🎁 Wrapper for `wgrib2`#

Herbie provides a very basic wrapper for wgrib2. Note that the wgrbi2 functionality requires wgrib2 is in your path. Note that wgrib2 can only perform actions on local GRIB2 files, not remote files.

Install wgrib2#

For Linux or Mac, the easiest way to install wgrib2 is installing from conda-forge. My preferred method is with Pixi:

pixi global install wgrib2

Unofficial wgrib2 docs#

I attempted to re-publish wgrib2 docs (unofficial) that may or may not be helpful to you.

[1]:

from herbie import Herbie, wgrib2
from pathlib import Path
import pandas as pd

import xarray as xr
from herbie.toolbox import EasyMap, ccrs, pc

Lets start by getting a GRIB2 file…

[2]:

H = Herbie("2023-02-01")

ss = "[U|V]GRD:10 m above ground"  # ss is for search
H.download(ss)

myFile = H.get_localFilePath(ss)
myFile, myFile.exists()

✅ Found ┊ model=hrrr ┊ product=sfc ┊ 2023-Feb-01 00:00 UTC F00 ┊ GRIB2 @ aws ┊ IDX @ aws

[2]:

(PosixPath('/home/blaylock/data/hrrr/20230201/subset_d7ef391b__hrrr.t00z.wrfsfcf00.grib2'),
 True)

You can print the raw GRIB2 inventory from wgrib2 like this:

[3]:

# Print the standard inventory of a grib2 file
print(wgrib2.inventory(myFile))

1:0:d=2023020100:UGRD:10 m above ground:anl:
2:2381615:d=2023020100:VGRD:10 m above ground:anl:

You can create and save the <filename>.grib2.idx inventory text files like this:

[4]:

# Create a standard inventory file

idxFile = wgrib2.create_inventory_file(myFile)

idxFile, idxFile.exists()

[4]:

(PosixPath('/home/blaylock/data/hrrr/20230201/subset_d7ef391b__hrrr.t00z.wrfsfcf00.grib2.idx'),
 True)

[5]:

# Read the index file (with pandas)
pd.read_csv(idxFile, delimiter=":", header=None)

[5]:

	0	1	2	3	4	5	6
0	1	0	d=2023020100	UGRD	10 m above ground	anl	NaN
1	2	2381615	d=2023020100	VGRD	10 m above ground	anl	NaN

You can check if the vector quantities are grid-relative or earth-relative vectors

[6]:

wgrib2.vector_relative(myFile)

All winds are grid-relative winds.

[6]:

{'winds(grid)'}

You can create a regional subset of the file. This uses the -small_grib option to trim the GRIB2 file to a bounding box. This method will, by default, also make a cooresponding .idx file.

[7]:

# Create a regional subset of the file

extent = (-100, -90, 30, 40)

subset_file = wgrib2.region(myFile, extent, name="myRegion")
subset_file

[7]:

PosixPath('/home/blaylock/data/hrrr/20230201/myRegion_subset_d7ef391b__hrrr.t00z.wrfsfcf00.grib2')

[8]:

# Look at the inventory file again to see it is different from before the region subset
pd.read_csv(str(subset_file) + ".idx", delimiter=":", header=None)

[8]:

	0	1	2	3	4	5	6
0	1	0	d=2023020100	UGRD	10 m above ground	anl	NaN
1	2	78653	d=2023020100	VGRD	10 m above ground	anl	NaN

Let’s look at that data, just to make sure it handeled the subset

[9]:

myFile, subset_file

[9]:

(PosixPath('/home/blaylock/data/hrrr/20230201/subset_d7ef391b__hrrr.t00z.wrfsfcf00.grib2'),
 PosixPath('/home/blaylock/data/hrrr/20230201/myRegion_subset_d7ef391b__hrrr.t00z.wrfsfcf00.grib2'))

[10]:

# Get crs from a regular Herbie object
ds = H.xarray(ss)
crs = ds.herbie.crs

# (can't use Herbie to open FILE yet, so just use xarray)
ds_region = xr.open_dataset(subset_file, engine="cfgrib")

/home/blaylock/GITHUB/Herbie/herbie/core.py:1052: UserWarning: Will not remove GRIB file because it previously existed.
  warnings.warn("Will not remove GRIB file because it previously existed.")

[11]:

ax = EasyMap(crs=crs).STATES(color="k").ax
ax.pcolormesh(ds.longitude, ds.latitude, ds.u10, alpha=0.5, transform=pc)
ax.pcolormesh(ds_region.longitude, ds_region.latitude, ds_region.u10, transform=pc)

ax.gridlines(xlocs=extent[:2], ylocs=extent[2:], color="k", ls="--", draw_labels=True)

[11]:

<cartopy.mpl.gridliner.Gridliner at 0x7f89f03461d0>

../../_images/user_guide_tutorial_wgrib2-wrapper_17_1.png

Well that is what we expected, so that is good.

[12]:

ds

[13]:

ds_region

🎁 Wrapper for wgrib2#

Install wgrib2#

Unofficial wgrib2 docs#

🎁 Wrapper for `wgrib2`#