A hierarchical model for the complexity of plant communities

A hierarchical model for ecological community structure is proposed. The higher levels are not simply summations of lower levels but represent organizations requiring their own sets of explanatory principles. Terms become redefined in content or context with a change in level. Disturbance at a low level of organization may be a stabilizing force at higher levels without contradiction because the differences between the levels keep the two descriptions disjunct. Levels are defined by the filters used by the observer. Here two levels of organization are displayed in two separate principal component analyses of prairie permanent quadrats. Analyses of cover data display a level where fire is a perturbation, while analyses of presence data focus upon a level in which fire is incorporated as a stabilizing factor of a healthy prairie. The general implication for data transformation is that each transformation may be profitably viewed as a filtering operation that emphasizes structure of different grain size in the data. The effect is to cut into the biological structure at different levels of resolution so as to display structure at different levels of organization.