Season, prey availability, sex, and age explain prey size selection in a large solitary carnivore

Prey selection is a fundamental aspect of ecology that drives evolution and community structure, yet the impact of intraspecific variation on the selection for prey size remains largely unaccounted for in ecological theory. Here, we explored puma (Puma concolor) prey selection across six study sites in North and South America. Our results highlighted the strong influence of season and prey availability on puma prey selection, and the smaller influence of puma age. Pumas in all sites selected smaller prey in warmer seasons following the ungulate birth pulse. Our top models included interaction terms between sex and age, suggesting that males more than females select larger prey as they age, which may reflect experiential learning. When accounting for variable sampling across pumas in our six sites, male and female pumas killed prey of equivalent size, even though males are larger than females, challenging assumptions about this species. Nevertheless, pumas in different study sites selected prey of different sizes, emphasizing that the optimal prey size for pumas is likely context-dependent and affected by prey availability. The mean prey weight across all sites averaged 1.18 times mean puma weight, which was less than predicted as the optimal prey size by energetics and ecological theory (optimal prey = 1.45 puma weight). Our results help refine our understanding of optimal prey for pumas and other solitary carnivores, as well as corroborate recent research emphasizing that carnivore prey selection is impacted not just by energetics but by the effects of diverse ecology.

Differential provisioning roles, prey size, and prey abundance shape the dynamic feeding behavior of gray wolves

The demands of raising dependent young can influence the feeding behaviors of social carnivores, especially for individuals that are primarily responsible for provisioning young. We investigated how the feeding and provisioning behavior of a social carnivore, gray wolves (Canis lupus), are connected and shaped by extrinsic and intrinsic factors, and whether and how these patterns changed throughout the pup-rearing season (April-August). We found breeding wolves had shorter handling times of prey, lower probability of returning to kills, and greater probability of returning to homesites after kills compared to subordinate individuals. However, the feeding and provisioning behaviors of breeding individuals changed considerably over the pup-rearing season. Wolves had longer handling times and returned to provision pups directly after kills less frequently as annual prey abundance decreased. These patterns indicate that adult wolves prioritize meeting their own energetic demands over those of their pups when prey abundance decreases. We suggest that differential provisioning of offspring based on prey abundance is a behavioral mechanism by which group size adjusts to available resources via changes in neonate survival.

Supervised versus unsupervised approaches to classification of accelerometry data

Sophisticated animal-borne sensor systems are increasingly providing novel insight into how animals behave and move. Despite their widespread use in ecology, the diversity and expanding quality and quantity of data they produce have created a need for robust analytical methods for biological interpretation. Machine learning tools are often used to meet this need. However, their relative effectiveness is not well known and, in the case of unsupervised tools, given that they do not use validation data, their accuracy can be difficult to assess. We evaluated the effectiveness of supervised (n = 6), semi-supervised (n = 1), and unsupervised (n = 2) approaches to analyzing accelerometry data collected from critically endangered California condors (Gymnogyps californianus). Unsupervised K-means and EM (expectation-maximization) clustering approaches performed poorly, with adequate classification accuracies of <0.8 but very low values for kappa statistics (range: -0.02 to 0.06). The semi-supervised nearest mean classifier was moderately effective at classification, with an overall classification accuracy of 0.61 but effective classification only of two of the four behavioral classes. Supervised random forest (RF) and k-nearest neighbor (kNN) machine learning models were most effective at classification across all behavior types, with overall accuracies >0.81. Kappa statistics were also highest for RF and kNN, in most cases substantially greater than for other modeling approaches. Unsupervised modeling, which is commonly used for the classification of a priori-defined behaviors in telemetry data, can provide useful information but likely is instead better suited to post hoc definition of generalized behavioral states. This work also shows the potential for substantial variation in classification accuracy among different machine learning approaches and among different metrics of accuracy. As such, when analyzing biotelemetry data, best practices appear to call for the evaluation of several machine learning techniques and several measures of accuracy for each dataset under consideration.

Large carnivores avoid humans while prioritizing prey acquisition in anthropogenic areas

Large carnivores are recovering in many landscapes where the human footprint is simultaneously growing. When carnivores encounter humans, the way they behave often changes, which may subsequently influence how they affect their prey. However, little research investigates the behavioural mechanisms underpinning carnivore response to humans. As a result, it is not clear how predator-prey interactions and their associated ecosystem processes will play out in the human-dominated areas into which carnivore populations are increasingly expanding. We hypothesized that humans would reduce predation risk for prey by disturbing carnivores or threatening their survival. Alternatively, or additionally, we hypothesized that humans would increase predation risk by providing forage resources that congregate herbivorous prey in predictable places and times. Using grey wolves Canis lupus in Jackson Hole, Wyoming, USA as a study species, we investigated 170 kill sites across a spectrum of human influences ranging from heavily restricted human activities on protected federal lands to largely unregulated activities on private lands. Then, we used conditional logistic regression to quantify how the probability of predation changed across varied types and amounts of human influences, while controlling for environmental characteristics and prey availability. Wolves primarily made kills in environmental terrain traps and where prey availability was high, but predation risk was significantly better explained with the inclusion of human influences than by environmental characteristics alone. Different human influences had different, and even converse, effects on the risk of wolf predation. For example, where prey were readily available, wolves preferentially killed animals far from motorized roads but close to unpaved trails. However, wolves responded less strongly to humans, if at all, where prey were scarce, suggesting they prioritized acquiring prey over avoiding human interactions. Overall, our work reveals that the effects of large carnivores on prey populations can vary considerably among different types of human influences, yet carnivores may not appreciably alter predatory behaviour in response to humans if prey are difficult to obtain. These results shed new light on the drivers of large carnivore behaviour in anthropogenic areas while improving understanding of predator-prey dynamics in and around the wildland-urban interface.

Standardizing protocols for determining the cause of mortality in wildlife studies

Mortality site investigations of telemetered wildlife are important for cause‐specific survival analyses and understanding underlying causes of observed population dynamics. Yet, eroding ecoliteracy and a lack of quality control in data collection can lead researchers to make incorrect conclusions, which may negatively impact management decisions for wildlife populations. We reviewed a random sample of 50 peer‐reviewed studies published between 2000 and 2019 on survival and cause‐specific mortality of ungulates monitored with telemetry devices. This concise review revealed extensive variation in reporting of field procedures, with many studies omitting critical information for the cause of mortality inference. Field protocols used to investigate mortality sites and ascertain the cause of mortality are often minimally described and frequently fail to address how investigators dealt with uncertainty. We outline a step‐by‐step procedure for mortality site investigations of telemetered ungulates, including evidence that should be documented in the field. Specifically, we highlight data that can be useful to differentiate predation from scavenging and more conclusively identify the predator species that killed the ungulate. We also outline how uncertainty in identifying the cause of mortality could be acknowledged and reported. We demonstrate the importance of rigorous protocols and prompt site investigations using data from our 5‐year study on survival and cause‐specific mortality of telemetered mule deer (Odocoileus hemionus) in northern California. Over the course of our study, we visited mortality sites of neonates (n = 91) and adults (n = 23) to ascertain the cause of mortality. Rapid site visitations significantly improved the successful identification of the cause of mortality and confidence levels for neonates. We discuss the need for rigorous and standardized protocols that include measures of confidence for mortality site investigations. We invite reviewers and journal editors to encourage authors to provide supportive information associated with the identification of causes of mortality, including uncertainty.

Kill rates and associated ecological factors for an apex predator

Kill rates and functional responses are fundamental to the study of predator ecology and the understanding of predatory-prey dynamics. As the most widely distributed apex predator in the western hemisphere, pumas (Puma concolor) have been well studied, yet a synthesis of their kill rates is currently lacking. We reviewed the literature and compiled data on sex- and age-specific kill rate estimates of pumas on ungulates, and conducted analyses aimed at understanding ecological factors explaining the observed spatial variation. Kill rate studies on pumas, while numerous, were primarily conducted in Temperate Conifer Forests (< 10% of puma range), revealing a dearth of knowledge across much of their range, especially from tropical and subtropical habitats. Across studies, kill rates in ungulates/week were highest for adult females with kitten(s) (1.24 ± 0.41 ungulates/week) but did not vary significantly between adult males (0.84 ± 0.18) and solitary adult females (0.99 ± 0.26). Kill rates in kg/day differed only marginally among reproductive classes. Kill rates of adult pumas increased with ungulate density, particularly for males. Ungulate species richness had a weak negative association with adult male kill rates. Neither scavenger richness, puma density, the proportion of non-ungulate prey in the diet, nor regional human population density had a significant effect on ungulate kill rates, but additional studies and standardization would provide further insights. Our results had a strong temperate-ecosystem bias highlighting the need for further research across the diverse biomes pumas occupy to fully interpret kill rates for the species. Data from more populations would also allow for multivariate analyses providing deeper inference into the ecological and behavioural factors driving kill rates and functional responses of pumas, and apex predators in general.

High frequency GPS bursts and path-level analysis reveal linear feature tracking by red foxes

There is a need to quantify and better understand how wildlife interact with linear features, as these are integral elements of most landscapes. One potentially important aspect is linear feature tracking (LFT), yet studies rarely succeed in directly revealing or quantifying this behavior. In a proof-of-concept study, we employed short-term intensive GPS monitoring of red foxes (Vulpes vulpes) in a multiple-use landscape in southern Norway. Using periodic bursts of high frequency GPS position fixes, we performed modified path selection analyses to estimate the propensity of foxes to track natural and man-made linear features (roads, forest edges, and streams) once they are encountered. Foxes in our study tracked primarily forest edges and roads. Forty-three percent of bursts that encountered any linear feature resulted in LFT. LFT, although prominent, was manifested as a short-lived behavior, with overall median times to linear feature abandonment around two minutes. Movement speeds were highest along roads, perhaps due to greater ease of travel or higher perceived risk. In the highly heterogeneous habitats that characterize human-dominated landscapes, LFT may be manifested at such a fine spatio-temporal scale that it would remain hidden during telemetry studies employing conventional position fix frequencies. The approach described here may aid others studying spatial behaviors that are manifested over very short durations, yet are biologically significant.