Petrological geodynamic modeling of mid-ocean ridges

Abstract

Mid-ocean ridges are the primary location where the Earth’s oceanic crust is formed. Beneath spreading ridges several processes such as dynamic melting and partial crystallization modify the petrology of the upper mantle and affect the Earth’s global geochemical evolution. A unified picture of the temporal and spatial evolution of melt and residual mantle, as well as crustal production and melt dynamics requires a comprehensive model that takes into account simultaneously the complexity of the physical processes involved and the petrological variations of the ridge system. Here we present the first results of a 2-D numerical approach applied to a spreading ridge that fully couples a two-phase flow model for melt and solid mantle and a chemical thermodynamic model which provides a spatial and temporal description of the minerals and melt abundance and composition. The most significant features found by this study are the following. (1) Accumulation of melt is observed at the base of the lithosphere in the off-axis region (<∼50 km from the ridge axis). (2) Crustal production (thickness) shows temporal variations which are mainly induced by periodic discharge of the melt accumulated underplate. (3) Magma waves develop between 10 and 30 km depth in proximity of the ridge axis. However to accurately resolve melt fluctuations, the grid size must be smaller than the compaction length for porous flow. Since in this study the compaction length decreases with depth, we have used a simplified 1-D melt model incorporating the two-phase flow dynamics and the thermodynamic formulation to show that the depth at which magma waves start to form increases by increasing the numerical resolution. Despite the limitation of the numerical grid resolution, we have observed that variations of the melt content do not appear to have significant influence on major elements composition of the residual solid and melt. (4) In the initial stage of the ridge evolution, a melting area detaches from the main melting region around the ridge axis. It is possible that this type of development may repeat over time beyond the duration of the simulation model of this study (∼15 Ma). Sluggish coupling between the dynamics of the lithosphere and the asthenospheric mantle flow suggests that accretion of the lithosphere by conductive cooling away from the ridge center involves portions of the upper mantle that not necessarily passed through the spreading ridge. (5) During the development of the spreading ridge, the asthenosphere affected by the melting process deflects downwards, creating in this way a chemical heterogeneity in the large mantle circulation. (6) Composition of major elements in the residual solid after partial melting is in agreement with the chemical pattern observed in abyssal peridotites. However, in order to explain the large variation of major elements content found in abyssal peridotite, a consistent petrological and geodynamic model of the evolution of the mid-ocean ridge, requires that partial crystallization of small amount of melt refertilizes the depleted mantle. The petrological model presented in this study accounts for the complexity of polybaric dynamic melting and the continuous reactions between the residual solid and melt, but it is limited by the assumption of local thermodynamic equilibrium within a domain defined by the numerical grid size. The interpretation of the petrological results needs to be carefully evaluated to ensure that the time and space scale of the numerical model complies with the constraints provided by solid–melt reactive experiments and the spatial scale of the petrological structures observed in mid-ocean ridges. (7) Melt distribution and thermal structure are revealed by the seismic shear wave map computed from the numerical model. Certain observations, such as the extent of the melting region, overall agree quite well with the evidences from seismic studies from various ridge settings.