The effect of bait on fine-scale habitat associations of reef fish investigated with remote underwater video systems

Establishing the associations between fish and their habitats can aid in the monitoring of fish stocks and the design of effective marine protected areas (MPAs). Baited remote underwater stereo-video systems (stereo-BRUVs) are now commonly used to asses fish populations. The habitats seen in the video footage of stereo-BRUVs can be used to link fish fauna to preferred habitat types. However, the application of bait potentially attracts fish from surrounding habitats, and might result in a biased understanding of fish–habitat associations. A field study was conducted in the Tsitsikamma National Park MPA to determine the effect of bait on fine-scale fish–habitat associations, using remote photographic and video methods. The study was conducted over the summer season of 2015 and 2016. Data were collected within a 1x1 km shallow (9–44 m) reef complex. Within the sampling area, 944 photo-quadrats of the macrobenthos were taken 30 m apart by means of a drop camera. By separating the macrobenthos into broad taxonomic groups, five habitat types were identified, namely Shallow Sand, Shallow Reef, Deep Reef, Deep Sand and Patch Reef. The results show that even on a fine scale, depth is an important predictor of macrobenthic distribution and assemblage structure. Baited (stereo-BRUVs) and unbaited (stereo-RUVs) surveys were then conducted to sample the fish community in the same area during the period under study. Higher abundances of fish were observed in reef than in sandy habitats, and bait was seen to have a positive effect on species richness and fish abundance. When comparing habitats, fish abundance and composition on reef habitats were significantly different from sand habitats. This was observed in both the stereo-RUVs and stereo-BRUVs methods. High counts of roman (Chrysoblephus laticeps), fransmadam (Boopsoidea inornata) and steentjie (Spondyliosoma emarginatum) in reef habitats were contrasted by high counts of white sea catfish (Galeichthys feliceps), evil-eye puffer (Amblyrhynchotes honckenii) and lesser guitarfish (Rhinecanthus annulatus) in sandy habitats. Overall, the underlying patterns in fish diversity recorded with the two video methods were generally comparable. However, stereo-RUVs appeared to be unable to detect species that were present in sand habitats, while stereo-BRUVs increased the number and abundance of species recorded in all habitat types. In the stereo-RUVs footage, differences between reef habitats were dampened by the presence of highly abundant fish species. In the stereo-BRUVs footage, although bait appeared to have an effect on the observed fish assemblage, this manifested in an increase in species richness, higher fish abundances and a better overall ability to detect fish–habitat relationships. As such, stereo-BRUVs are considered a robust, effective and recommended method for detecting fish–habitat relationships, even over a fine scale