How robust are estimates of coral reef shark depletion?

•Diver-based estimates of reef shark depletion have been controversial.•We evaluated the robustness of the estimates along a gradient of human interaction.•Diver surveys showed no evidence of bias across habitats or management zones.•Abundance estimates also were consistent with two alternative survey methods.•Findings support robustness of previous studies finding substantial shark depletion. On coral reefs, diver-surveys of shark abundance indicate that populations are severely depleted, even in no-take zones with low-levels of illegal fishing, but are protected by strictly enforced no-entry zones. These findings have been questioned, on the grounds that diver-surveys overestimate shark abundance. We evaluated whether divers encounter sharks at higher rates when they first enter the water, and whether these effects vary among reefs that are subject to different levels of human interaction due to management zoning. We also examined the consistency of abundance estimates derived from multiple survey methods. For timed-swim, towed-diver, and baited-remote-underwater-video (BRUV) surveys, encounter rates were constant over time. For audible-stationary-count (ASC) surveys, encounter rates were elevated initially, then decreased rapidly, but the extent of upward bias did not differ between management zones. Timed-swim, BRUV, and ASC surveys produced comparable estimates of shark density, however, towed-diver-surveys produced significantly lower estimates of shark density. Our findings provide no evidence for biases in diver-surveys: encounter rates with sharks were not elevated when divers first entered the water; behavioural responses of sharks were consistent across management zones; and diver-surveys yielded abundance estimates comparable to other stationary methods. Previous studies using underwater counts have concluded that sharks are vulnerable to low levels of illegal fishing in no-take management zones, and that additional measures are needed to protect species, which, like sharks, have demographic characteristics that make them vulnerable to low levels of exploitation. Our results support the robustness of the abundance estimates on which those conclusions have been based.