Distinct magnetic properties of marine sediments that record the Palaeocene-Eocene thermal maximum (PETM) have been suggested to be due to a bacterial magnetofossil signal that is linked to enhanced weathering conditions during the PETM. We document the dominance of bacterial magnetite in deep-sea sediments from southern Kerguelen Plateau (Ocean Drilling Program Hole 738C, southern Ocean) not only during the PETM, but also before and after the thermal event. This occurrence of magnetofossils throughout the PETM indicates that the occurrence of bacterial magnetosomes is not due to a preservation effect. Instead, we suggest that it is due to sustained mild iron-reducing conditions that dissolved the most labile aeolian-derived iron, which favoured continued magnetotactic bacterial activity without being strong enough to dissolve the less reactive magnetite and haematite. Enhanced aeolian haematite abundances at the beginning of the PETM indicate drier conditions on the neighbouring Antarctic continent at those times. Our results provide evidence that iron fertilisation by aeolian dust was the main limiting factor that conditioned proliferation of magnetotactic bacteria in the deep sea at the southern Kerguelen Plateau, with the exception of two short periods of rapidly changing palaeoenvironmental conditions at the onset and termination of the PETM. Increased iron supply from aeolian dust, that enhanced oceanic primary productivity and subsequent delivery of organic carbon to the seafloor, along with mild iron-reducing diagenetic conditions, seem to have been necessary to provide the iron needed for magnetite biomineralization by magnetotactic bacteria to drive their marked increase in abundance in the studied PETM record from southern Kerguelen Plateau. Our analyses of a deep-sea PETM record from Hole 1051B at Blake Nose (Atlantic Ocean) failed to identify magnetofossils despite evidence for the occurrence of magnetite and haematite of probable aeolian origin. Contrasting magnetic properties at these PETM sections indicate that further work is needed to understand the palaeoenvironmental and diagenetic factors whose interactions lead to production and preservation of magnetofossils in deep-sea sediments.
