Tephra layers in Quaternary sediments (<1 Ma) in the central Mariana Trough contain three different series of glass shards: (1) low-K basaltic to basaltic-andesitic series (LKB: SiO 2 = 47.3–53.6 wt.%; K 2O = 0.2–0.8 wt.%); (2) medium-to high-K basaltic to dacitic series (HKS: SiO 2 = 47.6–70.9 wt.%; K 2O = 0.8–3.2 wt.%); and (3) low-K dacitic to rhyolitic series (LKDR: SiO 2 = 63.9–77.9 wt.%; K 2O = 0.8–1.5 wt.%). LKB glasses are sideromelane shards which occur as minor component in two thirds of the tephra layers investigated. HKS glasses, mostly pyroclastically fragmented, dominate the discrete millimeter- to centimeter-thick fallout layers and are ubitiquous as dispersed shards. LKDR series comprise silicic pumice shards that prevail in decimeter-thick ash turbidites and crystal-poor, lapilli- to bomb-sized pumices. The composition of the LKB glass shards match those of coeval basalts and basaltic andesites of the Mariana Trough. LKB glasses are interpreted to have formed by the spalling of glassy rims of deep-submarine pillow and sheet lavas and subsequently became incorporated in the arc-derived tephra fallout and flow deposits. HKS and LKDR series both compositionally differ from coeval CIP lavas. Characteristic trace-element ratios, however, such as high large-ion-lithopile-elements (LILE)/high-field-strength elements (HFSE) ratios (Ba/La= 33–81; Cs/Nb= 0.3–0.8) and Nb depletion (La/Nb= 2–22) constrain their origin from explosive eruptions of the Central Island Province (CIP) of the Mariana arc. The relative enrichments of semi-incompatible elements preclude HKS and LKDR series from being linked by fractional crystallization and they are interpreted as derivative liquids from different magmatic series of the CIP. HKS glasses are suggested to be counterparts of CIP volcanos with medium-K trends (≈ 1.5–2 wt.% K 2O at 60 wt.% SiO 2), whereas the relatively LILE-depleted LKDR series are derived from CIP volcanic islands and seamounts with lower-K trends (≈; 0.5–1.5 wt.% K 2O at 60 wt.% SiO 2). The magmatic diversity of the CIP — and thus the variability of the distal tephra deposits — can be explained neither by crustal-level differentiation nor by variable degrees of partial melting of a homogenous mantle source. In view of the isotopic homogeneity of the CIP, it can best be explained by the Stolper and Newman (1994) model of binary mixing of source components, with one endmember being the NMORB source of Stolper and Newman (1994) and the other one being a compositionally variable slab-derived H 2O-rich component similar to the H 2O-component of Stolper and Newman (1994). This interpretation implies that the variation in K 2O-content in the arc-derived glasses (about a factor of 3 at 60 wt.% SiO 2) reflects a magmatic diversity that is inherent to the central Mariana arc and is not necessarily reflective of temporal changes of the composition of arc magma sources.
Read full abstract