抄録
There have been many models of the secular variation (SV) of the Earth's magnetic field proposed over the past few decades. Most if not all of them are models of Earth's magnetic field during so-called normal times, in other words, times during which the field is not reversing or undergoing an excursion. However, in studying data from Icelandic lava flows and from other areas, it has become obvious that there is no real distinction between so-called normal SV and the larger deviations of the Earth's field which are classified as excursions and/or reversals, except by the magnitude of the deviation, in which there is a continuous variation. Recently, reversal patterns have been studied in more detail, and it has been found that the exact start or end of a reversal is frequently only fuzzily defined, there being precursors to reversals, or events following reversals, which cannot be described either as belonging to the reversal per se or being a part of the normal SV pattern. In this paper, information from spherical harmonic models of the field during the past three centuries will be used to show that the field today has an abnormally large latitudinal variation of virtual geomagnetic pole (VGP) angular standard deviation, and should not be used to constrain SV models. Data for lava flows and other rock types from a number of different areas will be used to determine the scatter of the field at various observation latitudes, using VGPs from all latitudes, so that eventually a SV model may be produced which does not reject low latitude VGP information. Other factors will be described which any SV model should be able to produce.
