Distinct zonation of community assemblages among habitats is a ubiquitous feature of coral reefs. The distribution of roving herbivorous fishes (parrotfishes, surgeonfishes and rabbitfishes) is a particularly clear example, with the abundance of these fishes generally peaking in shallow-water, high-energy habitats, regardless of the biogeographic realm. Yet, our understanding of the factors which structure this habitat partitioning, especially with regards to different facets of structural complexity and nutritional resource availability, is limited. To address this issue, we used three-dimensional photogrammetry and structure-from-motion technologies to describe five components of structural complexity (rugosity, coral cover, verticality, refuge density and field-of-view) and nutritional resource availability (grazing surface area) among habitats and considered how these factors are related to herbivorous fish distributions. All complexity metrics (including coral cover) were highest on the slope and crest. Nutritional resource availability differed from this general pattern and peaked on the outerflat. Unexpectedly, when compared to the distribution of herbivorous fishes, none of the complexity metrics had a marked influence in the models. However, grazing surface area was a strong predictor of both the abundance and biomass of herbivorous fishes. The strong relationship between grazing surface area and herbivorous fish distributions indicates that nutritional resource availability may be one of the primary factors driving the distribution of roving herbivorous fishes. The lack of a relationship between complexity and herbivorous fishes, and a strong affinity of herbivorous fishes for low-complexity, algal turf-dominated outer-flat habitats, offers some cautious optimism that herbivory may be sustained on future, low-complexity, algal turf-dominated reef configurations.

GPS telemetry provides high-accuracy spatial data on animal movement; however, it has rarely been used with benthic organisms, such as stingrays, because of their irregular surfacing behaviour or bottom-dwelling habits. This study evaluated the use of towed-float GPS tags to assess movements of juvenile stingrays, with active tracking performed simultaneously for comparison. Four juvenile Urogymnus granulatus individuals (2 females and 2 males; average 32.2-cm disc width) were tracked in April 2016. Individuals travelled 1332.15±269.58m south-east across Pioneer Bay at an average speed of 6.87mmin–1 in 3.7h. Stationary tests demonstrated that the quality of the data obtained by towed-float GPS tags could not be matched by active, acoustic or ARGOS telemetry, reaching, on average, 99% of successful fixes and <15-m accuracy. Location error varied significantly based on the number of satellites detected, with error decreasing as the satellite number increased. This study demonstrated the potential of towed-float GPS telemetry for high-resolution assessment of movement patterns and habitat use of juvenile stingrays in shallow coastal water. If well applied, this technique can increase our knowledge of juvenile stingray ecology and their essential habitats.