eBird is a collective enterprise that takes a novel approach to citizen science by developing cooperative partnerships among experts in a wide range of fields: population ecologists, conservation biologists, quantitative ecologists, statisticians, computer scientists, GIS and informatics specialists, application developers, and data administrators. Managed by the Cornell Lab of Ornithology eBird’s goal is to increase data quantity through participant recruitment and engagement globally, but also to quantify and control for data quality issues such as observer variability, imperfect detection of species, and both spatial and temporal bias in data collection. eBird data are openly available and used by a broad spectrum of students, teachers, scientists, NGOs, government agencies, land managers, and policy makers. The result is that eBird has become a major source of biodiversity data, increasing our knowledge of the dynamics of species distributions, and having a direct impact on the conservation of birds and their habitats.

eBird basic dataset (https://science.ebird.org/en/use-ebird-data/download-ebird-data-products) query for Central California Drylands. These data were then associated with sites studied in ecological research regionally for plant-animal interactions and restoration ecology. Data were derived in 2022 from data product.

This R code is used to download species occurrence data from open datasets such as Global Biodiversity Information Facility, Atlas of Living Australia, Biodiversity Information Serving Our Nation, iNaturalist, eBird, and Integrated Digitized Biocollections.

This repository contains code and data for Callaghan et al. 2021. Global abundance estimates for 9,700 bird species. PNAS. (https://doi.org/10.1073/pnas.2023170118). This readme is intended to provide an overview of the data analysis. eBird data are free to download (https://ebird.org/data/download) and we used the eBird basic dataset (version ebd_relMay2019). The first thing to note is that the eBird dataset is a ‘live’ dataset whereby records from the past can be added or removed currently. This means that we could have used a record in our dataset that was present when we downloaded the data, but no longer present. This readme is set up to guide an interested user through the repository and only focuses on highlighting the main steps/workflow and does not highlight the function of every R script. It anticipates that a reader has read the Methods section of the paper in detail.

File List hb-migration.r (MD5: 1904c1692a02d984890e4575d0eeb4e6) R script that imports the eBird, map, and equal-area icosahedron data, summarizes the population-level migration patterns, runs the statistical analyses, and outputs figures.

    migration-fxns.r (MD5: a2ae2a47c066a253f18cad5b13cddcf6)
     R script that holds the relevant functions for executing the hb-migration.R script.

    BBL-Appendix.r (MD5: 370c701d6afb07851907922dcab51de4)
     R script that imports the Breeding Bird Laboratory data and outputs the figures for the Appendix.

    output-data.zip (MD5: 36e3a92a7d35e84b299d82c8bd746950)
    Folder containing the partially-processed text files (15 .txt files, 3 per species for centroids, migration dates, and migration speed) for the main analyses and figures in the paper. These text files can be used in part II of hb-migration.r and contain output data on the daily population-level centroids, migration dates, and migration speed. Part I of hb-migration.r relies on raw eBird data, which was queried from the eBird server directly. The raw eBird data can be requested through their online portal after making a user account (http://help.ebird.org/customer/portal/articles/1010524-can-i-download-raw-data-from-ebird-). The equal-area icosahedron maps are available at (http://discreteglobalgrids.org/). The BBL data, used in BBL-Appendix.R, can be requested from the USGS Bird Banding Laboratory (http://www.pwrc.usgs.gov/BBL/homepage/datarequest.cfm).

  Description
    The code and data in this supplement allow for the analyses and figures in the paper to be fully replicated using a data set of manipulated communities collected from the literature.
   Requirements: R 3.x, and the following packages: chron, fields, knitr, gamm4, geosphere, ggplot2, ggmap, maps, maptools, mapdata, mgcv, plyr, raster, reshape2, rgdal, Rmisc, SDMTools, sp, spaa, and files containing functions specific to this code (listed above).
    The analyses can then be replicated by changing the working directory at the top of the 
     file hb-migration.R to the location on your computer where you have stored the .R 
    and .csv files and running the code. Note that to fully replicate the analyses, the data will need to be requested from the sources listed above.
    Starting at Part II in hb-migration.R, it should take approximately 30 minutes to run all the code from start to finish. Figures should output as pdfs in your working directory. If you download the raw data and run the analyses starting at Part I, you will need a workstation with large memory to run the analyses in a reasonable amount of time since the raw eBird datafiles are very large.
    Version Control Repository: The full version control repository for this project (including post- publication improvements) is publicly available https://github.com/sarahsupp/hb-migration. If you would like to use the code in this Supplement for your own analyses it is strongly suggested that you use the equivalent code in the repositories as this is the code that is being actively maintained and developed.
    Data use: Partially-processed data is provided in this supplement for the purposes of replication. If you wish to use the raw data for additional research, they should be obtained from the original data providers listed above.

All data used in this assessment came from the World eBird Basic Dataset (Sullivan et al., 2009; publicly available at ebird.org/data/download). We downloaded the data from eBird on January 21, 2020 and filtered it using the R package ‘Auk’ (Strimas-Mackey, Miller & Hochachka, 2018) to include only records of interest (filtered by country and species). The result was a list of all eBird shorebird records in the insular Caribbean.We filtered to include the following countries:("Jamaica", "Cuba", "The Bahamas", "Haiti", "Dominican Republic", "Antigua and Barbuda", "Dominica", "Saint Lucia", "Saint Vincent and the Grenadines", "Barbados", "Grenada", "Trinidad and Tobago", "Anguilla", "Aruba", "Bahamas", "Caribbean Netherlands", "Cayman Islands", "Guadeloupe", "Martinique", "Sint Maarten", "Saint Kitts and Nevis", "Virgin Islands (British)", "Virgin Islands (U.S.)", "Puerto Rico", "Saint Martin (French part)", "Turks and Caicos Islands", "Curaçao", "Saint Barthélemy", "Montserrat")We filtered to include the following species:("Black-necked Stilt", "American Avocet", "American Oystercatcher", "Black-bellied Plover", "American Golden-Plover", "Pacific Golden-Plover", "Snowy Plover", "Wilson's Plover", "Semipalmated Plover", "Piping Plover", "Killdeer", "Upland Sandpiper", "Whimbrel", "Long-billed Curlew", "Hudsonian Godwit", "Marbled Godwit", "Ruddy Turnstone", "Red Knot", "Stilt Sandpiper", "Sanderling", "Dunlin", "Baird's Sandpiper", "Least Sandpiper", "White-rumped Sandpiper", "Buff-breasted Sandpiper", "Pectoral Sandpiper", "Semipalmated Sandpiper", "Western Sandpiper", "Short-billed Dowitcher", "Long-billed Dowitcher", "Wilson's Snipe", "Spotted Sandpiper", "Solitary Sandpiper", "Wandering Tattler", "Greater Yellowlegs", "Willet", "Lesser Yellowlegs", "Wilson's Phalarope", "Red-necked Phalarope", "Red Phalarope", "Double-striped Thick-knee", "Southern Lapwing", "Collared Plover", "Northern Jacana", "Wattled Jacana", "South American Snipe")We dropped the following columns in our filter:("global_unique_identifier", "last_edited_date", "taxonomic_order", "category", "subspecies_common_name", "breeding_bird_atlas_code", "breeding_bird_atlas_category", "age_sex", "country_code", "state_code","county", "county code", "usfws_code", "atlas_block", "locality_type", "time_observations_started", "project_code", "all_species_reported", "has_media", "trip_comments", "species_comments")

A dataset containing 275718645 species occurrences available in GBIF matching the query: DatasetKey: EOD - eBird Observation Dataset. The dataset includes 275718645 records from 1 constituent datasets: 445686535 records from EOD - eBird Observation Dataset. Data from some individual datasets included in this download may be licensed under less restrictive terms.

Dataset collection and processing is described in the methods section of the associated publication.

Climate change is expected to have a profound impact on species distributions, contracting suitable climate space. Biodiversity areas are important to mitigate these negative effects but are static by design and thus do not account for future projections of species distributions. The Harpy Eagle Harpia harpyja has a broad range across lowland Neotropical forests and thus its distribution could be negatively affected by climate change when combined with current rates of habitat loss. To test this hypothesis, we use spatial point process models fitted with climatic, topographic, and landcover covariates to identify current distribution. We then project to 24 future climate scenarios, using three General Circulation Models (GCMs), and two emission scenarios between the years 2021 and 2100 averaged over four 20-year periods. Our current model identified a core range across Amazonia and the Guiana Shield, with evergreen forest (71 %), mean diurnal temperature range (13 %), and elevation (6 %) the most important predictors. Reclassifying the current model to a binary prediction estimated a range size of ~7.6 million km2, with the Important Bird and Biodiversity Area (IBA) network covering 18 % of habitat (~1.4 million km2) within this range. By 2090, range size was predicted to decrease on average by 14.4 % under a higher emissions scenario, and 7.3 % under a lower emissions scenario. The IBA network would cover 14 % less area under a higher emissions scenario, and 3.3 % less distribution area under a lower emissions scenario by 2090. Southern Amazonia is predicted to have the greatest reduction in range size and subsequently highest loss of Harpy Eagle habitat within the IBA network. Our work demonstrates that the combination of climate change and subsequent habitat loss may result in substantial losses in distribution for this raptor across the southern edge of its range.

Data to support publication: Reduced range size and Important Bird and Biodiversity Area coverage for the Harpy Eagle (Harpia harpyja) predicted from multiple climate change scenarios

Species Occurrences:

GBIF Occurrence Download: https://doi.org/10.15468/dl.6ikhnj

eBird data: Download Data - eBird

Data from Miranda et al. 2019: Species distribution modeling reveals strongholds and potential reintroduction areas for the world’s largest eagle (plos.org)

Nest locations from the Darien region of Panama are available on request.

Environmental data available from: EarthEnv ENVIREM WorldClim

A dataset containing 65 species occurrences available in GBIF matching the query: Country: Iran, Islamic Republic of Year: 2001-2017 Geometry: POLYGON ((46.377410888671875 38.01228608167874, 46.31561279296875 38.03067754878772, 46.248321533203125 38.04473849713312, 46.220855712890625 38.06528425740672, 46.207122802734375 38.11176125672556, 46.26068115234375 38.14200946157744, 46.34033203125 38.109600191095176, 46.404876708984375 38.05230865877123, 46.399383544921875 38.015531970102984, 46.377410888671875 38.01228608167874)) HasCoordinate: true TaxonKey: Aves HasGeospatialIssue: false. The dataset includes 65 records from 1 constituent datasets: 65 records from EOD - eBird Observation Dataset. Data from some individual datasets included in this download may be licensed under less restrictive terms.

The AVONET database contains comprehensive functional trait data for all birds, including six ecological variables, eleven continuous morphological traits, and information on range size and location. Raw morphological measurements are available from 90020 individuals of 11009 extant bird species sampled from 181 countries. These data are also summarised as species averages in three taxonomic formats, allowing integration with a global phylogeny, geographical range maps, IUCN Red List data, and the eBird citizen science database. The full AVONET dataset including raw morphological measurements as well as species averages for each taxonomy is provided in 'AVONET Supplementary dataset 1' Data on duplicate measurements for a subset of individuals are provided in 'Supplementary dataset 2'Data and Code to reproduce the analyses and figures presented in Tobias et al 2022 (Ecology Letters doi: https://doi.org/10.111/ele.13898) is included in the 'ELEData' and 'ELECode' zip files.

The Global Biodiversity Information Facility (GBIF) indexes thousands of biodiversity datasets from Natural History Collections, citizen science initiatives (e.g., iNaturalist, eBird), and other sources. As part of the index process, GBIF associates at least two identifiers with indexed records: a record id (aka gbifID) and a dataset id (aka dataset key). These ids are central to do lookup, reference data, and package interpreted data products.

This publication contains an exhaustive list of GBIF IDs and ids associated by their data providers as derived from:

GBIF.org (01 March 2023) GBIF Occurrence Download https://doi.org/10.15468/dl.pk3trq

The resource (size: ~260GB) provided by GBIF had content id hash://sha256/c8bac8acb28c8524c53589b3a40e322dbbbdadf5689fef2e20266fbf6ddf6b97 and was used to generate the resource included in this publication using

preston cat 'zip:hash://sha256/c8bac8acb28c8524c53589b3a40e322dbbbdadf5689fef2e20266fbf6ddf6b97!/0015281-230224095556074.csv'
| cut -f 1,2,3,37,38,39
| gzip\

gbifid.tsv.gz

with the content id of gbifid.tsv.gz (size: ~35GB) being hash://sha256/a339e32e10edaad585f61f2ded06cbb23e0618c65a6360db18d7d729054940a8 .

the first 10 lines of gbifid.tsv.gz as extracted via

preston cat --remote https://zenodo.org/record/7789866/files,https://linker.bio hash://sha256/a339e32e10edaad585f61f2ded06cbb23e0618c65a6360db18d7d729054940a8
| gunzip
| head

are:

gbifID datasetKey occurrenceID institutionCode collectionCode catalogNumber 2997162320 c71c8000-9fc7-422c-804a-ce6abe751771 3399442 CEPEC CEPEC CEPEC00109669 2997162309 c71c8000-9fc7-422c-804a-ce6abe751771 2733085 CEPEC CEPEC CEPEC00000818 2997162317 c71c8000-9fc7-422c-804a-ce6abe751771 2733086 CEPEC CEPEC CEPEC00000888 2997162313 c71c8000-9fc7-422c-804a-ce6abe751771 3399443 CEPEC CEPEC CEPEC00109744 2997162306 c71c8000-9fc7-422c-804a-ce6abe751771 2733087 CEPEC CEPEC CEPEC00000889 2997162316 c71c8000-9fc7-422c-804a-ce6abe751771 3399440 CEPEC CEPEC CEPEC00109605 2997162324 c71c8000-9fc7-422c-804a-ce6abe751771 2733088 CEPEC CEPEC CEPEC00000890 2997162308 c71c8000-9fc7-422c-804a-ce6abe751771 3399441 CEPEC CEPEC CEPEC00109615 2997162303 c71c8000-9fc7-422c-804a-ce6abe751771 2733089 CEPEC CEPEC CEPEC00000891

Note that at time of writing, the html resource associated with the occurrence id 2997162320, and data set key c71c8000-9fc7-422c-804a-ce6abe751771 (extracted from of the first data row example above) are available via:

https://gbif.org/occurrence/2997162320

and

https://gbif.org/dataset/c71c8000-9fc7-422c-804a-ce6abe751771

respectively.

This resource was initially created to help integrate with Bionomia (https://bionomia.net) to help associate people identifiers provided by bionomia to their original records via their GBIF ids. Bionomia re-uses GBIF records ids as a way to define links between records and the people (e.g., curators, collectors, identifiers) that worked on them.

In other words, this resource provides a versioned translation table from the GBIF data universe (as defined by GBIF record ids, and dataset keys) to the data collections that exist (and evolve) independent of it.

Note that the resource identified by hash://sha256/c8bac8acb28c8524c53589b3a40e322dbbbdadf5689fef2e20266fbf6ddf6b97 was not included in this publication it was too big (260GB) to fit. You may be able to retrieve the resource from its original location at https://api.gbif.org/v1/occurrence/download/request/0015281-230224095556074.zip .

Quantifying habitat use is important for understanding how animals meet their requirements for survival and provides information for conservation planning. Currently, assessments of range-wide habitat use that delimit species distributions are incomplete for many taxa. The Harpy Eagle (Harpia harpyja) is a raptor of conservation concern, widely distributed across Neotropical lowland forests, that currently faces threats from habitat loss and fragmentation. Here, we use penalized logistic regression to identify species-habitat associations and predict habitat suitability based on a new International Union for the Conservation of Nature range metric, termed Area of Habitat. From the species-habitat model, we performed a gap analysis to identify areas of high habitat suitability in regions with limited coverage in the Key Biodiversity Area (KBA) network. Range-wide habitat use indicated that Harpy Eagles prefer areas of 70-75% evergreen forest cover, low elevation, and high vegetation species richness. Conversely, Harpy Eagles avoid areas of >10% cultivated landcover and mosaic forest, and topographically complex areas. Our species-habitat model identified a large continuous area of potential habitat across the pan-Amazonia region, and a habitat corridor from the Chocó-Darién ecoregion of Colombia running north along the Caribbean coast of Central America. Little habitat was predicted across the Atlantic Forest biome, which is now severely degraded. The current KBA network covered 18% of medium to high Harpy Eagle habitat exceeding a target biodiversity area representation of 10%, based on species range size. Four major areas of high suitability habitat lacking coverage in the KBA network were identified in north and west Colombia, western Guyana, and north-west Brazil. We recommend these multiple gaps of habitat as new KBAs for strengthening the current KBA network. Modelled area of habitat estimates as described here are a useful tool for large-scale conservation planning and can be readily applied to many taxa. Methods Environmental data available from: EarthEnv: www.earthenv.org ENVIREM: http://envirem.github.io/#downloads Occurrence data available from: GBIF: https://doi.org/10.15468/dl.6ikhnj eBird: https://ebird.org/data/download/ebd Miranda et al. (2019) database: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0216323#sec011

This data set contains basic observation information on Western Grebes in California and is provided by the Avian Knowledge Network (AKN). AKN is a joint venture including sponsors National Biological Information Infrastructure and National Science Foundation, and collaborators Cornell Lab of Ornithology, PRBO Conservations Science, Redwood Sciences Laboratory, Rocky Mountain Bird Observatory, and Bird Studies Canada. AKN collects, organizes, archives, and distributes avian data for various analytical purposes. This large (almost 20,000,000 observations to date) and comprehensive database provides a unified structure where data from multiple sources can be used together to produce more meaningful products. AKN also has developed applications to visualize temporal and spatial patterns of bird movements and numbers. These bird observations were made by professional birding consultants, ornithologists, researchers, and knowledgeable volunteers over the last few years. Most of their observations came from point counts from job sites, back yards, or from sites established to assess bird populations in specific locations or habitats throughout California. Protocols for the three programs that produce the most observations are on the web at: eBird - http://www.ebird.org/content/ Great Backyard Bird Count - http://www.birdsource.org/gbbc/ Project FeederWatch - http://www.birds.cornell.edu/pfw/ The dataset is limited in that the there is no observer provided for each record, the location is only accurate to a circular area equivalent to about one-sixteenth of a square mile, and a number of fields have only code values and the code definitions are not known.

Joint Ventures (JVs) are bird conservation partnerships, established to achieve the goals of the North American Waterfowl Management Plan (NAWMP). There are 22 bird habitat JVs in North America, and our JV, the Upper Mississippi/Great Lakes Region Joint Venture (UMGL JV) is located in the Midwest, in a pivotal location in the Mississippi Flyway, for people involved in migratory bird conservation. We are a collaborative, regional group of government agencies, tribes, nonprofit organizations, corporations, universities, and individuals that conserve habitat for the benefit of migratory birds, other wildlife, and people. Our landscape includes eastern Minnesota; all of Wisconsin and Michigan; eastern Nebraska and Kansas; western, southern and eastern Iowa; northern Missouri, Illinois and Indiana; and northwestern Ohio.To address Waterbird habitat limitation during the non-breeding (migration-staging and winter) periods, we aim to maximize focal species recruitment through conservation of limiting non-breeding habitats. To help achieve this objective, we used county-level eBird and Integrated Waterbird Management and Monitoring (IWMM; 20 National Wildlife Refuges) data for non-breeding focal species. These regional data sets were limited for several species but provide a basis for assessing distribution of non-breeding waterbirds. We assumed habitat area was more limited for non-breeding waterbirds where counts per wetland were highest. In each county, we appraised average eBird (2007–2016) and IWMM (2010–2017) waterbird counts relative to county size and percent coverage of wetlands important to distinct non-breeding guilds: Emergent, Forested, Aquatic Bed, and Unconsolidated Bottom and Shore (see table 1 in 2018 JV Waterbird strategic plan). Count neighborhoods (from kernel density analysis; see Soulliere and Al-Saffar 2017) were ranked by the density of waterbirds per wetland area, emphasizing non-breeding expanses with highest counts relative to wetland coverage. The resulting model highlighted the areas having the most important waterbird non-breeding habitats for retention and expansion. Finally, we converted the original raster datset of this model to TIFF format (this product) to improve data visualization, assessment, and interactivity online. The raster map was published in the JV Waterbird Habitat Conservation Strategy – 2018 Revision (Figure 10B).This product shows the distribution and relative density (scale of 0 – 1; values of low – high) of the most valuable areas (pixel neighborhoods) across the JV region, to acquire or manage whatever the required conservation action at the local scale is (i.e., retention, protection, and or restoration). Aggregate-priority-areas, emphasized mostly across Bird Conservation Region (BCR) 22, were predicted to produce the most value for the decision makers.This product can be scaled-down (i.e., stepped-down to state, watershed, county, or any smaller target area) and supplemented with additional local data for customized output maps depicting best areas for Waterbird non-breeding habitat restoration vs. retention vs. enhancement. See the JV website and related step-down publications for details.Links:Coming Soon: To download this TIFF dataset and related data, metadata, and documentation, please browse the UMGL JV online folder in the U.S. Geological Survey's ScienceBase repository.To view and explore this TIFF dataset and related data, please visit the online experience of the UMGL JV Decision Support Tools.References:Model codes and final steps (for internal use of UMGLJV staff): (i) AgsJvw_Hcs17YrSumFor8BrdNoWB_Ebd0716And21NWRs1017Pts_Us1034Cnty19St_1mRpts100mTKd1km31kmR_SumSppAvgsNrm2Times_RevisedV_Ready = albers_Jv18WbBrdNo. (ii) albers_Jv18WbBrdNo → converted to TIFF and projected on-the-fly to WGS84 → renamed to Jv18WbBrdNoR.Coming Soon: For internal use of USFWS staff: Data Management Plan (DMP) ID = 000; Title: UMGL JV Non-Breeding Waterbird Distribution.