Date of Award

Winter 12-27-2018

Level of Access

Open-Access Thesis

Degree Name

Master of Science (MS)

Department

Marine Biology

Advisor

Gayle Zydlewski

Second Committee Member

Emmanual Boss

Third Committee Member

Mark Baumgartner

Additional Committee Members

Danielle Cholewiak

Abstract

Passive acoustic monitoring, the recording and analysis of biological sound, is a standard method of research into the distribution and behavior of cetaceans worldwide. Acoustic monitoring is reliant upon a thorough reference catalog of species vocalizations and an understanding of the temporal and geographic parameters in which vocalizations occur. This study combined a standard cetacean passive acoustic monitoring survey with a concurrent visual survey at a known baleen whale summer feeding ground to determine the annual species and vocal composition, compare species detection rates using each method, and identify and attribute novel vocalizations to species. The survey took place at the Mount Desert Rock marine research station, Gulf of Maine, from July to October 2015 and 2016. Visual and acoustic detections of fin (Balaenoptera physalus), minke (Balaenoptera acutorostrata), humpback (Megaptera Noavaeangliae), right (Eubalaena glacialis), and sei (Balaenoptera borealis) whales were assessed. Minke whales were the most frequently seen species (71% of days), followed by fin (51%), humpback (40%), and right (4%). Visual detections of minke and fin whales were more restricted by distance than humpbacks, and fin and humpbacks were sighted significantly more frequently in the northeast quadrant of the survey space, suggesting each species may have fine scale spatial preferences within the survey space. Stereotyped vocalizations of fin, minke, right, and sei whales were recorded, and automatic template detectors were used to evaluate daily and annual occurrence. Fin whales were the most frequently detected (31% of days), followed by minke (22%), right (15%), and sei (2%). Fin and minke whales vocalized primarily in evening and night-time hours, and right whales vocalized primarily during the day. Instances of multi-hour vocalization bouts were also recorded for fin, minke, and right whales. The number of days where each species was detected both visually and acoustically was low, meaning the use of only one method or the other may have resulted in a less precise census. Five novel vocalizations were identified in the acoustic dataset that occurred on days when only fin or minke whales were visually detected. These were 74 to 34-Hz suspected fin whale downsweeps, 268 to 448-Hz short upsweeps, 415-Hz “whip-cracks”, 498 to 363-Hz long downsweeps, and 257 to 164-Hz long downsweeps. The 74 to 34-Hz downsweep was found to be statistically associated with fin whales (p = 0.040, Φ = 0.129), while all other statistical tests were inconclusive. Spatial comparisons between all visually detected fin and minke whales and all locations of novel vocalizations from the study period showed very little spatial overlap, a potential complication to the statistical results. Potential violations of statistical assumptions included non-vocalizing whales, whales vocalizing outside of the observation range, and vocalizations that occurred at night. The findings of this study suggest that while visual and acoustic surveys are both effective stand-alone techniques for detecting the presence and behavior of baleen whales, a more precise census is achieved when the two are combined. Concurrent visual and acoustic surveys are also an effective method for the identification and attribution of novel baleen whale vocalizations, though species composition, detection probability, and vocalization behavior must be accounted for.

Files over 10MB may be slow to open. For best results, right-click and select "save as..."

Included in

Life Sciences Commons

Share