Reconstructing Tropical Atlantic Hydrography Using Planktontic Foraminifera and an Ocean Model

In the tropical Atlantic, planktonic fomminfera species are vertically distributed with highest abundances occunfng in the photic zone (approximately 0-100 m). The tropical Atlantic thermocline dips from east to west and varies seasonally due to changes in the southeast and northeast trade winds. In the east, the thermocline is in the photic zone, and in the west, the well-mixed surface layer extends below the photic zone most of the year. As expected from species vertical distributions in plankton tows, the species assemblages on the seafloor are correlated to the hydrographic onditions of the overlying surface ocean layer. A new technique to reconstruct past tropical Atlantic (20øN to 20øS) photic zone hydrography and surface wind field uses faunal assemblage data from deep-sea cores. Planktonic foraminifera bundances incore tops correlate with observations of modern photic zone hydrography defined here as seasonal temperature variation and mixed layer depth. The hydrography is mathematically described using empirical orthogonal function (EOF) analysis of annual temperature ange as a function of depth. Factor analysis of 29 species of planktonic foraminifera from 118 core tops produces three factors. The factors correlate to mixed layer depth and the two EOF modes. The ocean model of the Atlantic ocean produces imilar map patterns of the EOF modes. Therefore the Copyright 1990 by the American Geophysical Union. Paper number 90PA00484. 08838305/90/90PA-00484510.00 model can be used to simulate hydrographic changes to compare with faunal predicted past hydrographic changes. Since the ocean model is wind driven, this approach provides a way of evaluating the validity of estimates of past wind stress changes and the contribution of these changes to the faunal changes in the past. A double wind stress run indicates that the central and eastern equatorial and southeast regions of the study area are most sensitive to wind stress increases. Factor analysis of the foraminifera abundances from the last glacial maximum (LGM) shows that species associations change downcore and demonstrates how the methods developed in this study can be applied. Comparison of the double wind stress experiment and the LGM faunal changes indicates ome areas of significant agreement suggesting that faunal changes may reflect thermocline su'ucmre response to the LGM wind field. Discrepancies may reflect the fact that uniform changes in the north and south trade wind strengths did not occur at the LGM.