Full Waveform Inversion of Shallow Seismic Data for Near-Surface Velocity Model Building in the Niger Delta

This study applies full waveform inversion to shallow seismic data for constructing high-resolution near-surface velocity models in the Niger Delta, addressing a fundamental challenge in seismic processing that limits the accuracy of depth imaging for petroleum exploration. Near-surface velocity variations caused by the Niger Delta's shallow gas-charged sands, shallow water channels, and highly heterogeneous deltaic sediments create static anomalies and wavefield distortions that degrade deep target imaging. Conventional refraction statics methods inadequately characterise these near-surface complexities at the resolution needed for modern migration algorithms. This study applies acoustic and elastic FWI to 2D seismic datasets from onshore Niger Delta test lines using academic-grade FWI code. Data preprocessing includes noise attenuation, wavelet estimation, and frequency-domain transformation. FWI is applied in a multi-scale frequency hierarchy from 5 to 30 hertz, starting from a smooth initial model derived from first-break tomography. Convergence is assessed using data misfit reduction and model update stability criteria. The FWI velocity model is compared with conventional near-surface models for migration performance assessment. Findings reveal that FWI-derived velocity models produce significantly sharper imaging of the shallow channel systems (15 to 80 metres depth) and reduce residual static anomalies by 35 percent compared to first-break tomography models. Imaging of deeper Agbada Formation targets improves measurably in areas with complex near-surface conditions. The study contributes an optimised FWI workflow for the Niger Delta near-surface environment and recommends implementation in production seismic processing workflows.

Keywords: full waveform inversion, near-surface velocity, Niger Delta, seismic processing, shallow model building.