Geomagnetic field intensity and reversal asymmetry in late Precambrian Keweenawan rocks

Abstract

Summary. Fifty-four palaeointensity values, measured by the Thellier technique, are used to study the cause of reversal asymmetry in late Precambrian Keweenawan rocks. These data, obtained mainly from diabase dykes carrying primary TRM and their baked host rocks, show that ‘reversed’ palaeointensities are on average 40 per cent higher than their ‘normal’ counterparts. This result cannot be explained by any overall difference in remanence characteristics, grain size or cooling rate (dyke width) between the rocks comprising the two polarity groups. A significant difference in the local intensity of the geomagnetic field between times of normal and reversed polarity is thus indicated. This conclusion is reinforced by positive baked contact tests for both normal and reversed Thellier palaeointensities. Secondary component and single dipole offset models as explanations for Keweenawan reversal asymmetry are in conflict with the palaeointensity results. The difference in palaeointensity between normal and reversed rocks disappears when the data are reduced to the palaeoequator in accordance with a geocentric axial dipole field. The palaeointensity results are therefore consistent with the idea that reversal asymmetry is caused by the North American plate occupying a lower palaeolatitude when the field had normal polarity. Acceptance of a geocentric axial dipole would imply that the Keweenawan geomagnetic field intensity was about 45 per cent higher than at present. However, evidence at one locality of possibly three asymmetric reversals poses problems for a plate movement interpretation, and forces consideration of a model whereby the asymmetry is caused by persistent non-dipole field components. Rock magnetic studies show that reliable palaeointensity estimates can be obtained for rocks which have stable NRM with single domain or pseudo-single domain grain sizes and with a high Koenigsberger ratio. Baked contact rocks are the most reliable whereas hydrothermally altered lavas are unsuitable for Thellier intensity determinations. It is suggested that af demagnetization studies on companion specimens should be performed before Thellier experiments in order to select specimens which have high values of median destructive field and which lack viscous or other overprints.