International Symposium on Seismic Risk Reduction

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