Sociality helps mitigate anthropogenic risks: evidence from elk crossing a major highway
Data files
Mar 03, 2025 version files 1.43 MB
Abstract
Phenotypic variation among individuals scales up when they associate with others, creating variation within and among groups that can shape group-level outcomes such as when and where groups move. While sociality is thought to be a fitness rewarding behavior, empirical evidence supporting how it influences individual behavior and the resulting fitness consequences (e.g., risk experienced) remains limited, especially in the context of human-modified landscapes.
Here, we use empirical observations to test whether sociality helps animals cross busy roads. Our data came from free-ranging elk in a population where > 75% of the adults were tracked, and in which group size and composition were highly variable. We combined field observations with GPS collar data to quantify four social phenotypes of individuals and groups: dominance (initiation of successful agonistic interactions), social connectedness (number and connectedness of social associates), social familiarity (frequency of association with group members in the past), and social stability (time since fusion with group members). We then investigated how these four social phenotypes influenced an individual's probability of crossing a major highway, and tested if particular social phenotypes made better road crossing decisions (i.e., crossed at lower traffic volume).
We found that who is in a group shapes the behavior of group members around anthropogenic risks. Individuals in groups that were more dominant, more connected, and to a lesser extent more familiar had a lower probability of crossing the highway. Individuals that had spent more time with group members had a higher probability of crossing the highway.
Importantly, our results suggest that sociality plays a role in safe movements around anthropogenic risks. Individuals in highly connected and familiar groups were less likely to cross the highway at high traffic volume.
Our work provides empirical evidence that sociality influences the movements of group-living individuals through anthropogenic disturbances, and helps individuals mitigate the risks associated with such disturbances. Developing a comprehensive understanding of animal sociality in human-modified landscapes is especially important as social behaviors are simultaneously threatened by human disturbances, which could be particularly detrimental for group-living species if those same behaviors help individuals mitigate risks.
https://doi.org/10.5061/dryad.wh70rxwzs
Description of the data and file structure
This dataset contains data on (i) agonistic interactions recorded between GPS-tracked, tagged elk in Yoho National Park (British Columbia, Canada) and on (ii) various metrics of sociality at the start of hourly travel steps by GPS-tracked elk in Yoho National Park.
Files and variables
File: PoulinEtAl2025_EloRatings.Rmd
Description: The following code allows to calculate Elo-ratings from agonistic interactions recorded between collared & tagged elk in Yoho National Park (British Columbia). The dataset, PoulinEtAl2025-interactionsBetweenIndividuals.csv, is described below.
File: PoulinEtAl2025-interactionsBetweenIndividuals.csv
Description: Agonistic interactions recorded between collared & tagged elk in Yoho National Park (British Columbia).
Variables
- field_nbr: Number identifying during which field season the interaction was recorded.
- field_season: Field season during which the interaction was recorded.
- obs_no: Unique identifier for group observations. Relates interactions to group size, group composition, and other info on the group.
- observer: Who recorded the interaction.
- date: Date when the interaction was recorded (DD-MM-YYYY).
- time_start: Time at which the interaction started.
- time_stop: Time at which the interaction ended.
- initiator: Tag number or age group of the elk that initiated the interaction.
- behavInit: Code indicating which behavior was displayed by the initiator [described below in Ethogram].
- recipient: Tag number or age group of the elk that received the interaction.
- behavRecip: Code indicating which behavior was displayed by the recipient [described below in Ethogram].
- type: Type of interaction (either vocalization, agonistic, or affiliative).
- winner: Tag number or age group of the elk that won the interaction.
- loser: Tag number or age group of the elk that lost the interaction.
- other: Additional notes on the interaction.
Ethogram (adapted from Franklin et al. 1974; Koutnik 1981; Weckerly 1999)
Aggressive behaviors
- Ear drop (ED): Low intensity threat in which the ears are pressed back alongside the neck with their orifices directed away from the neck; can be accompanied of short movements of the head.
- Head high threat (HH): An erect posture with head held high, forward and upward extension of the head and neck, nose tilted upward and ears laid back; sometimes done while standing or accompanied by several quick steps towards the recipient.
- Head low threat (HL): The head is lowered and neck extended with the ears laid back along the neck.
- Lunge (LG): Abruptly jerked its head forward towards the recipient and back without contact.
- Head raise (HR): The head, oriented toward the recipient, is quickly snapped up and backward, then brought back to a resting posture while the ears were held out horizontally; often used when a grazing animal is disturbed by a subordinate.
- Head shake (HS): Animal moves its head rapidly side to side and up and down.
- Head over (HO): Animal puts head and neck over the rump, shoulder, or neck of another animal.
- Charge (CHG): An individual runs rapidly at another animal with head erect or lowered and enlarged nostrils from a short distance but stop before making contact.
- Chase (CHS): Animal actively chasing another.
- Bite (BTE): Neck extended, ears laid back; hair is usually removed or bitten off, often directed at the back or hind quarters of the recipient.
- Front-leg kick (FLK): One foreleg is rapidly extended forward and back.
- Rake (RK): The foreleg is lifted forward about 0.5m above the ground with the ankle joint bent and the foot is scraped across the back of the bedded recipient.
- Poke (PK): The initiator gently touches the recipient with its snout; often contacts the rump of the recipient.
- Body push (BP): The initiator approaches the recipient from behind and pushes its front shoulder against the rear flank of the recipient while laying its throat on the back.
- Scissors kick (SK): Animal run up to or parallel with another and attempts to kick it with front legs.
- Mount (MT): Animal displays an apparent attempt to copulate.
- Rear up (RU): An individual rears up on its hind leg into a vertical position.
- Rear and flail (RF): Two individuals rear on hind leads and flail (i.e. move both forelegs alternately in a pedaling motion several times) at each other with front limbs.
Submissive behaviors
- Flight (F): Animal flees when charged or quickly moves away from antagonist.
- Displacement (DI): Recipient retreats and leaves; initiator takes its place.
- Orientation avoidance (OA): Animal slowly avoids the opponent by walking or changing orientation.
Affiliative behaviors
- Nursing (N)
- Suckling (SU)
- Licking (L)
- Mutual grooming (MG): Scratching, stroking other individuals.
Other behaviors
- Vocalizing (VOC)
References
Franklin, W.L., Mossmanm, A.S. & Dole, M. (1974). Social organization and home range of Roosevelt elk. J. Mammal., 56, 102–118.
Koutnik, D.L. (1981). Sex-Related Differences in the Seasonality of Agonistic Behavior in Mule Deer. J. Mammal., 62, 1–11.
Weckerly, F.W. (1999). Social bonding and aggression in female Roosevelt elk. Can. J. Zool., 77, 1379–1384.
File: PoulinEtAl2025_SupportingCode.Rmd
Description: The following code allows to run models to investigate the effect of sociality on the probability of crossing a major highway by elk in Yoho National Park (British Columbia). The dataset, PoulinEtAl2025-hourlyTravelSteps_wSocialPhenotypes.rds, is described below.
File: PoulinEtAl2025-hourlyTravelSteps_wSocialPhenotypes.rds
Description: This dataset includes all observed hourly steps classified as 'traveling' by the HMM, during which at least 2 collared elk were together (i.e., within 100-m of each other during simultaneous fixes within a 15-min interval). It also includes characteristics such as traffic volume recorded on the Trans-Canada highway during the step, season, and social phenotypes of the focal individual and of its group.
Variables
- elkID: Tag number of the focal individual (i.e., elk that performed the movement step).
- bioYear: Biological year, from the end of fall migration to the start of spring migration (i.e., period when elk were on their winter range or range near the road) // w2 = 2019-2020 and w3 = 2020-2021.
- elkID_bioYear: Unique combination of the focal individual and extended winter season.
- season: Identifies if the step was during Fall (August to November), Winter (December to February), or Spring (March to June).
- Timestamp_MT: Timestamp, in Mountain Time (.
- X and Y: X and Y coordinates in UTM zone 12N.
- hwyCros: Highway crossing // 0 = the step did not cross the Trans-Canada highway (TCH) and 1 = the step crossed the TCH.
- trafficVol: Hourly traffic volume on the TCH for the day and hour at which the step was observed.
- trafficVol100: Hourly traffic volume (in hundreds of cars per hour) for the day and hour at which the step was observed.
- nbrCollars: Number of collared elk in the group.
- propCollars: Proportion of all deployed collars that were in the group when the step was observed.
- groupSize: Predicted total group size, based on models of total group size from propCollars using field observations of groups.
- groupComp: Group composition as a vector of elkIDs.
- indivDominance: Mean elo-score of the focal individual, based on observations of agonistic interactions between pairs of collared elk.
- indivSocConnectedness: Scaled eigenvector centrality of the focal individual, based on proximity-based social networks of all collared elk for the 14 days prior to the step.
- indivSocFamiliarity: Median simple ratio index of dyadic association between focal individual and its group members, based on proximity-based social networks of all collared elk for the 14 days prior to the step.
- indivSocStability: Median number of hours since the fusion of dyads containing the focal individual and its group members.
- groupDominance: Mean elo-score of the most dominant individual in the group (i.e., maximum elo-score of group members), based on observations of agonistic interactions between pairs of collared elk.
- groupSocConnectedness: Median scaled eigenvector centrality of all group members, based on proximity-based social networks of all collared elk for the 14 days prior to the step.
- groupSocFamiliarity: Median simple ratio index of all dyadic associations between group members, based on proximity-based social networks of all collared elk for the 14 days prior to the step.
- groupSocStability: Median number of hours since the fusion of all dyads of group members.
Code/software
All analyses were conducted in R 4.2.1 (R Core Team, 2023).
Access information
Other publicly accessible locations of the data:
- The archived data are only available on Dryad.