Fine structures of the uppermost mantle beneath the northwestern United States revealed by deep-learning-based phase picking and Pn tomography

Le résumé fourni par la source

The northwestern United States is characterized by intense tectonic deformation and active deep-mass-transport processes, making it an ideal natural laboratory for studying the exchange of crustal and mantle materials and the transfer of energy. To further investigate the spatial variations in physical properties of the Cascadia subduction zone and the distribution of hot materials beneath the Yellowstone–Snake River Plain volcanic province, we inverted Pn-wave travel times extracted by deep-learning-based phase picking technique to obtain the Pn velocity and anisotropy model in the uppermost mantle beneath this region. Imaging results show that the Pn-wave velocity and anisotropy beneath the Cascadia subduction zone display significant along-strike segmentation, reflecting spatial heterogeneity of physical properties caused by the subduction of the Juan de Fuca plate, which provides new constraints on the segmented subduction model. The study clearly delineates an independent low-velocity anomaly in the uppermost mantle beneath Yellowstone volcano, suggests the existence of a magmatic gap in the uppermost mantle between Yellowstone and the low-velocity zone of the Snake River Plain to the west, and supports a model of inclined upwelling of hot materials beneath Yellowstone. The prominent high-velocity feature in the northern Columbia Plateau provides new seismological evidence for the subducted remnants of the ancient Farallon plate. This work not only demonstrates the significant value of deep-learning-based phase picking techniques for improving tomographic studies, but also establishes a high-quality model of uppermost mantle Pn velocity and anisotropy, thereby deepening the understanding of the complex deep structure of this region.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.