A Comparison of Rapid-Screen 14C Dates and U/Th Dates from Fossil Corals: Implications for Paleoclimate Reconstruction

Paleoclimate records from fossil corals provide valuable benchmarks for climate model simulations of tropical climate variability. Unfortunately, the number of such reconstructions is limited by the rarity of the relatively large fossil coral heads that are targeted for paleoclimate studies. Abundant, albeit shorter (5-10yrs-long) fossil coral sequences could be used to generate a more robust, quasi-continuous reconstruction of tropical climate with built-in uncertainty estimates, but the large number of radiometric dates required is costly. A new rapid-screen 14C dating method, with a nominal precision of ±1.8% (1σ) for young carbonates (Bush et al., 2013), is roughly 10 times faster than high-precision U/Th dating, making it well-suited to screening a large number of corals. In this study, we compare rapid-screen 14C dates to both high-precision 14C dates and U/Th dates from well-characterized mid- to late-Holocene fossil corals collected from Christmas and Fanning Islands (2-4°N, 157-160°W) (Cobb et al., 2013). Our results show that most rapid-screen 14C dates agree with high-precision 14C dates (N=3) and U/Th dates (N=42), within combined uncertainties. However, two samples that contain 15 and 23% calcite (as determined by XRD) are associated with large discrepancies in calibrated 14C vs published U/Th ages (5065-5579 yBP (2σ) versus 6350 ±13 yBP (2σ) and 4868-5424 yBP (2σ) vs 6598 ±13 yBP (2σ)), respectively (Cobb et al., 2013). The 14C-U/Th mismatches, as well as poor reproducibility of replicate U/Th dates from these samples, are consistent with diagenetic alteration indicated by the presence of calcite. Mass balance calculations indicate that the dating discrepancies can be explained by a combination of 14C addition and U removal, both of which can occur through calcite recrystallization during freshwater diagenesis. Results illustrate that the rapid-screen 14C dating method is well-suited to surveying a large number (~100's) of fossil coral samples, provided that samples are pre-screened for calcite. Beyond that, concordance between paired 14C and U/Th dates provides valuable confirmation of closed-system behavior and chronological accuracy.