Codes for filtering and analysing eBird data to calculate bird species' tolerances to human pressures and the calculated tolerances as dataset

README file created by Emma-Liina Marjakangas on December 8th 2023

---

We used binary observation data from eBird and model species’ occurrences as a function of the Human Footprint Index to quantify the world's bird species’ tolerances to human pressures. We included also the associated precision; values which can be applied to scientific research and conservation.

## Description of the data and file structure

Data for analysis steps are freely available from eBird (see below).

Dataset containing all calculated bird tolerances to human pressures are included in the excel-file "bird_tolerances". The first sheet "Tolerances" includes the tolerance values, while the second sheet "Metadata" includes the explanation for each column name and ranges of the presented values.

File: "bird_species_tolerances_dataset_240228.xlsx"

## Sharing/Access information

The data that support the analyses of this study are all freely available for research purposes, and can either be directly downloaded from the respective websites or are available on demand. Specifically, bird species checklists are available at http://ebird.org. Bird species’ distribution maps (BirdLife 2022) are available at http://datazone.birdlife.org. Bird species’ IUCN Red List statuses and population trend categories (IUCN Red List of Threatened Species, downloaded 24.04.2023) are available at http://iucnredlist.org. This analysis package does not include input data files as they are available freely available via the original platforms.

## Code/Software

All analyses were conducted in R software (version 4.0). The commented codes for the analysis steps are included in the data package. Each code file contains code for the analysis of one continent (Asia) as an example to reduce repetition. The code files are run sequentally from step1 to step5 as indicated in the code file names. Before the steps outlined below, eBird sampling event and species observation data need to be downloaded from eBird website. It is recommended to keep all necessary files in one directory folder or use a full path to the underlying datafiles needed for each step. The folder paths of the The general workflow of the code files is:

1. Filter eBird sampling location data to even out the spatial and temporal distribution of sampling event data
2. Obtain Human Footprint Index data for eBird sampling locations
3. Filter eBird species observation data to exclude poor-quality data

4. Quantify tolerances for each species within a continent

   1. Resample the data 50 times to obtain uncertainty estimates
   2. Fit Generalized Additive Models using eBird and Human Footprint Index data (core model structure occurrence ~ Human Footprint Index)
   3. Predict bird occurrences across the entire Human Footprint Index continuum to obtain a response curve
   4. Calculate three tolerance measures (Peak, Conservative and Maximum HTI) from the response curve

5. Summarize distributions of HTI values within continents

6. Summarize distributions of HTI values across continents and link them to species’ population trend and threat level information

Files:

step1_compile_ebird_data.R: Filter eBird sampling event and species observation data within continents

step2_resolve_taxonomies.R: Create a taxonomy crosswalk between different bird taxonomies used in different datasets

step3_quantify_tolerances.R: Loop through all species within a continent to fit occurrence~human pressure models, predict response curves and calculate tolerance values with uncertainty estimates

step4_summarize_tolerances_within_continent.R: Summarize tolerances of species within continents

step5_summarize_tolerances_across_continents.R: Summarize tolerances across continents and analyze links to population trends and threat statuses