Differences in paleomagnetic interpretations due to the choice of statistical, demagnetization and correction techniques: Kapuskasing Structural Zone, northern Ontario, Canada

The Kapuskasing Structural Zone (KSZ) reveals a section through the Archean lower crustal granoblastic gneisses. Our new paleomagnetic data largely agree with previous work but we show that interpretations vary according to the choices of statistical, demagnetization and field-correction techniques. First, where the orientation distribution of characteristic remanence directions on the sphere is not symmetrically circular, the commonly used statistical model is invalid Fisher, R.A., Proc. R. Soc. A217 (1953) 295]. Any tendency to form an elliptical distribution indicates that the sample is drawn from a Bingham-type population Bingham, C., 1964. Distributions on the sphere and on the projective plane. PhD thesis, Yale University]. Fisher and Bingham statistics produce different confidence estimates from the same data and the traditionally defined mean vector may differ from the maximum eigenvector of an orthorhombic Bingham distribution. It seems prudent to apply both models wherever a non-Fisher population is suspected and that may be appropriate in any tectonized rocks. Non-Fisher populations require larger sample sizes so that focussing on individual sites may not be the most effective policy in tectonized rocks. More dispersed sampling across tectonic structures may be more productive. Second, from the same specimens, mean vectors isolated by thermal and alternating field (AF) demagnetization differ. Which treatment gives more meaningful results is difficult to decipher, especially in metamorphic rocks where the history of the magnetic minerals is not easily related to the ages of tectonic and petrological events. In this study, thermal demagnetization gave lower inclinations for paleomagnetic vectors and thus more distant paleopoles. Third, of more parochial significance, tilt corrections may be unnecessary in the KSZ because magnetic fabrics and thrust ramp are constant in orientation to the depth at which they level off, at approximately 15-km depth. With Archean geothermal gradients, primary remanences were blocked after the foliation was tilted to rise on the thrust ramp. Therefore, the rocks were probably magnetized in their present orientation; tilting largely or entirely predates magnetization.