Modeling the Spatial and TemporalDistribution of PlanktonicForaminifera

Abstract

Planktonic foraminifera contribute substantially to the fossil record of marine sediments,and due to their excellent preservation in sediments, the fossil shells areof great paleoceanographic significance. Temporal variations in sedimentary faunalassemblages and geochemical composition of the individual fossil shells havebecome important proxies for past oceanic conditions. Different ecological preferencesbetween species cause distinct seasonal patterns, and the imprint of differingseasonal production is preserved in the sedimentary record. Accurate knowledge offoraminiferal ecology and seasonal succession is indeed crucial to correctly interpretcorresponding proxy-based reconstructions. Planktonic foraminifera are known tobe sensitive to temperature, and therefore climate changes may alter the seasonalpattern of species. Any change in the timing of the largest flux to the seafloor thatmay have occurred in the past will lead to a bias in estimated paleotemperature.Therefore, this dissertation focuses on the variability of planktonic foraminiferalseasonality through time, and discusses its implications for paleotemperature reconstructions.A numerical model simulating the population dynamics of planktonic foraminiferawas developed and coupled to an existing marine ecosystem model. This modelis forced with a global hydrographic dataset (e.g., temperature, mixed layer depth)and with biological information taken fromthe ecosystem model to predict monthlyconcentration of the following foraminifera species: Neogloboquadrina pachyderma(dextral and sinistral varieties), Globigerina bulloides, Globigerinoides ruber (white variety)and Globigerinoides sacculifer. These species are sensitive to sea-surface temperature,and due to their high spatial coverage and abundance are the most commonplanktonic foraminifera species used in paleoceanography. The model results forthe global distribution of planktonic foraminifera for modern conditions are comparedto available core-top and sediment-trap data. In the North Atlantic, modelprediction is compared to the living population collected by plankton-nets.The modeled spatial distribution of most of the species compares favorably withcore-top data. The model prediction indicates that polar regions are dominated byN. pachyderma (sin.); N. pachyderma (dex.) and G. bulloides are the most commonspecies in high productivity zones; and tropical-subtropical species like G. ruber andG. sacculifer are more abundant in oligotrophic waters. The predicted seasonal fluxpatterns coincided with sediment-trap records inmost of the locations, although thecomparison was hampered by interannual variability not captured by the model.Using the foraminifera model, we carried out sensitivity experiments to studythe response of foraminifera to different boundary conditions. A sensitivity experimentusing a constant temperature of 12C indicated that food availability is animportant factor controlling foraminiferal distribution.Another sensitivity experiment consisted of decreasing the temperature globallyby 2C and 6C, and assessing the influence of this temperature variation onthe recorded signal. In most of the regions at mid and high latitudes, due to thecooling and temperature sensitivity of the species, maximum production shiftedto a warmer season. Thus, the foraminiferal population as a whole recorded littlechange in the temperature. By contrast, in tropical waters, where temperature cyclehas relatively low amplitude, the recorded signal is close to annual mean SSTregardless of the timing of maximum production. Therefore, at low latitudes foraminiferarecorded the entire temperature change. These experiments emphasize theimportance of considering changes in seasonality through time, as they can maskthe total temperature variation.Finally, we studied the response of foraminifera to the boundary conditions ofthe Last Glacial Maximum. We forced the foraminifera model using the physical andchemical parameters predicted by coupled climate models. In tropical waters variationsin foraminiferal seasonality did not cause significant change in the recordedtemperature. By contrast, at high latitudes the foraminiferal flux to the sea floorhas a pronounced seasonal cycle, and the amplitude of temperature seasonality isalso high. Therefore, changes in the seasonality of foraminifera had large influenceon the seasonal imprint of the fossil record. The assessment and quantification ofseasonal bias on a global scale allows the improvement of foraminifera-based proxycalibrations.