CASCADIA MAGMATISM PROVIDES UNIQUE INSIGHTS INTO SUBDUCTION ZONE (SZ) PROCESSES

The Cascadia arc is associated with one of the warmest modern SZs known. Extensive dehydration and volatile losses from the slab occur at shallow depths below the frontal arc, and slab-derived volatile contributions are likely minimal below the arc as reflected in Cascades basalts. These comprise two distinct compositional groups (1: low-K tholeiite and OIB-like; 2: calcalkalic) that, locally, are spatially and temporally coeval. Both groups have low fluid-mobile element contents for arc magmas (e.g., for most: B/Nb <1, low d11B <2‰; Group I Ba/Nb <20,).

Mineral thermometry shows basalts of both groups were consistently hot (eruptive T ≥ 1200 C), suggesting they are not products of flux-melting in the lower mantle wedge but rather formed by decompression melting of ‘normal’ (i.e., volatile-poor) mantle. This is supported by low H2O melt inclusions (<1%, Group 1; <2-3%, Group 2). Chemical diversity among Cascades basalts likely reflects heterogeneous mantle sources. Estimated magma segregation depths support a vertical compositional layering in the mantle, wherein the most flux-modified sources are confined to shallower parts of the mantle wedge and the more normal mantle sources are confined to greater depths. That is, Group 2 magmas are derived in part from lithospheric mantle domains containing ‘fossil’ slab contributions accrued over the 40 Ma lifespan of the arc, whereas Group 1 magmas likely derived from upwelling asthenospheric mantle that was little modified by slab contributions. Along-strike variations in basalt distributions and compositions suggest the thermal and compositional structure of the underlying mantle varies laterally.

On a global scale, intensity of fluid-related subduction fluxes appears to inversely correlate with arc thermal structure. Volcanic arcs associated with 'cooler' subduction zones (most cases) display evidence for larger subduction inputs (e.g., high contents of fluid-mobile elements), whereas those associated with warm subduction zones (e.g., Cascades, Mexico) do not. At cooler SZs, strong subduction-related modification may overprint and homogenize the mantle wedge. Conversely, warm SZs potentially reveal more information concerning initial mantle wedge compositional heterogeneity.