Geophysical Prospecting in The Akropotamos Dam (N. Greece) by GPR and VLF Methods

JOURNAL OF THE BALKAN GEOPHYSICAL SOCIETY, Vol. 2, No 4, November 1999, p. 120-127, 6 figs.
Geophysical prospecting in the Akropotamos dam (N. Greece)

by GPR and VLF methods
Alexandros S. Savvaidis * , Gregory N. Tsokas * , George Vargemezis * and George Dimopoulos *

*
Geophysical Laboratory, Aristotle University of Thessaloniki, 54006, Thessaloniki, Greece

(E-mail: [email protected])
* *
Laboratory of Engineering Geology and Hydrogeology, Aristotle University of Thessaloniki, 54006 Thessaloniki, Greece
( Received 9 September 1999; accepted 30 October 1999 )
Abstract: The present work demonstrates the application of two geophysical methods in the studies for the
foundation of a dam in Northern Greece. The VLF and the Ground Penetrating Radar (GPR) were employed to
investigate specific problems concerning the near surface tectonic setting. The local geological setting is relatively
simple and it consists of granodiorite marbles and colluvial deposits.
The VLF method was mainly used for the detection of the fault zones in the area that intersects the axis of the
dam. The GPR method was used for a detailed study of the fractures in the area. The high-resolution capability of
the method resulted in recording the cracks of the main geological formation of the area. The combined use of the
two mentioned geophysical methods helped to reveal the large and small-scale fracture zones found in the area of
the dam.
Key Words: GPR, VLF, Dam foundation.
The geological setting of the dam foundation

mainly consists of granodiorite of Tertiary age. Also
marbles and colluvial deposits outcrop in a wider area.
The plutonic rock of diorite and tonalite also cover a
broad part of the study area. Depending on the degree
of metamorphism, the coarse-grained granodiorite can
be differentiated from the medium to coarse-grained
gneissed granodiorite. The granodiorite is ruptured
and the fracture zones are filled with mylonit. There
are two types of faults mapped in the area. The first
type are of low dip angles from 10 to 35 and the
second type are of high dip angles from 70 to 90.
The direction of the faults varies in all directions. The
geophysical methods were used in order to detect the
fracture zones and the cracks of the granodiorite.
INTRODUCTION
The high increase of the agricultural activity makes
necessary finding solutions to the problem of water
supplying. A common solution is the construction of
an irrigation dam in areas, which are close to the
upper and middle part of a river. An example is the
area of Akropotamos (N. Greece) which is presented
in Figure 1 (Wessel and Smith, 1995).
The use of the geophysical prospecting methods at
the site of the foundation of the dam is very common
especially at the initial stage of the whole study.
Geophysical methods are used for the detection of
inhomogeneities at shallow depth. Especially, they are
used to reveal concealed fracture zones.
For these reasons, the VLF and the GPR methods
were applied for the particular study (Fig. 2). The first
geophysical method is commonly used for the
accurate detection of fracture zones. However, the
GPR method is not conventionally used in such
problems. But, it was employed for this particular case
in order to detect cracks in the bedrock.
VLF DATA
The WADI of ABEM was used for the acquisition
of the VLF data. The measurements were carried out
on eight profiles of varying length. The exact
locations of profiles are shown in Figure (2). Readings
1999 Balkan Geophysical Society (access http:// bgs.ankara.edu.tr)

120
Savvaidis et al.
121
FIG. 1. Map of Greece showing the area of the foundation of Akropotamos dam in the inlet. The axis of the dam
has been also drawn. It is striking in the East-West direction approximately.
were taken along the profiles stepwise at 5 m

intervals. The profiles had an almost north-south
orientation with the exception of profile 3 that was
oriented along the river channel, i.e. at the northeastsouthwest direction.
Profile 1 lies along the diversion tunnel that is
almost perpendicular to the axes of the dam. The
original data is presented in Figure 3(a). The KarousHjelt filter was applied to the data (Karous and Hjelt,
1983) and the filtered data are given in Figure 3(b).
In Figure 3(c), the 2-D current density pseudo-section
is presented. A fracture zone is located between 33
and 105 m with a dip towards the South.
The profile 10 is presented in Figure 4. The
original data are given in Figure 4a, the filtered data
along the Karous-Hjelt scheme are shown in Figure
4b, while in Figure 4c shows the 2-D current density
pseudo-section. A fracture zone is detected between
117 and 137 m, which causes the high positive values
of the real part (Fig. 4b). An attempt to find a model

that describes the data was constructed (Fig. 4d). The
software VLFMOD (Edsen and Nissen, 1997) was
used for the modelling of the data. This software is
based on the routine of Nissen (1986) that performs
the modelling of 2-D structures. The Fraser filter
(Fraser, 1969) was first applied to the data. This filter
is designed for the noise suppression of the data.
Specifically the modified 5-point Fraser filter was
used to plot the output of the filter at the same
locations as the tilt angle measurements. The
concluded model consists of an ensemble of vertical
sided prisms of low resistivity (70 Ohm-m) which is
hosted in a half-space of high resistivity (1000 Ohmm). In the upper part of Figure 4d, the filtered data
are represented with the blue line while the green one
represents the filtered model response. The model
(lower part of Fig. 4d) represents a vertical fracture
zone, which extends till the depth of 70 m. Its width
122
GPR and VLF Methods
FIG. 2. Topographic map of the site where the Akropotamos dam is to be built. The profiles measured with both
GPR and VLF method are given with red lines. The profiles along which only GPR readings were taken are
represented with dashed black line. The dotted black lines show the profiles along which only VLF data were
taken.
varies with depth and it is 10 m till the 40 m depth, 7
m from 40 to 61 m and 5 m from 61 to 70 m. The
fracture zone is centred at the x position of 125 m.
GPR DATA
The instrument used for the GPR survey was a
Pulse Ekko 1000 from Sensors&Software having
antennas with central frequencies at 225 and 450
MHz. A radar trace was acquired every 0.2 m along
the profiles.
The GPR data were processed using the Sensors &
Software provided software (Annan, 1993; 1994). The
traces were edited wherever it was necessary. Editing
is essentially removing bad traces or a bad section of
the profile. Then, a signal saturation correction
(DEWOW) was applied. This is a slowly decaying
low frequency component on the trace, which is
superimposed on the high frequency reflections. Its
magnitude and decaying rate depend on the proximity
of the transmitter and the receiver as well as on the
electrical properties of the ground. In order to remove

this low frequency noise from the data, a running
average filter is applied on each trace. The power
spectrum was calculated and low pass filtering was
applied to each radargram depending on the frequency
with the highest amplitude. That was done in order to
enhance the desired signal and remove the high
frequency noise. Finally, a spatial high pass recursive
filter was applied to the data.
The use of this filter is to enhance localised
features at the expense of strong long wavelength
features such as stratigraphy or flat lying horizons.
The cut-off frequency was 10% of the Nyquist
frequency, i.e. 0.5 cyc/m. Figure 5a shows a part of
the radargram of profile S1. Its length is 23 meters
and its direction is from south to north. The antenna
with central frequency of 450 Mhz was used. In the
first 20 ns a lot of reflectors of low apparent dip are
presented (a, b, c). They are clearly seen in the area
between 11 to 23 meters from the north edge of the
Savvaidis et al.
123
FIG. 3. VLF profile 1 along the diversion tunnel of the dam (Fig. 2). Source frequency is equal to 18.3 kHz. (a)
Original data. In the x-coordinate the distance from the beginning of the profile is given. In the y-coordinate the
percentage of H sz H p is given. (b) Filtered data. The Karous-Hjelt filter was applied to the data. (c) 2-D
pseudosection.
124
GPR and VLF Methods
FIG. 4. VLF profile 10 along a road in the west part of the area of study. Source frequency is equal to 15.9 kHz. (a)
Original data. In the x-coordinate the distance from the beginning of the profile is given. In the y-coordinate the
percentage of H sz H p is given. (b) Filtered data. The Karous-Hjelt filter was applied to the data. (c) 2-D
pseudosection (d) Suggested model for the anomaly presented in (b). Resistivity is in ohm-m and the size of
squares is in meters. The blue line represents the filtered data while the green one is the filtered model response.
Savvaidis et al.
125
FIG. 4. (Continued).
FIG. 5. A part of the radargram of the S1 traverse, which is a part of profile along the diversion tunnel of the dam.
Reflectors at the upper part are easily recognised (a, b, c). In the deeper part, reflectors are very rare (d, e, f) and
they show a greater apparent dip.
126
GPR and VLF Methods
profile. In the deeper part, i.e. in the part between 2040 ns, the reflectors are not very frequent and not very
well defined. However at least three reflectors are
detected (d, e, f). All these reflectors are due to cracks
filled with low resistivity and/or high dielectric
permeability material. These reflectors indicate small
cracks at the geological formation.
Figure 6 illustrates the radargram of profile KIT4b.

Its length is 11 meters and its direction is northeast to
southwest. The antenna with central frequency of 450
MHz was used. Reflectors of low apparent dip are
also present in this profile in the upper part (0-20 ns)
of the section. At the part of the section between 2060 ns, reflectors (a, b, c, d) of high apparent dip are
FIG. 6. Radargram of the KITI4b traverse, which is a part of profile KITI that is almost along the route of the river.
Reflectors are easily recognised (a, b, c, d). Their apparent dip is greater than that of the reflectors revealed in
traverse S1.
Savvaidis et al.
indicating cracks that are longer than those on profile

S1 (Fig. 5).
CONCLUSIONS
The use of the various geophysical techniques
illustrates their importance in the study for the
foundation of dams. The current case history shows
that the joint use of ground radar and VLF methods
provide information to solve several issues concerning
the ability for the construction of a dam in the area.
The VLF method revealed the large-scale fracture
zones and helped on deciding the techniques that are
going to be used on the construction of the diversion
tunnel and the foundation of the dam. Also, it defined
the width and depth extent of these fracture zones so
these can be mapped along the axis of the dam.
The GPR data manage to give information on the
small-scale problems as revealing the cracks in the
bedrock. These are of low apparent dip in the upper
part (near surface) and of higher apparent dip in the
deeper part of the section.
The use of the geophysical methods at the initial
study stage and also their decisive role at the
construction stage of a dam is demonstrated. The low
127
cost of the geophysical methods in comparison to the

geotechnical boreholes, that are used for defining the
extent of fracture zones and the inhomogeneities,
show the necessity of the geophysical techniques
because of their cost effectiveness.
REFERENCES
Annan, A.P., 1993, Practical Processing of GPR data: Paper read
at the Proceedings of the Second Government Workshop on
Ground Penetrating Radar, 26-28 October, Columbus, Ohio.
Annan, A.P., 1994, GPR Frequency Selection: Paper read at
SAGEEP, 27-31 March, Boston, Massachusetts.
Edsen, N.A., and Nissen, J., 1997, VLFMOD, a free forward VLF
modeling software package: ftp://home1.swipnet.se/~w11019/ABEM-ftp/.
Fraser, D.C., 1969, Contouring of VLF-EM data: Geophysics, 34,
958-967.
Karous, M, and Hjelt S. E., 1983, Linear filtering of VLF dipangle measurements: Geophysical Prospecting, 31, 782-794.
Nissen, J., 1986, A versatile electromagnetic modeling program
for 2-D structures: Geophysical Prospecting, 34, 1099-1110.
Wessel, P., and Smith, W., 1995, New version of the Generic
Mapping Tools released: EOS Trans. Amer. Geophys. U., 76329.

Geophysical Prospecting in The Akropotamos Dam (N. Greece) by GPR and VLF Methods

Uploaded by

Copyright:

Available Formats

Geophysical Prospecting in The Akropotamos Dam (N. Greece) by GPR and VLF Methods

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Geophysical Prospecting in The Akropotamos Dam (N. Greece) by GPR and VLF Methods

Uploaded by

Copyright:

Available Formats

JOURNAL OF THE BALKAN GEOPHYSICAL SOCIETY, Vol. 2, No 4, November 1999, p. 120-127, 6 figs.

Geophysical prospecting in the Akropotamos dam (N. Greece)

Alexandros S. Savvaidis * , Gregory N. Tsokas * , George Vargemezis * and George Dimopoulos *

Geophysical Laboratory, Aristotle University of Thessaloniki, 54006, Thessaloniki, Greece

The geological setting of the dam foundation

1999 Balkan Geophysical Society (access http:// bgs.ankara.edu.tr)

were taken along the profiles stepwise at 5 m

of the real part (Fig. 4b). An attempt to find a model

GPR and VLF Methods

electrical properties of the ground. In order to remove

GPR and VLF Methods

GPR and VLF Methods

Figure 6 illustrates the radargram of profile KIT4b.

indicating cracks that are longer than those on profile

cost of the geophysical methods in comparison to the

You might also like