Paper submitted to IMAGE 2026, by Akshika Rohatgi, Andrey Bakulin, Sergey Fomel (University of Texas at Austin), Hassan Odhwani (TGS).
Abstract
Land seismic acquisition acknowledges offset-dependent noise, yet acquisition geometries are typically designed with uniform sampling across offsets, with limited quantitative feedback on how prestack data quality varies in practice. Here, we use data-driven metrics to evaluate offset- and frequency-dependent prestack data quality and to compare two 3D surveys acquired over the same subsurface. Using offset-dependent signal-to-noise ratio (SNR) and frequency-dependent phase stability, we apply the same diagnostics consistently to raw and processed data, allowing data quality to be tracked objectively through the processing sequence. The results show that systematic differences persist across offset ranges and frequency bands, demonstrating that prestack data quality can be quantified directly from the data at any stage, enabling unbiased comparison of acquisitions and realistic assessment of processing outcomes.
Read the full article here.