Individual herbivores take risks based on resource quality: stoichiometric distribution models with snowshoe hares

Isabella C. Richmond; Juliana Balluffi-Fry; Eric Vander Wal; Shawn J. Leroux; Matteo Rizzuto; Travis R. Heckford; Joanie L. Kennah; Gabrielle R. Riefesel; Yolanda F. Wiersma

Manuscript is being prepared for submission at the Journal of Mammalogy. All data and code for the manuscript are available here and published on Zenodo.

This repository contains code, data, and results for Richmond et al. Individual herbivores take risks based on resource quality: stoichiometric distribution models with snowshoe hares.

Zenodo Repository:

script
This folder contains all the R scripts necessary to reproduce our analyses.

• 1 - DataTriangulationRazimuth.R takes raw VHF relocation data and triangulates them with error ellipses for each collar and date using the package razimuth.
• 2 - kUDRazimuth.R uses triangulated data to calculate kernel utilization distributions for each individual using adehabitatHR. These kernel utilization distributions are NOT used in future analysis.
• 3 - aKDERazimuth.R uses triangulated data to calculate autocorrelated kernel density estimates for each individual using ctmm. These kernels are used for all future analysis.
• 4 - HorizontalComplexity.R processes raw horizontal complexity data collected in the field and produces finalized versions for analysis.
• 5 - RiskOrdination.R takes all habitat complexity data and analyzes understory and overstory habitat complexity using Principal Components Analyses. PCA axes are then extracted for future analysis.
• 6 - HomeRangeExtraction.R extracts kernel utilization distribution data for each individual hare, food quality data, and habitat complexity data at each habitat complexity sampling point.
• 7a - lme4Models.R models our data with intensity of space use (kernel utilization distribution) as a response variable and habitat complexity and food quality as explanatory variables using linear mixed effects models in the lme4 package.
• 7b - MCMCglmmModels.R models our data using a Bayesian approach with intensity of space use (kernel utilization distribution) as a response variable and habitat complexity and food quality as explanatory variables using a Markov chain Monte Carlo sampler for linear mixed effects models in the MCMCglmm package.
• 7c - MCMCglmmPlots.R plots of random slopes and correlations used in the manuscript from the Bayesian models produced using MCMCglmm in 7b - MCMCglmmModels.R.
• 7d - lem4PowerAnalysis.R uses the pamm and lme4 package to build a Frequentist version of our global model and test if our data has enough power to assess individual-level reaction norms as per Martin et al. 2011 (Methods in Ecology and Evolution)
• 7e - MCMCModelsBGRisk.R models our data using a Bayesian approach with intensity of space use (kernel utilization distribution) as a response variable and habitat complexity and food quality as explanatory variables using a Markov chain Monte Carlo sampler for linear mixed effects models in the MCMCglmm package. Adds year of capture interacting with risk to test if there is variation in background risk across years.
• 8 - Maps.R - plots maps used in the manuscript.
• function-plotOutliers.R -a function used to plot outliers in 1 - DataTriangulationRazimuth.R
• function-plotVariograms.R - a function used to plot variograms in 3 - aKDERazimuth.R.
• function-residPlots.R - a function used to produce diagnostic plots for Frequentist models in 7a- lme4Models.R.

graphics
This folder contains all the graphics and figures produced during the analysis.

input
This folder and all folders within contain the raw data used in analysis.

output
This folder contains all the output from our analysis.

• outliers
  • Contains outlier plots from triangulation of VHF data in 1 - DataTriangulationRazimuth.R
• variograms
  • Contains variograms from movement models produced using ctmm in 3 - aKDERazimuth.R