Some Issues On Seismology Based Ground Motion Attenuation Relationship

Seismic observation is to observe and record ground motion with seismic instruments.According to different requirements,different seismic networks are established,including large-scale seismological networks recording long distance earthquakes,regional seismological networks recording small and near earthquakes,and strong ground motion networks recording strong and near-field ground motions.Traditionally,the purpose,physical quantity and instrument system of seismological networks and strong networks are different,and the two kinds of data are applied in seismology and earthquake engineering,respectively.With the establishment and operation of broadband digital seismological observation networks,the difference between the two kinds of data is getting smaller.Once the obstacles are overcome,the data source will be greatly expanded,and the bottleneck of the lack of data from separated studies will be overcome as well.For the regions lack of strong motion records,our research group developed on seismology based ground motion attenuation relationship.In this framework,we mainly discuss three issues,including four parts.?1?The northeastern area of Japan?36°N-40°N,138°E-143°E?is selected as the target area,the difference between strong and weak ground motion was analyzed by collecting the records from F-net?Broadband Seismograph Network of Japan?stations and K-NET?Strong-motion Seismograph Networks?stations on bedrock site?M_W?3.5?.There is an inconsistency between the peak ground accelerations?PGAs?of the two datasets in M_W[3.5,4.5].The attenuation characteristics of the two databases of PGAs in other magnitude ranges,peak ground velocities?PGVs?and peak ground displacements?PGDs?are more consistent.The downhole data with M_W[3.5,4.5]from KiK-net stations or K-NET stations filtered by the fourth-order butterworth bandpass filtering[0.1Hz,10Hz)are supplemented,the attenuation characteristics of PGAs from F-net stations and K-NET stations in M_W[3.5,4.5]are improved to some extent.But the latter did not improve significantly.The K-NET has little records of PGA less than 1gal.The difference may be due to those conditions.?2?The effect of extended database on inversion results and attenuation is studied.In view of the fact that there is not enough strong motion data to establish the empirical attenuation relationship in some areas and there are a little strong motion data,now try to retrieve the regional parameters and establish the attenuation relation based on the strong motion record supplemented?M_W?4.5?.And results compared with the the previous attenuation relationship,it shows that the attenuation relationship is similar.?3?A commonly logical problem and a scheme for further coordination on seismology based ground motion attenuation relationship is discussed.In engineering seismology,ground motion is estimated with a given magnitude and distance,the distance of the attenuation of the ground motion is usually expressed by the geometrical spreading and inelastic attenuation;in seismology,the magnitude is calculated by the given ground motion and distance,and the diatance is usually expressed by the calibration function.This means that the distance attenuation terms in magnitude determination and ground motion prediction are quite different in forms with different parameters.There is a logical shortcoming.The possibility of resolving this logical problem is explained by the substitution of distance items based on K-NET data.Firstly,the magnitude difference between the magnitude is calculated by replacing the calibration function with the geometrical spreading or with the geometric spreading and the inelastic attenuation,and the magnitude differences are mainly concentrated in the-0.5-0.5.Then,the calibration function replaces the geometric spreading?the geometric spreading and the inelastic attenuation?to estimate the ground motion,and compared with the actual K-NET data,it is found that the error mean of the replacement geometric spreading in the PGAs is greater than that of the replacement the geometric spreading and the inelastic attenuation,but the error mean of PGV and PGD is exactly the opposite.There may be biased between the actual seismic record and the attention result,only the effect of before and after the replacement of the distance term on the establishment of the attenuation relation is discussed,and it is found that the replacement of the geometric spreading and the inelastic attenuation is more consistent.The result of the geometric spreading replaced by the calibration function is smaller beyond 100km.And both replacements are larger within 30km.?4?The feasibility of increasing the near-field saturation of the calibration function is discussed to the phenomenon that the near-field is on the high side due to the geometrical spreading and inelastic attenuation replaced by the calibration function.Through the analysis of near-field difference of amplitude before and after substitution,the sensitivity of f and R₀,the initial range of R₀ is determined,and the R₀ is determined by least squares.The improved calibration function is used to calculate magnitude,the magnitude differences are mainly concentrated in the-0.5-0.5,which has no obvious influence on the calculation results of magnitude.And it replaces the geometrical spreading and inelastic attenuation lead to an improvement in the PGA,PGV and PGD attenuation of the near-field.