Quantifying plate-boundary geometric control on apparent mantle–seismicity coupling

Global correlations between upper-mantle seismic structure and shallow earthquake occurrence are commonly interpreted as evidence of mantle–lithosphere coupling. However, both seismicity and low shear-wave velocity anomalies are strongly concentrated near tectonic plate boundaries, raising the possibility that apparent coupling may arise from their shared geometric organization rather than from an independent physical relationship. Here, I quantitatively evaluate the extent to which global mantle–seismicity correlations are controlled by plate-boundary geometry. Using exclusively open-access datasets, I combine shallow seismicity from the USGS ANSS Comprehensive Catalog (2000–2025; M ≥ 5.5; z ≤ 70 km), upper-mantle shear-wave velocity perturbations from the SPani tomography model, PB2002 global plate boundaries, and GSRM strain-rate fields within a fully reproducible framework. Mantle structure and seismicity are compared on a common global grid using non-parametric rank statistics, bootstrap uncertainty quantification, depth-dependent analyses, and distance-to-boundary diagnostics. Significant negative correlations are observed between low-velocity mantle anomalies and shallow seismicity, particularly for upper-crustal earthquakes (0–30 km). However, both variables exhibit a strong dependence on proximity to plate boundaries. After explicitly removing the first-order distance-to-boundary trend, the residual global correlation is substantially reduced. However, an additional GCMT-based faulting-style analysis shows that statistically significant residual correlations persist in mechanically distinct tectonic regimes, particularly for strike-slip seismicity. These results indicate that a substantial fraction of the apparent global mantle–seismicity coupling reflects tectonic geometric organization rather than an independent mantle-controlled signal. The study highlights the importance of explicitly accounting for plate-boundary geometry before drawing causal geodynamic inferences from large-scale mantle–seismicity correlations.