Application of Bayes Network analysis to RWGD siting: expert estimation of geological barrier effects due to climate change

Hincks, Thea, Aspinall, William P. and Sparks, R. Stephen J. (2017) Application of Bayes Network analysis to RWGD siting: expert estimation of geological barrier effects due to climate change. In, Apted, Michael and Ahn, Joonhong (eds.) Geological Repository Systems for Safe Disposal of Spent Nuclear Fuels and Radioactive Waste. (Woodhead Publishing Series in Energy) 2nd ed. Woodhead Publishing, pp. 551-582. (doi:10.1016/B978-0-08-100642-9.00019-0).

Abstract

This chapter presents the main methods and findings of a pilot study to demonstrate potential applications for Bayes Networks (BNs) in the management of uncertainty in the long-term safety case for a radioactive waste geological repository. The study, carried out with experts from the Radioactive Waste Management Directorate of NDA, demonstrates a method for determining uncertainty distributions for a probabilistic analysis of potential effects on geological barriers due to climate-driven changes in related processes and factors. Two complementary, decision support approaches are applied: (1) the Classical Model for expert judgment elicitation (introduced in Chapter 20) provides a structured basis for eliciting and aggregating expert judgments in a formal and auditable way and (2) a graphical Bayes Network is employed to assess the evidential worth of uncertain data, including expert judgments.

The BN was constructed by performing a systematic review of the key features, events, and processes relating specifically to the single issue of the extent to which climate change could affect biosphere radiological risks due to a release from a repository; in this exploratory exercise, the approach was applied to four different climate change scenarios over the next 1 million years. Although the model developed for this pilot study was necessarily highly simplified and considers a limited number of alternative climate scenarios, it demonstrates a methodology for identifying states and scenarios not currently captured by the reference case model in the post-closure performance assessment that could lead to significant increases in risk and/or uncertainty. An important outcome is the ability to evaluate relative changes in key risk factors, compared to a defined base climate scenario with a known level of risk.

This record has no associated files available for download.

Contributors

Author: Thea Hincks

Download statistics