Palaeointensity and Brunhes palaeomagnetic field models

Abstract

Statistical models of geomagnetic secular variations in terms of a Giant Gaussian Process (GGP) are compared with palaeontensity databases (Brunhes epoch, effusive rocks). We tested two GGP models TK and QC (suggested by Tauxe-Kent and Quidelleur-Courtillot) against 392 volcanic palaeointensity data, and found both models to be incompatible with the data. Even though at present there is no accepted GGP model of the Brunhes chron, we found interesting to clarify the details of this negative test. While such an incompatibility is clear enough for the TK model, because of its too small dipole coefficient, the incompatibility of the QC model was due the 12 per cent of the data with very small intensities; remaining 88 per cent of data is fully compatible with QC model QC. The troublesome, low-palaeointensity records belong to different sites and ages so their excess presence is not an effect of particular location or time. We found no evidence that the marked discrepancy between the empirical data and theoretical calculations on the QC model can be explained by the possibility that the lowintensity values belong to known excursions of the geomagnetic field. As an alternative, we conjecture that selected low-intensity values can possibly stem from acquisition of chemical remanent magnetization instead of the thermoremanence.