REMEDIATION OF THE FAULTLESS UNDERGROUND NUCLEAR TEST: MOVING FORWARD IN THE FACE OF MODEL UNCERTAINTY

摘要

The Faultless underground nuclear test, conducted in central Nevada, is the site of an ongoingrnenvironmental remediation effort that has successfully progressed through numerous technical challengesrndue to close cooperation between the U.S. Department of Energy, (DOE) National Nuclear SecurityrnAdministration and the State of Nevada Division of Environmental Protection (NDEP). The challengesrnfaced at this site are similar to those of many other sites of groundwater contamination: substantialrnuncertainties due to the relative lack of data from a highly heterogeneous subsurface environment.rnKnowing when, where, and how to devote the often enormous resources needed to collect new data is arncommon problem, and one that can cause remediators and regulators to disagree and stall progress towardrnclosing sites. For Faultless, a variety of numerical modeling techniques and statistical tools are used tornprovide the information needed for DOE and NDEP to confidently move forward along the remediationrnpath to site closure. A general framework for remediation was established in an agreement and consentrnorder between DOE and the State of Nevada that recognized that no cost-effective technology currentlyrnexists to remove the source of contaminants in nuclear cavities. Rather, the emphasis of the correctivernaction is on identifying the impacted groundwater resource and ensuring protection of human health and thernenvironment from the contamination through monitoring. As a result, groundwater flow and transportrnmodeling is the linchpin in the remediation effort. An early issue was whether or not new site data shouldrnbe collected via drilling and testing prior to modeling. After several iterations of the Corrective ActionrnInvestigation Plan, all parties agreed that sufficient data existed to support a flow and transport model forrnthe site. Though several aspects of uncertainty were included in the subsequent modeling work, concernsrnremained regarding uncertainty in individual parameter values and the additive effects of multiple sourcesrnof uncertainty. Ultimately, the question was whether new data collection would substantially reducernuncertainty in the model. A Data Decision Analysis (DDA) was performed to quantify uncertainty in thernexisting model and determine the most cost-beneficial activities for reducing uncertainty, if reduction wasrnneeded. The DDA indicated that though there is large uncertainty present in some model parameters, thernoverall uncertainty in the calculated contaminant boundary during the 1,000-year regulatory timeframe isrnrelatively small. As a result, limited uncertainty reduction can be expected from expensive characterizationrnactivities. With these results, DOE and NDEP have determined that the site model is suitable for movingrnforward in the corrective action process. Key to this acceptance is acknowledgment that the model requiresrnindependent validation data and the site requires long-term monitoring. Developing the validation andrnmonitoring plans, and calculating contaminant boundaries are the tasks now being pursued for the site. Thernsignificant progress made for the site is due to the close cooperation and communication of the partiesrninvolved and an acceptance and understanding of the role of uncertainty.