Gas transport across the low-permeability containment zone of an underground nuclear explosion.

Understanding the nature of gas transport from an underground nuclear explosion (UNE) is required for evaluating the ability to detect and interpret either on-site or atmospheric signatures of noble gas radionuclides resulting from the event. We performed a pressure and chemical tracer monitoring experiment at the site of an underground nuclear test that occurred in a tunnel in Nevada to evaluate the possible modes of gas transport to the surface. The site represents a very well-contained, low gas-permeability end member for past UNEs at the Nevada National Security Site. However, there is very strong evidence that gases detected at the surface during a period of low atmospheric pressure resulted from fractures of extremely small aperture that are essentially invisible. Our analyses also suggest that gases would have easily migrated to the top of the high-permeability collapse zone following the detonation minimizing the final distance required for migration along these narrow fractures to the surface. This indicates that on-site detection of gases emanating from such low-permeability sites is feasible while standoff detection of atmospheric plumes may also be possible at local distances for sufficiently high fracture densities. Finally, our results show that gas leakage into the atmosphere also occurred directly from the tunnel portal and should be monitored in future tunnel gas sampling experiments for the purpose of better understanding relative contributions to detection of radioxenon releases via both fracture network and tunnel transport.

The characteristic release of noble gases from an underground nuclear explosion.

Prompt release of gases at the ground surface resulting from explosively propagated vents or large operational releases has typically been considered to be the only mode of transport for detonation gases from an underground nuclear explosion (UNE) giving rise to detectable levels of radioxenon gases in downwind atmospheric samples captured at distances exceeding 100 km. Using a model for thermally and barometrically driven post-detonation transport across the broad surface of a simulated UNE site, we show in conjunction with the results of an atmospheric tracer-release experiment that even deep, well-contained UNEs, without prompt vents or leaks, are potentially detectable tens of kilometers downwind with current technology; distances that are significant for localizing the source of detected atmospheric signals during on-site monitoring or inspection. For a given yield, the bulk permeability of the UNE site and to a lesser extent the depth of detonation appear to be the primary source-term parameters controlling the distance of detection from the detonation point. We find for test-site bulk permeabilities exceeding 1 darcy (10-12 m2) that broad-area surface fluxes of radioxenon gas exhibit exponential dependence on permeability resulting in order-of-magnitude enhancements of surface flux for changes in permeability of only a darcy. Simulations of subsurface transport assuming a canonical detonation-depth-versus-nuclear-yield relationship generally resulted in larger atmospheric signals for shallower, lower-yield explosions allowing downwind detection at distances greater than 1000 km. Additionally, atmospheric simulations suggest that the lowest atmospheric boundary layer heights, such as occur at night, produced concentrations above minimum detectable levels at the greatest distances downwind.

International challenge to model the long-range transport of radioxenon released from medical isotope production to six Comprehensive Nuclear-Test-Ban Treaty monitoring stations.

After performing a first multi-model exercise in 2015 a comprehensive and technically more demanding atmospheric transport modelling challenge was organized in 2016. Release data were provided by the Australian Nuclear Science and Technology Organization radiopharmaceutical facility in Sydney (Australia) for a one month period. Measured samples for the same time frame were gathered from six International Monitoring System stations in the Southern Hemisphere with distances to the source ranging between 680 (Melbourne) and about 17,000 km (Tristan da Cunha). Participants were prompted to work with unit emissions in pre-defined emission intervals (daily, half-daily, 3-hourly and hourly emission segment lengths) and in order to perform a blind test actual emission values were not provided to them. Despite the quite different settings of the two atmospheric transport modelling challenges there is common evidence that for long-range atmospheric transport using temporally highly resolved emissions and highly space-resolved meteorological input fields has no significant advantage compared to using lower resolved ones. As well an uncertainty of up to 20% in the daily stack emission data turns out to be acceptable for the purpose of a study like this. Model performance at individual stations is quite diverse depending largely on successfully capturing boundary layer processes. No single model-meteorology combination performs best for all stations. Moreover, the stations statistics do not depend on the distance between the source and the individual stations. Finally, it became more evident how future exercises need to be designed. Set-up parameters like the meteorological driver or the output grid resolution should be pre-scribed in order to enhance diversity as well as comparability among model runs.