Geodynamic inversion to quantify the rheological parameters of the lithosphere

Tobias Baumann, Johannes Gutenberg University of Mainz, Institute for Geosciences, Mainz, Germany and Boris J P Kaus, Johannes Gutenberg University Mainz, Institute of Geosciences, Mainz, Germany
Abstract:
The dynamics of the crust and lithosphere is to a large extent controlled by its effective viscosity. Typically, rheological parameters are estimated from laboratory experiments and extrapolated over ten orders of magnitude to geological conditions, which has significant uncertainties. In this study, we propose a method to constrain these parameters directly from geophysical observables. The method links lithosphere-dynamics models with geophysical observations (e.g. horizontal and vertical surface (GPS-) velocities, topography and gravity data). A Bayesian inversion strategy is employed to estimate probabilities of model parameters that affect the rheology of the lithosphere. As this requires many forward models (104 - 106), an efficient and parallel implementation is necessary, which we recently showed for 3D models with linear viscous rheologies (Baumann et al. 2014).

Here, we examine how well this method is capable of resolving complex rheologies and address a few technical issues such as defining appropriate stopping criteria for models with a free surface. We use synthetic models of intra-oceanic subduction to demonstrate that the rheological parameters of the lithosphere can be recovered successfully even if rheologies are power-law viscous or plastic, provided that the temperature structure is known. A piece-wise linear lithospheric temperature parameterization as part of the inverse approach is shown to give reasonable results as well.

Furthermore, we apply the method to cross-sections of the Himalaya where we consider different geological interpretations as end-member cases. For each end-member, we estimate the probabilities of their rheological structure and plasticity parameterisation and also obtain suitable temperature distributions. Our results indicate that the Indian mantle lithosphere has large effective viscosities whereas the Tibetan lower crust has small effective viscosities. This is consistent with many published modelling studies of the region.

Acknowledgements
Funding was provided by ERC Grant agreement #258830

References
Baumann T.S., Kaus B.J.P., Popov A.A., 2014. Constraining effective rheology through parallel joint geodynamic inversion. Tectonophysics.

Back to: Linking Plate Tectonics and Mantle Convection to Wilson Cycles: Constraints from the Geological Record and Surface Processes III