INITIATION OF THE CASCADE ARC BY NORTHWARD MIGRATION OF THE FARALLON SLAB

The Cascade arc was one of the first magmatic arcs to be studied in detail, but the timing and mechanism of its initiation remain poorly known. Subduction began after accretion of the oceanic Siletzia terrane and consequent breakoff of the Farallon slab, both at ~48-50 Ma (Tepper & Eddy, 2016). However, because the Farallon slab was young and hot, mechanisms proposed for subduction initiation in other arcs (Stern & Gerya, 2016), are not tenable here. We propose an alternative process in which an already subducting segment of Farallon slab migrated northward along the continental margin, gradually reoccupying the no-slab gap left under Oregon and Washington by earlier breakoff. This model is consistent with plate motions (which indicate northward movement of oceanic plates relative to North America at that time), and eliminates the need to initiate subduction of oceanic lithosphere that is young and buoyant.

The best test of this model is provided by the locations and ages of Cascade arc rocks, which constrain the history of Farallon plate motion relative to North America since the mid-Eocene. In Washington we suggest the earliest rocks associated with the Cascade arc occur in a NW-trending belt that extends ~150 km from near Cle Elum in the central Cascades to near Chimacum on the NE Olympic Peninsula. Units that define this belt include the Mt. Persis, Tukwila, and Hansen Lake volcanics, portions of the Naches Formation, several small plutons (Youngs Creek, Granite Falls), and the adakites at Mount Zion and Chimacum. Generation of these rocks in a subduction setting is indicated by their calc-alkaline trends, HFSE depletions, and compositional diversity (B-A-D-R). LA-ICP-MS U-Pb zircon ages of these eight units cluster between 43-45 Ma, making them several m.y. younger than units associated with slab breakoff. Based on motion vectors for the Farallon slab (calculated from Babcock et al., 1992) the location and age of this belt (and its adakite affinities) are compatible with its having formed above the leading edge of the Farallon (now called Juan de Fuca) slab as it moved under Washington and its dip steepened. This model would predict older arc rocks to the south; their apparent absence in southern Washington and Oregon may reflect a shallow dip of the Farallon slab, a relict feature of its rapid subduction prior to 50 Ma.