Modeling density‐driven flow and radionuclide transport at an underground nuclear test: Uncertainty analysis and effect of parameter correlation

Abstract

The effects of uncertainty associated with input parameters and conceptual groundwater flow and transport models are studied and evaluated for radionuclide transport modeling at the Milrow underground nuclear test site on Amchitka Island, Alaska. Two‐dimensional numerical simulations in conjunction with a multiparameter uncertainty analysis are adapted and used to evaluate the effects of the uncertainties associated with the definition of the modeled processes and the values of the parameters governing these processes. Density‐driven flow caused by the mixing of fresh and salt water in a coastal aquifer and radionuclide mass transport resulting from the nuclear test are the two major processes analyzed. The parametric uncertainty analysis is performed in two stages. In the first modeling stage, one parameter is varied at a time, and the effects of this variability on transport results are evaluated. The analysis indicates that the solution to the flow problem, and thus the transport results, are very sensitive to the values of recharge and hydraulic conductivity. However, fracture porosity does not affect density‐driven flow but dramatically influences the transport results in terms of arrival times of radionuclides from the test cavity to the seafloor. Matrix diffusion is also crucial in affecting arrival times and mass flux values. The second modeling stage utilizes multiple realizations of the flow field generated by considering random combinations of recharge, conductivity, macrodispersivity, and fracture porosity drawn at random from their respective distributions. For each one of these realizations, transport parameters are also drawn at random from their assigned distributions. Different correlation models are considered and compared with scenarios with no parameter correlation. The results indicate that incorporation of correlation between hydraulic conductivity and recharge significantly reduces the uncertainty in both travel time and transverse plume location. This results from the impact of these parameters on the transition zone, which dominates flow velocity variability in this coastal aquifer setting. Correlation between fracture porosity and matrix diffusion has the largest impact on the transport results, with a positive correlation providing the critical scenario of most rapid transport (thus less radioactive decay) and a negative correlation leading to longer travel times with significant decay.