Magnetic susceptibility as a high-resolution correlation tool and as a climatic proxy in Paleozoic rocks - Merits and pitfalls: Examples from the Devonian in Belgium

Low-field magnetic susceptibility (χ_in) measurements are quick and sensitive enabling the creation of high-resolution records; making χ_in an attractive correlation tool and a proxy for paleoclimate and paleoenvironments. In geologically young material - foremost in Cenozoic sediments - χ_in belongs to the geoscientist's toolkit. However, χ_in is a convolved signal and may reflect other processes than the often implicitly inferred depositional conditions. Diagenesis, remagnetization and low-grade metamorphism, can potentially obscure the original, depositional, χ_in signal. This aspect is particularly important when interpreting χ_in records from Paleozoic rocks.Here, we review data obtained from a large sample collection of Middle to Upper Devonian sections in Belgium. Comparison of χ_in trends with paleoenvironmental indicators (facies) and with detrital input proxies (Zr, Th, Ti, Al) allowed to assess the persistence of depositional trends. Furthermore, the χ_in signal was deconvolved into its dominant mineralogical contributions with the help of magnetic-property analysis.The main results are pointing to a magnetic signal dominated by fine-grained magnetite, of which paleomagnetic analysis indicated a formation during Carboniferous remagnetization. This prompts a potentially strong influence of post-depositional processes and it complicates the interpretation of χ_in records in terms of depositional environments. However, in most of the sections, there is a relatively good relationship between χ_in trends and facies evolution and between χ_in and geochemical proxies for detrital inputs. This indicates that the newly formed magnetite grains would at least partly remain where they are formed, and this allows a relative preservation of the original signal, despite the strong influence of diagenesis. Two sections show a stronger impact of diagenesis, where for about half of these sections, the primary, depositional information is lost.The Eifelian-Givetian Monts de Baileux section was selected for time-series analysis of the χ_in series. The Average Spectral Misfit (ASM) method is applied to explicitly evaluate the null hypothesis of no orbital signal and in this section, there is 99.05% chance that the MS signal is reflecting an orbital imprint.