Ground Penetrating Radar

Introduction and physical properties

Ground penetrating radar signals respond to subsurface changes in dielectric permittivity and electrical conductivity by reflecting or scattering off boundaries where these physical properties change. Also, GPR signals are more rapidly attenuated when ground conductivity is higher, so variations in penetration depth may be informative. Therefore, information can be obtained from GPR reflection sections by looking for continuous reflecting horizons, for diffraction patterns, and for variation in the general character (caused by changes in scattering behaviour) or penetration depth.

Survey design

GPR data are most easily interpreted when measurement stations are close together. In fact, if target interfaces dip by more than a few degrees they may be invisible if station spacing is too far appart. Steeply dipping geological contacts were not expected, so lines were surveyed with a station spacing of 1m. The entire site was not surveyed owing to a shortage of time.  Target priorities included the north western portion of the grid where monitoring well PTW-3 and several test pits provided good ground truthing. Also, test pit 1 (TP-1) was a relatively large area with no pavement cover, and there was oil flow encountered in its central region. The large east-west EM anomaly under the eastern half of line 52N was also of interest. The complete GPR survey included north-south lines spaced 4m apart between 04E and 88E, all surveyed at 1m intervals between 28N and 100N. Two east-west lines were also surveyed along lines 52N and 88N.

The choice of antenna dictates the signal's centre frequency and hence the penetration depth and vertical resolution. In our case, 200 MHz and 100 MHz antennas were available, and 100MHz antennas were chosen to ensure maximum penetration depth at this site.

Raw GPR data

Results from all sixteen adjacent lines are presented on a separate page in HTML (or PDF format for more predictable printing results) because it is easier to visually correlate similarities and differences when all data are on one page. Several of the features were interpreted to be noise. For example, the diagonal feature on lines 4E to 20E that apparently gets deeper towards the east, is a reflection from a fence just outside the limits of our survey area. While interpretation of the entire data set was not part of our contract, we provide some initial remarks about what may be revealed by these data by commenting briefly on results from line 84E and the east-west line, 52N. No velocity analysis was carried out, so two way travel times have not been converted to depths.

Data from north-south line, 84E

Two curved reflection patterns (centred under stations 50N and 84N) have been highlighted in red on the "interpreted" image (click radio buttons under the image to the right). The pattern under station 50N is  likely the top of a buried concrete foundation wall containing quantities of reinforcing bar.  There is also some suggestion of associated reflections at greater depth. A second curved reflection  pattern under station 85N is likely related to the tank recovered later by digging (see photo). At this location there is  a change in echo character throughout the first 100ns, probably related to this tank's trench. There is also a possible contiguous layer at 45ns under the northern-most 10m of this line. However, this is roughly one pulse-width below the surface arrival, so this interpretation should be considered preliminary.

Raw data. Interpreted image

Also visible on this section under station 40E is a 4-5m region where signals are ringing (overlay in blue). This type of behaviour occured when antennas were in puddles. Also, the overlay outlines a 25m segment in yellow where all signals after the direct arrival (0-15ns) appear more scattered than elsewhere along the line. This may be evidence of a distinct region within the fill, perhaps delineating where there is an increased quantity of wood waste.

In other regions with prominent anomalies on magnetics and EM surveys, it was not always easy to identify discrete objects amoungst the scattering patterns on GPR data. This may be partly due to the course spatial  sampling of the GPR survey, and partly due to the quantity of coarse material (wood, construction  debris, etc) within the fill.

Data from east-west line, 52N

• From 0E to 20E there is a consistent near-surface reflector, and signals penetrate to between 100 and 150ns.
• From 20E to roughly 55E echo character is somewhat more chaotic and penetration depth is generally less than 90ns. This region also includes an apparent reflecting horizon at around 80ns.
• From 50E to the end, signals penetrate to around 100ns, and there are many possible features, some of which probably correspond to buried concrete foundations in the area.

As a final remark regarding use of GPR data, it should be remembered that qualitative characteristics of GPR data these are not usually easy to interpret directly. However, they point to recognizable zones that might benefit from more careful characterization using direct testing methods. In this way GPR data can help reduce the number of direct testing sites if it is assumed that features are consistent within a region of consistent GPR response. On this line, there appear to be three distinctive zones: