Acoustic documentation of temperate odontocetes in the Bering and Chukchi Seas

Warming in the Arctic region is three times the rate of the global average with summer sea ice declining 11.5% per decade since 1979 (Comiso and Hall 2014). This drives ice‐obligate and ice‐associated marine mammal species northward and opens space for temperate species to also shift poleward. Larger and more rapid shifts are to be expected, especially if the Arctic is predicted to be ice free in the summers by the 2030s (Kwok et al. 2009, Wang and Overland 2012). Previous research has shown that several marine mammal species that spend some or all of their life cycles in the Arctic (bowhead, gray, and beluga whales; and bearded and ribbon seals) adjust their distributions, mating, and migrating behaviors concurrently with ice cover changes, such as ice retreating midwinter compared to being continuously present (Grebmeier and Dutton 2000, Miksis‐Olds et al. 2013, Miksis‐Olds and Madden 2014, Hauser et al. 2016). Scientific names of these species include Balaena mysticetus (Linnaeus, 1758), Eschrichtius robustus (Lilljeborg, 1861), Delphinapterus leucas (Pallas, 1776), E. b. nauticus (Pallas, 1881), and Histriophoca fasciata (Zimmerman, 1783), respectively. Passive acoustic monitoring (PAM) was key in these studies because ship‐based visual surveys are not possible during ice cover, and even airborne visual surveys are dangerous to carry out in inclement weather. Furthermore, both types of visual surveys are only possible during daylight. PAM is the only safe way to monitor species presence in the Arctic year‐round and overnight.

Until recently, PAM was constrained by power capacity and storage‐limited sampling rates (typically up to 44.1 kHz), leading acoustical studies in the Arctic to use intermittent recordings limited to only those species that vocalize below 22 kHz. For any species that vocalizes higher than 22 kHz, distribution and presence studies have been limited to visual surveys, stranding data, and whaling records. In this study, Passive Acoustic Listeners (PALs) were used because they can sample at 100 kHz with year‐round duty cycling for power saving capabilities in cold waters (Nystuen 1998). To allot power throughout an entire year, the PALs record 4.5 s WAV files every 2 or 10 min, depending on whether a target signal was or was not detected in the previous file, respectively (Denes et al. 2014). This leads to very low duty cycles (0.75% and 3.75%), so presence is likely undersampled. It should be noted that duty cycling inherently leads to missing information since the environment is not monitored continuously. In killer whales, for example, it has been shown that lower resolution data with a high duty cycle is most efficient (Riera et al. 2012). While high‐frequency clicks and buzzes are short enough to be fully captured in 4.5 s files, PAL files being spaced at least two minutes apart is not always sufficient in providing context to aid in species identification. Therefore, a cautious manual identification methodology was used so as to not overstate the results.

The first high‐frequency acoustic data set collected by PALs from the Bering Sea is now a decade long. Some high frequency vocalizations of odontocetes that typically inhabit only temperate waters (referred to as “typically temperate” species in this note) have been found in this data set. In this note, we document acoustic detections of three typically temperate odontocete species at two sites in the Bering Sea and one site in the Chukchi Sea. While not prolific due to the

-Abstract Truncated-