Another approach to ice-core dating is seasonal stratigraphy. In ideal cases this technique offers a tree-ring-like accuracy. The prerequisite for dating by seasonal stratigraphy is a site with a reasonably high annual snow accumulation, which a priori excludes a whole range of drill sites, for example, the Vostok core on the East Antarctic plateau was dated by a combination of two models, one for the past annual accumulation and one for the ice flow (Lorius et al., 1985).
High accumulation areas do not suffer from underrepre-sentation of seasonal snowfall or mixing of annual snow layers by wind action, apart from certain marginal ice-sheet sites. Although the high accumulation areas are obvious candidates for dating by seasonal stratigraphy they are, however, also areas where the older layers are thinned dramatically at great depths: a consequence of the ice flow. This is the main reason why cores from the Greenland Ice Sheet can be dated by seasonal stratigraphy only back to approximately 100ka. The older annual layers are relatively close to the bedrock, which not only makes them very thin, but also increases the probability of breaking the ordered layering or they may simply have melted away at the bedrock.
The seasonal variations can be observed in many of the parameters studied along ice cores and their seasonality may have different explanations: Not all of them can be used as a dating tool to any depth of the core. In the literature some of the most used parameters are, for example, isotopic variation of the ice composition or concentration variations of nitrate, dust, calcium or chloride. It is, however, important to understand why the various parameters are changing with season in order to interprete the validity of the stratigraphical dating; not all apparently seasonal variations may actually turn out to be true seasonal variations. Multiple parameter dating and cross-dating may help to identify the true annual layers, but the accuracy will still be influenced by a number of'difficult' layers. It is not the intention here to discuss the various methods, but a number of techniques and their limitations can be found in Hammer et al. (1978) and Hammer (1989).
The estimation of the accuracy can be judged if events described in historical documents can be identified in the core, for example a well-known volcanic eruption. This will, however, not work for layers older than a few thousand years. For older layers the accuracy of the time-scale can be verified only if it is compared with other well-dated records containing comparable events. Presently this is a controversial subject, because no 'master chronology' exists for palaeorecords (SPECMAP, which is used to compare sea-sediment records is a type of master chronology for sea sediments, but it cannot be used for comparing the more detailed events in ice records, for example, DO events).
Was this article helpful?