Abstract
I show that for varying dipole sources in the core, the mean-square value over the surface of the Earth of the main field falls off more rapidly with the spherical harmonic degree, n, than does the corresponding factor for the secular variation field. The same general conclusion was reached in a 1974 study by F. J. Lowes from a consideration of the observed westward drift in the field and from consideration of known spherical harmonic coefficients. This means that a higher degree secular variation model is required than for a main field model for equally good representation of both fields. If the only use of the secular variation model is for updating a main field model in predictive modeling, then the secular variation model should be carried to the same degree as the corresponding main field model if possible. This has not usually been done because of the difficulty in obtaining adequate secular variation data. Satellite data may help, but variable fields from external sources, which are not considered as part of the secular variation field, limit their use in secular variation models.
The use of a secular variation model in studying mantle conductivity or dynamo theory may require a higher degree model than for the corresponding main field model in order to obtain details of the secular variation that may be critical to relevant theories.
