Geochemistry of volcanic glasses from the Louisville Seamount Trail (IODP Expedition 330): Implications for eruption environments and mantle melting

Abstract

Volcanic glasses recovered from four guyots during drilling along the Louisville Seamount Trail, southwest Pacific, have been analyzed for major, trace, and volatile elements (H 2 O, CO 2 , S, and Cl), and oxygen isotopes. Compared to other oceanic island settings, they are geochemically homogeneous, providing no evidence of the tholeiitic stage that characterizes Hawai'i. The degrees and depth of partial melting remained constant over 1-3 Ma represented by the drill holes, and along-chain over several million years. The only exception is Hadar Guyot with compositions that suggest small degree preferential melting of an enriched source, possibly because it erupted on the oldest and thickest lithosphere. Incompatible element enriched glass from late-stage volcaniclastics implies lower degrees of melting as the volcanoes moved off the melting anomaly. Volcaniclastic glasses from throughout the igneous basement are degassed suggesting generation during shallow submarine eruptions (<20 mbsl) or as subaerial flows entered the sea. Drill depths may no longer reflect relative age due to postquench downslope movement. Higher volatile contents in late-stage volcaniclastics indicate submarine eruptions at 118-258 mbsl and subsidence of the edifices below sea level by the time they erupted, or generation in flank eruptions. Glass from intrusion margins suggests emplacement 100 m below the surface. The required uplift to achieve these paleo-quench depths and the subsequent subsidence to reach their current depths exceeds that expected for normal oceanic lithosphere, consistent with the Louisville melting anomaly being <100 C hotter than normal asthenosphere at 50-70 Ma when the guyots were erupted.

Exploring the mineralogical heterogeneities of the Louisville Seamount Trail

“… Zr/Nb versus La/Yb for Louisville whole‐rock and glass samples [ Beier et al ., ; Nichols et al ., ], the liquids calculated from Louisville clinopyroxene cores, and Hawaiian whole‐rocks for reference [ Chen and Frey , ; West et al ., ; Frey et al ., , ; Wolfe et al ., ; Yang et al ., ; Huang and Frey , ]. Hawaiian tholeiitic shield stage rocks have low La/Yb and high Zr/Nb values.…”

“…The average phenocryst Al iv contents for the phenocrysts of Figure were used to determine clinopyroxene‐basalt partition coefficients from the equations in Bédard []. For both sets of partition coefficients, the LREE matched the whole‐rock and glass values [ Beier et al ., ; Nichols et al ., ], but the calculated liquids had slightly lower HREE concentrations than the whole‐rocks. Since Louisville liquid compositions are well represented by aphyric basalts and glasses, the cpx/basalt HREE partition coefficients were adjusted to produce calculated liquids that matched representative whole‐rock values and glasses (supporting information Table S3).…”

“…Dotted gray lines are whole‐rock values from Beier et al . [], gray lines are glasses from Rigil seamount (6A) and Burton seamount (6B) [ Nichols et al ., ]. Normalization constants from Anders and Ebihara [1982].…”

“…Most recently, Nichols et al . [] concluded that none of their analyzed glasses from Louisville rocks exhibit a transitional to tholeiitic composition but are alkalic basalts (sensu strictu) as did Vanderkluysen et al . [] for dredged samples from 22 Louisville seamounts.…”

Re‐Os isotope and platinum group elements of a FOcal ZOne mantle source, Louisville Seamounts Chain, Pacific ocean

“…Rhenium loss, leading to an insufficient age‐correction and thus a higher than expected initial 187 Os/ 188 Os ratio based on olivine data, may have been caused by shallow submarine to subaerial degassing [ Lassiter , ] as most of the drilled LSC lavas were degassed [ Nichols et al ., ]. On the other hand, Re gain during submarine alteration would have led to too much of an age correction for some Rigil and Burton seamounts data that plot off the isochron lines, leading to much lower than expected initial 187 Os/ 188 Os ratios indicated by the olivine data (Table ).…”