Absolute chronology of marine sediment beyond the C-14 age range provides a test for models of climate change and has many other applications. U-Th techniques have been used for such chronology by dating corals, but extending these techniques to marine sediment is complicated by the presence of significant initial Th-230-both in detrital material and scavenged from seawater. In this study, we investigate four methods of solving the initial Th-230 problem for a particular type of marine sediment-the aragonite-rich sediments of carbonate platforms and slopes. Bulk sediment U-Th analyses can be corrected for initial Th to yield ages with approximate to2 to 3 kyr precision for highstand periods when sediment aragonite contents are particularly high. Uncertainty on the corrections causes inadequate precision for sediment from other periods, however. Removal of scavenged Th before analysis would enable a dramatic increase in this precision but has not proved successful despite a range of chemical leach approaches. Using heavy liquids to separate the various carbonate minerals found in Bahamas sediment enables an isochron approach to correct for initial Th, but the presence of initial Th from two sources requires correction or removal of one source of initial Th before the other is deconvolved by the isochron. Quantitative removal of detrital material before isochron analysis proves a successful approach. Such isochron data demonstrate that, although sediment remains closed to U-Th on a centimetre scale, nuclides are moved from grain to grain by alpha -recoil. Such intergrain exchange is expected to be observed in all sediments containing mineral grains with different U concentrations. Measured U-234/U-238 allows the recoil movement to be corrected and results in isochron ages with precision sometimes as low as 3 kyr. The accuracy of this approach has been proved by dating samples within the 14C age range. Sediments spanning the penultimate deglaciation have also been dated. After a small correction for bioturbation, the age for this event is found to be 135.2 +/- 3.5 ka. TWs date is approximate to8 kyr before the peak in northern hemisphere insolation and suggests that deglaciation is initiated by a mechanism in the southern hemisphere or tropics. This isochron approach shows considerable promise for dating of sediments older than this event, which will provide further information about the timing and mechanisms of global climate change. Copyright (C) 2001 Elsevier Science Ltd.
462ZLTimes Cited:29Cited References Count:44