WAVEGUIDE PROPAGATION ALLOWS RANGE ESTIMATES FOR NORTH PACIFIC RIGHTWHALES IN THE BERING SEASean M. Wiggins1, Mark A. McDonald2, Lisa M. Munger1, Sue E. Moore3, John A. Hildebrand11Scripps Institution of Oceanography, 9500 Gilman Drive-0205, La Jolla, CA 92093-0205, [email protected] 2Whale Acoustics, 11430 Rist Canyon Road, Bellvue, CO 80512 3NOAA/National Marine Mammal Laboratory, 7600 Sand Point Way NE, Seattle, WA 98115 ABSTRACTThe shallow and uniform water depth of the eastern Bering Sea shelf results in an acoustic waveguide. Propagation within this waveguide produces waveform dispersion which is dependent upon range. We present a means for using dispersed waveforms to determine range to calling whales from a single autonomous acoustic recording instrument. The predominant North Pacific right whale (Eubalaena japonica) call is frequency upswept from about 90 Hz to around 160 Hz and lasts approximately 1 s. The regional bathymetry of the eastern Bering Sea middle shelf is relatively uniform and shallow (~ 70 meters deep). This geometry provides a plane-layered waveguide in which right whale upswept calls can be detected at ranges over 50 km and have multiple modal arrivals that become dispersed, displaying different propagation velocities for different frequencies. Dispersion characteristics of modal arrivals are dependent on the calling whale’s depth, the receiver’s depth, the water depth, the range from caller to receiver, and various environmental parameters including water and sediment density and sound velocity. A model of sound propagation for the eastern Bering Sea middle shelf is developed from right whale call dispersion recorded on sonobuoys and seafloor acoustic recording packages, using individual calls recorded at multiple instruments. After development of the model, waveform dispersion allows estimation of caller range based on single instrument recordings. Estimating range between instrument and calling whales provides a means to estimate minimum abundance for the endangered North Pacific right whale. RÉSUMÉL’eau peu profonde et uniforme de la rive Est de la mer de Béring produit un excellent guide d’ondes acoustiques. Dans ce guide de propagation, la dispersion des ondes sonores est dépendante de la distance. Nous présentons ici un moyen pour utiliser la dispersion des ondes sonores pour déterminer la portée de sons émis par des baleines à partir d’un unique instrument d’enregistrement du signal acoustique. La vocalisation prédominante de la baleine franche du Pacifique Nord (Eubalaena japonica) est une modulation ascendante d’environ 90 à 160 Hz et d’une durée approximative de 1 s. La bathymétrie régionale de la rive Est de la mer de Béring est relativement uniforme et peu profonde (~70 m de profondeur). Cette géométrie fournit un guide d’ondes à couches horizontales ou les vocalisations modulées de baleines franches peuvent être détectées à des distances supérieures à 50 km et ont de multiples arrivées modales qui deviennent dispersées, démontrant différente vitesse de propagation à différentes fréquences. Les caractéristiques de dispersion des arrivées modales sont dépendantes de la profondeur de la baleine, la profondeur du récepteur, la profondeur de l’eau, la distance de l’émetteur et du récepteur et une variété de paramètres environnementaux incluant la densité de l’eau et des sédiments, et la vitesse du son dans ces deux media. Un modèle de la propagation du son pour la rive Est de la mer deBéring est développé à partir de la dispersion des vocalisations des baleines franches enregistrées à partir de bouées acoustiques et de systèmes acoustiques ancrés sur le fond marin, en utilisant les vocalisations individuelles enregistrées à partir de multiples instruments. Après le développement du modèle, la dispersion de l’onde sonore permet l’estimation de la distance de la vocalisation basée sur l’enregistrement d’un seul instrument. Estimer la distance entre l’instrument et les vocalisations de baleines permet d’estimer l’abondance minimale de la baleine franche menacée d’extinction dans le Pacifique Nord. Canadian Acoustics / Acoustique canadienne Vol. 32 No. 2 (2004) - 146 Research article / Article de recherche147 - Vol. 32 No. 2 (2004) Canadian Acoustics / Acoustique canadienne1. INTRODUCTION The North Pacific right whale (Eubalaena japonica) is a critically endangered baleen whale. There is no reliable estimate for the eastern population, but it probably numbers less than 50 individuals (Clapham et al., 1999). Efforts to study these whales in the eastern Bering Sea have provided visual observations of them since 1996 (Fig. 1) (Goddard and Rugh, 1998); (Moore et al., 2000); (LeDuc et al., 2001); (Tynan et al., 2001). To complement these visual surveys, shipboard acoustic surveys have recorded North Pacific right whale calls in the eastern Bering Sea since 1999 (McDonald and Moore, 2002). In addition to providing the first descriptions of North Pacific right whale calls, the shipboard acoustic surveys provided the baseline acoustics needed to use long-term, autonomous acoustic recorders for passive monitoring of these endangered whales. Long-term autonomous acoustic recording provides a means for monitoring whale calling activity in poor weather conditions and during periods when ship-based visual and acoustic techniques are either impossible or cost prohibitive (Wiggins, 2003). By recording sound continuously for periods of more than one year, whale seasonal occurrence and minimum population estimates can be made. To do this requires an understanding of the relationship between calls recorded and total number of whales present within a given region. Knowledge of call detection range is critical. How far a call can be detected with an acoustic instrument depends on the characteristics of the call and the acoustic environment. Environmental noise from ships, storms or other calling whales may reduce the call detection range. In addition, acoustic propagation depends upon environmental factors such as water temperature profile and bathymetry. These factors can effectively enhance or decrease call detection range, and may distort call characteristics. Calls may be distorted by the environment in a range dependent way such that the distorted calls contain information about the caller’s location. For example, multi-path arrivals are common