Characteristics of benthic maps are controlled by the spatial and thematic resolutions used in map production. The implications of these production decisions on the inferences that can be drawn from the maps are poorly understood. I addressed this need by quantifying the differences among common map types, considering how map type affects inferences of fish and benthic communities at the patch level, and then evaluating the influence of map type on ecological neighborhood analysis of reef fish. Results indicated that hard bottom types, especially patch reefs and colonized pavement, were among the most sensitive to changes in spatial resolution of maps. In contrast, linear reef and continuous seagrass features were characterized quite consistently regardless of spatial resolution. Multivariate analyses indicated that both the fish assemblages and benthic characteristics of reef types overlapped considerably. In contrast, shelf position (inside versus outside of lagoons) showed clear differences in both environmental variables and fish assemblage composition. In general, the results of multivariate analyses suggest that knowledge of the overall fish

assemblage or fine-scale environmental characteristics could not be used to predict reef type or vice versa. Furthermore, spatial scale of benthic maps did not affect results when analyses were conducted at the patch level. In addition, a multi-scale landscape analysis was conducted wherein correlations between fish assemblages and surrounding landscape variables were measured using univariate linear regression for a range of scales between 25 and 800 m. The strength of the associations as a function of scale exhibited one of 6 response curve forms and was used to identify the scale that best correlates fish with their surrounding habitat. In these analyses, individual landscape variables explained a maximum of only 25% of the variability in fish distributions. Use of different input maps in many of these analyses resulted in a changed perception of either the strength of peak correlation at a given scale, or the scale at which peak correlations occurred. Overall, the findings revealed which aspects of coral reef ecosystems are sensitive to map scale and advise scientists and managers on map production and use in similar settings.