As we labour to rescue what can be saved, we shall continue from where we last were. Here again is that fabulous post by Albert, alas, without the original reader comments:
Volcanic eruptions have become major attractions, and even rather minor eruptions can make front page news. In modern days, any volcano deciding to erupt will find itself instantly monitored and Volcano-Cafe’d. But in the days before global coverage (and, dare I say it, Volcano Cafe), many eruptions went unnoticed. Thus, in May 1831 and again in August, parts of Europe and the coast of Africa were covered in a “dry fog” similar to (but not as extensive as) the one caused by Laki in 1783. But the sulphuric haze (if that is was what it was) was not identified as volcanic and the culprit has never been discovered. For older eruptions, the existence of records depends entirely on location. We have very good dates for Vesuvius or Mount Fuji, but none whatsoever for Mount Erebus. In 1915, Shackleton described seeing an iceberg with a clear volcanic dust layer embedded, but ascribing it to a particular volcano would have been difficult. (Given where he was at the time (the Weddell Sea), Deception Island with its ashy eruptions must be a suspect.)
The best records we have come from ice cores obtained from Greenland and from Antarctica. Ice cores are drilled out of the ice sheets and can be up to 3 km long (or deep). The yearly cycle of snow fall and summer melt leads to annual layers, which can be counted just like tree rings. The older layers are found at the bottom of the ice cores, and they can be as old as 123,000 years in Greenland and 800,000 years in Antarctica. The ice contains pollutants that became embedded in the snow, and small bubbles of air. These become isolated from the atmosphere as new layers accumulate above, and preserve a sample of conditions of the past. Dating ice by counting layers can be as accurate as one year for layers of a century old, and a few years for layers older than 1000 years. If a layer is seen in several ice cores, the date becomes more accurate. Ice cores have been used to measure past CO2 levels and temperatures, and lead from the 1850’s gold rush, but they also contain a history of the major volcanic eruptions.
Volcanoes produce two things that can end up in the polar ice: tephra (ash) and sulphate (SO4>). Both can reach the stratosphere after a major eruption and travel large distances. Tephra contained in the ice is mainly ash (particles less than 2 mm across, including glass shards) and these can often, but not always, be seen by eye in the core. Sulphate is measured either from analysis of the gas bubbles contained in the ice, or from measuring the electric conductivity of each layer which in effect measures the acidity of the ice.
Not all volcanic eruptions are recorded in the ice. It depends on how much sulphate and/or dust was ejected, but also where the eruption occured. Ash falls down more quickly and tends not to spread as far. Thus, ash in Greenland may come from Iceland (if the wind was right), Jan Mayen, or Alaska, but is less likely from (say) Java, unless the eruption was very large. Sulphate can stay up for several years and can spread out over an entire hemisphere, or both hemispheres if the eruption occured in the tropics. (Dispersal from south to north or vice versa is more difficult so non-tropical vocanoes mainly affect their own hemisphere.) But due to the vagaries of weather, it may miss one location and show up strongly in one a short distance away. The ejecta need time to spread to the polar regions, and this can take up to a year. The ice core date can therefore be a little after the volcanic date.
Several recent scientific papers have identified the volcanic layers seen in a number of different ice cores, and have tried to tie these to the guilty volcanoes. Sometimes this is easy: for instance, the layer at 1815/1816 is readily identified with Tambora. Sometimes the obvious answer is not fully right: the 1991/1992 layer is obviously attributed to Pinatubo, but in fact in the south may also contain material from Mount Hudson, a Chilean volcano only a little weaker than Pinatubo which erupted four months later. Older layers are more difficult to assign to a particular volcano. The list of volcanic events in the ice cores goes back almost 100,000 years. The famous Toba eruption, 73,000 years ago, and 100 times stronger than Tambora, has left a strong signal in the ice. The Phlegrean eruption in Italy, 34,000 years ago, is also clearly seen.
I have tried to compile a list of the ‘millennium volcanoes’ as seen in the ice, using recent papers . Where possible, the originating volcano is identified. The original papers list the dates for the sulphate and tephra layers, and give possible volcanoes if known. Many of the identifications are in the recent list of Michael Sigl from 2013 and 2014, others are in earlier work by Jihong Cole-Dai. But many of the layers are simply listed as ‘unknown’. I have tried to confirm the proposed identifications by correlating with a list of known eruptions of the past millennium, by Nicholas Deligne (2012). Sometimes there are significant uncertainties on the volcanic dates which makes a specific identification more difficult. Some dates come from radio carbon dating, either ‘calendric’ (real years) or ‘uncalibrated’: in the latter case they need to be adjusted to account for the fluctuating 14C content of the atmosphere. If they are listed as ‘BP’ (before present), they are counted back from 1950 (somewhat arbitrarily defined as ‘the present’). Other dates are from historical records. The list agrees with many of the proposed identifications, but in some cases the dates are not in good enough agreement. I also found about 10 new potential identifications.
The list of millennium volcanoes is shown below. Some large eruption may be missing if they left no ice record, and some eruptions seen in the ice cores may have been weakish (VEI 4) but were closer to the pole. All the largest eruptions should be in this list, but for comparison I also show some of the larger recent eruptions with no ice record. The eruption date is given if known (with month), or otherwise the ice core date, followed by the (suspected) originating volcano. The indicative VEI is given and the ejected volume in ‘dense rock equivalent’, i.e. the size of the hole it left (the volume after ejection can be several times larger because tephra expands so much). Both are taken from the literature if I could find this, or scaled down from the post-eruption volume given in Deligne’s catalog. Next, the amount of sulphate deposited in the ice cores is given, relative to that of Tambora (1815). Tambora deposited twice as much sulphate in the south as in the north (80 versus 40 kg of SO4 per square kilometre). Thus, equal numbers in both columns actually means twice as much in Antarctica. Numbers larger than 1 indicate more sulphate than from Tambora. That does not automatically mean that the eruption was bigger: it may have much closer to the north (or south) pole than Tambora was.
The last columns gives an estimate of the latitude of the eruption, based on the ratio of arctic to antarctic sulphate. This should be taken with a healthy grain of salt, and is certainly no more accurate than 10-20 degrees. Where the volcano is known, the actual value is also shown, for comparison. In most cases there is reasonable agreement. In some cases, such as 1477, the arctic and antarctic ice cores probably trace different volcanoes. Where no volcano is known, such as the large event of 1808, the estimated latitude gives an idea where the culprit might be located. In a few cases this calculation was used to identify a potential culprit, or rule out a proposed one.
Tambora is often mentioned as the largest eruption of the past millennium. This is however not clear. Two eruptions have left larger deposits, and probably exceeded Tambora: Kuwae (1458) and Rinjani/Samarand (1257). Tambora is now considered to be the 3rd largest eruption of the millenium. The three largest eruptions are all from the same part of the world, giving an indication where future disruption is most likely to come from.
One warning to end with. People have looked for evidence of past eruption mainly in volcanoes with known recent activity. The list of identified culprits is by no means complete. Some sleeping giants may have been overlooked. But in general, it may be wise to keep an eye on any volcano in this list.
- 1831: An identification with Babuyan is often proposed but seems unlikely. The dust/sulphur is seen only in Greenland, not Antarctica, and the dry fog reported in Europe that year also suggests a northern location, perhaps North America.
- The 1808/09 eruption remains a mystery. It has left a strong signal in Greenland, equal to Tambora, but a bit less than Tambora in Antarctica. This suggests equatorial, and possibly not Indonesia.
- Quilotoa, AD 1229: This is normally called the ‘800BP’ eruption (making it AD 1150), but a range of calibrated 14C dates indicates a bit younger ages, towards 1220-1260. The 1285 event could also be a strong candidate for this eruption.
- Cole-Dai et al., Journal of Geophysical Research, volume 102, page 16,761 – 16,771 (1997)
- Abbott and Davies, Earth-Science Reviews, volume 155, page 173 – 191 (2012)
- Plummer et al., Climate of the Past, volume 8, pages 1929 – 1940 (2012)
- Scott et al, GPA special paper (2005). http://www.geo.mtu.edu/~raman/papers2/ScottetalGSASP.pdf
- Stigl et al. Journal of Geophysical Research: Atmospheres, Vol. 118, 1151â€“1169 (2013)
- Deligne et al, Journal of Geophysical Research, Solid Earth, Vol 116, (2010)