Search

The characterization of individual magnetic components in sediments and sedimentary rocks is difficult, as these natural materials are often a complex mixture of magnetic mineral sources. The analysis of magnetization curves with model functions is the only practicable method of unmixing the magnetic components and characterizing their magnetic properties, if a priori information is not available. Unfortunately, such analyses rely on time consuming measurements and on the choice of appropriate model functions. The simplification of the unmixing problem is closely related to the number of parameters required to fully characterize a magnetic component, and the significance of these parameters in rock magnetic terms. A systematic analysis of synthetic and natural samples shows that a combination of four parameters, here called magnetic fingerprints, is sufficient for this purpose. The fingerprints of individual magnetic components isolated from a wide range of natural sediments and sedimentary rocks form well-defined groups with specific properties. These groups reflect common processes of formation, transport and dissolution of magnetic particles. A clear distinction can be made between two sorts of biogenic magnetite, atmospheric dust, urban pollution and ultrafine magnetite produced in soils and lacustrine/marine sediments. � 2004 Elsevier Ltd. All rights reserved.

The analysis of magnetization curves with model functions has been proposed independently in several works as a method to unmix and characterize magnetic minerals assemblages in sediments and sedimentary rocks. Unfortunately, a successful result of such analysis relies on time-consuming measurements and on the choice of appropriate model functions. However, once the magnetic properties of individual magnetic components have been determined on a small set of selected samples, a simpler and faster analysis of similar samples is possible. The fast analysis of a large number of samples allows investigation of the effect of natural processes on the properties of single magnetic components. The simplification of the unmixing problem proposed in this paper is based on an iterative linearization procedure, which considers the variability of magnetic components. Any simplification of the unmixing problem is limited by a minimum number of parameters, which are required to fully characterize a magnetic component. It has been shown that a combination of four parameters, so-called magnetic fingerprints, is sufficient for a complete characterization of the remanent magnetization of a component. The usefulness of magnetic fingerprints in tracking natural processes is demonstrated exemplarily for lake sediments. The response of Baldeggersee (Switzerland) to environmental changes has been investigated, with special regard to the role of bacterial magnetite in the iron cycle and its possible use as a sensitive paleoredox indicator. The relation between the magnetic properties of lake sediments on one hand, and climatic, tectonic- and human-driven environmental changes on the other, is strongly non-linear. Therefore, a classic correlation between so-called magnetic proxies and environmental signals should be considered with care.