L2 Lithium as a Potential Primary Mantle Conduit Tracer: Deep Lithospheric Electromagnetic Architecture and Cross-Commodity Non-Parametric Validation

Global mineral exploration is increasingly constrained by the challenge of targeting deposits beneath post-mineralization cover. While magmatic and hydrothermal ore genesis models emphasize shallow crustal or regional structural controls, the planetary-scale geodynamic driver that differentiates ore-forming systems from unmineralized baselines remains debated. Following the identification of a statistically anomalous vertical conductivity architecture beneath the GRID_78 lithium cluster (Echo of Axis, 2026L1), we present a global depth-stratified electromagnetic analysis across Ntotal = 162,937 georeferenced mineral occurrences sampled from the GlobalEM-2015 3D magnetotelluric model at 12 km, 35 km, and 75 km depth horizons. By deploying two-sample Kolmogorov-Smirnov (KS) tests, Monte Carlo permutation procedures (N = 10,000 iterations, seed = 42), and Dunn–Bonferroni post-hoc rank-sum matrices across 22 groups (21 MRDS commodity classes plus GRID_78 Li nodes), we establish an empirical spectrum of total vertical conductivity step-ups (Δσgrad (75→12) = σ75km - σ12km). The purely sedimentary baseline (Limestone, General; n = 2,096) yields the lowest observed gradient (mean Δσgrad (75→12) = 0.0525 log₁₀ S/m), while base and precious metal hydrothermal systems demand progressively deeper trans-crustal plumbing. Crucially, primary lithium network nodes (Group A, NA = 13 verified GRID_78 nodes) exhibit the highest mean vertical step-up across all 22 groups (Δσgrad (75→12) = 0.9963 log₁₀ S/m). Monte Carlo permutation testing against the non-GRID lithium background (Group B, NB = 94) yields z = 2.654, p = 0.0076; two-sample KS: D = 0.5802, p = 0.000393. Depth stratification reveals that upper-crustal conductivity (σ12km) is statistically indistinguishable between groups (mean: -2.7294 vs. -2.7231 log₁₀ S/m), whereas sub-Moho divergence drives the signal: mean σ75km = -1.7331 (Group A) vs. -2.0959 (Group B) log₁₀ S/m. The sub-Moho gradient, Δσgrad(75→35) = σ75km - σ35km, provides the strongest discriminant: Group A mean = 0.4949 vs. Group B mean = 0.1514 log₁₀ S/m (MC z = 3.279, p = 0.0011). Kruskal-Wallis testing across all 22 groups on absolute sub-lithospheric conductivity (σ75km) yields H = 6093.6, p ≈ 0.These results establish that GRID_78 lithium nodes are not merely embedded in high-gradient terrain - they occupy a physically distinct, deeply rooted electromagnetic regime, motivating their evaluation as potential primary mantle conduit tracers. Keywords: lithium deposits; mantle conduit; electrical conductivity; magnetotellurics; lithospheric architecture; conductivity gradient; MRDS; non-parametric statistics; commodity discrimination; GlobalEM