2. Method of Studies
[6] Usually, to reveal disturbances in ionospheric parameters, one
calculates their deviations from the mean or median values over
some period of time. In our study, we used running medians
obtained for each hour of the day on the basis of the 10 previous
days for each analyzed day of the earthquake preparation period.
This period was 7 days including the day of the earthquake. To
determine the input of the assumed seismogenic disturbance into
the quiet state of the ionosphere and exclude the dependence on the
local time and season, for each hour of the entire period under
study the value
Df/f m = (f - f m)/f m (where
f m is the running median
and
f is the critical frequency for the
F2 layer and blanketing
frequency for
Es ) was determined for the frequency parameters.
This is a significant difference from earlier publications on
detection of seismoionospheric disturbances
[Silina et al., 2001]
where the deviations relative to the critical frequency itself
averaged over several earthquakes were calculated. This led to an
underestimation of the disturbance relative value at the increase of
the critical frequency and to its overestimation at the frequency
decrease. The variations of the
Df/f m and
Dh'Es = h'- h' m value
from one hour to another show the dynamics of the disturbance in
the lower and upper ionosphere at the preparation of each
particular earthquake.
[7] Using this method, we processed the hourly data of the
measurements at Kokubunji ionospheric station
( j = 35.7 o N,
l =139.5 o E) for 40 earthquakes with
M 
6 and the hypocenter depth
h < 50 km having occurred in the Pacific region in 1985-2000.
Only the cases were considered when the ionospheric station was
within the zone of earthquake preparation ( r 
eM, km, where
r is
the radius of the preparation zone). The data on the earthquake
parameters were taken at http://www.neic.cr.usgs.gov. Kokubunji
station is the only station in this seismically active region for
which the most complete set of ionospheric parameters for many
years has been published
(Communications Research Laboratory, 1985-2000, Ionospheric
data in Japan,
http://wdc-c2.nict.go.jp/index_eng.html,
World Data
Cent. for Ionosphere, Tokyo).
|
|
Figure 1
|
[8] As an example of the above described data processing, Figure 1
shows the variations in the parameters of the
Es and
F2 layers in
the period of preparation of the 20 November 1996 earthquake
( M= 6.8;
r = 211 km is the distance from the epicenter to the station).
Time is counted from the beginning of the minus sixth day (0000 LT)
prior to the day of the earthquake (the seventh day). The
dashed line shows the earthquake moment. It follows from Figure 1
and the analysis of other earthquakes that during the preparation
period at Kokubunji station, parameter deviations of both signs are
detected but sharp positive "spikes" 2-3 hours long prevail. Just
such spikes were discussed by
Korsunova et al. [1996],
Liperovsky et al. [1992], and
Silina et al. [2001].
[9] The value of some spikes in frequency and height parameters can
reach 50% and more and 40-60 km, respectively. The mean values
of the deviations from the running median are
20% and
10 km
for the frequencies and virtual heights, respectively.
[10] As should be expected, parameters of the sporadic
E layer are more
variable than parameters of the regular
F2 layer and anomalous
spikes in the
Es layer with duration 2-3 hours are noticed within
the whole period of observations. Such spikes can be caused by
various meteorological factors that do not influence the behavior of
the
F2 layer. That is why the excess by the spike amplitude the
averaged value or mean-square deviation cannot be chosen as a
criterion for ionospheric precursor, especially in the
Es layer. From
the other side, for all considered earthquakes, as well as for that
one presented in Figure 1, there is only one group of spikes in all
ionospheric parameters that manifest itself in 24-hour interval. For
the earthquake on 20 November 1996 this group manifest itself 5.5
days prior the moment of the earthquake (black peaks indicate the
moment of the group appearance). For other earthquakes this
period comprises 1-6 days.
[11] During the preparation period, the appearance of anomalous
spikes was close in time for all three parameters only once. The
time of appearance of these spikes relative to the earthquake
moment differs considerably for earthquakes of different energetic
class and different epicenter distances. The performed calculations
showed that the value of the deviations of the ionospheric layer
parameters differs slightly if one uses 27-day running medians or
monthly mean values, but this choice does not influence on the
appearance of the anomalous spikes and their features.
[12] Earlier,
Korsunova and Khegai [2005]
have shown that spikes
groups with duration 2-3 hours that appear simultaneously in the
Es and
F2 layers are not connected with magnetic disturbances,
because for the considered earthquakes at the stage of the
preparation the daily mean index was
Ap < 15 nT. All mentioned
above allows us to suppose that the discussed groups of
perturbations in ionospheric parameters are caused by the
processes connected with the preparation of the earthquake. Hence
they can be considered as ionospheric precursors to earthquakes.
[13] On the basis of the analysis of ionospheric parameter deviations in
the period of preparation of 40 earthquakes in the vicinity of
Kokubunji station, we proposed a method of detection of
ionospheric precursors of earthquakes using the following set of
morphological signs: (1) the presence in the hourly deviations of
ionospheric parameters of spikes group (that appear in the
Es and
F2 layers simultaneously) with duration 2-3 hours; (2) the excess
by the spikes amplitudes of the 20% and 10 km values of
deviations in the frequency parameters of the ionospheric layers
and in
Dh'Es, respectively.
[14] We consider the disturbances in ionospheric parameters satisfying
these requirements as medium-term earthquake precursors. These
disturbances will be discussed below.
Powered by TeXWeb (Win32, v.2.0).