A diatom based transfer function for reconstructing sea ice concentrations in the North Atlantic

A diatom based sea ice transfer function is developed using 99 surface sediment samples from the North Atlantic and the associated modem sea ice concentrations. Canonical correspondence analysis (CCA) is applied to the species assemblages of the surface sediment samples and the association of the species with two environmental parameters, August sea surface temperature and May sea ice concentration, is assessed. The results of this analysis indicate negative correlation between sea ice and sea surface temperature and that a group of diatom species is strongly associated with sea ice, especially May sea ice concentration. The results of the CCA legitimate the development of a diatom based sea ice transfer function. The maximum likelihood method has been applied as the transfer function method, as it has been proven most suitable with this particular data set. The newly developed transfer function is then used to reconstruct May sea ice concentration in three cases, each focusing on a different time period: the Last Glacial Maximum, the Younger Dryas and the Holocene. In all three cases the transfer function produces reasonable results when compared to other paleoclimatic proxy results. This suggests that the sea ice concentration reconstructed by the diatom based sea ice transfer function is a valid and reliable method, which can be applied as a valid proxy for May sea ice concentration. (c) 2007 Elsevier B.V. All rights reserved.