GEOLOGICA CARPATHICA, 49, 1, BRATISLAVA, FEBRUARY 1998
ELECTROMAGNETIC VARIATIONS ASSOCIATED
WITH THE SEISMICITY OF THE FRONTAL HELLENIC ARC
and FILIPPOS VALLIANATOS
Technological Educational Institute of Athens, Greece
Technological Educational Institute of Crete, Chania, Crete, Greece
(Manuscript received March 18, 1997; accepted in revised form December 11, 1997)
Abstract: In the present paper we report the results of almost three years of measurements of the electromagnetic
variations in 3 and 10 kHz, 41 and 53 MHz. The data are collected using a telemetric network installed along the island
of Crete, i.e. in the central segment of the South Hellenic arc. The experiment indicates that electromagnetic perturba-
tions occur prior to seismic activity and supports previous observations that the time sequence of the variations
appears in an invariant pattern.
Key words: seismoelectromagnetism, Greece, Hellenic arc.
A number of electromagnetic phenomena associated with
earthquakes have been reported in recent time in different fre-
quency ranges. There are slow magnetic field variations
which are claimed as a useful tool for earthquake prediction
(Shapiro et al. 1994). Ultra low frequency (ULF) electromag-
netic emissions (0.0110 Hz) (Fraser-Smith et al. 1990),
ELF-VLF electromagnetic radiation (0.1100 kHz), (Gokh-
berg et al. 1982; Fujinawa & Takahashi 1990; Nomikos et al.
1994) and HF emission (f>1 MHz) (Nomikos et al. 1997) are
also associated with earthquakes.
In order to study the aforementioned preseismic electro-
magnetic phenomena, a telemetric system has been installed
on the island of Crete (South Aegean, Greece). Structurally,
our test area is linked with the frontal part of the central seg-
ment of the Hellenic arc. The south Aegean is an appropri-
ate place for our experimentation, because in this area the
existence of shallow and intermediate depth seismicity has
long been recognized (Hatzfeld & Martin 1992).
The latter multipoint network records the earths electro-
magnetic field variations in four field stations installed
along the island of Crete. In each field station, we measure,
using tuned loop antennas, the two horizontal components
of the electromagnetic field variations, in low frequencies
(LF), i.e. in 3 and 10 kHz. Using half-wavelength dipoles
we measure high frequencies (HF), i.e. in 41 and 53 MHz.
The central station communicates with a datalogger in the
field station and collects the data through a standard tele-
phone line (Fig. 1). In this work we present data recorded
from the telemetric network during the time period from Oc-
tober 1992 to December 1995.
Instrumentation and data selection
For the observation of electromagnetic variation in each
field station, the following instruments are used:
a) Four receivers appropriate for measuring the electro-
magnetic field variations at 3 and 10 kHz into EW and NS
directions. These receivers are constructed using wide band
and low noise amplifiers and switching band-pass filters
tuned by crystal oscillators. The final stage of the receiver is
an RMS to DC converter. Thus, the output V
of the re-
ceiver is a DC voltage proportional to the power spectrum
of the magnetic field which excites the antenna
and is given by the expression
are the number of turns, the radius and
the quality factor of the antenna, respectively.
bandwidth of the filter, G is the total gain of the system,
H/m the magnetic permeabilitty of free space and
the angular frequency of the electromagnetic wave.
b) Two receivers for measuring the electric field varia-
tions at 41 MHz and 53 MHz. The receivers are constructed
using double supper heterodyne technology and the output
in each of them is a DC voltage which is proportional to the
Fig. 1. Instrumentation arrangement of the telemetric system.
* Corresponding author: F. Vallianatos, 111 Homirov Str., Moschato 18 345, Athens,
Greece, E-mail: firstname.lastname@example.org
electric field which appears on the antenna. The anten-
nas used for these very high frequencies are horizontal
half-wavelength dipoles tuned at the above frequencies.
c) a datalogger that is the main instrument for reading the
analog information from the electromagnetic receivers. The
sampling rate was taken on a channel basis every second, and
the average value of 60 samples for each channel saved in the
final memory. The signal from the field station was transmit-
ted through a telephone line to the Central one (Fig. 1).
We would like to point out that the most crucial point in
our experiment was to determine noise free observation fre-
quencies. A number of transmitting stations radiate electro-
magnetic signals at almost all frequencies over and around
the island of Crete. To make sure that the observation fre-
quencies were silent, they were checked using a radio re-
ceiver, for several months. We emphasize that a criterion for
the selection of an electromagnetic variation as a preseismic
signal, associated with an event of magnitude Ms approxi-
matly grater than 5.0, should be its existence in both compo-
nents simultaneously, otherwise they could be artificial in-
terference. Any interference from mobile transmitters at
these two frequencies obviously will produce spikes on the
recordings for at most a few minutes and in the specific fre-
quency of the transmitter and not simultaneously in both of
them. Furthermore, in order to localize the anomalous pat-
tern, artificial intelligence techniques have recently applied
(Yalouris et al. 1996). Among other criteria the latter tech-
nique is checking the change with time of the mean level
and the sampling variance of the background noise.
Recording of the electromagnetic variations started in Oc-
We proceed now to the presentation of the recordings ob-
tained from October 1992 to December 1995. Table 1 shows
all the earthquakes with Ms magnitude greater than 5 in
the vicinity of Crete (latitude: 3437, longitude: 2228)
during the above time window. The stations of our net-
work and the frequencies in which electromagnetic emis-
sions were observed prior to earthquakes are shown in
Table 1. Fig. 2 shows the earthquakes during the afore-
mentioned period. Furthermore, the field stations of the
telemetric network indicated by their initial (i.e. N = Ni-
pos, I = Ierapetra, H = Heraklion and D = Drapanias).
Figures 3 and 4 show typical recordings of an electromag-
netic variation prior an earthquake.
In a recent paper the time sequence of seismoelectromag-
netic events studied constructed time charts (Nomikos et al.
1997). The conclusion was that a sequence of electromag-
netic events which precede an earthquake exists. In all the
examined cases the LF variations (i.e. in 3 and 10 kHz)
Fig. 2. Map showing the sites of the stations of the telemetric net-
work and the distribution of epicentres of earthquakes with Ms>5
in the vicinity of Crete, from October, 1992 to December, 1995
(see Table 1).
Table 1: List of the earthquakes with electromagnetic variations collected at the stations of the network. The position of the observation
station inticated by its initial i.e. N = Nipos, I = Ierapetra, H = Heraklion and D = Drapania.
NOP= Station not in operation during this period
ELECTROMAGNETIC VARIATIONS ASSOCIATED WITH THE SEISMICITY 59
always precede the HF variations (i.e. in 41 and 53
In the present paper we will discuss more specificly the
electromagnetic variations which preceded the November
21st 1992 and May 23rd 1994 intermediate depth earth-
quakes which were the strongest in magnitude events in the
The first one on November 21st 1992 with Ms 6, epicen-
ter at (35.9
E) and focal depth 70 km belongs to
the seismogenic source zone 1A, as identified by Papaza-
chos (1990); this source region lies under the eastern part of
Peloponnese but extends to the north up to 39
to the south up to the western corner of the island of Crete.
The second earthquake, on May 23st 1994 with magnitude
Fig. 3. Electromagnetic variation at 3 and 10 KHz in EW direction, recorded prior to November 21st 1992 earthquake.
Ms = 6, epicenter at (35.48
E) and focal depth
66 km belongs to source volume 1B, which lies under the
island of Crete (see Fig. 1 of Kiratzi & Papazachos 1995).
a) Electromagnetic variations prior to the November 21st
(i) At the low frequencies, 3 and 10 kHz, on November
17th 1992 and at 13:15 GMT, for a time period of 15 min-
utes a simultaneous electromagnetic variation was recorded,
at Drapania, Nipos and Heraklion and in both measuring di-
rections. The strongest signal appeared in Heraklion and the
weakest in Drapania. Ierapetra being the furthest from the
epicenter, did not record any signal.
Fig. 4. Electromagnetic variation at 41 and 53 MHz, recorded prior to November 21st 1992 earthquake.
Fig. 5. Time chart depicting the electromagnetic variations in as-
sociation to the Ms = 6 earthquakes of November 21st 1992 and
May 23rd 1994 to December 1995. Both of them have epicenters
in the vicinity of Crete (see Fig. 1). The arrows indicate the earth-
(ii) At high frequencies the electromagnetic variations
started mainly from 1:00 GMT of November 19th 1992 and
continued until 22:00 GMT of the same day, even though
slight variations did continue through the following day. It
can be seen that the strongest signals were recorded at the
nearest (to the epicenter) station of Drapania.
b) Electromagnetic variations prior to the May 23rd 1994
(i) At the low frequencies, 3 and 10 kHz, on May 17th 1994
and at 6:00 to 16:00 GMT a simultaneous electromagnetic
variations was recorded, at Drapania, Nipos and Ierapetra and
in both measuring directions. The strongest signal appeared
in Drapania and the weakest in Ierapetra. The strange thing is
that station Heraklion being quite close to the epicenter, did
not record any signal, either at low or high frequencies.
(ii) At high frequencies the electromagnetic variations
were detected at 14:00 to 18:00 GMT of May 22nd 1994 at
Drapanias and Nipos stations.
In Fig. 4 we see the time charts describing the sequence of
electromagnetic events. We observe the already reported time
pattern, i.e. LF variations HF variations Earthquake event.
In the present contribution recordings of the electromag-
netic anomalies that precede to earthquakes in the time win-
dow October 1992December 1995 are presented.
The experimental results indicate the presence of electro-
magnetic variations in the frequencies of 3 and 10 kHz, 41
and 53 MHz, associated with shallow and intermediate
depth earthquakes in the vicinity of Crete Island (South Ae-
The study of the time sequence of seismoelectromagnetic
events the constructed time charts shows that a sequence of
electromagnetic events that precede the earthquakes exists.
The recently collected data strengthen the conclusion that
the electromagnetic variations appears to follow an invari-
ant time pattern (i.e. LF variations HF variations Earth-
quake event). The latter pattern suggests that the appearance
of HF variations as follows of an LF anomaly observation is
a strong indicator that an earthquake process is in the final
stage of preparation.
Fraser-Smith A.C., Bernardi P.R., McGill P.R., Ladd M.E., Halli-
well B.A. & Villard O.G., 1990: Low-frequency magnetic
field measurements near the epicenter of the 7.1 Loma Prieta
earthquake, Geophys. Res.Lett., 17, 14651467.
Fujinawa Y. & Takahashi K., 1990: Emission of electromagnetic
radiation preceding the Ito seismic swarm of 1989. Nature,
Gokhberg M.B., Morgounov V.A., Yoshino T., & Tomozawa I.,
1982: Experimental measurement of electromagnetic emis-
sions possibly related to earthquakes in Japan. Journal of
Geophysical Research, 87, B9, 78247828.
Hatzfeld D. & Martin C., 1992: The Aegean intermediate depth
seismicity defined by ISC data. Earth Planet. Sci. Lett., 113,
Kiratzi A. & Papazachos C., 1995: Active deformation of the shal-
low part of the subducting lithospheric slab in the southern
Aegean. J. Geodynamics, 19, 6578.
Nomikos K., Bakatsakis M., Paterakis D., Kogionis T., Sideris S.,
Zaxaropoulos B., Cristou C., Kaliakatsos J. & Vallianatos F.,
1994: Development of a telemetric system for observation of
radioemission associated with earthquakes in Crete island. In:
Proceedings of the XXIV General Assembly, European Seis-
mological Commission,vol. II 11121120, Athens, Greece.
Nomikos K., Vallianatos F., Kaliakatsos J.,Sideris S. & Bakatsakis
M., 1997: Latest aspects of telluric and electromagnetic vari-
ations associated with shallow and intermediate depth earth-
quakes in South Aegean. Annali di Geophysica (in press).
Papazachos B., 1990: Seismicity of the Aegean and surrounding
area. Tectonophysics, 178, 287308.
Shapiro V.A., Muminov V. & Abtullabecov K.N., 1994: High pre-
cision magnetometry for earthquake prediction in Uzbeki-
stan: Ninety- one forecasts between 1982 and 1982. In:
Electromagnetic Phenomena related to Earthquake prediction,
Terra Pub., Tokyo, Japan.
Yalouris C., Vallianatos F., Nomikos K. & Sideridis A., 1996: Rec-
ognition of preseismic electromagnetic anomaly using Artifi-
cial Inteligence techniques. 1
Congress of the Balkan
Geophysical Society, Athens, Greece, 8485.