FIN-PRINT a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales

Biometric identification techniques such as photo-identification require an array of unique natural markings to identify individuals. From 1975 to present, Bigg’s killer whales have been photo-identified along the west coast of North America, resulting in one of the largest and longest-running cetacean photo-identification datasets. However, data maintenance and analysis are extremely time and resource consuming. This study transfers the procedure of killer whale image identification into a fully automated, multi-stage, deep learning framework, entitled FIN-PRINT. It is composed of multiple sequentially ordered sub-components. FIN-PRINT is trained and evaluated on a dataset collected over an 8-year period (2011–2018) in the coastal waters off western North America, including 121,000 human-annotated identification images of Bigg’s killer whales. At first, object detection is performed to identify unique killer whale markings, resulting in 94.4% recall, 94.1% precision, and 93.4% mean-average-precision (mAP). Second, all previously identified natural killer whale markings are extracted. The third step introduces a data enhancement mechanism by filtering between valid and invalid markings from previous processing levels, achieving 92.8% recall, 97.5%, precision, and 95.2% accuracy. The fourth and final step involves multi-class individual recognition. When evaluated on the network test set, it achieved an accuracy of 92.5% with 97.2% top-3 unweighted accuracy (TUA) for the 100 most commonly photo-identified killer whales. Additionally, the method achieved an accuracy of 84.5% and a TUA of 92.9% when applied to the entire 2018 image collection of the 100 most common killer whales. The source code of FIN-PRINT can be adapted to other species and will be publicly available.

Passive acoustic monitoring of killer whales (Orcinus orca) reveals year-round distribution and residency patterns in the Gulf of Alaska

Killer whales (Orcinus orca) are top predators throughout the world’s oceans. In the North Pacific, the species is divided into three ecotypes—resident (fish-eating), transient (mammal-eating), and offshore (largely shark-eating)—that are genetically and acoustically distinct and have unique roles in the marine ecosystem. In this study, we examined the year-round distribution of killer whales in the northern Gulf of Alaska from 2016 to 2020 using passive acoustic monitoring. We further described the daily acoustic residency patterns of three killer whale populations (southern Alaska residents, Gulf of Alaska transients, and AT1 transients) for one year of these data. Highest year-round acoustic presence occurred in Montague Strait, with strong seasonal patterns in Hinchinbrook Entrance and Resurrection Bay. Daily acoustic residency times for the southern Alaska residents paralleled seasonal distribution patterns. The majority of Gulf of Alaska transient detections occurred in Hinchinbrook Entrance in spring. The depleted AT1 transient killer whale population was most often identified in Montague Strait. Passive acoustic monitoring revealed that both resident and transient killer whales used these areas much more extensively than previously known and provided novel insights into high use locations and times for each population. These results may be driven by seasonal foraging opportunities and social factors and have management implications for this species.

Southern resident killer whales encounter higher prey densities than northern resident killer whales during summer

The decline of southern resident killer whales (Orcinus orca) may be due to a shortage of prey, but there is little data to test this hypothesis. We compared the availability of prey (Chinook salmon, Oncorhynchus tshawytscha) sought by southern residents in Juan de Fuca Strait during summer with the abundance and distribution of Chinook available to the much larger and growing population of northern resident killer whales feeding in Johnstone Strait. We used ship-based multifrequency echosounders to identify differences in prey fields that may explain the dynamics of these two killer whale populations. Contrary to expectations, we found that both killer whale habitats had patchy distributions of prey that did not differ in their frequencies of occurrence, nor in the size compositions of individual fish. However, the density of fish within each patch was 4–6 times higher in the southern resident killer whale habitat. These findings do not support the hypothesis that southern resident killer whales are experiencing a prey shortage in the Salish Sea during summer and suggest a combination of other factors is affecting overall foraging success.