Volcanoes, Ice-Cores and Tree-Rings: One Story or Two?

Article excerpt


A tree-ring chronology is normally built back from the present, anchored by the sampling date of modern trees. A more dangerous game would be to choose an anchor further back in rime, for example, the felling date of timbers from a great medieval cathedral that is historically 'dated'. If you did this, no matter how good the historical documentation might appear to be, you would be leaving yourself open to the possibility of error.

When Ernst Hollstein (1980) was constructing his German oak chronology at Trier there came a point when be had a continuous chronology pieced together back into the third century AD. At that point he ran into trouble because he had oak timbers from a Roman bridge across the Rhine at Koln that appeared to have been constructed in 310. He linked this bridge chronology into his existing chronology at 310. This was fine until independent evidence showed that the extended chronology now produced dates that were historically impossible (Hollstein 1980). Moreover an independent German chronology (Becker 1981) showed that Hollstein's chronology actually ended 26 years later in AD 336. Hollstein was wrong because he had based his chronology, not on replicated tree-ring pattern cross-matches (as be should have done, and as Becker did), bur on an assumption that be could rely on the archaeological identification of a specific bridge mentioned in an ancient source.

Another example occurred in England. In the 1970s there was no continuous oak chronology in Britain over the last 1000 years. John Fletcher at Oxford, cross-dating the oak panels of late medieval paintings, assumed that the panels had been cut from elite (straight grown and narrow ringed) English oaks (Fletcher 1978). As be did not have ah English oak chronology against which to date his ring patterns, he attempted cross-matching against another art-historical chronology that had already been constructed in Germany, using mostly oak panels of Dutch paintings (Eckstein et al. 1975). Things went well until a wider suite of oak chronologies started to become available from across Northern Europe, Britain and Ireland. Correlations showed that the oaks Fletcher had assumed to be English had in fact been imported from the Baltic region. Worse still, the 'German' art-historical oak chronology against which the allegedly 'English' chronology had been cross-dated turned out to be from the same Baltic source (explaining why they cross-matched) (Baillie et al. 1985; Fletcher 1986). In turn it transpired that the 'German' art-historical chronology had not been correctly dated against genuine German oak chronologies (Eckstein et al. 1986). The dangers in making assumptions in chronology building could not be better demonstrated.

Similar problems may have been encountered in ice-core dating. While tree-rings record ah environmental downturn from 43 to 40 BC, following the eruption historically recorded as a dust veil in the year of Caesar's death, European ice-cores put a large volcanic acid layer at 50 BC. Similarly there are tree-ring effects in the 1620s BC, consistent with radiocarbon dates for the eruption of Thera; ice-core acidity suggests a large volcano in the 1640s BC. There are really only two possibilities: either the ice-core dates are slightly too old compared with the precisely dated tree-ring and/or historical dates, or our understanding of the relationship between the acid output of large volcanoes and their environmental effects is highly inadequate.

This article examines the history of the ice-core chronology and identifies one area that could explain how the ice-core workers might have built ah error into their overall chronology. To be specific, the European ice-core chronology--based primarily on the Dye-3 ice-core from Greenland--may have been erroneously placed in time by mis-identification of a layer of acid as being due to the Vesuvius eruption of AD 79. The possibility exists that the ice chronology would have been better fixed by placing the 50 BC acidity at the historical 44 BC event. …