Oxygen isotope composition of the final chamber of planktic foraminifera provides evidence of vertical migration and depth-integrated growth

The translation of the original seawater signal (i.e. ambient temperature and delta O-18(sw)) into distinct chambers of a single shell of a foraminifer during calcification can influence our interpretation of surface ocean conditions of the past, when based upon oxygen and carbon stable isotope geochemistry. In this study three different hypotheses were tested to gain more insight into biological and ecological processes that influence the resultant composition of stable isotopes of oxygen (delta O-18) in the shells of planktonic foraminifera. These hypotheses were related to the shell size; the differences in isotopic composition between the final chamber and the remaining shell; and the differences between different species. Shells of Trilobatus sacculifer, Globigerinoides ruber white and Neogloboquadrina dutertrei were picked from the top of multi-core GS07-150-24, of modern age, offshore of north-eastern Brazil (3 degrees 46.474' S, 37 degrees 03.849' W) and analysed for single-shell and single-chamber stable isotope analysis. We show that the mean value of delta O-18 of the final chambers (delta O-18(F)) is 0.2% +/- 0.4% (1 sigma) higher than the mean value delta O-18 of the test minus the final chamber (delta O-18 (< F)) of T. sacculifer. The formation of the final chamber happens at temperatures that are approximately 1 degrees C cooler than the chambers formed prior, suggesting both ontogenetic depth migration to deeper water and a potential offset from the surface signal. Furthermore, we show that there is no statistical difference in the delta O-18(sacculifer) values of shells of three different size classes of T. sacculifer, although the pattern between the different size classes indicates depth migration during the life and growth of T. sacculifer. Comparison of vital effect corrected delta O-18(shell) between T. sacculifer, G. ruber white and N. dutertrei suggests that G. ruber has a slightly shallower depth habitat (similar to 90-120 m) compared to the other two species (similar to 100-130 m). Disentangling depth vs. seasonal habitat is complicated given the commonality between isotope values from similar depths but different seasons; for instance, the same average isotope value will have a shallower depth habitat in May than September. Calculation of seasonal-depth habitat was therefore tested. Our results highlight the complicated nature of interpreting oxygen isotopes even for the modern record.