A Ternary Mixing Model Approach Using Benthic Foraminifer δ13C‐δ18O Data to Reconstruct Late Pliocene Deep Atlantic Water Mass Mixing

Abstract

AbstractLate Pliocene deep Atlantic δ13C data have been interpreted as evidence for enhanced Atlantic Meridional Overturning Circulation (AMOC) compared to the present, but this hypothesis is not supported by the Pliocene Model Intercomparison Project (PlioMIP). Here, we adopt a new approach to assess variability in deep ocean circulation based on paired stable carbon (δ13C) and oxygen isotopes (δ18O) of benthic foraminifera, both (semi)conservative water mass tracers. Assuming that deep Atlantic benthic δ13C‐δ18O variability is predominantly driven by mixing, we extrapolate the δ13C‐δ18O data outside the sampled range to identify the end‐members. At least three end‐members are needed to explain the spatial δ13C‐δ18O variability in the deep North Atlantic Ocean: two Northern Component Water (NCW) and one Southern Component Water (SCW) water masses. We use a ternary mixing model to quantify the mixing proportions between SCW and NCW in the deep Atlantic Ocean. Our analysis includes new data from Ocean Drilling Program Sites 959 and 662 in the eastern equatorial Atlantic and suggests that the AMOC cell was deeper during the M2 glacial than during late Pliocene interglacials. Moreover, we identify a new cold and well‐ventilated water mass that was geographically restricted to the southeast Atlantic Ocean between 3.6 and 2.7 Ma and did not contribute significantly to the δ13C‐δ18O variability of the rest of the basin. This high‐δ13C high‐δ18O water mass has led to the misconception of a reduced latitudinal δ13C gradient. Our analyses show that the late Pliocene δ13C gradient between NCW and SCW was similar to the present‐day value of 1.1‰.