**ADVANCES IN GROUND
MOTION STUDIES IN CHINA**

*XU Lisheng,,
YU Yuanxian *and*
CHEN Yuntai*

Institute of Geophysics,
China Seismological Bureau, Beijing 100081, China

All of the seismic source, transmitting media and site condition are the important factors that affect strong ground motion. Engineering seismologists pay more concerns to the effects of site, and seismologists take more care about the effects of transmitting path and seismic source. Engineering seismologists and seismologist have gradually realized the necessities of cooperative research. If the cooperation of this kind is conducted, a lot of problems will hopefully be solved better and sooner, such as parameterization of the strong ground motion, techniques for estimating the parameters of strong ground motion, establishment of the dataset of strong earthquakes, characteristics of the long-period ground motion, characteristics of the great velocity pulse near the fault, parameterization of the ground motion field, attenuation law of the region with shortage of acceleration recordings, characteristics of the source time function, assessment of earthquake risk, uncertainty in the assessment of seismic activity, improvement in the methods of probability and certainty, structure codes considering the duration of ground motion, topograghical condition and designing standard, and others (Hu, 2001). In this paper, we briefly summarize the progresses made by engineering seismologists and seismologists of China in recent years.

I. PROCESSING OF THE STRONG GROUND MOTING DATAIt is very important to scientifically use observation data in scientific research. Improper use of data will result in mistakes in conclusion. Thus, it is highly necessary and very important to process the data for keeping the true signal and throwing the fault signal before it is put in use. The problem that is often present in the data of strong ground motion is the constant shift from the base line or zero-point shift. One of the reasons is that the instrument has been tilted by the ground deformation near the fault even before the earthquake process is over. In the recordings of strong ground motion of the September 21, 1999, Ji-Ji, Taiwan,

M_{S}7.6 earthquake popularly existed the problem of the zero-point shift. In order to make the data reliable, Wang (2001) proposed a technique for correcting the shifted base-lines, and applied it to 210 recordings of 70 stations of the Ji-Ji TaiwanM_{S}7.6 earthquake. He used this technique to correct the acceleration recordings, integrated them to obtain velocity recordings and displacement recordings, and calculated the response spectra. It was noticed that the response spectra in the periods less than 10s were hardly affected by the base-line correction, while the response spectra at the low frequencies were highly affected. Thus, one must pay a special attention to the effect caused probably by the base-line shift as calculating the peak velocity, peak displacement and permanent displacement.

II. FACTORS AFFECTING THE NEAR-SOURCE STRONG GROUND MOTIONBoth of the distance to the seismogenic fault and the rupture propagation direction are important factors that affect the strong ground motion near the fault. In studying the September 21, 1999, Jiji (Chi-Chi), Taiwan,

M_{S}7.6 earthquake, Wang (2001) found that the strong ground motions recorded at the sites less than 3km from the fault were much larger than those recorded in other sites, and the components of ground motion parallel to the slipping direction were much larger than that perpendicular to the slipping direction. However, the analysis of the response spectra indicated that the rupture propagation direction affected only the strong ground motion of short period (<0.6 s). Also, the geometrical and kinematical characteristics of the fault are the important factors that affect the strong ground motion. The September 21, 1999, Jiji (Chi-Chi), Taiwan,M_{S}7.6 earthquake was an event of mainly thrust faulting with a shallow dip angle. For this event, Yu and Gao (2001) compared the empirical attenuation relations of the horizontal and vertical peak ground accelerations (PGA) set up by regression method and that set up with the real recordings. The comparison showed that a systematic difference existed between PGA on the hanging-wall and that on the footwall. The peak accelerations recorded on the hanging-wall were evidently higher than those recorded on the footwall. The asymmetry of the PGA distribution to the fault trace could be clearly seen on the observed PGA contour map, with the PGA attenuating slower on the hanging-wall than on the footwall.The strong ground motion generated by the real earthquakes is complicated generally. The ground motion of the September 21, 1999, Ji-Ji, Taiwan, earthquake had some robust and special features. Firstly, the velocity pulses were quite large. The maximum PGVs of the vertical, NS and EW components generated by this earthquake were 228.6, 292.2 and 280.5 cm/s, respectively. The largest velocity pulse was 177cm/s before, which was generated by the 1994 Northridge earthquake. Secondly, the permanent displacements were large. From the recordings of the strong-earthquake instrument, the maximum vertical displacement was close to 10m, and the maximum horizontal displacement was 9m. Thirdly, It had a plenty of long-period components. The response spectra of the horizontal components from a large number of recordings of the broad band digital instruments (8 sets of recordings obtained from the observation points near the fault) had the average value larger than 0.2 g at the 6 s, and the average value plus the squared residual was still larger than 0.2 g even at the 8 s (Wang, 2001).

There are many factors affecting the strong ground motion. Seismologists and engineering seismologists have been making varieties of efforts in understanding and explaining the complicated ground motion. Jin et al. (2001) quantitatively analyzed the effect of seismic source on spatial correlation of strong ground motion by means of seismic source theory and expansion of the Fourier spectrum of ground motion with the space variables, and obtained a relation of the spectra of ground motion between two neighboring positions in a relatively simplified case that the seismic source is a linear fault with finite length and the rupture is specified a unilateral rupture of constant velocity.

Besides the stable factors affecting the strong ground motion, there are random factors. The effects of random factors can only be treated with the random analysis. Wang (2001) analyzed the strong ground motion data of the September 21, 1999, Ji-Ji, Taiwan,

M_{S}7.6 earthquake using the concept of the sample assemble. The analysis indicated that no essential difference existed in the statistical feature of amplitude and period of the horizontal ground motions while the relatively larger difference existed in the feature of correlation, and that the evident difference existed in both of statistical feature and correlation feature of the vertical and horizontal ground motions. The correlation study of both sample function and sample assemble indicated that the correlation function of acceleration time series contained a plenty of random components. According to the characteristic of the correlation matrix of the acceleration time series, the ground motion was proposed to be a super random procedure. The preliminary study of the velocity and displacement time series for randomness characteristic indicated that the randomness of the velocity time series was rather weaker than that of the acceleration time series, and that the displacement time series was nearly certain (Wang, 2001).

III. MODELING AND PREDICTION OF THE STRONG GROUND MOTIONThe strong ground motion depends on the seismic source of the earthquake generating the seismic wave, and the transmitting media as well as the site condition. Some researchers focus on the study of the relation between the ground motion and the characteristic of the seismic source, the other researchers focus on the study of the relation between the ground motion and the transmitting media. In recent years, a great progress has been made in these two directions. Luo (2000) considered a scenario earthquake of M6.5 which occurs 100km away from Shanghai down-town, and predicted the possible ground motion of the Shanghai area using both empirical Green's function method and theoretical Green's function method. The average peak accelerations predicted by the two methods are 25 Gal and 33 Gal, respectively, which are in good agreement with the value 24 Gal estimated using the method of hazard analysis in doing which an exceeding probability of 63.2% within 50 years were used. Ding et al. (2003a,b) simulated the ground motion in Beijing City by an algorithm for the calculation of synthetic seismograms in laterally heterogeneous anelastic medial. The synthetic signals were compared with the few available seismic recordings of the 1998 Zhangbei earthquake and with the distribution of observed macroseismic intensity of the 1976 Tangshan earthquake. The synthetic 3-component seismograms have been computed for the Xiji area and Beijing town. The modeling of the seismic ground motion for both the Tangshan and the Zhangbei earthquakes showed that the thick Quaternary sedimentary cover amplified the peak values and increased the duration of the seismic ground motion in the northwest part of the City. Such a result was well correlated with the abnormally high macro-seismic intensity zone in the Xiji area associated with the 1976 Tangshan earthquake as well as with the ground motion recorded in Beijing town as the 1998 Zhangbei earthquake occurred. Therefore the thickness of the Quaternary sediments in Beijing City was considered as the key factor controlling the local ground motion. The four zones were defined on the base of the different thickness of the Quaternary sediments. The response spectra for each zone were computed. The peak spectral value as high as 0.1g was considered to be compatible with past seismicity and be exceeded if an earthquake similar to the 1697 Sanhe-Pinggu occurs.

IV. BROAD-BAND RESPONSE SPECTRUMMore and more earthquake examples have shown that some large structures even far away from earthquake epicenters were damaged. Therefore, people have to take into account the effects of the distant earthquakes and the long-period ground motion as they design modern structures like very tall building and/or long bridges whose natural periods are larger than 5 or 10 s. The currently available ground motion data from analog instruments have shortcomings in the period range because of the limitation of the observation technique and the response spectrum based on the dataset of this kind are lack of the reliable information at the low frequencies. In the recent years, the improvement of observation technique has enabled us to use broad-band seismic data and the techniques of modern seismology in calculating response spectrum. Hu and Yu (2000) proposed a seismological-engineering approach for estimating the broad spectrum for distant earthquakes. They used the broad-band digital seismic recording and took into account the relations between magnitude and displacement in estimating long period (2-20s) part of the spectrum, while used the currently available strong ground motion data from the analog accelerograph in estimating the short period (0.1-3s) part. By using this technique, the response spectrum up to the period of 10 to 20 s can be obtained, and the precision and reliability have proved to be high enough (Yu and Hu, 2001).

V. INTERNATIONAL COOPERATION AND FUTURE DEVELOPMENTIn the era when the information is globalized, the international cooperation in scientific research is becoming more and more important. In the last years, the Chinese experts have been involved in the international cooperation and made a great progress. In the period from August of 1998 to March of 1999, the Chinese experts took part in the cooperative research on strong ground motion simulation of the San Francisco Bay. In this cooperation, a new method was proposed for calculating ground motion, and applied this method to three scenario earthquakes for verification. The main effectiveness of the cooperation lay in that the existing database and various ground motion models were applied and checked (Qi, 2001). Our seismologists and engineering seismologists agreed that it will be very necessary to push forward the international cooperation in the future, and suggested the future cooperation will be focused on the establishment of seismic array consisting of the strong-earthquake acceleration instruments, observation technology, data processing and application of observation data. In particular, the following work should be taken into account, (1) the long-period strong ground motion; (2) the ground motion near the fault; (3) multi-component ground motion and correlation; (4) the attenuation of the strong ground motion; (5) the relation between the surface rupture and seismic source, transmitting media; (6) the technique for modeling the rupture; (7) stochastic procedure for strong ground motion simulation; (8) broad-band Green function for strong ground motion simulation; (9) earthquake response analysis of large sediment basin; and (10) GIS system for active fault prospecting and evaluating(Qi,2001).

REFERENCESDing Z., Chen Y., Panza G. F., 2003a, Estimation of site effects in Beijing City, PAGEOPH, (in press).Ding Z., Romanelli F., Chen Y. T., Panza G. F., 2003b, Realistic modeling of seismic wave ground motion in Beijing City, PAGEOPH, (in press).Hu Y. X., Yu Y. X., 2000. A seismological–Engineering approach to the estimation of Wide response spectrum for distant earthquake.Advances in Structural Dynamics, Vol. I, J. M. Ko and Y. L. Xu (Eds), 193-200.Hu Y. X., 2001. Strengthen the link between engineering and seismology,Earthquake Engineering Frontiers in the New Millennium, B. F. Spencer Jr and Y. X. Hu (Eds), 103-108Jin X., Chen C., Ding H. P., Zhang M. Y., 2001. Research on effects of seismic source on space correlation.Earthquake Engineering frontiers in the New Millennium, B. F. Spencer, Jr and Y. X. Hu (Eds), 153-158.Luo Q. F.,2000. Estimation of Ground motions affecting Shanghai by long distance earthquake.Advances in Structural Dynamics,Vol. I, J. M. Ko and Y. L. Xu,(Eds), 225-232.Luo Q. F., 2000. Probability-consistent scenario earthquake and its application in estimation of ground motions,Proceedings of 12, Auckland, New Zealand, Paper ID0010.^{th}WCEEQi X. Z., 2001. An overview of US-PRC cooperation on earthquake engineering at IEM and future prospects of the program.Earthquake Engineering Frontiers in the New Millennium, B. F. Spencer Jr and Y. X. Hu (Eds), 127-131.