Predicting Rates and Distribution of Carbonate Melting in Oceanic Upper Mantle: Implications for Seismic Structure and Global Carbon Cycling

Coupling a global mantle flow model with a parameterized carbonate solidus, we examine the global distribution and extent of carbonate melting beneath the ocean basins. We predict carbonate melting in spatially heterogeneous patterns throughout the oceans. The rate of CO2 segregation from the mantle by off‐axis carbonate melting (~1.1 × 1012 mol/year) is comparable to the global ridge flux (~1.2 × 1012 mol/year). As the generation of carbonate melts should be enhanced in regions of mantle upwelling, we compare upwelling patterns with seismic detections of the G‐discontinuity. The upwelling velocities in areas where the G‐discontinuity is detected are not statistically different from areas with no detections—implying that the generation and pooling of carbonate melts are not dominant mechanisms in forming the G‐discontinuity. However, detections of a deeper seismic discontinuity are correlated with enhanced upwelling velocities, suggesting locally higher melt fractions at the transition between carbonate‐only and carbonate‐enhanced silicate melting. Plain Language Summary Despite support from indirect observations, the existence of a layer of carbon‐rich, partially molten rock (~60 km) below oceanic crust, made possible by the presence of CO2, remains uncertain. In particular, abrupt decreases in the velocity that seismic waves propagate at depths of 40–90 and 80–180 km beneath the ocean basins remain unexplained. In this study, we test whether these seismic discontinuities can be attributed to the presence of a layer of carbon‐rich melt. Melt generation occurs only where the mantle is upwelling; thus, we predict the locations of carbonate‐enhanced melting using a mantle convection model and compare the resulting melt distribution with the seismic observations. We find that the shallower seismic discontinuities (at 40‐ to 90‐km depth) are not associated with regions of predicted melting but that the deeper discontinuities (80–180 km) occur preferentially in areas of greater mantle upwelling—suggesting that these deep observations may reflect the presence of localized melt accumulation at depth. Finally, we show that carbonate melting far from mid‐ocean ridges produces an additional CO2 flux previously overlooked in deep carbon cycle estimates, roughly equivalent to the flux of CO2 due to seafloor volcanism. Key Points Using calculated mantle upwelling patterns, we predict the presence of low‐degree carbonate melts throughout the oceanic basins Globally, the