Description: Using bathymetric transects of surface sediments underlying similar sea surface temperatures but exposed to increasing dissolution, we examined the processes which affect the relationship between foraminiferal Mg/Ca and d18O. We found that Globigerinoides saccculifer calcifies over a relatively large range of water depth and that this is apparent in their Mg content. On the seafloor, foraminiferal Mg/Ca is substantially altered by dissolution with the degree of alteration increasing with water depth. Selective dissolution of the chamber calcite, formed in surface waters, shifts the shell's bulk Mg/Ca and d18O toward the chemistries of the secondary crust acquired in colder thermocline waters. The magnitude of this shift depends on both the range of temperatures over which the shell calcified and the degree to which it is subsequently dissolved. In spite of this shift the initial relationship between Mg/Ca and d18O, determined by their temperature dependence, is maintained. We conclude that paired measurements of d18O and Mg/Ca can be used for reconstructing d18Owater, though care must be taken to determine where in the water column the reconstruction applies.

Description: The Mg/Ca of planktic foraminifera Globigerinoides ruber (white) is a widely applied proxy for tropical and sub-tropical sea-surface temperature. The accuracy with which temperature can be reconstructed depends on how accurately relationships between Mg/Ca and temperature and the multiple secondary controls on Mg/Ca are known; however, these relationships remain poorly quantified under oceanic conditions. Here, we present new calibrations based on 440 sediment trap/plankton tow samples from the Atlantic, Pacific and Indian Oceans, including 130 new samples from the Bay of Bengal/Arabian Sea and the tropical Atlantic Ocean. Our results indicate temperature, salinity and the carbonate system all significantly influence Mg/Ca in G. ruber (white). We propose two calibration models: The first model assumes pH is the controlling carbonate system parameter. In this model, Mg/Ca has a temperature sensitivity of 6.0 ± 0.8 %/°C (2 Sigma), a salinity sensitivity of 3.3 ± 2.2%/PSU and a pH sensitivity of -8.3 ± 7.7%/0.1 pH units; The second model assumes carbonate ion concentration ([CO3**2-]) is the controlling carbonate system parameter. In this model, Mg/Ca has a temperature sensitivity of 6.7 ± 0.8%/°C, a salinity sensitivity of 5.0 ± 3.0%/PSU and a [CO3**2-] sensitivity of -0.24 ± 0.11/µmol kg**1. In both models, the temperature sensitivity is significantly lower than the widely-applied sensitivity of 9.0 ± 0.6%/°C. Application of our new calibrations to down-core data from the Last Glacial Maximum, considering whole ocean changes in salinity and carbonate chemistry, indicate a cooling of 2.4 ± 1.6 °C in the tropical oceans if pH is the controlling parameter and 1.5 ± 1.4 °C if [CO3**2-] is the controlling parameter.

Description: Global cooling and the development of continental-scale Antarctic glaciation occurred in the late middle Eocene to early Oligocene (~38 to 28 million years ago), accompanied by deep-ocean reorganization attributed to gradual Antarctic Circumpolar Current (ACC) development. Our benthic foraminiferal stable isotope comparisons show that a large d13C offset developed between mid-depth (~600 meters) and deep (〉1000 meters) western North Atlantic waters in the early Oligocene, indicating the development of intermediate-depth d13C and O2 minima closely linked in the modern ocean to northward incursion of Antarctic Intermediate Water. At the same time, the ocean's coldest waters became restricted to south of the ACC, probably forming a bottom-ocean layer, as in the modern ocean. We show that the modern four-layer ocean structure (surface, intermediate, deep, and bottom waters) developed during the early Oligocene as a consequence of the ACC.