Page 89 - Global Tectonics

76 CHAPTER 4

Figure 4.3 Interpretation of a magnetic anomaly profile across the Juan de Fuca ridge, northeastern Pacific Ocean,
in terms of normal and reversed magnetizations of two-dimensional rectangular blocks of oceanic layer 2. The arrow
marks the ridge crest (redrawn from Bott, 1967, with permission Blackwell Publishing).

supercomputers in the 1990s. The first results of numer- in either of two states: one which generates a fi eld of
ical integrations of full three-dimensional, nonlinear, constant polarity for tens of millions of years, and one
geodynamo models were published in 1995 (e.g. Glatz- during which the fi eld reverses in polarity at least once
maier & Roberts, 1995). These, and other comparable every million years. This is surprising in that convective
simulations through to 2000, were reviewed by Kono & overturn in the core is thought to be on a timescale of

Roberts (2002). The models simulate many of the fea- hundreds of years. It is difficult to imagine processes or

tures of the Earth’s field, such as secular variation and conditions in the core that could account for two
a dominant axial dipole component, and in some cases different states, which, once attained, persist for tens of
magnetic reversals. Some of the latter are very similar millions of years. This timescale is characteristic
in duration and characteristics to those deduced from of convection in the mantle. Changes in the pattern of
paleomagnetic studies (Coe et al., 2000). convection in the mantle could produce changes in the
The rates at which geomagnetic reversals have physical conditions at the core–mantle boundary on the
occurred in the geologic past is highly variable (see Figs appropriate timescale. Small changes in seismic veloci-
4.4, 4.13). There has been a gradual increase in the rate ties in the mantle, revealed by seismic tomography, are
of reversals during the Cenozoic, following a period interpreted in terms of temperature variations associ-

during the Cretaceous when the field was of constant ated with convection, although they could in part be due
normal polarity for 35 Ma. Paleomagnetic studies reveal to chemical inhomogeneity (Section 12.8.2). This raises

a similar prolonged period of reverse polarity in the Late the possibility that the heat flux at the core–mantle
Carboniferous and Permian (McElhinny & McFadden, boundary is nonuniform, and changes signifi cantly over
2000). This seems to imply that the geodynamo can exist periods of 10–100 Ma. The low viscosity and relatively