Insight into climate debate at the Volcanism and the Atmosphere conference

Last week was the American Geophysical Union (AGU) Chapman Conference on Volcanism and the Atmosphere in Selfoss, Iceland. It covered topics such as explosive eruptions, satellite detection of volcanic ash, aviation hazards and climate modelling. Unlike larger meetings, where sessions run in parallel like the stages at a music festival, all the presentations happened in one room and everyone went to all of them. This way, instead of sticking to the geology sessions (normally filled with pictures of hammers in exotic locations), I saw a lot of talks from other fields. These included a lively debate about how the effects of volcanic eruptions are preserved in the tree ring climate record, which is the subject of this post.

Live tweeting from the conference

Much of the conference was reported live on Twitter.  You can find the conversation by searching for the hashtag #AGUVolcAtm.   After the speakers have agonized over their presentations in order to fit them into the 15 minute time slot, it’s fun to see them subsequently mashed into less than 140 characters by the audience. There are also 300 word summaries (abstracts) of the presentations available online here.

Volcanic eruptions and climate

The connection between volcanoes and climate is a result of the gases produced during eruptions.  These include sulphur dioxide and carbon dioxide. It is the sulphur dioxide (SO2) that provides the main influence on climate, as it reacts in the atmosphere to form sulphuric acid aerosol (H2SO4). An aerosol is a suspension of tiny solid or liquid particles in a gas. The sulphuric acid particles reflect incoming radiation from the Sun back into space. If the gas is injected into the stratosphere, the aerosol can remain aloft for years. In this way, large volcanic eruptions cool the surface of the Earth.  Of course, it’s a bit more complicated than I have just explained, and many of the presentations explored the details of the process.

Climatically speaking, volcanic carbon dioxide (CO2) is of minor importance, as the amount of gas that volcanoes emit is dwarfed by human emissions; the average annual global volcanic CO2 emission rate is equivalent to that of a moderately-sized country such as Poland. Other volcanic gases such as chlorine and fluorine have atmospheric effects such as breaking down ozone in the stratosphere.

Genuine climate debate

A highlight of the conference was a pair of consecutive talks by Michael Mann and Rosanne  D’Arrigo about the signals from past volcanic eruptions in the tree ring record. As a geologist, the exchange gave insight into the topics being debated at the cutting edge of climate science.

That debates exist between climate scientists is sometimes reported as an indication that the foundations of the whole field are unsteady, but this is a misunderstanding of how science works. Arguments over details are common, and indeed necessary to refine our understanding, but often reflect just a fragment of a bigger picture. Two palaeontologists may argue over whether Tyrannosaurus Rex had feathers, but both would agree on the bigger point that they share a common ancestor with the beast.

Here, both Mann and D’Arrigo agree on long term trends in the tree ring record (they must, because Mann’s study uses data produced by D’Arrigo) and that the Earth is warming. But each scientist’s passionate defence of their own ideas shows that consensus on the short term effects of volcanic eruptions on the tree ring record is yet to be reached. Both talks were clear, logical, detailed, and absorbed everyone in the room. This is genuine climate debate and it is fascinating to watch.

Underestimation of Volcanic Cooling in Tree-Ring Based Reconstructions of Hemispheric Temperatures

The <140 character version of Michael Mann’s talk is:

.@MichaelEMann: Tree rings miss volcanic cooling spikes. Cold limits growth, but diffuse light from atmos aerosol boosts it. #AGUVolcAtm

It described his recent paper that suggests that tree rings underestimate volcanic cooling. Mann used a computer simulated climate, which he had shown to do a good job of estimating the cooling effect of recent eruptions, to calculate global temperature from 1200-1980 (red line below). It shows clear spikes associated with eruptions in 1258, 1452, 1809 and 1815. He also plotted temperatures from the same period as estimated from studies of tree ring widths (blue line). The spikes are missing. The new study tried to explain why the tree ring data were underestimating the cooling and missing the spikes.

Figure from the Mann et al paper

Modified version of Figure 2d from the paper by Mann and coworkers published in Nature. Click to visit the journal.

The tree ring width data came from forests that are so high up mountains or so close to the poles that the trees are clinging on to life at the very edge of where trees can survive. The growth of such trees has been shown to respond more to temperature changes than to other effects e.g. rainfall. Mann and colleagues used equations describing tree growth at different temperatures to predict what the trees would record given the temperatures in the computer-simulated climate (green line). They included a threshold temperature below which growth stops, a description of how diffuse light caused by atmospheric aerosols can help trees grow, and random local variations in weather conditions.

The result is that the recorded cooling is reduced and, because the no-growth threshold resulted in some years with missing rings, the cooling appears delayed relative to the eruption. The calculated tree rings now show good agreement with the measured ones, leading Mann to conclude that his equations are describing real effects.

Volcanic Signals in Tree-ring Records for the Past Millennium

Next, Rosanna D’Arrigo took to the podium in defence of dendrochronology (tree ring dating) and launched into a point-by-point rebuttal of a number of Mann’s arguments. In 140 characters, it goes like this:

D’Arrigo: BOOM! Tree ring widths aren’t as good as density and diffuse effect was measured on different forest type to rings. #AGUVolcAtm

Her main point was that Mann’s use of tree ring width data was inappropriate, because tree ring width data are best suited to measuring longer-term trends in temperature. To look at volcanic cooling spikes, they should have used tree ring density data (maximum latewood density: MXD), which is more sensitive to short term changes. She described a study by Briffa and co-workers that picked up the cooling following the Tambora eruption in 1815. Mann had not mentioned this study in his paper.

D’Arrigo said that the diffuse effect was recorded in forests with a thick canopy, which is unlike the areas where the tree rings were measured. She also pointed out studies showing the trees can grow at temperatures below Mann’s threshold, and that the missing rings were less common than he had suggested.

The discussion continues

D’Arrigo and her fellow dendrochronologists have prepared a formal response, so the debate will continue, in full public view, in the pages of scientific journals. At the meeting, it spilled over onto Twitter, with Mann agreeing that tree ring density (TRD) measurements are better than widths (TRW), but that “TRW dominate tree-ring temp recons”. Unfortunately, tree ring density data is more difficult and expensive to collect.

From the sidelines, it doesn’t seem that the scientific problems are so serious. In time, more density measurements will be collected and the reconstructions will be improved.  Meanwhile, Mann’s ideas can be tested further and accepted or discarded depending on how well they stand up.  The real heat of the argument appears to result from Mann’s failure to emphasise that that tree rings CAN measure volcanic cooling spikes.  His study used tree ring widths because the data were most abundant, even though better methods exist. In the wider media, this made tree ring dating sound less useful than it is, which understandably annoyed the dendrochronologists.

Other highlights in 140 characters or less

Here are Twitter-sized summaries of some of the other talks at the meeting:

  • .@volcanofile: tropospheric volcanic sulphate -> whiter clouds -> global cooling. Better estimates of past emissions needed. #AGUVolcAtm
  • Alan Robock, Rutgers: 2011 eruption of Nabro, Eritrea, was largest sulfate producer since 1991’s Pinatubo. #AGUVolcAtm (sent by @alexwitze)
  • Foelsche: GPS signals between satellites bend as pass thru atmosphere; temperature controls bending -> can calculate atmos temp. #AGUVolcAtm
  • Foelsche: Now we need a big eruption to see if we can detect the effects. #AGUVolcAtm
  • Thor Thordarson: 560 cubic km of magma erupted in Iceland in last 11,000 years, since ice age ended. (That’s a lotta magma.) #AGUVolcAtm (sent by @alexwitze)
  • Miller: Baffin glaciers retreating -> 14C-date newly uncovered moss -> shows rapid lowering of snowline in 1450s -> LittleIceAge #AGUVolcAtm
  • Lavigne: named the 1257 (1258) ‘mystery’ #eruption but can’t reveal name due to ’embargo’ I suspect … #AGUVolcAtm (sent by @volcanofile)
  • Lavigne misunderstands journal embargoes. Nature & Science v clear. Talks allowed. #AGUVolcAtm (sent by @alexwitze)
  • Prata: If ash conc < 200 microns/m3, can’t really detect w/satellite, but that’s OK b/c it’s not that dangerous to planes. #AGUVolcAtm (sent by @alexwitze)
  • Prata: #Eyjafjallajokull yielded some 10 sci papers per teragram of ash emitted. #AGUVolcAtm (sent by @alexwitze)
  • Krueger: Strong eruption -> increased winds in Southern ocean -> limits transport to Antarctica -> reduced sulphate in ice core. #AGUVolcAtm
  • Elkins-Tanton: Siberian Traps magma chambers in hydrocarbon+evaporite basin -> adds extra S+Cl+F. To 3000000km3 basalt! Nasty! #AGUVolcAtm
  • Brian Toon: ‘noctilucent #clouds after large #eruptions could indicate that water was injected in stratosphere’ #AGUVolcAtm (sent by @volcanofile)
  • Graf: ‘romantic sunsets after big #eruptions … Need to watch birth rates … with implications for geo-engineering’ #AGUVolcAtm (sent by @volcanofile)

Other reports from the conference

Categories: Uncategorized


  1. Pingback: Stuff we linked to on Twitter last week | Highly Allochthonous

  2. Pingback: How do satellites map volcanic ash clouds? | Volcan01010

Leave a comment:

Your email address will not be published. Required fields are marked *