International Symposium on Seismic Risk Reduction
The JICA Technical Cooperation Project in Romania
JICA Project – Paper number 05
NCSRR DIGITAL SEISMIC NETWORK IN ROMANIA
A. Aldea1, N. Poiata2, T. Kashima3, E. Albota1, S. Demetriu1
SUMMARY
Digital seismic instrumentation donated by Japan International Cooperation Agency (JICA) to the National
Center for Seismic Risk Reduction (NCSRR, Romania) allowed the installation in 2003 of a new Romanian
seismic network. In 2005-2006 the network was developed by investments from NCSRR within the budget
ensured by Ministry of Transports, Construction and Tourism (MTCT). The NCSRR seismic network contains
three types of instrumentation: (i) free-field stations - outside the capital city Bucharest (8 accelerometers),
(ii) instrumented buildings - in Bucharest (5 buildings), and (iii) stations with free-field and borehole
sensors - in Bucharest (8 sites with ground surface sensor and sensors in 15 boreholes with depths up to
153m). Since it's installation, the NCSRR network recorded more than 170 seismic motions from 26
earthquakes with moment magnitudes ranging from 3.2 to 6.0. The seismic instrumentation was accompanied
by investigations of ground conditions and site response: PS logging tests, single-station and array
microtremor measurements. The development of seismic monitoring in Romania is a major contribution
of JICA Project, creating the premises for a better understanding and modelling of earthquake ground
motion, site effects and building response.
INTRODUCTION
Seismic activity in Romania is due to Vrancea intermediate-depth subcrustal source (focal depth between 60
and 170 km), and to several shallow crustal sources (Banat, Fagaras, etc.). Vrancea source dominates seismic
hazard not only in Romania but also in Republic of Moldova and affects large areas in Bulgaria and Ukraine.
Strong Vrancea subcrustal earthquakes produced significant damage and victims in Romania and neighbouring
countries.
Worldwide, three major ways are used for reducing seismic risk: (i) better design codes accompanied by more
performing construction methods and high quality construction materials, (ii) seismic rehabilitation of existing
buildings, and (iii) disaster education of population accompanied by intervention preparation of authorities.
The first two ways are strongly supported by the acquisition of strong seismic ground motions, the development
of seismic networks being of major concern for engineering community and authorities. Seismic instrumentation
is essential for the proper establishment of input ground motion for design and for the seismic evaluation and
rehabilitation of existing buildings. United States of America and Japan are the major examples of countries
understanding the need for improved seismic instrumentation and taking serious actions in this direction.
United States Geological Survey clearly states [USGS, 1995]: "Strong-motion data collected by the USGS
have contributed to the improvement of building codes over the decades. These improved codes have saved
many lives and reduced damage in recent earthquakes. A growing network of instruments will provide even
more extensive data in earthquakes to come. Using this information, scientists and engineers will be able to
suggest further improvements to building codes. These improvements will help protect citizens of the United
States from loss of life and property in future earthquakes”.
1
National Center for Seismic Risk Reduction and Technical University of Civil Engineering, Bucharest, Romania, aldea@utcb.ro,
albota@utcb.ro, demetriu@utcb.ro
2
National Center for Seismic Risk Reduction, Bucharest, Romania, natasa@cnrrs.ro
3
Building Research Institute, Tsukuba, Japan, kashima@kenken.go.jp
143
The same understanding exists in Japan where not only there are dense national seismic networks, but local
authorities, education and research institutions, and companies also developed their own networks. K-NET
[Kyoshin Net] is one of the most impressive networks in seismic instrumentation. After 1995 Great Hanshin
Earthquake, as a proof of understanding the necessity of making available for research seismic ground motions,
1000 free-field digital stations were deployed all over Japan, with an average station to station distance of
~25km. The records are acquired by telemetry and data is available via Internet.
At the First International Workshop on Vrancea Earthquakes [Wenzel and Lungu, 1999] held in Bucharest in
1997, the working group "Strong Ground Motion" chaired by Prof. B.Bolt made the following statement:
"We recommend the establishment of a National Strong-Motion Program to provide an earthquake recording
capability that is vital for earthquake risk reduction and public earthquake safety. The distribution of strong
motion equipment should follow the main seismotectonic and geologic features, including local soil
condition, and also focus on the instrumentation of representative buildings, industrial structures.”
The development of seismic instrumentation in terms of quantity and quality represents a continuous concern
and effort of Romanian and foreign institutions and/or projects. Significant efforts were done within German
Research Foundation SFB461 Project [SFB461], Japan International Cooperation Agency JICA Project
[JICA], and with investments by State Inspectorate for Constructions, these activities being implemented by
in Romania by National Institute for Earth Physics (NIEP), National Center for Seismic Risk Reduction
(NCSRR), and National Institute for Building Research (INCERC). The present paper presents the
instrumentation efforts done within the JICA project.
NCSRR SEISMIC NETWORK
Within the JICA Project in Romania entitled "Seismic risk reduction for buildings and structures" [JICA, 2002],
NCSRR received seismic instrumentation equipments (Kinemetrics). OYO Seismic Instrumentation Corp. and
NCSRR installed the equipments in 2003. In 2005-2006 NCSRR network was enlarged with Romanian
investment (within the budget ensured by Ministry of Transports, Construction and Tourism MTCT), other
sites being instrumented with Geosig equipments and technical support. NCSRR network [Aldea et al., 2004,
2006a] contains 3 types of instrumentation: free-field stations (outside Bucharest), instrumented buildings
and stations with ground surface and boreholes sensors (in Bucharest).
Free-field seismic stations for ground motion attenuation analysis
Six Kinemetrics ETNA stations were installed in 2003 on the SW direction starting from Vrancea epicentral
area toward Bucharest, for ground motion attenuation analysis. All of them are in buildings with 1 or 2
storeys, which can be considered as a free field condition. Ground conditions are not yet known. Two Geosig
IA-1 accelerometers were installed in 2006 and 2007, on a perpendicular axis to the SW. Details about the
free-field stations are given in Table 1, and their distribution is in Figure 1.
Table 1: NCSRR seismic network - free-field stations in Romania
No.
1
2
3
4
5
6
7
8
Site
Giurgiu
Ploiesti
Focsani
Buzau
Ramnicu Sarat
Urziceni
Constanta
Brasov
Station ID
GRG
PLO
FOC
BUZ
RMS
URZ
CST
BRV
Sensor location
Ground Floor of 2 storey bldg.
Equipment
ETNA (Kinemetrics)
GF of 1 storey bldg.
Free-field
IA-1 (Geosig)
Seismic stations for structural monitoring
Two residential buildings and two public buildings were instrumented in 2003. In 2006 the Technical University
of Civil Engineering Bucharest UTCB main building was also instrumented. Details about NCSRR building
instrumentation are given in the present proceedings [Aldea et. al, 2007a].
144
Figure 1: NCSRR free-field seismic network in Romania
Seismic stations for site effects assessment in Bucharest
NCSRR installed in 2003 in Bucharest seven (7) Kinemetrics K2 stations with sensors at ground surface (close
to free-field conditions) and in boreholes at two levels of depth: the first level at about 30m depth and the
second level between 50m and 153m depth. In 2005 another site was instrumented with Geosig equipments
(free-field and a 30m depth borehole). At all the stations the soil profile of the boreholes is known, and NCSRR
and Tokyo Soil Corp. (Japan) performed down-hole tests. A brief description of the borehole instrumentation
is given in Table 2 and their location within Bucharest in Figure 2.
Table 2: NCSRR Bucharest seismic stations with sensors at ground-surface and in boreholes
No.
1
2
3
4
5
6
7
8
Depth of sensor
Depth of
Station Surface sensor
in shallow
sensor in deep
Equipment
ID
location
borehole, m
borehole, m
UTCB Tei
UTC1
free field
-28
-78
UTCB Pache
UTC2 1 storey building
-28
-66
NCSRR/INCERC
INC 1 storey building
-24
-153
K2 + FBA-23DH
Civil Protection Hdq. PRC 1 storey building
-28
-68
(Kinemetrics)
Piata Victoriei
VIC
free field
-28
-151
City Hall
PRI
free field
-28
-52
Municipal Hospital SMU
free field
-30
-70
GSR24+AC23 DH
UTCB Plevnei
UTC3
free field
-30
(Geosig)
Site
Sampling rate is set at 100Hz, pre-trigger time is 30s, post-trigger time is 60s, and full scale is ±2g for all
stations. For Kinemetrics stations time is set by GPS, for Geosig stations it is set by internet-time-servers.
Kinemetrics stations are stand-alone stations, but recently GSM data retrieval is implemented with support
from JICA and Orange. Geosig IA-1 stations are internet based, while GSR station is stand-alone.
145
Figure 2: NCSRR seismic network in Bucharest
EARTHQUAKE RECORDS
Since its installation in 2003, the NCSRR network recorded more than 170 seismic motions from 26 earthquakes
with moment magnitudes ranging from MW=3.2 to 6.0.
Between the earthquakes recorded by NCSRR network, 21 are from Vrancea subcrustal source, 2 from Vrancea
crustal source, 2 from shallow sources in Bulgaria and 1 from North-Dobrogea shallow source. A synthesis
of the distribution of the records with regards to the seismic source and type of instrumentation is given in
Table 3, the main characteristics (as indicated on NIEP and EMSC websites) of the recorded seismic events
are presented in Table 4 and the location of these events is shown in Figure 5. It can be observed that 95% of
the seismic records are due to Vrancea subcrustal source.
Table 3: Distribution of NCSRR seismic records with source type and instrumentation type
Earthquake source
No. of
recorded earthquakes
No. of
records
Vrancea subcrustal
Vrancea crustal
Bulgaria shallow
North Dobrogea shallow
Total
21
2
2
1
26
165
2
5
2
174
146
Distribution of no. of records with
instrumentation type
Free-field
Buildings
Boreholes
51
41
73
2
5
1
1
54
41
79
Table 4: Characteristics of the seismic events recorded by NCSRR seismic network
No.
Date
Seismic source
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
05/10/2003
17/12/2003
24/12/2003
21/01/2004
07/02/2004
30/04/2004
14/05/2004
10/07/2004
27/09/2004
03/10/2004
27/10/2004
17/11/2004
14/05/2005
18/06/2005
05/09/2005
08/09/2005
10/09/2005
26/10/2005
13/12/2005
18/12/2005
16/02/2006
06/03/2006
19/03/2006
23/09/2006
17/01/2007
14/02/2007
Vrancea subcrustal
Bulgaria crustal
Vrancea subcrustal
Vrancea crustal
Bulgaria crustal
Vrancea subcrustal
North-Dobrogea
Vrancea subcrustal
Vrancea crustal
Vrancea subcrustal
Coordinates
Origin time
lat. long.
(UTC)
(oN) (oE)
21:38:18 45.57 26.46
23:15:15 43.19 27.44
13:44:59 45.06 26.08
05:49:10 45.6 26.4
11:58:22 45.72 26.64
09:19:36 45.6 27.13
11:09:37 43.5 26.5
00:34:58 45.52 26.49
09:16:23 45.69 26.32
09:02:01 45.16 29.09
20:34:32 45.83 26.77
11:31:02 45.74 26.72
01:53:22 45.67 26.47
15:16:40 45.79 26.91
14:23:35 45.76 26.62
16:35:50 45.52 26.37
04:56:55 45.25 27.21
22:51:21 45.66 26.57
12:14:45 45.78 26.79
15:09:43 45.41 26.04
02:49:40 45.71 26.66
10:40:46 45.69 26.53
11:05:53 45.65 26.48
05:44:07 45.54 26.40
13:17:22 45.56 26.41
06:56:34 45.49 26.52
Focal Moment
Depth magnitude
(km)
Mw
145.6
60.0
86.0
117.7
146.0
16.9
10.0
150.4
166.1
5.0
98.6
127.4
144.0
135.0
90.0
140.0
22.0
141.8
144.0
60.0
130.0
145.0
157.8
130.8
120.0
159.4
4.6
4.6
3.8
4.1
4.4
3.2
*
4.3
4.6
5.1
6.0
4.4
5.2
5.0
4.4
4.3
3.8
4.3
4.8
3.7
4.1
4.8
4.4
4.3
4.3
4.2
* not reported by agencies
In Figure 4 are represented the magnitudes versus the focal depths of the recorded Vrancea subcrustal earthquakes.
The average magnitude of recorded events is ≈4.5, and the average focal depth is 130 km.
Moment magnitude Mw
3.5
4
4.5
5
5.5
6
6.5
0
20
Focal depth, km
40
60
80
100
120
140
160
180
Figure 4: Distribution of magnitudes versus focal depth of recorded earthquakes
147
Bucharest
Seismic events recorded by NCSRR
seismic network (2003-2007)
Figure 5: Epicentre location map of seismic events recorded by NCSRR seismic network
(fault plain solutions from Harvard CMT and ETHZ)
Earthquake records outside Bucharest
The free-field seismic stations recorded 54 seismic motions (number of records per station: BUZ - 9, FOC 11, GRG - 5, PLO - 8, RMS - 9, URZ - 12) from which 51 motions due to Vrancea subcrustal earthquakes. In
Figure 6 is presented the distribution of magnitudes with epicentral distances corresponding to the existing
records (only for the events from Vrancea subcrustal source).
6.5
Moment magnitude Mw
6
5.5
5
4.5
4
3.5
0
50
100
150
200
250
Epicentral distance ∆, km
Figure 6: Distribution of magnitudes with epicentral distances of records
148
30
3
25
2.5
20
2
IJMA
2
M ax. horiz.PGA (cm/s )
In Figure 7 are shown the distribution of maximum horizontal peak accelerations (left) and the distribution of
Japan Meteorological Agency seismic intensity (right) with epicentral distance, for the records outside Bucharest
(to be used in ground motion attenuation studies), due to Vrancea subcrustal events.
15
1.5
10
1
5
0.5
0
0
10
60
110
160
210
Epicentral distance ∆ (km)
260
10
60
110
160
210
Epicentral Distance ∆ (km)
260
Figure 7: Distribution of max. PGA and IJMA with epicentral distance for the records outside Bucharest
The data from the strongest earthquake (Oct.27, 2004) are not represented in Figure 3 for scaling reasons;
those values are indicated in next chapter. One can notice that for the seismic events with moment magnitudes
up to 5.2, the maximum peak accelerations are bellow 30cm/s2 (the majority of data bellow 15cm/s2). The
JMA seismic intensities are bellow 3 for all these small earthquakes.
The attenuation law for strong ground motions due to Vrancea subcrustal earthquakes developed by Lungu et
al. [1999], is used for predicting PGA attenuation for two scenarios: (i) MW=4.5 and h=130km (average
values for the recorded database), and (ii) MW=6 and h=98.6km (the largest recorded event). The attenuation
relation format is:
ln PGA = c0 + c1 MW + c2 lnR +c3R +c4 h + ε
(1)
where: PGA is peak ground acceleration at the site, MW- moment magnitude, R - hypocentral distance to the
site, h - focal depth, c0, c1, c2, c3, c4 - data dependent coefficients and ε - random variable with zero mean and
standard deviation σε = σln PGA.
The regression was performed using data from three strong Vrancea events (04/03/1977, 30/08/1986 and
30/05/1990),i.e., the max. peak ground acceleration from 80 records from Romania, Moldova and Bulgaria
[Lungu et al., 1999]. In Figure 8 are compared the max PGA recorded within NCSRR network and the attenuation
curves corresponding to the scenarios described above. A satisfactory match can be observed.
Max. PGA
2
Max. horiz. PGA (cm/s )
100
Mw=4.5, h=130km
80
Mw=6, h=98.6km
60
40
20
0
10
60
110
160
Epicentral distance ∆ (km)
210
Figure 8: Comparison of recorded data with attenuation curves
149
260
Earthquake records inside Bucharest
The seismic stations inside Bucharest recorded a total of 120 motions, from which 41 at building stations and
79 at stations with boreholes (number of records per station: UTC1 - 11, UTC2 - 16, UTC3 - 2, INC - 14,
PRC - 11, VIC - 4, PRI - 6, SMU - 15). Information about records in buildings is given in the present
proceedings [Aldea et al., 2007a].
In Figure 9 is presented the distribution with earthquake magnitude of the maximum horizontal peak accelerations
recorded at the stations with boreholes. The data from the strongest earthquake (Oct.27, 2004) is not represented
for scaling reasons; those values are indicated in next chapter.
20
UTC1
Max. PGA , cm/s 2
UTC2
15
INC
PRC
VIC
10
PRI
SMU
5
0
3.6
3.9
4.2
4.5
4.8
5.1
5.4
Moment magnitude M W
Figure 9: Distribution of max. PGA with magnitude for the records inside Bucharest
For the earthquakes with magnitude Mw<4.5, all accelerations are smaller than 15cm/s2. For the earthquakes
with magnitudes between 4.5 and 5.2 the peak accelerations are generally in between 5 and 15 cm/s2. It can
be also noticed that in general the largest peak accelerations were recorded at UTC2 and SMU stations. However,
the data from these small earthquakes and limited number of stations does not indicate a clear difference
between different zones within the city, as it was observed clearly during the Aug.30, 1986 strong event
(Mw=7.2-7.3), as shown in the microzonation map (Lungu et al, 1997, Aldea, 2002).
27th OCTOBER 2004 VRANCEA SUBCRUSTAL EARTHQUAKE (Mw=6.0)
The October 27, 2004 Vrancea earthquake (Mw=6.0, focal depth 98.6km) is the strongest recorded until now
by the NCSRR seismic network and is the strongest event since 1990. The earthquake was felt on large areas
and produced almost no damage as reported by news agencies. A summary of news is presented here.
"Authorities said there were no immediate reports of injuries or damage. It struck at 11:34 p.m. and was felt
in several Romanian cities, including Iasi, Bacau and the capital, Bucharest, where it knocked out telephone
service. The quake also rattled portions of Turkey, Moldova and Ukraine, Turkey's private NTV television
reported. In some Istanbul neighbourhoods, people rushed out of their homes in panic, NTV said. The
observatory's telephone lines were jammed with people calling seeking information." Associated Press
"An earthquake hit Romania on Wednesday night, shaking buildings in the Bucharest but apparently causing
little damage, Reuters witnesses said. Hundreds of people called emergency services soon after the quake
struck at 11:24 p.m., but ambulance officials told Reuters nearly all of them were suffering panic attacks and
it seemed there were no serious casualties. Bucharest's streets were mainly calm, but some people gathered
on main avenues in their nightclothes, waiting for news and trying to find out what had happened.
"No victims or collapsed buildings have been reported in Bucharest so far but we have been flooded with
phone calls from people asking if they should abandon their homes," a police emergency center spokesman
150
said. An ambulance controller in the capital added: <We got about 300 phone calls, mostly from people having
panic attacks in the first hour after the quake.> A spokesman for the Interior Ministry, D. Marcel, said so far
not there had been no reports of significant damage. Witnesses said cracks had appeared in Bucharest's
historic City Hall and plaster was falling off. Residents feared they would see more evidence of damage at
first light. In Braila, eastern Romania, the wall of an abandoned building collapsed and the windows of office
building had been shattered, the Interior Ministry's Marcel added. Bulgarian civil defense officials said the
tremors had been felt on their side of the border, but it was too early to tell if there was any significant damage
in the country." Reuters
All NCSRR seismic stations ouside Bucharest recorded the event, the peak ground accelerations are given in
Table 5. One may notice the highest horizontal PGAs at Focsani, Buzau and Ploiesti stations, with the maxima
of 86cm/s2 (EW comp.) at Buzau. A peculiar powerful vertical motion was recorded at Ramnicu Sarat with a
P waves peak of 219.6cm/s2.
Table 5: Oct. 27, 2004 Vrancea earthquake - PGA at NCSRR free field stations
Station name
Focsani
Ramnicu Sarat
Buzau
Urziceni
Ploiesti
Giurgiu
Station
ID
FOC
RMS
BUZ
URZ
PLO
GRG
Peak ground acceleration PGA, cm/s2
NS
EW
Vertical
62.9
64.6
82.2
41.4
49.0
219.6
67.9
86.0
80.8
33.7
44.3
38.9
64.5
49.2
34.8
22.9
30.6
15.3
The Japan Meteorological Agency seismic intensity had values below 4: 3.9 (FOC), 3.6 (RMS and BUZ), 3.5
(PLO), 3.4 (URZ), and 2.8 (GRG). JMA seismic intensity 3 is described as follows: "Felt by most people in
the building. Some people are frightened. Dishes in a cupboard rattle occasionally. Electric wires swing
slightly.", and JMA intensity 4 "Many people are frightened. Some people try to escape from danger. Most
sleeping people are awakened. Hanging objects swing considerably and dishes in a cupboard rattle. Unstable
ornaments fall occasionally. Electric wires swing considerably. People walking on a street and some people
driving automobiles feel the tremor." Even the JMA intensity is computed with a formula calibrated using
Japanese records and corresponding damage, these descriptions generally match the newspaper data.
In Bucharest, all stations with boreholes recorded the Oct.27, 2006 earthquake (Mw=6) except VIC one (due
to the absence of electric supply). The recorded peak ground accelerations PGA are presented in Table 6. All
NCSRR instrumented buildings in Bucharest recorded the earthquake. Details are given in the present proceedings
[Aldea et. al, 2007a].
Table 6: Oct. 27, 2004 Vrancea event - peak accelerations at NCSRR stations in Bucharest
Station
Surface PGA
Shallow Depth
sensor PGA
Deep Depth
sensor PGA
UTC1
NS EW V
34.9 58.4 34.4
-28 m
28.5 14.6 11.1
-78 m
16.5 23.1 9.8
UTC2
NS EW V
41.6 40.9 24.8
-28 m
21.6 16.8 11.5
-66 m
15.6 23.5 7.0
INC
NS EW V
29.7 29.6 24.9
-24 m
13.9 12.5 8.3
-153 m
11.3 11.4 6.7
PRC
NS EW V
29.0 49.2 34.0
-28 m
20.3 13.1 11.2
-68 m
12.7 19.4 8.8
PRI
NS EW V
29.8 79.0 33.1
-28 m
16.6 37.7 11.8
-52 m
13.2 22.2 9.6
SMU
NS EW V
54.6 44.5 50.8
-30 m
11.6 18.5 8.2
-70 m
12.6 18.1 8.9
A certain variability within the city can be observed, with highest values in the vicinity of Dambovita river
(PRI and SMU), and with the lowest value in eastern Bucharest (INC). It can be noticed that the top ~30m of
soil had the most important contribution for the level of the PGA by at least doubling the recorded values, while
from the deep borehole to the shallow one there are no major changes. The JMA intensity computed from the
ground surface records was: 3.4 (SMU), 3.3 (UTC2), 3.2 (UTC1 and PRI), 3.0 (INC and PRC).
151
The surface/borehole spectral ratio (SBSR) method has pro and contra arguments (Atakan, 1998 and Safak,
1997). Since the NCSRR sensors in deep boreholes are not located on/in the bedrock, the SBSR does not give
a complete image on the site response, but gives information on the response during earthquakes of the soil
column between two sensors. The surface (S) over deep borehole (B2) spectral ratios for the 27/10/2004
event are presented in Figure 10 (where the two lines indicate the B2/S ratios for EW and NS components).
The frequencies corresponding to the first identified peaks in Figure 10 are given in Table 7.
Figure 10: SBSR at NCSRR seismic stations in Bucharest (27/10/2004 earthquake)
Table 7: First frequency peak identified from SBSR (27/10/2004 earthquake)
Station
B2 borehole depth
f1 (Hz) from SBSR
PRI
-52 m
1.53
UTC2
-66 m
1.35
PRC
-68 m
1.21
SMU
-70 m
1.32
UTC1
-78 m
1.19
INC
-153 m
0.7
It can be observed that, in general, deeper the borehole sensor lower the frequency corresponding to the first
spectral peak. In case of sites with deep sediments (as in the case of Bucharest), the site response and the site
conditions should be analysed by considering the whole thickness of sediments until the bedrock, and not
only the top 30m. INCERC site gives an instrumental proof that the soil category and the design spectrum
selected by using only the upper 30 m of soil can be a misleading approach, the strong earthquakes of 1977
(Mw=7.5) and 1986 (Mw=7.2-7.3) having response spectra with large values at long periods, much larger
than those from EC 8 spectra (corresponding to the site category based on the top 30m of soil). The long-period
phenomenon was explained by the contribution of two factors: the source characteristics in case of strong
Vrancea events and the site conditions (thick sediments with probably non-linear behaviour during strong
earthquakes). Even the SBSR presented in Figure 4 are characterising the soil column response between the
two sensors in elastic range, they show that soil thickness has an important contribution in the site response,
and that the ground vibration fundamental period tends to be closer or even higher than 1 sec.
As an example, in Figure 11 is presented the evolution with depth of the accelerograms recorded at INC station
(EW components). In Figure 12 are presented the evolutions with depth of acceleration and velocity response
spectra, of Fourier amplitude spectra and of H/V Fourier amplitude spectral ratios. In higher frequency domain
the upper 24m have a significant impact. In the low frequency domain, the impact of the upper 24 m
becomes insignificant, except for the velocity spectra. It should be noticed that even at 153m depth there is
an important content of low frequencies, and a clear peak appears around 0.4-0.5Hz.
152
Acc. (cm/s/s)
Acc. (cm/s/s)
Acc. (cm/s/s)
30
Acceleration
0
Surface EW (peak:- 29.6 cm/s/s)
-30
30
0
-24m EW (peak:- 12.5 cm/s/s)
-30
30
0
-153m EW (peak:- 11.4 cm/s/s)
-30
0
20
40
60
80
100
120
Time (sec)
140
160
180
200
220
Figure 11: Evolution with depth of the accelerograms recorded at INC station (EW comp.), 27 Oct.2004
100
Acc. Response Spectrum (h=5%)
5
Vel. Response Spectrum (h=5%)
1
Velocity Response (cm/s)
Acceleration Response (cm/s/s)
50
10
5
1
0.5
0.1
0.05
0.5
Surface EW
-24m EW
-153m EW
0.1
0.05
20
0.1
0.5
1
Period (sec)
Surface EW
-24m EW
-153m EW
0.01
5
10
0.005
0.05
20
Fourier Spectrum (Time:0-223s, Parzen:0.2Hz)
10
0.1
0.5
1
Period (sec)
5
10
20
Fourier Spectral Ratio (Time:0-223s, Parzen:0.2Hz)
10
5
Spectral Ratio
Fourier Amplitude (cm/s)
5
1
0.5
1
0.5
0.1
0.02
0.2
Surface EW/Surface V
-24m EW/-24m V
-153m EW/-153m V
Surface EW
-24m EW
-153m EW
0.05
0.5
1
Frequency (Hz)
5
0.1
0.2
10
0.5
1
Frequency (Hz)
5
10
Fig.12: INC: evolution with depth of acceleration & velocity response spectra, of Fourier spectra & H/V ratio
Since a reference nearby rock site is not available in Bucharest area, a non-reference site technique (the single
station H/V Fourier amplitude spectral ratio) can be used for estimating the site response characteristics.
Despite a lack in theoretical justification, the single station spectral ratio was tested successfully for soil sites
by an increasing number of authors [example Lermo and Chavez-Garcia, 1993]. In Figure 13 are comparatively
presented the H/V spectral ratios at ground surface, at the shallow sensor B1 (-24m) and at the deep sensor
153
B2 (-153m), computed using the whole record, the S-wave part, the coda/surface-wave part and the P-wave
part. A first peak (not of largest amplitude) appears constantly around 0.4-0.5Hz, when using the whole
record, the S-wave part and the coda/surface-wave part.
Fourier Spectral Ratio (Time:0-223s, Parzen:0.2Hz)
10
1
0.5
1
0.5
0.5
SEW/SV
SNS/SV
SEW/SV
SNS/SV
0.1
0.2
0.1
0.2
10
10
10
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:30-47s, Parzen:0.2Hz)
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:0-223s, Parzen:0.2Hz)
1
0.5
1
0.5
SEW/SV
SNS/SV
0.1
0.2
1
0.5
B1EW/B1V
B1NS/B1V
0.1
0.2
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:80-200s, Parzen:0.2Hz)
10
B1EW/B1V
B1NS/B1V
0.1
0.2
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:30-47s, Parzen:0.2Hz)
10
5
Spectral Ratio
5
1
0.5
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:0-223s, Parzen:0.2Hz)
5
Spectral Ratio
10
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:47-80s, Parzen:0.2Hz)
5
Spectral Ratio
5
Spectral Ratio
1
0.5
1
0.5
B1EW/B1V
B1NS/B1V
B1EW/B1V
B1NS/B1V
B2EW/B2V
B2NS/B2V
0.1
0.2
0.1
0.2
0.1
0.2
10
10
10
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:47-80s, Parzen:0.2Hz)
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:80-200s, Parzen:0.2Hz)
5
Spectral Ratio
5
1
0.5
1
0.5
0.5
1
Frequency (Hz)
1
0.5
B2EW/B2V
B2NS/B2V
B2EW/B2V
B2NS/B2V
0.1
0.2
0.5
1
5
10
Frequency (Hz)
Fourier Spectral Ratio (Time:30-47s, Parzen:0.2Hz)
5
Spectral Ratio
Spectral Ratio
1
0.1
0.2
5
Spectral Ratio
Fourier Spectral Ratio (Time:80-200s, Parzen:0.2Hz)
5
SEW/SV
SNS/SV
Spectral Ratio
10
5
Spectral Ratio
Spectral Ratio
5
Fourier Spectral Ratio (Time:47-80s, Parzen:0.2Hz)
Spectral Ratio
10
5
10
0.1
0.2
0.5
1
Frequency (Hz)
B2EW/B2V
B2NS/B2V
5
10
0.1
0.2
0.5
1
Frequency (Hz)
5
10
Figure 13: INC (27/10/2004): H/V spectral ratios at ground surface, at the shallow sensor B1 (-24m) and at
the deep sensor B2 (-153m), computed using the whole record, the S-wave part, the coda/surface-wave part
and the P-wave part of the record.
154
PS LOGGING TESTS AT THE SEISMIC STATION SITES
At all the stations the soil profile/stratigraphy of the boreholes is known, and NCSRR and Tokyo Soil
Research Co., Ltd. performed in 2003 down-hole tests for the estimation of the seismic velocities profiles at
all sites (Aldea et al., 2006b).
In Table 8 are presented the weighted average (UBC formula) shear wave velocity (Vs) for seven sites (using
the upper 30m, the upper 52 m and the whole investigated depth). It can be observed that the differences
between the sites are not significant when looking at the average on 52m and some differences (up to 50%)
exists at the average on 30m. The sites are classified as "hard soil"-class D according to UBC 1997, and
"Deep deposits of dense or medium dense sand, gravel or stiff clay with thickness from several tens to many
hundreds of m" class C according to Eurocode 8. In the mentioned codes, these ground classes are associated
with a control period of response spectra Tc≈0.6s, which is not in agreement with the real situation in Bucharest,
where during August 30, 1986 Vrancea earthquake the values of Tc were higher than 0.6s all over the city
(Lungu et al, 1997, Aldea, 2002). In the future editions of the Romanian earthquake resistant design code,
the ground categories may also be related to the predominant period of the site, supplementary to the
characteristics of the upper 30m of soil profile, since the important thickness of medium-to-hard sediments
can also induce long periods motions with high spectral amplifications beyond 1 second. The influence of
soil thickness on the predominant periods in Table 8 is self-explanatory.
Table 8: Average shear wave velocity at NCSRR stations based on down-hole tests
Station
Average (30m) Vs, m/s
Predominant frequency (30m), Hz
Average (52m) Vs, m/s
Predominant frequency (52m), Hz
Average Vs, m/s
(whole borehole depth, m)
Pred. freq. (whole investigated depth), Hz
f1 (Hz) from SBSR (27/10/2004 event)
PRI
219
1.83
258
1.24
258
(52m)
1.24
1.53
UTC2
288
2.40
318
1.53
332
(66m)
1.26
1.35
PRC
293
2.44
309
1.49
324
(68m)
1.19
1.21
SMU
245
2.04
281
1.35
303
(69m)
1.10
1.32
UTC1 INC
VIC
309
270
284
2.58
2.25
2.37
326
302
310
1.57
1.45
1.49
349
364
354
(78m) (140m) (110m)
1.12
0.65
0.81
1.19
0.7
-
In Table 8 are also computed the predominant frequencies using an approximate formula (average Vs over
four times the total considered thickness). When estimating the predominant frequency by using the whole
investigated depth, the predominant frequency values start to be similar to those obtained from SBSR ratio
(Table 7).
The H/V ratios of the ground surface records from Oct.27, 2004 indicate in Bucharest a ground vibration
rather rich in low frequencies, all the ratios displaying significant amplitudes around 1Hz and/or below [Aldea
et al., 2007b]. A first peak (not of largest amplitude) appears constantly around 0.4-0.5Hz, a second one around
0.7-0.8Hz, and a third one again quite constantly around 1Hz. In a rough approximation, considering a quaternary
layer with h=300m thickness and VS=450m/s shear wave velocity, the vibration frequency f=VS/4h=0.38Hz.
Bonjer et al. (1999) indicated for all over the city a peak of ~0.7Hz obtained from microtremor measurements,
and also reported a 0.5Hz peak using records from May 30, 1990 Vrancea event. It should be noticed that at
INCERC site (INC station) a peak at 0.7Hz-0.8Hz was reported by Lungu et al., [1997] using power spectral
density H/V ratio for March 4, 1977 earthquake records. Also, at INCERC site the same 0.7Hz-0.8Hz peak
was obtained from H/V Fourier spectral ratios and H/V displacement response spectra ratios for 1977 earthquake
records, (Aldea and Okawa, 2000).
CONCLUSIONS
The development of seismic monitoring in Romania is an important achievement of JICA Project in Romania,
creating the premises for a better understanding and modelling of earthquake ground motion attenuation, site
effects and building response. In future, it is necessary to characterise site/soil conditions at all the seismic
stations sites (from all networks in Romania) in order to establish the link between site/soil conditions and
155
ground motion characteristics used for design. The remarkable borehole instrumentation now available in
Bucharest, together with the increasing number and quality of soil data will allow an improved understanding
and predictive modelling of site response within the city. The functioning and development of the existing
Romanian seismic networks should be continuously supported, since recorded ground motions are fundamental
for improving the seismic input for design of new buildings and for evaluation and rehabilitation of existing
ones, items that finally contribute to the reduction of seismic risk.
AKNOWLEDGEMENTS
JICA is gratefully acknowledged for the donation of equipment for seismic instrumentation toward the NCSRR
in the frame of the JICA project in Romania. Acknowledgements to OYO Seismic Instrumentation Corp. and
Geosig for their technical support in installation and maintenance. Orange (Romania) is acknowledged for
supporting the implementation of GSM data transmission. All the institutions and persons that accepted and
supported the installation of NCSRR seismic stations are also gratefully acknowledged.
REFERENCES
Aldea, A., Demetriu, S., Albota, E., Kashima, T., (2007a). Instrumental Building Response Studies within JICA Project in Romania,
Proceedings of International Symposium on Seismic Risk Reduction ISSRR2007, Bucharest, 14p. (present volume)
Aldea, A., Kashima, T., Okawa, I., Poiata, N., Koyama, S., Demetriu, S., (2007b). Free-Field And Borehole Strong Motion Array In
Bucharest, Romania, 4th International Conference on Earthquake Geotechnical Engineering, Paper No. 1521, Thessaloniki
Aldea, A., Kashima, T., Poiata, N., Kajiwara, T (2006a), A New Digital Seismic Network in Romania with Dense Instrumentation in
Bucharest, Proceedings of First European Conference on Earthquake Engineering and Seismology, Geneva, 2006, 10p., CD
Aldea, A.,Yamanaka, H., Negulescu, C., Kashima, T., Radoi, R., Kazama, H., Calarasu, E.,(2006b). "Extensive seismic
instrumentation and geophysical investigations for site-response studies in Bucharest, Romania", ESG 2006 Third International
Symposium on the Effects of Surface Geology on Seismic Motion, Grenoble, France, 30 Aug.-1 Sept., Paper Nr: 69, p.415-426
Aldea, A., Yamanaka, H., Takahashi, T., (2006c). "Microtremor array measurements in northern Bucharest for estimation of site
response", Proceedings of First European Conference on Earthquake Engineering and Seismology, Geneva, 3-8 Sept., 10p., CD,
Aldea, A., Kashima, T., Lungu, D., Vacareanu, R., Koyama, S., Arion, C., (2004). "Modern Urban Seismic Network in Bucharest,
Romania", Proc. of the First International Conf. on Urban Earthquake Engineering, Tokyo Inst. of Techn., Yokohama, 8p.
Aldea, A., (2002). “Vrancea source seismic hazard assessment and site effects”, PhD thesis Technical University of Civil Engineering
Bucharest, Bucharest, 256p.
Aldea A., Okawa, I., (2000). "Soft soil site response analysis in Eastern Bucharest, Romania", 6ICSZ Sixth International Conference
on Seismic Zonation, Palm Springs, California, USA, p.581-586
Atakan, K., (1998). "A review of the type of data and the techniques used in empirical estimation of local site response", Proceedings
of International Symposium on the Effects of Surface Geology on Seismic Motion 1998, Japan, p.1451-1460,
Bonjer, K.-P., Oncescu, M.-C., Driad, L., Rizescu, M., (1999). "A note on empirical site responses in Bucharest", in Vrancea
Earthquakes. Tectonics, Hazard and Risk Mitigation, Kluwer Academic Publishers b.v., Wenzel F., Lungu D., Editors, Proc. of
First International Workshop on Vrancea Earthquakes, p.149-162
Kyoshin network (http://www.kyoshin.bosai.go.jp/)
Lermo, J., F.J.Chavez-Garcia, (1993). Site effect evaluation using spectral ratios with only one station, Bull. Seism. Soc. Am., Vol.83,
No.5, p.1574-1594,
Lungu, D., Cornea, T., Aldea, A., Nedelcu, C., Demetriu, S., (1997). "Uncertainties in mapping frequency content of soil response to
various magnitude earthquakes". EUROMECH 372 Colloquium of the European Mechanics Society, Reliability in nonlinear
structural mechanics, Université Blaise Pascal, Institut Français de Mécanique Avancée, Clermont-Ferrand, France, p. 39-48
Lungu, D., Arion, C., Aldea, A., Demetriu, S., (1999), “Assessment of seismic hazard in Romania based on 25 years of strong ground
motion instrumentation”, NATO ARW Conference on strong motion instrumentation for civil engineering structures, Istanbul,
Turkey, June 2-5, 1999.
JICA (2002) Project "Reduction of Seismic Risk for Buildings and Structures" with Romania, 2002-2007, Ref.#7241011E0
Safak E., "Models and methods to characterize site amplification from a pair of records", Earthquake Spectra, Vol.13, No.1,
February, p.97-129, 1997
SFB 461 (http://www-sfb461.physik.uni-karlsruhe.de/)
U.S. Geological Survey (1995). Fact Sheet 176-95/Online v1.0. Reducing Earthquake Losses Throughout the United States. Saving
Lives Through Better Design Standards.
Wenzel, F., Lungu, D., Editors (1999), Vrancea Earthquakes. Tectonics, Hazard and Risk Mitigation, Contributions from the First
International Workshop on Vrancea Earthquakes, Bucharest, Romania, 1997, Kluwer Academic Publishers, 374 p.
156