Ground motion variability and its effect on the probabilistic seismic hazard analysis

Abstract

The majority of injuries and casualties during earthquakes occur as a result of partial or complete collapse of buildings. The assessment of possible seismic ground motions for the purposes of earthquake-resistant design can be performed by following the deterministic or probabilistic methodology. Chapter 1 presents an overview of the current practice in seismic hazard analysis with emphasis on PSHA. At present, the Cornell-McGuire method prevails in PSHA studies. Despite significant development and modifications, this method has several controversial aspects. Absence of an upper bound of the seismic hazard curve is one of the most disputable aspects of the method, as it leads to unrealistic ground motion estimates for very low probabilities of exceedance. This problem stems from using the unbounded log-normal distribution in the modelling of the ground motion variability. The main objective of the study was to investigate this variability and suggest a more realistic probability distribution which would allow accounting for the finiteness of the ground motion induced by earthquake. Chapter 2 introduces the procedure that is suitable for studying the ground motion variability. Given the data sample, this procedure allows selecting the most plausible probability distribution from a set of candidate models. Chapter 3 demonstrates the application of this procedure to PGA data recorded in Japan. This analysis demonstrated the superiority of the GEVD in the vast majority of considered examples. Estimates of the shape parameter of the GEVD were negative in every considered example, indicating the presence of a finite upper bound of PGA. Therefore, the GEVD provides a model that is more realistic for the scatter of the logarithm of PGA, and the application of this model leads to a bounded seismic hazard curve. In connection with a revival of interest in seismic intensity as an analogue for physical ground motion parameters, the problem of accounting for anisotropy in the attenuation of MMI is considered in Chapter 4. A set of four equations that could account for this anisotropy was proposed and the applicability of these equations was demonstrated by modelling the isoseismal maps of two well-recorded seismic events that have occurred in South Africa. The results demonstrated that, in general, the new equations were superior to the isotropic attenuation equation, especially as regards to the pronounced anisotropy. As several different PSHA methods exist, it is important to know how the results of application of these methods corresponded to each other. Chapter 5 presents the comparative study of three major PSHA methods, namely, the Cornell-McGuire method, the Parametric-Historic method, and the method based on Monte Carlo simulations. Two regions in Russia were selected for comparison, and the PGA estimates were compared for return periods of 475 and 2475 years. The results indicated that the choice of a particular method for conducting PSHA has relatively little effect on the hazard estimates when the same seismic source model was used in the calculations. The considered PSHA methods would provide closely related results for areas of moderate seismic activity; however, the difference among the results would apparently increase with increasing seismic activity.