The pattern of flow in the mantle cannot be observed directly but is informed by implementing models of known physical behavior of Earth materials in computer simulations. There are uncertainties associated with this approach and this project targets one of them- the imperfect knowledge about how mantle minerals deform. Laboratory studies provide constraints but both the direct measurements and the inferred material flow laws have uncertainty. A quantitative description of this uncertainty in microscopic behavior will, for the first time, be included directly in the macro-scale mantle flow simulations. This will result in a probabilistic description of the model spaces relevant for the Earth for a range of likely constituent architectures. Local mineral deformation mechanism, grain size, potential temperature, pressure, water, and melt are all known to affect stress-strain relations in mantle rock. A previously developed Bayesian method will be modified to allow various architectures to be included. A graduate student will receive training at a forefront of numerical modeling in geophysics and will gain insight to laboratory results- a cross disciplinary background that is not typical in most departments. A website will be developed for researchers to query for self-consistent descriptions of mantle flow law specification.
