Correcting relative paleointensity records for variations in sediment composition: Results from a South Atlantic stratigraphic network

Abstract

Marine sediments record direction and intensity of the Earth's magnetic field by the alignment of magnetic particles during deposition. For determining relative paleointensity (RPI) from sediment records it is commonly assumed that their natural remanent magnetization (NRM) is proportional to the Earth's magnetic field during deposition, and also proportional to the concentration of remanence carriers in the sediment layer. However, little is known how varying sediment composition and environmental conditions during deposition influence the NRM. Here we try to identify and quantify such sedimentary influences for eight sediment series from the subtropical and subantarctic South Atlantic. The cores were recovered in a constraint area crossing the subtropical front (STF). They have widely different sediment lithologies, which can be divided into three lithologic groups. Due to their mutual proximity, they have experienced approximately the same magnetic field history, and differences in their RPI signals must be caused by their varying sediment composition and recording properties. Based on high resolution rock magnetic and compositional data from two previous studies it is possible to quantitatively test and compare the influences of different sediment properties upon the NRM. It is found that magnetic grain size, as measured by the magnetic parameter ARM/IRM, is most influential among the parameters tested. Weak to moderate reductive diagenesis, as measured by the parameter Fe/ κ, turns out to have minor impact. By comparing the sensitivity of different normalization procedures for RPI determination, it is found that induced remanent magnetization (IRM) is most robust. Based on an extended linear RPI theory, we can calculate a corrected RPI stack for the investigated cores. This correction improves the correlation with independent global paleointensity stacks in comparison to our previous uncorrected RPI stack (Hofmann, D., Fabian, K., 2007. Rock-magnetic properties and relative paleointensity stack for the last 300 ka based on a stratigraphic network from the subtropical and subantarctic South Atlantic. Earth Planet. Sci. Lett. 260, 297–312.). The ratio between corrected and uncorrected RPI stacks reveals a hidden global climate signal, which indicates that climatic variations in sediment composition are inevitably present in non-ideal sediment sequences.