Alkenone-derived UK'37 data and sea surface temperatures (SST) of sediment cores from the Gulf of Taranto (Italy)

Auxiliary data include one file with alkenone-derived UK'37 data and sea surface temperatures (SST). On these data Figs. 7 and 8 of the manuscript are based. The SST are derived from UK'37 by using the transfer function: SST = 29.876 UK'37 - 1.334 of Conte et al. (2006). The data are against the ages (in A.D.) of samples derived from cores GT91-1 (39[deg]59'23"N, 17[deg]45'25"E), GT89-3 and GT90-3 (both 39[deg]45'43"N, 17[deg]53'55"E ). Also included are composite records for UK'37 and SST. For creating the composite records, GT-89-3 was taken as reference core. In the overlapping period the GT89-3 data seem in general lower than the GT91-1 data. To accommodate for this in the composite record, the average difference (0.0343 UK'37 units; equivalent to 1.023 [deg]C) was subtracted from the GT91-1 record. Hereafter, for each depth in the overlapping interval the respective values (UK'37 or SST) of GT89-3 and GT91-1 were averaged. We have also averaged with 16 additional alkenone measurements, from 1793 to 1851, performed in the GT90-3 core.

(Table 1) Summary of coral geochemical data and reconstructed SST

We analyzed strontium/calcium ratios (Sr/Ca) in four colonies of the Atlantic coral genus Montastrea with growth rates ranging from 2.3 to 12.6 mm/a. Derived Sr/Ca-sea surface temperature (SST) calibrations exhibit significant differences among the four colonies that cannot be explained by variations in SST or seawater Sr/Ca. For a single coral Sr/Ca ratio of 8.8 mmol/mol, the four calibrations predict SSTs ranging from 24.0° to 30.9°C. We find that differences in the Sr/Ca-SST relationships are correlated systematically with the average annual extension rate (ext) of each colony such that Sr/Ca (mmol/mol) = 11.82 (±0.13) - 0.058 (±0.004) * ext (mm/a) - 0.092 (±0.005) * SST (°C). This observation is consistent with previous reports of a link between coral Sr/Ca and growth rate. Verification of our growth-dependent Sr/Ca-SST calibration using a coral excluded from the calibration reconstructs the mean and seasonal amplitude of the actual recorded SST to within 0.3°C. Applying a traditional, nongrowth-dependent Sr/Ca-SST calibration derived from a modern Montastrea to the Sr/Ca ratios of a conspecific coral that grew during the early Little Ice Age (LIA) (400 years B.P.) suggests that Caribbean SSTs were >5°C cooler than today. Conversely, application of our growth-dependent Sr/Ca-SST calibration to Sr/Ca ratios derived from the LIA coral indicates that SSTs during the 5-year period analyzed were within error (±1.4°C) of modern values.

Age model, raw data and interpolated raw data from proxy records of sediment cores GeoB6007-1 and GeoB6007-2

Here we present a 1200 yr long benthic foraminiferal Mg/Ca based temperature and oxygen isotope record from a ~900 m deep sediment core off northwest Africa to show that atmosphere-ocean interactions in the eastern subpolar gyre are transferred at central water depth into the eastern boundary of the subtropical gyre. Further we link the variability of the NAO (over the past 165 yrs) and solar irradiance (Late Holocene) and their control on subpolar mode water formation to the multidecadal variability observed at mid-depth in the eastern subtropical gyre. Our results show that eastern North Atlantic central waters cooled by up to ~0.8± 0.7 °C and densities decreased by Sigma theta=0.3±0.2 during positive NAO years and during minima in solar irradiance during the Late Holocene. The presented records demonstrate the sensitivity of central water formation to enhanced atmospheric forcing and ice/freshwater fluxes into the eastern subpolar gyre and the importance of central water circulation for cross-gyre climate signal propagation during the Late Holocene.