NOAA/WDS Paleoclimatology - Kiritimati Island 4.3ka Fossil Coral d18O Data

STUDY NOTES: Data files updated 5/2/2014 to include Sr/Ca, Sr/Ca-SST, and Dd18O data from fossil coral XM35 and modern coral XM22, as plotted in Supplementary Figure 9, McGregor et al. 2013 Nature Geoscience. Fossil coral oxygen isotope (d18O) data, and fossil and modern coral Sr/Ca data from near Kiritimati Island, eastern tropical Pacific. Living and fossil Porites sp. microatolls were collected at Kiritimati Island (1°50'N, 157°25'W). The well-preserved fossil coral (XM35) grew continuously for 175 years between ~4.2 and ~4.4 kyBP, as indicated by an average U-series age of 4,243 ± 9 years BP, and by clear annual d18O cycles and density bands. Skeletal d18O was measured in fossil coral XM35 at approximately monthly resolution. The age model for coral XM35 was established in the same way as for the modern coral microatoll record (XM22; McGregor et al. 2011), by assigning d18O maxima to early February. The d18O data were interpolated to give 12 values/year. The overall age model uncertainty for the coral time-series is estimated as ± 1 year. For comparison a compilation of Kiritimati Island modern coral d18O records (modern coral d18O stack; Evans et al., 1998; Woodroffe et al., 2003; McGregor et al., 2011), spanning March 1938 to May 2007 was constructed. Comparisons of the annual cycle were made using a sub-set of the modern coral d18O stack (WM_stack; Woodroffe et al., 2003; McGregor et al., 2011) modern coral d18O record for the period February 1978 to May 2007. Time series were broken down into Trend, Annual, and Residual (interannual) components, which occupy distinct spectral bins. Statistical analyses were performed on these components (Supplementary Methods). Coral Sr/Ca was measured on the youngest 30 years of coral microatoll XM35 (coral years 147-177), and for 1994-2007 in modern coral microatoll XM22 to verify the dominant contribution of SST to Kiritimati coral d18O in non-El Nino years. The within-run precision was +/- 0.02 mmol/mol (0.2% RSD). The JCp-1 coral reference material (Okai et al., 2002) was included in every run and the standard deviation for repeat measurements was +/- 0.03 mmol/mol (n = 24). The average standard error for repeat analyses of Sr/Ca sample solutions was 0.005 mmol/mol (n = 56). The coral d18O-based age model for XM22 and XM35 also applies to the Sr/Ca results, since the Sr/Ca analyses were performed on the same coral powders. Sr/Ca-SST was calculated using the calibration equation Sr/Ca-SST = -12.056*Sr/Ca + 138.2681 derived from reduced major axis regression of modern microatoll XM22 Sr/Ca and ERSSTv3b SSTs for 158W, 2N over the period 1994 to 2007. The Sr/Ca and ERSSTv3b show excellent correspondence (R2 = 0.64). Dd18O was calculated using the centering method (Cahyarini et al., 2008) with a d18O-SST slope of 0.15 per mil per degC (McGregor et al., 2011) and the Sr/Ca-SST slope above. XM22 d18O data are from McGregor et al. 2011, and XM35 d18O data are from McGregor et. al. 2013, and the full d18O datasets are archived with WDC-Paleoclimatology. Coral XM35, age: ~4.3 kyBP, location: Kiritimati Island (01°44.212'N, 157°12.522'W).
ABSTRACT SUPPLIED BY ORIGINATOR: Earth's interannual climate variability is dominated by El Niño/Southern Oscillation (ENSO). Palaeoclimate records indicate a lower ENSO variance during the middle Holocene compared with today; however, model simulations have not reproduced the full magnitude of the changes, and whether external forcing drives large intrinsic ENSO variability is therefore a matter of considerable debate. Here we present a 175-year-long, monthly resolved oxygen isotope record, obtained from a Porites coral microatoll located on Kiritimati (Christmas) Island, in the NINO3.4 region of the central equatorial Pacific. Our quantitative record of ENSO variability about 4,300 years ago shows that ENSO variance was persistently reduced by 79%, compared with today, and it exhibits a dominant annual cycle. Season-specific analysis shows that El Niño events were damped during their September-November growth phase, and delayed relative to the climatological year. We suggest that the higher boreal summer insolation at the time strengthened the tropical Pacific zonal winds as well as the gradients in sea surface temperature, and thereby led to an enhanced annual cycle and suppressed ENSO. As the weak ENSO is subject to interdecadal amplitude modulation, we conclude that amplitude modulation is likely to remain robust under altered climates. Our findings show that ENSO is capable of responding to external forcing.