Imaging Near Surface Scatterers by Scattered Surface Waves - Revamping an Inverse Scattering Code

Modelling 3D elastodynamic wave scattering due to density and Lamé parameter contrasts of near‐surface scatterers

“…The modelling method that we propose here is meant for a full‐waveform inversion scheme that will comparatively be more efficient in terms of the computational cost. For example, using the full frequency band in an FD‐based inversion scheme is cost‐prohibitive, while it is cheaper with our method, as demonstrated in Harmankaya and Kaslilar (2019). As for this study, we did our modelling on a workstation with Intel Xeon 2.20 GHz processor.…”

“…For near‐surface characterizations, fluid saturation is of interest in environmental and hydrologic studies (Konstantaki et al ., 2013; Solatzzi et al ., 2021), where the P wave velocity, sensitive to λ , is the one of the main indicators between unsaturated and saturated media (Socco and Strobia, 2004). Another point is that the density is the most difficult parameter to resolve (Butzer, 2015), although for our method we were able to successfully resolve the positive and negative density contrasting scatterers with a density only imaging scheme (Harmankaya and Kaslilar, 2019). One approach for multi‐parameter inversion is to consider coupling of these parameters, such as P‐ and S‐wave velocities or their impedances.…”

“…Using the above‐mentioned parameters, we calculate and store Green tensor elements for subsequent modelling efforts. As the calculation of the integrals in Equations (19) to (21) is computationally expensive, parallel programming is considered to improve the efficiency of both ‘on the fly’ and table calculations (Harmankaya and Kaslilar, 2019). With the benefit of parallel calculations, we are able to calculate incident and scattered wavefields rather efficiently, especially with the precalculated Green tensor tables.…”

“…(2020). To accelerate the calculations further, we make use of parallel programming techniques for the calculation of the wavefields (Harmankaya and Kaslilar, 2019). To examine the effects of each parameter contrast on the scattered wavefield as preparation for the forthcoming inversion study, we use a homogeneous half‐space model with an embedded near‐surface scatterer with several cases of parameter contrasts.…”

“…For scatterer imaging through inversion, even with only a single source location and few receiver lines, either perpendicular or parallel to each other, the method can provide reasonable results, even though more shot gathers would improve the parameter estimations. In Harmankaya and Kaslilar (2019), it was demonstrated that the previous (density contrasts only) version of the method could even be used with only a single shot gather for imaging the scatterer. Traditionally, due to its high computational cost, 3D full waveform inversions for elastic or viscoelastic cases are rare among subsurface imaging studies (Teodor et al ., 2021).…”