Reconstructing Mantle Circulation and Surface Tectonics Using Ensemble Data Assimilation

Abstract:

Plate tectonic reconstructions are estimates of the surface history of mantle circulation. Combining this information with a physical model of mantle convection allows to infer the history of the whole mantle circulation for the last hundreds of million of years. So far, the reconstructions of mantle circulation have been done by imposing plate reconstructions as surface conditions on numerical mantle convection models. This technique does not allow any dynamical feedback of mantle dynamics on surface tectonics. Mantle convection models are now able to generate plate-like tectonics self-consistently: they describe both the effect of plates on mantle convection, and the effect of mantle convection on the surface tectonics. We propose to use these convection models in combination with plate reconstructions to produce a new generation of mantle circulation models. For this purpose, we use the ensemble Kalman filter. This method has been developped and successfully applied to meteorology, oceanography and even more recently outer core dynamics. It consists in integrating sequentially a time series of data into a numerical model, starting from an ensemble of possible initial states. The initial ensemble of states is designed to represent an approximation of the probability density function (pdf) of the a priori state of the system. Whenever new observations are available, each member of the ensemble states is corrected considering both the aproximated pdf of the state, and the pdf of the new data. Between two observation times, each ensemble member evolution is computed independently, using the convection model. This technique provides at each time an approximation of the pdf of the state of the system, in the form of a finite ensemble of states. We perform synthetic experiments to assess the efficiency of this method for the reconstruction of mantle circulation.