A hybrid workflow combining the Plus–Minus method and tomography to resolve sharp interfaces in near‐surface seismic refraction

Abstract Near‐surface seismic refraction tomography is a powerful tool for imaging shallow subsurface structures, yet conventional approaches often fail to resolve sharp velocity contrasts at geological interfaces due to inherent smoothness constraints. We present a hybrid methodology that combines Hagedoorn's Plus–Minus ( T ±) method with refraction tomography, constructing geologically plausible 2D starting models with predefined interfaces to overcome these limitations. Validation through synthetic benchmarks and field applications—performed with identical iteration counts and inversion parameters for fair comparison—demonstrates consistent superiority in scenarios involving abrupt velocity transitions (permafrost boundaries, water tables and bedrock interfaces). Synthetic tests show that although conventional tomography fails to recover plausible models for interfaces with significant topography, the hybrid approach accurately resolves both interface geometries and velocity distributions. Field applications confirm these advantages. In a hydrological survey, the method delineated horizontal water tables at 5.2 ± 1.0 m depth versus smooth, non‐physical solutions from conventional tomography. In Alpine permafrost zones, it resolved extreme lateral velocity contrasts (such as 2.1–5.3 km/s at 20 m depth) while maintaining inversion stability. This work establishes a practical and robust workflow for generating geologically constrained starting models directly from refraction data, significantly advancing the resolution of sharp interfaces in near‐surface seismic tomography for both academic and industrial applications.