Understanding the Role of Biogenic Magnetite in Geomagnetic Paleointensity Recording: Insights From Ontong Java Plateau Sediments

Abstract

AbstractMarine sediments can preserve continuous paleomagnetic intensity records. Different magnetic minerals may acquire remanent magnetizations differently, so that compositional variations of magnetic mineral assemblages in sediments may hinder extraction of reliable relative paleointensity (RPI) records. To better understand this issue, we conducted a paleo‐ and rock magnetic study of a sediment core from the Ontong Java Plateau, western equatorial Pacific Ocean. RPI estimated by normalizing natural remanent magnetization (NRM) with anhysteretic remanent magnetization (ARM) decreases downcore with an inverse correlation with the ratio of ARM susceptibility to saturation isothermal remanent magnetization. This relationship indicates that the RPI signal weakens as the proportion of biogenic magnetite increases. The NRM–ARM demagnetization diagrams obtained have concave‐down curvature. These observations indicate that the RPI recording efficiency of the biogenic component is lower than that of the terrigenous component if we assume that the magnetizations of the high‐ and low‐coercivity windows are carried dominantly by biogenic and terrigenous components, respectively. This assumption is supported by first‐order reversal curve measurements, transmission electron microscope observations, low‐temperature measurements, and extraction of silicate‐hosted magnetic inclusions from the sediments. Previous studies have suggested that the RPI recording efficiency of biogenic magnetite is higher than that of the terrigenous component, which disagrees with our results. Different concentrations of silicate‐hosted magnetic inclusions in different sedimentary environments might explain this contradiction. We conclude that biogenic magnetite contributes to RPI records with lower efficiency than unprotected terrigenous magnetic minerals in the studied sediments. Changing biogenic magnetite proportions distort ARM‐normalized RPI estimations.