Paradise Lost: Uncertainties in melting and melt extraction processes beneath oceanic spreading ridges

Abstract:

In many ways, decompression melting and focused melt transport beneath oceanic spreading ridges is the best understood igneous process on Earth. However, there are remaining – increasing – uncertainties in interpreting residual mantle peridotites.

Indicators of degree of melting in residual peridotite are questionable. Yb concentration and spinel Cr# are affected by (a) small scale variations in reactive melt transport, (b) variable extents of melt extraction, and (c) “impregnation”, i.e. partial crystallization of cooling melt in pore space. Roughly 75% of abyssal peridotites have undergone major element refertilization. Many may have undergone several melting events.

The following three statements are inconsistent: (1) Peridotite melt productivity beyond cpx exhaustion is > 0.1%/GPa. (2) Crustal thickness is independent of spreading rate at rates > 2 cm/yr full rate (excluding ultra-slow spreading ridges). (3) Thermal models predict, and observations confirm, thick thermal boundary layers beneath slow spreading ridges. If (a) melt productivity is << 0.1%/GPa beyond cpx-out, and (b) cpx-out occurs > 15 km below the seafloor beneath most ridges, then the independence of crustal thickness with spreading rate can be understood. Most sampled peridotites from ridges melted beyond cpx-out. Cpx in these rocks formed via impregnation and/or exsolution during cooling. Most peridotites beneath ridges may undergo cpx exhaustion during decompression melting. This would entail an upward modification of potential temperature estimates. Alternatively, perhaps oceanic crustal thickness does vary with spreading rate but this is masked by complicated tectonics and serpentinization at slow-spreading ridges.

Dissolution channels (dunites) are predicted to coalesce downstream, but numerical models of these have not shown why > 95% of oceanic crust forms in a zone < 5 km wide. There may be permeability barriers guiding deeper melt toward the ridge, but field studies have not identified them. Permeable “shear bands” may guide melt to the ridge, but their nature in open systems at natural grain size and strain rates is uncertain. 2D and 3D focused solid upwelling due to melt buoyancy deep in the melting region, where pyroxenes are abundant and permeability is low, may warrant renewed attention.