Experimental Research On Ground Motion Attenuation With Seismograms From Seimic Monitoring Network

The ground motion attenuation relationship plays an indispensable role in seismic zonation and site seismic hazard assessment, which has been an important subject among engineering seismology. There are uncoordinated problems existing in the two processes of magnitude determined and seismic amplitude estimated, which is found in the research of using the regional parameters inversed with seismograms from seismic monitoring network to establish attenuation relationship. For this problem, this dissertation does some preliminary research to discuss the possibility of estimating attenuation relationship of strong ground motion directly based on seismograms from seismic monitoring network, and then does some experimental studies for making better use of broadband records from seismic monitoring network to build ground motion attenuation relationshi p, shown as follows.Northeast area of Japan( 36-40 N, 138-143 E) is taken as the target area because of the richness of earthquake records. As the data base of this dissertation, 1382 records from 106 small earthquake s and 540 records from 56 moderate earthquakes are selected from F-net database, and 706 records from 72 small earthquakes and 1496 records from 86 moderate earthquakes are selected from K-NET database. Then, these data are divided into four magnitude grou ps, including 3.5≤Mw<4.5(the moment magnitude of 4), 4.5≤Mw<5.5(the moment magnitude of 5), 5.5≤Mw<6.5(the moment magnitude of 6) and 6.5≤Mw<7.5(the moment magnitude of 7), and reveals that there is consistency between the two types of data by overall comparison of the attenuation characteristics for different observed physical parameters, different distance ranges, different amplitudes and different frequency bands. The consistency is more obvious with larger magnitude or lower frequency. Then, the attenuation relationship of ground motion is fitted, and the result shows that the attenuation coefficient of the two kinds of data is not very different, however, the coefficient of the magnitude has a significant difference. 15 earthquakes are further selected, each of which has 10 or more records from both F-net and K-NET, to compare their respective attenuation characteristics with the overall attenuation characteristics for the corresponding magnitude. It is pointed that their difference is not very obvious, and the difference is smaller with lower frequenc y.This paper analyzes the difference between the JMA’s magnitude and magnitude calculated by two types of observed data, according to the relevant provisions of JMA, points that the difference of magni tude is from-0.3(0.4) to 0.3. Then, this dissertation compares the compliance of PGD attenuation characteristics of observed data and the attenuation relationship calculated with calibration functions in the magnitude formula, points that they are consistent except magnitude of 7. It indicates the feasibility of estimating ground m otion with calibration function used in the determination of magnitude.Five regional parameters are inversed by Micro-Genetic algorithm, including stress drop(35)s, quality factor parameters 0Q and η, geometric attenuation parameters R1 and R2 with the expanding range of data, such as the F-net’s observed data of magnitude from 3.5 to 7.5, or the F-net and K-NET’s observed data of magnitude from 3.5 to 4.5. The PGA attenuation relationship is calculated with the region parameters obtained by inversion, and compared with the existing attenuation relationship and the observed data of K-NET. The result shows that the attenuation relationship established on larger range of magnitude is enlarged in the near field, the attenuation is accelerated, and it is more consistent with the observed data. The attenuation relationship established on larger range of data does not change obviously, and it is more consistent with observed data from far filed. Above all, it is helpful to improve the estimation of ground motion attenuation on expanding the range of the inversion data. Compared with the empirical attenuation relation ship published by the two Japanese scholars, the mean value of the error in this paper is the middle of them, and it is smaller for the magnitude of 6 or the larger. In general, the standard deviation was between two empirical attenuation relationships, and it is less than two empirical attenuation relationships for the small earthquake. Compared with the attenuation of calibration function, the difference is not obvious in the medium distance, and it is higher in the near site than this paper, which the rate of attenuation is larger. It is worth studying the method of estimating ground motion attenuation with calibration function directly.