Reproducibility-Based Field Validation of Complementary Subsurface Sensing Using a Quantum Diamond Magnetometer and Multi-Frequency Ground-Penetrating Radar

Underground infrastructure inspection requires sensing methods capable of characterizing both ferromagnetic and non-ferromagnetic subsurface features. This study evaluated a portable nitrogen-vacancy (NV) quantum diamond magnetometer (QDM) together with multi-frequency ground-penetrating radar (GPR; 250, 500, and 800 MHz) at a 90 m full-scale pavement test site containing documented subsurface targets. Repeated stationary QDM measurements were first conducted to characterize temporal variability and establish site-specific magnetic baselines, followed by a 1 m interval spatial survey and corresponding GPR measurements. The QDM showed reproducible spatial magnetic patterns despite substantial variability in reinforcement-dense sections, and all 10 cross-session measurements remained within ±1.5 standard deviations of their repeated-measurement baselines. Ferromagnetic targets produced localized magnetic responses, whereas post hoc classification using disturbance-normalized magnetic anomalies showed limited target-level discrimination (AUC = 0.495). GPR likewise showed limited post hoc target classification (500 MHz AUC = 0.472), but section-level radar responses differed significantly across pavement environments. The two modalities exhibited contrasting section-level responses while showing weak and inconsistent point-level correspondence. These results indicate that NV-diamond magnetometry and GPR provide complementary physical information, while repeated magnetic measurements are important for interpreting QDM observations under heterogeneous field conditions.