Search

Emerging technologies support a new era of applied wildlife research, generating data on scales from individuals to populations. Computer vision methods can process large datasets generated through image‐based techniques by automating the detection and identification of species and individuals. With the exception of primates, however, there are no objective visual methods of individual identification for species that lack unique and consistent body markings. We apply deep learning approaches of facial recognition using object detection, landmark detection, a similarity comparison network, and an support vector machine‐based classifier to identify individuals in a representative species, the brown bear Ursus arctos. Our open‐source application, BearID, detects a bear's face in an image, rotates and extracts the face, creates an "embedding" for the face, and uses the embedding to classify the individual. We trained and tested the application using labeled images of 132 known individuals collected from British Columbia, Canada, and Alaska, USA. Based on 4,674 images, with an 80/20% split for training and testing, respectively, we achieved a facial detection (ability to find a face) average precision of 0.98 and an individual classification (ability to identify the individual) accuracy of 83.9%. BearID and its annotated source code provide a replicable methodology for applying deep learning methods of facial recognition applicable to many other species that lack distinguishing markings. Further analyses of performance should focus on the influence of certain parameters on recognition accuracy, such as age and body size. Combining BearID with camera trapping could facilitate fine‐scale behavioral research such as individual spatiotemporal activity patterns, and a cost‐effective method of population monitoring through mark–recapture studies, with implications for species and landscape conservation and management. Applications to practical conservation include identifying problem individuals in human–wildlife conflicts, and evaluating the intrapopulation variation in efficacy of conservation strategies, such as wildlife crossings. [ABSTRACT FROM AUTHOR]

Snow is a major, climate-sensitive feature of the Earth's surface and catalyst of fundamentally important ecosystem processes. Understanding how snow influences sentinel species in rapidly changing mountain ecosystems is particularly critical. Whereas effects of snow on food availability, energy expenditure, and predation are well documented, we report how avalanches exert major impacts on an ecologically significant mountain ungulate - the coastal Alaskan mountain goat (Oreamnos americanus). Using long-term GPS data and field observations across four populations (421 individuals over 17 years), we show that avalanches caused 23-65% of all mortality, depending on area. Deaths varied seasonally and were directly linked to spatial movement patterns and avalanche terrain use. Population-level avalanche mortality, 61% of which comprised reproductively important prime-aged individuals, averaged 8% annually and exceeded 22% when avalanche conditions were severe. Our findings reveal a widespread but previously undescribed pathway by which snow can elicit major population-level impacts and shape demographic characteristics of slow-growing populations of mountain-adapted animals., (© 2024. The Author(s).)

Implementation of ecosystem-based fisheries management (EBFM) requires a clear conceptual and quantitative framework for assessing how different harvest options can modify benefits to ecosystem and human beneficiaries. We address this social-ecological need for Pacific salmon fisheries, which are economically valuable but intercept much of the annual pulse of nutrient subsidies that salmon provide to terrestrial and aquatic food webs. We used grizzly bears, vectors of salmon nutrients and animals with densities strongly coupled to salmon abundance, as surrogates for "salmon ecosystem" function. Combining salmon biomass and stock-recruitment data with stable isotope analysis, we assess potential tradeoffs between fishery yields and bear population densities for six sockeye salmon stocks in Bristol Bay, Alaska, and British Columbia (BC), Canada. For the coastal stocks, we find that both bear densities and fishery yields would increase substantially if ecosystem allocations of salmon increase from currently applied lower to upper goals and beyond. This aligning of benefits comes at a potential cost, however, with the possibility of forgoing harvests in low productivity years. In contrast, we detect acute tradeoffs between bear densities and fishery yields in interior stocks within the Fraser River, BC, where biomass from other salmon species is low. There, increasing salmon allocations to ecosystems would benefit threatened bear populations at the cost of reduced long-term yields. To resolve this conflict, we propose an EBFM goal that values fisheries and bears (and by extension, the ecosystem) equally. At such targets, ecosystem benefits are unexpectedly large compared with losses in fishery yields. To explore other management options, we generate tradeoff curves that provide stock-specific accounting of the expected loss to fishers and gain to bears as more salmon escape the fishery. Our approach, modified to suit multiple scenarios, provides a generalizable method to resolve conflicts over shared resources in other systems., Competing Interests: The authors have declared that no competing interests exist.