Modelling the magnetic properties of natural and environmental materials

Abstract

Magnetic properties have been used to characterize natural and environmental materials. An evaluation of magnetic properties, for the modelling of sources of materials and minerals, has been completed. A methodological framework has been developed for the application of magnetic techniques to studies involving the quantification of sources of materials and minerals in any environment. The framework includes the identification of sources using magnetic reconnaissance and multivariate statistical classification techniques. Magnetic measurements used are susceptibility (both field and laboratory), remanence and magnetization measurements. The linear additivity of magnetic measurements, and classification and linear modelling techniques, have been tested using data for artificial laboratory mixtures and hypothetical mixing experiments. The limitations of using magnetic properties with these statistical and mathematical techniques are defined. The framework allows for the testing of suitability of magnetic modelling techniques in any sourcing study.


To test the methodological framework three environmental applications have been completed. The first, a coastal study, has allowed assessment of the methodology in sourcing beach sediments in a dynamic system. In the second application, a catchment study, evaluation of the spatial variability of sources is undertaken; including statistical definition of the source properties for modelling techniques. The final application covers modelling of magnetic minerals in several contrasting soil profiles. Hysteresis loop data are used for the purpose of modelling diamagnetic, paramagnetic, canted-antiferromagnetic and ferrimagnetic minerals in the soils.

The methodological framework is re-assessed according to the needs and results of each individual application. It has largely been the case that there are great limitations in the use of magnetism for sourcing environmental materials, and minerals quantitatively. At best only four potential sources of material in a 'mixture' can be successfully modelled. These are mainly defined by the four mineral magnetic properties listed above. Limitations and errors involved in each step of the methodology are tested, and quantified, and a final definition of the positive application of quantitative source modelling using magnetism is made.

From the experimental work it has been found that _Xlf is the most linearly additive measurement and field kappa can be widely used for mapping magnetic properties indicating soil, land-use, geology and topographical/drainage features. _Xlf versus HIRM+100mT bi-variate plots give vital preliminary information on the modelling suitability of the potential sources of mixtures. From the case study applications it has been found that environmental variability of sources can be incorporated into linear modelling routines. Often this gives source contributions which have a wide, often overlapping, range of values. In the coastal and mineral modelling work dominant ferrimagnetic minerals have been found to mask weaker magnetic signals which makes unmixing of the mineral contributions very difficult. However, also in the coastal study field tracing, using kappa has been utilised and gives information on short term beach and wave processes. In the catchment study sources of finer sediments have been quantified successfully and the reliability of answers tested. Ranges of source contributions overlap when the natural variability of sources is taken into account. In the future, wider incorporation of hysteresis parameters may improve characterization of minerals other than those which are ferrimagnetic.

Date of Award	1994
Original language	English
Awarding Institution	Coventry University
Sponsors	Coventry University Research Initiative Grant
Supervisor	John Dearing (Supervisor) & Peter Lockett (Supervisor)