Section 6.2 above presents an analysis related to the thermally-induced geochemical alteration of the buffer as an example of a deviation in performance, the potential for which are explored by modelling. Such deviations in performance could potentially arise in certain “PA scenarios”, e.g., upper bound heat production in the HLW canisters. A wide range of PA scenarios is considered, primarily through modelling, including the expected performance of the repository, which is also categorised as a PA scenario.
The results from performance assessment (for details see NTB 24‑22 Rev. 1, Nagra 2024u) indicate that most PA scenarios do not affect the broad descriptions of the evolution of the pillars of safety in any substantive way, i.e., performance targets are met even for combinations of unfavourable parameters explored using the probabilistic workflow, generally with a large safety margin, as illustrated by the example in Fig. 6‑10. This shows that the repository system is highly robust regarding uncertainties related to the repository system performance and that the repository is likely to evolve according to its “expected performance” as described e.g., in NAB 24‑20 Rev. 1 (Nagra 2024m). PA scenarios that result in only minor deviations in performance are propagated, along with the PA scenario for expected performance, to the “reference safety scenario” and its variants, described further in Section 7.2.
There are several PA scenarios that relate to the properties of the CRZ. Compared to the engineered barriers, which are assumed to adhere to their design specifications, certain aspects of the CRZ involve some uncertainty. As noted in Section 6.2.1, PA modelling generally assumes that diffusion is the dominant transport mechanism in the CRZ, and that advection along steeply dipping faults does not represent a relevant transport mechanism. Furthermore, it is assumed that sedimentary features, such as hard beds in the upper and lower confining geological units, do not act as lateral release paths. While these assumptions are well supported, some uncertainty remains due to inevitable incompleteness in geological characterisation. Thus, there are PA scenarios that assume the unlikely presence of an undetected, water-conducting fault, or the presence of water-conducting sedimentary features within the confining geological units, which cannot be excluded a-priori, even in the repository construction phase. If they were to occur, such features could affect (at least locally) the conceptualisation of radionuclide transport in the geosphere as a diffusion-dominated process. These PA scenarios are propagated to alternative safety scenarios, as described in Section 7.3, or, if extreme and hypothetical properties for such features are assumed, to the “what-if” cases described in Section 7.4.
Finally, some PA scenarios, such as the creation of new faults, or the re-activation of existing faults, in the post-closure period by repository-induced effects, i.e., by the effects of repository heat and repository-generated gas, or by geological phenomena such as earthquakes, have been shown in the PA in NTB 24‑22 Rev. 1 (Nagra 2024u) and in the assessment of earthquake scenarios in NAB 24‑28 (Nagra 2024g) to be either implausible or exceedingly unlikely. It has also been shown that the excavation of the repository by erosion, within a million-year time frame, by erosive processes is hypothetical. These effects are, nonetheless, considered in the “what-if?” cases in safety scenario development, as described in Section 7.4.
Future human actions (FHA) and their potential radiological consequences do not fall within the scope of PA. Rather, these are considered separately, as reported in NAB 24-09 (Nagra 2024r). FHA safety scenarios are described in Section 7.5.