On the Nature of Logic Trees in Probabilistic Seismic Hazard Assessment

Abstract

An objection sometimes made against treating the weights of logic tree branches as probabilities relates to the Kolmogorov axioms, but these are only an obstacle if one believes that logic tree branches represent a seismic source model or ground motion model as being “true.” Models are never true, but some models are better than others. It is argued here that a logic tree weight represents the probability that the model in question is better than the others considered. Only one branch can be the best one, and one branch must be the best one. It is also argued that there are situations in PSHA where uncertainty exists but the analyst lacks the means to express it. Therefore it is not necessarily the case that more information increases uncertainty; it may be that more information increases the possibility of expressing uncertainty that was previously unmanageable.