In this First Break article from January 2025, TGS experts Terence Krishnasamy*, James Beck and Cristina Reta-Tang present an effective model-building workflow that incorporates FWI for both diving waves and reflections, utilising different cost functions.

 

Introduction

Near-surface characterisation plays a crucial role in accurately imaging deeper targets in onshore exploration seismic data. However, conventional model building techniques, such as reflection tomography, have limitations when it comes to updating near-surface velocities due to
insufficient offset coverage at shallow depths. To overcome this, alternative methods like diving-wave or first arrival tomography have been widely used, offering updates to the shallow velocity model, albeit often lacking the necessary resolution.

In contrast, full waveform inversion (FWI) is a robust algorithm used to derive velocity models with high resolution and fidelity. FWI, based on direct solutions of the two-way wave equation, has been intensively developed in recent years because it provides a superior way to build high-resolution velocity fields, especially under complex geological settings. The inversion is performed by iteratively updating the model parameters by reducing the data differences between the observed and synthetic data. In recent years, it has been widely adopted in the industry, with successful examples of its application to marine data, especially for surveys acquired with rich azimuth long offsets and recorded with low frequency. However, onshore examples are not as abundant (Mei & Tong, 2015; Lemaistre et al., 2018; Tang et al., 2021; Masclet et al., 2021; Sheng et al., 2022, Krishnasamy et al., 2023).

There are several challenges for FWI in onshore applications. Onshore seismic data are acquired on non-flat datum (topography). As such, wavefield propagation from this non-flat surface must be addressed to achieve accurate results and be computationally efficient. Additionally, the low signal-to-noise ratio (SNR) of onshore data poses a challenge, mainly due to near-surface heterogeneity which results in strong near-surface scattering. Furthermore, recent successful applications of FWI to onshore data have shown that it remains a challenge to take full advantage of the recorded energy below 5 Hz due to the highly variable SNR, even with nonlinear sweeps starting as low as 2 Hz (Durussel et al., 2022). Lastly, the weathering layer (i.e., the layer at or near the surface, mostly made up of unconsolidated and heterogeneous material) results in strong elastic effects such as surface waves and converted waves, which are not accounted for during acoustic simulation. While one could utilise an elastic FWI workflow to address the elastic challenge, in practice it is still resource intensive, especially with very low shear velocities present in the model.

Figure 1 Conceptual representation of the FWI velocity model building workflow.

We have developed an effective model building workflow that incorporates FWI for both diving waves and reflections, utilising different cost functions. Figure 1 shows a conceptual representation of the workflow, which has been applied to many onshore surveys from a variety of geological settings. Thus far, the workflow produces geologically plausible and consistent models for data acquired with legacy/conventional onshore acquisition setups.

Read the full article HERE.