Theory of viscous remanent magnetization acquisition and its application to flood dating

• Main Author: Berndt, Thomas
• Other Authors: Muxworthy, Adrian R., Imperial College London; University of Minnesota; British Council; International Union of Geodesy and Geophysics (IUGG)
• Format: Doctoral or Postdoctoral Thesis
• Language: unknown
• Published: Imperial College London 2017
• Subjects: Sólheimajökull; Iceland
• Online Access: http://hdl.handle.net/10044/1/51502; https://doi.org/10.25560/51502

Most rocks contain magnetic minerals that acquire a natural remanent magnetization (NRM) in the direction of the geomagnetic field during their formation. Floods may quarry, transport and rotate boulders such that their NRM will subsequently be randomly aligned. The magnetic minerals then gradually acquire a viscous remanent magnetization (VRM) at ambient temperatures in the direction of the ambient field that partially overprints the original NRM. According to Néel's (1949) theory of fine magnetic particles, the demagnetization temperature of a VRM is a function of acquisition time and can be used to date the movement of the rock and, hence, the flood. No generally accepted standard method of VRM dating exists yet. In this thesis, a theory of VRM acquisition is rigorously derived and experimentally tested, including heating/cooling rate effects, temperaturedependence of the attempt time, and the effect of mixtures of magnetic minerals. A protocol for VRM flood dating based on continuous thermal demagnetization (CTD) of flood boulders is proposed and evaluated on five cases of known ages: three jökulhlaups in Iceland (Sólheimajökull, Kotarjökull and Markarfljot), a storm flood deposit in Scotland, and a tsunami in Cape Verde. The protocol encompasses sampling procedure, CTD of the NRM, magnetic viscosity experiments, criteria for sample selection and a full statistical treatment of the obtained age estimates and uncertainties. Two Icelandic dating attempts were successful, while the remaining three attempts were unsuccessful, but were clearly identifed as such. In the two successful cases gave a good agreement with known ages of 72 years (error limits 11-360 years) versus a known age of 155 years for the Sólheimajökull flood and 290 years (error limits 80-2300 years) versus a known age of 288 years for the Kotarjökull flood. This suggests that CTD of VRMs can be a useful order-of-magnitude tool for flood dating. Open Access