Herbaceous plant community responses to fluctuations in hydrology: Using Mississippi River islands as models for plant community assembly

MOORE, J. E. (Department of Biological Sciences, The University of Memphis, Memphis, TN 38152), S. B. FRANKLIN (Department of Biological Sciences, The University of Northern Colorado, Greeley, CO 80639), AND J. W. GRUBAUGH (Department of Biological Sciences, The University of Tennessee at Martin, Martin, TN 38238). Herbaceous plant community responses to fluctuations in hydrology: Using Mississippi River islands as models for plant community assembly. J. Torrey Bot. Soc. 138: 177-191. 2011.-To understand how plant communities respond to fluctuations in hydrology, we conducted a natural study on five riverine islands located within the lower Mississippi River. Specifically, we addressed whether species richness and diversity peak at intermediate elevations of flooding and whether plant assembly on these islands follow a niche-derived continuum pattern or develop due to chance processes. Elevations of islands were standardized into 1 m zones providing replicate island gradients of flood frequency and intensity. We examined effects of changing frequency and duration along the gradient and between years, and changes in intensity due to flow based on island side (main channel sides had greater intensity floods). We analyzed diversity data using nonparametric general linear models and compositional dissimilarity using Sorenson's index. Less frequent, more intense disturbances decreased both landscape level and local diversity with the frequency and intensity (alone) of floods affecting community composition. Both temporally and spatially, 'average' disturbances beget higher diversity, supporting the intermediate disturbance hypothesis. Compositional data support multiple controls of community assembly. Clear gradients due to niche drivers were found in relation to typical flooding regimes and so specific species have adapted to various frequencies and intensities of disturbance. However, chance played some role throughout the gradient, resulting in either very high similarity or very low similarity where diversity was low and generally low similarity (not different from random), albeit also more permutations of communities, where diversity was high along the gradient.