Study On The Ground Motion Uncertainty Based On The Chinese Flatfile

As been well know,the different sites show the different ground motions during the same earthquake,the ground motions from different earthquakes are also different on the same site.Therefore,there is a great uncertainty in prediction the ground motion intensity of a site using the attenuation relationship.The result of ground motion uncertainty will directly affect the reliability of seismic hazard analysis.In recent years,with the rapid development of digital strong-motion instruments,strong-motion data has increased rapidly in the world,a large number of strong-motion records have been obtained,providing the basic data for studying the ground motion uncertainty.Since the operation of the Digital Strong-motion Observatory Networks in China,a large number of high-quality strong-motion records have been collected,which provides sufficient data for relevant research in earthquake engineering in China.In this paper,the strong-motion data from China as research subjects,and the first China's strong-motion data Flatfile is constructed.Based on this database,the uncertainty of ground motion intensity measures is studied.The specific contents are as follows:(1)Building the China strong-motion data Flatfile.About 10000 ground-motion records collected by National Strong-Motion Obersvation Network System from 2005 to 2015 were processed and complied,and building the China strong-motion Flatfile,including source metadata,seismic event metadata,station metadata,and propagation path metadata,ground-motion intensity measures metadata and recrodings information metadata.The empirical conversion relationships between different magnitude forms(Mw,ML,Ms,and Mb)and different distance metrics(Rrup and Rhyp,RJB,and Repi)are given;For the VS30 values of station without borehole velocity profile were derived from the the recommented values;this study also provide the the high-pass cutoff frequency in different magnitudes ranges,when the strong-motion data were filtered.(2)Analyzing the horizontal ground motion directionality.Based on the China strong-motion data Flatfile,the relationships between values for different definitions of the horizontal component of ground motion were analyzed,the RotD50 that not dependent on the in situ orientations of the sensors is used as a reference.The empirical model of Sai/SaRotD50 is given to convert between different definitions of the horizontal component of ground motion.In addition,different definitions of horizontal ground motion prediction equations have been derived by performing regression analyses,and the between-events standard deviation,within-event standard deviation,and total standard deviation.The ratio between the aleatory variabilities that result from using different different definitions of the horizontal component(?Sai/?SaRotD50,(?)Sai/(?)SaRotD50 and ?Sai/?SaRotD50)indicate that there is a difference between different definitions,and the within-event uncertainty is significantly gereater than the between-events uncertainty.The empirical relationship between RotD100 and RotD50 depending on the period is also given in this study.(3)Analyzing the site effects varibiltiy.Based on the strong-motion records in Sichuan region,China,the components of single-station standard deviation were evaluated.At present,the published ground motion prediction euqations applicable to the Sichuan region are not appreciate for the non-ergodic assumption standard deiations analysis,due to their model variables or data limitations.The ground motion prediction euqation was firstly established,and the RotD50 is adopted as ground-motion intensity measures.The residual analyses confirm that there are not magnitude-scaling,site-scaling,and distance-scaling errors,which ensure that the prediction model is applicable to evaluate the single-station standard deviation.The results show that the non-ergodic assumption standard deviation is reduced by about 10%-20%compared with the ergodic assumption.Compared the event-corrected single-station standard deviation(?)ss in this study with other regions(Califomia,Greece,Chile,Iran,Japan,Italy,New Zealand,etc.),the results show that the phiss for Sichuan region,California and Iran were significantly larger than other regions,which could be a sort of warning against the usual assumption that(?)ss is regionally independent.(4)Building the empirical equations of the prediction of the significant duration of ground motion,and its uncertainty was analyzed.This study gives the prediction equations for horizontal and vertical ground motion significant duration(Ds-75?D5-95 and D20-80)applicabled to China.The residual analyses confirm that there are not magnitude-scaling,site-scaling,and distance-scaling errors,which ensure that the prediction model is reliable.The results of uncertainty analysis show that the total standard deviation of D5-75?D5-95 and D20-80 for both of horizontal and vertical ground motion ordered by D5-95?D5-75?D20-80,the uncertainty of significant duration of verical ground motion is less than that of horizontal ground motion.(5)Analyzing the regional characteristics of ground motion uncertainty.For the regional characteristics of ground motion directionality,the results indicate that SaENN/SaGM is weakly dependent for short period in the near field,however,there is a strong regional dependence for long period.For the Significant duration of ground motion,the results show that there is greater difference in within-event uncertainty when the Mw<5.75.This study investigated the regional characteristics of ground motion in six Mw?6.0(the Minxian,Ludian,Jinggu,Kangding,two aftershocks of the 2008 Wenchuan earthquake)earthquakes that occurred within China's North-South Seismic Zone.The average source effect on ground motion of the Mw?6.0 in this study region was weaker than the average level of global shallow crust earthquakes considered by the prediction model that were presented by Boore et al.(2014);the modified coefficient ?c3 values of inelastic attenuation indicate that the regional adjustment factor of inelastic attenuation in the model is only applicable to some local regions(ie.the Longmenshan fault zone),and is not applicable to the entire North-South Seismic Zone.