High-precision and high-efficiency first-arrival slope tomography via eikonal solvers and the adjoint-state method

Abstract First-arrival slope tomography (FAST) introduces first-arrival slopes, corresponding to the horizontal components of slowness vectors, at receiver and source positions supplement traveltime for better guiding ray propagation in media until best match is achieved with observed data. FAST can recover velocity model higher resolution precision than (FATT) but computationally intensive. In this context, we propose a high-precision high-efficiency approach, referred as HFAST. HFAST redefines one slopes using reciprocity principle simultaneously employs ensure high-quality building. On other hand, extracts calculated data derives gradient misfit function from solutions relatively limited forward inverse problems, resulting low computational cost. The cost proportional minimum between receivers sources, whereas scaled sum sources. Numerical experiments involving checkerboard SEAM II Foothill models demonstrate that achieve inversion FATT, especially recovery small-scale anomalies presence reversal. Moreover, more efficient suitable managing large sets. Therefore, be regarded valuable current first-arrival-based building methods has potential applied static corrections, prestack depth migration, waveform future.