Jason R. McKenna

Tele-infrasonic studies of hard-rock mining explosions

The Lac-du-Bonnet infrasound station, IS-10, and the Minnesota iron mines 390 km to the southeast are ideally located to assess the accuracy of atmospheric profiles needed for infrasound modeling. Infrasonic data from 2003 associated with explosions at the iron mine were analyzed for effects of explosion size and atmospheric conditions on observations with well-constrained ground truth. Noise was the determining factor for observation; high noise conditions sometimes prevented unequivocal identification of infrasound arrivals. Observed arrivals had frequencies of 0.5 to 5 Hz, with a dominant frequency of 2 Hz, and generally had durations on the order of 10 s or less. There was no correlation between explosive amount and observability. Tele-infrasonic propagation distances (greater than 250 km) produce thermospheric ray paths. Modeling is based upon MSIS/HWM (Mass Spectrometer Incoherent Scatter/Horizontal Wind Model) and NRL-G2S (Naval Research Laboratory Ground to Space) datasets. The NRL-G2S dataset provided more accurate travel time predictions that the MSIS/HWM dataset. PE modeling for the NRL-G2S dataset indicates energy loss at higher frequencies (around 4 Hz). Additionally, applying the Sutherland/Bass model through the NRL-G2S realization of the atmosphere in InfraMAP results in predicted amplitudes too small to be observed.

Phase and amplitude inversion of crosswell radar data

Phase and amplitude inversion of crosswell radar data estimates the logarithm of complex slowness for a 2.5D heterogeneous model. The inversion is formulated in the frequency domain using the vector Helmholtz equation. The objective function is minimized using a back-propagation method that is suitable for a 2.5D model and that accounts for the near-, intermediate-, and far-field regions of the antennas. The inversion is tested with crosswell radar data collected in a laboratory tank. The model anomalies are consistent with the known heterogeneity in the tank; the models relative dielectric permittivity, which is calculated from the real part of the estimated complex slowness, is consistent with independent laboratory measurements. The methodologies developed for this inversion can be adapted readily to inversions of seismic data (e.g., crosswell seismic and vertical seismic profiling data).

Modeling and analysis of the response of a triaxial, frequency-domain electromagnetic induction sensor to a buried linear conductor

This paper presents analytical modeling results for a triaxial frequency-domain electromagnetic-induction (EMI) sensor over a homogeneous earth containing a long linear conductor. Although the conductor studied is intended to represent an underground wire or pipe, it can represent any subsurface, linear geologic structure that can channel current. Treating the sensor transmitter as a vertical magnetic dipole, the model combines the well-known solution for the magnetic field arising from the interaction with the earth with the solution for the induced magnetic field from the excited subsurface conductor. Expressions for the three components of the magnetic field at an arbitrary point above the earth are presented. Two types of coupled, moving transmitter-receiver configurations (coaxial and coplanar) wereconsidered, and the model is sufficiently flexible to allow for many other sensor variations to be studied. Characteristics of the sensor signals were explored through several parametric modeling studies t...

Topographic effects on infrasound propagation

Infrasound data were collected using portable arrays in a region of variable terrain elevation to quantify the effects of topography on observed signal amplitude and waveform features at distances less than 25 km from partially contained explosive sources during the Frozen Rock Experiment (FRE) in 2006. Observed infrasound signals varied in amplitude and waveform complexity, indicating propagation effects that are due in part to repeated local maxima and minima in the topography on the scale of the dominant wavelengths of the observed data. Numerical simulations using an empirically derived pressure source function combining published FRE accelerometer data and historical data from Project ESSEX, a time-domain parabolic equation model that accounted for local terrain elevation through terrain-masking, and local meteorological atmospheric profiles were able to explain some but not all of the observed signal features. Specifically, the simulations matched the timing of the observed infrasound signals but underestimated the waveform amplitude observed behind terrain features, suggesting complex scattering and absorption of energy associated with variable topography influences infrasonic energy more than previously observed.

Near-surface Void Identification Using MASW And Refraction Tomography Techniques

Subsurface voids may manifest themselves as natural or anthropogenic dissolution features, illegal cross-border tunnels, or abandoned mines. Detection of these voids using geophysical methods has often proven difficult due to multiple factors including depth-to-diameter ratio, lack of resolution, non-uniqueness, etc. Experiments were conducted at a test site with a known subsurface void to determine the capability of multiple near-surface seismic methods as applied to void detection. In this study, refraction tomography and multichannel analysis of surface wave methods successfully identified a man-made void approximately three meters deep. Results of these experiments correlated well with the expected results, exhibiting decreased velocity, the absence of seismic ray coverage, and a high shear-wave velocity halo above the void.

Frequency-domain Green’s functions for radar waves in heterogeneous 2.5D media

Green’s functions for radar waves propagating in heterogeneous 2.5D media might be calculated in the frequency domain using a hybrid method. The model is defined in the Cartesian coordinate system, and its electromagnetic properties might vary in the x - and z -directions, but not in the y -direction. Wave propagation in the x - and z -directions is simulated with the finite-difference method, and wave propagation in the y -direction is simulated with an analytic function. The absorbing boundaries on the finite-difference grid are perfectly matched layers that have been modified to make them compatible with the hybrid method. The accuracy of these numerical Green’s functions is assessed by comparing them with independently calculated Green’s functions. For a homogeneous model, the magnitude errors range from −4.16% through 0.44%, and the phase errors range from −0.06% through 4.86%. For a layered model, the magnitude errors range from −2.60% through 2.06%, and the phase errors range from −0.49% through 2....

Automatic Detection of a Subsurface Wire Using an Electromagnetic Gradiometer

A model-based correlation detection scheme is presented with the aim of detecting and localizing subsurface tunnel infrastructure in an automated fashion. Our goal is to develop a comprehensive detection technology that can be fielded and successfully used by nonexperts, while simultaneously being sufficiently robust as to be effective. Our correlation detection algorithm relies on a library of model signals that are generated using an analytical model of a thin subsurface wire in a homogeneous half-space. The wire is illuminated using an active transmitter source (12, 20, or 200 kHz), and its response is sensed using a man-portable electromagnetic gradiometer (EMG) system. The performance of the detector is assessed using synthetic data and receiver operating characteristic (ROC) analysis as well as experimental data collected during a field test. Preliminary ROC results indicate that at sufficient signal-to-noise ratio, the detector can achieve detection probabilities greater than 0.9 with corresponding false alarm rates of less than one every 1000 m. Results from the field tests revealed that the responses from the EMG can be used to detect and localize (to within 0.5 m in the horizontal) a wire target down to a depth of at least 7 m. We believe the EMG system and correlation detector combine to form a promising technology for detecting tunnel infrastructure that can be used by experts and, more importantly, nonexperts as well.

Response of an Electromagnetic Gradiometer to a Subsurface Wire

The response of an electromagnetic gradiometer (EMG) system to a subsurface wire is analyzed in terms of experimental and analytical modeling results. Our objective is to explore characteristics of the response and assess the fidelity of our model. The EMG system consists of a static transmitter and a man-portable sensor, which uses a pair of receivers that yield a gradiometric measurement. Experimental results were collected over a range of wire depths from 3.4 to 8.5 m. A number of different transmitter positions were explored, and the tests studied were conducted at 200 kHz. Modeling results were consistent with the experimental results and supported a number of key findings. Results are presented showing that, in order to maximize the strength of the wire response, the transmitter should be positioned approximately 5 m off the wire axis. Furthermore, in order to avoid unwanted transmitter influence on the response, the EMG should be at least 30 m from the transmitter. Using the experimental and modeling results, we found a linear relationship between the width of the magnitude response peak and the wire depth. Based on our experimental results, the EMG is able to yield a discernible target response at a depth of at least 7 m. Lastly, an example of how the model can be used to optimize survey planning is presented. This paper illustrates how an EMG can be used to locate underground wires with applications ranging from underground utility mapping to the detection of shallow subsurface tunnels.

Frequency domain surface nuclear magnetic resonance forward modeling on an adaptive octree mesh

Nuclear magnetic resonance is unique among geophysical methods in its ability to directly detect hydrogen in liquids. In near surface applications wire loops may be deployed on the surface, both as transmitters and receivers, and used to investigate groundwater in the top 100 or so meters. There are numerous applications for the method, and fast, robust methods to forward model the data are needed. Most formulations assume coincident transmitter and receiver loops, 1D layering of water, and are formulated either in the time domain or solve for the initial amplitude of the signal only. We present an alternative scheme that makes use of efficient frequency-domain calculations that is generalized for 3D water distributions. Separate transmitter and receiver loops are supported. An adaptive octree mesh is used that splits the solution domain into cells such that a specified tolerance of error is not exceeded.

Partially saturated soil causing significant variability in near surface seismic signals

The behaviour of dry, moist, and saturated soils has been studied for over a century without adequately investigating the behaviour associated with transient saturation in the near surface, i.e. the upper 1 m of overburden, including the effects of rapid meteorological changes, dynamic fluid flow, and variability of saturation on shallow seismic sensors. This paper presents observational data wherein the geophysical instrumentation response was significantly influenced by near-surface post-precipitation saturation and additional laboratory experimentation on the effects of saturation on shear wave velocity. The lack of partially-saturated data is primarily because transient meteorological events have not been critically important to the types of long-term deployments performed in the past, where sensors were situated in hard-rock, collecting data under idealized conditions, as opposed to sedimentary settings. Shorter-duration deployments and smaller system architectures, e.g. persistent monitoring, now necessitate detailed a priori knowledge of meteorological impacts to system design and performance. The purpose of this persistent monitoring geophysical instrumentation is to continually monitor the near surface and relate small perturbations to a specific source type(s) and distance(s) from the receiver. As such, the received signal is compared to known sources within a predetermined geological/ meteorological condition. Presented herein is the calibration signal generated by a 3.63-kg (8-lb) sledgehammer prior to and post 36 hours of steady precipitation. The resulting subsurface seismic velocity time-histories show a significant increase in signal amplitude, change in frequency content and no change in duration. Thus, the amplification effects of near-surface moisture variability combined with dynamic pore fluid could be interpreted as false positives of a specific source signature and/or instrument failure.