Background: Recreational scuba diving is a popular activity of the coral reef tourism industry. In practice, local diving centers recommend interesting sites to help visiting divers make their plans. Fish are among the major attractions, but they need to be listed with care because the temporal occurrence of a fish species is difficult to predict. To address this issue, we propose methods to categorize each fish species based on its long-term occurrence and likelihood of being seen.
Methods: We assume that there are K categories of occurrence of a fish assemblage and propose two methods [an arithmetic-mean method (AM) and a geometric-mean method (GM)] to define the range of species in each category.
Results: Experiments based on long term datasets collected at three underwater stations (each having 51-53 surveys and totals of 262-284 fish species) on coral reefs in southern Taiwan showed that when K = 4 (rare, occasional, frequent and common categories), 11-14 species were concurrently assigned to the common category by AM for data sets based on surveys 10, 15, 20, 25, 30, 35, 40, 45, or 51-53 in contrast to the 18-26 species assigned as common by GM. If a similarity index of 0.7 (compared to the total pool of fish species) was the minimum threshold for diver satisfaction, then 20-25 surveys provide sufficient data for listing the common species at a given dive spot.
Conclusions: Common fish species, are the most temporally stable, and thus are more appropriate for attracting divers. These can be effectively differentiated by either AM or GM with at least 25 surveys. We suggest regular updating of each fish's category through periodic surveys to assure the accuracy of information at a particular dive spot.
Keywords: Coral reef; Dive tourism; Ecological service; Long-term data; Long-term monitoring.