questions to ask, but is an activity that can guide and shape the decisionmaking process itself by bridging science and policy needs (Figure 3.2). The
questions also are not answered all at once. Instead, answers evolve through
the modeling and decision-making processes as more is learned about the
water body, fish populations, and technology options.
The pattern that has emerged for selecting modeling approaches in
316(b) assessments is to start with a simple screening model and to increase
in model complexity as the situation merits. This pattern makes good sense,
knowing that a decision needs to be made in a short time frame and that
perfect knowledge is neither required nor attainable in any case. This
approach has meant, for example, starting with models of individual losses
or fractional losses without including density dependence (EPRI 1999;
EPRI in press). The next-more-complex approach has been age-based or
stage-based matrix projection modeling, with or without density dependence. Even more complicated and realistic modeling approaches have
emerged in highly contested cases where millions of dollars for retrofitting
cooling towers are at stake. The increased complexity may involve the
modification of existing code, new computer codes, and new modeling
approaches (e.g., individual-based modeling). A similar tiered sequence
of increasingly complex modeling approaches is typical in other fields (e.g.,
1-D, 2-D, and 3-D hydrodynamic and water quality models).
56
Webb Van Winkle and John Kadvany
Figure 3.2. Conceptual view of a proposed decision analysis framework for fisheries management, including risk assessment and risk management components
(Lane and Stephenson 1995, 1998).