coral-bleach-recovery
This repository contains R code for the Bayesian models accompanying Robinson, Wilson and Graham "Abiotic and biotic controls on coral recovery 16 years after mass bleaching", Coral Reefs 2019.
https://link.springer.com/article/10.1007/s00338-019-01831-7
The following R packages were used to analyse data.
install.packages(c("tidyverse", "rethinking", "here"))
data
recovery_year_model.Rdata
contains model structure (rec.year.m) and parameter effect sizes (rec.params) for predicting recovery yearrecovery_trajectory.Rdata
contains logistic model strucutre (rec.trajectory.m), predicted recovery years (base) and recovery trajectory for each reef over 100 years (rec.trajectory)posterior_sims.Rdata
contains posterior samples for each recovery year predictor covariate (depth, complex, init_cover, wave, herb, coral_juv, nitrogen)jacknife/
contains model structures and posterior samples for jacknife sensitivity analysis
figures
logistic_models_predictions.pdf
are model predictions for each candidate logistic model, with uncertainty intervals and plotted against observed datarecovery_diagnostic_jackknife.pdf
are parameter effect sizes for each jacknife subsample
scripts
1_logistic_growth_model.R
fits Bayesian logistic growth models2_recovery_year_model.R
fits Bayesian linear model to predict recovery year3_jacknife_analysis.R
runs jacknife sensitivity analysis on recovery year model (2)scaling_function.R
is generic function for scaling covariates to mean of 0 (for continuous) or creating dummy variables (for categorical)