Uncertainty and sampling issues in tank characterization
- and others
A defensible characterization strategy must recognize that uncertainties are inherent in any measurement or estimate of interest and must employ statistical methods for quantifying and managing those uncertainties. Estimates of risk and therefore key decisions must incorporate knowledge about uncertainty. This report focuses statistical methods that should be employed to ensure confident decision making and appropriate management of uncertainty. Sampling is a major source of uncertainty that deserves special consideration in the tank characterization strategy. The question of whether sampling will ever provide the reliable information needed to resolve safety issues is explored. The issue of sample representativeness must be resolved before sample information is reliable. Representativeness is a relative term but can be defined in terms of bias and precision. Currently, precision can be quantified and managed through an effective sampling and statistical analysis program. Quantifying bias is more difficult and is not being addressed under the current sampling strategies. Bias could be bounded by (1) employing new sampling methods that can obtain samples from other areas in the tanks, (2) putting in new risers on some worst case tanks and comparing the results from existing risers with new risers, or (3) sampling tanks through risers under which no disturbance or activity has previously occurred. With some bound on bias and estimates of precision, various sampling strategies could be determined and shown to be either cost-effective or infeasible.
- Research Organization:
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE Office of Environmental Restoration and Waste Management, Washington, DC (United States)
- DOE Contract Number:
- AC06-76RL01830
- OSTI ID:
- 548889
- Report Number(s):
- PNNL-11529; ON: DE97053945; TRN: 98:000922
- Resource Relation:
- Other Information: PBD: Jun 1997
- Country of Publication:
- United States
- Language:
- English
Similar Records
Toward a lean analysis for waste to support the decommissioning operation - 15321
An Iterative Uncertainty Assessment Technique for Environmental Modeling