Options to download GBIF occ data with rgbif. Some attention is also given to get back the species not present in the data.

download_data_rgbif_example.R

library(rgbif)
library(tidyverse)

# Create list of taxa we are interested to.
taxa <- tibble(
  species_name = c("Amelanchier lamarckii",
                   "Mareca sibilatrix",
                   "Ursus maritimus"
  )
)

# Countries we are interested to.
country <- c("BE", "LU")

# All types of occurrences, except `FOSSIL SPECIMEN` and `LIVING SPECIMEN`, 
# which can have misleading location information (e.g. location of captive animals).
basis_of_record <- c(
  "OBSERVATION", 
  "HUMAN_OBSERVATION",
  "MATERIAL_SAMPLE", 
  "LITERATURE", 
  "PRESERVED_SPECIMEN", 
  "UNKNOWN", 
  "MACHINE_OBSERVATION"
)

# Occurrences with a valid year: from 1000, up to now.
year_begin <- 1000
year_end <- year(Sys.Date())

# Occurrences with valid geographic coordinates.
hasCoordinate <- TRUE

# We want only presences.
occ_status <- "PRESENT"

# Option 1: get data directly in your R session (attention: 100k occs max, hard limit)

# This search will fail if one name is not found
keys <- sapply(taxa$species_name, 
               function(x) name_suggest(x)$data$key[1], USE.NAMES=FALSE
)
# Add keys to taxa. Useful later.
taxa$key <- keys


# It could takes a lot of time! Better to use it for previews (default `limit`
# argument value = 500)
# or if you expect very low amount of data. In my case it is ~ 8k occs.
occs <- occ_data(
  taxonKey = paste0(taxa$key, collapse = ";"),
  country = paste0(country, collapse = ";"),
  occurrenceStatus = occ_status, # Actually "PRESENT" is also the default value
  hasCoordinate = TRUE,
  basisOfRecord = paste0(basis_of_record, collapse = ";"),
  year = paste(c(year_begin, year_end), collapse = ","),
  limit = 50000 # high enough, but less than 100k
)

# Data are in `data` slot.
occs$data

# Which species from the original list are not present in our data?
# First, get all taxon keys we have in the data
accepted_synonyms_subtaxa_keys <- unique(
  c(
    unique(occs$data$taxonKey),
    unique(occs$data$speciesKey),
    unique(occs$data$acceptedTaxonKey)
    # add genusKey if you are providing one more genuses as taxa
  )
)

# Original keys not present in our cllection of keys from the data
absent_keys <- keys[! keys %in% accepted_synonyms_subtaxa_keys]

# And so the species:
taxa %>% filter(.data$key %in% absent_keys)


# Option 2: trigger a download. Note: GBIF credentials are required.
gbif_download_key <- occ_download(
  pred_in("taxonKey", keys),
  pred_in("country", country),
  pred_in("basisOfRecord", basis_of_record),
  pred_gte("year", year_begin),
  pred_lte("year", year_end),
  pred("hasCoordinate", hasCoordinate),
  user = rstudioapi::askForPassword("GBIF username"), # needed if not saved in R environment
  pwd = rstudioapi::askForPassword("GBIF password"), # needed if not saved in R environment
  email = rstudioapi::askForPassword("Email address for notification") # needed if not saved in R environment
)


# Check status of download
metadata <- occ_download_meta(key = gbif_download_key)
metadata$key
metadata$status

while (metadata$status != "COMPLETED") {
  message(paste("Download not ready. Status", metadata$status))
  Sys.sleep(time = 60)
}
if (metadata$status == "COMPLETED") {
  # Once metadata$status = "COMPLETED", you can download the data:
  path_dwc_occs <- "./" # current directory: change it as you wish
  occ_download_get(metadata$key,
                   path = path_dwc_occs # default: current directory
  )
}