IS THE TRANSITION ZONE AN EMPTY WATER RESERVOIR? INFERENCES FROM NUMERICAL MODEL OF MANTLE DYNAMICS

Abstract

Water is probably present everywhere in the Earth’s mantle today, with abundances ranging between scales of percent (%) to the parts per million (ppm). Mantle total water content is estimated to be between 10% to several times that of the present-day hydrosphere. Numerous studies have been devoted to the determination of water solubility in mantle material [D.R. Bell, G.R. Rossmann, Science 255 (1992) 1391–1397; J. Ingrin, H. Skogby, Eur. J. Mineral. 12 (2000) 543–570]. They all show strong solubility variations from one mineral phase to another. Principally, water partitioning has made the transition zone a probable trap for water from the Earth’s mantle [N. Bolfan-Casanova et al., Earth Planet. Sci. Lett. 182 (2000) 209–221; D.L. Kohlstedt et al., Contrib. Mineral. Petrol. 123 (1996) 345–357]. Nevertheless, water distribution within the mantle is still debated. We have studied the role of mantle dynamics in water distribution by modeling water transport and mantle convection in a two-dimensional (2-D) cartesian geometry. The model takes into account water partitioning between the mantle’s transition zone and the upper mantle of 10:1 and between the lower mantle and the transition zone of 1:100 (i.e. respectively between olivine and spinel and spinel and post-spinel). We have modeled the mantle temperature field using depth-dependent viscosity and plate-like surface conditions. Water injection at the trench has also been simulated. Our numerical experiments suggest that diffusivity of water has to be very high, at least two orders of magnitude higher than the one experimentally determined [D.R. Bell, G.R. Rossmann, Science 255 (1992) 1391–1397; J. Ingrin, H. Skogby, Eur. J. Mineral. 12 (2000) 543–570] to significantly influence water distribution in Earth’s mantle. In fact, the diffusion process is not efficient enough to balance the mixing due to mantle dynamics and to force water into the transition zone. We show that the distribution of water should be quite homogeneous throughout the mantle if advection and diffusion are the only processes involved in water transport in the mantle. This homogeneity implies that water below the transition zone could be in excess according to the lower mantle rocks solubility. This addresses the question of stability of free water in the lower mantle and its mobility by percolation process, which could be a very efficient transport process, previously unconsidered in this field of research.