Uncertainties in environmental radiological assessment models and their implications
Environmental radiological assessments rely heavily on the use of mathematical models. The predictions of these models are inherently uncertain because these models are inexact representations of real systems. The major sources of this uncertainty are related to biases in model formulation and parameter estimation. The best approach for estimating the actual extent of over- or underprediction is model validation, a procedure that requires testing over the range of the intended realm of model application. Other approaches discussed are the use of screening procedures, sensitivity and stochastic analyses, and model comparison. The magnitude of uncertainty in model predictions is a function of the questions asked of the model and the specific radionuclides and exposure pathways of dominant importance. Estimates are made of the relative magnitude of uncertainty for situations requiring predictions of individual and collective risks for both chronic and acute releases of radionuclides. It is concluded that models developed as research tools should be distinguished from models developed for assessment applications. Furthermore, increased model complexity does not necessarily guarantee increased accuracy. To improve the realism of assessment modeling, stochastic procedures are recommended that translate uncertain parameter estimates into a distribution of predicted values. These procedures also permit the importance of model parameters to be ranked according to their relative contribution to the overall predicted uncertainty. Although confidence in model predictions can be improved through site-specific parameter estimation and increased model validation, risk factors and internal dosimetry models will probably remain important contributors to the amount of uncertainty that is irreducible.
- Research Organization:
- Oak Ridge National Lab., TN (USA)
- DOE Contract Number:
- W-7405-ENG-26
- OSTI ID:
- 5986310
- Report Number(s):
- CONF-830460-3; ON: DE83014019
- Resource Relation:
- Conference: Annual meeting of the National Council of Radiation Protection and Measurements, Washington, DC, USA, 6 Apr 1983; Other Information: Portions are illegible in microfiche products
- Country of Publication:
- United States
- Language:
- English
Similar Records
Geostatistical and Stochastic Study of Flow and Transport in the Unsaturated Zone at Yucca Mountain
The role of uncertainty analysis in dose reconstruction and risk assessment
Related Subjects
RADIOACTIVE EFFLUENTS
RADIATION HAZARDS
RISK ASSESSMENT
MATHEMATICAL MODELS
ACCURACY
ENVIRONMENTAL EXPOSURE PATHWAY
MAN
PROBABILITY
RADIOISOTOPES
RADIONUCLIDE MIGRATION
VALIDATION
ANIMALS
ENVIRONMENTAL TRANSPORT
HAZARDS
HEALTH HAZARDS
ISOTOPES
MAMMALS
MASS TRANSFER
MATERIALS
PRIMATES
RADIOACTIVE MATERIALS
RADIOACTIVE WASTES
TESTING
VERTEBRATES
WASTES
560151* - Radiation Effects on Animals- Man
560161 - Radionuclide Effects
Kinetics
& Toxicology- Man