Effects of training on observer performance in assessing benthic cover by means of the manta tow technique

Abstract

The manta tow technique is used to obtain broadscale information on the distribution of Acanthaster planci and corals on the Great Barrier Reef. A field experiment was used to investigate how experience and training affects visual estimates of benthic cover made by observers using the manta tow technique. For live coral and sandlrubble cover, experienced observers obtained significantly higher levels of precision in their estimates than inexperienced observers; training increased the precision of these estimates for both experienced and inexperienced observers. Dead coral cover was sparse (<5%), making it difficult to assess the performance of the manta tow technique for this form of benthic cover. For both live and dead coral, and for sandlrubble, the levels of agreement between observers exceeded that expected by chance by between 5 to 1 and 10 to 1.