In CA 111 Chris Scarre pointed out that the explosion of Thera could be dated to 1626 BC. This may, however, only be the beginning. There are at least 4 other prehistoric dates that the readers of CA should learn by heart; I believe that our work on tree-rings has revealed several major volcanic eruptions which may have caused climatic upset on a world wide basis.
When these major eruptions occurred, the climate of the northern hemisphere may have been altered for several years. We have to envisage the possibility of failed harvests, famine – and no doubt plague and pestilence as well. Empires could have been destroyed and vast tracts of land rendered uninhabitable. In such circumstances the survivors would have been those who were more war-like than their neighbours – the consequences of such “aggression means survival” could have changed social behaviour for centuries.
While such claims may seem like rather fanciful extrapolation from a few dated “events” in the Irish oak record, they need to be suggested in the form of an hypothesis for archaeological testing. The most dramatic claim, which can ultimately be tested, is that the dust veils recognised in the Irish tree-rings may date the start and end of the Chinese Shang Dynasty.
Before we begin let us recap on the methodology. There are three major lines of research which appear to provide evidence for environmental catastrophes. The earliest to surface was the work of Danish glaciologists led by Dr C U Hammer. They analysed long cores from the Greenland ice sheets and found layers of sulphuric acid deposited by volcanic eruptions.
These could be broadly dated because the layers in the ice sheets are basically annual in character (see Nature, 288  and 328 ).
Then in 1984 an American dendrochronologist, Val Lamarche, who worked at the Tucson tree ring laboratory, drew attention to frost damaged rings which occurred in a lot of his high-altitude bristlecone pines. In recent times such frost rings tended to occur in the years following major volcanic eruptions e.g. the eruption of Krakatoa was in 1883, the bristlecones showed frost damage in their growth rings for 1884. One notable frost-ring event occurred in 1626 BC and he suggested that this might well be the result of the explosion of Santorini – if so, the actual eruption would have occurred in 1626 BC “or one or two years earlier”.
To clarify one point at this stage, all the American prehistoric dates have to be moved back by one year. This was due to La-Marche’s use of a “zero year” between AD 1 and 1BC. No such year is recognised in the historical calendar hence the frost ring actually relates to the growth ring for 1627 BC.
In Ireland our approach was rather different. One has to be clear that there are two populations of oaks. Archaeological timbers are assumed to be “land grown” oaks. The oaks we studied divided down fairly neatly by period – the last two millennia were mostly landgrown timbers while the prehistoric trees were almost exclusively bog oaks. These bog trees grew rooted on deep peat and hence their existence was marginal at best.
It became clear that it was these trees which were sensitive recorders of worsening conditions. The starting point was the observation that some of our bog trees showed very narrow bands of rings in the decade of the 1620s BC – tending to support the notion that something big had happened at that time.
When the effect was quantified it was discovered that seriously reduced growth episodes, indicated by the “widespread occurrence of narrowest growth rings”, were infrequent and tended to occur at the same times as Hammer’s ice-core acidity peaks i.e. clusters of narrowest rings in Irish bog oaks appeared to be related to large volcanic dust veil events (see Nature 332 ). Interestingly recent climate research by the Climatic Research Unit at the University of East Anglia has shown that in the months following eruptions in recent times, we tend to get low pressure anomalies over Britain and Ireland.
One can imagine that the more extreme conditions which might have been related to much larger eruptions in the past could have supplied just the right conditions to cause stress to the marginal bog trees.
The date of 1628 for Thera (or Santorini) presents a problem. Although it is backed up by the radiocarbon dates from Akrotiri on Thera, the conventional dating – based on cross-links with Egypt – is around 1500 BC. One possible solution is to look closely at the conventional dates. It is tempting to see the Thera explosion in dramatic terms as marking the “end” of the Minoan civilisation. However the dates and the linkage has always been a problem ever since the connection between Thera and the Minoan civilisation was first mooted.
The problem is that the Thera eruption is dated – by the pottery on Thera itself – to Late Minoan 1A, while the destructions on Crete are LM1B. In the original article in Antiquity in 1939 when Marinatos linked Thera with the end of Crete, he suggested in effect that LM1A and LM1B were merely two different styles and could be contemporaneous.
However our date of 1628 for Thera would make this impossible. The best current compromise – adopted as a resolution at the recent Thera Congress – is that Thera marks the end of LM1A, when there was considerable damage on Crete that was rapidly repaired. The “end” of the Minoan
empire came at the end of LM1B, some time later. There are still chronological problems: the end of LM1B is still usually dated around 1500 BC by cross links with Egypt, and this is usually considered to be “a generation” later than LM1A. The problem undoubtedly needs further examination.
In many ways, an even more interesting catastrophe was that of 1159 BC. Hammer had noted a significant ice signature at HOO —50 BC. The Irish trees show a very spectacular narrow band of rings beginning in 1159 BC – 43% of the trees from six sites have their narrowest rings during this period.
The event is almost certainly the third great eruption of the Icelandic volcano Hekla. The case for something unpleasant happening in Britain
has been noted by John Barber, director of the Scottish Central Excavation Unit. He has noted both extensive abandonment of upland sites and a decline in the numbers of “burnt mounds” – related to cooking at temporary hunting sites – in Scotland in the 12th century BC.
This can be coupled with Andrew Dugmore’s discovery of a fine layer of volcanic dust first in South Uist and the Shetlands and more recently in a wide area of northern Scotland. This tephra can be positively identified to the Hekla 3 eruption. So the evidence, precisely dated by the tree-rings, is accumulating to present a firm picture of effects on human populations.
Colin Burgess has long argued (notably in his book “The Age of Stonehenge”) that around this same date the Earlier Bronze Age came to a sudden end, and was replaced by a very different Later Bronze Age. It looks as if, for want of a better suggestion, that we may be able to argue that the Earlier Bronze Age ended, and the Later Bronze Age began in 1159 BC – precisely.
Coming forward, there is another major event in AD 540. The most obvious association with this date is the occurrence of the Justinian plague. If there really was a dust veil event at that time – and European dry fogs in 536 and 537 testify to that – then what better to follow any famine but plague.
The Neolithic and beyond
Earlier than 1628 BC there are two other events which may feature in the archaeological record. These are at 3195 BC and 4375 BC. It is
Sections of three bog oaks reveal the 1159 lacuna. Ring patterns from three sites, Gortgole, Toome and Tullyroan, show the catastrophic reduction in ring widths in the 1150s and 1140s. It is tempting at first to link the 3195 BC event to the elm decline, but unfortunately that doesn’t work.
Since 3195 BC is a tree-ring date, it has to be de-calibrated to convert it to the radiocarbon time scale. So this suggested dust veil event could be related to happenings around 2500 bc, to use the old lower case notation that everyone understands – distinctly too late for the elm decline. How about trying it out for the interface between the Early and Middle Neolithic – a nice catastrophe to end the causewayed camp/chambered tomb complex?
The same goes for 4375 BC which de-calibrates to something like 3600 bc in radiocarbon years. This could have made life difficult for Mesolithic people – or indeed for any really early farmers who had just sneaked in. Only time will tell how we interpret this date!
Further back still there is an event which should be noted even though there is no tree-ring evidence for it! Hammer’s 5400 + 100 BC ice core date almost certainly related to Mt. Mazama better known as Crater Lake, Oregon. This was a very large eruption and it is interesting that the Irish bog oak sequence begins at 5289 BC.
Again the de-calibrated date is somewhere around 4600 bc in radiocarbon years. Having uncovered these events/dates in the Irish oak record and then having found the coincidence with the ice-core dates I have been amazed how additional information has tended to reinforce the
catastrophic concept. I am convinced that these dates mean something in human terms. However I also recognise that the poor chronological control normally available in archaeology presents a problem. Any sloppily dated archaeological event, within a century or so, tends to be “sucked in” to the precisely dated tree-ring events. We all have to be on our guard against circular arguments.
If people really want their catastrophes they will have to work for them by refining their archaeological chronologies.
This is an extract from an article published in our ‘Disasters’ special issue, CA 117