A COMPARISON OF RAPID-SCREENING 14C DATES AND U-TH DATES FROM FOSSIL CORALS: IMPLICATIONS FOR PALEOCLIMATE RECONSTRUCTION

Paleoclimate reconstructions from the central tropical Pacific provide much-needed targets for climate models that are used to simulate future projections of ENSO variability under enhanced greenhouse emissions. To date, paleo-ENSO records rely on rare but decades-long fossil coral sequences that date to the last 7,000 years (Cobb et al., 2013). By turning to abundant but shorter (5-10yrs-long) fossil coral rubble samples, we seek to produce a more statistically robust reconstruction of central tropical Pacific climate through the Holocene. To take full advantage of the rubble archive, we need to date hundreds of samples accurately and quickly. As a first step, we test a rapid-screening ¹⁴C dating method, developed at UC Irvine (Bush et al., in press), and compare its performance against high-precision ¹⁴C and U/Th dates. This method yields ages that differ by ±1.8% (1σ) from high-precision ¹⁴C dates for young (<10,000 years old) carbonates (Bush et al., in press) and is roughly 10 times faster than U/Th dates, making it well-suited to screen samples for time periods of interest.

We compare 69 rapid-screening ¹⁴C dates with new and published U/Th dates from fossil corals spanning the last 7,000 years from the Line Islands (~4N, 160W; Cobb et al., 2013). Results show that rapid-screening ¹⁴C dates are reproducible and differ from high-precision ¹⁴C dates by 1-4% (N=3). Sixty-seven of the 69 rapid-screening ¹⁴C dates agree with the U/Th dates within the relatively large uncertainties of the calibrated ¹⁴C dates. However, two samples with U/Th ages of ~6,500 years have ¹⁴C ages ~1000 years younger. Duplicate samples from these corals yield differing U/Th ages by ~300 years, indicating geochemical heterogeneity consistent with open-system behavior. Specifically, ¹⁴C addition (via precipitation of younger carbonates) and/or U removal (via preferential leaching of U during freshwater diagenesis) could explain the observed discrepancies. While the precision of the rapid-screening ¹⁴C dates translates into multi-century age errors for samples from the last millennium, this efficient dating technique could likely replace costly and time-consuming U/Th dates for older samples. At the very least, concordance between paired ¹⁴C and U/Th dates for fossil corals provides valuable confirmation of closed-system behavior.