Abstract

The advent of high-speed computing has facilitated a rapid change in the way that ecological systems are modeled. Population models can now explicitly represent the complex interplay between the local environment and the individual. This complexity is also the major liability of these models because parameter estimation, sensitivity to particular functional forms, and error propagation become important problems. In addition, it becomes increasingly difficult to understand which interactions contribute most to large-scale dynamics. A complex model of forest dynamics (SORTIE) was explored through a series of model manipulations and experiments. The model was subjected to several experiments, including propagation of uncertainty in the model parameters, introduction of large clear-cuts in the forest, increase of individual-tree mortality, removal of spatially explicit interactions, and approximation of the functional responses of species. The results from these model experiments were compared in order to evaluate the predictive capability of the SORTIE model. It is also shown how visualization of its behavior can be an integral part of model exploration and evaluation. The visualizations are viewed as fundamental steps in identifying the relevant detail in this forest model. Insofar as this mechanistic model mirrors nature, it provides insight into the critical detail controling the emergence of forest patterns from the interactions of trees. This quasi-experimental approach with a detailed, mechanistic model derived from data is a powerful method for addressing questions of relevant detail, emergent properties, and scale.

(Back Home)

(Next Page)

Copyright © 1997 by the American Association for the Advancement of Science.