Melt transport in the upper mantle and lower crust beneath oceanic spreading ridges
Melt transport in the upper mantle and lower crust beneath oceanic spreading ridges (last updated in 2000)
• In adiabatic regions of the upper mantle, reaction between ascending melt and peridotite dissolves pyroxene, creates olivine, and increases the mass of melt. This leads to unstable formation of a coalescing network of high porosity dissolution channels. Some channels, from which all pyroxene has dissolved, are observed as dunite veins within peridotite in the mantle section of many ophiolites. Dunites have a power-law width/frequency relationship, as predicted for a coalescing network of channels that conserves flux, in which width is related to flux. Flow of ascending melt through dunite channels allows liquids equilibrated with peridotite at relatively high pressure (> 1 GPa) to escape equilibration with pyroxene at depths less than 30 km.
• In the conductive region (crust and uppermost mantle), reaction generally consumes olivine, producing pyroxene and plagioclase, and decreases the mass of melt. Where crystallization in pore space is more rapid than viscous relaxation, this leads to diverging, random porous flow, ponding of melt beneath a permeability barrier, increasing melt pressure, and hydrofracture. Periodic pressure changes associated with hydrofracture may explain some modal layering in gabbros as well as formation of sheeted dikes. Where crystallization is slower than viscous relaxation, the mantle may "decompact" to maintain porosity, and evolving melt may ascend into the thermal boundary layer. Incompatible element concentrations increase as melt mass decreases, while Ni and Mg# remain high, in equilibrium with mantle olivine. Ultimately, enriched melts may become fluid saturated.
• In summary, flow in regions where dissolution increases pore space becomes increasingly organized into a coalescing channel network, while flow in regions where precipitation decreases pore space becomes increasingly diffuse, punctuated by the formation of transient, single conduits fed by an upstream reservoir. This may have analogues in the morphology of drainage networks formed via erosion and deposition of sediments.
• ONGOING RESEARCH on this topic includes characterizing the multi-scale geometry of dunite networks in ophiolite mantle sections, investigating the nature of chemical variation in layered gabbros from ophiolite lower crustal sections, observation of drainage networks on beaches at low tide, various laboratory scale experiments on channel formation and evolution, and theoretical modeling of (a) geochemical output from a network of channels formed by reactive porous flow in a partially melting mantle, and (b) processes at the transition from dissolution reactions in the adiabatically convecting mantle, to crystallization reactions in the thermal boundary layer.