Marine Mammal Science最新文献

Acoustic monitoring is an essential tool for investigating animal communication and behavior when visual contact is limited, but the scalability of bioacoustic projects is often limited by time-intensive manual auditing of focal signals. To address this bottleneck, we introduce orcAI—a novel deep learning framework for the automated detection and classification of a broad acoustic repertoire of killer whales (Orcinus orca), including vocalizations (e.g., pulsed calls, whistles) and incidental sounds (e.g., breathing, tail slaps). orcAI combines a ResNet-based Convolutional Neural Network (ResNet-CNN) with Long Short-Term Memory (LSTM) layers to capture both spatial features and temporal context, enabling the model to classify signals and to accurately determine their temporal boundaries in spectrograms. Trained on a comprehensive dataset from herring-feeding killer whales off Iceland, the framework was designed to be adaptable to other populations upon training with equivalent data. Our final model achieves up to 98.2% accuracy on test data and is delivered as an open-source tool with an easy-to-use command-line interface. By providing a ready-to-use model that processes raw audio and outputs annotations, orcAI serves as a useful tool for advancing the study of killer whale vocal behavior and, more broadly, for understanding marine mammal communication and ecology.

Monitoring live and stranded cetaceans can be expensive and logistically challenging, resulting in knowledge gaps. Very high-resolution (VHR) optical satellites are considered a potential solution to addressing some of these gaps. Despite success at detecting live and stranded cetaceans, satellites have only been trialed on restricted spatiotemporal scales. This project presents a framework for assessing the feasibility of using VHR optical satellite-based monitoring of cetaceans at high temporal frequency and local to global scales, focusing on the UK and UK Overseas Territories as a case study. We assess the primary environmental conditions necessary for the successful application of this technology: cloud cover and wind speed. Five-year monthly median “Total cloud cover” and “10m wind speed” ERA5 global reanalysis data were analyzed to map the spatial feasibility of satellite monitoring. We found that for the United Kingdom, VHR optical satellites could complement existing monitoring methods to achieve greater spatial and temporal coverage of live cetacean surveys, particularly, offshore, during the boreal spring and summer. However, satellites cannot address gaps in UK live cetacean monitoring in winter due to high wind speeds reducing whale detection probability. Based on environmental conditions, the tropics hold the greatest promise for achieving year-round satellite-based cetacean monitoring. In the Falkland Islands, particularly, the remote, unpopulated coastlines of West Island, satellites have the potential to improve strandings monitoring, opportunistically complementing existing stranding monitoring efforts.

Baleen whales are among the largest animals ever and engineer marine ecosystems by transporting nutrients both vertically through the water column and across vast oceanic distances. Hidden underwater, their feeding habits often remain unseen and hence must be studied indirectly based on stomach contents and stable isotopes. Here, we explore the potential of bone-bound calcium isotopes as a chemical proxy for extant baleen whale foraging ecology. Unlike bulk nitrogen and various trace elements (Mg, Sr, and Ba), calcium isotopes clearly distinguish gulp (n = 4 species) from skim feeders (n = 4 species) and provide a powerful tool to infer diet and feeding strategies. The difference in calcium isotope composition plausibly reflects incidental ingestion of seawater, which may be more pronounced in skim feeding whales than previously thought.

Ringed seals (Pusa hispida) in Svalbard, Norway, are suspected to be declining due to a significant reduction in land-fast sea ice, which serves as an essential breeding habitat, but were last surveyed in 2002. We address this data gap by conducting Uncrewed Aerial Systems (UAS) aerial surveys throughout Isfjorden (including adjacent small fjords) on the island of Spitsbergen, in the Svalbard Archipelago, during the molting season in 2023. We also provide an assessment of the size/age structure of the ringed seal population within Isfjorden using body lengths measured from aerial images. We show a decrease of 77% in land-fast sea ice area and a concomitant decline in ringed seal abundance of approximately 46% since 2002. During the same period, the density of hauled-out seals increased from 1.86 seals/km² in 2002 to 2.41 seals/km² in 2023. Length-based estimates of age suggest that 65% of the contemporary population consists of immature individuals. Together, the results from this updated population survey underscore the need for more frequent monitoring of ringed seals in Svalbard to better understand the drivers of population decline and highlight the broader implications of environmental change on Arctic marine biodiversity. The innovative use of UAS technology in this study demonstrates its utility in wildlife research, offering a less invasive, more cost-effective, efficient, and safer method for surveying marine mammals in polar regions compared to crewed aircraft.

Studies of cetacean diving behavior in multiple locations in different ocean basins allow for an assessment of variability within and among populations. We examine foraging dive behaviors of goose-beaked whales (Ziphius cavirostris) and dense-beaked whales (Mesoplodon densirostris) using data from 132 tagged whales in seven locations in the Pacific and Atlantic Oceans and the Mediterranean Sea. Acoustic recording tags are used to identify foraging dives by the presence of echolocation. For other tag types, foraging dives are identified based on maximum dive depth. Five parameters are used to characterize foraging dives: maximum dive depth, foraging dive duration, dive cycle duration, and the mean and standard deviation of echolocation depths. We find that differences among dives within one tagged individual are typically larger than the differences among individuals or among locations, and that differences among individuals are typically similar in magnitude to differences among locations. Regression is used to estimate the mean and standard deviation of echolocation depths from maximum dive depth for dives without acoustic data. Composite values of foraging dive parameters (and standard deviations) are estimated as the average of all study locations.