An experimental and analytical assessment of geomagnetic intensity variation since the Devonian : links with global geological processes
This study was broadly concerned with acquiring geomagnetic palaeointensity estimates and interpreting existing data in the wider context of the global geodynamic system. Two-hundred and ninety-eight samples from nine suites of Mesozoic and Permian intrusions (comprising thirty-one individual rock units) from eastern Australia were subjected to rock magnetic and palaeodirectional analyses. The grain size distributions of these were generally observed to contain a significant fraction of large pseudo-single domain (PSD) and multi-domain (MD) grains. This frequently allowed a substantial part of their blocking temperature spectrum (- 350°C) to be overprinted by a thermal event affecting this region during the mid-Cretaceous. The modified Thellier palaeointensity method was employed on samples from seven of the intrusion suites producing virtual dipole moment (VDM) estimates generally lower than the present value (8 x 1022 Am2) and variable in quality. Mean values of VDM were obtained for the following periods: 90 Ma « 5.9 x 1022 Am2); 127 Ma (7.9 x 1022 Am2); 172 Ma (1.2 x 1022 Arn); 200 - 178 Ma (6.3 x 1022 Am2); 255 Ma (7.5 x 1022 Arrr'). Despite being in good agreement with previously acquired data, few of these estimates satisfied all conventional acceptance criteria and consequently a reliability (R) factor was invoked to allow qualitative comparisons of estima~es tobe made. A simulated Thellier experiment was also performed on pre-treated rock samples containing magnetic grains dominantly within the single domain (SD) range. This produced a surprising amount of variable non-ideal behaviour. In particular, a number of samples significantly overestimated the 'palaeointensity' when only a low temperature, seemingly ideal, portion of their blocking temperature spectra was used. Five proposed mechanisms of non-ideal behaviour succeeded in explaining most of this non-ideal behaviour and correcting the results of 50% of the samples. One of these, 'demagnetisation bias,' could be developed into an extremely useful tool for correcting convex-down NRM-TRM plots produced by assemblages of Ml) grains and allow this widespread problem to be overcome in future palaeointensity studies. The global dipole moment record for the period 400 - 10 Ma was subject to a detailed statistical analysis allowing it to be segmented into periods defined by the distribution of the data themselves. Additionally, a system of grouping estimates into 'rock suites' was developed to avoid over-representation of secular variation (SV) in the record. Variation of poloidal field strength (PFS; recorded in VDM estimates) since the Devonian was shown to be largely decoupled from geomagnetic reversal frequency except during superchrons when its lower limit may have been raised. It was discovered that periods of intense true polar wander (TPW) and large igneous province (LIP) emplacement, which have previously been associated with changes in reversal, frequency, may perturb the geodynamo's capacity to generate poloidal field but certainly do not control it An excellent time correlation between PFS and the supercontinent amalgamation-dispersion cycle since the Devonian was observed. The following model, comprising four' phases, was developed to explain this. (l) Prior to the formation of Pangaea (> 350 Ma) the 'upper and lower mantle convected separately, the latter was hot allowing only a small heat flux across the CMB, and consequently PFS was kept low. (2) During the final assembly of Pangaea (350 - 325 Ma), there was a catastrophic avalanche of cold material through the 660 km transition into the lower mantle which increased CMB heat flux and PFS dramatically. (3) The maintenance of Pangaea through the period 325 - 180 Ma allowed both the upper and lower mantle to wann gradually, causing PFS to fall steadily. (4) In the mid-Jurassic, the continents began to disperse and the high temperature contrast between lithosphere and mantle provided the subducting slabs with sufficient momentum to penetrate the 660 km transition and gradually cool the lower mantle, PFS has risen (steadily or in small jumps) since this time as a consequence. This model is entirely consistent with long-term trends in PFS since the Devonian, and provides a benchmark to be tested by the addition of more palaeointensity data and quantitative mantle modelling.