Nature of the 400-km Discontinuity in the Earth's Mantle

Abstract

In a recent paper Liebermann (1975) determined the elasticity data of olivine (a), beta ( B ) and spinel (y) polymorphs of germanates and silicates and made a comparison of these new elasticity data with recent earth models in the vicinity of the 400-km discontinuity. From this comparison he concluded that only model 1066 B of Gilbert & Dziewonski (1975) is compatible with the experimental data for the a-y and a-B transformations. This conclusion does not seem to be justified because his experimental results for phase transformations show on the average the velocity jumps of about 15-20 per cent whereas model 1066 B shows only 5-7 per cent velocity increase. In fact, not only model 1066 B, which is chosen by Liebermann as the most viable model, several other earth models listed in Table 1 also show such smaller velocity jumps (7-10 per cent) which are only about half in percentage magnitude as due to a-y or aj l phase transformations determined experimentally. This is obviously a very significant difference between the two which needs a satisfactory explanation. This discrepancy between the earth models and the experimental results has also been pointed out by Liebermann (1973) while examining the UTD (Dziewonski & Gilbert 1972) and Bl (Jordan 1972; Jordan & Anderson 1974) earth models. Liebermann (1973, 1975) also concluded that the experimental data should be used to provide constraints in future inversion attempts to derive geophysical models for the Earth’s interior. They are: at the 4Wkm discontinuity (1) the velocity jumps for V,, Vs and V+ should be of comparable percentage magnitude and approximately twice the percentage jump in density and (2) the Vp/Vs ratio should be constant or increase across the discontinuity according to Liebermann. We feel that before thinking of applying such constraints it may be worthwhile to investigate the real cause for the discrepancies observed between the earth models and the experimental data for phase transformations in the olivine. Matching the laboratory elasticity data with the seismic data to draw inferences about phase or compositional change in the mantle, it will be more realistic if the laboratory data are obtained at temperature and pressure conditions prevailing in the upper mantle. Even then, we cannot expect that the percentage velocity jumps determined in the laboratory due to olivine-spinel or olivine-8 phase transformation alone would be really of equivalent magnitude to those determined by seismological studies. In our opinion, considering the 400-km discontinuity as merely a phase boundary may be an over simplification and perhaps there is also some chemical change taking place simultaneously along with phase change at the discontinuity so that the actual velocity jumps are much smaller. Probably an increase of Fayalite content in the olivine might