The ease with which the uncertainty can be quantified

The ease with which the uncertainty can be quantified

Quantifying parameter uncertainty can be relatively simple with collection of

appropriate data. Model uncertainty and scenario uncertainty are more difficult to

quantify because they cannot be analyzed using straightforward statistical or

mathematical techniques (Suter 1993). When uncertainty is not quantifiable (e.g.,

it cannot be described by a distribution or a range and there is no suitable

analytical solution), then it becomes difficult to assess the magnitude of the

effect. For example, choosing the contaminants of concern is partly restricted by

what is possible to detect in the laboratory.

The ease with which uncertainty associated with different sources can be

reduced was judged using the following criteria:

a. Is the uncertainty source categorized as parameter uncertainty?

b. Are data, models, or other types of quantitative information available to

estimate uncertainty or can this information be obtained easily?

c. Is qualitative information available (e.g., expert opinion) to determine

importance of uncertainty?

d. Have uncertainty analyses been performed?

For each source of uncertainty, a "yes" response to at least three of these

questions resulted in a ranking of "easy." However, occasionally, these sources

have been ranked "moderate" to account for limited validation of available

models. A "yes" response to two of these questions resulted in a ranking of

"moderate," and a "yes" response to none or one of these questions resulted in a

ranking of "difficult."

This criterion was not used in the overall preliminary ranking; however, the

results of this ranking will be useful in the future for planning efforts to reduce

uncertainty. Overall preliminary scores can range from 1 to 6. Chapter 6

summarizes results of the preliminary ranking of uncertainty sources.

Chapter 2 Preliminary Ranking Approach