Characterization of marine mammal recordings from the Hawaii Range Complex

Citation

Baumann-Pickering, S., L.K. Baldwin, A.E. Simonis, M.A. Roche, M.L. Melcon, J.A. Hildebrand, E.M. Oleson, R.W. Baird, G.S. Schorr, D.L. Webster, and D.J. McSweeney. 2010. Characterization of marine mammal recordings from the Hawaii Range Complex. Naval Postgraduate School NPS-OC-10-004CR.

Executive summary

This report summarizes work conducted in FY2009-FY2010 with Navy support to characterize marine mammal sounds related to passive acoustic monitoring in the Hawaii Range Complex (HRC). Existing acoustic data from the Hawaii Range Complex area were analyzed to provide better descriptions of acoustic signals by species. Recordings were either from a boat-based hydrophone during small boat-based surveys, or from an autonomous bottom-moored Highfrequency Acoustic Recording Package (HARP). Recordings were made of pygmy killer whales (Feresa attenuata) during four encounters, melon-headed whales (Peponocephala electra) during three encounters, Risso’s dolphins (Grampus griseus) during one encounter, and roughtoothed dolphins (Steno bredanensis) during one encounter. Echolocation click parameters were calculated for single species recordings during visual and acoustic surveys by boat-based hydrophones, as well as by using sightings from small boat surveys and locations of satellite tagged individuals in the vicinity of the HARP. False killer whales and short-finned pilot whales had the lowest peak frequencies (15-21 kHz) in comparison to the other species. Pygmy killer whale echolocation clicks showed a bimodal distribution of peak frequencies (in the range of 35 to 50 kHz or 75 to 100 kHz). Melon-headed whales had peak frequency in the range of 31 to 35 kHz. Risso’s dolphins showed a distinct peak/notch frequency structure in their echolocation clicks. (Peaks appear at 24.5, 26.7, 34.6 and 40.3 kHz.) Automatic classification of echolocation clicks of false killer whales and short-finned pilot whales was performed using a Gaussian mixture model. This method resulted in a mean misclassification of 10.7 ± 0.7%. Two unknown but distinct echolocation click types were observed in the HARP data. One was a high frequency click that had its minimum frequency around 70 kHz and extended beyond the frequency range of the recorder (100 kHz). The other click type was a low frequency click that had a distinct banding pattern with peak structure at 12.2, 16.4 and 23.8 kHz, close to the peaks seen for short-finned pilot whales.
An acoustic analyst manually screened the HARP data collected off the west coast of the island of Hawaii during the time period of February 10, 2009, until March 9, 2009. Distinct call types were found for: beaked whales with frequency modulated upsweep echolocation pulses (particularly those previously noted at Cross Seamount); sperm whales; high frequency clicks of unknown origin; low frequency banded echolocation clicks; and a large number of unidentified echolocation clicks. Odontocetes were acoustically active every day of the recording period. (65% of total hours had echolocation clicks.) Beaked whales were detected on 41% of the recording days, but only during short periods per day (4% of total hours). Events of anthropogenic noise were logged and categorized as ship noise or echosounder.

Download PDF

Associated projects