Constraints on melt production rate beneath the mid-ocean ridges based on passive flow models

Abstract

We present a model for computing the total melt production rate from the decompression partial melting region beneath a mid-ocean ridge, and the maximum oceanic crustal thickness created at the ridge axis assuming an ideal melt migration mechanism. The calculations are based on a self-consistent numerical model for the thermal structure and steady-state mantle flow field at a mid-ocean ridge. The model includes the effect of decreasing the melt production rate within the partial melting region by melt extraction as the residual mantle matrix becomes increasingly difficult to melt. Thus the melt fraction depends not only on temperature and pressure determined by the location beneath the ridge axis (the Eulerian description) but also on the accumulated melt extraction since the upwelling mantle matrix enters the partial melting region determined by the location along the flow-line path (the Langrangian description). This effect has been neglected by previous models. The model can predict the size of the melting region and the locations of the boundaries between mantle, residual mantle, and the partial melting region for a given spreading rate, also the distribution of the melt depletion and the mean melting depth. Given the observed average thickness of oceanic crust (∼6 km), which is relatively independent of spreading rate, the model results also provide a constraint on the overall efficiency of melt migration to the ridge axis; the efficiency must decrease from 100% at 10 mm/yr to about 60% at fast spreading rates (>50 mm/yr). Although this reduction may be partially due to the increasing size of the melting region with increasing spreading rate, it still requires less efficient melt migration near the ridge axis at fast spreading rate. We found that the calculated crustal thickness is very sensitive to the mantle temperature. For a normal mantle temperature of 1350°C, the model can generate the observed 6 km oceanic crust over the global range of spreading rates, while the anomalous thicker crusts of the Iceland hotspot and the Reykjanes Ridge are related to higher mantle temperatures associated with the hotspot. Finally, by comparing our model results with previous ones we found that neglecting variations of the melting relations of the residual mantle matrix with melt removal will overestimate the crustal thickness by at least a factor of 1.7.