Data from: Seasonal macro-demography of North American bird populations revealed through citizen science monitoring
Data files
Nov 25, 2024 version files 225.56 MB
-
data_code.zip
225.56 MB
-
README.md
3.96 KB
Abstract
Avian population sizes fluctuate and change over vast spatial scales, but the mechanistic underpinnings remain poorly understood. A key question is whether spatial and annual variation in avian population dynamics is driven primarily by variation in breeding season recruitment or by variation in overwinter survival. We present a method using large-scale eBird citizen-science data to develop species-specific indices of net population change as proxies for survival and recruitment gain, based on twice-annual, rangewide snapshots of relative abundance in spring and fall. We demonstrate the use of these indices by examining spatially explicit annual variation in survival and recruitment gain in two well-surveyed nonmigratory North American species, Carolina Wren (Thryothorus ludovicianus) and Northern Cardinal (Cardinalis cardinalis). We show that, while interannual variation in both survival and recruitment gain is slight for Northern Cardinal, eBird abundance data reveal strong and geographically coherent signals of interannual variation in the overwinter survival of Carolina Wren. As predicted, variation in wintertime survival dominates overall interannual population fluctuations of wrens and is correlated with winter temperature and snowfall in the northeastern United States, but not the southern United States. This study demonstrates the potential of volunteer-collected big datasets like eBird for inferring variation in demographic rates and introduces a new complementary approach towards illuminating the macrodemography of North American birds at comprehensive continental extents.
Seasonal macro-demography of North American bird populations revealed through citizen science monitoring
https://doi.org/10.5061/dryad.69p8cz991
Description of the data and file structure
This zip file contains underlying dataset, R script and R package that accompany the following project: Socolar, J., B. Galtbalt, A. Johnston, F. L. Sorte, O. Robinson, K. Rosenberg, and A. Dokter. 2024. Seasonal macro-demography of North American bird populations revealed through citizen science monitoring. Ecography.
The zipped folder (data_code) contains:
- “carwre.rds” and “norcar.rds” - the bootstrapped abundance data for Carolina wren and Northern cardinal derived from eBird Basic Dataset (EBD).
- “tmax_winter-tmax_summer-swe.rds” - the daymet weather dataset, extracted from: https://developers.google.com/earth-engine/datasets/catalog/NASA_ORNL_DAYMET_V4
- “residents.R” script specifying the workflow
- “macrodemography” R package containing the necessary functions to run the code.
The raw eBird data are freely available online from the Cornell Lab of Ornithology.
Sharing/Access information
Links to other publicly accessible locations of the data:
- https://science.ebird.org/en/use-ebird-data/download-ebird-data-products
Data was derived from the following sources:
- https://science.ebird.org/en/use-ebird-data/download-ebird-data-products
Code/Software
Datasets: “carwre.rds” and “norcar.rds” contains the data to run the workflow, “residents.R”, starting from the section named “Calculate demographic indices” after loading the necessary packages. This section in the script loads the data as per below. It is recommended to put these species data inside the output directories as specified below:
file_species <- paste0(params$output_path, "/abun_data/", species_code , ".rds")
data <- readRDS(file_species)
Both datasets contain a large list with two components (“grid” and “abun”). Only component “abun” (i.e., abundance) is used for downstream analyses. The abundance data for all hexagon cells for each season and year is structured as individual dataframe within the list components of year, and contains following variables:
- cell : Large hexagon cell ID
- tot_small: total number of small hexagon cells within a large hexagon cell
- mean: mean abundance estimate for the focal species in given large hexagon cell for given year and season
- rep_1 - 100: Hundred replicates of bootstrapped abundance estimate from which above mean is calculated.
The dataset ”tmax_winter-tmax_summer-swe.rds” contains the weather data to run weather regressions as specified in “residents.R”, and loaded by calling:
weather_file <- paste0(params$input_path, "weather/",paste0(params$daymet$label, collapse = "-"),".rds")
data_daymet <- readRDS(weather_file)
This weather data contains following variables:
- cell: Large hexagon cell ID
- year: respective year of the weather variables
- tmax_winter: maximum winter temperature in given cell and year (Celsius)
- tmax_summer: maximum summer temperature in given cell and year (Celsius)
- swe: snow water equivalent, the amount of water contained within the snowpack, kg/m^2
R is required to run “residents.R” script which installs the package “macrodemography” as part of the workflow; the script was created using version 4.4.0. Annotations are provided throughout the script through 1) defining parameters and input/output paths, 2) library loading, 3) importing raw data, 4) bootstrapping abundance, 5) calculating the demographic indices, 6) comparing recruitment and mortality variances, 7) running weather regressions, 8) smoothing regression estimates using ICAR (Intrinsic Auto-Regressive) models. Empty folders present are generated during the analysis.