The consequences of hierarchy for modeling in forest ecosystems

An overview of forest ecosystem models in the context of hierarchies is presented, considering spatiophysical, mechanistic, temporal, usage, and range resolutions. We argue that models of forest ecosystem dynamics can be treated as inherently hierarchical structures of discrete submodels or modules. These submodels represent distinct but connected processes, such as growth, mortality, or regeneration. This approach has important ramifications for model building, fitting, criticism, and application. It provides new strategies for dealing with common modeling problems, such as component choice, as well as placing in context established modeling strategies. A system of categories that avoids popular and simplistic dichotomies is provided to assist in model classification. The need for more sophisticated techniques at each stage of model construction is demonstrated, and candidate solutions are suggested.