Non‐steady Geomagnetic Dynamo Models

Summary Bullard proved that a disk dynamo performs an oscillation of characteristic nature, but no reversal of magnetic field is achieved by that model. A system of coupled‐disk dynamos (Rikitake) was proved to perform complicated oscillations with occasional reversals. The Rikitake model has been extensively examined by many workers. The model is capable of performing oscillations similar to those as revealed by palaeomagnetic study. As the Rikitake model seems to be highly idealized, time‐dependent behaviour of an Inglis model, which simulates Bullard dynamo action, was studied (Rikitake). The model seems highly stable; a field reversal occurs only when an incredibly large kick of magnetic field is given to its steady state. A small magnetic field given to the zero‐field state of the model grows, eventually reaching a steady state. Analogy between the model and the actual earth suggests that the time constant of reversal amounts to a few thousand years. A non‐steady state Herzenberg dynamo model consisting of two rotating spheres embedded in a conductive medium has been studied by solving simultaneous non‐linear integral equations. No great success has been achieved because of computational difficulties. Similar studies have been extended to other models though it is no easy matter to take the equation of motion into account properly. On the basis of the time‐dependent behaviour of these models, the mechanism of field reversal is discussed. A few mechanisms based on statistical fluctuations of fluid motion (e.g. Nagata) do not seem workable. It is concluded that geomagnetic polarity reversals take place as a result of complicated exchange of energy between units consisting of dynamo models.