Tan Qin

Full-waveform inversion (FWI) of multi-offset surface ground-penetrating radar (GPR) data has the potential to provide high-resolution subsurface electromagnetic models. Conventional GPR-FWI uses the l₂ objective function, which is more sensitive to near-offset data and early arrivals. In this study, we propose a random multi-objective waveform inversion (R-MOWI) to balance the contributions of near- and far-offset data, as well as early and late arrivals. R-MOWI adopts a multi-objective framework that integrates the l₂, envelope, normalized, and auto-gain control (AGC) objective functions. R-MOWI randomly selects an objective function for each radargram at each iteration. Our 2-D synthetic examples demonstrate that R-MOWI outperforms the four single-objective FWIs in reconstructing permittivity and conductivity models by leveraging multiple objective functions to constrain convergence. R-MOWI mitigates overestimation by offering greater flexibility in the search space. Furthermore, it exhibits robust performance and facilitates uncertainty estimation of the inversion results. In summary, this study highlights the effectiveness of R-MOWI in processing multi-offset surface GPR data.

Understanding ice thickness distribution and estimating glacier volume are crucial for gaining insights into the dynamics of temperate glaciers. However, comprehensive measurements of these parameters remain limited for the Baishui River Glacier No. 1 (BRG1) on the Yulong Snow Mountain, Lijiang, China. To address this issue, we performed a ground-penetrating radar (GPR) survey on BRG1, employing low-frequency antennas (with center operating frequencies of 25 and 50 MHz, respectively) coupled with a high-voltage transmitter and numerous stacks on the recorded data to enhance the signal-to-noise ratio (SNR). On the glacier surface, we collected common-offset profiles and processed the GPR data using a specified workflow to further improve the SNR. We proposed a fast coherent enhancement (FCE) technique to reduce clutter scattering and enhance event continuity. Compared with the traditional method, FCE enhanced the processing speed of the 25 and 50 MHz GPR data by approximately 120 and 50 times, respectively. After data processing, we identified the ice–bedrock interface within the 25 MHz GPR profiles, with ice thicknesses ranging from 40 to 120 m. The survey area was 0.2758 km², and we calculated the ice volume to be 0.0232 km³ with an average ice thickness of 83 m. By performing GPR numerical simulations and comparing BRG1 with another temperate glacier, we observed that water inclusions substantially contribute to clutter scattering and signal loss in our GPR data. In brief, this study confirms the effectiveness of low-frequency GPR in investigating complex temperate glaciers.

Ground-penetrating radar (GPR) full-waveform inversion (FWI) can determine high-resolution electromagnetic (EM) properties of the subsurface and has gained increasing attention in near-surface geophysics. However, the application of GPR-FWI to field surface data is limited due to the high computational costs. Here we apply for one of the first times a 2-D time-domain gradient-based FWI to synthetic and field multioffset surface GPR dataset. We use subset FWI (SFWI) to reduce computational costs by implementing the simulation on a model subset rather than the whole model. Thus we obtain theoretical speedup and memory saving factor equal to the size ratio of the model to its subset. The properties of the model subset depend on the illumination of the chosen acquisition geometry, based on which we provide rules of thumb for selecting the model subset. Our study reveals that SFWI can be further improved by parallelizations, where source parallelization allows for higher efficiency than model domain parallelization. Both 2-D synthetic data and field data validate that SFWI provides results comparable to FWI but without the redundant computations present in FWI. In a field example with surface geometry, SFWI provides a speedup of more than six times for a 45-m survey line. In another field example with cross-hole geometry, SFWI achieves more than a three times acceleration for a 21-m multiborehole plane. Our study proves that SFWI has the potential for high-performance computation to solve large-scale GPR survey problems.

SUMMARY: Full-waveform inversion (FWI) of ground-penetrating radar (GPR) data has received particular attention in the past decade because it can provide high-resolution subsurface models of dielectric permittivity and electrical conductivity. In most GPR FWIs, these two parameters are regarded as frequency independent, which may lead to false estimates if they strongly depend on frequency, such as in shallow weathered zones. In this study, we develop frequency-dependent GPR FWI to solve this problem. Using the τ-method introduced in the research of viscoelastic waves, we define the permittivity attenuation parameter to quantify the attenuation resulting from the complex permittivity and to modify time-domain Maxwell’s equations. The new equations are self-adjoint so that we can use the same forward engine to back-propagate the adjoint sources and easily derive model gradients in GPR FWI. Frequency dependence analysis shows that permittivity attenuation acts as a low-pass filter, distorting the waveform and decaying the amplitude of the electromagnetic waves. The 2-D synthetic examples illustrate that permittivity attenuation has low sensitivity to the surface multioffset GPR data but is necessary for a good reconstruction of permittivity and conductivity models in frequency-dependent GPR FWI. As a comparison, frequency-independent GPR FWI produces more model artefacts and hardly reconstructs conductivity models dominated by permittivity attenuation. The 2-D field example shows that both FWIs reveal a triangle permittivity anomaly which proves to be a refilled trench. However, frequency-dependent GPR FWI provides a better fit to the observed data and a more robust conductivity reconstruction in a high permittivity attenuation environment. Our GPR FWI results are consistent with previous GPR and shallow-seismic measurements. This research greatly expands the application of GPR FWI in more complicated media.

Both full-waveform inversion (FWI) of shallow-seismic and ground-penetrating radar (GPR) surface recordings have received particular attention in the past decade since they can reconstruct seismic and electromagnetic properties at high resolution. For consistent near-surface imaging, seismic and GPR wavefields can be combined by joint petrophysical inversion (JPI) using classical time-domain FWI. In conventional JPI of wavefields, both shallow-seismic and multi-offset GPR data contribute to reconstructing the same petrophysical parameters. In this paper, we show that seismic and GPR wavefields have different sensitivities to these parameters assuming the widely established petrophysical model combining the Gassmann fluid substitution model, the complex refractive index model and Archie’s equation. Based on this observation, we propose a new coupling strategy of petrophysical parameters which we call indirect JPI. In indirect JPI, seismic data is primarily used for porosity reconstruction, while GPR data is used only for saturation reconstruction. Unlike conventional JPI, we first update the seismic and GPR parameters using non-petrophysical parametrizations and then transform the most reliable estimates to petrophysical parameters. 2-D synthetic tests show that indirect JPI can provide more accurate and consistent results than conventional JPI. In addition, indirect JPI is more robust when uncertainties exist in petrophysical a priori knowledge. More importantly, indirect JPI has the flexibility to integrate different types of seismic and electromagnetic waves and acquisition geometries depending on the target of interest, resulting in the best solution. Indirect JPI has been proven to be a promising approach for multiparameter reconstructions (seven parameters in this study).

Shallow-seismic surface wave and ground penetrating radar (GPR) are employed in a wide range of engineering and geosciences applications. Full-waveform inversion (FWI) of either seismic or multi-offset GPR data are able to provide high-resolution subsurface characterization and have received particular attention in the past decade. Those two geophysical methods are involved in the increasing requirements of comprehensive site and material investigations. However, it is still challenging to provide an effective integration between seismic data and electromagnetic data. In this paper, we investigated the joint petrophysical inversion (JPI) of shallow-seismic and multi-offset GPR data for more consistent imaging of near surface. As a bridge between the seismic parameters (P-wave velocity, S-wave velocity, and density) and GPR parameters (relative dielectric permittivity and electric conductivity), the petrophysical relationships with the parameters namely porosity and saturation are employed to link two data sets. We first did a sensitivity analysis of the petrophysical parameters to the seismic and GPR parameters and then determined an efficient integration of using shallow-seismic FWI to update porosity and GPR FWI to update saturation, respectively. A comparison of several parameterisation combinations shows that the seismic velocity parameterisation in shallow-seismic FWI and a modified logarithm parameterisation in GPR FWI works well in reconstructing reliable S-wave velocity and relative dielectric permittivity models, respectively. With the help from the petrophysical links, we realized JPI by transforming those well recovered parameters to the petrophysical parameters and then to other seismic and GPR parameters. A synthetic test indicates that, compared with the individual petrophysical inversion and individual FWI, JPI outperforms in simultaneously reconstructing all seismic, GPR, and petrophysical parameters with higher resolution and improved details. It is proved that JPI would be a potential data integration approach for the shallow subsurface investigation.