On the climatic contamination of relative paleointensity estimations from sediments

A fundamental problem of relative paleointensity (RPI) estimations using marine sediments is that magnetic compositional variations induced by environmental changes may influence RPI estimations, which is known as climatic contamination or lithological contamination. Valet et al. (2011) demonstrated this problem by isolating a climatic contamination component from some RPI records using principal component analysis. In this presentation, we discuss the origin of the climatic contamination and a possible way to mitigate it. Recent RPI studies often show a negative correlation between RPI and magnetofossil abundance estimated from the ratio of ARM susceptibility to SIRM and FORC diagrams. This indicates that PRI recording efficiency of the magnetofossil component is lower than that of the detrital magnetic component. This implies that changing proportion of the magnetofossil component to the detrital component in magnetic mineral assemblages of sediments can be a source of the climatic contamination. When the coercivity distributions of the magnetofossil and detrital components are different, NRM-ARM demagnetization diagram (a variant of the pseudo-Thellier plot of Tauxe et al., 1995) shows curvature, and less contaminated RPI estimations may be possible by choosing a coercivity fraction reflecting the magnetofossil or detrital components on calculating a best-fitting slope. Linear NRM-ARM demagnetization diagram was sometimes used as a criterion for reliable RPI estimations, but largely overlapping coercivity distributions of the two components yield linear NRM-ARM demagnetization despite changing proportion of the two components results in contaminated RPI.