Migratory birds modulate niche tradeoffs in rhythm with seasons and life history.

Movement is a key means by which animals cope with variable environments. As they move, animals construct individual niches composed of the environmental conditions they experience. Niche axes may vary over time and covary with one another as animals make tradeoffs between competing needs. Seasonal migration is expected to produce substantial niche variation as animals move to keep pace with major life history phases and fluctuations in environmental conditions. Here, we apply a time-ordered principal component analysis to examine dynamic niche variance and covariance across the annual cycle for four species of migratory crane: common crane (Grus grus, n = 20), demoiselle crane (Anthropoides virgo, n = 66), black-necked crane (Grus nigricollis, n = 9), and white-naped crane (Grus vipio, n = 9). We consider four key niche components known to be important to aspects of crane natural history: enhanced vegetation index (resources availability), temperature (thermoregulation), crop proportion (preferred foraging habitat), and proximity to water (predator avoidance). All species showed a primary seasonal niche "rhythm" that dominated variance in niche components across the annual cycle. Secondary rhythms were linked to major species-specific life history phases (migration, breeding, and nonbreeding) as well as seasonal environmental patterns. Furthermore, we found that cranes' experiences of the environment emerge from time-dynamic tradeoffs among niche components. We suggest that our approach to estimating the environmental niche as a multidimensional and time-dynamical system of tradeoffs improves mechanistic understanding of organism-environment interactions.

Artificial night light helps account for observer bias in citizen science monitoring of an expanding large mammal population.

The integration of citizen scientists into ecological research is transforming how, where, and when data are collected, and expanding the potential scales of ecological studies. Citizen-science projects can provide numerous benefits for participants while educating and connecting professionals with lay audiences, potentially increasing the acceptance of conservation and management actions. However, for all the benefits, collection of citizen-science data is often biased towards areas that are easily accessible (e.g. developments and roadways), and thus data are usually affected by issues typical of opportunistic surveys (e.g. uneven sampling effort). These areas are usually illuminated by artificial light at night (ALAN), a dynamic sensory stimulus that alters the perceptual world for both humans and wildlife. Our goal was to test whether satellite-based measures of ALAN could improve our understanding of the detection process of citizen-scientist-reported sightings of a large mammal. We collected observations of American black bears Ursus americanus (n = 1,315) outside their primary range in Minnesota, USA, as part of a study to gauge population expansion. Participants from the public provided sighting locations of bears on a website. We used an occupancy modelling framework to determine how well ALAN accounted for observer metrics compared to other commonly used metrics (e.g. housing density). Citizen scientists reported 17% of bear sightings were under artificially lit conditions and monthly ALAN estimates did the best job accounting for spatial bias in detection of all observations, based on AIC values and effect sizes ( ß^  = 0.81, 0.71-0.90 95% CI). Bear detection increased with elevated illuminance; relative abundance was positively associated with natural cover, proximity to primary bear range and lower road density. Although the highest counts of bear sightings occurred in the highly illuminated suburbs of the Minneapolis-St. Paul metropolitan region, we estimated substantially higher bear abundance in another region with plentiful natural cover and low ALAN (up to ~375% increased predicted relative abundance) where observations were sparse. We demonstrate the importance of considering ALAN radiance when analysing citizen-scientist-collected data, and we highlight the ways that ALAN data provide a dynamic snapshot of human activity.

Camera trapping expands the view into global biodiversity and its change.

Growing threats to biodiversity demand timely, detailed information on species occurrence, diversity and abundance at large scales. Camera traps (CTs), combined with computer vision models, provide an efficient method to survey species of certain taxa with high spatio-temporal resolution. We test the potential of CTs to close biodiversity knowledge gaps by comparing CT records of terrestrial mammals and birds from the recently released Wildlife Insights platform to publicly available occurrences from many observation types in the Global Biodiversity Information Facility. In locations with CTs, we found they sampled a greater number of days (mean = 133 versus 57 days) and documented additional species (mean increase of 1% of expected mammals). For species with CT data, we found CTs provided novel documentation of their ranges (93% of mammals and 48% of birds). Countries with the largest boost in data coverage were in the historically underrepresented southern hemisphere. Although embargoes increase data providers' willingness to share data, they cause a lag in data availability. Our work shows that the continued collection and mobilization of CT data, especially when combined with data sharing that supports attribution and privacy, has the potential to offer a critical lens into biodiversity. This article is part of the theme issue 'Detecting and attributing the causes of biodiversity change: needs, gaps and solutions'.

Improving the accessibility and transferability of machine learning algorithms for identification of animals in camera trap images: MLWIC2.

Motion-activated wildlife cameras (or "camera traps") are frequently used to remotely and noninvasively observe animals. The vast number of images collected from camera trap projects has prompted some biologists to employ machine learning algorithms to automatically recognize species in these images, or at least filter-out images that do not contain animals. These approaches are often limited by model transferability, as a model trained to recognize species from one location might not work as well for the same species in different locations. Furthermore, these methods often require advanced computational skills, making them inaccessible to many biologists. We used 3 million camera trap images from 18 studies in 10 states across the United States of America to train two deep neural networks, one that recognizes 58 species, the "species model," and one that determines if an image is empty or if it contains an animal, the "empty-animal model." Our species model and empty-animal model had accuracies of 96.8% and 97.3%, respectively. Furthermore, the models performed well on some out-of-sample datasets, as the species model had 91% accuracy on species from Canada (accuracy range 36%-91% across all out-of-sample datasets) and the empty-animal model achieved an accuracy of 91%-94% on out-of-sample datasets from different continents. Our software addresses some of the limitations of using machine learning to classify images from camera traps. By including many species from several locations, our species model is potentially applicable to many camera trap studies in North America. We also found that our empty-animal model can facilitate removal of images without animals globally. We provide the trained models in an R package (MLWIC2: Machine Learning for Wildlife Image Classification in R), which contains Shiny Applications that allow scientists with minimal programming experience to use trained models and train new models in six neural network architectures with varying depths.