Enhancing 3D Broadband Land Seismic Data with Smart Super Groups for Processing and FWI

Land seismic data in Saudi Arabia is always challenging thus hindering any technique relying on pre-stack information such as for instance full-waveform inversion (FWI). While many sophisticated algorithms for noise removal exist, the greatest power remains with final stack for imaging. We show how power of stacking can be harvested to enhance raw broadband 3D land data for various pre-stack applications using smart super grouping. We design large super groups both in inline and crossline directions and devise smart stacking procedures adapted to variable spatial sampling in each direction. We demonstrate on synthetic data that at low frequencies (< 8Hz) reflections and diving waves are identical on point-source and super-group data. On real 3D broadband data, super grouping greatly enhances signal-to-noise ratio and produces data ready for FWI starting from lowest sweep frequency. Advanced stacking with complex data-driven weights can also enhance high-frequency reflection data for processing and imaging. We illustrate both of these points by applying smart shot-based super groups on a 3D seismic dataset from Saudi Arabia acquired with broadband sweep 2–90 Hz.