Near field ground motions of intermediate-size earthquakes: statistical analysis of response spectra Diethelm Kaiser 1 Introduction In Central Europe intermediate-size earthquakes with magnitudes in the range of approximately M L = 5 – 6 and focal depths down to 20 km at short distances up to 20 km contribute most significantly to the seismic hazard. Due to a severe lack of strong motion recordings in Central Europe we have to rely on data from other regions to predict ground motions. 2 Strong motion data For this study we selected 54 strong motion recordings according to the criteria outlined above from earthquakes with magnitudes 5.0 ≤ M L ≤ 6.0 in Friuli and Lazio-Abruzza (Italy), California, Taiwan, Georgia (Caucasus), Greece, Iran, Azores Islands, and San Salvador recorded on soil sites. Fig. 1 shows the distribution of the selected strong motion records with respect to the relevant parameters having mean values of magnitude M L = 5.4, epicentral distance ∆ = 10 km, hypocentral distance r = 13 km, focal depth h = 7 km. The three- component accelerograms were converted to the SMC format, corrected and processed using BAP (Converse 1992). 1 10 Frequency [Hz] 0.1 1 10 Spectral Acceleration [m/s 2 ] resp. spectra 84% 50% 16% 1 10 Frequency [Hz] 0.2 0.4 0.6 0.8 1 V/H Ratio of Spectral Acceleration median mean 5 5.2 5.4 5.6 5.8 6 ML 0 2 4 6 8 10 12 Number of Records Fig. 1 Distribution of strong motion records used with respect to the magnitude M L , epicentral and hypocentral distance, and focal depth of the earthquake 5,0 5,2 5,4 5,6 5,8 6,0 0 5 10 15 20 25 Epicentral Distance [km] ML 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Focal Depth [km] Epicentral Distance [km] 5,0 5,2 5,4 5,6 5,8 6,0 2 4 6 8 10 12 14 ML Focal Depth [km] 5,0 5,2 5,4 5,6 5,8 6,0 2 4 6 8 10 12 14 16 18 20 22 24 26 ML Hypocentral Distance [km] Fig. 2 Response spectra (5 % damping) of the 107 horizontal components in the magnitude range 5.0 ≤ M L ≤ 6.0, and 16th, 50th (median), and 84th percentile. Maximum horizontal spectral amplitudes of 3 m/s 2 in the median response spectrum occur at 5 – 6 Hz. The 16th and 84th percentile are a factor of 3 below resp. above the median spectrum. Fig. 3 Ratio of vertical to horizontal component response spectra. The ratio strongly depends on the frequency with average V/H ratios ranging from 0.3 at 1.3 Hz to 1 at 15 Hz. Application of a constant ratio of 0.5 or 0.7 is not justified. 3 Statistical analysis of response spectra Response spectra of horizontal and vertical components were statistically analysed. The results are shown in Fig. 2, 3, and 4; more details are given in Kaiser (1999). 1 10 Frequency [Hz] 0.1 1 Spectral Acceleration [m/s 2 ] M = 5.4 dist = 10km (soft) soil this study Ambraseys et al. 96 Boore et al. 97 1 10 Frequency [Hz] 0.1 1 Spectral Acceleration [m/s 2 ] this study, ML = 5.4 Hosser et al. 91, I = VIII Hosser et al. 91, I = VII 0 0.2 0.4 0.6 0.8 1 Cumulative Distribution 0.1 1 10 Spectral Acceleration [m/s2] 1 hz data 1 Hz fit 10 Hz data 10 Hz fit Fig. 4 Cumulative distribution of spectral acceleration amplitudes of the response spectra shown in Fig. 2 at examplary selected frequencies of 1 Hz and 10 Hz and fitted log-normal distribution functions with the mean and the standard deviation of the samples. The spectral acceleration amplitudes do not deviate significantly (at a level of 5 %) from a log-normal distribution. Fig. 5 Comparison of median horizontal response spectra for magnitude 5.4 earthquakes recorded at a distance of 10 km at (soft) soil sites. If we assume that in the magnitude range 5.0 – 6.0 both the differences in the magnitude scales M L , M S , and M W , and the difference between epicentral distance and the closest distance to the vertical projection of the rupture are small, the spectra show good agreement, except for frequencies below 1 Hz. Fig. 6 Comparison of median horizontal response spectra at soil sites for earthquakes with magnitude M L = 5.4 (range 5.0 – 6.0) at a mean distance of ∆ ∆ ∆ = 10 km (range 2 – 24 km) with response spectra for site intensities I = VII, M L = 6.0 (range 5.5 – 6.3), ∆ ∆ ∆ = 34 km (range 14 - 45 km) and I = VIII, M L = 6.3 (range 5.9 – 6.7), ∆ ∆ ∆ = 23 km (range 15 - 34 km). The spectra differ in amplitude and shape because of differences in the predominant magnitude and distance used. 4 Discussion and conclusions The median response spectra derived in this study agree well with recently published strong ground motion attenuation models at frequencies above 1 Hz (Fig. 5). However, our spectra deviate both in shape and in amplitude from response spectra used in the past for the design of nuclear power plants, dams, and buildings (Fig. 6). The spectra published by Hosser et al. (1991) were derived largely from recordings of earthquakes with magnitudes above the magnitude range of engineering importance in Central Europe. In particular, our results show that at high frequencies (f > 2 Hz) larger strong ground motion amplitudes should be expected. References Ambraseys, N. N., Simpson, K. A., Bommer, J. J., 1996. Prediction of horizontal response spectra in Europe. Earthquake Engineering and Structural Dynamics 25, 371 - 400 Boore, D. M., Joyner, W. B., Fumal, T. E., 1997. Equations for estimating horizontal response spectra and peak acceleration from western North Amercian earthquakes: a summary of recent work. Seismological Research Letters 68, No. 1, 128 – 153 Converse, A. M., 1992. BAP: Basic Strong-Motion Accelerogram Processing Software. US Geological Survey Open-File Report 92-296A Hosser, D., Keintzel, E., Schneider, G., 1991. Seismische Eingangsgrößen für die Berechnung von Bauten in deutschen Erdbebengebieten. Abschlußbericht zum Forschungsvorhaben Harmonisierung europäischer Baubestimmungen, Eurocode 8 - Erdbeben. Universität Karlsruhe. Institut für Massivbau und Baustofftechnologie Kaiser, D., 1999: Bodenbewegungen in der Nähe mittelgroßer Erdbeben. In Savidis, S. A. (Ed.): Entwicklungsstand in Forschung und Praxis auf den Gebieten des Erdbebeningenieurwesens, der Boden- und Baudynamik (D-A-CH-Tagung, Berlin 1999). Deutsche Gesellschaft für Erdbebeningenieurwesen und Baudynamik DGEB-Publ. Nr. 10 Acknowledgements Strong Motion Data was provided by the following organizations: Ente Nazional per l’Energia Elettrica, Rom; Imperial College of Science, Technologie and Medicine, London; Lamont Doherty Earth Observatory / National Center for Earthquake Engineering Research, Columbia University, Palisades; Southern California Earthquake Center, Institute for Crustal Studies, University of California, Santa Barbara; Swiss Seismological Service, ETH Zürich; United States Department of Commerce, National Geophysical Data Center, NOAA, Boulder, Colorado; U.S. Geological Survey, Dallas. - I thank Lothar Hahn, Peter Hergert, Jürgen Kopera, Patrick Smit, Christine Wassilew-Reul and Rainer Zinn for their help and numerous discussions. Friedrich-Schiller-University Jena Institute of Geosciences Burgweg 11, D-07749 Jena, Germany kaiser@geo.uni-jena.de XXVII General Assembly of the European Seismological Commission, Lisbon, Portugal, September 10-15, 2000, Session SSF-4 Strong Ground Motion Analysis and Prediction