Quantifying Features of Arctic Odontocete Echolocation and Marine Habitat Variation in West Greenland

Abstract

Monitoring Arctic cetacean habitat use enables identification of biologically important regions and predictions of climate change effects, both of which are critical to conservation efforts. The beluga (Delphinapterus leucas) and narwhal (Monodon monoceros) are the only toothed whale (odontocete) species that occupy Arctic waters year-round. As social animals, these species produce a variety of sounds to communicate with conspecifics and echolocate to forage and navigate. In recent decades, underwater recordings of their sounds have been collected to track seasonal movements and detect phenological shifts in response to sea ice declines. Yet, the vocal repertoires of belugas and narwhals are diverse, occur in similar frequency bands, and are only partially described, which can make it challenging to identify them in recordings. Furthermore, many questions remain surrounding what physical features of the marine environment explain why belugas and narwhals occupy certain areas. This thesis takes a multidisciplinary approach to investigate Arctic odontocete echolocation and their marine habitat in West Greenland with goals to improve sustained observation methods. First, beluga echolocation click parameters were calculated using data from a 16-channel hydrophone array. Results revealed belugas had a narrow sonar beam emitted at high intensities, which was concordant with previous work showing the highly directional beam of narwhals. Second, beluga and narwhal click parameters were compared and found to be differentiable where beluga echolocation was distinguished by higher frequency content. Following this comparison, machine learning models were developed to classify the two species’ clicks and the best model yielded highly accurate predictions (97.5% correct classification). Finally, this classifier was tested with a large acoustic dataset collected from seafloor-mounted moorings near glacier fronts in Melville Bay, Northwest Greenland. We present novel guidelines for visual identification of beluga and narwhal clicks using spectrograms from the moored recordings. Beluga clicks were characterized by frequencies >30 kHz whereas dominant narwhal click frequencies ranged from 20 to 30 kHz. Further, spectral energy in narwhal clicks quickly dropped below 20 kHz, providing a distinct lower frequency limit that was not visible in beluga clicks. These differences were reflected in one-third octave levels (TOL) calculated from click spectra. Narwhal clicks had a large increase between the 16 and 25 kHz TOL bands that was absent in beluga click spectra, and beluga clicks had a large increase between the 25 and 40 kHz TOL bands that was absent in narwhal click spectra. Compared to other variables, TOL ratios provided the strongest predictions to identify beluga and narwhal clicks. Our results provide compelling evidence to reliably identify beluga and narwhal clicks in long-term recordings and can be applied to future passive acoustic monitoring studies. Often passive acoustic recording platforms also contain multiple oceanographic sensors to collect synchronous data on ocean conditions. In the final chapter of this thesis, synchronous year-round oceanographic observations were examined from the same moorings introduced in the preceding acoustic analyses. Results showed consistent seasonal hydrographic variability across sites. Differences in upwelling from subglacial plume mixing were found between sites where deep glaciers with higher freshwater runoff produced larger plumes. Maximum temperatures were observed in spring, priming deep glaciers for increased submarine melt during summer. Physical oceanographic variability has important implications for biological processes, including primary production, prey availability, and predator communities. Future work will continue to relate physical and biological processes at the glacier-ocean boundary and evaluate impacts from a warming ocean on ecosystem structure. Altogether, results from this thesis fill key knowledge gaps in Arctic odontocete click classification and ocean variability along Greenland’s coast, enabling advancements in future passive acoustic and oceanographic observation procedures in the Arctic.