- Research Article
- 10.1186/s40462-025-00621-x
Inferring behavioural states from tracking data with hidden Markov models - a validation study using GPS video-camera collars.
- Jan 24, 2026
- Movement ecology
- Benjamin Larue + 11 more +11
Publications from 2021 to 2026
Showing 10 of 45 papers
Inferring behavioural states from tracking data with hidden Markov models - a validation study using GPS video-camera collars.
Validation of grizzly bear hair hormone profiles as a tool to monitor population demographics
Abstract A common approach to monitoring bear populations is the estimation of population sizes through non‐invasive DNA sampling by means of hair snags. Although successful in acquiring population size estimates, genetic capture‐recapture‐based population surveys lack provision of the demographic parameters needed to inform a multi‐dimensional understanding of a population. We hypothesized that grizzly bear ( Ursus arctos ) hair hormone concentrations could be used to determine key demographic parameters, specifically whether animals were male or female, sub‐adult or adult, pregnant or non‐pregnant females, lactating or non‐lactating females, and females with or without cubs. We measured hormone profiles (16 steroid and thyroid hormones relating to reproduction, stress, and nutrition) from 130 grizzly bear hair samples collected during live‐capture events in Alberta, Canada, from 2008‐2019, for which we were able to determine sex, age class, and several measures of reproductive status in females (pregnancy, lactation, and presence of cubs). We used random forest models to predict demographic parameters based on different combinations of hormone values, with a predictive accuracy ranging from 53% to 94%. Our best performing models were those developed to predict sex (80% accuracy when applied to all bears and 94% accuracy when subset to adult bears). Age class models performed better on male bears (86% accuracy) than on female bears (73% accuracy). Our analyses of female reproductive status were constrained by sample size limitations but resulted in a predictive model able to determine whether an adult female was accompanied by cubs of the year with 66% accuracy. By testing this methodology on hair samples collected from captured grizzly bears where these demographic parameters were known, we demonstrate its potential applicability to non‐invasive monitoring approaches. Pairing hair hormone concentration analysis with genetic capture‐recapture surveys has the potential to provide multi‐dimensional population data to wildlife managers, better informing evidence‐based decisions.
Read moreCollaborative strategies for wildlife health: case studies from the Canadian North
Abstract Wildlife health and conservation are increasingly recognized as key to improving human, animal, and environmental health (One Health) and detecting and addressing threats such as altered distribution and transmission of zoonotic diseases due to climate change. Wildlife, and therefore wildlife management, are crucial for the livelihood and well‐being of people in the Canadian North, a vast geographic area with low population density and socio‐economic disparities, which can make widespread program implementation challenging. We analyzed and compared 4 case studies on collaborative programs on wildlife health management from 4 distinct jurisdictions in the Canadian North: the Yukon, chronic wasting disease surveillance in ungulates; Northwest Territories, rabies in arctic fox ( Vulpes lagopus ); Nunavut, foodborne diseases in harvested wildlife; and Nunavik, Québec, ringed seal ( Pusa hispida ) health and Trichinella in walrus ( Odobenus rosmarus ). The case studies differed in whether they focused on a specific pathogen, transmission route, or health of a specific wildlife population. Despite these differences, 3 main themes were common to all case studies: collaboration, infrastructure limitations, and adaptation. Collaboration promotes greater community buy‐in and investment as community members help shape the program and is key to long‐term success, sustainability, and local relevance. Laboratory infrastructure and human resource capacity are limited in most regions, highlighting the importance of collaboration among Indigenous‐managed boards and organizations, community members, government agencies at multiple levels, academic institutions, and the national wildlife health non‐governmental organization (Canadian Wildlife Health Cooperative). Examples of adaptations include capitalizing on wildlife harvest for fur and food to obtain samples for surveillance, developing regulations to train and permit veterinary paraprofessionals and community volunteers to facilitate co‐existence of pets and people with wildlife reservoirs of rabies, jointly developing regulations to prevent introduction of non‐endemic diseases, and adapting hunting locations based on the results of wildlife disease monitoring. Combining Indigenous knowledge and a One Health framework is fundamental to co‐managing zoonotic and food‐borne diseases in wildlife while respecting Indigenous ways of life. The case studies provide meaningful examples of collaboration and solutions to address complex problems at the One Health interface and to use a holistic and community‐based approach to facilitate the creation of regionally acceptable and culturally relevant programs, with better outcomes for wildlife, human, and environmental health.
Read more<scp>DNA</scp> metabarcoding and video camera collars yield different inferences about the summer diet of an arctic ungulate
Abstract The diets of wild ungulates are a foundational component of their ecology, influencing their behavior, body condition, and demography. With changing environmental conditions, there is a significant need to identify important forage items for ungulates, but this has often proved challenging. Declines in several barren‐ground caribou herds across the North American Arctic have raised concerns about the influence of climate change on caribou forage conditions. Shifts in plant phenology, biomass, quality, and composition may be influencing caribou diets and subsequently affecting their body condition and demographic rates. Although forage is a primary driver of barren‐ground caribou behavior and population dynamics, there is limited recent information about the specific foods they consume, and uncertainty about appropriate methods for identifying those foods. Investigators are increasingly using fecal DNA metabarcoding and video camera collars to assess ungulate diets, but comparative studies of these approaches are lacking. To examine the summer diets of barren‐ground caribou, we used both fecal metabarcoding and video camera collars to identify forage used by the Porcupine caribou herd, which spans the Alaska–Yukon border. In 2021, we sampled the diets of adult females by collecting fecal samples and observing collar videos during 4 sampling occasions. We found that caribou consumed very specific forage items, and those items varied markedly across the growing season. Caribou predominantly consumed graminoids and lichens during early summer, and shrubs and forbs later in the season. Metabarcoding and video data provided significantly different estimates of diet for all taxonomic levels we evaluated, and inferences from the two approaches were often disparate. Metabarcoding failed to detect some items frequently consumed in videos, such as lichens, and indicated high use of other items rarely consumed, such as mosses. We found that video data provided greater taxonomic diversity and resolution for vascular plants and lichens, and more closely aligned with past research and biological expectations than data from fecal metabarcoding. Additional research is needed to be able to use these methods to identify the biomass of different forage items consumed.
Read moreA digital twin-based integrated optimization strategy for cascade refrigeration systems to boost energy efficiency
Why does animal home range size decrease with population density?
Spatial confinement to a home range is theorized to be a more energetically efficient method of acquiring resources than random searching due to spatial memory. Intraspecific studies that have compared home range size at different population densities have found that home ranges shrink as population density increases. This negative trend could be due to increased conspecific competition via population density increase or due to correlations between resource density and population density. We use the 10-year population cycle of snowshoe hares (Lepus americanus) and individual-level food-add experiments as a case study to assess whether the mechanism of the relationship between home range size and population density is related to competition from increased conspecific density or confounds between population density and resource density. Over six winters (1 December-31 March) and a 50-fold change in population density, we estimated weekly home range sizes (n = 464; 90% minimum convex polygons) of 88 radio-collared hares, of which 26 were food-supplemented. We found a negative relationship between home range size and population density in controls; home ranges decreased by 2.5 ha as hare density increased from 0.24 to 1.2 hare/ha. Food-supplemented hares showed a more negative response to population density than controls (4.0 ± 0.56 ha decrease per 1 hare/ha increase). Our results suggest that the negative trend between home range size and population density is not due to confounds between population and resource density. Likely, there is a trade-off between resource acquisition and some other density-driven constraint when foraging at high densities, which we suggest is a reduction in resource sharing to minimize competition and maintain resource familiarity at high densities.
Read moreA Thirty-Year Precipitation Record at Wolf Creek Research Basin, Yukon and the Importance of Bias-Corrected, Sub-Daily Measurements in a Changing Climate
Precipitation (snow and rainfall) is an essential climate variable for hydrological modelling, flood forecasting, avalanche preparedness and assessing permafrost stability and ecological change. In data sparse regions, such as the Canadian Sub-Arctic, long-term sub-daily precipitation measurements are rare, yet imperative to understanding environmental feedback and the impact of extreme events. The Wolf Creek Research Basin (WCRB) in the southern Yukon, Canada, provides a unique long-term hydrological and climate record across forested, shrub and alpine ecozones. This study presents hourly precipitation recorded in WCRB since 1993 using a variety of instruments. The diversity in measurement techniques and range of monitoring elevations allows for thorough consideration of precipitation phase and lapse rate.We outline the challenges of maintaining and compiling in-situ, remote monitoring data spanning decades of change. This study facilitates discussion around best practices for cold-region precipitation data products by using transparent data filtering, correction and in-filling. We consider the efficacy and uncertainty of measurement techniques and bias correction methods for wind-induced losses at a site equipped with multiple concurrent instruments, shields and gauges. Our results explore spatiotemporal trends in the preliminary dataset and compare to available data in the southern Yukon. This work provides critical insights into the improvement and longevity of cold region, remote precipitation monitoring and the importance of long-term data sets in a changing climate.
Read moreThe “Centering Justice” Symposium: A Call to Climate and Environmental Justice Centers Within Higher Education
In January 2024, the Tishman Environment and Design Center (Tishman Center or TEDC) at The New School hosted the Centering Justice Symposium to explore how the rapid uptake of environmental and climate justice missions at university centers, together with increased federal and philanthropic funding opportunities, is shaping the relationship between higher education and the environmental justice (EJ) movement. The symposium addressed grounding principles, opportunities, and challenges that can accompany this emergent interest and explored ways to foster equitable partnerships with EJ communities. This article offers a summary of the event and reflects on some of the key aspects and outcomes of the symposium, including (1) investing in inclusive planning practices and agenda setting grounded in EJ principles, (2) relationship-building as a cornerstone of EJ-focused missions, (3) committing to changing the rules of higher education to center EJ movement needs, (4) institutionalizing and modeling practices that center justice, (5) creating a manifesto to disseminate and act on shared commitments, and (6) building a community of practice for co-learning and accountability.
Read moreAssessing indicators of harvest sustainability in northern populations of marten (Martes americana)
Descriptive epidemiology and phylogenetic analysis of highly pathogenic avian influenza H5N1 clade 2.3.4.4b in British Columbia (B.C.) and the Yukon, Canada, September 2022 to June 2023
ABSTRACT Surveillance data from wildlife and poultry was used to describe the spread of highly pathogenic avian influenza (HPAI) H5N1 clade 2.3.4.4b in British Columbia (B.C.) and the Yukon, Canada from September 2022 – June 2023 compared to the first “wave” of the outbreak in this region, which occurred April – August 2022, after the initial viral introduction. Although the number of HPAI-positive poultry farms and wildlife samples was greater in “Wave 2”, cases were more tightly clustered in southwestern B.C. and the most commonly affected species differed, likely due to an influx of overwintering waterfowl in the area. Eight HPAI genetic clusters, representing seven genotypes and two inter-continental viral incursions, were detected, with significant variation in the relative abundance of each cluster between the waves. Phylogenetic data suggests multiple spillover events from wild birds to poultry and mammals but could not rule out transmission among farms and among mammals.
Read more