Video transects are the most appropriate underwater visual census method for surveying high-latitude coral reef fishes in the southwestern Indian Ocean

Abstract

Pilot studies are used to assess different sampling and monitoring protocols when developing long-term monitoring programs (LTMPs). Considering that the ultimate goal of LTMPs is preservation-orientated, non-destructive sampling methods are favoured. For coral reef fishes, underwater visual censuses (UVCs) have become the method of choice for surveying shallow, highly-diverse reefs. This study assessed three different UVC transect types to monitor high-latitude western Indian Ocean coral reefs; these were slate, photo and video transects. Slate data were analysed with CountN (all fish observed), while footage from reviewable photo and video transects was analysed by both CountN and MaxN (a single frame with the most fish per species), resulting in an assessment of five transect-type analysis technique combinations; slate, photo-CountN, video-CountN, photo-MaxN and video-MaxN. Video-CountN transects had the highest estimates of richness and abundance, the lowest variability, and the most representative diversity. Video-CountN transects were also the most efficient in that the time taken to analyse footage was comparable to the other combinations considered, the proportion of unidentifiable individuals was low, and it required the lowest sample size to achieve an acceptable statistical power. Video-CountN transecting is therefore recommended as the most appropriate method for conducting UVCs of high-latitude coral reef fishes in the southwestern Indian Ocean.