Rheology of the deep upper mantle and its implications for the preservation of the continental roots: A review

The longevity of deep continental roots depends critically on the rheological properties of upper mantle minerals under deep upper mantle conditions. Geodynamic studies suggest that the rheological contrast between the deep continental and oceanic upper mantle is a key factor that controls the longevity of the continental roots. Current understanding of rheological properties of deep upper mantle is reviewed to examine how a large enough rheological contrast between the continental and oceanic upper mantle develops that leads to the longevity of the deep continental roots. Based on the microstructures of naturally deformed deep continental rocks as well as on the observations of seismic anisotropy, it is concluded that power-law dislocation creep dominates in most of the deep upper mantle. Deformation by power-law creep is sensitive to water content and therefore the removal of water by partial melting to form depleted continental roots is a likely mechanism to establish a large rheological contrast. The results of experimental studies on the influence of temperature, pressure and water content on plastic flow by power-law dislocation creep are reviewed. The degree of rheological contrast depends critically on the dependence of effective viscosity on water content under “wet” (water-rich) conditions but it is also sensitive to the effective viscosity under “dry” (water-free) conditions that depends critically on the influence of pressure on deformation. Based on the analysis of thermodynamics of defects and high-temperature creep, it is shown that a robust estimate of the influence of water and pressure can be made only by the combination of low-pressure (< 0.5 GPa) and high-pressure (> 5 GPa) studies. A wide range of flow laws has been reported, leading to nearly 10 orders of magnitude differences in estimated viscosities under the deep upper mantle conditions. However, based on the examination of several criteria, it is concluded that relatively robust experimental data are now available for power-law dislocation creep in olivine both under “dry” (water-free) and “wet” (water-saturated) conditions. These data show that the influence of water is large (a change in viscosity up to ∼ 4 orders of magnitude for a constant stress) at the depth of ∼ 200–400 km. I conclude that the conditions for survival of a deep root for a few billions of years can be satisfied when “dry” olivine rheology with a relatively large activation volume (> (10–15) × 10 − 6 m 3/mol)