described in detail and that important steps in the development of a model
be referenced. This specification allows users of the model to trace the
development of its mathematical formulations and conceptual underpinnings to ensure that the model is properly applied. Such documentation
allows models to be categorized by application, dimensionality, spatial
discretization strategy, solution scheme, and temporal strategy. Within
each category, efforts should be made to standardize ecological models to
increase their ease of use and to increase their reliability. All model applications should undergo a rigorous, documented confirmation process
involving parameterization (estimating optimum values for model parameters), calibration (adjusting model parameters and model formulation to
match observed data with model predictions), and validation (verifying that
the model works correctly on a data set different from the data set used for
model calibration).
8.4.1.5 Visual Output
Visualization is a very powerful form of communication, as epitomized in
the adage that “a picture is worth a thousand words.” For models with a
spatial component, GIS provides tremendous communication potential by
placing model inputs and intermediate and final results in a spatial context.
A good example is a model predicting gray wolf habitat in the northern
lake states (Mladenoff et al. 1995). By showing the spatial distribution of
input variable values and the results of model calculations, the authors
make a compelling case for the utility and validity of their model.
The GIS also provides a framework for integrating information from
different modeling paradigms.An example is the development of integrated
forest management models, where a GIS provides the integration for timber
optimization models and process models predicting wildlife habitat and
biological diversity (Naesset 1997). The optimization model produces treatment schedules for forest stands, the locations of which are tracked in the
GIS. A spatial model that can access the GIS can assess the potential effects
on wildlife when those specific stands are harvested. Finally, GIS can act as
a catalyst for stakeholder involvement (Cornett 1994). People find it much
easier to relate to visualizations of data and concepts than to text and
numbers. Because “seeing is believing,” spatial representations of model
results can lower skepticism and increase the involvement of stakeholders
in the decision-making process. Maps, animations, or virtual reality pictures
are understood by most users (Shepard 2000). For example, FORSYS (a
cooperative for forest systems engineering) is developing graphical systems
to represent the data gathered by the national forests to visually demonstrate alternative management practices (McGaughy 2001).
While model and data visualizations may be very useful, there are limitations. Just as graphs can be constructed in ways that are misleading,
the huge variety of color schemes available can cause the same data to be
8. Evolving Approaches and Technologies
153
be referenced. This specification allows users of the model to trace the
development of its mathematical formulations and conceptual underpinnings to ensure that the model is properly applied. Such documentation
allows models to be categorized by application, dimensionality, spatial
discretization strategy, solution scheme, and temporal strategy. Within
each category, efforts should be made to standardize ecological models to
increase their ease of use and to increase their reliability. All model applications should undergo a rigorous, documented confirmation process
involving parameterization (estimating optimum values for model parameters), calibration (adjusting model parameters and model formulation to
match observed data with model predictions), and validation (verifying that
the model works correctly on a data set different from the data set used for
model calibration).
8.4.1.5 Visual Output
Visualization is a very powerful form of communication, as epitomized in
the adage that “a picture is worth a thousand words.” For models with a
spatial component, GIS provides tremendous communication potential by
placing model inputs and intermediate and final results in a spatial context.
A good example is a model predicting gray wolf habitat in the northern
lake states (Mladenoff et al. 1995). By showing the spatial distribution of
input variable values and the results of model calculations, the authors
make a compelling case for the utility and validity of their model.
The GIS also provides a framework for integrating information from
different modeling paradigms.An example is the development of integrated
forest management models, where a GIS provides the integration for timber
optimization models and process models predicting wildlife habitat and
biological diversity (Naesset 1997). The optimization model produces treatment schedules for forest stands, the locations of which are tracked in the
GIS. A spatial model that can access the GIS can assess the potential effects
on wildlife when those specific stands are harvested. Finally, GIS can act as
a catalyst for stakeholder involvement (Cornett 1994). People find it much
easier to relate to visualizations of data and concepts than to text and
numbers. Because “seeing is believing,” spatial representations of model
results can lower skepticism and increase the involvement of stakeholders
in the decision-making process. Maps, animations, or virtual reality pictures
are understood by most users (Shepard 2000). For example, FORSYS (a
cooperative for forest systems engineering) is developing graphical systems
to represent the data gathered by the national forests to visually demonstrate alternative management practices (McGaughy 2001).
While model and data visualizations may be very useful, there are limitations. Just as graphs can be constructed in ways that are misleading,
the huge variety of color schemes available can cause the same data to be
8. Evolving Approaches and Technologies
153