making a better or more-defensible or lower-risk decision. In other words,
the availability of the model, and any costs associated with delaying decisions, need to be weighed just as uncertainties are weighed.
5.2.10 Scenarios and Sensitivity Analyses
Do the models have the ability to assess the efficacy of specific management options? An important use of models is to assess the implications of
alternate management options or other changes in the system of concern.
Properly developed models can provide tremendous insights by being
scenario friendly (i.e., able to be used in a straightforward manner to test
different input or parameter conditions). For example, a model being used
to assess the potential consequences of global change on a wetland ecosystem would be much more useful (and therefore more used) if it is easy
to change the specific weather drivers such as inter- and intra-annual
variability in precipitation, combinations of elevated temperatures and
altered humidity, or potential evapotranspiration. Allowing pairwise comparison of model outputs of selected endpoints with scenarios that are identical except for one specific factor is a powerful tool for assessing
alternatives and for assessing the importance of one variable over another.
Sensitivity analyses, which examine the relative change in selected output(s)
with change in one or more variables, can be enormously helpful in identifying important uncertainties or major research needs. Conversely, they can
also be important for identifying those aspects that do not greatly influence
the outcome. The more a model can be used for scenario and sensitivity
analyses and the more that visualization techniques are incorporated into
the model to illustrate the message of these analyses, the more useful the
model will be in terms directly relevant to the decision process.
5.2.11 Absolute versus Relative Results
Can the questions being asked of the model be put into a comparative risk
context? Generally it is much easier to be accurate in predicting a relative
result than an absolute one. That is, many times a model is more reliable in
comparing the relative risks between two options than it is in predicting the
absolute outcome for either option. Thus, the demand placed on a model
in a comparative risk assessment may be less, effectively reducing the
importance of uncertainties in the model results. Another way to look at a
comparative ecological risk assessment is as a special case of scenario analyses, except that each major option may itself be examined in the context
of a suite of scenarios. For example, if two alternate fuels were being considered for a power plant on a bay and the ecological risks being assessed
are associated with spills of the fuels in transportation to the power plant,
then the risks of each fuel type can be assessed with a suite of scenarios
of the specific conditions of the fuel spills (e.g., different weather or tide
5. Overcoming Barriers to the Use of Models
101
the availability of the model, and any costs associated with delaying decisions, need to be weighed just as uncertainties are weighed.
5.2.10 Scenarios and Sensitivity Analyses
Do the models have the ability to assess the efficacy of specific management options? An important use of models is to assess the implications of
alternate management options or other changes in the system of concern.
Properly developed models can provide tremendous insights by being
scenario friendly (i.e., able to be used in a straightforward manner to test
different input or parameter conditions). For example, a model being used
to assess the potential consequences of global change on a wetland ecosystem would be much more useful (and therefore more used) if it is easy
to change the specific weather drivers such as inter- and intra-annual
variability in precipitation, combinations of elevated temperatures and
altered humidity, or potential evapotranspiration. Allowing pairwise comparison of model outputs of selected endpoints with scenarios that are identical except for one specific factor is a powerful tool for assessing
alternatives and for assessing the importance of one variable over another.
Sensitivity analyses, which examine the relative change in selected output(s)
with change in one or more variables, can be enormously helpful in identifying important uncertainties or major research needs. Conversely, they can
also be important for identifying those aspects that do not greatly influence
the outcome. The more a model can be used for scenario and sensitivity
analyses and the more that visualization techniques are incorporated into
the model to illustrate the message of these analyses, the more useful the
model will be in terms directly relevant to the decision process.
5.2.11 Absolute versus Relative Results
Can the questions being asked of the model be put into a comparative risk
context? Generally it is much easier to be accurate in predicting a relative
result than an absolute one. That is, many times a model is more reliable in
comparing the relative risks between two options than it is in predicting the
absolute outcome for either option. Thus, the demand placed on a model
in a comparative risk assessment may be less, effectively reducing the
importance of uncertainties in the model results. Another way to look at a
comparative ecological risk assessment is as a special case of scenario analyses, except that each major option may itself be examined in the context
of a suite of scenarios. For example, if two alternate fuels were being considered for a power plant on a bay and the ecological risks being assessed
are associated with spills of the fuels in transportation to the power plant,
then the risks of each fuel type can be assessed with a suite of scenarios
of the specific conditions of the fuel spills (e.g., different weather or tide
5. Overcoming Barriers to the Use of Models
101