Bibliographie complète

Abstract Based on the analysis of fish otolith assemblages from surface sediments of the Lomonosov Ridge (Arctic Ocean), we demonstrate that the very low Holocene sedimentation rates and winnowing of fine sediments result in the mixing of the whole Holocene populations at the sediment surface. Specimens from the Marine Isotope Stage (MIS) 3 or older could even be recovered in the surface due to a sedimentary hiatus at some locations in the central Arctic during the last glacial maximum. Two examples illustrate that 14 C‐stratigraphies from planktic foraminifers in underlying cored sediments reflect the mixing between Holocene and MIS 3 or older populations, thus invalidating continuous age‐depth inferences based on 14 C ages. Hence, much caution is required when attempting to set paleoceanographic reconstructions based on 14 C chronologies in a low sediment accumulation rate environment such as the central Arctic Ocean. Already published paleoceanographic reconstructions from this area might thus require some revisions. , Plain Language Summary Radiocarbon ages of microfossils (fish otoliths) collected at the surface sediments of the Lomonosov Ridge, in the central Arctic Ocean, indicate that all populations that developed during the present interglacial are mixed within the approximately 1 cm‐thick surface layer. Fossil assemblages occasionally include specimens from older warm intervals. The stacking of fossil spanning thousands of years is due to the very low sediment accumulation rate of the area, the post‐depositional winnowing of fine sediments and mixing by benthic organisms. These process result in the impossibility to document the faunal evolution in the central Arctic Ocean during the last few tens of thousands of years using such fossils. , Key Points Fish otolith radiocarbon age distributions in surface sediments illustrate the mixing of Holocene and pre‐Last Glacial Maximum populations Low sedimentation rates, particle winnowing and sedimentary gaps may impact microfossil mixing and 14 C chronologies Published paleoclimate/paleoceanographic records from similar sites might thus require some reinterpretation