ABSTRACT Upper trophic‐level predators are often a target of conservation and effective protection, and population recovery requires limiting exploitation. Maintaining prey abundance may also be critical, but we typically have little information on prey needs and how this driver compares in importance to other factors. This lack of data is true for most reef shark species, despite widespread conservation efforts. Here we consider how prey biomass affects the probability of seeing a Caribbean reef shark ( Carcharhinus perezi ) on reefs in The Bahamas, where establishment of a shark sanctuary in 2011 allows quantification of relatively natural spatial variability in abundance. Shark presence or absence was quantified from 631 baited remote underwater video station (BRUVS) deployments across eight island groups. We then built a model of reef fish prey biomass from 85 underwater visual surveys and 23 biophysical covariates. This model allowed predictions of prey biomass in the likely foraging areas of C. perezi around each BRUVS. A generalized linear model of C. perezi presence or absence revealed that sightings were more frequent close to reef escarpments and with increasing depth, which is consistent with known habitat use. The model also included significant terms for both mean per 4 ha and total prey biomass within foraging areas as covariates. C. perezi individuals were more commonly sighted where mean prey biomass per 4 ha was higher but, surprisingly, sightings were negatively correlated with total prey abundance (meaning sharks preferred smaller reef patches with high prey biomass). It is unclear why larger reef patches with higher total prey biomass were less used, but potential explanations include seascape context and the presence of larger‐bodied sharks that may prey on C. perezi . This study highlights the need to consider managing prey abundance for effective shark conservation, particularly ensuring adequate three‐dimensional reef structure upon which prey fish rely.
Harborne et al. (Wed,) studied this question.