Abundance estimation of blacklip abalone ( Haliotis rubra) I. An analysis of diver-survey methods used for large-scale monitoring

Abstract

A comparative evaluation of radial transect and timed collection methods used in large-scale monitoring of abalone stocks in Victoria was undertaken. Methods were assessed according to their accuracy, precision, robustness and sensitivity (ability to detect change). Sample size required to detect a 50% decrease was estimated using data from a pilot study, and three replicate stock depletion experiments were undertaken in a simple before/after design testing the null hypothesis of no change at the impacted sites. Harvesting was undertaken by commercial abalone divers. On all occasions, both methods were sensitive enough to detect the specified change and robust to both a significant effect of observer and violation of the assumption of equality of variances. No significant differences in precision occurred, however the 30 m 2 transect estimated the actual number of abalone harvested to within 10% of the true value. Only a weak linear relationship (pre-recruits r 2 = 0.12; post-recruits r 2 = 0.39) was observed between relative abundance (timed collection data) and absolute abundance (30 m 2 transect data). Importantly, the fishing behaviour (targeting of aggregations) of commercial abalone fishers resulted in much reduced variances for both transect estimates and timed collections, and consequently, increased power to detect a decrease in population size. In the long term however, this phenomenon may be offset by re-aggregation of populations, the extent of which needs to be throughly determined. Overall, the 30 m 2 transect was considered the safer, more robust method for long-term monitoring of abalone stocks.