Estimates of species richness have been used to assess how community diversity changes across space and time. Here, we elucidate the bias that can arise when applying a popular sampling method (i.e., fixed number of belt-transects) to census coral community richness when the size of the reef and the regional species pools vary. Based on surveys of 148 patch reefs in 3 sub-regions of the Western Atlantic, we show that a fixed subsampling approach underestimated coral species richness of the reef as reef size increased; though this bias is minor for the entire region. The percentage of the true reef richness captured by sub-sampled transects declined by 6% for every 10-fold change in reef size. However, in Belize, the most speciose sub-region sampled, coral richness was underestimated by 15% solely as a result of a 10-fold increase in reef size. Increasing sampling effort (nr of transects) per reef was not able to correct for this underestimation. Furthermore, we demonstrate that this sampling bias is not an artifact of larger reefs simply being more diverse, and hence, requiring greater sampling efforts. Rather, these results suggest that coral species in diverse regions are distributed in accordance with the variety of spatially structured microhabitats present on a reef, rather than distributing randomly across the reef surface. Furthermore, these patterns seem to be highly scale-dependent. As such, sampling protocols should consider the size of the reef to be surveyed as well as the regional species pool to ensure accurate estimates of coral diversity.
ASJC Scopus subject areas
- Ocean Engineering