Dynamic topography as constraints on stress and viscosity in the mantle and lithosphere

Abstract:

Mantle convection generates stress in the mantle and lithosphere. The lithosphere stress is responsible for localized deformation including seismic deformation at plate boundaries, and localized stress highs in lithosphere are also suggested to cause dynamically self-consistent generation of plate tectonics and continental lithosphere instability, as the stress exceeds a threshold or yield stress. Modeling load-induced deformation at oceanic islands (e.g., Hawaii) constrains lithospheric stress at 100-200 MPa in the plate interiors, leading to a lower limit on lithospheric yield stress (Zhong and Watts, 2013). However, convection-induced lithospheric stress is poorly understood, ranging from 500 MPa to tens of MPa as reported in mantle convection studies. The magnitude and distribution of lithospheric and mantle stress depend critically on buoyancy and viscosity, particularly the latter. Unfortunately, lithospheric and mantle viscosity is also poorly constrained. For example, the inferred lower mantle viscosity from post-glacial rebound and geoid modeling studies ranges from 10²³ Pas to 10²² Pas (e.g., Mitrovica and Forte, 2004; Simons and Hager, 1996; Paulson et al., 2007). In addition to the stress, the lower mantle viscosity may also affect the time evolution of mantle structure including sinking rate of slabs and formation of the degree-2 mantle seismic structure. Therefore, it is important to develop independent constraints on mantle viscosity and convection-induced stress. In this study, I demonstrate that dynamic topography can be used to place first-order constraints on both lithospheric stress and mantle viscosity. For a given superadiabatic temperature difference across the mantle (e.g., 2500 K), a larger mantle viscosity (or a smaller Rayleigh number) leads to a larger lithospheric stress and a larger dynamic topography. To be consistent with the inferred dynamic topography, the lower mantle viscosity is constrained to be significantly smaller than 10²³ Pas, and convection-induced lithospheric stress is unlikely to exceed 200 MPa (i.e., the yield stress for plate interiors). I will discuss the general implications of the results for mantle dynamics including dynamic generation of plate tectonics and the nature of mantle structures (i.e., slabs and plumes) as imaged seismically.