Epistemic uncertainty in PSHA and seismic hazard characterization using the logic tree approach: part I, developing the framework

Abstract

Epistemic uncertainty offers alternatives on decision making and various possibilities of computing the hazard integral. Generally, the logic tree approach is used while treating the epistemic uncertainty. Logic tree weight calculation is a subjective decision based on the degree of belief of the analyst on the possible contributors to the epistemic uncertainty and often leads to a different set of values by different researchers. This paper aims to develop a framework of accounting for the epistemic uncertainty in probabilistic seismic hazard analysis (PSHA) by minimizing the subjectivity involved in weight calculation. Guidelines/rules are developed for the weight calculation at each node of the logic tree. Recurrence parameters, magnitude and distance probability distributions, maximum magnitude, and selection of ground motion predictive equations (GMPEs) are considered the possible sources of epistemic uncertainty. A GMPE rule is proposed to be used with the PSHA framework to account for the propagation of epistemic uncertainty. The north-east region of India is chosen for the purpose of illustration. The study region is divided into seven seismic source zones (five in the active crustal region and two in the subduction zone). Seismic hazard is characterized in terms of the weighted mean and fractile representation of hazards using the logic tree approach. Only one sample illustration of the results are reported in terms of weighted mean and fractile representation of hazard curves and uniform hazard spectra (UHS). Further illustration of the PSHA results with possible implications from the epistemic uncertainty is reported in the companion paper.