PROBLEMS WITH LOGIC TREES IN EARTHQUAKE HAZARD EVALUATION

Logic trees make sense when used in earthquake risk analysis where cos ts are compared for specific outcomes. However logic trees fail in ear thquake hazard analysis when they are used to develop earthquake groun d motions for applications in engineering. The failure stems from a mi sguided attempt to assign numbers for degrees of belief which are pers onal and indefinable, almost like love or taste, and for which there a re neither tests nor measurements. The result is a complicated jumble of egocentric impressions. In contrast, the need in engineering is to have values that are based as much as possible on evidence.