EXPERIMENTAL STUDY OF TOPOGRAPHIC AMPLIFICATION USING THE ISRAEL SEISMIC NETWORK

Earthquakes and microtremor records are used for estimating the site response of hard rock sites comprising four three-component seismic stations which operate as part of the Israel Seismic Network. The response functions are determined by implementing the horizontal-to-vertical component spectral ratio of earthquake shear-waves (receiver function estimates) and microtremors (Nakamura's estimate) observed simultaneously at the site. The sites of seismic stations ATZ (Mt. Atzmon), MBH (Mt. Berech) and MRNI (Mt. Meron) exhibit amplification attributed to topography effects. At ATZ, within the 1.3–2.0 Hz range, the amplification is in the order of factor 4. At MBH amplification levels of 3.0–3.5 are observed in the frequency range 1.5–4.0 Hz. Station MRNI exhibits a relatively strong amplification effect (up to 4) in the frequency range of about 2.5 to 3.5 Hz. Slight amplification around 5 Hz is observed at ATR (the proposed site for a nuclear power plant). These effects were correlated with the thickness of the weathered layer above unweathered chalk. A comparison between the amplification factor observed during earthquakes and those inferred from microtremors shows that these are, in general, in agreement. However, details of the spectral ratios from different microtremor recordings are not exactly the same. Differences appear mainly in the frequency at which the maximum amplification occurs. These observations demonstrate the usefulness of non-reference technique in estimating the topographical effects of ground shaking. These methods may be used in the process of seismic hazard assessment for ridges and mountain tops, common sites for settlements, communication relay stations, bridges, rope-drive and power transmission towers.     

MICROZONATION OF THE EXPECTED SEISMIC SITE EFFECTS ACROSS PORT VILA,VANUATU

The city of Port Vila, Vanuatu, is located in one of the most active seismic regions on earth. Earthquakes are felt frequently and, due to very rapid plate convergence rates, return period of large earthquakes (M>6) in the New Hebrides Benioff zone can be less than 10 years. Even though Port Vila does not lie on an identified seismic fault zone, strong motions by nearby earthquakes have to be expected due to the city's geographical location close to the plate boundary of New Hebrides convergence zone. An accurate estimation of the seismic ground motion across the city is of prime importance for urban developments and mitigation of earthquake risk. Following many examples of monitored strong earthquakes in the current century, it is evident that the local site effects may have a dominant contribution to the intensity of damage and destruction. In this study we focussed on the first stage of associating site effects and seismic hazard by preparing a microzonation map for Port Vila. The seismic microzonation of the city has been carried out to provide a detailed map of the zones that exhibit site effects in terms of resonance frequencies and approximated amplification of the ground shaking. Having in mind that these data will be used in improving building design to sustain strong ground motions, our analysis is limited to the frequency band of 1–10 Hz, corresponding to the expected resonance of different types of buildings in Port Vila. The Nakamura technique has been used to estimate site amplification effects from single station noise recordings. Interestingly, excluding one site located on an old dump zone, the amplification factors at about the 100 sites surveyed in Port Vila remain below 3 with an average well below 2 in the 1 to 10 Hz frequency band. These results suggest that there is no significant Vs velocity change in consequently layered material and that the uppermost sedimentary layers in the surveyed down town area are relatively thin. These observations are in agreement with the mapping of limestone terraces throughout Port Vila area. However, both the surface geology and results from seismic zonation indicate a thicker (up to several tens of meters) sedimentary cover around the Bauerfield airport and in the Mele terrace zone. Low resonance frequencies (around and below 1 Hz) and amplification factor of the order of 5 were observed over this large area, immediately outside Port Vila. Any building development in this area should take these results into account.     

SEISMIC HAZARD ASSESSMENT IN FLORENCE CITY ITALY

A seismic hazard analysis of Florence city was performed in the frame of a project concerning the dynamic behaviour of cable-stayed bridges. Both a probabilistic approach and a methodology based on the use of a local macroseismic catalogue were applied. A local catalogue was expressly compiled for this purpose, to collect the macroseismic intensities actually observed at the site as a result of past earthquakes. This sort of catalogue is an independent tool to verify the assumptions of the probabilistic approach (seismic zoning, earthquake recurrence relation, attenuation model), though it can supply results in terms of macroseismic intensity only and reflects the effective seismic history at the site, without taking into account any variability. The Cornell' methodology was used to assess probabilistic hazard in terms of macroseismic intensity, peak ground acceleration, peak ground velocity, and pseudovelocity uniform response spectra. The local catalogue points out level VII of the Mercalli-Cancani-Sieberg scale (MCS) as the maximum intensity historically observed in Florence. The probabilistic approach leads to the consideration of intensity VIII MCS as the maximum credible for the city. The probabilistic analysis in terms of ground motion was performed using attenuation relations estimated for alluvium sites, since the geology of Florence area is represented by fluvial and lacustrine deposits of various thickness. Peak ground acceleration values with 90% non exceedence probability in 50 and 500 years are respectively 145 and 219 cm/s's for a shallow alluvium site, and 95 and 157 cm/s's for a deep alluvium site; the corresponding peak ground velocity values for sites located on alluvium are 6.41 and 11.76 cm/s. Uniform response spectra are provided for shallow and deep alluvium sites, according to frequency-dependent attenuation relations estimated from strong Italian earthquakes.     

SITE AMPLIFICATION FACTOR IN ANCHORAGE,ALASKA BASED ON LONG PERIOD MICROTREMORS

Long period microtremors with periods ranging from 0.5 to 10 seconds were measured in the Anchorage metropolitan area. Two horizontal components of motion were recorded at 81 sites uniformly distributed throughout the basin with spatial resolution of about 2 km. Recording at each site was done for 300 seconds with a sampling rate of 20 Hz. Repeated measurements were performed at a bedrock reference site simultaneously with the measurements in the field. The measurements were completed in six days. In addition, multiple recordings were obtained concurrently at the reference bedrock site and a sediment site. Based on these measurements the Fourier spectra were calculated for each of the site. Ground motion amplification is determined in terms of spectral ratio of horizontal spectral amplitudes at a sediment site and the reference bedrock site. Mean spectral ratio contours were evaluated for different period bands. The results show that for period band 3 to 5 seconds the spectral ratio contours agree well with the ground failure susceptiblity map of Anchorage.     

SEISMIC AMPLIFICATION AT A SOFT SOIL SITE WITH LIQUEFIABLE LAYER

It is well known that local soil conditions play a key role in the amplification of earthquake waves. In particular, a liquefiable shallow soil layer may produce a significant influence on ground motion during strong earthquakes. In this paper, the response of a liquefiable site during the 1995 Kobe earthquake is studied using vertical array records, with particular attention on the effects of nonlinear soil behaviour and liquefaction on the ground motion. Variations of the characteristics of the recorded ground motions are analysed using the spectral ratio technique, and the nonlinearity occurring in the shallow liquefied layer during earthquake is identified. A fully coupled, inelastic finite element analysis of the response of the array site is performed. The calculated stress-strain histories of soils and excess pore water pressures at different depths are presented, and their relations to the characteristics of the ground motions are addressed.     

VERTICAL EARTHQUAKE GROUND MOTION RECORDS: AN OVERVIEW

Recent research clearly shows the importance of including the vertical component of earthquake ground motion in seismic analysis and design. In addition, pioneering studies [e.g., Elnashai and Papazoglou (1997)] have explored and documented the characteristics of available near-field vertical ground motion records. As a follow-up, this paper complements earlier studies, and investigates additional far-field records and available downhole array vertical motion records. A total of 111 free-field strong motion records (from California) and available downhole array records are employed. Compared to near-field records, far-field records generally contain more energy at longer periods. Based on the available data, response spectra are presented for near-field and far-field records respectively. The currently scarce downhole-array vertical motion records show that significant amplification may occur within the top 10-20 m soil layers. A simple one-dimensional (ID) vertical wave propagation model did not appear adequate for modelling the observed downhole array response. In using such a simplified model, very high viscous damping in the range of 15-25% was needed to match the recorded downhole vertical response, even for small tremors. Additional data and research are required [Beresnev et al., 2002] towards the development of a rational vertical motion site response analysis procedure.     

Design Earthquake Ground Motion Prediction for Perth Metropolitan Area with Microtremor Measurements for Site Characterization

Perth is the largest city in Western Australia and home to three-quarters of the state's residents. In recent decades, there have been a lot of earthquake activities just east of Perth in an area known as the South-West Seismic Zone. Previous numerical results of site response analyses based on limited available geology information for PMA indicated that Perth Basin might amplify the bedrock motion by more than 10 times at some frequencies and at some sites. Hence, more detailed studies on site characterization and amplification are necessary. The microtremor method using spatial autocorrelation (SPAC) processing is a useful tool for gaining thickness and shear wave velocity (SWV) of sediments and has been adopted in many previous studies. In this study, the response spectrum of rock site corresponding to the 475-year return period for PMA is defined according to the probabilistic seismic hazard analysis (PSHA) based on the latest ground motion attenuation model of Southwest Western Australia. Site characterization in PMA is performed using two microtremor measurements, namely SPAC technique and H/V method. The clonal selection algorithm (CSA) is introduced to perform direct inversion of SPAC curves to determine the soil profiles of representative PMA sites investigated in this study. Using the simulated bedrock motion as input, the responses of the soil sites are estimated using numerical method based on the shear-wave velocity vs. depth profiles determined from the SPAC technique. The response spectrum of the earthquake ground motion on surface of each site is derived from the numerical results of the site response analysis, and compared with the respective design spectrum defined in the Australian Earthquake Loading Code. The comparison shows that the code spectra are conservative in the short period range, but may slightly underestimate the response spectrum at some long period range.     

EMPIRICAL EVALUATION OF SITE EFFECTS BY MEANS OF H/V SPECTRAL RATIOS AT THE LOCATIONS OF STRONG MOTION ACCELEROMETERS IN ISRAEL

Acceleration data from local and regional earthquakes is of prime importance in evaluating the seismic hazard. Consequently, strong motion accelerometers are currently installed at more than 60 locations in Israel. We have explored the possibility of site amplification effects at 10 sites where local earthquakes triggered strong motion accelerometers by integrating empirical and analytical estimations. Implementing H/V spectral ratio techniques using 15 accelerograms from nine earthquakes, 105 seismograms shear-wave records of 35 local and regional earthquakes and seismograms of microtremors were used in the empirical evaluations. The subsurface models were constructed by integrating available geological and geophysical information at the analysed site with empirically evaluated site response functions. Amplification effects of factor 3-6 are observed at various frequencies in the 0.8-6.0 Hz band. Through the analysis process it became evident that the instant availability of many useful time windows of microtremors provides systematic estimations of the fundamental resonance frequency of each site and their associated amplification levels, which are similar to those obtained from H/V spectral ratios of seismograms and accelerograms and to those inferred from the subsurface geology. Analytical transfer functions should be reviewed with respect to empirical site response evaluations. Estimations that are based on only one approach may be totally misleading.     

EFFECTS OF SOIL LAYERING ON THE CHARACTERISTICS OF BASIN-EDGE INDUCED SURFACE WAVES AND DIFFERENTIAL GROUND MOTION

This paper studies the effects of soil layering in the basin based on the characteristics of basin-edge induced surface waves and associated differential ground motion. Seismic responses of various basin-edge models were simulated using software based on parsimonious finite difference staggered grid approximation of 2.5D eiastodynamic wave equation. Seismic responses of various models with different number of soil/soft rock layers but for a fixed thickness of deposit, fundamental frequency and impedance contrast revealed a decrease of surface wave amplitude with an increase in the number of layers in the basin. Shifting of dominant frequency towards the higher values was obtained with an increase of number of layers. An increase of dispersion of surface waves with an increase of number of soil layers in the basin was observed. A minor increase of Rayleigh wave velocity with an increase of number of soil layers was also obtained, but in the case of Love wave it was almost negligible.

Spectral analysis of the edge-induced surface waves revealed that the anomalous earthquake intensity may arise in a zone of width of 2.5–3.0 km, parallel to basin-edge and at an offset of 0.5–0.7 km from the edge. Maximum horizontal differential ground motion (HDGM) developed by Love wave (≈4.9×10^?2) was more than that of Rayleigh wave (≈9.4×10^?3). Large variation in HDGM caused by the surface waves was obtained with a change in the number of layers in the basin and maximum HDGM was observed when there were only two layers in the basin. It was inferred that the effect of soil layering in the basin was more on the Rayleigh wave as compared to the Love wave. Development of large HDGM near the basin-edge and its dependency on the number of soil layers reveals that basin-edge induced surface waves need special attention during seismic microzonation or seismic hazard prediction.     

AN EVALUATION OF THE STRONG GROUND MOTION RECORDED DURING THE MAY 1, 2003 BINGÖL TURKEY,EARTHQUAKE

An important record of ground motion from a M6.4 earthquake occurring on May 1, 2003, at epicentral and fault distances of about 12 and 9 km, respectively, was obtained at a station near the city of Bingöl, Turkey. The maximum peak ground values of 0.55 g and 36 cm/s are among the largest ground-motion amplitudes recorded in Turkey. From simulations and comparisons with ground motions from other earthquakes of comparable magnitude, we conclude that the ground motion over a range of frequencies is unusually high. Site response may be responsible for the elevated ground motion, as suggested from analysis of numerous aftershock recordings from the same station. The mainshock motions have some interesting seismological features, including ramps between the P-and S-wave that are probably due to near- and intermediate-field elastic motions and strong polarisation oriented at about 39 degrees to the fault (and therefore not in the fault-normal direction). Simulations of motions from an extended rupture explain these features. The N10E component shows a high-amplitude spectral acceleration at a period of 0.15 seconds resulting in a site specific design spectrum that significantly overestimates the actual strength and displacement demands of the record. The pulse signal in the N10E component affects the inelastic spectral displacement and increases the inelastic displacement demand with respect to elastic demand for very long periods.     

ATTENUATION RELATIONSHIP FOR LOW MAGNITUDE EARTHQUAKES USING STANDARD SEISMOMETRIC RECORDS

Northwestern Italian weak-motion data were used to study attenuation characteristics of horizontal peak ground acceleration (PGA) and horizontal peak ground velocity (PGV) from earthquakes of local magnitudes (M _l ) up to 5.1. Data have been provided by the RSNI (Regional seismic network of Northwestern Italy) and RSLG (Regional seismic network of Lunigiana-Garfagnana) waveform database. The database consists of more than 14000 horizontal components recorded in the period 1999-2002 by both broadband and enlarged band seismometers. The accuracy of the procedure used to extract PGA values from the velocity recordings was verified comparing observed and derived PGA values at station STV2, which was equipped with both a temporary K2 Kinemctrtcs accelerometer and Guralp CMG40 broadband sensor. The attenuation of both peak ground acceleration and peak ground velocity was found to be logarithmically distributed with a strong attenuation for low distances (less than 50 km) and low M _l values (<3.0). The resulting equations are:

Log(PGA)=?3.19+0.87M?0.042M ²?1.92 Log(R)+0.249S,

Log(PGA)=?4.23+0.76M?0.018M²?1.56 Log(R)+0.230S,

where PGA is expressed in g, PGV is expressed in m/s, M is local magnitude, R is the hypocentral distance in kilometers and S is a dummy variable assuming values of 0 and 1 for rock and soil respectively. For increasing distance and magnitude, both PGA and PGV values show a linear distribution. The validity range of the obtained attenuation relationships is 0–200 km for distances and M _l up to 4.5. Sensitivity studies performed by analysis of residuals, showed that predicted PGA and PGV values are stable with respect to reasonable variations of the model and distances providing the data. Comparisons with attenuation relationships proposed for Italian region, derived from strong motion records, are also presented.     

Seismic Microzonation and Damage Assessment of Bam City,Southeastern Iran

This article aims to study the dynamic characteristics of soil in Bam, southeastern Iran. Fundamental frequency and amplification factor of soil sediment were estimated by microtremor measurement. This procedure was performed at 49 sites in the city. Two 120-second data were recorded at each site. Segmental cross spectra were applied to calculate spectrum, Nakamura's method (H/V) was used for analyzing data and fundamental frequency and amplification factor values were derived. Iso-frequency and iso-amplification maps of the city were prepared. Results show that soil type in Bam city is mainly stiff, although amplification factor is relatively large. There is a short period zone in northwest to southeast direction. Sediment depth was estimated using a correlation between fundamental frequency and sediment thickness. Results obtained from microtremors were compared to the geotechnical boreholes and it was shown that microtremors can be used for a rough estimation of sediment thickness. Damage distribution map of Bam due to the Bam earthquake on December 26, 2003 was prepared by both field and aerial investigation. Quality of buildings in Bam was evaluated, as a result a crude zoning map on the building quality was prepared. Finally, it was concluded that several factors have contributed to damage intensity variation; earthquake characteristics, local site effect, and buildings' quality. All mentioned parameters affected the damage variation.     

Archaeological Site or Natural Marine Community? Excavation of a Submerged Shell Mound in Ninigret Pond,Rhode Island

ABSTRACT

Submerged shell midden sites and natural shell deposits can have similar characteristics and can be difficult to distinguish archaeologically. We excavated two test units from a large (at least 35 m×70 m) submerged shell mound in Fort Neck Cove in southern Rhode Island to assess whether it was natural or cultural in origin. This mound had been recognized as a potential archaeological feature as early as the 1970s. Excavation, radiocarbon dating, and subsequent laboratory analysis of excavated materials suggest that the mound was a natural oyster reef rather than a submerged archaeological site. No artifacts were found; there was no clear evidence for human modification of any shells; small species that would not have been targeted as food were present; and δ¹³C values of oyster shells from the mound were consistent with freshwater input into their growth environment, suggesting that they grew in an estuarine environment that did not exist prior to the inundation of the ponds. The stratigraphically oldest radiocarbon date we could obtain (430–190 cal BP, 2σ range), from 70 cm below the pond floor, placed deposition of shells at least 3,000 years after the inundation of the pond. The excavation methods that we used and the process of testing, irrespective of whether the feature is cultural, are valuable contributions to the methodological literature on submerged site archaeology and help provide insight for other researchers working to discern natural from cultural shell midden sites.     

SEGMENTAL CROSS-SPECTRUM AS A NEW TECHNIQUE IN SITE RESPONSE ESTIMATION USING SPECTRAL RATIO ANALYSIS

Different aspects of spectral analysis for site response evaluation are investigated in this study. The segmental cross-spectrum is proposed in spectral analysis of earthquake ground motions. The performance of segmental cross-spectrum in contrast with the conventional methods is investigated through the mathematical modelling, numerical analysis and application to earthquake data recorded at Chiba and Shinfuji downhole arrays in Japan. In analysis of earthquake data, the soil amplification function is identified using both uphole/downhole (U/D) and H/V spectral ratios. The advantage of seg-mental cross-spectrum is assessed by comparing identified amplification functions using different spectral methods and theoretical soil response. The reliability of site response estimations obtained by H/V spectral ratio using segmental cross- and Fourier spectra is also examined by means of cross-validation with the U/D spectral ratio of earthquake motion and theoretical soil response. Furthermore, the application of segmental cross-spectrum in nonlinear soil response is examined by comparing the amplification function of weak and strong motions for both methods. The results validate the advantage of segmental cross-spectrum in both linear and nonlinear soil response, particularly, when it used with H/V technique.     

Site Effects Analyses in the Old City Center of Trieste (NE Italy) Using Accelerometric Data

The influence of local geology and soil conditions on the intensity and the amplification of ground shaking are well known. Part of the old city center of Trieste is built on the site of a former salina, placed at a river mouth and is characterized by soft sediments several tens of meters thick. A new accelerometric station has been recently installed in a historical building, in order to analyse earthquake-induced site amplifications. This station has recorded five regional earthquakes and the related records are compared to those obtained at a nearby bedrock-installed accelerometeric station. Fourier and response spectra for all components are computed and both the H/V ratio and the reference station techniques are used to assess site effects. Noise measurements performed in the historical building, where the accelerometer is located, confirm these results. Relevant amplifications are detected in the frequency range of 2–4 Hz, particularly important for the type of buildings present in this part of the city.     

EUROCODE 8 DESIGN RESPONSE SPECTRA EVALUATION USING THE K-NET JAPANESE DATABASE

The Eurocode 8 (EC8) currently proposes two standard shapes for the design response spectra. Type 1 spectra are enriched in long period and are suggested for high seismicity regions. Conversely, Type 2 spectra are proposed for low to moderate seismicity areas (like France), and exhibit both a larger amplification at short period, and a much smaller long period contents, with respect to Type 1 spectra. These propositions, however, were constrained using few events mostly recorded on analogical instruments. In the present study, we use the Japanese high quality digital K-net array in order to evaluate the proposed ECS response spectra. Furthermore, all K-net stations have geotechnical characterisation. We first constructed a database of shallow events, depth less than 25 km, to avoid subduction related records. The database spans six years of seismicity from 1996 until 2003. Thus, 591 events were selected with moment magnitude between 4 and 7.3, recorded at 691 stations, giving a total of 6812 two horizontal components accelerograms. Using these records, we computed spectral ground-motion prediction equations and we used them to review the shape of the proposed EC8 spectra. In particular, we studied the plateau-PGA ratio level, the period interval where this plateau is constant, and site amplification effects. The results show surprisingly that the Type 2 rock better envelope the Japanese data. Another interesting observation is that the K-net data corresponding to all soil classes are rich in short periods around 0.1 s. This characteristic has not been observed in other worldwide databases. Normalised empirical predictions show a widening of the plateau as the soil conditions degrade. This suggests that the Type 2 EC8 spectra do not cover enough the long periods for EC3-soil classes C, D and E. Finally, the computed ground-motion prediction equations show that the peak ground acceleration (PGA) is nearly invariant to the soil conditions. Soil effects are mainly seen in the shape and plateau level.     

RADIOCARBON DATING OF KILL SITES

Age determinations of kill sites, where charcoal is generally absent, has usually depended on radiocarbon analyses of bone material. However, it has been seen with known-age samples that these dates are often in error due to contamination by plant products or ground water carbonates. The use of soil samples from strata both above and below, as well as the level containing the artifacts, offers important advantages. The soil profiles present sequences of dates, whereas the bones are only individual samples. The distribution of the soil date values can indicate upper and lower limits for the age of the kill site. The Taima-taima paleo-indian site in the state of Falcón, Venezuela is presented as an illustration of the application of radiocarbon dating of soil to archaeological problems. A total of 22 samples were processed. It is seen that soils from two profiles indicate a dating of 13 000 years for the kill. Taking into account reasonable extents of error, uncertainties of ± 2000 years can be estimated. The non-carbonate fraction dates are in agreement with the earth samples, but this is probably due to the organic matter extracted from these materials being mostly soil contamination in the porous bone matrix. Fluorine measurements on the bones confirm the Pleistocene origin.     

Numerical Evaluation of Seismic Soil Pressures on Rigid Walls Fixed to the Bedrock

Seismic soil pressures developed on a 7 m rigid retaining wall fixed to the bedrock are investigated using a finite element model that engages nonlinear soil intended materials available in OpenSees. This allows incorporation of the inelastic behavior of the soil and wave propagation effects in the soil-wall system seismic response. The nonlinear response of the soil was validated using the well-stablished, frequency-domain, linear-equivalent approach. An incremental dynamic analysis was implemented to comprehensively examine the effect of soil nonlinearity and input motion on the induced seismic pressures and to evaluate current code equations/methodologies at different levels of earthquake intensity. The results show that soil nonlinearity and seismic wave amplification may play an important role in the response of the soil-wall system. Therefore, methodologies that rely only on peak ground acceleration may introduce large bias on the estimated seismic pressures in scenarios where high nonlinearity and site amplification are expected.     

Lead concentrations in bones and soil

Lead concentrations were measured in 41 rib samples from the Romano-British cemetery at Poundbury and in two sets of soil samples from the graves yielding up the bones. The levels of lead in the soil were in no way exceptional (means for the two series were 16·7 and 24·8 μg/g dry weight) and there was no correlation between soil lead and bone lead concentrations.Since bone lead concentrations vary independently of those in the soil, it is unlikely that the bones are contaminated by lead in the soil, at least under the conditions prevailing at this particular site.     

ATTENUATION RELATION FOR WEST EURASIA DETERMINED WITH RECENT NEAR-FAULT RECORDS FROM CALIFORNIA,JAPAN AND TURKEY

Strong ground motion close to a fault can be expected to be very large, so its estimation is essential for human safetv. Although a few strong-motion data exist for the west Eurasian region, we proposed in a previous work [Berge-Thierry et al., 2003] an attenuation relation for spectral acceleration using strong-motion data recorded in west Eurasia (mainly in Europe) and some in the western United States: this relationship was derived for the French Safety Rule, which is applied for seismic hazard assessment at nuclear power plants. In this study, we propose a constraining of the amplitude saturation term related to the proximity of the fault, and an adding of an amplitude saturation term in the regression model. We add, to the data-set previously used to derive the west Eurasian attenuation relationship strong-motions recorded during recent large earth-quakes: the 1995 Hyogo-ken Nanbu (Kobe) event in Japan and the 1999 Kocaeli (Izmit) event in Turkey. The regression analysis, adopted from Fukushima and Tanaka [1990], is non-linear, so an iterative procedure is applied. The determined regression coefficients lead to a prediction of a peak ground acceleration of about 0.7 g for soil site conditions at a fault distance of 0.5 km. The Q coefficient deduced, from the distance coefficient is in agreement with scattering Q models. The introduction of the saturation term leads to significantly lower predictions of average spectral accelerations at short distances as compared with using the Berge-Thierry et al. [2003] empirical model.