Rates of change of the Earth's magnetic field measured by recent analyses.

Abstract

The availability of accurate representations of the Earth's magnetic field for the past 25 years allows us to study the typical rates of change of the field as a function of the degree of the spherical harmonic. It is shown that there is a general increase in the relative rate of change as the degree of harmonic is increased. The analysis was only carried out to degree eight harmonics, because knowledge of the rate of change of the higher degrees is poor. If the projected rate of change continues in the same manner for higher degrees, then it is predicted that harmonics of, say, degree 14 will change at a rate of about 6% per year. This will mean that it should be easy to determine whether harmonics higher than degree fourteen originate in the core or whether they are caused by crustal magnetization. Although westward drift of the field does not necessarily tell us about motions of the core fluid at the core mantle boundary, it is probably true that these motions depend on the same parameters as those used to determine the westward drift. Since the westward drift depends on the longitudinal gradient of the component of field under consideration, it generally depends more on the higher orders of harmonic within each degree than on the lower orders, all other things being equal. In fact, not all other things are equal, because it is shown that in general the power in the higher orders of harmonic is less than the power in the lower orders. Another important factor in determining longitudinal motions is that it is the field at the core mantle boundary which is important, not the field at the Earth's surface. This results in the higher degrees being relatively more important than at the surface of the Earth. But because of the longitudinal differentiation, it turns out that in order to determine a major portion of the longitudinal drift it is necessary to have information about harmonics up to about degree 20. Therefore, with knowledge only of degrees eight or less, we really know rather little about longitudinal motions. Evidence for secular acceleration was found. Not surprisingly, the average rate of secular acceleration

Self-Affine Time Series: II. Applications and Models

“…This is identical to the spectrum observed in sediment cores and historical data discussed earlier in the chapter. The crossover time scale is the diffusion time across the diameter of the core, estimated to be between 10 3 yr [Harrison and Huang, 1990] and 10 4 yr [McLeod, 1996]. These values are somewhat higher than the time scale of 10 2 yr identified as the crossover in the sediment core and historical data.…”

Mapping the Gauss Coefficients to the Pole and the Models of Paleosecular Variation.

“…1) has a definite structure which cannot be reproduced by a homogeneous model such as that of Constable and Parker. In a study of the recent geomagnetic field, Harrison and Huang (1990) and Harrison (1994) noted that lower order harmonics tend to be larger than the higher order harmonics of the same degree f. This shows that the assumption of Constable and Parker is not consistent with the recent field.…”

“…In order to reduce the parameters to some reasonable ASD for Single Gauss Coefficient number, we may further assume that the variables of the same order f (ge and r7t) have the same variance aq, as was suggested by Constable and Parker (1988), although Harrison and Huang (1990) and Harrison (1994) observed that low-order coefficients tend to have larger magnitudes than high-order coefficients. Then we finally obtain m 2 m 2 (BP) = Eap I r (cos0_dB + 2 2 1 sinePe) + (smB) (14) C I…”

Statistical palaeomagnetic field modelling and dynamo numerical simulation

“…This property can then be used to decide which γ( x m n , x ′ m ′ n ′ , τ) may a priori be set to zero in the matrix, if we happen to be able to estimate the τ m n . Hulot & Le Mouël (1994; see also Harrison & Huang 1990) and Hongre et al (1998) showed that historical and archeomagnetic data could provide enough information for this to be possible. Also, it turns out that the timescales involved are in fact quite similar to those found here for models g and h (see also Christensen & Olson 2003).…”