Ground-penetrating radar, or GPR, is a means of exploring the shallow subsurface with electromagnetic waves (radar), usually in the 10 to 1000 MHz band. The two-way traveltimes of reflected radar waves give the depths where changes in electrical properties occur. Also called georadar, ground probing radar, and surface penetrating radar. See Figure G-9.
Introduction to ground penetrating radar
While the operation of ground penetrating radar is rather simple, complexity in the geophysical technique arises with qualitative comprehension of how the system functions. The GPR signal is explained by equations depicting the wave nature of EM fields and particle response to EM fields. The application of GPR techniques are based on electromagnetic (EM) theory and Maxwell’s equations. Short bursts of electromagnetic energy emitted by the transmitting antenna projecting an EM field into the ground. Molecular response to the projected EM field is defined by the material properties of the subsurface. Depending on the materials conductivity, permittivity or permeability the EM field will diffuse or propagate as a radio wave. Responses to the EM field that propagate a wave are measured as signal by the receiving antenna.
Achieving a desirable signal from the subsurface depends on the transmitting antennas’ frequency and conductivity of the subsurface material. Low conductivity materials produce a significant signal depth. While high conductivity environments hinder the penetration of the EM field. The velocity and attenuation of a propagated wave signal relies on the frequency of the transmitter. For high frequency broadcasts, the velocity and attenuation of signal is equivalent, preserving the signal shape without dispersion. Signal dispersion is observed in low frequency broadcasting where a linear diffusion of the signal occurs distorting the signal shape.
Imaging of the subsurface is a result of a planar wavefront encountering material boundaries whose geometry cause a wave reflection. Snell’s law describes the transmitted field’s direction change because of the medium. The contact between transmitted signal and boundaries produces a reflection or refraction according to Snell’s law and Fresnel coefficient. Fresnel’s coefficient quantifies amplitude variation of an EM field crossing a border separating two materials. Reflected signal is evaluated based on its amplitude versus time of emission, allowing visual definition of various boundaries in the subsurface.
Fundamental to a ground penetrating radar (GPR) survey is defining the problem and survey intent, this is crucial to a focused interpretation. The effectiveness of GPR comes into question when considering target depth. If points of interests lie below the ideal depth of radar penetration for a scenario, GPR is rendered inept. Influential to depth of penetration and spatial resolution is antenna-operating frequency. In order to increase depth of penetration spatial resolution is decreased. For scenarios concerned about target depth over vertical resolution, a lower operating frequency is preferred.
Commonly, application of GPR surveying involves reflection profiling. A system with fixed antenna geometry is moved along a survey line to record reflections versus instrument position. Defining parameters for a reflection survey involves selecting antenna orientation and separation as well as station spacing and survey line location. Antenna orientation is chosen so that the electric field emitted is polarized parallel to the long axis or strike direction of a target. For circumstances where the target is equal-dimensional, orthogonal or parallel orientation produces the same result. Antenna separation determines the reflectivity of horizontal planar targets. In practice, antenna separation is set equal to a distance smaller than 20% of target depth. Station spacing is defined as the distance between discrete radar measurements. The determining factor to assure that subsurface response is not aliased is ensuring the Nyquist sampling interval is greater than station spacing. This allows data collection to define steeply dipping reflections. Based on the Nyquist sampling equation, antenna-operating frequency is the primary factor in selecting station spacing. For high frequency operations in most subsurface materials, Station spacing is small. i.e. a calculated Nyquist sampling interval of above 5 cm allows for a 5 cm station spacing, ensuring detection of steep dipping reflections.
Establishment of a survey grid with a coordinate system precedes selection of survey line locations. Radar survey lines are ran perpendicular to the trend of the feature under investigation. Dictating the number of survey lines is the degree of target variation in the trend direction. i.e. if small targets are sought then survey line intervals should be closely spaced. Maximizing target detection is the priority in selecting the spacing between separate survey lines.
Processing and interpretation
Once field data is collected, processing and interpretation are conducted to address the problem that required GPR surveying. After exporting files from a GPR system to a desktop computer, profiles can be viewed and altered using a MATLAB-based program matGPR. Unprocessed data offers minimal information to the interpreter, thus it is necessary to preform minor processing to emphasize features of interest. The steps for simple processing are remove DC, remove global background, trimming time window and adjusting signal position. Removing the DC component eliminates low-frequency ground coupling in GPR profiles. Remove global background eliminates horizontal banding caused by standing and direct air waves. Adjusting signal position is the process of setting time-zero for the moment a radar pulse leaves the antenna. Trimming the signal position corrects the location of surface reflections. Determination of time-zero is based on antenna separation divided by the speed of light. Trim time window is an action used to eliminate unnecessary standing waves consuming the lower half of a GPR profiles. By discarding portions of the window, the matrix size for GPR data is reduced, allowing the interpreter to focus on targets of interest.
The vertical axis for GPR data is measured in nanoseconds, which is undesirable in interpreting target depth. In order to convert time to depth, a common midpoint test to determine the wave velocity in the subsurface is necessary. Preform a common midpoint test by taking manual trace readings while repeatedly separating the antennas in 10 cm intervals until receiving a hyperbolic reflection curve. By plotting the square of arrival time of the reflection versus the square of the source to receiver antenna separation distance, the wave velocity is obtained.
Interpretation of reflection profiles requires an understanding of the implication of a reflection. Radar reflections are produced by electrical property changes associated with boundaries describing subsurface stratigraphic units. Reflections are also generated by contacts between differing rock types. Therefore, potential points of interests in a GPR profile are reflection curve groupings that may indicate a buried feature such as a historic structures, rock grouping, or utility.
Application of GPR in archeological exploration
Archaeological research involves excavation of past cultures remains. The nature of the studies are destructive to surrounds and can be ineffective without direction. The use of a GPR survey aids in location of targets of interest and minimization to destruction of surroundings. The effectiveness of applying GPR surveying in archaeology is demonstrated by Montana Tech students involved in subsurface exploration at a prehistoric Native American site near Dewey, Montana. Prior to the GPR survey, the site revealed prehistoric Native American activity based on a pedestrian survey conducted by archeologists from Montana Tech and the Montana Bureau of Land Management. Desirable inhabitation features for this location include a natural spring and ravine landscape yielding game animals and protection.
The level landscape adjacent to the natural spring became the focus of a GPR survey. With the identified goal being to use GPR surveying to determine the location of buried fire hearths, which appear as highly reflective, closely spaced reflections with steeply dipping diffraction curves to their sides. Native Americans would typically work their tools and weapons with the aid of fire. As a result, cultural remains accumulate around fire hearths and are established as a target for archeological excavation. Prior to GPR surveying, an excavation grid size of 6 m wide and 16 m long was chosen for the archeological operation. Survey line spacing was set to 50 cm, for better data collection in archeological settings, use a closer spacing (i.e. 10 cm). Taking into consideration excavation time, a focus was set on detecting targets up to a depth of one meter. In order to achieve a resolution capable of discerning possible targets an antenna frequency of 400 MHz was applied. Antenna separation was set at 60 cm. Station spacing was set to 5 cm in order to record steep reflections indicative of contact with rocks. Trace stacking was set to four (i.e. four measurements are made at each station and then averaged into a single trace).
A hip chain can be used to record survey distance and trigger discrete station measurements. After surveying the excavation grid, GPR profiles are processed using a compatible software (i.e. MATLAB-based program matGPR). The location of anomalies on profiles directly correlated to the discovery of animal bones and a fire hearth. As suggested earlier, the unearthed fire hearth had an assemblage of artifacts located around it. A point of interest surrounding prehistoric bone accumulations is the consistent discovery of related artifacts in surrounding strata. This realization aids in unearthing more cultural remains as bone groupings appear distinctly on GPR profiles as an area of closely spaced vertical striations. While GPR profiles do not display small artifacts such as projectile points, the technique is reliable in detecting anomalies associated with artifact assemblages.
Ground penetrating radar relies on the interaction of subsurface material with EM fields. Therefore, there are a considerable extent of applications for GPR in near-surface exploration. The ability to select resolution versus depth of penetration enables GPR users to dictate survey design. For mapping utilities such as buried pipes and electric lines, high frequency application of GPR enables high-resolution mapping of such features. If detection of a certain target is not the priority, low frequency GPR applications can map soil stratigraphy, water table depths, and changes in rock type. This is useful for geochemical sampling, geotechnical investigations, and determining existence of groundwater. Conducting a GPR survey preceding construction activities can offer valuable information on subsurface conditions, and ensure that structures or roads are built on desirable ground.
- Annan, P. and Cosway, S. W., 1992, Ground-penetrating radar survey design, in Bell, R. S., Ed., Proc. Symp. on application of geophysics to engineering and environmental problems: Soc. Eng. Mining Expl. Geophys., 329–351.
- J. L. Davis, A. P. Annan, 1989, Geophysical prospecting: Ground-Penetrating Radar for High-resolution Mapping of Soil and Rock Stratigraphy, Vol. 37, issue 5. http://dx.doi.org/10.1111/j.1365-2478.1989.tb02221.x
- H. M. Jol, Ground Penetrating Radar Theory and Applications: Introduction, 4-27, ISBN: 978-0-444-53348-7
- W. A. Wilson, J. Clarke, 2017, A Ground Penetrating Radar Survey of the Unexcavated 24BE2206 Site near Dewey, in the Big Hole Valley of Montana.