Key points are not available for this paper at this time.
The problem of uniqueness of earth structures obtained by the inversion of geophysical data is still unsolved. Monte Carlo methods offer the advantage of exploring the range of possible solutions and indicate the degree of uniqueness achievable with currently available geophysical data. This procedure was applied to the earth by using the following data: 97 eigenperiods, travel times of compressional and shear waves, and mass and moment of inertia of the earth. Indirect use was made of dt/dΔ data obtained from arrays. Five million models have been examined, and six have passed all tests. Results are as follows: (1) The earth's core is inhomogeneous, consistent with Fe(15–25%)-Si for the outer core and Fe(20–50%)-Ni alloy for the inner core. (2) The radius of the core is increased by 5 to 20 kilometers. (3) Large density and velocity gradients are found in the transition region without prior assumption of an equation of state. The transition region is a compositional boundary as well as a zone in which phase changes occur. The lower mantle shows an increase of the FeO/MgO + FeO ratio by a factor of 2 compared with the upper mantle. This could inhibit mantle-wide convection. (4) Upper mantle models fit better than ‘standard’ models when they are more complex, showing large fluctuations in shear velocity and density. Such complexity might be expected if the mantle is chemically and mineralogically zoned, if there are high thermal gradients, and if partial melting takes place. However, the magnitude of the fluctuations suggests that the zoning is lateral in nature and that the mantle is variable in composition laterally ranging from pyrolite to eclogite.
Frank Press (Thu,) studied this question.