R/download_worldclim_present.R

In EvolEcolGroup/pastclim: Manipulate Time Series of Palaeoclimate Reconstructions

#' Download a WorldClim modern observations.
#'
#' This function downloads annual and monthly variables from the WorldClim 2.1 dataset.
#' @param dataset the name of the dataset
#' @param bio_var the variable name
#' @param filename the file name (full path) of the file to be saved
#' @returns TRUE if the requested WorldClim variable was downloaded successfully
#'
#' @keywords internal

download_worldclim_present <- function(dataset, bio_var, filename) {
  # reset warnings for gdal to terra default
  # this is necessary if sf was loaded in the mean time
  terra::gdal(warn = 3)

  # get resolution from the dataset name and convert it to the original
  res_conversion <- data.frame(
    our_res = c("10m", "5m", "2.5m", "0.5m"),
    wc_res = c("10m", "5m", "2.5m", "30s")
  )
  wc_res <- res_conversion$wc_res[res_conversion$our_res == substr(dataset,
    start = 15,
    stop = nchar(dataset)
  )]

  # function to grab the number from the raster layer
  if (grepl("bio", bio_var)) {
    postfix <- "bio.zip"
    var_prefix <- "bio"
  } else if (grepl("temperature_min", bio_var)) {
    postfix <- "tmin.zip"
    var_prefix <- "temperature_min_"
  } else if (grepl("temperature_max", bio_var)) {
    postfix <- "tmax.zip"
    var_prefix <- "temperature_max_"
  } else if (grepl("temperature_", bio_var)) {
    postfix <- "tavg.zip"
    var_prefix <- "temperature_"
  } else if (grepl("precipitation_", bio_var)) {
    postfix <- "prec.zip"
    var_prefix <- "precipitation_"
  } else if (grepl("altitude", bio_var)) {
    postfix <- "elev.zip"
    var_prefix <- "elevation"
  }

  base_url <- "https://biogeo.ucdavis.edu/data/worldclim/v2.1/base/wc2.1"
  full_url <- paste(base_url, wc_res, postfix, sep = "_")
  destfile <- tempfile()
  # download this zip file into a temp file
  curl::curl_download(full_url,
    destfile = destfile,
    quiet = FALSE
  )

  # unzip it to a temporary directory
  destpath <- file.path(tempdir(), "to_unzip")
  utils::unzip(destfile, exdir = destpath)
  wc_rast <- terra::rast(dir(destpath, pattern = ".tif", full.names = TRUE))
  # sort out variable names
  if (!(grepl("altitude", bio_var))) {
    # digits at the end of the name are the key identifier of each variable
    digits_at_end <- sprintf(
      "%02d",
      as.numeric(substr(
        names(wc_rast),
        regexpr("_\\d+\\b", names(wc_rast)) + 1,
        nchar(names(wc_rast))
      ))
    )

    # now we need to rename the layers
    names(wc_rast) <-
      paste0(var_prefix, digits_at_end)
  } else {
    names(wc_rast) <- "altitude"
  }

  # and finally we save it as a netcdf file
  time_bp(wc_rast) <- rep(35, nlyr(wc_rast))

  terra::writeCDF(wc_rast,
    filename = filename, compression = 9,
    split = TRUE, overwrite = TRUE
  )

  # fix time axis (this is a workaround if we open the file with sf)
  nc_in <- ncdf4::nc_open(filename, write = TRUE)
  ncdf4::ncatt_put(nc_in, varid = "time", attname = "axis", attval = "T")
  ncdf4::nc_close(nc_in)

  # clean up
  unlink(file.path(destpath, "*"))
}