Analysis of emplaced waste data and implications of non-random emplacement for performance assessment for the WIPP

The WIPP Land Withdrawal Act recognized that after the initial certification of the WIPP and start of disposal operations, operating experience and ongoing research would result in new technical and scientific information. The Environmental Evaluation Group (EEG) has previously reported on issues that it considers important as the Department of Energy (DOE) works towards the first recertification. One of these issues involves the assumption of random emplacement of waste used in the performance assessment calculations in support of the initial certification application. As actual waste emplacement data are now available from four years of disposal, the EEG performed an analysis to evaluate the validity of that initial assumption and determine implications for performance assessment. Panel 1 was closed in March 2003. The degree of deviation between actual emplaced waste in Panel 1 and an assumption of random emplacement is apparent with concentrations of 239Pu being 3.20 times, 240Pu being 2.67 times, and 241Am being 4.13 times the projected repository average for the space occupied by the waste. A spatial statistical analysis was performed using available Panel 1 data retrieved from the WWIS and assigned room coordinates by Sandia National Laboratories. A comparison was made between the waste as emplaced andmore » a randomization of the same waste. Conversely, the distribution of waste as emplaced is similar to the distribution of waste in the individual containers and can be characterized as bi-modal and skewed with a long high-concentration tail. The distribution of randomized waste is fairly symmetrical, as would be expected from classical statistical theory. In the event of a future drilling intrusion, comparison of these two distributions shows a higher probability of intersecting a high-concentration stack of the actual emplaced waste, over that of the same waste emplaced in a randomized manner as was assumed in the certified performance assessment calculations. This suggests that the methodology used during the certification performance assessment calculations underestimated potential releases by cuttings and cavings. That methodology sampled each layer in a stack separately and used the mean concentration for each waste stream.« less