The physics of creep and attenuation in the mantle

Dislocations contribute to both seismic wave attenuation and steady‐state creep in the mantle. The two phenomena involve quite different strains and characteristic times but they can both be understood with simple dislocation models. The most satisfactory model for creep involves the glide of dislocations across subgrains and rate limited by the climb of jogged dislocations in the walls of a polygonized network. The jog formation energy contributes to the apparent activation energy for creep, making it substantially larger than the activation energy for self–diffusion. The theory leads to either a σ^2 or σ^3 law for creep rate, depending on the length of the dislocations relative to the equilibrium spacing of thermal jogs. Attenuation in the mantle at seismic frequencies is probably caused by the glide of dislocations in the subgrains. Kink and impurity drag can both contribute to the glide time constant. The kink‐formation, or Peierls barrier, model for dislocation glide appears to be a low‐temperature, high‐frequency mechanism most appropriate for pure systems. A small amount of impurity drag brings the dislocation glide characteristic time into the seismic band at upper‐mantle temperatures. The attenuation and creep behavior of the mantle are related through the dislocation structure. Discussion of the various possible mechanisms is facilitated by casting them and the geophysical data in terms of a pre‐exponential characteristic time and an activation energy. The relaxation strength is an additional parameter that can be used to identify the attenuation mechanism. Mobile dislocations in subgrains, rather than cell walls, have the appropriate characteristics to explain the damping of seismic waves in the upper mantle. The grain boundary peak may be responsible for attenuation in the lithosphere