Method‐dependent influence of environmental variables on reef fish assemblages when comparing trap and video surveys

Despite substantial survey effort and a large body of literature on abiotic and biotic factors in temperate reef ecosystems, knowledge of the complex and interactive effects of environmental variables on those communities is limited. Various survey methods have been developed to study environmental predictors of biodiversity, but there remains a gap in our understanding of how survey results are influenced by environmental factors. Here, we surveyed the fish assemblage associated with southeastern U.S. temperate marine reefs with simultaneous, paired trap, and camera gears throughout a ~50,000 km2 area during 2011–2013 and assessed the influence of environmental variables on the trap‐ and video‐surveyed assemblages. Predictor variables in the multivariate general linear models included depth, temperature, month, year, location, substrate relief, percent sessile biota, biota type, and turbidity. Depth and latitude had the greatest influence on the fish assemblage for both gears. The influence of habitat variables differed between methods and percent biota explained more variation in the fish assemblage when assessed by traps, while substrate relief and biota type explained more variation in the fish assemblage when assessed by video. In general, habitat complexity was positively related to the abundance of fishes in the video survey, but there was a negative relationship in the trap survey. Differences between gears were species‐specific and the influences of environmental variables were similar for some species such as Haemulon plumierii and Hyporthodus niveatus. The methods presented here can be used to assess method‐dependent differences in fish assemblages, which is a necessary precursor to assess the effect of environmental variables on the accuracy of surveys.