Effective conservation of animal species depends on accurate identification of their critical habitat. Marine mammals, however, often transit through heterogeneous habitats and perform various activities within short periods of time. The predictive power of habitat modelling techniques can thus suffer from variability in behaviour and the use of multiple habitat types. We used data loggers and ecological-niche factor analysis (ENFA) modelling techniques to determine blue whale Balaenoptera musculus associations with underwater topography on a feeding ground in the St. Lawrence River estuary, Canada. We compared a naïve model that had no knowledge of behaviour with a model that used the locations of feeding events inferred from specific velocity signatures. Blue whales travelled over several habitat types with different characteristics, which confounded modelling efforts when pooled together. The model based on the feeding set had considerably higher predictive power but could not highlight all suitable habitats at the same time. Using cluster analysis, we identified 4 habitat types used for feeding, each corresponding to distinct underwater topographies. Feeding depth and behaviour varied across these habitats, which were used preferentially at different times of the tidal cycle and appeared linked to known prey aggregation mechanisms. Our results suggest that failure to identify feeding activity or to take into account the existence of multiple foraging habitats at a fine scale could result in spurious modelling results.KEY WORDS: Habitat modelling · Feeding behaviour · Habitat suitability · ENFA · Blue whale · St. Lawrence River · Marine mammals 17: 255-268, 2012 rectly via a submersible video camera attached to animals, but this technology is greatly limited by turbidity and darkness (Marshall 1998). An alternative method has been to use speed or acceleration data to investigate feeding behaviour in large whales (Goldbogen et al. 2006, Simon et al. 2009). Blue whales Balaenoptera musculus, for instance, perform feeding lunges characterised by rapid speed changes (Goldbogen et al. 2011), which can be used to pinpoint the exact moment of each feeding attempt (Doniol-Valcroze et al. 2011).
Resale or republication not permitted without written consent of the publisher
Contribution to the Theme Section 'Beyond marine mammal habitat modeling'
OPEN PEN ACCESS CCESSEndang Species ResAlthough little is known of their winter distribution and migration routes (Reeves et al. 2004), North Atlantic blue whales are found during summer on a range of high-latitude feeding grounds from eastern Canada to Greenland (Sears & Calambokidis 2002). Whaling records (Mitchell 1974) and field studies (Sears et al. 1990, Kingsley & Reeves 1998 indicate that portions of the Gulf of St. Lawrence are important seasonal feeding sites for blue whales from late May to December. They are found as far upstream as the St. Lawrence Estuary (SLE), where the largest concentrations have been recorded in August and September (Edds & Macfa...