Sequencing seismograms: A panoptic view of scattering in the core-mantle boundary region
Sequencing for seismic structures Structures illuminated by seismic waves at the core-mantle boundary of the Earth are traditionally found by focusing on a specific target area. Kim et al. used an unsupervised manifold learning algorithm called “the Sequencer” to automatically detect anomalies in seismic data (see the Perspective by Miller). Using this technique, they uncovered structures at the core-mantle boundary across the entire Pacific region all at once. They found many structures previously identified, but also a new, ultra-low-velocity zone beneath the Marquesas Islands. Science, this issue p. 1223; see also p. 1183 An unsupervised machine learning algorithm uncovers the structure of the core-mantle boundary region under the Pacific. Scattering of seismic waves can reveal subsurface structures but usually in a piecemeal way focused on specific target areas. We used a manifold learning algorithm called “the Sequencer” to simultaneously analyze thousands of seismograms of waves diffracting along the core-mantle boundary and obtain a panoptic view of scattering across the Pacific region. In nearly half of the diffracting waveforms, we detected seismic waves scattered by three-dimensional structures near the core-mantle boundary. The prevalence of these scattered arrivals shows that the region hosts pervasive lateral heterogeneity. Our analysis revealed loud signals due to a plume root beneath Hawaii and a previously unrecognized ultralow-velocity zone beneath the Marquesas Islands. These observations illustrate how approaches flexible enough to detect robust patterns with little to no user supervision can reveal distinctive insights into the deep Earth.