Why Volcanoes Don't Matter Much

Guest Post by Willis Eschenbach

The word “forcing” is what is called a “term of art” in climate science. A term of art means a word that is used in a special or unusual sense in a particular field of science or other activity. This unusual meaning for the word may or may not be logical, but each field has its terms of art, and it’s useless to complain that they don’t make sense. The IPCC defines “radiative forcing” as follows:

Radiative forcing 

Radiative forcing is the change in the net, downward minus upward, radiative flux (expressed in W m–2) at the tropopause or top of atmosphere due to a change in an external driver of climate change, such as, for example, a change in the concentration of carbon dioxide or the output of the Sun. Sometimes internal drivers are still treated as forcings even though they result from the alteration in climate, for example aerosol or greenhouse gas changes in paleoclimates.

Now, the current climate science paradigm says that regarding the things that affect temperature, everything averages out in the long run except for any changes in total forcing. The current paradigm further says that the future evolution of the climate can be forecast by the simple linear relationship given as:

Change in temperature equals climate sensitivity times change in total forcing.

Me, I think that’s simplistic nonsense, but let’s set my opinion aside for a bit and compare their forcing claims to the actual observations of the changes in forcing. As an example, let me use the forcings that are the result of volcanic eruptions. The larger eruptions blast aerosols (various molecules and minerals) into the stratosphere, reducing the incoming sunshine. These forcings have been estimated by Sato  as being of the following amounts:

Figure 1. Volcanic forcings estimated by Sato et al.  http://data.giss.nasa.gov/modelforce/RadF.txt  Forcings are negative because they represent a reduction in available solar energy due to volcanic aerosols. The large eruption at the far right is Pinatubo in the Philippines, 1991, and the eruption to its left is El Chichon, Mexico, in 1982.

You can see that some eruptions, like that of El Chichon, produced a much larger aerosol cloud in the northern hemisphere (red) than in the southern (blue), while others like Pinatubo were more equal in the distribution of the aerosols between the hemispheres. In all cases, the hemisphere where the volcano is located shows the greatest effect from the eruption.

As I mentioned above, I think that the idea that the temperature slavishly follows the changes in forcings to be a fundamental misunderstanding of how the climate system operates. Instead, I say that although initially the temperature responds to the forcings, soon the climate system responds to the resulting changes in temperature by changing the forcings themselves, often in very non-linear ways. In particular, I have presented plenty of evidence that the climate system responds to increasing tropical temperatures by varying the timing and strength of the daily emergence of the cumulus cloud field. Part of the climate system response works like this:

On warmer days, the emergence of the tropical cumulus cloud field is both earlier and stronger. This cuts down on the available solar energy by reflecting more of it back to space. The high cloud albedo means that less sunlight reaches the surface, so the surface cools.

And on cooler days, the opposite occurs. The tropical cumulus field emerges later, and is weaker. As a result, the day warms up more than it would otherwise, because there is less cloud albedo and thus more available solar energy.

All that is required to show that this effect exists is to show that tropical albedo is positively correlated with temperature … as I have done here, here, and here.

Now, if we assume for the moment that my theory is correct, what kind of climate response would we expect to find from a volcanic eruption large enough to put aerosols into the stratosphere and cause some global cooling? Well, eruptions reduce available solar energy in two ways—increased reflection from white aerosols, and increased absorption from dark aerosols.

So the first thing to happen after the eruption would be the reduction in incoming sunlight from the increased albedo and increased stratospheric absorption. Then after the decreased sunlight actually starts to cause widespread cooling, the climate system would respond. We’d expect the climate system response following such an eruption to have the following characteristics:

• Right after the eruption, there would be a reduction in available solar energy, due to the volcanic aerosols in the stratosphere.

• This initial eruption-induced reduction in available solar energy would be both deeper and sooner after the eruption in the hemisphere where the eruption occurred than in the opposite hemisphere.

• As a result, the corresponding climate reaction in the eruption hemisphere would also both be deeper and occur sooner than the climate reaction in the opposite hemisphere. In other words there will be a dose-related effect, where a larger reduction is met with a larger climate reaction.

• The form of the climate reaction will be an albedo reduction, which will cause increase in available solar energy. The increase in available energy will be of the same order of magnitude as the corresponding decrease due to volcanic aerosols.

With those predictions derived from my theory about the nature and timing of the climate response, we can compare them to what actually happened when Mount Pinatubo erupted. I’ve taken the albedo records for the globe and for each hemisphere individually, and analyzed what happened after the eruption of Pinatubo in June of 1991. This gave me the anomaly in the amount of solar energy that is actually available to the climate system. Figure 2 shows three variables for the period 1984-1997, which includes the eruption of Mt. Pinatubo on June 15, 1991.

First, in black, is a closer look at the same dataset shown in Figure 1. Black shows the global average of the Sato volcanic forcing data for the period 1984-1997.

Second, in violet, is the aforementioned anomaly in the amount of incoming sunshine, in watts per square metre. This is the “available energy”, meaning the solar energy that remains after the albedo reflections.

Third, in gold, is the amount of incoming solar energy that is absorbed in the stratosphere. Recall that volcanoes affect the sunshine in two ways—changes in reflection (violet line) and changes in absorption (gold line). The gold line shows the reductions from absorption of solar energy by stratospheric aerosols.

Figure 2. Sato estimated volcano forcing (black), available solar energy anomaly after albedo (violet), and stratospheric absorption forcing (gold). The observed values (violet and gold) are expressed as anomalies around the value they had the month before the eruption. See below for methods and data sources.

Now, the first thing I noticed is that immediately after the eruption, all three datasets agree with each other—as we would expect, there is a precipitous drop in downwelling solar radiation. However, after that they go their separate ways, so it’s hard to tell what the overall effect of the absorption and the reflection might be.

For that kind of comparison, I use a running post-eruption average. This is the average forcing over the period from the date of the eruption to the date in question. So for example, the data point for January 1996 represents the average forcing from the date of the eruption until January of 1996. Figure 3 shows that type of post-eruption average applied to Figure 1, with the actual Figure 1 data shown grayed out in the background for reference.

Figure 3. Post-eruption averages. Total observed eruptive forcing [reflection (violet) plus absorption (gold)] is shown in yellow. Other colors as in Figure 1 — black is the Sato estimate of total volcanic forcing; violet is available solar anomaly after albedo reflections; gold is stratospheric absorption anomaly. Each point on the graph represents the average forcing from the eruption until that date.

The important thing to note is that from the eruption to the end of the record (end of 1997) the Sato forcing estimate (black line) has an average forcing of about minus one watt per square metre (W/m2). However, the observed change in total forcing of the period (yellow line, sum of purple (albedo forcing) and gold (absorption forcing) is a bit more than plus one watt per square metre.

Also, the speed of the climate response is visible in Figure 3. The total forcing (yellow line) follows the Sato forcing estimate (black line) for the first four months or so after the eruption. But after that, while the Sato calculated forcing continues to become more and more negative, the observations show that the total observed forcing does not ever become much more negative than it was at four months after the eruption. Instead, it runs level for about a year, and then rapidly increases. By the end of 1993, the observed post-eruption average forcing has returned to pre-eruption values … while the Sato theoretical forcing is still at minus two W/m2.

Now, Figures 2 and 3 show the global situation. We also have data for each hemisphere separately. This will let us observe the difference in the response of the climate in the two hemispheres. Here are the observed forcing and the Sato theoretical forcing for the northern and southern hemisphere.

Figure 4. As in Figure 2 but by individual hemisphere. The two panels show the Sato estimated forcing (black), the solar absorption forcing (gold), and the available solar energy after albedo (upper panel, red, northern hemisphere; lower panel blue, southern hemisphere)

The most notable difference between the hemispheres is the deep drop in available solar energy in the northern hemisphere (red line, upper panel) during the months immediately following the eruption. I note also that following that initial drop, the amount of available energy in the NH steadily increases in both the absorption (gold) and reflection (red) datasets.

To conclude this analysis I looked at the post-eruption averages for the individual hemispheres. Figure 5 shows those results:

Figure 5. As in Figure 3 but by individual hemisphere. These show the running average starting at the time of the eruption and moving forwards.

In the northern hemisphere we can see that the initial drop in forcing was almost as large as the Sato estimate. However, from there, the climate response kicked in, and the amount of available energy started to rise rapidly. In the southern hemisphere, on the other hand, the response was smaller and initially slower.

However, once the SH response began, the available solar energy rose very quickly. Both hemispheres took about the same amount of time, about two years, for the average forcing over the post-eruption interval to return to zero.

And in both hemispheres, the eventual response was nearly identical—the average change in total available sunshine at the end of the record is about plus a watt and a half per square metre, compared to the Sato estimate which has an average change to the end of the record of minus one watt per square metre.

Conclusions: The main conclusion that I draw from this is that the central paradigm of modern climate science is wrong—temperature does not slavishly follow the forcings.

To the contrary, when the tropical temperature changes, the solar forcing subsequently changes in the opposite direction, negating much of the effect of the volcanoes.

And in particular, the observations agree with the theoretical predictions, which were:

• Right after the eruption, there would be a reduction in available solar energy, due to the volcanic aerosols in the stratosphere.

• This initial eruption-induced reduction in available solar energy would be both deeper and sooner after the eruption in the hemisphere where the eruption occurred than in the opposite hemisphere.

• As a result, the corresponding climate reaction in the eruption hemisphere would also both be deeper and occur sooner than the climate reaction in the opposite hemisphere. In other words there will be a dose-related effect, where a larger reduction is met with a larger climate reaction.

• The form of the climate reaction will be an albedo reduction due to the temperature reduction, which will cause an increase in available solar energy. The increase in available energy will be of the same order of magnitude as the corresponding decrease due to volcanic aerosols.

These theoretical predictions are all visible in the graphs above, and they lead back to the title of this piece. The reason volcanoes don’t matter much is that the climate rapidly responds to re-establish the status quo ante. Yes, eruptions do put loads of aerosols into the stratosphere; and yes, these aerosols do cut down available solar energy; and yes, this does have local effects in space and time … but because available solar energy in the tropics goes up as the temperature goes down, the balance is quickly restored. As a result of this and other restorative phenomena, the climate system has proven to be surprisingly insensitive to such variations in forcing.

My best regards to everyone,

w.

The Usual Request: If you disagree with someone, please quote the exact words that you disagree with, so we can all be clear both who and what you are objecting to.

Methods and Data

Sato Theoretical Forcing: The Sato data is from here. Following Sato, I have used the aerosol optical depth (AOD) to estimate the forcing. Sato says that the forcing is estimated as a linear function of the AOD, which seems reasonable. I have used his formula for the “instantaneous” forcing (as opposed to the “equilibrium” or other forcings), since we are discussing the immediate effects of the eruptions.

Available Solar Energy Anomaly After Albedo: For the albedo data, I digitized the albedo shown in Figure 5(b) of the most interesting study, Long-term global distribution of Earth’s shortwave radiation budget at the top of atmosphere,  by Hatzianastassiou et al.  I multiplied the monthly (1 – albedo) by the monthly TOA solar to get the absolute value of the available solar energy after albedo reflections. Then I  subtracted the “climatology”, which means the monthly averages, from that dataset to get the anomaly in available solar energy

Stratospheric Absorption: While researching for this post, I had an interesting insight about the increase in stratospheric absorption of solar energy after an eruption. This was that I could use the change in stratospheric temperature to calculate the amount of additional sunlight being absorbed, using the Stefan-Boltzmann relationship. For the stratospheric temperatures, I used the UAH satellite based estimate of the lower stratosphere, Version 6.0beta2, available here. Yes, I am aware that this is an uncertain estimate, but it’s accurate enough for a first-order analysis such as this one.

Sensitivity to Assumed Emissivity: I used the most conservative assumption, that of a blackbody relationship with emissivity=1. If we assume a graybody, the change in solar absorption corresponding to a given temperature difference goes down in proportion to the change in emissivity. This reduces stratospheric absorption forcing. And this in turn increases the difference between the observed (yellow line) and the Sato theoretical forcings (black line) in the period immediately after the eruption, but makes little difference in the later years because the stratospheric absorption term is small. For an example of the change in the early years, using an emissivity of 0.5 reduced the largest total forcing decrease (reflected plus absorbed) to about minus one W/m2, rather than the approximately minus 1.75 W/m2 as shown after the eruption in Figure 3.

Data: One of the bad things about this is that the dataset is so short. Can’t be helped, because as far as I know there’s no hemispheric estimate of the albedo during the time of the previous eruption, El Chichon in 1982. (If you know of such a dataset, please post a link). But the good side of short data is there’s not much of it, so it’s easy to move around … for example, I’ve been looking at one-minute radiation measurements from Mauna Loa, 31 million data points per year since 1980. That’s hard to download, and too big for me to put up on something like photobucket.

But here we only have 14 years at 12 months per year = 168 records, so it’s small enough to put into an Excel spreadsheet, which I’ve done in .csv format here. The spreadsheet contains the TOA solar values, the albedo values, the Sato forcing values, the stratospheric temperature values, and as a special bonus, the hadCRUT4 records for the period both globally and for individual hemispheres. Enjoy.