dochvam / GLMM_Nmix_reproducible

Reproducible code for paper comparing GLMM and N-mixture outcomes from the eBird dataset.

Reproducing results for "Comparing N-mixture models and GLMMs for relative abundance estimation in eBird"

Reproducible code for paper comparing GLMM and N-mixture outcomes from the eBird dataset.

How to use this repository

This repository contains code for processing eBird code, retrieving data for environmental covariates and running the analysis in the associated paper.

If you encounter any problems or have any questions, feel free to contact the corresponding author Ben Goldstein.

Due to copyright, this repository does not include the eBird dataset, the EltonTraits dataset, nor the LANDFIRE landscape covariate dataset. To use these scripts, these files must be downloaded and added to the data/ directory. Instructions for filepathing can be found in the files where these inputs are read. They can be retrieved from the following public repositories:

• eBird basic dataset: https://ebird.org/data/download
• EltonTraits: https://esapubs.org/archive/ecol/E095/178/
• LANDFIRE: https://www.landfire.gov/evt.php

Once that's been done, numbered files can be run in sequence. The following scripts are included:

• 00_process_raw_eBird.R - converts the .txt eBird file into a CA-level relational database
• 01_prepare_eBird_data.R - filters eBird data for quality, chooses species-subregion datasets (SSRs)
• 02_prepare_covariate_data.R - associates checklists with spatial covariates (climate, landcover)
• 03_fit_one_ssr_final.R - contains the functions called for fitting models. This is the bulk of the work
• 04_batch_many_ssr.R - batch-runs the functions in 03 over all SSRs in parallel
• 05_visualize_model_results.R - Creates figures for primary results
• 06_post_tests.R - contains code defining post-tests (goodness-of-fit, stability, autocorrelation)
• 07_batch_posttests.R - runs the posttests over all SSRs
• 08_visualize_model_results.R - visualize posttest results

The following auxilliary scripts are also included:

• model_code.R - contains code defining NIMBLE models
• nmixgof_manual.R - support code for computing RQ residuals. Code adapted from Knape et al. 2018 (DOI: 10.1111/2041-210X.13062)
• nmixgof_manual.cpp - c++ code for above
• read_results_helper_file.R - contains code for translating output .RDS files in a summary df