EXPERIMENTAL-ANALYSIS OF PATCH DYNAMICS IN TALLGRASS PRAIRIE PLANT-COMMUNITIES

Previous research has indicated that patch structure at small spatial scales (< 100 m 2) in tallgrass prairies was defined by a diverse array of infrequent species because dominant species occurred in all samples at this scale. Also, patch structure was not significantly different from that derived from random species associations. Based on these results, we hypothesized that removal of adominant species would have no effect on patch structure in these prairies. We tested this hypothesis by removing a dominant grass, Schizachyrium scoparium (Poaceae), from half of each of four 10 m x 10 m study blocks, and comparing differences in patch structure between control and removal halves before and after removal. The minimum resolution in our study was 1 m2. Patches of similar species composition were defined by cluster analysis of presence/absence data and cover data. Patch sizes ranged from 1 to 34 m2. Following the removal of S. scoparium there was an overall increase in the number of species in the removal half of each block compared to pre-treatment levels. However, the number of patch types and number of spatially mapped groups, based on presence/absence or cover data, did not change between control and removal plots after the removal of S. scoparium. This supports the hypothesis that removal of a large, dominant species would have no effect on patch structure at this scale of resolution in these prairies. Thus, patch structure, as defined here, is an emergent property in these grasslands that is not predictable from changes in species composition. This property of stochastic patch structure results from interactions of processes operating at scales both larger and smaller than our scale of resolution. Stochastic models may provide a reasonable approach to modelling small-scale patch dynamics in tallgrass prairie communities.