and space, and interact at numerous scales. Developing appropriate parallel implementations to model these interactions is quite difficult, and only
limited research has been conducted. However, the availability of parallel
architectures for ecological modeling allows one to conceptualize models
that may be considerably more realistic than strictly serial implementations would be. For example, Mellott et al. (1999) investigated parallel
methods for an individual-based predator–prey model and point out that
the parallel implementation involved quite different assumptions about
individual movements and interactions than were necessary in a prior serial
implementation.
8.3.2 Relevance of Advances in Model Implementation
for Decision Making
The advances in model implementation outlined above will enhance decision making in the long term by allowing modelers to improve the sophistication and relevance of models. Public expectations of resource managers
are steadily increasing, requiring more-definitive abilities to predict the
consequences of management actions. Much of the information currently
needed by managers is not available because the models have not yet been
developed or provide inadequate information. This void exists, at least
partly, because of limitations in computing power or analytical and
conceptual-modeling capabilities. A combination of technological advances
and improved dialogue between modelers and managers is needed to
fully realize the potential of ecological models to enhance environmental
decision making.
8.4 Communicating Model Structure and Output
Managers are reluctant to use model results for making decisions unless
they are confident that they understand how the model works and that the
model, in fact, accurately produces the information they require. Models
that are perceived as an incomprehensible black box will not be widely used
by managers. Consequently, it is critical that an implemented modeling
system be adequately explained and communicated both to managers and
to stakeholders affected by management decisions. A number of techniques
are available to enhance the communication of models to decision makers,
making their structure and function more transparent.
8.4.1 Approaches and Technologies
8.4.1.1 Artificial Intelligence
Artificial intelligence (AI) refers to a branch of computer science focused
on problems associated with the acquisition, representation, and utilization
150
Eric Gustafson et al.
limited research has been conducted. However, the availability of parallel
architectures for ecological modeling allows one to conceptualize models
that may be considerably more realistic than strictly serial implementations would be. For example, Mellott et al. (1999) investigated parallel
methods for an individual-based predator–prey model and point out that
the parallel implementation involved quite different assumptions about
individual movements and interactions than were necessary in a prior serial
implementation.
8.3.2 Relevance of Advances in Model Implementation
for Decision Making
The advances in model implementation outlined above will enhance decision making in the long term by allowing modelers to improve the sophistication and relevance of models. Public expectations of resource managers
are steadily increasing, requiring more-definitive abilities to predict the
consequences of management actions. Much of the information currently
needed by managers is not available because the models have not yet been
developed or provide inadequate information. This void exists, at least
partly, because of limitations in computing power or analytical and
conceptual-modeling capabilities. A combination of technological advances
and improved dialogue between modelers and managers is needed to
fully realize the potential of ecological models to enhance environmental
decision making.
8.4 Communicating Model Structure and Output
Managers are reluctant to use model results for making decisions unless
they are confident that they understand how the model works and that the
model, in fact, accurately produces the information they require. Models
that are perceived as an incomprehensible black box will not be widely used
by managers. Consequently, it is critical that an implemented modeling
system be adequately explained and communicated both to managers and
to stakeholders affected by management decisions. A number of techniques
are available to enhance the communication of models to decision makers,
making their structure and function more transparent.
8.4.1 Approaches and Technologies
8.4.1.1 Artificial Intelligence
Artificial intelligence (AI) refers to a branch of computer science focused
on problems associated with the acquisition, representation, and utilization
150
Eric Gustafson et al.