Long-term prediction of spent-fuel performance in a geologic repository

The US government has designated Yucca Mountain, Nevada, as the sole candidate site for permanent disposal of high-level nuclear waste (HLW). The HLW includes reprocessing wastes from defense activities and spent fuel from commercial power reactors. Most of the radioactivity associated with the HLW that may be placed in a Yucca Mountain repository is present in the spent nuclear fuel (nominally, UO[sub 2]). Spent nuclear fuel contains elements and isotopes of high toxicity. Safe, permanent disposal of such material requires a thorough knowledge of its behavior in the repository environment over time periods that exceed all historical records. The US Environmental Protection Agency regulations call for containment to meet specified cumulative release requirements over a 10,000-yr period. Reliance on laboratory data for assessment of the long-term safety of a HLW repository requires extrapolation of the experimental data to time periods many orders of magnitude beyond the measurement period. Extrapolation of short-term laboratory data to such great time periods is extremely uncertain. One alternative to such extrapolation is to study natural systems that have operated for long times (in many cases longer than the required 10000-yr period) and that are analogous in some respect to portions of an HLW repository system.more » This type of reasoning, though uncommon in traditional and current design and engineering activities, is routinely applied in the earth sciences. Geologists interpret the early history of the earth (on the order of 10[sup 9] yr) and study ongoing but extremely slow processes such as continental deformation (on the order of 10[sup 6] to 10[sup 8] yr) without the benefit of being able to conduct experiments on those time scales. Although such reasoning has uncertainties of its own, there is no alternative to the study of natural systems to obtain observational knowledge of the long-term behavior of HLW components.« less