Quantifying paleosecular variation: Insights from numerical dynamo simulations

Thanks to advances in dynamo modelling, it is now possible to produce numerical sequences of magnetic polarity reversals over periods of time equivalent to several tens of millions of years. Using a set of five numerical models integrated over the equivalent of 40–50 Myr, we generated synthetic data analogous to paleomagnetic data derived from volcanic flows. We find that the distributions of directions are remarkably similar to those observed in the paleomagnetic database. Paleosecular variation among the five models is best discriminated by the relative variability in paleointensity ( ε F ) and the relative dispersion of directions or poles as defined by the precision parameter (k). Whether the geodynamo operated in different regimes in its past can be best tested with these parameters in combination. Roughly one million years of time with 200 time-independent samples is required to achieve convergence of ε F and k. The quantities ε F and k correlate well with the average chron duration ( μ chr ), which suggests that excursions and reversals are an integral part of paleosecular variation. If applicable to the geodynamo, the linear dependence of k on μ chr could help to predict μ chr for the Earth during geologic times with no available reversal frequency data; it also predicts much higher average k for directions during superchrons ( k ≈ 2500 for the Cretaceous normal superchron) than during actively reversing times ( k ≈ 35 for the last 80 Myr). As such high k values are not observed, either this family of dynamo models is not applicable to the geodynamo, or the geodynamo regime acting during superchrons is independent of that acting during times with frequent reversals.