Direct measurement of U and Th by mass spectrometry recently provided a spectacular improvement in the precision of coral dating by the U-Th disequilibrium method . Mass spectrometric data also provide a much better resolution for the examination of subtle diagenetic effects revealed by small variations of the U-234/U-238 ratio. Such perturbations may strongly affect the accuracy of the U-Th chronometer.A compilation of all the corals analyzed to date by mass spectrometry shows that most of the corals from terraces of the last interglacial have initial U-234/U-238 ratios higher than present-day seawater, in contrast to modern, Holocene and last Glacial corals. Some samples that have very high U-234/U-238 initial ratios, up to 1.2, and high U concentrations, up to 4 ppm, were probably contaminated by continental groundwaters. However, even apparently pristine samples have U-234/U-238 initial ratios which are still slightly higher than present-day seawater (mean value: 1.160, compared to 1.140-1.150 ), with little overlap between the two distributions.This difference in the U initial ratio raises some uncertainty about the accuracy of the U-Th age determinations of these corals. In spite of the fact that the U-234-Th-230 ages cluster in a narrow range between 122 and 133 kyr, the data could also be interpreted as resulting from slight contamination of corals that are significantly older than 125 kyr.Two possible explanations may explain these data: (1) All these samples may have been diagenetically altered, since they all come from surface outcrops which have been directly exposed to precipitations or soil waters for 125 kyr. In this case, the true age of these corals remains uncertain, depending on the timing of the alteration process (i.e., initial, late or continuous). (2) The second possibility is that some of the differences in U-234/U-238 ratios measured in approximately 125 kyr cold corals, compared to modern seawater, may be due to a higher U-234/U-238 ratio in seawater 125 kyr ago. This could result from temporal variations in the weathering regimes of continental land masses.
Gm343Times Cited:83Cited References Count:39