MAFIC MAGMATISM AND NEW PERSPECTIVES ON CASCADIA SUBDUCTION

Widespread Cascadia mafic magmatism comprises two compositionally distinct groups: [1] CA basalts show relative LILE-enrichments and HFSE- depletions seen in many arc lavas, and commonly attributed to influx of slab-derived fluids to a mantle source. [2] WIP lavas resemble within-plate basalts in lacking the above features; two distinct variants are MORB-like low-K tholeiite (LKT), and OIB-like basalts with elevated K and incompatible elements. Quaternary-Pliocene variants of these basalts are distributed over a >160 km long E-W transect (~46°N), with LKT and ne-normative OIB erupted from some westernmost vents. CA variants occur in all but the easternmost part of the transect.

To address basalt genesis, P-T of magma segregation was estimated for the most primitive analyzed samples (sparsely ol+sp±cpx phenocrystic, Mg# >65, MgO > 7-10%). Eruptive Ts >1180-1300°C and olivine compositions (~Fo84-91) attest to only slight removal from primitive mantle melts. Correction to Fe-Mg equilibrium with model mantle (Fo89-92) requires addition of small amounts of equilibrium olivine. Two important findings are: [1] Higher Mg# and Fo in equilibrium olivines for CA basalts reflects more refractory sources than for LKT or OIB types. [2] LKTs have clear compositional distinctions, and also appear to segregate from systematically greater depths than CA types (OIB overlap both).

While CA magmas could originate near the subducting slab and subsequently reequilibrate in the shallow mantle, this scenario is problematic in that such ‘juicy’ magmas would have to ascend through and have negligible impact on the unenriched LKT sources while retaining the respective chemical identities. Moreover, the distribution of LKT and OIB basalts at 46°N implies their sources must be present scarcely above the Cascadia slab, yet they lack a slab-fluid signature. A simple though controversial alternative is that Cascadia CA types represent melts of shallow lithospheric mantle (metasomatized during early Cascades magmatism). In this scenario, there is little modern slab-flux involved (consistent with weak fluid mobile element and 10-Be signatures). Instead, decompression melting in a warm mantle wedge could produce LKTs that in turn ascend, warm the shallow mantle, and promote remelting (CA and some OIB-type basalts).