Global prediction of planktic foraminiferal fluxes from hydrographic and productivity data

Understanding and quantifying the seasonal and spatial distribution of planktic foraminiferal fluxes reflected in sedimentary assemblages is key to interpret foraminiferabased proxies in paleoceanography. Towards this goal we present an empirical model to predict foraminiferal fluxes on a global scale. A compilation of planktic foraminiferal flux and export production data from globally distributed sediment traps together with environmental data of sea-surface temperature and mixed-layer depth from online databases is used to calibrate the model that calculates monthly foraminiferal fluxes for the 18 most common species. The calibrated model is then forced with a global data set of hydrographic and productivity data to predict monthly foraminiferal fluxes worldwide. The predictive skills of the model are assessed by comparing the model output with planktic foraminiferal assemblages from globally distributed surface sediments as well as with measured foraminiferal fluxes of sediment traps not included in the calibration data set. Many general distribution patterns of foraminiferal species recognized from the model output compare favorably with observations from coretops or sediment traps, even though the model still produces problematic results in some places. Among others, meridional gradients in species richness and diversity, increased relative abundances of Neogloboquadrina pachyderma (dex.) in upwelling areas, and peak abundances of Globigerinella siphonifera in oligotrophic subtropical gyres show good agreement between model and coretops. Absolute foraminiferal fluxes are significantly underestimated in most cases, while seasonal variations can be reproduced for some species. Interannual differences in foraminiferal fluxes are not reflected by the model which might partly be due to a lack of actual environmental data for the calibration and model experiments. Correspondence to: S. Žarić ([email protected]) The limited predictive skills of the model suggest that additional parameters should be considered. Export production should probably be replaced by a more realistic representation of food availability for planktic foraminifera. This could be achieved by adding a dynamic component to the model and linking it to an ecosystem model.