マントル対流論の現状

Abstract

I review the recent topics of the mantle convection, especially with regard to the scales of mantle flows. The studies of 3-D velocity structure of the earth's mantle, geoid, gravity and isotope geochemistry provide a new perspective of the flow in the mantle. Several workers show a clear correlation between the distribution of hotspots and the geoid anomalies obtained by subtracting the contribution of the subducted slabs. Assuming a spherically symmetric viscous earth, several researchers could constrain the form of mantle flow and viscosity structure in the earth by correlating the calculated geoid with the observed one. Seismological studies suggest the existence of aseismic slab beneath 650km discontinuity. If this is a case, the aseismic slab may be explained by several ways on the basis of convection theories, that is, whole mantle convection, layered convection with thermal coupling and the penetrative convection. 3-D velocity structure obtained by several workers show the correlation between the geoid and lower mantle heterogeneity at low degrees (l=2-3) and this correlation was explained by the flow models. Possible seismological indicators of mantle layering are proposed. Seismic anisotropy found by the studies of the long period surface wave correlates with some mantle flow model and may suggest the flow direction or stress field in the real earth. The topography of core-mantle boundary revealed by recent seismological and geodetic studies may constrain the viscosity and thermal properties of mantle near CMB. The heterogeneity in the CMB may be closely related to the dynamics near CMB, for example, entrainment of dense material by the convection and the double diffusive convection. Isotope geochemistry shows that there are many geochemical source regions which have various sizes and age distrbution. Convective mixing was investigated and two opposite views have been presented. In one view, the convection is strong enough to homogenize the geochemical heterogeneities within a geologically short time; in another, both small and large scale heterogeneities persist for 1-2 b. y. The studies mentioned above are combined into the flow models by several workers.