Geoscience Reference
In-Depth Information
There are two obvious dielectric constant discontinuities. One is the interface between the drain-
age pipe and surrounding soil material and the second is the interface between the drainage pipe
and the air and water within it (Figure 29.1). However, if the drainage pipe wall thickness is small
relative to the radar pulse wavelength (definitely the case for corrugated plastic tubing and likewise
for clay tile), then as a result of constructive or destructive interference between the radar pulses
reflected off the outer and inner walls of the pipe, the effective GPR reflection response essentially
becomes governed by the dielectric constant values of the surrounding soil material and the air and
water inside the pipe.
The dielectric constant ranges in value from 1 for air to 80 for water with dry soil closer to the
lower end of this range, ~5 to 15, and very moist or saturated soils near the middle of the range, ~30
to 40 (Conyers and Goodman, 1997; Reynolds, 1997; Sharma, 1997; Sutinen, 1992). The dielectric
constant, ε, of soil material is directly dependent on the volumetric moisture content, θ. A relation-
ship between ε and θ was empirically developed by Sutinen (1992) for glacial materials, similar to
those found in the Midwest United States, and is expressed as follows:
2
3
ε
=+ +
32 35 4
.
. ()
θ
101 7
.( )
θ
−
63
(
θ
)
(29.1)
The GPR unit used predominantly for this research was the Sensors & Software Inc. Noggin
plus
with 250 MHz center frequency antennas. In order to investigate the effect of different antenna
frequencies on drainage pipe detection, other Sensors & Software Inc. GPR systems were tested
in a more limited manner. These included a Noggin
plus
unit employing 500 MHz center frequency
antennas and a pulseEKKO 100A unit equipped with 100 MHz center frequency antennas.
For data collection during this project, the distance between measurement points along a
transect was 0.05 m, and thirty-two signal traces were averaged at each point location. For each
test plot, GPR measurements were typically collected along two sets of parallel transects oriented
perpendicular to one another and forming a rectangular grid covering the test plot. The spacing
distance between adjacent GPR measurement lines was usually 1.5 m. These transect measurements
were then used to produce GPR images of the soil profile and time-slice amplitude maps showing
drainage pipe patterns.
As discussed in Chapter 7, GPR profiles are constructed by plotting side by side the sequential
signal traces collected along a line of measurement. GPR profiles represent the amount of reflected
radar energy returning to the surface from different depths beneath the line along which data were
collected. The vertical axis on a GPR profile is given in two-way radar signal travel time unless soil
water content or soil dielectric constant data are available to allow converting the two-way travel
time values to depth values. Information from all of the measurement transects at a test plot were
integrated to generate the GPR time-slice amplitude maps, which represent the amount of reflected
radar energy returning to the surface over an area from a specified interval of two-way travel time
(or depth).
A certain amount of computer processing was employed to produce the GPR profiles and ampli-
tude maps presented in this case history. Computer processing was essential in order to enhance
the GPR drainage pipe response embedded in the raw data. The computer-processing steps for gen-
erating GPR profiles included a signal saturation correction filter and a spreading and exponential
compensation gain function. For GPR time-slice amplitude maps, a signal saturation correction fil-
ter, two-dimensional migration, signal trace enveloping, a high-frequency noise filter, and a spatial
background subtraction filter were all used.
The impacts of antenna frequency (100, 250, and 500 MHz), soil hydrologic conditions, pipe
construction material, and drain line orientation on GPR detection of agricultural drainage pipes
were evaluated at one test plot. This test plot was built specifically for the overall GPR drainage pipe
detection project and is located behind the ElectroScience Laboratory (ESL) at Ohio State Univer-
sity in Columbus, Ohio. The surface soil (2.5 to 15 cm depth) texture at the ESL site, as determined
