Prediction of anisotropy from flow models: A comparison of three methods

Abstract

Observations of anisotropy in Earth are used regularly as constraints for models of deformation, using various assumptions about the relationship between deformation and the resulting anisotropic fabric. We compare three methods for calculating fabric from velocity fields: tracking of finite strain ellipses, a kinematic crystallographic code, and the evolution of directors. We find that the use of the finite strain ellipse provides only limited prediction capabilities, as it cannot reproduce experimental observations that involve recrystallization. The crystallographic code we tested (a variant of the popular code D-Rex) provides a more complete fabric prediction, but at a much higher computational cost. The directors method provides an intermediate solution: while it does not include some of the more complex crystallographic processes that D-Rex does, the results of this method closely resemble those of D-Rex, at a lower computational cost. The directors are also capable of tracking anisotropy at much larger strains than DRex. We conclude that when computation speed is important, for example, in self-consistent geodynamic models with anisotropic rheology, the directors method provides an appropriate approximation. Copyright 2008 by the American Geophysical Union.