Low-velocity zone atop the 410-km seismic discontinuity in the northwestern United States

The seismic discontinuity at 410 km depth in the Earth’s mantle is generally attributed to the phase transition of (Mg,Fe)2SiO4 (refs 1, 2) from the olivine to wadsleyite structure. Variation in the depth of this discontinuity is often taken as a proxy for mantle temperature owing to its response to thermal perturbations. For example, a cold anomaly would elevate the 410-km discontinuity, because of its positive Clapeyron slope, whereas a warm anomaly would depress the discontinuity. But trade-offs between seismic wave-speed heterogeneity and discontinuity topography often inhibit detailed analysis of these discontinuities, and structure often appears very complicated. Here we simultaneously model seismic refracted waves and scattered waves from the 410-km discontinuity in the western United States to constrain structure in the region. We find a lowvelocity zone, with a shear-wave velocity drop of 5%, on top of the 410-km discontinuity beneath the northwestern United States, extending from southwestern Oregon to the northern Basin and Range province. This low-velocity zone has a thickness that varies from 20 to 90 km with rapid lateral variations. Its spatial extent coincides with both an anomalous composition of overlying volcanism and seismic ‘receiver-function’ observations observed above the region. We interpret the lowvelocity zone as a compositional anomaly, possibly due to a dense partial-melt layer, which may be linked to prior subduction of the Farallon plate and back-arc extension. The existence of such a layer could be indicative of high water content in the Earth’s transition zone. Spatial variations in topography of the 410-km and the 660-km discontinuities (referred to here as the 410 and 660) are often inferred from SS precursors (length scale of about 1,500–2,000 km), near subduction zone depth phase precursors, and receiver function analyses (length scale of about 100–300 km). The latest efforts consider simultaneous inverting for both mantle velocity and discontinuity topography. In general, the 660 is depressed under cold regions (slabs) as expected, but the 410 appears to be far more complicated. The stacked converted P-to-S phase (Ps) (receiver function) from the 410 is rather weak, complicated and sometimes shows negative pulses above the 410 (ref. 7). We first compute one-dimensional (1D) full-waveform synthetics, model the direct S-wave triplications at epicentral distances of 14–178 and explain the timing and amplitude of multiple arrivals coming from fine structures near the 410 (ref. 10). To resolve the trade-offs between discontinuity topography and mantle velocity directly above or below the discontinuity, we model S-wave triplications at epicentral distances of 21–248. A low-velocity zone (LVZ) on top of the 410 produces a secondary pulse not normally seen at these distances. Using this secondary pulse as a proxy for the existence of an LVZ atop the 410, we examine events located offshore of Washington–Oregon, USA, and recorded by the TriNet broadband network and several temporary PASSCAL broadband arrays (Fig. 1). Perturbed velocity structures are shown at the turning point for a given great-circle path because the sensitivity to 410 structure is greatest there. The size of the perturbed structure is estimated as the size of the Fresnel zone in both along-path and cross-path directions. We determine a LVZ directly above the 410 in the northwestern US and cross-validate our finding with receiver-function profiles. Record section A shows typical waveform characteristics sampling this area at epicentral distances of 14–178 (Fig. 2a). It samples the region beneath the California-Oregon border (Fig. 1).