PARENT SESSION
Symposium #25: Scaling information from plots or regions to landscapes: when do details matter?.

Organized by: DPC Peters and JE Herrick
Thursday, August 8. 9:00 AM to 11:30 AM. Crystal Ballroom, TCC.

Model-based extrapolation in ecology: when does space really matter?

Urban, Dean^*,1, Gardner, Robert², ¹ Duke University, Durham, NC² University of Maryland Center for Environmental Science, Frostburg, MD

ABSTRACT- A range of modeling approaches can be used to make ecological predictions at broad scales based on fine-scale data. There is a great deal of confusion concerning the role of "space" in models but we can distinguish three broad categories of models: nonspatial, spatially implicit, and. spatially explicit models. Much of the confusion concerning these models arises because predictions from any of these can be rendered as a map—giving the appearance that the model is spatial. Adding spatial complexity to a model increases its realism and its ability to capture a few truly spatial processes and feedbacks, at the cost of a larger investment in model development and application as well as greater potential for error propagation. By contrast, many complex models can be parameterized as simpler spatially implicit models, at the cost of decreased generality and robustness to novel conditions. We discuss general cases that do call for explicit spatially models, offering these as guidelines to model selection for ecological applications.

KEY WORDS: simulation, contagion, forecasting, complexity