Methods for computing internal flattening, with applications to the Earth's structure and geodynamics

After general comments (Section 1) on using variational procedures to compute the oblateness of internal strata in the Earth and slowly rotating planets, we recall briefly some basic concepts about barotropic equilibrium figures (Section 2), and then proceed to discuss several accurate methods to derive the internal flattening. The algorithms given in Section 3 are based on the internal gravity field theory of Clairaut, Laplace and Lyapunov. They make explicit use of the concept of a level surface. The general formulation given here leads to a number of formulae which are of both theoretical and practical use in studying the Earth's structure, dynamics and rotational evolution. We provide exact solutions for the figure functions of three Earth models, and apply the formalism to yield curves for the internal flattening as a function of the spin frequency. Two more methods, which use the general deformation equations, are discussed in Section 4. The latter do not rely explicitly on the existence of level surfaces. They offer an alternative to the classical first‐order internal field theory, and can actually be used to compute changes of the flattening on short timescales produced by variations in the LOD. For short durations, the Earth behaves elastically rather than hydrostatically. We discuss in some detail static deformations and Longman's static core paradox (Section 5), and demonstrate that in general no static solution exists for a realistic Earth model. In Section 6 we deal briefly with differential rotation occurring in cylindrical shells, and show why differential rotation of the inner core such as has been advocated recently is incompatible with the concept of level surfaces. In Section 7 we discuss first‐order hydrostatic theory in relation to Earth structure, and show how to derive a consistent reference Earth model which is more suitable for geodynamical modelling than are modern Earth models such as 1066‐A, PREM or CORE11. An important result is that a consistent application of hydrostatic theory leads to an inertia factor of about 0.332 instead of the value 0.3308 used until now. This change automatically brings ‘hydrostatic’ values of the flattening, the dynamic shape factor and the precessional constant into much better agreement with their observed counterparts than has been assumed hitherto. Of course, we do not imply that non‐hydrostatic effects are unimportant in modelling geodynamic processes. Finally, we discuss (Sections 7–8) some implications of our way of looking at things for Earth structure and some current problems of geodynamics. We suggest very significant changes for the structure of the core, in particular a strong reduction of the density jump at the inner core boundary. The theoretical value of the free core nutation period, which may be computed by means of our hydrostatic Earth models CGGM or PREMM, is in somewhat better agreement with the observed value than that based on PREM or 1066‐A, although a significant residue remains. We attribute the latter to inadequate modelling of the deformation, and hence of the change in the inertia tensor, because the static deformation equations were used. We argue that non‐hydrostatic effects, though present, cannot explain the large observed discrepancy of about 30 days.