Comparison of a rapid visual and a strip transect technique for censusing reef fish assemblages

Two visual census techniques frequently employed for quantifying coral reef fish assemblages, the rapid visual technique (RVT) and the strip transect technique (STT), were compared in terms of precision, observer bias, and species enumeration. Spearman rank correlation coefficients calculated between replicate RVT censuses were lower than for STT censuses, indicating that the RVT was less precise. The correlation coefficients for RVT censuses conducted by the same observer were greater than those for censuses conducted by different observers. Observer bias may be evident in the RVT because the order in which species are noticed is critical. Since there were no significant differences between the number of species recorded in each STT census and in the corresponding RVT census, the number of censuses required to assess species composition was equal for the two techniques. The correlation between averaged RVT and STT censuses was statistically significant. Both the RVT and the STT are valid visual census techniques. When quantitative estimates of abundance are desired, the STT must be used. When available field time is minimal and the main objective is to qualitatively characterize fish assemblages so that sites can be distinguished and compared on that basis, the RVT is more cost-effective.