Application of generalized Pareto distribution for modeling aleatory variability of ground motion

Abstract The lognormal distribution is commonly used to characterize the aleatory variability of ground-motion prediction equations (GMPEs) in probabilistic seismic hazard analysis (PSHA). However, this approach often leads results without actual physical meaning at low exceedance probabilities. In paper, we discuss how calculate PSHA with a probability. Peak ground acceleration records from NGA-West2 database and 15,493 residuals calculated by Campbell-Bozorgnia using GMPE were applied analyze tail shape residuals. showed that generalized Pareto (GPD) captured characteristics better than distribution. Further study shapes tails distributions different magnitudes varied significantly due heteroscedasticity magnitude; larger had smaller upper limit on right side. Moreover, three magnitude ranges given more consistent GPD parameters fitted all residuals, leading bounded curve. Therefore, representative up determined threshold, characterizes for calculation.