Modeling Stream Fish Assemblages with Niche Apportionment Models: Patterns, Processes, and Scale Dependence

—Understanding patterns in terms of the mechanistic processes that produce them is the essence of ecology. However, many studies merely document nonrandom patterns of species coexistence without providing insight into the structuring mechanisms and their scale dependence. Moreover, few studies address the mechanisms driving functional diversity. The objectives of this study were to document empirical patterns of species coexistence within stream fish assemblages across a broad geographic region, addressing both the functional and taxonomic organization of those assemblages, and to determine whether the structuring mechanisms are scale dependent. We gathered data from a report published by the Texas Parks and Wildlife Department that provided empirical species abundance distributions for 62 assemblages from 7 ecoregions and 18 river basins. For each assemblage, we simulated expected species abundances using five previously published models of niche apportionment. These models generate relative abundances in proportion to niche breadth and are constrained to maintain observed species richness, but they do not include information on species identity. These models included four stochastic variants of a geometric series (dominance preemption, random fraction, broken stick, and dominance decay) in which species abundances relate directly to niche breadth and one model (random assortment) in which abundances are independent of niche partitioning altogether. To assess scale dependence, we repeated the analyses after pooling assemblages by ecoregion, river basin, and the entire state. The patterns of niche apportionment for assemblages pooled according to river basin are more consistent with local assemblages than with assemblages pooled according to ecoregions or the entire state. These results suggest that niche apportionment plays a crucial role in structuring stream fish assemblages and that functional groups divide niche space more evenly than do species. Two of the major goals in ecology are to determine the factors responsible for species coexistence and to use that information to formulate predictive models of community assembly (May 1975; Drake 1991; Strange and Foin 1999). Species coexistence in stream fishes is often a result of niche partitioning in which species differ along axes of food, habitat, or activity times (Ross 1986; Nishikawa and Nakano 1998; Herder and Freyhof 2006). In this case, species coexistence depends on either the amount of niche space that each species sequesters (i.e., niche breadth) or the total amount of available niche space. For small-scale studies (e.g., stream reach), estimating niche breadth is feasible by quantifying diet selection, habitat selection, and patterns of activity. This is not the case for large-scale studies, however , for which it is more practical to use species abundance distributions as indicators of niche breadth. Species abundance distributions are a fundamental characteristic of communities. Not only do they provide the basis for most measures of biodiversity, they also provide insight into how species subdivide niche space to coexist (Whitaker 1965; May 1975; Sugihara 1980; Tokeshi 1990; Magurran 2004). A suite of mechanistic models has been developed to specifically address niche partitioning (Tokeshi 1990). Each of these models conceptualizes niche space as being sequentially divided among the species as they enter the community; however, the amount of niche space that each new species sequesters differs with each model. These models are stochastic variants of a geometric series n i 1⁄4 NC k K(1 k) , where n i 1⁄4 the number of individuals in the ith species, N 1⁄4 the total number of individuals in the community, k 1⁄4 the proportion of the available niche space sequestered by each new species, and C k 1⁄4 [1 (1 k)] 1 is a constant that ensures that the total number of individuals equal N (Magurran 2004). (Although these models are based on the concept of sequential introduction of species, in practice the abundances of all species are estimated simultaneously.) Although these models were originally developed to assess the taxonomic organization of assemblages, they can also be used to assess functional organization. The functional organization of assemblages is becoming an increasingly important topic in stream fish ecology. First, functional characteristics allow * Corresponding author: [email protected] 1 Present address: Department of Biological Sciences, Tarleton State University, Box T-0100, Stephenville, Texas 76402, USA. Received March 19, 2007; accepted November 5, 2007 Published online April 10, 2008 696 Transactions of the American Fisheries Society 137:696–706, 2008 Copyright by the American Fisheries Society 2008 DOI: 10.1577/T07-061.1 [Article]