The region in and around Qiryat Shemona is subjected to large seismic risk.
An accurate estimation of the seismic ground motion across this area is of prime
importance for urban developments and mitigation of seismic risk. Following many
examples of monitored strong earthquakes in the current century, it is evident 
that thelocal site effects may have a dominant contribution to the intensity of 
damage and destruction. In this study we focused on the first stage of associated
site effects and seismic hazard by preparing two maps that reflected the 
fundamental characteristics of site effects: distribution resonance frequency and 
maximum relative amplification.
    We investigated site effects using the records of ambient noise collected by
short period digital seismic stations during a period of six months. The ambient 
noise survey was carried out at 285 sites. The measurement points were selected to 
provide good coverage of the area, considering different surface sedimentary 
deposits,thickness of sediments and the shear-wave velocity contrast between 
sediments and bedrock, using a 0.25 km grid.
    The site response is obtained in terms of spectral ratio calculated by dividing 
the horizontal versus the vertical components of motion observed at the same site 
and in terms of ratio of the horizontal components spectra at the investigated 
site and those at the reference site. In some cases the shape of spectral ratios 
and level of spectral amplifications obtained from two methods are similar. 
However, in many cases, the soft-to-hard spectral ratios do not reproduce the 
frequency at which the maximum amplification occurs. The obtained results support
the application of horizontal-to-vertical spectral ratios, using ambient noise 
measurements, which couldbe very important on urban and land planning.
    The spectral ratios indicate site amplifications range from 2.0 to 7.0 in the 
frequency band 2.0-8.0 Hz within the city and from 2.0 to 5 in the frequency band
0.7-5.0 Hz outside of the city. These results suggest that there is significant 
shear-wave velocity alteration and considerable variation of sediments thickness.
     As the next step in present project the estimations of fundamental resonance 
frequency and amplification must be calibrated with independent geophysical data.
This task can only be achieved by a detailed study of the geotechnical parameters 
for the soil and rock present.

     Earthquake hazard is presently accepted to be a combination of three factors:
the magnitude, location and frequency of earthquakes, the effect of the wave path 
of seismic energy, and the degree to which local geological conditions contribute 
to damage. Destructive earthquakes have demonstrated that damage is often more 
severe over unconsolidated deposits than on firm rock outcrops. Since sedimentary
valleys are often the prime locations for the development of urban areas, local 
amplification is a major concern in earthquake-prone regions (e.g. San Francisco,
Lima, Bogota, Kobe) but also in moderate seismicity areas where the mid-size 
cities developed could be struggle with future damaging events due to the 
combination of site effects and urban development.
     Given the greatly increased level of damage that can be produced, site effects 
must be measured or estimated especially when choosing the location and design of
critical and essential facilities. One of the goals of engineering seismology has 
beento quantitatively measure amplification of ground motion in earthquake-prone 
areas. 
     Site response can be estimated through an empirical approach (using seismic 
recordings) or a theoretical approach (using synthetics generated with modeling 
methods). The numerical approach requires a good knowledge of the local structure
responsible for site effect. Both geometric and mechanical properties have to 
accurately model the seismic response of soil. A more complex model requires more
precision. Although theoretical approaches are instructive, conducting the 
necessary sensitivity tests (with respect to different locations and sizes of 
possible events) and incorporating the inherent uncertainty (with respect to our
limited knowledge of the Earth’s structure), is usually impractical. For example,
hundreds of thousands of dollars were spent on geotechnical site-
characterizationat Turkey Flat, located near the Parkfield section of the San 
Andreas Fault. According to Field and Jacob (1993) “Turkey Flat will be the most
extensively studied sediments field in the word”. However, these authors reached
the conclusion that “the average spectral ratios of earthquake recording provide
a better estimate of the weak motion site response at 
Turkey Flat than do theoretical prediction”. The superiority of empirical 
observationswas also suggested by Seed et al., (1988) in stating, “…it is 
desirable to refine direct measurements of shear wave velocities with data that 
may be obtained from actualearthquake records”.
      Various site-response estimation techniques already have been studied through 
the comparison of earthquakes, or seismic noise recordings. These techniques can be
separated into two categories: those that use at least two sites and those that 
only use one site. Borcherdt (1970) introduced the sediment-to-bedrock ratio (that
is still the most common approach) that consists of dividing the spectrum of a site
by that of a nearby reference site (rock site) using earthquakes. This approach 
identifies the site response function of a site in most cases (e.g., Jarpe et al., 
1988; Darragh and Shakal,1991; Zaslavsky et al., 2000) and is often considered to 
be the most reliable. However, the choice of the reference site remains critical: 
rock sites can have their own site response that can lead to an underestimation 
of the site response (e.g., Steidl et al., 1996; Zaslavsky, 2002). According to
Steidl (1993), when it is possible, the reference ground motion should be 
calculated by averaging several rock sites.
     Several studies (e.g., Kagami et al., 1982, 1986; Tanaka et al., 1968;
Zaslavsky,1987; Katz, 1976; Katz and Bellon, 1978; Ohta et al., 1978; Bard, 1995)
applied theapproach of Borcherdt, using ambient seismic noise instead of 
earthquakes. These studies found that the ratios from long-period seismic noise
(microseisms) and strong motion records produced similar predominant periods. 
Shorter-period noise (microtremor) has also been used to find dominant periods.
This research tool is especially applicable in urban areas where the risk is 
highest but where the high cultural noise level make routine recording of weak
earthquake motions problematic.
     An alternative method, requiring only one recording station, consists of 
dividing the spectrum of the horizontal component by that of the vertical 
component.Nakamura (1989, 2000) introduced the horizontal-to-vertical (H/V)
noise ratio, using noise to estimate site amplification. Most studies show that
the H/V ratio obtained from microtremors coincides with first mode of response
functions of near surface structures to incident shear wave (Ohmachi et al., 
1991; Lermo and Chavez-Garcia,1994; Zaslavsky et al., 1995; Seekins et al., 
1996; Gitterman et al., 1996; Konno and Ohmachi 1998; Mucciarelli and Monachesi,
1998; Chavez-Garcia and Cuenca, 1998;Toshinava et. al., 1997; Shapira et al.,
2001).  In addition, Lachet and Bard (1994)
concluded that the amplitude of the H/V noise ratio is rather sensitive to the 
sourcereceiver distance while being independent of the source excitation 
function. Empirical studies (e.g., Lermo and Chavez-Garcia, 1994; Field and 
Jacob, 1995; Lachet et al. 1996; Seekins et al., 1996; Zaslavsky et al., 2001b)
showed that although the fundamental frequency can be determined with a 
reasonable precision, higher modes couldn’t. A main purpose of our study is to
make quantitative estimates of site effects in the Qiryat Shmona area, which is
located close to a major seismogenic zone, by measuring of ambient noise. 
The area of 25 square kilometers has been analyzed using records at 280 sites
over a six months period. Two ground ambient noise instrumental techniques were
used reference site technique (classical approach) and non-reference site
technique (Nakamura’s technique). We preparing two maps that reflected the 
fundamental characteristics of site effects: distribution resonance 
frequency and maximum relative amplification.
The resulting soil response maps are then discussed. We also compare the 
estimate of site effects as derived from instrumental data for selected sites 
in investigated area to those estimated theoretically from local 1 D models.

     The investigated area of Qiryat Shemona occupies a territory of 6 km long and 
4 km wide between geographical coordinates: South 288.500, North 294.500, West 
202.750, East 206.750. The geological mapping of the region was done by 
Rosenberg(1960), Glikson (1966), Heimann and Ron (1986), Kafri (1991), Sneh  (1998),
and presented in a new version of geological map of Israel to a scale of 1:200,000 
(Sneh et al., 1998). Information about faults in the region was obtained from Ron 
(1987, 1997) and Sneh (1998). Field excursions in the QiryatShemona area were 
conducted prior to the measurements for planning of seismic stations location. 
During our field tripslandslides and talus cones were revealed and contoured. A 
compiled on the basis ofafore-mentioned sources as well as our field observations 
Geological map is presented (see Fig 1).




Figure 1.Geological map and location of the observation points 
 




  

  

    

Tectonic pattern

    


     The sedimentary fill of the basin in the Hula Valley area composed mainly of
lake deposits interbeded with Pleistocene basalts (Picard, 1965). The Qiryat Shemona
investigated area is located in the Metula high tectonic block, close to the Hula 
Valley graben of the structural part of the Dead Sea Rift system and the Beka’a 
syncline, notfar from the border between the eastern Jordan fault and the Tunur 
fault. The evolution of the Hula Pull-Apart graben is attributed to the early 
Pliocene to Pleistocene activity of the Dead Sea transform (Garfunkel et al., 1981).
In the north part of the Hula Valley transform fault splits into five subordinate
faults. At least three of them are of strike-slip character. The main fault, named
the Yammuneh Fault trending northward, is situated to the north of the investigated
area (Heimann, Ron,1987). In accordance with Heimann and Ron in the Qiryat Shemona 
area are located two young fault systems: Shehumit fault and Ma’ayan Baruch fault. 
The Shehumit Fault trends in a north-south direction, and consists of a few segments
arranged in a right-stepping enechelon pattern. The northern tip of each segment is 
bent northeastward and usually continues to the next segment. The northern side of 
each bend is uplifted and forms a local three elevated zone:Shehumit Hill (289.500N),
Barom Hill (292.150 N) and Gidem Hill (294.300N). The Ma’ayan Baruch fault 
system consists of two east-west normal faults. The Tanur fault trends between Roum
and Tel-Hai faults and it is serving the boundary of the Qiryat Shemona graben. At
the hanging side of this graben occur the conglomerates of the Qiryat Shemona Fm., 
at the lying side the loose Quaternary sediments are deposited.  
     Tel Hai fault is located west of the Shehumit ridge in the Qiryat Shemona area 
(Sneh A., 1996) and is concealed by thick alluvium decreasing to the south.






  

  

    

Stratigraphy and lithology

    


      The geological units are presented on the Geological map (see Fig. 1). The
stratigraphy of the northern Naftali Mountains, as first described by Rosenberg
(1960), may be summarized as follows.
      The Lower Cretaceous sequence is of 426 m thick composite of sandstone and 
shale. They are followed by marl, detric limestones and fossiliferous calcarenite
of Aptain age which are followed by the Albian “Knemiceras marl” (Glikson, 1966),
and the Zalmon Fm. (Kafri, 1991). The sandstone of the Kurnub Gr. is presented in 
the southwestern corner of the investigated area. The limestone and marl of the 
Zalmon Fm., as well as dolomite of the Kefira Fm. are found on western border of 
the Qiryat Shemona area. All described above geological units are located at the 
upthrown side of the Roum fault.
     Horst located in the central part of the north area between Tel Hai and
Shehumit faults are covered marls (Mt.Scopus Gr.) of Senonian age ( Kafri ,1991).
The Tanur fault, marked by Ron (1997), separates conglomerate outcropping in the 
northeast part of Qiryat Shemona town and loose alluvial deposits. A series of 
fluvio-lacustrine conglomerates and limestones, and clays more than 300 m thick 
of Neogen age are described by Picard (1952). These series are named according 
to Geological Map of Israel (Sneh, 1998) of the Kefar Gil’adi Fm. In the 
northwestern part of the area investigated Kefar Gil’adi Fm. is represented by 
the Tanur conglomerates composed of subangular to well-round limestone pebbles,
well cemented by calcarenite by Glikson (1966). In the central and northeastern 
parts outcrops of basalt flow covers the Shehumit ridge, which is elevated 60m 
above the surrounding area (Heiman and Ron, 1987). These basalts are distributed
in the northeastern part of Qiryat Shemona area at the Nahal Ayoun. The Hasboni 
basalt was dated recently at 0.88±0.15m.y (Sneh A., 1998). Late Pleistocene 
travertine up to 8 m thick overlaysthe Hasbani basalt with some pebble of 
conglomerates at the bottom (Heiman, 1985).In the investigated area we observe 
two patches of exposed loose sedimentary deposits. First patch is in the Qiryat 
Shemona graben located in the central part of Qiryat Shemona town along main 
road No.90. Second patch is the Hula valley graben located in the southeastern
part of the investigated area (south of the airport). The Notera-3 well 
drilled at central part of the Hula Valley penetrated the Pleistocene 
deposits down to 1675 m (Kashai and Goldberg, 1984). The Pliocene deposits
predominantly consist of basalts flows with tuffs and agglomerates, which could 
becorrelated to the “Cover basalt”. Between the repeated sections of thick 
basaltic rocks occur interbeds of clay, marls, and chalks, occasionally rich 
in peat, and lignite beds and some fresh water limestones. The upper 1200m of
 the penetrated section was deposited during the glacial Pleistocene regime. 
According to Hula-1 boreholesituated in the southeastern part of the area 
investigated three basalt flows of approximately 32, 36 and 27 m thick at 
the depths of 165, 212 and 285 m correspondingly are interbedded with the glacial 
Pleistocene sequence (see Fig.2).Below the soil are deposited the brown clay of 
Holocene age with thickness of 20m. 
     The patch of loose sediments in the Qiryat Shemona graben based on borehole
information: Qiryat Shemona-1 borehole (see Fig. 2), BH-1, 2 borehole (Ezersky, 
1999) and refraction survey data (Ezersky and Shtivelman, 1999) are characterized by
detrial and argillaceous deposits, which forms landslides. These deposits of 10-80 m
thick overlay basalt in the southern part of the area. There are talus cones composed
of loam and talus breccia of 0-20 m thick (by BH-2 borehole data) at the Roum fault 
cliff situated in southwestern side of the area. Field of the loose deposits of 
Qiryat
Shemona graben and Hula Valley graben merge in the southern part of the city and
deepen down to 150 m and deeper.





Figure 2. Lithological sections of Hula-1 and Qiryat Shemona-1 boreholes
 

     Traditional spectral ratio site-response estimate (Kagami et al., 1982)is to 
divide the spectrum observed at the site in question by that observedat a nearby 
reference site (competent bedrock).
              Rk(ω)= |Hs(ω)| / |Hr(ω)|         (1)
where Hs and Hr denote spectral amplitudes of the horizontal components of motion
at the investigated site and those of the reference site, respectively.
If the two sites have similar source and path effects (i.e. when the distance to the
reference site is small compared to the source-site distance) and if the reference 
site has negligible site response, then the resulting spectralratio constitute a 
reliable estimate of the site response.


     Nakamura (1989) developed the technique by formulating three man 
     hypotheses:
     1.Ambient noise is generated by reflection and refraction of shear 
       waves within superficial soil layers and by surface waves.
     2.Local superficial sources of noise do not affect ambient noise at
       the bottom of the unconsolidated structure.
     3.Soft soil layers do not amplify the vertical component of ambient
       noise.
     Hence, site response can be expressed as the spectral ratio of the 
     horizontal and vertical components of ambient noise at the surface,i.e.

              Sn(ω)=|Hs(ω)| / |Vs(ω)|         (2)
     where horizontal component of microseisms (Hs) and vertical component 
     of microseism (Vs) recorded at the same location.

     During the summer and fall of 2002 in order to determine the microtremor
characteristics at different sites in Qiryat Shemona. 285 of field deployments 
of the portable seismic instruments were carried out.
     We planned the grid of measurement points of 250m*250m increasing a density
at sedimentary deposits and decreasing it in places of hard rock outcropping.
Distribution of investigated sites is shown in Figure 1. In Table 1 we presented
coordinates of seismic station locations and time of measurements.
     Ground motions were recorded using a multi-channel, PC-based, digital 
seismic data acquisition system (see Shapira and Avirav, 1995) designed for site 
response field investigations. The seismometers used were sensitive velocity
transducers with a natural frequency of 1.0 Hz and damping at 70% of critical. 
Each of the stations was equipped with one vertical and two horizontal seismometers 
(oriented north-south and east-west). The seismometers installed on leveled metal
ground plates and connected to the data acquisition system via cables. The 
acquisition equipment included: 12-channel amplifier with a band pass filter 
0.2-25 Hz, GPS (for timing) and a laptop computer with analog-to-digital 
conversion card. Digital recordings were made with a sampling rate of 100 samples 
per second. The recorder has 16-bit data word. Prior to and during the 
measurements we checked and determined the transfer functions of the 
instrumentation in order to facilitate transformation of the recorded signals 
into ground motion data, i.e., particle velocity. The instrument characteristics
of the stations are summarized in Table 2. All seismometers were wired directly
to the recording sites. In the Figure 3 we present the pictures of the locations
of the seismic stations during the site investigations.





Figure 3. Location of the seismic stations during the site investigations

Based on many previous site investigations, we concluded that a window length of
25-30 sec for spectral calculations is sufficient to provide stable results. The 
selected time windows were Fourier transformed using cosine tapering before 
transformation. The spectra were then smoothed with a triangular moving Hanning 
window (0.4 Hz).After data smoothing and in order to obtain spectral ratios, the 
spectra of an EW or NS channel at a site were divided by the spectra of the 
corresponding channel of a reference site (Kagami ratios) and averaged:

A(f)=
=1/2n[(i=1..n)(SNS(f)isite/SNS(f)irock)+(i=1..n)(SEW(f)isite/SEW(f)irock)] (3)
or spectra of an EW or NS channel at a site were divided by the spectra of the 
vertical channel at the same site (receiver function and Nakamura estimate). 
The arithmetical average of each individual ratio was also computed:

A(f)=
   =1/2n[(i=1..n)(SNS(f)i/SV(f)i)+(i=1..n)(SEW(f)i/Sv(f)i)]    (4)

  
     Separately we present in Table 2 information on time of records and locations
of seismic stations where microtremor measurements were carried out by reference
technique. In the Table 3 are shown instrument characteristic of the seismic 
stations.

     The determinations of the site effects using ambient noise data were carried
out by applying two techniques. One of them is horizontal-to-vertical spectral ratio
and other the sediment-to-bedrock spectral ratio.

Figure 4. Individual and average H/V spectral ratios for different
 kind of rock sites: (a) Limestone, Point 32; (b) Basalt, Point 183; 
(c) Sandstone, Point 155; (d) Travertine, Point 121; (e) Marl, Point 143;
 and (f) Conglomerate, Point 33. Black line is average spectral ratio.


     Individual and average ambient noise H/V spectral ratios for six different rock 
sites are shown in Figure 4.The scatter between individual curves is high but the
variations in the averaged functions are small. All curves are approximately flat 
in thefrequency range 0.6 to 10 Hz. The bedrock data illustrate that Nakamura’s 
method of ambient noise analyzing failed to produce a peak at stations where we did
not expectone. It must be mentioned that results obtained on bedrock this is not 
trivially,because the near-surface weathering and cracking of the bedrock may be 
affects the recorded ground motion at frequencies of engineering interest, even at
sites appear to be located on competent crystalline rock. For example, data 
collected from earthquake and ambient noise on weathered and cracked granite bedrock
to the north of Eilat city show a spectral amplification in the frequency range 6 to
7 Hz, with a factor about 4 (Zaslavsky et al., 2002). On the other hand, results 
obtained in the Dead Sea rift area (Zaslavsky et al., 2000) show that marl overlying
a dolomite can have a site response of their own. In another locations the frequency
-dependent character of site response may be obtained for travertine and 
conglomerate, too.







Figure 5. Examples of average Fourier spectra and H/V 
spectral ratios in different sites providing true evaluation
 of site effects: (a) Point 19A-1, (b) Point 169-1, and (c) 
Point 44. Black line is average spectral ratio.




     Our observations revealed consistent and clear peaks of spectral ratios, which
we could trace from site to site. We would like to demonstrate three cases that may
produce them. Figure 5a illustrates the character average spectra of microtremors
recorded at Point 19A-1 and its horizontal-to-vertical spectral ratios. An increase in
the spectral levels of the horizontal components is clear in frequency near 5.0 Hz,
while the spectrum of vertical component is flat. Therefore, spectral ratios show a
prominent peak at about 5.0 Hz with an amplification of about factor 7. In the second
case (Fig. 5b), if we compare the average spectra horizontal and vertical motions at
Point 169-1, we can see that in the vertical spectrum there is a narrow-bandwidth 
trough at frequency near 1.0 Hz. Hence, the general character of the spectral ratios 
is clear amplification at a frequency of about 1.0 Hz. Consequently, the high levels 
of amplification obtained from H/V spectral ratios are controlled not only by peaks in 
the spectra of the horizontal components but also by narrow-bandwidth "holes" (troughs)
in the spectra of the vertical components. In third case (Fig. 5c) the general 
character of the horizontal Fourier spectra at Point 44 is a gradual amplification 
of high-frequency energy starting at the frequency of 5 to15 Hz, while energy of 
verticalcomponent in this frequency range is very low. We can see that using the 
vertical Fourier spectra as reference for Point 44 we have following site resonance 
estimates:resonant frequency is 8.0 Hz with amplification factor about 8.






Figure 6. Sequential expression of H/V spectral ratios along line
 “A-A” in the southeast part of the investigated area




     In Figure 6 we depict the H/V spectral ratio of some site closely located along 
line "A-A" (for the position see Fig. 1). The fundamental resonance frequency is
clearly shifted from south to north with decreasing sediments thickness. This is fact
that for distance of two kilometers resonance frequency increases from 0.7 Hz to
5 Hz.Some recent studies (Yamanaka et al., 1994, Ibs-von Seht and Wohlenberg, 1999,
Zaslavsky, 2001, Zaslavsky et al., 2002) showed that ambient noise measurement
could be used to map the thickness of soft sediments. However, now we have not 
acquired available borehole logging data and, therefore, we have not been able to
derive a calibrated relationship between resonance frequency and thickness of the
sediments. On the other hand, we have not information about subsurface distribution
of shear wave velocities obtained from seismic refraction surveys. Therefore, we
could not calculate the thickness of sediments for all sites in this area.








Figure 7. Individual and average (black line) H/V spectral 
ratios for points located on scarp slope of talus.




     In south-west part of the Qiryat Shemona town (see Fig. 1) there are landslides
(talus). Estimation of local amplifications obtained on the talus is displayed in 
Figures 7 and 8. For Points 1, 2, 14 and 155 (Fig. 7) located on the scarp slope of 
the talus (angle of dip about 400) the H/V ratios do not present a peak, though the 
alluvium deposits according to the BH-1, 2 boreholes have thickness of 20 m, and the
impedance contrast is relatively strong (up to 3.0-3.5). Points 5, 15, 29, 102, 103 
and 104 located in the foot of talus on platform exhibit the clear amplifications by 
a factor around 3.5-5.0 in the frequency range 2.5-5.5 Hz (Fig. 8).









Figure 8. Individual and average (black line) H/V spectral 
ratios for points located at foot of talus





It is important to be reported that we are not known S-wave velocity structure of
surface, therefore these results (resonance frequencies and amplifications) not 
can be calibrated with theoretical model though it is very important from 
engineering point of view.




Figure 9.Examples of average H/V spectral ratios
 for points located on sediments of Qiryat Shemona town.


     Figure 9 shows examples of H/V spectral ratios for sites on alluvial sediments 
in Qiryat Shemona town. As shown from Figure 6, in general, sites have wide range 
values of amplifications (from 3.5 to 7.0) in wide range of frequencies between 1.5 to
8.0 Hz. Significant variations of site effects it is quite possible to explain a wide
variety of the underground structure. On other hand, we observe an agreement in the
shape of NS and EW components for the ratios at sites. This suggests that for
theoretical site response function in the Qiryat Shemona may be to use 1D models.
The extensive database of microtremor recording can be used to obtain 
information about the morphologies of the hard rock basement. Examples of the H/V 
spectral ratios of ambient noise at sites located on both sides of Shehumit fault in
southern part of investigated area are displayed in Figure 10.









Figure 10. Individual and average (black line) H/V spectral ratios 
for points located at the both sides of the Shehumit 
fault in its south part 





All the spectral ratios at Points160, 161, 162 (to the west of the fault) show 
similar characteristics,having a singlespectral ratio peak in a frequency 0.7 
Hz with amplification about 5. 
The spectral ratios at Points 113, 114 and 159 (to the east of the fault) have similar
amplifications at the frequency 1.2 Hz. The fair agreement with amplification level of
west and east parts of fault allow to suppose that the velocity configuration of
sedimentary layers and basement (half space) is the same for two parts. Therefore, we
believe that difference in the main peak of the H/V ratios obtained systematically may
be dueto a thickness of soft sediments. If to take into account (Frieslander, 2002) 
that the sediment thickness to the east of fault gradually increases, vertical shift 
in the basement may reach of 50-70 m.
     Fundamental resonance frequencies and maximum amplification factor at all
measurement sites across the Qiryat Shemona area are summarized in Table 4.

     The relationship between microtremor (ambient noise) characteristics and 
local site response has been a topic of research for many years (Kanai and 
Tanaka, 1954),because noise measurements are much easier than earthquake 
observation. The ratio of the horizontal components of the velocity spectra 
at the sediment site to those at the rock site (Kagami et al., 1982, 1986) 
can be used as a measure of site effects.The main problem of this method 
(hereafter referred to as H/H ratio) is a problem regarding the areas over 
which the microtremor H/H ratios should be used.Because noise is generated 
by human activities, especially in urban and suburban regions, the intensity
of noise source seems to significantly vary from place toplace. On the 
otherhand, the validity of noise ratios is based on an assumption that the
intensity of noisesource is the same between sites.
     The field observations were conducted in the following manner using two 
stations. One station permanently located at a “reference point” to monitor the
variations of ambient noise throughout the duration of the experiment. The other
station was moved from point to point to record ambient noise along the 
observation lines. The maximum length of the observation lines was 0.4 km. 
An hour and a half noise recording was made at each observation point. The 
overallobservation schedule is given in Table 2.




Figure 11. Individual and average (black line) H/V 
spectral ratios for reference Points 169R obtained during 
different times of day




     In Figure 11 we present the H/V spectral ratios for reference Point 169R 
recorded during different time of measurements at different sites. Analyzing 
the sixestimates of site effects we can see that the scatter between 
individual curves is high but average spectral ratios haven’t frequency-
dependent peak that characterizes of siteeffect. Moreover, in frequency 
range 0.6 to 3 Hz all average H/V spectral ratios are relatively flat, with
unit amplification. Hence, the result based on the H/V ratio technique 
forstations 169R Line suggest that this station may serve as a reference
station. The “amplification” in the frequency range 5 to 10 Hz is difficult 
to explain.One possibility is that cultural features of some sort have 
affected the signal invertical and horizontal component in this frequency
range.




Figure 12. Comparison between the average H/V 
spectral ratios and the average H/H ratios obtained from
 noise motion for stations 169R Line. Red line is H/H ratio,
 and black line is H/V ratio

Figure 12 shows the comparison between the average H/H and H/V spectral ratios 
for different points along Line 169R. The observation results can be 
interpreted as follows. In all six cases, the fundamental frequency and 
amplification factor obtainedfrom the noise H/V average spectral ratios are in
good agreement with H/H average spectral ratios. It is important to note that 
there are no significant differences in shapes of spectral ratios. The marked
rise in the H/V spectral ratios observed in the frequency near 0.3 Hz is 
normally associated with ocean waves. Therefore, in the H/H spectral ratios 
this peak is negligible.






Figure 13. Results obtained for stations Line 237R: 
(a) Average H/V ratios for reference Points 237R obtained 
during different times of day. (b) Comparison between the average H/V
 spectral ratios and the average H/H ratios with respect to reference 
points. Red line is H/H ratio and black line is H/V ratio. 

     In Figure 13 are presented results obtained from measurements at stations 
along the 237R Line. As Figure 10 demonstrates the variation of the average H/V 
spectral ratios obtained in the different time of measurement at reference point 
is high but these spectral ratios does not indicate amplification effects across
the frequency range 0.6 to 10 Hz. We think that this point can be used as the 
reference motion. In Figure 11b-d, we show H/V and H/H spectral ratios for 
different station located at 237R line. Between these two different windows, 
frequencies of the dominant peak of H/V ratio coincide with H/H ratio and the 
overall shapes of two ratios have roughly some similarities. On the other hand,
maximum amplification of H/V ratios is smaller than the one of H/H ratios and 
frequency range of maximum amplification of H/H ratio is noticeably wider than
H/V ratio. The difference between the two spectral ratios may be caused by two
reasons. First, the density of microtremor sources at sites higher than at 
reference point. Second, reference station 237R located on conglomerate, 
while investigated sediments overlying basalt.





Figure 14. Results obtained for stations Line 222R:
 (a) Individual and average H/V ratios for reference Points 222R.
 (b) Comparison between the average H/V spectral ratios and the 
average H/H ratios. Red line is H/H ratio and black line is H/V ratio 

    In Figure 14a we show H/V spectral ratios for reference Point 222R. From the 
average function it may be concluded that in the frequency range of 0.2 to 3.0 Hz 
there is no site amplification effect and this station may serve as a reference 
station in frequency range, as already mentioned. Figure 14b shows the comparison 
between H/V and H/H average spectral ratios for station 222R Line. These plots 
demonstrate a similarity between the various estimates, implying resonance 
frequency and amplification.

     After processing of each record to obtain main parameters of site response,
themaps of fundamental frequency and amplification factor were produced to show
its spatial variation over the region of interest (see Figs. 15, 16). The main 
geological structure is reflected in the measurements results. We obtained 
almost flat response function with no amplification for sites where basalt, 
conglomerates, limestone, marl sandstone and travertine are outcropped. 
Response functions of the soil sites exhibit peaks at dominant frequencies 
between 0.7 to 8 Hz. The area where we determined site effect can be separated
into two zones. First zone, the Qiryat Shemona town proper, is located at the 
Qiryat Shmona graben limited by the Roum fault in the west, by Tanur fault in 
the northwest and by Tel Hai fault in the east. This zone is characterized by 
predominant frequency from 2.5 to 4.5 Hz in the northern and southern parts of
the town. In the central part of the town predominant frequency is increasing 
up to 6.5-8.5 Hz. Such variability in the dominant frequency may be explained
in first approximation by the alteration in the depth of the basalt basement
from 10 up to 50 m. In the southern part of the town loose deposits of Qiryat
Shemona graben and Hula Valley graben merge and deepen. Corresponding falling
down of basalt basement to the south provides the decreasing of the fundamental
frequency to 0.9 Hz. Significant increasing of amplification up to factor 8 in
the central part of Qiryat Shemona town is probably connected with 
heterogeneities of loose alluvium sediments deposited at the trough axes of the
Qiryat Shemona graben. Maps of fundamental frequency and amplification factor
combined with topographic map within Qityat Shemona town are displayed in Fig.17.
     Second zone where site effects were observed we distinguished in the 
southern and southeastern parts of the area investigated. This zone is attached
to the northwestern part of the Hula basin graben. Gradual decreasing of resonant
frequency from 6 Hz to 0.7 Hz that we detected from the north to the south is 
correlated with dipping of cover basalts from the surface in the north of the zone
to the depth of 300m and deeper in the south. The sediments overlaying basalt and
producing site response are represented by Holocene-Pleistocene deposits. A sharp
resonant frequency alteration across the line with NS coordinate of 204500 of 1 km
long is observed in the southern part of the area investigated. These observations
previously described (see Results and Discussion, Fig.7) agree with geological data
on the Shehumit fault (Heiman, 1987).





Figure 15. Distribution of fundamental
 frequency in the investigated area 
 





Figure 16. Distribution of maximum amplification factor in the 
investigated area 
 





Figure 17. Contour maps showing: (a) distribution of fundamental frequency, 
(b) distribution of maximum amplification and 
(c) topography of the Qiryat Shemona town.
 

     Analysis of the observed differences, when comparing the analytical results 
with previous empirical studies, provides a useful feedback to establish site 
dependence suitability and reliability of methods. Soil-column models based on 
interpretation of the shear-wave data, existing geotechnical and borehole information,
were developed only for limited sites, because we have the very small amount 
available data. 
     Hula-1 borehole situated in the southeast part of the area investigated with 
geographical coordinates NS 206.206 EW 288.864 was drilled to 697 m depth. Its 
subsurface structure is characterized by interbedded layers of Quaternary sediments 
(clay, silt) with gravels (see Fig. 2). Three cycles of basalt layers are found at the
depths of 165, 212 and 285 meters, respectively. Each of these basalt layers may be 
considered a half-space. 
     In order to construct 1D model we assumed required as input parameters shear-
wave velocity for each layers based on geotechnical data collected from previously 
published reports (Ezersky, 1999; Ezersky and Shtivelman, 1999; Zaslavsky et al., 
2001). Geotechnical parameters used to calculation of response functions at site 86 
near the Hula-1 borehole are summarized in Table 5.
     We estimated the shear wave velocity averaged for the layers overlying the half-
space by relation 
        ___
        VS = (i=1..n)hi/(i=1..n)(hi/Vsi)     (5)

where Vsi is the shear wave velocity of any one of the layers between 1 and n, hi is
the thickness of any layer.
     Value of averaged S-wave velocity, we obtained, is 770 m/sec. In the other 
hand average P-wave velocity by reflection survey data (ôøéæìðãø åîãáãééá, 2002)
is 1820 m/sec. Consequently Vp/Vs ratio is 2.4. This value is very close to the one 
obtained in refraction survey carried out at the Qiryat Shemona site (Ezersky, 1999). 
Further we computed using the Joyner’s programs (1977) the analytical response 
function and compared its with the empirical estimations. In initial model, we tried
to use as a half-space upper basalt layer bedding at a depth of 165 m (see Fig. 2).






Figure 18. Analytical transfer functions for Point 86 calculated
 using two different basalt layers as a half-space. 
Green line – for basalt layers at a depth of 165 m; red line – at a depth of 285 m.

     If we examine the theoretical transfer function of the trial model for Point 86 
situated at Hula-1 borehole (green line on Fig. 18) we can see that maximum 
amplification appears at frequency 1 Hz while the fundamental frequency at 
measurement sites surrounding the Hula borehole is 0.7 Hz. In attempt to achieve 
better agreement with experimental data we recalculated the transfer function using 
the basalt layer at a depth of 285 m as a half-space. This trial yielded much more 
satisfying result, namely fundamental frequency is about 0.7 Hz (red line on Fig. 18).
Comparing the results of modeling with depth section along reflection of seismic 
survey displayed in Figure 19 we can see that defined by our model depth of half-
space corresponds to the lower layer marked on the depth profile.

     To verify our assumption concerning the reflector depth we estimated for 
Point 82 located 1 km northwest from the Hula-1 borehole along the GP0170 line. 
Observed resonant frequency at this site is 0.8 Hz. It is known that for simple 1-D
 of a simple layer over a half-space the average shear-wave velocity (Vs) could be
 directly calculated from measured resonant frequency of this layer (fr) and its 
thickness (H) by equation
                         Vs= 4H * fr       (6)

From the assumption that reflector traced by reflection survey (see Fig. 19) appears
at a depth of 140 m we obtained Vs is about 450 m/s, and correspondingly Vp/Vs 
ratio is 4. It is difficult to imagine such a striking change (from 2.4 to 4.4) 
for a distance of 1 km and ratio itself does not seem realistic for actual geological 
conditions.





Figure 19.  Comparison between the top of Hasbani basalt sequence
 derived from reflection survey (Frieslander and Medvedev, 2002) 
and depth of half space obtained from noise measurements 

	The procedure of model development was repeated for some sites situated 
along the reflection line GPO170, where the ambient noise measurements were
carried out. The morphology of the half-space structure constructed on the basis of 
these measurements is displayed on Figure 19. It is quite possible that between points
199 and 85 we observe the vertical shift in the basalt layer roughly 50 m. For Point 
92 located to the west from the horst using lower basalt layer as a half-space we 
obtained the fundamental frequency of 0.9 Hz while experimental fundamental frequency 
characteristic for this site is 2.7 Hz. A good coincidence of both frequency and 
amplification factor with the empirical ones we obtained only using the first basalt 
layer as a half-space.
	We would like to emphasize that our estimations of depth of half-space were 
based on extremely poor data on velocity structure in the area investigated. 
Experimental determination of shear wave layer velocities could enable direct 
calculations of sediment thickness estimates from the frequency of main peak in the 
microtremor spectral ratios.

      The high number of noise measurements performed in the Qiryat Shemona 
area allowed us to infer the following conclusions:
      1. Ambient noise measurements proved to be valuable tool to determine 
dominant frequency of motion at a soft soil site with good reliability estimate of 
amplification level in different geological conditions and elastic behaviors. Our 
results support the use of these measurements on microzonation studies, which could
 be important on urban and land planning. However, a few details are not yet well 
clarified, and further theoretical research is needed for a better understanding 
of some problem, such as the interpretation of the higher frequencies and 
amplifications.
      2. The comparison between the site effects observed from horizontal-to-
vertical spectral ratios and inferred from sediment-to-bedrock using noise data show 
generally consistent results. However, experimental study of site effect by sediment-
to-bedrock spectral ratio can be successful only under particular circumstances 
because noise is generated by human activities, especially in urban and suburban 
regions and the intensity of noise source seems to significantly vary from place to 
place. On other hand, the validity of noise ratios is based on an assumption that the 
intensity of noise source is the same between sites. Therefore, the H/H ratio should
be used within limited area (a diameter of several hundred meter).
      3. As demonstrated in this study, the site response variations may be 
significant over short distances, thus we strongly recommend that prediction of 
different seismic shaking characteristics during large earthquakes should be based on 
the experimental site response functions obtained over a relatively dense gird of 
measurement points.
      4. Ambient noise measurements in combination with Nakamura's method can 
be a powerful tool to map sedimentary cover layers. Particularly in regions of 
unknown basement morphology, such a procedure may by a way to quickly obtain a 
general idea of the subsurface structure. The requirement for obtaining quantitative 
thickness values is either knowledge of the velocity depth structure of the shear wave.
The ambient noise survey expense is relatively small because only one seismic 
recorder is required and data processing routines are standard.
      5. It would be recommended that multidisciplinary studies including 
seismological, geophysical, and geological observations be conducted to validate 
results obtained in present study. This knowledge allowed the computation of 
theoretical seismic response, which is very important to prediction of ground motion 
to be expected in the occurrence of a large event. The application of this methodology
is very important for investigated area where big earthquake present along return 
period, but which exhibit a high seismic risk according to historical reports.
      6. The site response functions that were discussed in this study are associated 
with weak motions at the range where the behavior of the soils is linear. In this 
respect, the site response functions do not represent the site effects under strong 
ground motions that cause the soils to behave nonlinearly. However, based on the 
results presented in the above the nonlinear site response can be determined by 
different analytical models. 

Bard, P.-Y. and. Tucker, B. E, 1985. Underground and ridge and site effects: 
     comparison of observation and theory, Bull. Seism. Soc. Am. 75: 905-922.
Borcherdt, R.D., 1970. Effects of local geology on ground motion near San Francisco 
     bay, Bull. Seism. Soc. Am., 60: 29-61.
Chavez-Garcia, F. J., and Cuenca, J., 1998. Site effects and microzonation in 
     Acapulco, Earthquake Spectra, Vol. 14, No. 1: 75-93.
Darragh, R.B., Shakal, A.F., 1991. The site response of two rock and soil station
     pairs to strong and weak motion, Bull. Seismol. Soc. Am., 81: 1885-1889.
Ezersky M., 1999. Seismic borehole studies at the Qiryat Shemona, Beit Shean and 
     Hagoshrim sites, GII Report No.803/35/99.
Ezersky, M. and Shtivelman, V., 1999. Seismic refraction surveys for updating the 
     microzonation map of Israel, GII Report No. 802/43/99.
Field, E.H., and Jacob, K.H., 1993. Monte-Carlo simulation of the theoretical site 
     response variability at Turkey Flat, California, given the uncertainty in the 
     geotechnically derived input parameters, Earthquake Spectra, V 5, No. 9: 669-
     701.
Field, E. H., and Jacob, K. H., 1995. A comparison and test of various site-response 
     estimation techniques, including three that are not reference-site dependent, 
     Bull. Seism. Soc. Am. 85: 1127-1143.
Garfunkel, Z., Zak, I. and Freund, R., 1981.Active faulting in the Dead Sea Rift, 
     Tectonophysics, 80: 1-26. 
Gitterman, Y., Zaslavsky, Y., Shapira, A. and Shtivelman, V., 1996. Empirical site 
     response evaluations: case studies in Israel, Soil Dynamics Earthquake 
     Engineering 15: 447-463.
Glikson Y., 1966. The lacustrine in the Kefar Gil’adi area, Northern Jordan valley, 
     Israel journal of Earth-Sciences, vol. 15: 85-100.
Heimann, A., and Ron H., 1987. Young faults in the Hula Pull-Apart Basin, central 
     Dead Sea Transform, Tectonophysics, 141, No. 1-3, September: 117-124.
Ibs-von Seht, M., and Wohlenberg J., 1999. Microtremor measurements used to map 
     thickness of soft sediments, Bull. Seism. Soc. Am. 89: 250-259.
Jarpe, S.P., Cramer, C.H., Tucker, B.E., and Shakal, A.F., 1988. A comparison of 
     observations of ground response to weak and strong motion at Coalinga, 
     California, Bull, Seismol. Soc. Am., 78:421-435.
Kafri U., 1991. Litostratigraphic and lithofacial map northern Naftali Mountains 
     (GSI-1991).
Kagami, H., Duke C.M., Liang, G.C., and Ohta, Y., 1982. Observation of 1- to 5-
     second microtremors and their application to earthquake engineering. Part II. 
     Evaluation of site effect upon seismic wave amplification deep soil deposits, 
     Bull. Seism. Soc. Am. 72: 987-998.
Kagami, H., Okada, S., Shiono, K., Oner, M., Dravinski, M., and Mal, A. K., 1986. 
     Observation of 1- to 5-second microtremors and their application to earthquake 
     engineering. Part III. A two-dimensional study of site effects in the San 
     Fernando Valley, Bull. Seism. Soc. Am. 76: 1801-1812.
Kanai, K., Tanaka, T., and Osada, K., 1954. Measurement of the micro-tremor. I, 
     Bull. Earthquake Res. Inst., Tokyo Univ. 12: 192-210.
Kashai, E. and Goldberg Y., 1984. Notera-3, Geological Completion Report. OE(i)L 
     Report No.84/20.
Katz, L. J., 1976. Microtremor analysis of local geological conditions, Bull. Seism. 
     Soc. Am. 66: 45-60.
Katz, L. J., and Bellon, R.S., 1978. Microtremor site analysis study at Beatty, 
     Nevada,Bull. Seism. Soc. Am. 68: 757-565.
Konno, K., Ohmachi, T., 1998. Ground-motion characteristics estimated from spectral 
     ratio between horizontal and vertical components of microtremors, Bull. Seism. 
     Soc. Am. 88, 228-241.
Lachet, C., and Bard P-Y., 1994. Numerical and theoretical investigations on the 
     possibilities and limitations of the Nakamura’s technique, J. Phys. Earth, 42: 
     377-397.
Lachet, C., Hatzfeld, D., Bard, P-Y., Theodulidis, N., Papaioannou, Ch., and 
     Savvaidis, A., 1996. Site effects and microzonation in the City of Thessaloniki 
     (Greece). Comparison of different approaches. Bull. Seismol. Soc. Am., 86 (6): 
     1692-703.
Lermo, J., and Chavez-Garcia, F.J., 1994. Are microtremors useful in site response 
     evaluation? Bull. Seismol. Soc. Am., 84: 1350-1364.
Mucciarelli M., Monachesi, G., 1998. A quick survey of local amplifications and their 
     correlation with damage observed during the Umbro-Marchesan (Italy) 
     earthquake of September 26, 1997. Journal of Earthquake Engineering, vol. 2, 
     No 2: 325-337.
Nakamura, Y., 1989. A method for dynamic characteristics estimation of subsurface 
     using microtremor on the ground surface, QR of RTRI, Vol. 30, No.1: 25-33.
Nakamura, Y., 2000. Clear identification of fundamental idea of the Nakamura’s 
     technique and its applications, 12th World Conference of Earthquake 
     Engineering, Auckland, New Zeeland, Proceedings, No. 84:8.
Ohmachi, T., Nakamura, Y., Toshinawa, T., 1991. Ground motion characteristics of 
     the San Francisco bay area detected by microtremor measurements. Second 
     International conference on recent advances in geotechnical earthquake 
     engineering and soil dynamics, St. Louis, Missouri, 1643-1648.
Ohta, Y., Kagami, H., Goto, N., and Kudo, K., 1978. Observation of 1- to 5-second 
     microtremors and their application to earthquake engineering. Part I: 
     Comparison with long period accelerations at the Tokachi-Oki earthquake of 
     1968, Bull. Seism. Soc. Am. 68: 767-779.
Picard L., 1958. The Geology of Kefar Gil’adi ,in Zalman Lif Memorial Volume 
     Israel Exploration Society, 2 Jerusalem.
Picard L., 1965.The geological evolution of the Quaternary in the Central-Northern
     Jordan Graben, Israel.Geol.Soc.Am., Spec.Pap., 84:337-366.
Rosenberg E, 1960, Geologische Untersuchungen in den Naftalibergen , Inaug, PhD 
     Thesis, Zurich.
Rotstein Y., and Bartov Y., 1989. Seismic Reflection Across a Continental 
     Transform: An Example from a Convergent Segment of the Dead Sea Rift.1989, 
     Journal of Geophysical research, vol. 94, NO.B3: 2902-2912.
Seed, H.B., Romo, M.R., Sun, J.I., Jaime, A., and Lysmer, L., 1988. The earthquake 
     of September 19, 1985 – relationship between soil conditions and earthquake 
     ground motion, Earthquake Spectra, 4: 687-729.
Seekins, L.C., Wennerberg, L., Margheriti, L., and Liu, H-P., 1996. Site 
     amplificationat five locations in San Francisco, California: a comparison 
     of S waves, codas and microtremors, Bull. Seismol. Soc. Am., 86:627-635.
Shapira, A., Feldman, l., Zaslavsky, Y., Malitzky, A., 2001. Application of a 
     stochastic method for the development of earthquake damage scenarios: Eilat, 
     Israel test case, The Problems of Lithosphere Dynamics and Seismicity, 
     Computational Seismology, 32, 58-73.
Sneh A., 1998. The Dead Sea Rift: Lateral displacement and downfaulting phases, 
     Tectonophysics, 263: pp 277-292.
Sneh A., Bartov Y. and Rosensaft M., 1998. Geological Map of Israel 1: 200,000. 
     Geological Survey of Israel, Sheet 1.
Steidl, J, 1993. Variation of site response at the UCSB dense array of portable 
     accelerometers, Earthquake Spectra 9:289-302.
Steidl J., Tumarkin, A. and Archuleta R., 1996. What is reference site? Bull. Seism. 
     Soc. Am. 86:1733-1748.
Tanaka, T., Kanai, K., Osada, K., and Leeds D. J., 1968. Observation of 
     microtremors. XII. (Case of the U.S.A.) Bull. of the Earth. Research Inst. Vol. 
     46:1127-1147.
Toshinawa, T., Taber J. J., and Berrill, J. B., 1997. Distribution of ground motion 
     intensity inferred from questionnaire survey, earthquake recordings, and 
     microtremor measurements- a case study in Christchurch, New Zealand, during 
     1994 Arthurs pass earthquake. Bull. Seism. Soc. Am. 87: 356-369.
Yamanaka, H., Takemura, M., Ishida, H., and Niva, M., 1994. Characteristics of long-
     period microtremors and their applicability in exploration of deep sedimentary 
     layers, Bull. Seism. Soc. Am. 84:1831-1841.
Zaslavsky, Y., 1987. Analysis microtremor at the Dushanbe seismological 
     observatory, Instruments and Method of Earthquake registration (Moscow), 152-157
Zaslavsky, Y., Gitterman, Y. and Shapira, A., 1995. Site response estimations in 
     Israel using weak motion measurements, The fifth International conference on 
     Seismic Zonation, Nice, France, 1713-1722.
Zaslavsky, Y., Shapira, A., Arzi, A.A., 2000. Amplification effects from earthquakes 
     and ambient noise in Dead Sea Rift (Israel), Soil Dynamics and Earthquake 
     Engineering, V. 20/1-4: 187-207.
Zaslavsky, Y., 2001. Variation of site response and thickness of sediments in Yagur 
     fault zone detected by microtremor measurements, Annual Meeting, Geological 
     Society of Israel.
Zaslavsky, Y., Gorstein, M., Giller, V., Livshits, I., Giller, D., Dan, I., Shapira,
     A.,Leonov, J., and Peled, U., 2001. Microzoning of the earthquake hazard in 
     Israel. Seismic microzoning of Lod and Ramla. GII Report 569/143/01.
Zaslavsky, Y., Shapira, A., and Arzi, A. A., 2002. Earthquake site response on hard 
     rock – empirical study, Proceeding of 5th International Conference on analysis
     of Discontinuous Deformation, ICADD, 6-10 October 2002, Beer-Sheva, Israel 
     133-144.
 
        øåï, ç., ùîéø, â., àééì, é., 1997
 îòååúéí ááìå÷ îùâá-òí-îøâìéåú áéï ùáø øåí ìùáø îøâìéåú
       ëðñ äçáøä äâéàåìåâéú, îãøéê ñéåøéí, 33-36
ôøéæìðãø,à.  ñ÷ø  øôì÷öéä ñééñîéú     áäôøãä âáåää áàæåø 
    ëôø éåáì – ëôø âìòãé. ãå"ç îâ"é 643/100/96  
ã"ø à. ôøéæìðãø åá. îãáãééá,   ñ÷ø øôì÷öéä áäôøãä   
                 ã"ø à. ôøéæìðãø åá. îãáãééá
     ñ÷ø øôì÷öéä áäôøãä âáåää áàæåø ÷éøéú - ùîåðä ãå"
             ç îâ"é áéðééí îñ'' 398/243/02
ùôéøà, à., æñìáñ÷é, é., âéèøîï, é.,   1994
        îç÷øé úâåáú àúøéí åñô÷èøåí úâåáä á÷øéú ùîåðä
          D12/527/93(2) ãå"ç îâ"é îñëí îñé