Calibrating the backbone approach for the development of earthquake ground motion models

Douglas, J. (2018) Calibrating the backbone approach for the development of earthquake ground motion models. In: Best Practice in Physics-based Fault Rupture Models for Seismic Hazard Assessment of Nuclear Installations : Issues and Challenges Towards Full Seismic Risk Analysis, 2018-05-14 - 2018-05-16, CEA - Cadarache- Château.

Text (Douglas-PSHANI-2018-Calibrating-the-backbone-approach-for-the-development-of-earthquake-ground-motion-models)
Accepted Author Manuscript

Download (539kB)| Preview


    The backbone approach is becoming increasingly employed to develop ground-motion models for use within probabilistic seismic hazard assessments, particularly for nuclear facilities. The backbone approach has a number of attractions, including: transparency over the level of uncertainty implied by the ground-motion model, a clearer understanding of the meaning of the weights of the logic tree (because each branch is mutually exclusive and collectively exhaustive) and an ability to make the model specific for a given site. This is in contrast to the classic method of selecting (using various approaches) a suite of ground motion prediction equations from the literature, which may appear easier but suffers, for example, from the difficulty of understanding whether epistemic uncertainty in future ground motions at the site is sufficiently captured. One of the principal challenges in applying the backbone approach is its calibration so that the branches of the ground-motion logic tree capture the appropriate level of epistemic uncertainty. This is particularly difficult for regions with limited strong-motion data, which are generally areas of lower seismicity. In this article, I summarize previous uses of the backbone approach in the literature before investigating calibration using the stochastic method, which is particularly useful when there are few or no local strong-motion records. I show that the scaling factors developed from the stochastic models roughly imply the expected variations in epistemic uncertainty given the amount of data available from different tectonic regimes.