Bull Earthquake Eng (2014) 12:237–254DOI 10.1007/s10518-013-9580-y

ORIGINAL RESEARCH PAPER

Evaluation of the national strong motion networkin Greece: deployment, data-processing and sitecharacterization

B. Margaris · I. Kalogeras · Ch. Papaioannou ·A. Savvaidis · N. Theodoulidis

Received: 16 September 2013 / Accepted: 23 December 2013 / Published online: 16 January 2014© Springer Science+Business Media Dordrecht 2014

Abstract In this paper a brief introduction regarding the latest advancements of strongmotion instrumentation programs in Greece since the first accelerograph was installed byNOA-IG is given. Analytical description of various utilized analog and digital accelerographicsensors is presented and the evolution of the different types of data processing techniquesapplied by the two research centers is given using recordings from both Institutes. The sitecharacterization of the strong motion recording sites and the available information for thosestations are presented as well. A final evaluation and correlation of the previous proposedconfiguration of the Hellenic accelerographic networks ITSAK and NOA-IG resulted fromprevious study (Theodoulidis et al. in Planning of strong motion network. pp 86–04, 1986)in regard to the new strong-motion installations within the framework of EPPO’s project isaccomplished.

Keywords National accelerographic network · Data processing · Site characterization

1 Introduction

Earthquake strong-motion observation networks are being expanded with the worldwide con-tinuing addition of strong-motion recording stations. Thousands of strong-motion recordingsare now available and the inventory continues to grow. Web-based data mining technologiesnow permit efficient distribution of strong-motion data over the Internet by linking multipleindependent database providers through a main data dissemination hub. To support optimumuse of these important advances in strong motion monitoring and data dissemination capa-bilities by earthquake scientists, standardized formatting and processing of strong-motionrecordings is essential.

B. Margaris (B) · Ch. Papaioannou · A. Savvaidis · N. TheodoulidisEPPO-ITSAK: Institute of Engineering Seismology and Earthquake Engineering,Thessaloniki, Greecee-mail: margaris@itsak.gr

I. KalogerasNOA-IG: Institute of Geodynamics National Observatory of Athens, Athens, Greece

123

238 Bull Earthquake Eng (2014) 12:237–254

Strong-motion processing of accelerograms is a highly specialized procedure which isperformed to raw data by applying worldwide-distributed software. Many record processingissues are commonly documented and independently dealt with by each code. While thephysical separation of programs has impeded the development of common approaches, it isclear that more uniformly processed recordings would increase the confidence of users andfacilitate the broad exchange of strong-motion data. Current international practices and pro-cedures for processing strong-motion recordings, identify especial record processing issues,and develop recommendations for uniform routine and special record processing guidelinesthat can be adopted by data providers. Strong-motion data processing is an application ofsignal processing analysis to accelerograms recorded after a strong seismic event.

Digital strong-motion instruments allow routine recovery of ground motions at periodsmuch longer than those of older analog instruments. In spite of the theoretical possibility ofrecovering even the residual displacements in the vicinity of earthquakes, in practice digitalrecordings are often plagued by drifts in the velocity and displacement traces obtained byintegrating acceleration traces. Baseline-correction schemes may not give stable results, andremoval of low—frequencies by filtering is often required. However, values of filter cor-ners can be so low, that almost none of engineering interest information is lost. Recent datafrom several earthquakes shows that for displacement response spectra, the transition fromincreasing to constant spectral levels occurs at significantly longer periods than in Eurocode8; the transition period is in good agreement, however, with the recent 2003 NEHRP code(Akkar et al. 2005).

The basic aim of strong motion instrumentation program in Greece is to adopt uniformstandards for strong motion data collection and processing, enriching the Greek accelerogramdata bank with reliable data for engineering seismology research needs and further improve-ment of seismic codes. Gradual installation of the first analog strong motion instruments beganin early 70’s by the Institute of Geodynamics, National Observatory of Athens (NOA-IG)followed by the Institute of Engineering Seismology and Earthquake Engineering (ITSAK) inearly 80’s, in high seismicity areas of Greece. The development and the spatial distribution ofthe accelerographic network in Greece was mainly based on Theodoulidis et al. (1986) study,in terms of seismotectonic, seismicity and seismic hazard assessments of the region. Afteralmost 25 years and two major network upgrading projects, the 3rd instrumentation projectfinancially supported by EPPO (Earthquake Planning Protection Organization) reached itslast phase deploying almost 400 strong motion instruments throughout Greece (Theodoulidiset al. 2013a). Simultaneously a unified strong motion database of Greek accelerographic data,spanning early 70’s up to the end of 2009, will be available to the geoscientists and engineeringcommunity.

In this paper a brief introduction regarding the latest advancements of strong motion instru-mentation programs in Greece since the first accelerograph installed by NOA-IG, is presented.Analytical description of various utilized analog and digital accelerographic sensors is pre-sented and the evolution of the different types of data processing techniques applied by bothresearch centers is given using recordings of the national strong motion network. Site char-acterization of strong motion recording sites and available information for those stations arepresented as well. Evaluation of strong motion data acquired from the national strong motionnetwork during the last about 40 years is attempted and their scientific potential is discussed.

2 Evolution of strong motion network in Greece

The strong motion instrumentation program in Greece started in early 70’s within the project“Balkan region seismicity study”. The Institute of Geodynamics of the National Observatory

123

Bull Earthquake Eng (2014) 12:237–254 239

Fig. 1 Record of a SR-100 Wilmot seismoscope (left) and the SMA-1 (right) record of the November 4, 1973,M5.8 earthquake near Lefkada island (Western Greece)

in Athens (NOA-IG) installed 2 SMAC-B accelerographs, which in a short time were replacedby a network of 15 SMA-1 accelerographs. This first network came to substitute the installedseismoscopes (SR-100 of Wilmot), which although were not accelerographs, they could bedeemed as simplified instruments for recording strong earthquakes (Cloud and Hudson 1961).

There were at least three occasions mentioned for the same earthquake being recorded byseismoscopes and accelerographs located at the same site; the Cephalonia Island main shockof September 17, 1972, M6.3 and its aftershock and the Lefkada island main shock of Novem-ber 4, 1973, M5.8 (Drakopoulos and Rousopoulos 1973; Galanopoulos and Drakopoulos1974) (Fig. 1).

This limited geographical extension of strong motion instruments covered almost allthe 70’s period. After the destructive earthquake, which occurred in Thessaloniki, northernGreece (Jun. 20, 1978, M6.5), the Greek Ministry of Public Works established a new orga-nization (the Institute of Earthquake Engineering and Engineering Seismology—ITSAK),with main research activities in engineering seismology, soil dynamics and earthquake engi-neering. In order to serve these activities, ITSAK undertook the installation, maintenance,and operation of a permanent strong motion network in Greece. The deployment of ITSAKaccelerographic network in Greece was accomplished after an analytical study, which tookinto account all available seismological, seismotectonic and seismic hazard assessment results(Theodoulidis et al. 1986). By 90’s ITSAK strong motion network has been expanded to over65 installations Greece, covering free-field and in basements of mainly low rise buildings,various site conditions and tectonic environments, while NOA-IG, based on the State financialsupport, expanded its network reaching a number of 45 permanently installed instruments ofSMA-1 type.

The two institutes established a non-formal cooperation (Theodoulidis et al. 2004) on thesubjects of strong motion record processing, specifications of instruments installations andmaintenance. The growing number of strong motion data collected, forced the scientists tounderstand the importance of the applied processing and correction techniques as a step beforeany final engineering use of recordings. Thus, ITSAK and NOA-IG developed a digitizationand correction procedure for their accelerograms (Margaris 1986; Trifunac and Lee 1973).

Due to the rapid evolution of digital technology, installation of digital instruments oflow resolution (12-bit) started by mid 90s (A-800/A-900 of Geotech), while the analogunits of ITSAK and NOA-IG have been replaced by common digital accelerographs with an“almost” 11-bit A/D converter (QDR of Kinemetrics) by 2001 and after the September 7,1999 M5.9 Athens earthquake. The dynamic range of those instruments can not be consideredas better than the analog ones, however the main advantage of these accelerographs is theelimination of digitization processing errors and remote data accessibility through modem.

123

240 Bull Earthquake Eng (2014) 12:237–254

The data analysis and processing of the aforementioned digital recordings can be carriedout by computer software provided by Kinemetrics (1999). At that time, the first Greekunified accelerographic relational database was developed, including records from the twonetworks, which were selected under specific criteria, processed uniformly and after carefulmetadata selection resulted to the new Greek Ground Motion predictive Equations GMPE’s(Theodoulidis et al. 2004; Skarlatoudis et al. 2003a).

By 2006, NOA-IG became a member of the Syzefxis network (the State network for inter-net connections which provides various services like data transfer, e-mail, VOIP, teleconfer-encing, digital signature, a helpdesk and a users’ forum). Within Syzefxis, NOA-IG developedthe necessary hardware and software in order to enhance “traditional” accelerographic instru-ments equipped with serial interfaces and exhibited various operational characteristics withmodern network functionality. This requires not only several upgrades in terms of physicalinterconnections but designing and implementing the necessary intermediate software mod-ules as well. This upgrade offered a continuous 24/7/365 monitoring of the digital instruments,a rapid data download and processing after a strong earthquake (Kalogeras et al. 2008).

In 2007–2008, the Regional Authority of Central Macedonia, Greece, funded the purchaseand installation of 32 high resolution recorders (GSR-24), all equipped by Guralp CMG-5Tsensors, covering the largest settlements of Central Macedonia and the city of Thessalonikiitself. This network is remotely maintained via dial-up and Internet connections procedureby the EPPO-ITSAK.

Since 2008, the Greek strong motion network consisted totally from digital instruments(A-800/A-900, QDR, ETNA, K2, CMG-5TD, GSR-24). At that time a national project ledto a tremendous increase of quantity and quality of the national strong motion instruments,namely the purchase of more than 200 CMG-5TD-EAM units by the two Institutes and theirinstallation at permanent sites. The technical specifications of this modern instruments offera lot of services, like the continuous recording (in addition to triggering and STA/LTA algo-rithms), the site ambient noise monitoring via the PQLX software, the incorporation of therecording to the daily seismicity monitoring and an enhancement at the seismic parameterestimation (improvement of epicenter location, magnitude determination and focal mecha-nism solution). The up to date deployment of the free-field strong motion stations in Greeceis shown in Fig. 2.

In addition, real time continuous data streaming offers the possibility of additional publicservices, like Shakemap generation. As one of the outcomes of the modernized HellenicAccelerographic Network with real time transfer of continuous streams is the deploymentof a ShakeMap Service. This service is running in CLOUD computing infrastructure, pro-viding high sustainability (24/7) and can be accessed at http://portal.ingeoclouds.eu/sitools/client-user/Shakemaps/project-index.html. This application is based on the Shake-Map prod-uct of the U.S. Geological Survey Earthquake Hazards Program (http://earthquake.usgs.gov/shakemap/) in conjunction with regional seismic network operators (Wald et al. 2006).ShakeMap sites provide near-real-time maps of ground motion following significant earth-quakes. In the current approach for the Aegean area Ground Motion Predictive Equations(GMPE’s) provided for the region have been used and more specific for the PGA, Skarlatoudiset al. (2003b, 2004) for shallow earthquakes with M > 4.5 and 4.5 ≥ M ≥ 3, respectively,have been adopted.

In Fig. 3, the ShakeMap (Peak Acceleration Map) for an earthquake event in CentralGreece of M5.5 is presented. The source parameters are provided as an automatic locationsolution from the Seismological Station of the Aristotle University of Thessaloniki. Theiso—lines correspond to Peak Ground Acceleration and the yellow triangles to the positionsof the Accelerographic Stations.

123

Bull Earthquake Eng (2014) 12:237–254 241

Fig. 2 National strong motion Network running by two research centers (EPPO-ITSAK and NOA-IG: opencircles represent the sites equipped with strong motion instruments working on trigger mode while the greycircles depict the accelerographic broadband network working on continuous recording)

3 Strong motion sensors and data processing of Hellenic accelerograms

A significant number of strong motion data has been collected by the two national accelero-graphic networks in Greece ITSAK and NOA-IG, and soon became obvious that the appliedprocessing and correction techniques would be an important step before the final dissemina-tion of those recordings. ITSAK recognizing the significance of strong motion data processingdeveloped a digitization and data processing procedure for its strong-motion data (Margaris1986). For this reason in the mid of 80’s a manual digitization procedure for strong motionrecords was developed and data processing technique was installed in collaboration withENEA-ENEL Italy (Tinelli et al. 1985). Results of the aforementioned digitization and dataprocessing approach were presented at the European Strong Motion Meeting at ImperialCollege in 1990 organized by Pr. N. N. Ambraseys. Because of the hardware and softwareadvancements, at the beginning of 90’s a scanner based film accelerogram digitization proce-dure using PC was applied (Nigbor and Kodama 1990) in both Institutes ITSAK and NOA-IG(Kalogeras 2002; Skarlatoudis et al. 2003a). The analog accelerograms were digitized usinga high-resolution scanner and an appropriate software tool (Nigbor and Kodama 1990), thusreducing the digital noise introduced compared to previous (manual or semi-automatic) dig-itization techniques.

The data processing of the aforementioned semi-digital and digital recordings were carriedout by a software provided by Kinemetrics (1999) and the smoothed Fourier AmplitudeSpectra (FAS) of the 3-component recordings and the corresponding fixed traces were utilized

123

242 Bull Earthquake Eng (2014) 12:237–254

Fig. 3 Peak acceleration map (in %g) for the earthquake event of August 7th, 2013 in Central Greece withM = 5.5. Iso-lines denote Peak Ground Acceleration and Yellow Triangles the position of AccelerographicStations

for the determination of filter cut-offs in data processing (Skarlatoudis et al. 2003a). To furtherimprove the quality of the strong motion dataset, a new correction procedure was applied inthe uncorrected accelerograms. This procedure involves the estimation of the characteristicfrequencies of the digital band-pass filter, based on comparisons of the FAS of the digitizedhorizontal components with the FAS of the corresponding fixed traces. An Internet site forEuropean strong-motion data was established in four different European research centers,(ICST, ITSAK, Univ. of Trieste and Univ. of Reykjavik) under Pr. N. N. Ambraseys’ auspicesand coordination (Ambraseys et al. 2004).

Digital strong motion instruments have many benefits over the older analog ones and thecorresponding recordings mainly within the frame of the quality of processing procedureand their frequency content. The typical resolution of a digitized analog record has a maxi-mum of 2 cm/s2, while for digital instruments with ±2 g full scale recording the respectivemaximum value reaches the 1 cm/s2 for 11-bit ADC and 0.0002 cm/s2 for 24-bit ADC.

123

Bull Earthquake Eng (2014) 12:237–254 243

Fig. 4 Focal mechanism solution of the Limnos earthquake (Jan 08, 2013 Mw5.7) corresponding to a strike-slip seismogenic fault. For this high quality solution nine strong motion recordings were used (LIA, EFSA,PRK, SMTH, PSRA, ALXA, ALNA, THSA, KVLA) together with three broadband seismographic recordings(SKY, ALN, ATH) (http://bbnet.gein.noa.gr/mt_solution/2013/130108_14_16_08.00_MTsol.html)

The Fourier acceleration spectrum decreases at long periods, and thus the higher resolutionshould allow reliable calculations to be obtained at longer periods. These longer period dataare useful in defining residual displacements, long-period site response and basin waves,as well as the long-period behavior of displacement response spectra and the strains withinstructures (Boore 2001). Since the mid of 2000 decade such 24-bit accelerographs initiatedto be installed all over Greece from ITSAK and NOA-IG and recently the project for thedeployment and the beginning of operation of the Hellenic Strong Motion Network has beencompleted after the financial support of EPPO. Within this project over 400 instruments are inoperation all over Greece in nowadays, the majority of which record in real-time continuousmode.

As accelerometric stations become upgraded, new deployed and operate in real-time,they are included in the seismic network monitoring within the frame of NOA-IG routineoperational procedure. Since some accelerometric sensors are co-located with seismic broad-band sensors, this extends today further the ability of the operated system to act more reliably,if a very strong earthquake takes place within the region covered by the network, as the broad-band seismometers will be saturated (clipped) and the accelerometric data may replace themin the routine procedure.

Moreover real-time accelerometric data are used within the daily routine procedure forthe calculation of deviatoric moment tensors. Figure 4 is an example for the focal mechanismsolution of the Limnos earthquake (Jan 08, 2013 Mw5.7) corresponding to a strike-slip seis-mogenic fault. For this high quality solution nine strong motion recordings were used (LIA,EFSA, PRK, SMTH, PSRA, ALXA, ALNA, THSA, KVLA) together with three broadbandseismographic recordings (SKY, ALN, ATH). The real time accelerographic stations are also

123

244 Bull Earthquake Eng (2014) 12:237–254

monitored through the PQLX software tool, for evaluating station performance and dataquality (the level of noise across all frequency bands, station malfunctions and quality of theinstallations.

It is noteworthy to show this evolution of strong motion data processing through the timetaking into account the technological advancements. For this reason the accelerogram (coded:PEL1_19841025094917) recorded by the seismic event occurred at SW Peloponessus Greecein Oct. 25, 1984 and time 09:49:17, Mw5.1 and Rep = 15 km, is adopted. This accelero-gram was recorded by a SMA-1 type instrument and was one of the first records, which hadbeen collected by the ITSAK strong motion network. The present record was manually digi-tized and processed by ITSAK (Margaris et al. 1989, 1990) applying the ENEA-ENEL dataprocessing procedure (Tinelli et al. 1985). This processed record is hereafter presented asPEL184_MNL_ITSAK.A. In addition the same record was automatically digitized by CON-TRAVES Italy and processed by the aforementioned ENEA-ENEL data processing procedure(Tinelli et al. 1985). This processed record is hereafter presented as PEL184_AUT_ENEA.A.After the establishment at ITSAK of a scanner based film accelerogram digitization proce-dure using a PC, the same record was automatically processed (Skarlatoudis et al. 2003a).This processed record is hereafter presented as PEL184_AUT_ITSAK.A. Finally withinthe framework of the project National Accelerographic Network and the attempt of a unifiedstrong motion data catalog from 1970 to 2010 the same record was processed by the softwareproposed by Boore (2012) and presented as PEL1_1841025094917. A.

In Fig. 5, 3-component uncorrected acceleration time histories of the various digitizationand data processing techniques applied by different approaches are presented based on theaforementioned characterization. No significant discrepancies in the uncorrected accelerationtime histories can be shown in Fig. 5. Table 1 shows the various digital filtered windows,which have been applied during the various data processing procedures that time period.

In order to see the effect of the digitization and data processing technique in longer periodground motions e.g displacement, the corresponding time histories are depicted in Fig. 6a,b. The uncorrected displacement values of the aforementioned time histories are presentedin Fig. 6a, while the corrected ones, applying the digital filters given in the Table 1 are alsoshown in Fig. 6b. Even in the same record, large differences can be noticed in the uncor-rected displacement time histories applying different digitization and correction techniques.Applying digital filtering for the different correction procedures the corrected displacementamplitudes seem to give more realistic results. The filtered windows were determined empir-ically for different correction techniques. For the case of PEL1_1841025094917.A recordthe most efficient technique based on both time and frequency domain analysis was adopted(Boore 2005, 2012). For this reason, by visual inspection of the FAS of all components ofground motion a preliminary frequency window was defined, in order to estimate the cut-off frequency, fc, of the high-pass filter applied to each record. For this filtered frequencywindow, the displacement time series were calculated for ten specific fc (logarithmicallyequally spaced). After visual inspection the appropriate, fc, was selected in which the dis-placement time history seems to be stable without including long period effects or transients.The filtering procedure involved zero-padding in the beginning and at the end of each recordbased on the order of the applied digital filter and the use of causal band-pass filters for dig-ital noise removal (Boore 2005). This above mentioned approach was applied in the strongmotion record PEL1_1841025094917.A, which is depicted in the last three time histories(3-components of the record) in the Fig. 6b deriving the most reliable time series keeping allthe frequency content.

Using the aforementioned processed strong motion data (Table 1), the uncorrected 5 %-pseudo-acceleration spectral values (Fig. 7; upper line plots) and the FAS (Fig. 7; lower line

123

Bull Earthquake Eng (2014) 12:237–254 245

Fig. 5 The 3-component uncorrected acceleration time histories processed and digitized by various tech-niques applied from different research centers, PEL184_MNL_ITSAK.A: manual digitization by ITSAKand data processing by ENEA-ENEL, PEL184_AUT_ENEA.A: automatic digitization by Contraves Italyand data processing by ENEA-ENEL, PEL184_AUTO_ITSAK.A: automatic digitization by ITSAK and dataprocessing by Kinemetrics, PEL1_ 1841025094917.A: automatic digitization by ITSAK and data processingby Boore (2012)

plots) of the two horizontal components (L-Longitudinal and T-Transversal) of the samestrong motion records are presented. The spectral comparisons are limited in a frequencyrange less than 30 Hz, being compatible with the applying high cut corner frequencies(Table 1). The spectral values of all strong motion records are in a satisfactory agreement

123

246 Bull Earthquake Eng (2014) 12:237–254

Table 1 Strong motion recordsof various digitization and dataprocessing techniques and thecorresponding low-cut ( fLc) andhigh-cut ( fHc) cornerfrequencies

Strong motion record fLc (Hz) fHc (Hz)

PEL184_MNL_ITSAK.A 0.05–0.16 25–27

PEL184_AUT_ENEA.A 0.60–0.70 40–41

PEL184_AUT_ITSAK.A 0.394–0.804 25–27

PEL1_1841025094917.A 0.300 (nslope = 8) –

in the high frequency (short period) range. On the other hand, PEL184_AUT_ENEA.A(red line) and PEL184_AUT_ITSAK.A (blue dashed line) are in a good agreement in shortperiod range but the applied filtered windows are too high and significant information ishigh-cut from the corrected time histories. The record adopting manual digitization proce-dure PEL184_MNL_ITSAK.A (cyan line) presents high spectral values in a long periodrange and this can be attributed in the way of the digitization technique which is now nonusable. The record, which was processed by efficient technique based on both time and fre-quency domain analysis (Boore 2005, 2012), seems to give the most realistic results keepingall the required information of the strong ground motion for seismic design purposes of longperiod structures.

4 Site characterization of accelerographic stations

Accelerographic stations must be well documented in terms of geophysics, geology, geo-morphology, type of instrument and its housing so that recordings can be exploited to theirfull potential. Lack of such metadata may lead to restricted or/and misuse of strong motionrecordings affecting consequently seismic hazard assessment. For instance, previous gener-ation but even more next generation GMPE’s adopted site effect factor either in a simplegeology descriptor or recently in the average shear wave velocity of the upper 30 m, theso-called Vs30.

Accelerographic station characterization in Greece followed the international standards. In‘80s geologic description and material properties (e.g. hardness, age) were taken into accountmainly based on geologic maps of scale 1:50.000. In order to characterize a site geologistsor/and geotechnical engineers, gave a simple classification of soil conditions, namely, “rock”,or “stiff soil”, or “soft soil”, after they were provided for each site a geology description andformation age. Such a kind of classification was mainly used in GMPEs to control influenceof geologic site conditions on strong ground motion (Theodoulidis et al. 1988; Theodoulidisand Papazachos 1992).

In ‘90s in situ active geophysical methods, mainly cross-hole and down-hole, were per-formed for seven accelerograph stations throughout Greece (AUTH-SMF & ITSAK &MEPPW 1996) as well as for selected sites where specific 2D/3D arrays were deployed(e.g. Euroseistest strong motion array). The high cost of performing in situ geophysicalmeasurements was prohibitive for widespread application of the method. However, site char-acterization got into a quantitative approach upgrading thus quality of certain recordings interms of site parameters.

In early ‘00s, another five accelerograph sites were investigated performing in situ cross-hole and down-hole methods (ITSAK & AUTH-GL 2003). During the first decade of 21stcentury a non-invasive geophysical method, namely the ambient noise array measurements,has been performed at several tens of accelerograph stations in Greece, in the frameworkof several EC projects (SESAME 2002; EMERIC-CRINNO 2004; FRANCO-HELLENIC

123

Bull Earthquake Eng (2014) 12:237–254 247

Fig

.6T

he3-

com

pone

ntun

corr

ecte

d(a

)an

dco

rrec

ted

(b)

disp

lace

men

ttim

ehi

stor

ies

digi

tized

and

proc

esse

dby

vari

ous

tech

niqu

esas

desc

ribe

din

Fig.

5

123

248 Bull Earthquake Eng (2014) 12:237–254

0.01 0.10 1.00 10.00

Period (sec)

0.10

1.00

10.00

100.00

1000.005%

-PSA

(cm

/sec

2 )

Manual Digit. by ITSAK 1985Correct. by ENEA-ENEL1985

Auto. Digit. by Contraves 1985Correct. by ENEA-ENEL1985

Auto. Digit. by ITSAK 1992Correct. by ITSAK 1993

Auto. Digit. by ITSAK 2003Correct. by Boore (2012).

Frequency (Hz)

0.001

0.010

0.100

1.000

10.000

100.000

FA

S (c

m/s

ec)

0.01 0.10 1.00 10.00 100.00 0.01 0.10 1.00 10.00 100.00

Frequency (Hz)

Period (sec)

Comp: -L-

Comp: -L-

Comp: -T-

Comp: -T-

0.01 0.10 1.00 10.00

Fig. 7 Upper line uncorrected 5 %-pseudo-acceleration spectral values and (lower line plots) Fourier Ampli-tude Spectra of the two horizontal components (L-Longitudinal and T-Transversal) of the strong motion records,PEL184_MNL_ITSAK.A: manual digitization (cyan line), PEL184_AUT_ENEA.A: automatic digitizationby Contraves Italy and data processing by ENEA-ENEL (red line), PEL184_AUTO_ITSAK.A: automaticdigitization by ITSAK and data processing by Kinimetrics (dashed blue line) and PEL1_1841025094917.A:automatic digitization by ITSAK and data processing by Boore (2012: grey line)

2005; SEISIMPACT-THES 2005; SyNaRMa 2007; NERIES 2007; ITSAK.GR 2009). Ambi-ent noise measurements were applied at selected sites where geophysical profiles existedbased on “reference” VS30 estimation methods (e.g. cross-hole, down-hole). Similaritiesand differences as well as advantages or disadvantages of the non-invasive method are dis-cussed (NERIES, JRA4 Final Report 2010; Hobiger et al. 2013).

In early ‘10s on-going project is set up for site characterization in Greece (Theodoulidiset al. 2013b; Stewart et al. 2013) in order to cover as much as possible representative siteconditions. This new approach based on geology description, topographic and terrain basedproxies of VS30, aims to expand site characterization to all accelerograph stations in Greece.For this purpose, a dataset has been created, including more than three hundred sites in Greecewhere shear wave velocity profiles with depth, Vsz, were available along with informationof local geology, in scale 1:50.000, topography and terrain category. Classification of localgeology in groups according to criteria such as geologic age, depositional environment andmaterial gradation, is undertaken to find any possible correlation with VS30. In addition,

123

Bull Earthquake Eng (2014) 12:237–254 249

Fig. 8 Site information database for accelerograph stations in Greece (example of Korinthos station)

correlations of VS30 with other proxies descriptive of topography and terrain is consideredto investigate their relative effectiveness for VS30 estimation (Stewart et al. 2013).

123

250 Bull Earthquake Eng (2014) 12:237–254

Fig. 9 Cumulative number of accelerographs of the national strong motion network in Greece since early‘70s (left); Cumulative number of strong motion recordings during the period 1973–2009 (right)

In order to accomplish the on-going site characterization work a database that has beeninitially developed during the NERIES EC project (monographs.itsak.gr), greatly facilitatesmetadata compilation and organization. This database is continuously fulfilled by new siteinformation regarding strong motion stations in Greece (Fig. 8).

5 Strong motion data during the period 1973–2009

During the last 40 years several moderate to large magnitude events, some of them destructive,occurred in Greece (among others; Lefkas 1973; Thessaloniki 1978; Corinth 1981; Kalamata1986; Griva 1990; Kozani 1995; Aegion 1995; Athens 1999; Lefkas 2003; Kythera 2006;Achaia-Ilia 2008). These events together with aftershocks and smaller magnitude ones (M <

5.0) provided a large dataset of strong motion records. Certainly, the record number dependson the strong motion instruments installed at the time of the earthquake occurrence as wellas on their sensitivity.

In Fig. 9(left), the cumulative number of accelerographs of the national strong motionnetwork in Greece is shown. A rapid increase of instruments in early ‘80s is observed. Then,a gradual increase of accelerographs during the period 2002–2009 is apparent while after2009 the rate of the cumulative number is much higher. In Fig. 9(right) the cumulative numberof recordings acquired during the period 1973–2009, exhibits a respective increase rate. Inearly ‘80s a rapid increase of records till almost mid of ‘90s is observed while a secondturning point of increase is placed in 2002.

In Fig. 10 (presented by D. Boore in ITSAK 2013), the distribution of the recordings, for theperiod 1973–2009, in magnitude–epicentral distance space, is shown. The data set consistsof about 900 recordings homogeneously processed and exhibits sufficient distribution for1 km ≤ R ≤ 200 km and for M ≤ 5.2. For M > 5.2 there is reasonable distribution of datafor 15 km ≤ R ≤ 200 km. The lack of larger magnitude events (M > 5.2) data in the nearfield is evident.

The magnitude in the catalogue is equivalent moment magnitude (M) and the distanceis epicentral and hypocentral. About 70 % of the 900 recordings of the data come fromevents with known focal mechanism. The majority of recordings (more than 70 %) come

123

Bull Earthquake Eng (2014) 12:237–254 251

Fig. 10 Distribution of strong motion data in M–R space (Boore 2013, personal communication)

from instruments installed in ground level or in the basement of relatively light structures(less than 4 storeys).

In Fig. 11 (presented by D. Boore in ITSAK, 2013) the distribution of strong motionrecordings in ground motion intensity measures (GMIM) versus epicentral distance (REPI)

is shown. As GMIMs the PGV, PGA and spectral accelerations for periods T = 0.1, 0.2,1.0 and 2.0 s have been chosen. Distance decay is clearly defined in all plots while strongermagnitude dependence for longer periods is evident. However, a few outliers apparent in allplots need to be carefully checked.

6 Conclusions

The main goal of installing strong motion accelerographs is to acquire important strong-motion data from significant earthquakes. The planning and deployment of the strong motionnetwork in Greece were mainly based on various advancements, which have been carriedout during the last four decades in Greece. A dense Hellenic accelerographic network wasdeployed in the whole territory, with the next target being to install additional stations inorder to study near field and site effects issues. The homogeneous strong motion recordingsprocessing with adopted techniques, contribute to a better adjustment of the collected datawhile compared to past commonly used techniques the recently adopted ones can afford morereliable information for geoscientists and engineers.

The new generation accelerographic network in Greece may provide high quality accelero-metric data which in turn are going to expand the frontiers of research in the fields of engineer-ing seismology and earthquake engineering. Given the research interest of the two research

123

252 Bull Earthquake Eng (2014) 12:237–254

Fig. 11 Distribution of data in ground motion intensity measures (GMIM) and epicentral distance (REPI)space, for various magnitude-ranges (Boore 2013, personal communication)

organizations (EPPO-ITSAK and NOA-IG) it is assured that the data will be exploited at thehighest level. In addition, data dissemination may contribute worldwide to research in seis-mology and earthquake engineering. The continuous recording of ground motion guaranteesthe accumulation of acceleration values by many instruments at large distances even frommoderate magnitude earthquakes. This will contribute to the improvement of the empiricalpredictive equations. An example is the recordings of recent earthquakes with 5.5 ≤ M ≤ 6.0occurred at the borders of Greek territory and recorded at distance of about 600 km. Further-more it will also add to the increase of the national seismological network density contributingto the improvement of reliability for earthquake parameter estimation, especially for earth-quakes close to populated areas. Furthermore, estimation of ground motion intensity measures(e.g. PGA, PGV, Spectral values) in real time is of very high operational importance for localor/and regional authorities in Greece in making almost real time effective decisions towardsseismic risk mitigation measures.

The future installation of digital free-field accelerographs nearby to instrumented R/Cstructures, (ITSAK project under development), will allow research in the influence of struc-

123

Bull Earthquake Eng (2014) 12:237–254 253

tures and soil-structure interaction in recorded ground motions. Such strong ground motionstations are to accurately record the combined effects of earthquake source, propagation path,and site effects within the range of amplitudes (.001–2 g) and frequencies (0–50 Hz). Thisdata, needed for the various public safety, engineering and scientific applications, may sig-nificantly improve estimation of seismic code actions and in turn contribute to seismic riskreduction.

Acknowledgments Many people helped with their contribution in the development of the Hellenic Strongmotion Network. Among those we wish to thank S. Zacharopoulos Civil. Eng. in charge of EPPO-ITSAKnetwork, K. Konstantinidou IT, A. Marinos and N. Adam technicians, for their assistance in deploying andmaintaining the strong motion network. The technical staff of EPPO—ITSAK and NOA-GI is greatly appre-ciated for their efforts to keep the national network in an operational status. I. Kalogeras is gratetful to Drs. N.Melis and C. Evangelidis (NOA-GI) for the productive discussions and their crucial contribution to incorporatethe strong motion instruments within the operation of NOA-GI seismographic network. The researchers, Drs.V. Lekidis, Ch. Karakostas from EPPO-ITSAK and G. Stavrakakis from NOA-IG, for their important partici-pation in the different phases of those projects are strongly acknowledged. We are profoundly indebted to Profs.B. Papazachos (AUTh-Greece), late D. Papastamatiou (NTUA-Greece), late I. Drakopoulos (KUA-Greece),S. Anagnostopoulos (TUP-Greece), late N. N. Ambraseys (ICST-UK) and Dr. D. Boore (USGS) for their sig-nificant contribution in the strong motion instrumentation programs in Greece. We would like to thank, Prof.J. Stewart (UCLA), Prof. N. Klimis (DUTH-Greece), Profs. G. Mylonakis and G. Athanasopoulos (TUPatras-Greece), Dr. R. Kayen (USGS) and Pr. S. Foti (Pol. Torino—Italy) for their valuable help in site classificationof Greek accelerographic sites. Dr. D. Boore and the anonymous reviewer for their careful reading and con-structive comments improving the manuscript are deeply acknowledged. The financial support from EPPO forfunding the purchase of accelerographs is also gratefully acknowledged. This research has been co-financed bythe European Union (European Social Fund—ESF) and Greek national funds through the Operational Program“Education and Lifelong Learning” of the National Strategic Reference Framework (NSRF)—Research Fund-ing Program: Thales. Investing in knowledge society through the European Social Fund, THALIS—AUTH,85330, by THALIS-UC 80198 and by the Black Sea project CN18316, MIS-ETC2614.

References

Akkar S, Boore DM, Gülkan P (2005) An evaluation of the strong ground motion recorded during the May 1,2003 Bingöl, Turkey, earthquake. J Earthq Eng 9:173–197

Ambraseys NN, Smit P, Douglas J, Margaris B, Sigbjornsson R, Olafsson S, Suhadolc P, Costa G (2004)Internet site for European strong-motion data. Boll di Geofisica Teor Ed Appl 45(3):113–129

AUTH-SMF & ITSAK & MEPP (1996) Final Report on “Study of the influence of local site conditions,geomorphology and dynamic soil-structure interaction on recordings of the national network of accelero-graphs”, Funded by EPPO

Boore DM (2001) Effect of baseline corrections on displacements and response spectra for several recordingsof the 1999 Chi-Chi, Taiwan, earthquake. Bull Seismol Soc Am 91:1199–1211

Boore DM (2005) On pads and filters: processing strong-motion data. Bull Seismol Soc Am 95:745–750Boore DM (2012) TSPP—a collection of FORTRAN programs for processing and manipulating time series,

U.S.G.S Open File Report 2008–1111, Ver4.3,09 Oct 2012, 47 ppCloud WK, Hudson DE (1961) A simplified instrument for recording strong motion earthquakes. Bull Seism

Soc Am 51(2):159–174Drakopoulos J, Rousopoulos A (1973) Analysis of strong motion records of the Cephalonian shock occurring

on Sept. 17, 1972 and of its larger aftershock. Annali di Geofisica, 1973: XXVI, 23 ppGalanopoulos A, Drakopoulos J (1974) A T-phase recorded on an accelerogram. Bull Seism Soc Am 64(3):717–

719Hobiger M, Cornou C, Wathelet M, Di Giulio G, Knapmeyer-Endrun B, Renalier F, Bard P-Y, Savvaidis A,

Hailemikael S, Le Bihan N, Ohrnberger M, Theodoulidis N (2013) Ground structure imaging by inversionsof Rayleigh wave ellipticity: sensitivity analysis and application to European strong-motion sites. GeophysJ Int. doi:10.1093/gji/ggs005 (in press)

ITSAK & AUTH-GL (2003) Final report on “Study of the influence of local site conditions, geomorphologyand dynamic soil-structure interaction on recordings of the national network of accelerographs”, Fundedby EPPO

Kalogeras IS (2002) A strong motion data databesa in Greece. Nat Haz 26:265–277

123

254 Bull Earthquake Eng (2014) 12:237–254

Kalogeras I, Loukatos D, Melis N, Boukouras K, Stavrakakis G (2008) The Hellenic accelerographic net-work: Enhancements towards automated real-time data acquisition. 1st Euro-Mediterranean meeting onAccelerometric data exchange and archiving, NERIES NA5, JRA4, Grenoble, March 10–11

Kinemetrics Inc. (1999) KMI strong motion, analyst, Doc302415, CMargaris B (1986) Digitizing errors and filters. Report ITSAK, Report: 86-03Margaris B, Papastamatiou D, Theodulidis N (1989) ITSAK: strong motion data processing. Report ITSAK,

Report: 89-02Margaris B, Marinos A, Theodulidis N, Vorrias E, Akritidis A (1990) ITSAK strong motion bulletin (1980–

1988). Report ITSAK, Report: 90-01NERIES: Network of Research Infrastructures for European Seismology (2010) European project. http://

www.neries-eu.org/Nigbor RL, Kodama DA (1990) Scanner-based film accelerogram digitization system user’s manual. kinemet-

rics/systems, 1.1-6.2Skarlatoudis AA, Papazachos CB, Margaris BN (2003a) Determination of noise spectra from strong motion

data recorded in Greece. J Seismol 7:533–540Skarlatoudis AA, Papazachos CB, Margaris BN, Theodoulidis NP, Papaioannou ChP, Kalogeras I, Scordilis

EM, Karakostas VG (2003b) Empirical peak ground motion predictive relations for shallow earthquakesin Greece. Bull Seismol Soc Am 93:2591–2603

Skarlatoudis A, Theodoulidis N, Papaioannou Ch, Roumelioti Z (2004) The dependence of the peak horizontalacceleration on magnitude and distance for small magnitude earthquakes in Greece 2004. In: 13th WCEE,Vancouver, Canada, 1–6 Aug 2004, Paper 1857

Stewart J et al (2013) Compilation of a local vs profile database and its application for inference of Vs30 fromGeologic- and Terrain-Proxies. Bull Seismol Soc Am (submitted for publication)

Tinelli G, Zonetti L, Basili M, Berardi R (1985) Manual d’ uso del codice PRACMA con esempi di applicazione.Commissione ENEA-ENEL, Roma 1985

Theodoulidis N, Margaris B, Papastamatiou D (1986) Planning of strong motion network. Report ITSAK,Report: 86-04

Theodoulidis N, Margaris B, Papastamatiou D (1988) Influence of geologic site conditions on strong groundmotion. Report ITSAK 88-02, 76 p (in Greek)

Theodoulidis N, Papazachos B (1992) Dependence of strong ground motion on magnitude-distance, sitegeology and macroseismic intensity for shallow earthquakes in Greece: I, Peak horizontal acceleration,velocity and displacement. Soil Dyn Earth Eng 11:387–402

Theodoulidis N, Kalogeras I, Papazachos C, Karastathis V, Margaris B, Papaioannou Ch, Skarlatoudis A(2004) HEAD 1.0: a unified hellenic accelerogram database. Seismol Res Lett 75:36–45

Theodoulidis N, Papaioannou Ch, Margaris B, Savvaidis A, Rovithis M, Lekidis V, Karakostas C, Demos-thenous M, Salonikios T, Makarios T, Makra K, Morfidis K, Konstantinidou K, Kanaki M, ZacharopoulosS, Adam N, Marinos A, Akritidis A (2013a) Final Report of the project “Funding of ITSAK and Geo-dynamic Institute of NOA, for the creation and launch of the national network of Accelerographs”, 35 p(in Greek)

Theodoulidis N, Savvaidis A, Margaris B, Zargli E, Klimis N, Lazaridis T, Stewart JP, Athanasopoulos G,Pelekis P, Vlachakis V, Batilas A (2013b) Geologic and terrain based proxy estimation of Vs30 in Greece,Proceedings, Joint Assembly of IAHS-IAPSO-IASPEI, 22–26 July (abstract)

Trifunac MD, Lee VW (1973) Routine computer processing of strong- motion accelerograms. EERL, 73-03Wald D, Worden B, Quitoriano V, Pankow K (2006) ShakeMap® Manual, USGS, 156 pp

123