The seasonal stratigraphy could in principle have been measured in situ as long as a(z) was at least a few millimetres thick, but in practice this was not possible as the measurements became too time consuming as a(z) approached 1cm; hence only selected sequences of the deeper ice strata were measured with a seasonal resolution. The GRIP continuous seasonal stratigraphy therefore ends around 60kyrBP. For older layers the sequential analysis can be used to establish a chronology, but the accuracy cannot compete with other palaeoclimatic chronologies. This may change in the future, when the North GRIP core becomes fully analysed. Such an extended ice-core master chronology will be limited to the past 100-130 kyr simply because this seems to be the limit for obtaining old ice with continuous and well resolved seasonal stratigraphy in the north polar region.
A major problem in estimating the stratigraphical chronology between 15 and 60kyrBP is the LGM period covering 15-30 kyr BP. During this period the dust concentration in the ice is up to 100 times the Holocene values and there is a risk that the interpretation of the seasonal signal is biased. In order to estimate the accuracy back to 60kyrBP we are presently left with only one possibility for verification: the Z2 volcanic tephra layer. This layer has been traced in several deep-sea sediment records from the North Atlantic and has also marked the GRIP ice with a visible layer (Gr0nvold et al., 1995). The layer has been dated by interpolation in the DSDP 609 core to 57.5kyrBP (Bond et al., 1993). In the GRIP core it is found at 2431 m depth with a stratigraph-ical age of 58,400 yr BP. The ss09 model time-scale dates the Z2 layer to 52 kyr BP. The ss09 time-scale is again underestimating the age, as for the onset of the B0lling, but now the difference is 6 kyr. The ss09 time-scale was later refined by including the change of the isotopic composition of seawater (Johnsen et al., 2001) and the new time-scale was called ss09sea. Recently a paper on speleothems in France by Genty et al. (2003) claimed a higher precision than the ss09 and ss09sea time-scales. The authors noted that the model chronologies differed by an entire DO event, i.e. indicating that the accuracy of the model time-scale could be out by several thousand years. The claimed precision of the new speleothem time-scale, dated by the U/Th technique, is in itself questionable, because it is based on a not too detailed comparison between DO events in the ice and a speleothem record; the latter record needs to be verified and so does the estimation of the claimed accuracy.
The accuracy of the stratigraphic dating of the GRIP core at 60 kyrBP is estimated to ±1-2kyr, but it rests on the interpolation of the DSDP 609 core, which can be questioned.
The continuous Greenland ice-core record reaches some 100kyr back in time and offers a very high time resolution. The last glacial stage is covered by nearly 1200 m of annual ice deposits containing information on regional and global changes of atmospheric composition and climatic changes. Some of the global changes have their counterpart in Antarctic ice cores and therefore may serve as reference horizons tying the well-dated Greenland record and the Antarctic cores into a tight chronological framework.
An interesting automated layer counting technique is also on the way, but it remains to be seen if such a technique is more accurate than the classic counting by 'eye'.
The chronological accuracy of the pre-100ka ice layers reached by deep Antarctic ice cores will probably depend on the dating accuracy offered by marine and terrestrial records. The unique information stored in the kilometre-long deep ice cores will, however, add to our understanding of the Earth's changing climate and environment.
Was this article helpful?