Our understanding of the dynamic evolution of the Earth hinges critically on our knowledge of the mantle's viscosity structure. The post-glacial rebound process provides the most important and direct constraints on that structure. In the last three decades, a large number of studies have been devoted to constraining mantle viscosity by comparing model predictions and post-glacial rebound data. Almost all these studies have assumed that the Earth's viscoelastic structure is one-dimensional with dependence only on radius -- a condition that greatly simplifies the modeling process. However, important laterally varying viscoelastic structures exist in the mantle and lithosphere, including lithospheric thickness variations, faulted plate boundaries, continental keels, and other structures implied by a variety of seismic tomography models. The investigators propose to develop spherical finite element models to study the post-glacial rebound process for a mantle with realistic 3-D viscoelastic structures that include continental keels, and other structures implied by a variety of seismic tomography models. They will study the coupling between spherical harmonics with different degrees and orders induced by the laterally varying structure, and will examine the possible effects of that coupling on post-glacial rebound observations. For the proposed research, theywill use a finite element software CitcomS that has been developed to model thermal convection with variable viscosity in a spherical shell. They will also extend CitcomS to incorporate a viscoelastic (e.g., Maxwell) rheology and a deformable grid.