Convection and melting at mid‐ocean ridges

Open Access
Publisher Website
Google Scholar
Abstract
We present a thermodynamically self‐consistent model of plate and buoyancy driven flow and melt generation beneath mid‐ocean ridges. Mantle flow is driven by a rigid lithosphere and by buoyancy forces resulting from melting, depletion, and melt extraction. Melt is generated using the solidus of Kinzler and Grove (1992a). Constant viscosity models without melt buoyancy forces show that no significant narrowing of the melting region or pressure gradient focusing of the melt is achieved. Temperature‐dependent viscosity models show that pressure gradients in a high viscosity lithosphere are insufficient for focusing melt to the ridge axis. Constant viscosity models with melt buoyancy forces show that significant narrowing of the melting region is possible at slow spreading rates but not at fast spreading rates. Melt buoyancy forces cause the crustal thickness to decrease with spreading rate. Aggregate melts are similar to those required to form mid‐ocean ridge basalt (MORB) but are too depleted in incompatible elements. They are also insensitive to spreading rate and melt region shape but are sensitive melting rate distribution. Model residuum trends are similar to those in abyssal peridotites and imply that abyssal peridotites result from partial melting and not refertilization. Massif peridotites appear to result from refertilization of harzburgite with MORB primary melt. Vertically integrated melts show very similar trends to data from the off‐axis Lamont seamounts on the flanks of the East Pacific Rise. This implies the melting region at fast spreading ridges is at least 80–100 km wide.