Temporally variable diagenetic overgrowth on deep-sea nannofossil carbonates across Palaeogene hyperthermals and implications for isotopic analyses

Abstract

Abstract Calcareous nannofossil assemblages of deep-sea sediments were subjected to intensive diagenetic alterations during early Palaeogene hyperthermal events. These alterations may have significantly modified bulk isotopic and trace metal signals and nannofossil preservation, thus biassing palaeoceanographic and palaeoecological interpretation. We present a detailed characterisation of the temporal variation in degree of diagenetic overgrowth on nannoliths during the PETM and Early Eocene Thermal Maximum (ETM2) using scanning electron microscopy (SEM), and explore in detail the consequences of these changes in overgrowth for interpretation of nannofossil assemblages and geochemical records covering the ETM2 at ODP Site 1265 where the event is well recognised. Results show that the nannofossil genera Discoaster and Zygrhablithus are particularly receptive to significant amounts of diagenetic calcite overgrowth, which was confirmed by Sr/Ca variations within single discoasters. Overgrowths show a strong correlation with changes in sediment carbonate content across the hyperthermals, with notably less overgrowth in low carbonate intervals. This secondary calcite affects stable isotope, notably oxygen isotopes, and assemblage composition modifying the dissolution susceptibility of taxa. In particular, the size fraction with a high contribution of overgrown discoasters has heavier δ18O values. Size fractions that are mainly composed of primary calcite give lighter δ18O values, and reveal a conspicuous early warming trend across ETM2 that appears to be attenuated in fractions with a major contribution of secondary calcite. Hence, coccolith-based indices based on the degree of overgrowth may be used to evaluate the degree of diagenetic alteration, as to improve precision and accuracy of the interpretation of nannofossil-based stable isotope records and abundance data.