, even a transient one. There are a couple of spikes observable in the late 90’s and close to 2010 — but there are no specific eruptions one can attribute these two as all of the significan eruptions in the 2000’s are 4’s, and 4’s produce almost no spike at all unless they are incidentally sulphur-rich eruptions (which happens, BTW, and is a variable almost independent of VEI). To illustrate the problem, Mt. St Helens, a 5, produced no visible spike at all in transmission, but sulphur-rich El Chichon at 5 produced a spike almost as noteworthy as the double whammy of Pinatubo and Mt Hudson at or close to 6. If you look for correlations between these spikes and global temperature dips, if you squint and look a bit cross-eyed then one can convince oneself (and R) that there is a weak, transient correlation between VEI 5 and up eruptions and temperature, but it is more difficult than seeing the 3-D unicorn in a dot stereogram because the effect is comparable to the natural noise and trend and is often confounded by (cancelled by) that noise.
I think that both Willis and I are in agreement that the effect of volcanoes of VEI less than 6 is almost completely irrelevant to the climate, and one could argue that even 6’s have a pretty ignorable, highly transient effect on long term trends, with any effect they produce completely disappearing within 1-3 years. That is, volcanoes by definition have no effect whatsoever on the “climate” if we define climate in terms of decadal or longer trends, but they can have an effect on global weather. Obviously this would not be true if e.g. Yellowstone or the Siberian Traps erupted as a supervolcano — a major sulphur-rich eruption that basically never stops on a century time scale — but Pinatubo? Krakatoa? Fooey. I spit on your Krakatoa as having more than a transient effect on the weather, not the climate.
The second thing is that I’ve fit a linear trend to the mlo data over the entire interval (to avoid cherrypicking) with R. Note that with the big divots, this makes R sad — R-squared is basically 0 because the data is not, not, not linear with those huge nonlinear holes. But aside from this, R finds that the data has no linear trend — the best fit slope is, well statistically indistinguishable from zero, weighing in at
Indeed, it is perfectly obvious that on average, the broadband transmittivity of the stratosphere has basically not varied over the entire time frame of the Mauna Loa data. That makes looking for a correlation between volcanism and the change in this transmittivity rather moot, because it has not changed in any statistically resolvable way for almost 60 years. To put it bluntly, it is like looking for the effect of stock market prices on the mass of a brick. The mass of the brick doesn’t change much, right? In fact, some would say that it doesn’t change at all. You know that even before you look at the stock market data. So why would one publish a paper suggesting that the generally rising stock prices represented by the Dow Jones are causing the mass of the brick to decrease, which is why you should worry about your house floating away if the market ever really takes off?
Well, OK, there are three things in this graph that are worthy of note. The third thing is this. The mean transmittivity of the stratosphere is (according to R, again) — around 0.92, INCLUDING the fortunately balanced spikes from major volcanoes. Obviously, the equilibrium to which transmittivity regresses is a bit higher than this, so I drew a dashed line by hand at 0.93 that does a pretty good job of splitting the noise on what appears to be the nearly completely insensitive equilibrium transmittivity — the data isn’t worthy of an better treatment as it is pretty obviously flat outside of the major divots and their sort-of exponential return.
Now look. Both El Chichon and the Pinatubo-Hood double drove top-of-troposphere insolation down well below 0.9 — arguably down to maybe 0.88 — for at least one full year. Let’s put that in perspective. TOA insolation is roughly 1300-something watts/m^2 — it varies by 91 watts/m^2 over a year so there is little point in doing much better than that in this argument. Mauna Loa measured a drop in insolation at the top of the troposphere of five percent that lasted for a full year. That is, top of troposphere insolation was reduced by a staggering 68 watts/m^2 (transiently, much more!) for at least one full year.
Let’s consider what effect this should have had on the Earth’s temperature if any of the naive high-sensitivity models were correct. Over all of the Earth’s land surface area, temperatures are dominated by short-timescale relaxation processes. It warms up right now when the sun shines. It cools down quickly at night, often by as much as 20 to 40 C. If you drop insolation by 5%, you should see profound, immediate cooling, worldwide, in the entire land surface area. After all, the total change in forcing expected from doubling CO_2 is around 5 watts/m^2 — this is an order of magnitude greater (although still smaller than the annual variation with its anticorrelation with global surface temperture).
Yet the actual observed effect is almost invisible against the noise. Here’s my own solution to Willis’ “Spot the volcano” game, played with R and an ad-hoc model based on fitting the Mauna Loa transmittivity divots to converted average VEI (using the same method he describes above, but including only 4’s or higher as 3’s and below are truly just noise when even 5’s have almost no effect on temperature). First the recent time spanned by the actual ML transmittivity data:
http://www.phy.duke.edu/~rgb/cCO2-to-T-volcano-recent.jpg
Can you spot the volcanoes? Note well that I really do try to help you by providing a best-fit prediction of their effect, and in the thick black line you can see the early 60’s volcanoes (the Tsar Bomba, which was effectively a 5, Mt Agung 5, and Shiveluch 4.5 but included as honorable mention), Mt St Helens/El Chicon, and Pinatubo/Hood. Can you resolve the signal from the noise? I sure can’t. As I said, if you squint a bit you can maybe convince yourself that Pinatubo caused the small, slightly lagged blip in temperature immediately following, but I count at least 9 other downward fluctuations in temperature with almost exactly the same appearance, magnitude, and duration and none of them are correlated with, well, anything. That’s the point of Willis’ game — if you were given the data and told that there were two events that dropped top of troposphere insolation by 5% for over a year and you were to make your best guess as to what years they were, there are at least 4 years you would guess before getting around to Pinatubo. Basically 100% of all people, armed with the entire statistical panoply of R or just their own eyes, would get it wrong. You’d have a better chance of getting it right if you labelled the divots in temperature with numbers and rolled dice to guess.
This is precisely how our minds and biases trick us in to thinking that there is a causal correlation between events. We somehow focus on the local correlation and deselect all events that don’t fit the pattern with our brains and then unsurprisingly find a good correlation within the selected subset! It is why the gods of statistical reasoning get angry when you pick cherries to make cherry pie, as you are likely to conclude that your pet hypothesis that all red berries are good to eat because cherries are is true, and poison a whole lot of people when they try to eat any of the numerous poisonous red berries available in the wild.
Let’s play Spot the Volcano one last time, this time for the entire range of arguably good data represented by HadCRUT4, using this same model (built with 60 years of Mauna Loa data, and now hindcast into the past for validation):
http://www.phy.duke.edu/~rgb/cCO2-to-T-volcano.jpg
I’m not going to list the minor events, but the volcano data contains four clearly discernible VEI 6’s: Krakatoa, Santa Maria, Novarupta and Pinatubo (in temporal order). Each of these “should” have produced a substantial divot in global temperature that one would expect is resolvable from the noise and natural variation. Again, if you let your brain only look at the local correlation between the predicted divots and the events, you can convince yourself that there was a sudden cooling in 3 out of the four cases. But now invert the question. If you were told that there were 4 VEI 6 eruptions that occurred during the timespan of this data and asked to make your best guess as to when they happened and how long afterwards they affected temperatures, 100% of all rational guessers would guess wrong. I doubt that given all four chances independently that people would get even one guess right, although a few might hit Krakatoa or Santa Maria. But Novarupta? Temperatures rose strongly all the way through Novarupta, which is the kind of thing that gives cherrypickers migraines as they try to resolve the cognitive dissonance between their pet hypothesis and something that more or less directly refutes it. Pinatubo? Don’t make me laugh — people wouldn’t even guess a single event from the 1980s on given the wealth of sharp drops in the 50’s and 60’s and elsewhere in the record. And even Krakatoa and Santa Maria are far from the most rapid or disruptive of coolings in the record.
To conclude, it will be very difficult for me to be convinced by Santer et. al.’s paper, because no matter how hard you try, you aren’t going to succeed in convincing me that there is a strong causal relationship between leaving a brick out in the sun and its mass, even with a perfectly good physical basis for that belief. I even believe the truth of the assertion — a sun-warmed brick will almost certainly have slightly more mass than the same brick, cold. The problem is measuring that difference in a 1 kilogram brick, because hey, come on, the mass of the brick doesn’t change on any measuring apparatus you are likely to be able to afford to build or apply and even if you do a really careful job and confirm (again) relativity theory, the variation is so small compared to the total mass of the brick that nobody cares.
That’s what the graphs above ever so clearly show. Sure, why not, maybe it is true that volcanoes weakly affect weather (but not climate). They certainly “should”. R can — with a bit of help — manage to fit a model that explains a tiny part of the total variation in global T slightly better than one that omits volcanoes altogether. But what we learn from this is that in the end we do not care if it is technically true, because it is a nearly irrelevant effect compared to the other, much much larger factors driving the climate.
We learn one last thing. The Earth’s climate is awesomely stable. Every year it manages to countervary average global temperature and a 7% annual variation in top of atmosphere insolation. It manages to shrug off 5% sustained variations in top of troposphere insolation with anywhere from a response the opposite of what one expects (Novarupta) where any cooling is swamped by natural warming driven by things we Do Not Understand to no effect at all (El Chichon, and actually many other VEI 5’s in the climate record) a mere 1-3 year response, at best, to the three VEI 6 events in the climate record that one can squint hard and convince one’s self are relevant to transient changes in global weather (but that have no effect at all on the trajectory of the climate.
rgb
Willis’ author rule has a software analog. The more developers working on a software project, the worse the quality. The is partly due to the more challenging communications, and the difficulty of making more and more separately written pieces work well together. It is also because poor and incompetent programmers can only be found in large groups because they cannot survive elsewhere, which also causes good programmers to avoid large groups when they can choose to do so. Finally, the group dynamics of peer acceptance and attempts at group harmony mean that stupidity is not challenged in larger groups.
I believe these forces are also at play in science groups.
The quote from the Tao of Programming:
Oops, sorry mod — could you remove the spurious spaces in the last two web links to graphs on my website? My mouse was apparently greedy by one space when I grabbed the text and they hence failed to expand inline.
rgb
Robert, as always your comments move the discussion forwards. As you point out, the MLO data is a clear refutation of their claims.
I’ve removed the spaces in your links, so the graphs are now part of your most interesting exposition.
Finally, I’ve never read “The Tao of Programming”, but if that is any example, the guy definitely understands both programming and the dynamics of large groups …
w.
It’s on my personal website, under “general”. I didn’t write it, but I wish that I did. It’s pretty funny and oh, so very true (and I say this as a professional coder, among my many and various hats).
rgb