Researchers used deep learning on decades of seismic data to identify six previously undocumented heterogeneous zones (or strong scatterers) near Earth’s core-mantle boundary (CMB), roughly 2,900 km (about 1,800 miles) deep.

The work, led by scientists including Yurui Guan and colleagues at the Institute of Geology and Geophysics, Chinese Academy of Sciences, was published in 2026 in the Journal of Geophysical Research: Solid Earth (Guan et al.).

It analyzes PKP precursors—faint seismic signals that arrive slightly ahead of the main PKP/PKIKP waves. These precursors arise when seismic waves from large earthquakes scatter off small-scale irregularities (heterogeneities) just above the CMB.

A supervised deep-learning system (combined with iterative human-guided optimization) processed more than 2 million vertical-component waveforms from earthquakes (typically Mw ≥ 6.0) recorded between about 1990 and 2024.

It automatically detected roughly 174,929–175,000 high-quality PKP precursor signals—more than ten times the total from all previous global studies combined. This dramatically improved sampling density and coverage.

Earlier maps showed relatively sparse, isolated scattering patches. The expanded dataset reveals many of these connecting into broader, continuous belts or bands in the lowermost mantle.

The analysis independently confirmed known scattering regions and, for the first time, highlighted six new high-potential strong-scattering zones (labeled B1–B6 in some reports). These lie in previously undersampled areas, including regions beneath high-latitude Eurasia, Central Asia, the South Atlantic, parts of the circum-Antarctic domain, and other locations.

These small-scale structures act as “deep-seated scatterers” that affect how seismic waves travel.

Possible origins (within the study’s interpretive limits) include remnants of multi-episode subducted slabs (ancient oceanic crust and related material dragged into the deep mantle), localized partial melting, mineral phase transitions under extreme pressure/temperature, and interactions with larger known features such as large low-shear-velocity provinces (LLVPs, the continent-sized “blobs” under Africa and the Pacific) or ultra-low-velocity zones (ULVZs).

The CMB is a critical interface for heat transfer and material exchange between the mantle and core.

Mapping these fine-scale heterogeneities improves understanding of mantle convection, possible plume roots, chemical cycling, and how deep processes may ultimately influence surface phenomena such as volcanism and tectonics. The six new zones are flagged as priority targets for future multi-phase seismic imaging and other studies.

_____________________________________________________________________________________

Global Distribution of PKP Precursors Derived From Three Decades of Seismic Data With Deep Learning

“Global Distribution of PKP Precursors Derived From Three Decades of Seismic Data With Deep Learning” is a 2026 research article by Yurui Guan and colleagues (primarily from the Institute of Geology and Geophysics, Chinese Academy of Sciences).

The study presents the most comprehensive global map to date of small-scale seismic scatterers near Earth’s core-mantle boundary (CMB, ~2,900 km depth). It does this by systematically detecting and analyzing PKP precursors—faint seismic signals that arrive slightly before the main PKP/PKIKP core phases. These precursors are generated when seismic waves scatter off fine-scale lateral heterogeneities (compositional or thermal variations) in the lowermost mantle.

Methods and data:

• Researchers combined supervised deep learning with iterative human-guided optimization.
• They processed over 2 million vertical-component waveforms from earthquakes with magnitude Mw ≥ 6.0 recorded between 1990 and 2024.
• The system automatically identified 174,929 high-quality PKP precursor signals—roughly an order of magnitude more than all previous global compilations combined. (Related conference abstracts sometimes cite slightly different totals, such as ~227,000, depending on quality thresholds.)
• They introduced a dual-probability framework that integrates:
  • Precursor occurrence probability (how consistently signals appear in a region).
  • Scatterer location probability (likely positions of the scattering features).

This framework evaluates the spatial stability of observations and helps pinpoint localized strong scatterers.

Key results:

• Sampling density and geographic coverage improved dramatically compared with earlier manual or smaller-scale studies.
• Previously isolated scattering patches (e.g., beneath the Pan-American region) now appear connected into continuous bands or belts.
• The expanded dataset confirms known scattering regions associated with deep anomalies and, for the first time, identifies six high-potential strong-scattering zones in previously undersampled areas. These include regions beneath the South Atlantic, high-latitude Eurasia, Central Asia, and circum-Antarctic domains.
• Small-scale scatterers occur in both high-velocity and low-velocity domains of the lower mantle, indicating diverse origins that are not limited to the edges of large low-shear-velocity provinces (LLVPs).

Possible sources of the heterogeneities include remnants of multi-episode subducted slabs, localized partial melting, mineral phase transitions, and interactions with larger-scale features such as LLVPs or ultra-low-velocity zones (ULVZs). The results underscore the complexity of the lowermost mantle and provide prioritized targets for future multi-phase joint inversions and high-resolution imaging of CMB structure.

This work advances understanding of mantle convection, core-mantle interactions, heat and material exchange, and the potential links between deep-Earth features and surface processes. Related presentations (e.g., at EGU 2026) under titles such as “Global Mapping of Small-Scale Heterogeneities at the Core-Mantle Boundary” describe the same dataset and findings.

Published:  Journal of Geophysical Research: Solid Earth

DOI: 10.1029/2025jb033195

Authors: Yurui Guan, Juan Li, Zhuowei Xiao, Wei Wang, Tao Xu

Provided: Institute of Geology and Geophysics, Chinese Academy of Sciences

Abstract

Small-scale lateral heterogeneities in the lowermost mantle are crucial for understanding mantle convection and core-mantle interactions. Owing to their high sensitivity to fine scatterers, PKP precursors provide a powerful probe of deep Earth heterogeneities. However, manual identification of these precursors is inefficient, subjective, and insufficient for vast global seismic data sets. Here, we integrate deep learning with iterative manual optimization to process over 2 million waveforms from earthquakes (Mw ≥ 6.0) between 1990 and 2024, automatically identifying 174,929 precursors, an order of magnitude more than prior data sets, to build the most comprehensive global map of potential scattering distributions. We propose a dual-probability framework integrating precursor occurrence and scatterer location probabilities to assess the global spatial stability of precursor observations and to constrain the probable locations of localized strong scatterers. By applying this extensive catalog within the dual-probability framework, we achieved an order-of-magnitude improvement in global sampling density and a substantial expansion of spatial coverage over previous individual-detection studies, connecting once-isolated sampling zones (e.g., beneath Pan-America) into continuous bands. Cross-validation with independent seismic phases confirms the robust embedding of multiple ultra-low velocity zones (ULVZs) within diverse velocity heterogeneity backgrounds, suggesting that these structures represent thermochemical piles shaped by some combination of multi-episode subducted slab remnants, localized partial melting, and interactions with large low-velocity provinces (LLVPs). Extension to undersampled regions reveals, for the first time, six high-potential scattering zones, laying the groundwork for prioritized target areas in subsequent multi-phase joint inversions and fine-scale imaging of core-mantle boundary (CMB) heterogeneity.