Planktonic foraminifera oxygen isotope and trace metal data ODP Hole 172-1056A, 46-54 ka BP

Salinity increase in the subtropical gyre system may have pre-conditioned the North Atlantic Ocean for a rapid return to stronger overturning circulation and high-latitude warming following meltwater events during the Last Glacial period. Here we investigate the Gulf Stream - subtropical gyre system properties over Dansgaard-Oeschger (DO) cycles 14 to 12, including Heinrich ice-rafting event 5. During the Holocene and Last Glacial Maximum a positive gradient in surface dwelling planktonic foraminifera d18O (Globigerinoides ruber) can be observed between the Gulf Stream and subtropical gyre, due to decreasing temperature, increasing salinity, and a change from summer to year-round occurrence of G. ruber. We assess whether this gradient was a common feature during stadial-interstadial climate oscillations of Marine Isotope Stage 3, by comparing existing G. ruber d18O from ODP Site 1060 (subtropical gyre location) and new data from ODP Site 1056 (Gulf Stream location) between 54 and 46 ka. Our results suggest that this gradient was largely absent during the period studied. During the major warm DO interstadials 14 and 12 we infer a more zonal and wider Gulf Stream, influencing both ODP Sites 1056 and 1060. A Gulf Stream presence during these major interstadials is also suggested by the large vertical d18O gradient between shallow dwelling planktonic foraminifera species, especially G. ruber, and the deep dwelling species Globorotalia inflata at site 1056, which we associate with strong summer stratification and Gulf Stream presence. A major reduction in this vertical d18O gradient from 51 ka until the end of Heinrich event 5 at 48.5 ka suggests site 1056 was situated within the subtropical gyre in this mainly cold period, from which we infer a migration of the Gulf Stream to a position nearer to the continental shelf, indicative of a narrower Gulf Stream with possibly reduced transport.

Chemical composition of ocean-floor basalts and separated magnetites DSDP Legs 37, 45 and 46

Titanomagnetites separated from 15 different rock samples (including ocean-floor basalts from DSDP Legs 37, 45 and 46) were analyzed together with whole-rock samples by instrumental neutron-activation analysis for Sc, Cr, Co, Zn, Hf, Ta, Th and the REE La, Ce, Nd, Sm, Eu, Gd, Tb, Dy, Tm, Yb and Lu. In titanomagnetities from ocean-floor basalts and some other rocks, REE are enriched with respect to the whole-rock composition by factors of between 1.5 and 3 for light REE and between 1.0 and 1.9 for heavy REE; that is, REE with larger ionic radii are preferentially incorporated into the magnetite lattice. Three magnetite samples are REE depleted. Their whole-rock samples contain P in appreciable amounts, so apatite, an important REE-accumulating mineral, could have captured REE to some extent. All titanomagnetites show a marked negative Eu anomaly, this is most probably caused by discrimination of Eu(2+) from the magnetite lattice. Co, Zn, Hf and Ta are significantly enriched in magnetites. The distribution behaviour of Sc and Cr is masked chiefly by the crystallization of clinopyroxene and therefore is not easy to estimate. Ulvöspinel contents of about 70% for the titanomagnetites from ocean-floor basalts were estimated from qualitative microprobe analysis. Ulvöspinel contents of all other samples varied in a wide range from 20% to about 90%. No correlation could be observed between this and the REE contents of the magnetites. Ilmenite exsolution lamellae could only be observed in titanomagnetites from a doleritic basalt from Leg 45.