ABSTRACT
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.
ABSTRACT
In Bimini, Bahamas, the consistent employment of longlines, beginning in 1982, provided a rare opportunity to explore population trends for large resident sharks. This study assessed three shallow water longline survey periods at this location; 1982-1989, 1992-2002 and 2003-2014, with the aim of determining trends in annual catch per unit effort (CPUE) for an IUCN listed near-threatened species, the lemon shark Negaprion brevirostris. A general additive model (GAM) was used to analyse the non-linear annual CPUE values over the entire 32-year research period. The GAM displayed high variability of annual CPUE, with a peak value of 0·026 N. brevirostris per hook day (hooks day(-1) ) in 2000. The temporal pattern of CPUE indicated an abundance trend with a complete cycle, from trough to trough, occurring over a period of approximately 18 years. The 1982-1989 survey period saw the highest proportion of mature individuals (19·8%) and the smallest average pre-caudal length (LPC ; 124·8 cm). The 1992-2002 survey period had the highest average annual CPUE (0·018 hooks day(-1) ), while the 2003-2014 research period saw largest average LPC size (134·8 cm) and the lowest average CPUE values (0·009 hooks day(-1) ) of the entire research period. The long-term trend identified in this study provides a baseline for future assessment.
Subject(s)
Sharks/physiology , Animal Migration , Animals , Atlantic Ocean , Bahamas , Ecosystem , Endangered Species , Population Density , Sexual Behavior, AnimalABSTRACT
A longline survey was conducted from 2004 to 2014 to investigate the demographic population structure and seasonal abundance of the blacktip shark Carcharhinus limbatus in the Bimini Islands, the Bahamas. All individuals sampled (n = 242) were sub-adult or adults [70·1-145·1 cm pre-caudal length (LPC) range] with no neonates or YOY recorded in Bimini. Carcharhinus limbatus abundance peaked in September, coincident with the largest ratio of female to male sharks and a peak in fresh mating wounds on females. Mitochondrial control region (mtCR) DNA sequences were obtained from C. limbatus at Bimini to test whether Bimini C. limbatus are most closely related to geographically proximate populations sampled on the south-eastern coast of the U.S.A., the closest known nursery areas for this species. Nine mtCR haplotypes were observed in 32 individuals sampled at Bimini [haplotype diversity (h) = 0·821, nucleotide diversity (π) = 0·0015]. Four haplotypes observed from Bimini matched those previously found in the northern Yucatan (Mexico)-Belize and two matched a haplotype previously found in the U.S.A. Four haplotypes were novel but were closely related to the northern Yucatan-Belizean haplotypes. Pair-wise ΦST analysis showed that Bimini was significantly differentiated from all of the populations previously sampled (U.S.A. Atlantic, U.S.A. Gulf of Mexico, northern Yucatan, Belize and Brazil). This indicates that C. limbatus sampled from Bimini are unlikely from the described, proximate U.S.A. nurseries.
