Biotic interactions and habitat drive positive co‐occurrence between facilitating and beneficiary stream fishes

Abstract

AbstractAimThe role of biotic interactions in determining species distributions is difficult to decouple from abiotic factors. Most research to date has focused on negative biotic interactions, but the importance of positive interactions such as mutualism and facilitation at large scales is less understood. We used a two‐species occupancy modelling approach to decouple the relative effects of abiotic factors and biotic interactions between a habitat‐modifying North American stream fish (Nocomis leptocephalus) and two of its nearly‐obligate beneficiary species (Chrosomus oreas and Clinostomus funduloides).LocationSixty‐one sites on tributaries to the New River basin in the Appalachian mountains of the eastern United States.MethodsWe sampled fishes using backpack electrofishing in the summers of 2012–2014. We gathered 10 habitat covariates from the National Hydrography Dataset and from instream measurements. We reduced dimensionality in habitat variables, corrected them for spatial autocorrelation and selected two eigenvectors as habitat covariates. We then used an information‐theoretic approach to compare two‐species occupancy models representing hypotheses specifying the importance of only habitat, only biotic interactions, or combinations of both.ResultsThe best model for each species combination specified the importance of biotic interactions. For both associate species, probabilities that associates would occur in the absence of their host were considerably lower than probabilities of co‐occurrence. Species interaction factors indicated positive patterns of co‐occurrence between hosts and associates. Models suggested that habitat variables mediated host‐associate interactions for C. oreas but not for C. funduloides.Main conclusionsThis study provides some of the first large‐scale quantitative evidence of positive co‐occurrence among vertebrates, and demonstrates the importance of abiotic context for mediating interspecific interactions. Two‐species occupancy modelling may be superior to traditional co‐occurrence analyses for parsing out the relative importance of biotic interactions and habitat variables for determining species distributions. However, experiments and small‐scale behavioural observations will also be necessary to confirm mechanisms.