Trimming the UCERF2 Hazard Logic Tree

Abstract

Listen

Translate article

### Assessing Which Sources of Hazard Uncertainty Matter The Uniform California Earthquake Rupture Forecast 2 (UCERF2) is a fully time‐dependent earthquake rupture forecast developed with sponsorship of the California Earthquake Authority (Working Group on California Earthquake Probabilities [WGCEP], 2007; Field et al. , 2009). UCERF2 contains 480 logic‐tree branches reflecting choices among nine modeling uncertainties in the earthquake rate model shown in Figure 1. For seismic hazard analysis, it is also necessary to choose a ground‐motion‐prediction equation (GMPE) and set its parameters. Choosing among four next‐generation attenuation (NGA) relationships results in a total of 1920 hazard calculations per site. The present work is motivated by a desire to reduce the computational effort involved in a hazard analysis without understating uncertainty. We set out to assess which branching points of the UCERF2 logic tree contribute most to overall uncertainty, and which might be safely ignored (set to only one branch) without significantly biasing results or affecting some useful measure of uncertainty. The trimmed logic tree will have all of the original choices from the branching points that contribute significantly to uncertainty, but only one arbitrarily selected choice from the branching points that do not. Figure 1. Branches of the UCERF2 logic tree (after WGCEP, 2007). Dotted lines (…) indicate that the tree continues parallel to the branches that are shown, but that these branches are omitted from the figure for clarity. Black numbers below branches are the branch weights. Risk analyses that use the trimmed tree should produce approximately the same results as those that use the full tree. Risk, as used here, refers to the relationship between some undesirable outcome and its likelihood of occurrence, typically from some particular decision‐maker’s perspective. Risk can be measured in terms of an uncertain quantity of loss, such as the relationship between future earthquake‐related repair costs to a particular asset or group …