Accuracy and precision in sediment chronology

Abstract

Several complications are encountered in radiocarbon dating of limnological samples. The contamination problem is obvious. Since lake sediments are usually composed of both allochthonous and autochthonous material the origin must be considered. The organic content may give some guidance. Graphite in the catchment area and a low organic content indicate a potential risk of major errors in the dates. Organic material from the surroundings may be displaced into a lake by creeks, snowslides and wind. Such material may be older than, or contemporaneous with the sediment deposited at the same time. Water-level changes and wave action may cause erosion and thus admixture of old material. Such contamination can sometimes be traced via different fractions extracted mechanically or chemically. The accelerator technique has made it possible to date small samples and thus macrofossils can be selected as the material to be used for dating far more frequently than is possible with the conventional technique for activity measurement.

Another problem is the reservoir effect — thus the lower ¹⁴C activity of the dissolved bicarbonate and carbon dioxide than of the atmosphere. There may be several reasons besides dissolved carbonates in hard-water lakes. Some plants even use carbon dioxide in the sediment. It is thus to be expected that submerged plants will be dated too old. Selection of plant remains for dating is thus no guarantee of an accurate date. The material to be preferred is terrestrial, such as tree leaves or remains of emergent plants.

Bioturbation must not be forgotten since this means that old material may be found too high in a sediment, and young material can be brought downwards in the sediment. A single date may be very misleading. Under rather stable conditions the sediment accumulation rate can be roughly estimated and acceptable dates obtained despite the complications.

It may be worth sieving gyttja samples to remove roots and rootlets. Roots in peat should also be removed. Diagrams are available of the errors due to certain degrees of contamination by younger or older material.

The ¹³C content must be measured to allow a normalisation. The present knowledge of the secular variations of the ¹⁴C activity allows calibration of dates made on wood etc. The same curves cannot be used for submerged material because of the lower ¹⁴C activity. Sudden changes of the atmospheric ¹⁴C activity are smoothed. Certain age ranges give better resolution than others. Recent curves indicate that strange results are to be expected for Holecene samples with a real age of 7000 years or older. Finally bacterial influence during storage may influence the isotopic composition.