Impact of observers experience level on counts of fishes in underwater visual surveys

Underwater visual surveys are widely used to estimate reef fish abundance. Using data from repeated surveys of coral reef sites on West Hawai'i, we assessed one possible factor driving variation among and within observers' counts, namely the level of observers' previous survey expe- rience. Increasing experience was associated with higher estimates of species richness and of abun- dance for some but not all taxa, but no systematic change in intra-observer variability. Least experi- enced observers counted 66 to 71% as many Acanthuridae, Labridae, and benthic Pomacentridae as the most experienced observers, and only 44% as many Holocentridae, those taxa being numerically dominated by small, mobile, or cryptic fishes. In contrast, there was no effect of experience level on counts of more conspicuous or slow-moving taxa: Chaetodontidae, Chromis, and Cirrhitidae. Experi- ence bias is unlikely to be a major problem for survey programs that pool data from multiple observers with broadly similar levels of experience, other than that all sources of difference among observers contribute to increased data variability and consequently lowered statistical power and precision. However, experience bias clearly could be a serious issue for survey or monitoring pro- grams that begin with largely inexperienced survey divers who gain experience and, consequently, tend to count more fish as the program matures. While experience level is not the only factor causing differences among observers, it is relatively easy to recognize and therefore to mitigate.