Mid-Pleistocene (800-400 ka) records of benthic and planktic foraminfer d18O and d13C, planktic foraminifer based SST, and IRD of IODP Site 306-U1314

Stable carbon and oxygen isotopes from benthic and planktic foraminifers, planktic foraminifer assemblages and ice rafted debris from the North Atlantic Site U1314 (Integrated Ocean Drilling Program Expedition 306) were examined to investigate orbital and millennial-scale climate variability in the North Atlantic and its impact on global circulation focusing on the development of glacial periods during the mid-Pleistocene (ca 800-400 ka). Glacial initiations were characterized by a rapid cooling (6-10 °C in less than 7 kyr) in the mean annual sea surface temperature (SST), increasing benthic d18O values and high benthic d13C values. The continuous increase in benthic d18O suggests a continuous ice sheet growth whereas the positive benthic d13C values indicate that the flow of the Iceland Scotland Overflow water (ISOW) was vigorous. Strong deep water formation in the Norwegian Greenland Sea promoted a high transfer of freshwater from the ocean to the continents. However, low SSTs at Site U1314 suggest a subpolar gyre cooling and freshening that may have reduced deep water formation in the Labrador Sea during glacial initiations. Once the 3.5 per mil threshold in the benthic d18O record was exceeded, ice rafting started and ice sheet growth was punctuated by millennial-scale waning events which returned to the ocean part of the freshwater accumulated on the continents. Ice-rafting events were associated with a rapid reduction in the ISOW (benthic d13C values dropped 0.5-1 per mil) and followed by millennial-scale warmings. The first two millennial-scale warm intervals of each glacial period reached interglacial temperatures and were particularly abrupt (6-10 °C in ~3 kyr). Subsequent millennial-scale warm events were cooler probably because the AMOC was rather reduced as suggested by the low benthic d13C values. These two abrupt warming events that occurred at early glacial periods were also observed in the Antarctic temperature and CO2 records, suggesting a close correlation between both Hemispheres. The comparison of the sea surface proxies with the benthic d18O record (as the Southern sign) indicates the presence of a millennial-scale seesaw pattern similar to that seen during the Last Glacial period.

(Figure 3d) Export production reconstruction based on planktonic foraminiferal sampled in surface sediments

To reconstruct Export Productivity (Pexp), 27 taxonomic categories of the planktonic foraminifera census data were used with the modern analog technique SIMMAX 28 (Pflaumann et al., 1996, doi:10.1029/95PA01743; 2003, doi:10.1029/2002PA000774). To the 26 taxonomic groups widely used and listed in Kucera et al. (2005, doi:10.1016/j.quascirev.2004.07.014), Turborotalita humilis was added in our calibration as it is associated with the PCC source region (Meggers et al., 2002, doi:10.1016/S0967-0645(02)00103-0). The modern analog file is based on the Iberian margin database (Salgueiro et al., 2008, doi:10.1016/j.marmicro.2007.09.003) combined with the North Atlantic surface samples used by the MARGO project (Kucera et al., 2005). This results in a total of 999 analogs for Pexp. Modern oceanic primary productivity (PP) is obtained for each site by averaging 12 monthly primary productivity values for a 8-year period (1978-1986) that were estimated from satellite color data (CZCS) and gridded at 0.5° latitude - longitude fields (Antoine et al., 1996, doi:10.1029/95GB02832). Export Productivity (Pexp) was calculated from the PP values following the empirical relationship Pexp = PP**2/400 for primary production below 200 gC/m**2/yr, and Pexp = PP/2 for primary production above 200 gC/m2/yr (Eppley and Peterson, 1979, doi:10.1038/282677a0; Sarnthein et al., 1988, doi:10.1029/PA003i003p00361). The residuals gives the differences between satellite based Pexp and foraminiferal Pexp.