readDWD.grib2: read nwp forecast data
In brry/rdwd: Select and Download Climate Data from 'DWD' (German Weather Service)

readDWD.grib2

R Documentation

read nwp forecast data

Description

read gridded numerical weather prediction data. Intended to be called via readDWD().

Usage

readDWD.grib2(file, pack = "terra", bargs = NULL, quiet = rdwdquiet(), ...)

Arguments

`file`	Name of file on harddrive, like e.g. cosmo-d2_germany_regular-lat-lon_single-level_2021010100_005_T_2M.grib2.bz2
`pack`	Char: package used for reading. One of "terra" or "stars". "rgdal" (for the deprecated cosmo-d2 data) is no longer available, see issue. DEFAULT: "terra"
`bargs`	Named list of arguments passed to `R.utils::bunzip2()`, see `gargs` in `readDWD.raster()`. DEFAULT: NULL
`quiet`	Silence readGDAL completely, including warnings on discarded ellps / datum. DEFAULT: FALSE through `rdwdquiet()`
`...`	Further arguments passed to `terra::rast()` or `stars::read_stars()`.

Value

terra or stars object, depending on pack

Author(s)

Berry Boessenkool, berry-b@gmx.de, Jan 2021.

Examples

## Not run:  # Excluded from CRAN checks, but run in localtests
nwp_t2m_base <- "ftp://opendata.dwd.de/weather/nwp/icon-d2/grib/15/soiltyp"
nwp_urls <- indexFTP("", base=nwp_t2m_base, dir=tempdir())
# for p instead of soiltyp, icosahedral_model-level files fail with GDAL errors,
# see https://github.com/brry/rdwd/issues/28
# regular-lat-lon_pressure-level files work with pack="terra" or "stars"

nwp_file <- dataDWD(tail(nwp_urls,1), base=nwp_t2m_base, dir=tempdir(), 
                    joinbf=TRUE, dbin=TRUE, read=FALSE)
nwp_data <- readDWD(nwp_file)
terra::plot(nwp_data)
addBorders() # the projection seems to be perfectly good :)

# index of GRIB files
if(FALSE){ # indexing takes about 6 minutes!
grib_base <- "ftp://opendata.dwd.de/weather/nwp/icon-d2/grib"
grib_files <- indexFTP("", base=grib_base, dir=tempdir())
for(f in unique(substr(grib_files, 1,3))) print(grib_files[which(substr(grib_files, 1,3)==f)[1]])
View(data.frame(grep("regular",grib_files, value=TRUE)))
}

# Project_TRY bz2 netcdf data:
url <- "hourly/Project_TRY/air_temperature_mean/TT_201102.nc.bz2" # 97 MB
file <- dataDWD(url, base=gridbase, joinbf=TRUE, dir=tempdir(), read=FALSE)
nc <- readDWD(file) # should also be using readDWD.grib2
# Setting layer=1:2 takes 4 minutes, just go one at a time, I guess:
ncp <- plotRadar(nc, main=paste(terra::longnames(nc), terra::time(nc)), layer=1,
                 col=berryFunctions::seqPal(), proj="nc", extent="nc")

## End(Not run)

brry/rdwd documentation built on Jan. 26, 2025, 11:33 a.m.