Discerning behaviours of free-ranging animals allows for quantification of their activity budget, providing important insight into ecology. Over recent years, accelerometers have been used to unveil the cryptic lives of animals. The increased ability of accelerometers to store large quantities of high resolution data has prompted a need for automated behavioural classification. We assessed the performance of several machine learning (ML) classifiers to discern five behaviours performed by accelerometer-equipped juvenile lemon sharks (Negaprion brevirostris) at Bimini, Bahamas (25°44′N, 79°16′W). The sharks were observed to exhibit chafing, burst swimming, headshaking, resting and swimming in a semi-captive environment and these observations were used to ground-truth data for ML training and testing. ML methods included logistic regression, an artificial neural network, two random forest models, a gradient boosting model and a voting ensemble (VE) model, which combined the predictions of all other (base) models to improve classifier performance. The macro-averaged F-measure, an indicator of classifier performance, showed that the VE model improved overall classification (F-measure 0.88) above the strongest base learner model, gradient boosting (0.86). To test whether the VE model provided biologically meaningful results when applied to accelerometer data obtained from wild sharks, we investigated headshaking behaviour, as a proxy for prey capture, in relation to the variables: time of day, tidal phase and season. All variables were significant in predicting prey capture, with predations most likely to occur during early evening and less frequently during the dry season and high tides. These findings support previous hypotheses from sporadic visual observations.Electronic supplementary materialThe online version of this article (10.1007/s00227-018-3318-y) contains supplementary material, which is available to authorized users.

No abstract

Shark populations have declined across the Caribbean region, with negative associations between shark abundance and human population density, open access to fishing, and proximity to large markets (‘market gravity’). This decline is frequently attributed to fishing mortality, which increases in areas closer to humans and outside marine reserves. Although it is difficult to disentangle the effects of fishing mortality from other anthropogenic pressures on sharks, comparing shark abundance and diversity in jurisdictions with near zero fishing mortality versus prevalent shark fishing can demonstrate the role of overfishing. We used baited remote underwater video systems to compare shark abundance and diversity on coral reefs in 2 Caribbean nations with contrasting levels of shark exploitation: Belize (shark fishing) and The Bahamas (shark sanctuary). The abundance of targeted shark species and diversity were significantly higher in The Bahamas than in Belize. Caribbean reef and nurse shark abundance in Belize were best predicted by fishing-related factors (marine reserves, market gravity, their interaction). In The Bahamas, abiotic factors (depth, sea surface temperature) best predicted nurse shark abundance, while depth, market gravity, and its interaction with marine reserves predicted Caribbean reef shark abundance. These results indicate that fishing mortality reduces shark abundance and diversity in Belize, while lower fishing mortality in The Bahamas has greatly reduced but not eliminated human impacts on sharks. Future work should elucidate the indirect effects of humans to develop holistic shark conservation plans. We suggest minimizing shark fishing through multinational management plans to improve shark abundance and diversity, especially on reefs near densely populated areas.