Magmatic plumbing at Lucky Strike volcano based on olivine‐hosted melt inclusion compositions

Abstract

Here we present volatile, major, and trace element concentrations of 64 olivine-hosted melt inclusions from the Lucky Strike segment on the mid-Atlantic ridge. Lucky Strike is one of two locations where a crustal melt lens has been seismically imaged on a slow-spreading ridge. Vapor-saturation pressures, calculated from CO 2 and H 2 O contents of Lucky Strike melt inclusions, range from approximately 300-3000 bars, corresponding to depths of 0.5-9.9 km below the seafloor. Approximately 50% of the melt inclusions record crystallization depths of 3-4 km, corresponding to the seismically imaged melt lens depth, while an additional 35% crystallize at depths > 4 km. This indicates that while crystallization is focused within the melt lens, significant crystallization also occurs in the lower crust and/or upper mantle. The melt inclusions span a range of major and trace element concentrations from normal to enriched basalts. Trace element ratios at all depths are heterogeneous, suggesting that melts are not efficiently homogenized in the mantle or crust, despite the presence of a melt lens. This is consistent with the transient nature of magma chambers proposed for slower-spreading ridges. To investigate the petrogenesis of the melt inclusion compositions, we compare the measured trace element compositions to theoretical melting calculations that consider variations in the melting geometry and heterogeneities in the mantle source. The full range of compositions can be produced by slight variations in the proportion of an Azores plume and depleted upper mantle components and changes in the total extent of melting.

Spreading rate‐dependent variations in crystallization along the global mid‐ocean ridge system

“…Intermediate‐spreading rate melt inclusions are from two segments along the Juan de Fuca Ridge [Cleft and Vance segments; Wanless and Shaw , ] and two segments on the Galapagos Spreading Center [ Colman et al ., ]. Slow‐spreading melt inclusions are from the Lucky Strike segment [ Wanless et al ., ], a magmatically robust section of the Mid‐Atlantic Ridge. The ultraslow‐spreading melt inclusions are from five volcanic centers along the Eastern Volcanic Zone of the Gakkel Ridge [ Wanless et al ., ].…”

“…Vapor‐saturation pressures of melt inclusions, which are derived from CO 2 and H 2 O contents using experimental results of Dixon and Stolper [], provide an estimate of the pressure of melt entrapment or crystallization in MOR systems [e.g., Saal et al ., ; Wanless and Shaw , ] and thus do not record a polybaric crystallization history. Vapor‐saturation pressures of olivine‐hosted melt inclusions from several MORs show that crystallization occurs over a wide range of pressures from the uppermost mantle to the seafloor [ Saal et al ., ; Shaw et al ., ; Wanless and Shaw , ; Wanless et al ., ; Colman et al ., , Wanless et al ., ; Le Voyer et al ., ].…”

Petrogenesis and structure of oceanic crust in the Lau back-arc basin

“…very narrow magmatic system beneath ELSC4 that extends to shallow depths (∼2 km) with the seismically-imaged melt lens near the top, while the low-velocity zones underlying Domain III are broader and deeper (up to ∼3-4 km depth). Results from mantle Rayleigh wave tomography also suggests mantle melt extraction is more efficient for the ridge segments near the arc, as the low seismic velocity anomalies are weaker in the south indicating lower melt retention (porosity) despite an increase in melt supply to the ridge (Wei et al, 2015).…”

“…There it is proposed to form via underplating of ultramafic material due to an anomalously hot asthenosphere during the opening of the basin (Sato et al, 2014;Hirahara et al, 2015), similar to high velocity lower crust adjacent to continental margins (e.g., Kodaira et al, 1995). While high mantle potential temperature is thought to be a mechanism for generating ultramafic lower crust (e.g., Kelemen and Holbrook, 1995;Korenaga et al, 2000Korenaga et al, , 2002, this is unlikely in the Lau basin, since estimated mantle potential temperatures are lower in the south where the anomalous Domain II crust is currently forming Wiens et al, 2006;Wei et al, 2015). An analysis by Kelley et al (2006) suggests that, while mantle potential temperature seems to be responsible for most excess melt generation (and correspondingly thick crust) beneath many back-arc basins, the Lau back-arc appears to be an exception, showing a steep linear increase in extent of melting with mantle source H 2 O 0 .…”

“…Sato et al (2014) propose that the anomalous back-arc crust in the Japan Sea is the result of higher mantle potential temperature (e.g., Kelemen and Holbrook, 1995;Korenaga et al, 2002). However, this is likely not the underlying mechanism in the Lau basin, where mantle potential temperature is thought to be hottest in the northern areas of the basin, decreasing to the south where the anomalous crust is actually observed Wiens et al, 2006;Wei et al, 2015).…”