抄録
The Earth possesses its intrinsic magnetic field generated by fluid motion in the electrically conducting core, known as the geodynamo. Provided the spatial distributions of the Earth's magnetic field and its temporal variations are known at the core surface, it seems to be possible to estimate the fluid motion there by solving an inverse problem. The magnetic diffusion term is much smaller than the advection term in the magnetic induction equation, and therefore it can be neglected for the time scale much shorter than the magnetic diffusion time. This is called as the frozen-flux approximation, and magnetic lines of force behave as if they are frozen-in fluid elements. Because of its fundamental non-uniqueness, additional constraints, for example, toroidal flow, steady flow, geostrophic flow, or a combination there of, are imposed in estimating the flow at the core surface. Estimated fluid motions have common features : large vortices at midlatitudes and westward flow near the equatorial region. It is now possible to carry out numerical simulations of three-dimensional magnetohydrodynamic dynamos in rotating spherical shells. Inversion methods are then tested for synthetic data of numerical models. It is concluded that some meaningful information on the core flow can be recovered, but that the resolution of the magnetic field at the core surface has a crucial effect on the flow structure. Hence, to construct a realistic geodynamo model, it is important to determine the magnetic field precisely with smaller length scales at the core surface. It is required to monitor the geomagnetic field by launching satellites for magnetic field measurements periodically, for example every five years.
