Visual censuses of tropical fish aggregations on artificial reefs: slate versus video recording techniques

Artificial coastal reefs were initially installed off Reunion Island (SW Indian Ocean) in order to attract and catch small pelagic fish of economic importance for the local fishery. These artificial reefs also attract a high density of juvenile coral reef fish, especially after recruitment. Because of high species richness and density, a suitable method of evaluating diversity and abundance in such assemblages is needed. An experiment was conducted in situ for I year in order to test which of the two visual techniques, state or video recording, would be the most suitable for sampling artificial reefs. For the quantitative study, a rope framework (four belt transects each measuring 24 x 3 in) was used to ease the movement of the divers. To avoid bias due to time variations, the study was conducted under limited time in situ. The advantages and disadvantages of the two techniques regarding the accuracy of qualitative and quantitative results are discussed. The qualitative study results show that the slate technique is the most accurate. one for determining species richness. In the quantitative study, factors such as species characteristics (ecological type), attractiveness (relationship between fish and diver) and abundance of fish communities around the artificial reefs appear to influence our results on the comparison of the two techniques. Of 13 'permanent' (observed in more than 75% of dives) species, 6 returned a similar frequency of detection by slate and video (Lutjanus kasmira, Lutjanus bengalensis, Priacanthus hamrur, Chaetodon kleinii, Heniochus diphreutes, Mulloides vanicolensis); these were then selected for comparison of the techniques using quantitative data. The correlation between video and slate census results is high, but, in comparison with the slate census, the video census underestimates abundance, especially for low values. A corrected index is proposed to remedy this underestimation; it could be used for video surveys conducted on species responding well to video detection (as below). (C) 2004 Elsevier B.V. All rights reserved.