金曜日セミナー：(2020年2月14日) Dr. Geoffrey Abers (Cornell University)

Searching for the deep roots of arc volcanoes: seismic imaging in the Washington Cascades

Many arc volcanoes erupt mantle-sourced basalts and high-temperature lower-crustal magmas, yet seismic images of this deep plumbing are almost non-existent. Are deep partial-melt bodies too small for seismology to see, or are typical seismic arrays too limited to detect them? We address these questions with a 2014-16, 70-station broadband array around Mount St. Helens (MSH), termed iMUSH (imaging Magma Under St. Helens). Receiver functions image 35-40 km of upper-plate crust and a subducting plate 65-68 km directly beneath MSH. This leaves very little mantle to melt, yet basalts are present. As elsewhere in the Cascades, the upper-plate Moho vanishes in the forearc (west of MSH), an observation usually interpreted as serpentinized mantle. Ambient-noise tomography shows low wavespeeds here consistent with hydrated mantle, requiring temperatures low enough for serpentines to be stable, as does low forearc heat flow. Much of the reduction in Moho strength is a consequence of upper-plate geology: very high wavespeeds are observed in the lower crust west of MSH, greatly reducing the Moho velocity contrast. We see little evidence for melt except far east of MSH and overall these observations show little sign of the deep magma plumbing system. In fact, they imply that the lower crust and uppermost mantle are too cold directly beneath the edifice to generate the observed melts. Seismic body-wave attenuation also confirms large temperature contrasts between the forearc and back arc. To reconcile these observations with the volcanism at MSH requires significant lateral melt transport within the crust. These pathways are not directly visible with seismic imaging techniques currently available, motivating the need for other approaches.