Constraining Paleo-ENSO with Coral Based Sr-U Thermometry

Nathaniel Rust Mollica, Woods Hole Oceanographic Institution, Woods Hole, MA, United States, Anne L Cohen, Woods Hole Oceanographic Institution, Woods Hole, United States and Forrest Horton, Woods Hole Oceanographic Institution, Department of Geology & Geophysics, Woods Hole, United States
Global Climate Model simulations project an increase in the frequency and intensity of El Niño under 21st century global warming, with profound implications for marine and terrestrial ecosystems. However, calibration and validation of model projections is constrained by the short duration of observational data used to index El Niño. Specifically, estimates of sea surface temperature (SST) in the central equatorial Pacific (CEP) rely on shipboard measurements and mooring data that are sparse in this region prior to the satellite era. Consequently, studies that attempt to evaluate changes in frequency and intensity of modern ENSO have constrained their analyses to the last few decades. The skeletons of long-lived corals are important archives of SST that can fill these critical spatial and temporal gaps, and those growing in the CEP are well-situated to capture ENSO variability. However, traditional coral proxies, i.e. Sr/Ca and δ18O, have not consistently reproduced instrumental SST due in part to vital effects that distort Sr/Ca ratios or uncertainties in historical seawater δ18O variability. A new coral thermometer, Sr-U, shows promise in overcoming vital effects. However, current Sr-U measurements require over a year of paired Sr/Ca and U/Ca data to yield a single SST, restricting the proxy’s temporal resolution. Here we present a new method for deriving monthly Sr-U SST from coral skeletons that allows resolution of ENSO variability. Using laser ablation ICP-MS, we generated numerous Sr/Ca and U/Ca data points within a single month of skeletal growth, as constrained by dissepimental sheets within the skeleton. Geochemical variability within each month supported an estimate of Sr-U, based on the regression of Sr/Ca versus U/Ca. Applying this method to two corals collected on atolls in the CEP, we generated ~10 years of bimonthly Sr-U from each core, spanning the 2010 and the 2015 El Niños. We compared the Sr-U records against satellite SST and in-situ logged temperature data. In both cases, Sr-U captured the timing and amplitude of SST variability and trends to within ±0.45 degrees (RMSE), including peak temperatures during both El Niños. Current analyses focus on reconstructing the history of SST in the CEP over the last 100+ years to enable quantification of the timing and amplitude of El Niño variability.