Relationships among map resolution, fish assemblages, and habitat variables in a coral reef ecosystem

Abstract

Benthic maps are broad-scale characterizations that lack the detailed environmental attributes which have been the focus of most prior empirical studies linking reef fish and habitat. We used multivariate analyses to quantify correlations among fish assemblages, local habitat variables, and reef types depicted in benthic maps. Benthic maps of a study system in the U.S. Virgin Islands with high (100 m2 minimum mapping unit) and low (4,048 m2 minimum mapping unit) spatial resolution, respectively, were evaluated. Benthic maps depicted six reef types and two shelf positions (lagoon vs. shelf). Fish assemblages and local environmental variables were quantified at random sites throughout the landscape by diver surveys. Multivariate ordination based on either fish assemblages or environmental data did not result in well-separated groups of sites. Mapped reef types were not associated with distinct values of either local environmental variables or fish assemblages. Reef types exhibited substantial overlap in ordination plots based on benthic characteristics with groupings based on fish assemblages showing even greater overlap. Ordination patterns involving reef type were largely the same for both low- and high-resolution maps. In contrast, sites showed clear groups for lagoon and shelf in ordinations based on both environmental variables and fish assemblage composition, respectively. These results suggest that knowledge of the overall fish assemblage or fine-scale environmental characteristics could not be used to predict reef type depicted in benthic maps or vice versa. In contrast, reef zone could be used to predict fish assemblage or fine-scale environmental variables and vice versa.