Time-lapse Ground-penetrating Radar Data Research Articles

Advanced Time-Lapse Ground-Penetrating Radar Data Processing for Quantitatively Monitoring of Small-Scale Fluid Infiltration

Advanced Time-Lapse Ground-Penetrating Radar Data Processing for Quantitatively Monitoring of Small-Scale Fluid Infiltration

Continuous Automatic Estimation of Volumetric Water Content Profile During Infiltration Using Sparse Multi-Offset GPR Data

Ground-penetrating radar (GPR) is a non-destructive and non-invasive geophysical survey method that has been used to characterize soil volumetric water content (VWC) dynamics. An array antenna GPR system was used to collect nearly seamless, time-lapse multi-offset GPR data during an in-situ infiltration test on sand dunes with limited traces. Because the data volume was significant, an approach was utilized to automatically determine electromagnetic wave velocities from sparse common midpoint (CMP) data using standard velocity analysis, such as semblance analysis. The objective of this study was to develop a methodology that allows one to automatically perform velocity analysis by interpolating sparse CMP data obtained with the array GPR system. The proposed method determined the optimal normal moveout velocity values and the removal range of the F-K zone pass filter that minimized errors between the original and interpolated CMP data using leave-one-out cross-validation (LOOCV). After interpolating the sparse CMP data with the F-K zone pass filter, semblance analysis was used to determine the time-lapse velocity structure of the soil profile during water infiltration. The velocity data were converted to VWC data based on the Topp equation, which relates the soil VWC to the soil dielectric constant. The proposed method was tested using CMP data obtained via numerical simulation and experiments. The VWC profile from the proposed approach matched well with the independently observed VWC profiles obtained from an invasive probe-type soil moisture sensor.

Fast Segmentation and Dynamic Monitoring of Time-Lapse 3D GPR Data Based on U-Net

As the amount of ground-penetrating radar (GPR) data increases significantly with the high demands of nondestructive detection methods under urban roads, a method suitable for time-lapse data dynamic monitoring should be developed to quickly identify targets on GPR profiles and compare time-lapse datasets. This study conducted a field experiment aiming to monitor one backfill pit using three-dimensional GPR (3D GPR), and the time-lapse data collected over four months were used to train U-Net, a fast neural network based on convolutional neural networks (CNNs). Consequently, a trained network model that could effectively segment the backfill pit from inline profiles was obtained, whose Intersection over Union (IoU) was 0.83 on the test dataset. Moreover, segmentation masks were compared, demonstrating that a change in the southwest side of the backfill pit may exist. The results demonstrate the potential of machine learning algorithms in time-lapse 3D GPR data segmentation and dynamic monitoring.

On the usage of diffractions in ground-penetrating radar reflection data: Implications for time-lapse gas migration monitoring

Time-lapse ground-penetrating radar (GPR) measurements are used to image/monitor, for example, water infiltration, water table changes, and fluid/gas flow patterns. Although crosshole GPR is often preferred over surface-reflection GPR in such studies, its application is limited by the selected borehole geometry, which may be difficult to define in an optimal way, especially in experiments in which flow pathways are difficult to predict. Surface-reflection GPR data sets are generally faster to collect over relatively large areas and are therefore more efficient for covering the volume when a fast-moving tracer (e.g., gas) may infiltrate a heterogeneous subsurface medium. We have used a diffraction imaging approach on time-lapse surface-reflection GPR data to detect changes in radar wave velocity associated with gas (CO2) injected into a heterogeneous chalk succession. We initially test and evaluate the diffraction imaging approach on synthetic GPR data. Afterward, we apply the methodology to time-lapse GPR field data, and we evaluate the robustness of using information from diffractions in light of the obtained data quality. The synthetic tests indicate that diffractions provide essential information for delimiting the area affected by gas in the heterogeneous chalk section studied. Our field experiment shows that using the diffraction information alone allows for detection of gas-affected zones and, therefore, potential flow characteristics of gas movement. We infer that the CO2 flow patterns in our study most likely are determined by small-scale fractures rather than the porosity/permeability of the rock matrix. Moreover, the approach used may serve as an initial study for future more targeted experiments or for further detail-retrieving full-waveform inversion.

Reflection tomography of time-lapse GPR data for studying dynamic unsaturated flow phenomena

Abstract. Ground-penetrating radar (GPR) reflection tomography algorithms allow non-invasive monitoring of water content changes resulting from flow in the vadose zone. The approach requires multi-offset GPR data that are traditionally slow to collect. We automate GPR data collection to reduce the survey time significantly, thereby making this approach to hydrologic monitoring feasible. The method was evaluated using numerical simulations and laboratory experiments that suggest reflection tomography can provide water content estimates to within 5 % vol vol−1–10 % vol vol−1 for the synthetic studies, whereas the empirical estimates were typically within 5 %–15 % of measurements from in situ probes. Both studies show larger observed errors in water content near the periphery of the wetting front, beyond which additional reflectors were not present to provide data coverage. Overall, coupling automated GPR data collection with reflection tomography provides a new method for informing models of subsurface hydrologic processes and a new method for determining transient 2-D soil moisture distributions.

Stochastic inversion for soil hydraulic parameters in the presence of model error: An example involving ground-penetrating radar monitoring of infiltration

Proxy forward solvers are commonly used in Bayesian solutions to inverse problems in hydrology and geophysics in order to make sampling of the posterior distribution, for example using Markov-chain-Monte-Carlo (MCMC) methods, computationally tractable. However, use of these solvers introduces model error into the problem, which can lead to strongly biased and overconfident parameter estimates if left uncorrected. Focusing on the specific example of estimating unsaturated hydraulic parameters in a layered soil from time-lapse ground-penetrating radar data acquired during a synthetic infiltration experiment, we show how principal component analysis, conducted on a suite of stochastic model-error realizations, can for some problems be used to build a sparse orthogonal basis for the model error arising from known forward solver approximations and/or simplifications. Projection of the residual onto this basis during MCMC permits identification and removal of the model error before calculation of the likelihood. Our results indicate that, when combined with an informal likelihood metric based on the expected behaviour of the ℓ2-norm of the residual, this methodology can yield posterior parameter estimates exhibiting a marked reduction in bias and overconfidence when compared to those obtained with no model-error correction, at reasonable computational cost.

Ground penetrating radar for fracture mapping in underground hazardous waste disposal sites: A case study from an underground research tunnel, South Korea

Secure disposal or storage of nuclear waste within stable geologic environments hinges on the effectiveness of artificial and natural radiation barriers. Fractures in the bedrock are viewed as the most likely passage for the transport of radioactive waste away from a disposal site. We utilize ground penetrating radar (GPR) to map fractures in the tunnel walls of an underground research tunnel at the Korea Atomic Energy Research Institute (KAERI). GPR experiments within the KAERI Underground Research Tunnel (KURT) were carried out by using 200MHz, 500MHz, and 1000MHz antennas. By using the high-frequency antennas, we were able to identify small-scale fractures, which were previously unidentified during the tunnel excavation process. Then, through 3-D visualization of the grid survey data, we reconstructed the spatial distribution and interconnectivity of the multi-scale fractures within the wall. We found that a multi-frequency GPR approach provided more details of the complex fracture network, including deep structures. Furthermore, temporal changes in reflection polarity between the GPR surveys enabled us to infer the hydraulic characteristics of the discrete fracture network developed behind the surveyed wall. We hypothesized that the fractures exhibiting polarity change may be due to a combination of air-filled and mineralogical boundaries. Simulated GPR scans for the considered case were consistent with the observed GPR data. If our assumption is correct, the groundwater flow into these near-surface fractures may form the water-filled fractures along the existing air-filled ones and hence cause the changes in reflection polarity over the given time interval (i.e., 7days). Our results show that the GPR survey is an efficient tool to determine fractures at various scales. Time-lapse GPR data may be essential to characterize the hydraulic behavior of discrete fracture networks in underground disposal facilities.

Attribute-based analysis of time-lapse ground-penetrating radar data

Analysis of time-lapse ground-penetrating radar (GPR) data can provide information regarding subsurface hydrological processes, such as preferential flow. However, the analysis of time-lapse data is often limited by data quality; for example, for noisy input data, the interpretation of difference images is often difficult. Motivated by modern image-processing tools, we have developed two robust GPR attributes, which allow us to distinguish amplitude (contrast similarity) and time-shift (structural similarity) variations related to differences between individual time-lapse GPR data sets. We tested and evaluated our attributes using synthetic data of different complexity. Afterward, we applied them to a field data example, in which subsurface flow was induced by an artificial rainfall event. For all examples, we identified our structural similarity attribute to be a robust measure for highlighting time-lapse changes also in data with low signal-to-noise ratios. We determined that our new attribute-based workflow is a promising tool to analyze time-lapse GPR data, especially for imaging subsurface hydrological processes.

Coupled hydrogeophysical inversion of time-lapse surface GPR data to estimate hydraulic properties of a layered subsurface

[1] A major challenge in vadose zone hydrology is to obtain accurate information on the temporal changes of the vertical soil water distribution and its feedback with the atmosphere and groundwater. A state of the art coupled hydrogeophysical inversion scheme is applied to evaluate soil hydraulic properties of a synthetic model and a field soil in southern Ontario based on time-lapse monitoring of soil dynamics with surface ground-penetrating radar (GPR). Film flow was included in the hydrological model to account for noncapillary water flow in a sandy medium during dry conditions. The synthetic study illustrated that GPR data contain sufficient information to accurately constrain soil hydraulic parameters within a coupled inversion framework and led to an accurate estimation of the soil hydraulic properties. When film flow was not accounted for within the inversion, an equally good fit could still be achieved. In this case, errors introduced by neglecting film flow were compensated by different hydraulic parameters. For the field data, the coupled inversion reduced the overall misfit compared to an uncalibrated model using hydraulic parameters obtained from laboratory data. Although the data fit improved significantly for water content in the deeper soil layers, accounting for film flow in the uppermost subsurface layer did not lead to a better fit of the GPR data. Further research is needed to describe the processes controlling water content in the dry range, in particular coupled heat and vapor transport. This study illustrates the suitability of surface GPR measurements combined with coupled inversion for near-surface characterization of soil hydraulic parameters.

Examining the information content of time-lapse crosshole GPR data collected under different infiltration conditions to estimate unsaturated soil hydraulic properties

Time-lapse geophysical data acquired during transient hydrological experiments are being increasingly employed to estimate subsurface hydraulic properties at the field scale. In particular, crosshole ground-penetrating radar (GPR) data, collected while water infiltrates into the subsurface either by natural or artificial means, have been demonstrated in a number of studies to contain valuable information concerning the hydraulic properties of the unsaturated zone. Previous work in this domain has considered a variety of infiltration conditions and different amounts of time-lapse GPR data in the estimation procedure. However, the particular benefits and drawbacks of these different strategies as well as the impact of a variety of key and common assumptions remain unclear. Using a Bayesian Markov-chain-Monte-Carlo stochastic inversion methodology, we examine in this paper the information content of time-lapse zero-offset-profile (ZOP) GPR traveltime data, collected under three different infiltration conditions, for the estimation of van Genuchten–Mualem (VGM) parameters in a layered subsurface medium. Specifically, we systematically analyze synthetic and field GPR data acquired under natural loading and two rates of forced infiltration, and we consider the value of incorporating different amounts of time-lapse measurements into the estimation procedure. Our results confirm that, for all infiltration scenarios considered, the ZOP GPR traveltime data contain important information about subsurface hydraulic properties as a function of depth, with forced infiltration offering the greatest potential for VGM parameter refinement because of the higher stressing of the hydrological system. Considering greater amounts of time-lapse data in the inversion procedure is also found to help refine VGM parameter estimates. Quite importantly, however, inconsistencies observed in the field results point to the strong possibility that posterior uncertainties are being influenced by model structural errors, which in turn underlines the fundamental importance of a systematic analysis of such errors in future related studies.

Efficient Electromagnetic Imaging of an Artificial Infiltration Process in the Vadose Zone Using Cross-Borehole Radar

Cross-borehole ground-penetrating radar (GPR) has been widely used to characterize the shallow subsurface and to monitor hydrogeologic processes. To investigate an infiltration process in the vadose zone, an artificial groundwater infiltration test was conducted in Nagaoka, Japan. Time-lapse cross-borehole GPR data were collected using zero-offset profiling (ZOP) mode. The infiltration process was observed as a variation of GPR traveltimes, which can be transformed into a dielectric constant, and further converted to volumetric water content. A standard ZOP analysis, for which all first arrivals are assumed to be direct waves, results in an underestimation of the dielectric constant because of the existence of critically refracted waves. This letter presents an efficient algorithm using the maximum first-cycle amplitude to approximately determine the traveltime of direct arrival, deriving a dielectric constant model more accurately than the standard ZOP analysis from ZOP data. Tests on synthetic and real field data show that the proposed approach is effective in building accurate water content profile without iterative calculations as in the standard ZOP analysis.

Time-lapse borehole radar monitoring of an infiltration experiment in the vadose zone

Ground-penetrating radar (GPR) is an effective tool for imaging the spatial distribution of water content. An artificial groundwater recharge test was conducted in Nagaoka City in Japan, and time-lapse crosshole GPR data were collected to monitor the infiltration process in the vadose zone. Since electromagnetic wave velocities in the vadose zone are largely controlled by variations in water content, an increase in traveltime is interpreted as an increase in saturation. In the test zone, the infiltrated water penetrated downward with an average velocity of about 2.7 m/h. A finite-difference time-domain method using two-dimensional cylindrical coordinates is applied to simulate radargrams associated with the advancing wetting front and to quantify the effects of critical refraction. Standard zero-offset profiling for which all first-arrivals are assumed to be direct waves results in an underestimation of water content in the transition zone above the wetting front. As a result, correct velocity analysis requires identification of first-arriving critically refracted waves from the traveltime profile to accurately determine a water content profile.

Closed loop GPR data inversion for soil hydraulic and electric property determination

We propose an integrated electromagnetic‐hydrodynamic inverse modeling approach for identifying field‐scale unsaturated soil hydraulic properties and electric profiles from off‐ground time‐lapse ground‐penetrating radar data. Hydrodynamic modeling based on the one‐dimensional Richards' equation with homogeneous soil hydraulic parameters is used to physically constrain the full‐wave radar electromagnetic inverse problem in a closed loop. Significant effects can be observed of water dynamics on the time‐lapse radar data and numerical studies demonstrate the uniqueness of the inverse solution. The results suggest that the method is promising for non‐invasive characterization of the shallow subsurface and monitoring of its dynamics.