Identify 'lv_' variables as vertical coordinates.

[ahijevyc]

ncl_convert2nc starts its vertical coordinate variables with the
string "lv_". Added this to the set of regular expressions to match
in metpy.xarray.coordinate_criteria.

For example, when you download NARR grib files from https://nomads.ncdc.noaa.gov/data/narr/
and convert them from grib to netCDF, the 10-m zonal wind is stored in a variable called U_GRD_221_HTGL with a vertical coordinate lv_HTGL5.

Here are its attributes:
int lv_HTGL5(lv_HTGL5) ;
lv_HTGL5:long_name = "fixed height above ground" ;
lv_HTGL5:units = "m" ;

Unfortunately, because lv_HTGL5 doesn't have units of pressure, nor the "positive" attribute indicating positive is "up". metpy.parse_cf() doesn't recognize this as a vertical coordinate. It would be nice if ncl_convert2nc did this for you, but I think its development is over.

Could we have metpy.parse_cf() recognize this prefix? i.e. "lv_"

Otherwise you get an Attribute error
AttributeError: vertical attribute is not available.

p.s. sorry about empty line in line 194. not needed. Couldn't figure out an easy way to remove.

Description Of Changes

Checklist

• Closes #xxxx
• Tests added
• Fully documented

[CLAassistant]

All committers have signed the CLA.

[jthielen]

Good catch! Would you be able to add a test for this by adding an entry like ('lv_HTGL5', 'vertical'), to the regex_matches list in the test file? See

MetPy/tests/test_xarray.py

Lines 372 to 410 in 82e5258

	regex_matches = [
	('time', 'time'),
	('time1', 'time'),
	('time42', 'time'),
	('Time', 'time'),
	('TIME', 'time'),
	('bottom_top', 'vertical'),
	('sigma', 'vertical'),
	('HGHT', 'vertical'),
	('height', 'vertical'),
	('Altitude', 'vertical'),
	('depth', 'vertical'),
	('isobaric', 'vertical'),
	('isobaric1', 'vertical'),
	('isobaric42', 'vertical'),
	('PRES', 'vertical'),
	('pressure', 'vertical'),
	('pressure_difference_layer', 'vertical'),
	('isothermal', 'vertical'),
	('y', 'y'),
	('Y', 'y'),
	('lat', 'lat'),
	('latitude', 'lat'),
	('Latitude', 'lat'),
	('XLAT', 'lat'),
	('x', 'x'),
	('X', 'x'),
	('lon', 'lon'),
	('longitude', 'lon'),
	('Longitude', 'lon'),
	('XLONG', 'lon')
	]
	@pytest.mark.parametrize('test_tuple', regex_matches)
	def test_check_axis_regular_expression_match(test_ds_generic, test_tuple):
	"""Test the variety of possibilities for check_axis in the regular expression match."""
	data = test_ds_generic.rename({'e': test_tuple[0]})
	assert check_axis(data[test_tuple[0]], test_tuple[1])

[ahijevyc]

Yes, I will add a test for this in the test_xarray.py file, as suggested!

[zbruick]

Thanks for the contribution!

[dopplershift]

My preference would also be to fix NCL to be CF-compliant, but alas. I think this is a reasonable way to make things more useful. Thanks for the PR!