Comparing naive and experienced observers estimations of fish size and abundance using DOVS footage

Underwater Visual Samples are used widely in marine ecology to determine an accurate population and biomass of reef fishes, however different observers are thought to produce significantly different results. This study looks at the ex situ processing of Diver Operated Video Stereo (DOVS) samples of reef fish within Marine National Parks, along Victoria's coast, by observers of varying skill levels. Three observers rated experienced, intermediate and naive processed at least 11 sites. It was hypothesised that naive observers would be unsuitable to process DOVS footage at the level of the experienced observers. The observers' results were analysed in several ways to determine any difference in species detection, abundance, time taken, as well as looking for biases and differences in precision in estimates of lengths of fish. It was found that inexperienced observers may be used to process DOVS footage, as the differences in results proved to be minimal. There was no statistical difference between observers for the abundance of common species and simple measures of species diversity across the sites. The time taken by the naive observer to process the sites also reduced, without compromising accuracy, as processing progressed and experience was gained. It was found there was no statistical difference for estimated lengths of seven common species by observers, except for Notolabrus tetricus. Notolabrus fucicola was the only instance of statistically significant differences in the precision of measurements. The naive observer was also shown to achieve a higher overall abundance and measurement count compared to the experienced observers and any sampling errors would be expected to moderate with post sampling debriefing and training