Wolf presence disrupts seasonal variation in hair cortisol among free-ranging beef cattle
Data files
May 13, 2026 version files 37.22 KB
-
df.csv
12.81 KB
-
README.md
4.88 KB
-
WolfPresenceBeefCattle.Rmd
19.53 KB
Abstract
Free-ranging livestock face increasing exposure to recovering carnivore populations, yet the physiological consequences of predator reintroduction remain poorly understood. Here, we examined hair cortisol concentrations in beef cattle from nine herds across California's Sierra Nevada to assess how wolf presence affects stress physiology over seasonal transitions. We collected hair samples before and after summer grazing periods, comparing herds exposed to the Lassen Wolf Pack with unexposed herds across a natural temperature gradient. Using Bayesian multilevel regression models, we tested for effects of wolf presence and found that hair cortisol concentrations decreased with increasing minimum temperatures (13.5% reduction per 1°C increase), but this relationship was disrupted in wolf-exposed herds. We also found that wolf presence moderated the cortisol-temperature relationship, with wolf-exposed herds showing inverted patterns during summer compared to unexposed herds. These results suggest that wolf presence disrupts normal physiological regulation rather than simply elevating hair cortisol concentrations, suggesting dysregulation indicative of sustained allostatic load in wolf-naïve cattle populations. Our findings demonstrate that predator reintroduction can fundamentally alter livestock physiology indirectly via metabolic regulation, with implications for animal welfare, productivity, and carnivore-livestock coexistence strategies in working landscapes.
Dataset DOI: 10.5061/dryad.s7h44j1mk
Description of the data and file structure
This repository contains the data and analysis code for the study examining how wolf presence affects hair cortisol concentrations in free-ranging beef cattle across seasonal transitions.
Study Overview
Study Period: May-November 2022
Location: 780,000 acres across western Lassen and northwestern Plumas Counties, California
Sample Size: 79 cattle from 9 herds, 158 hair samples total
Files and variables
## File Descriptions
Data Files
df.csv - Main dataset containing:
-
Animal.ID: Unique identifier for each individual cattle (anonymized) -
Herd: Herd identifier (A-I, anonymized) -
Study.Period: Sampling period (Winter, Summer) -
wolfImpact: Wolf exposure status (N = no exposure, Y = wolf exposure) -
pgCort: Hair cortisol concentration (pg/mg) -
avg_daily_min: Average daily minimum temperature (°C) for ~3 months preceding sampling -
avg_daily_min.s: Standardized average daily minimum temperature (mean=0, sd=1) -
breedColor: Combined breed and coat color variable -
Date.Collected: Sample collection date (MM/DD/YY format)
Code Files
WolfPresenceBeefCattle.Rmd - R Markdown file containing all analyses and figure generation, including:
-
Data preprocessing and standardization
-
Prior predictive checks
-
Bayesian multilevel regression modeling using brms
-
Model diagnostics and validation
-
Figure generation for all manuscript figures
Methods Summary
Sample Collection
Hair samples were collected using a shave-reshave protocol from the tail switch of mature cows. Samples were collected twice per animal: before movement to summer grazing (winter sample) and after return from summer grazing (summer sample), with mean interval of 20.2 ± 3.9 weeks.
Temperature Data
Daily minimum temperature data were obtained from the nearest weather station to each study location for approximately 3 months preceding each sampling date (median = 107 days, sd = 27.7 days). Values were averaged across each study period and standardized for analysis.
Cortisol Assay
Hair cortisol was extracted from 3-cm samples and analyzed using enzyme immunoassays (Salimetrics, Carlsbad, CA, USA). Assay validation included parallelism and spike-and-recovery tests. Interassay CV averaged 13.02 ± 1.21%, intraassay CV averaged 1.27% ± 1.17%.
Wolf Exposure Classification
-
Wolf-exposed herds (G, H, I): Grazed in areas with documented Lassen Wolf Pack activity and experienced confirmed wolf-related impacts including depredations
-
Control herds (A-F): Grazed in areas with no documented wolf activity
Code/software
Statistical Analysis
Bayesian multilevel regression models were implemented using the brms package in R. The model included:
-
Outcome: Log-transformed hair cortisol concentration
-
Predictors: Wolf impact, study period, standardized temperature, breed/color
-
Interactions: Three-way interaction between wolf impact, study period, and temperature
-
Random effects: Animal ID nested within Herd
-
Priors: Weakly informative normal(0,1) priors for all coefficients
-
Family: Log-normal distribution
Software Requirements
R Packages Required:
-
brms (Bayesian modeling)
-
ggplot2 (visualization)
-
dplyr (data manipulation)
-
tidyr (data reshaping)
-
nord (color palettes)
-
bayestestR (Bayesian diagnostics)
-
bayesplot (posterior predictive checks)
-
lubridate (date handling)
-
performance (model diagnostics)
R Version
Analysis conducted in R version 4.02 with Stan backend for Bayesian computation.
Data Usage Notes
-
Anonymization: All herd and animal identifiers have been anonymized to protect rancher privacy
-
Temperature standardization: Use provided standardized temperature values (avg_daily_min.s) for model replication
-
Missing values: Random effects structure accounts for unbalanced data across herds
-
STRANGE considerations: Sample limited to adult cows from voluntarily participating ranches with consistent management practices
Model Validation
The analysis includes comprehensive model validation:
-
Prior predictive checks
-
Posterior predictive checks using density overlays and scatter plots
-
Leave-one-out probability integral transformation (LOO-PIT) checks
-
R-hat and effective sample size diagnostics
-
Parameter correlation assessment via pairs plots
Funding
Western Sustainable Agriculture and Research Education (#SW22-931) and Russell L. Rustici Rangeland & Cattle Research Endowment
Ethics and Permits
UCD Institutional Animal Care and Use Committee (IACUC) Approved Protocol #22895