An open-system model for U-series age determinations of fossil corals

The source of excess U-234 in fossil corals and its relationship to U-series age determinations has been an outstanding problem in geochronology for more than 20 years. With increasing numbers of U-series isotope measurements in corals, and significant improvements in analytical precision through mass spectrometry, it is increasingly apparent that a substantial fraction of observed isotope ratios cannot be reasonably explained by closed-system decay. Moreover, observations of a positive correlation between U-234/U-238 and Th-230/U-238 ratios in corals from the same terrace are difficult to explain. However, the decay of dissolved uranium and a-recoil mobilization of uranium daughters produce particle-reactive Th-234 and Th-230, and the coupled addition of these Th isotopes could simultaneously increase coral U-234/U-238 and Th-230/U-238. Here we present a quantitative model, based on decaydependent redistribution of Th-234 and Th-230, permitting calculation of open-system coral ages. These equations provide a general solution to the a-recoil redistribution problem, applicable to any alpha decay series. While measured isotope ratios of corals from the three youngest stratigraphically defined Barbados terraces are inconsistent with closed-system decay, they fall in broadly linear arrays agreeing with model predictions. Isotopic arrays of older Barbados corals, and corals from terraces around the world, are also consistent with model predictions suggesting the open-system model is generally applicable. Corals with extreme isotopic compositions that are impossible to produce by closed-system decay are consistent with the limited range of isotopic compositions predicted by the model at ages older than 600 ka. For corals from a single terrace, Th-234 and Th-230 redistribution appears to be a source of systematic conventional age error, even for corals with slightly elevated U-234. However, open-system ages are consistent, even for corals with extremely elevated U-211. For the youngest three Barbados terraces, mean open-system terrace ages are consistent with mean conventional terrace ages calculated from pristine samples. If the most accurate conventional ages are from corals with an initial U-234/U-238 identical to modern seawater, then the open-system model will improve the accuracy of coral U-series age determinations and dramatically increase the number of reliable ages. (C) 2003 Elsevier Science B.V. All rights reserved.