Analysis of uncertainty

Mishra, S., Parker, J. C., and Kaluarachchi, J. J., 1989b, Analysis of Uncertainty in Predictions of Hydrocarbon Recovery from Spill Sites Journal of Contaminate Hydrology, in review. 

Beck MB. 1987. Water quabty modeling a review of the analysis of uncertainty. Water Resour Res 23 1393-1442. 

Heugens, E.H.W, Hendriks, A. J., Reede, T., Straalen, N.M. van, Admiraal, W. (2001). A review on the effects of multiple stressors on aquatic organisms and analysis of uncertainty factors for use in risk assessment. Toxicology 31, 247-... 

Tiedeman, C., and Gorelick, S. M. (1993). Analysis of uncertainty in optimal contaminant capture design. Water Resources Research, 29(7), 2139-2153. 

Tier 0 exposure assessments, which are commonly used for first-tier screening purposes, do not require an analysis of uncertainty on every occasion, provided they include appropriate conservative assumptions or safety factors to take account of uncertainty. 

In view of the often considerable limitations of available data supporting exposure assessment, which sometimes limit the extent of uncertainty quantification and the need to explicitly identify sources of uncertainty prior to their quantification, this section provides an overview of existing concepts and proposes a harmonized approach for the qualitative analysis of uncertainty in exposure assessment. 

The level of sophistication with which uncertainty is assessed is necessarily dependent on the use that will be made of the information (Pate-Comell, 1996a). Sometimes, a level zero analysis of uncertainty is all that is needed, simply asking are there any major sources of uncertainty At a slightly higher level, we can ask what are the controversial sources of uncertainty If we believe that we can afford to incur this uncertainty, the analysis can end there. 

This section provides an overview of common methods for quantitative uncertainty analysis of inputs to models and the associated impact on model outputs. Furthermore, consideration is given to methods for analysis of both variability and uncertainty. In practice, commonly used methods for quantification of variability, uncertainty or both are typically based on numerical simulation methods, such as Monte Carlo simulation or Latin hypercube sampling. However, there are other techniques that can be applied to the analysis of uncertainty, some of which are non-probabilistic. Examples of these are interval analysis and fuzzy methods. The latter are briefly reviewed. Since probabilistic methods are commonly used in practice, these methods receive more detailed treatment here. The use of quantitative methods for variability and uncertainty is consistent with, or informed by, the key hallmarks of data... 

The general equation for risk assessment is the product of exposure and toxic potential. The above discussion, focused on human health exposure assessment, is clearly applicable to the analysis of uncertainty in the determinants and predictions of toxic effect per unit exposure. Indeed the complete assessment of risk and its uncertainty comes from the integrated evaluation of uncertainty associated with both toxicity and exposure. It is simply an extension of the technique demonstrated above with the addition of algorithms for toxic effect yielding a forecasted distribution of predicted risk. 

It is important to clearly describe and quantitatively estimate the assumptions and uncertainties involved in the evaluation, where possible. Examples include natural variability in ecological characteristics and responses and uncertainties in the test system and extrapolations. The description and analysis of uncertainty in characterization of ecological effects are combined with uncertainty analyses for the other ecological risk assessment elements during risk characterization. 

The analysis of uncertainty sources and components reveals that the uncertainty value of the spray cabinet corrosivity depends on the measurements of the surface area of the RS and mass loss in the corrosion processes. 

While the scientific basis for risk assessment has been described in the previous chapters, certain critical issues were left up in the air. Scientists have not yet been able, for example, to establish with certainty the relevance to humans of animal toxicity findings, and generally have poor or even no empirical data regarding dose-response relations at the human dose levels typically associated with environmental chemicals. Moreover, toxicity data gaps of one sort or another exist for all chemicals sometimes the gaps are not of great importance, but often they will seriously hamper the risk assessment process. So, while risk assessors may know much, they are always faced with significant scientific uncertainties. Indeed, some observers think risk assessment is best described as the analysis of uncertainty. 

To identify sources of uncertainly in deriving the current MeHg RfD, ERA conducted an analysis of uncertainties (ERA 1997b, Vol. VI Appen... 

Aity refinement of the current RfD wiU require consideration of sources of uncertainty. The committee has evaluated the body of evidence, focusing on the prospective epidemiological studies of neurotoxicity in children exposed in utero. Refinement of the current RfD based on results from these studies wiU require both quantitative and qualitative analysis of uncertainties to guide the application of uncertainty factors. 

According to the Battelle-NBS report (1, 2, 8), the total US cost of metallic corrosion was estimated to be 70 billion, which comprised 4.2% of GNP in 1975, and 15% or 10 billion was estimated to be avoidable by using the available corrosion prevention technology. The final results based on NBS analysis of uncertainty in the Battelle input/output model estimated the metallic corrosion cost of 82 billion or 4.9% of 1,677 trillion GNP in 1975. Nearly 33 billion or 2% ofGNPwas estimated to be avoidable. 

Wang Chi 2002. Analysis of uncertainty in measurement of flow. Beijing China Measurement Publishing House. 

Thanks in large part to work done in several of the programs in technology and policy, today modern policy analytic work is much improved, both in terms of the way in which problems are framed and the analytical tools that are employed, than was the case 30 years ago. For example, techniques such as decision analysis, the systematic characterization and analysis of uncertainty, and methods in quantitative risk analysis, that were pioneered in several of these programs, are now almost ubiquitous. 

---