Unit conversion fails in declarative wind barb plot

[sgdecker]

I thought this was supposed to work, but subtleties with units often get me. In the code below, I can get a plot if I only include the 500-mb heights (which includes a unit conversion), but no dice when I try wind barbs (also with a unit conversion).

from datetime import datetime

import xarray as xr
from metpy.io import GempakGrid
from metpy.units import units
from metpy.plots.declarative import ContourPlot, BarbPlot, MapPanel, PanelContainer

gem_file_name = '/ldmdata/gempak/model/nam/22092112_nam211.gem'
gem_file = GempakGrid(gem_file_name)

plot_time = datetime(2022, 9, 21, 12)
ht500 = gem_file.gdxarray(parameter='HGHT', date_time=plot_time, level=500)[0]
ht500 = ht500 * units('m')

u500 = gem_file.gdxarray(parameter='UREL', date_time=plot_time, level=500)[0]
v500 = gem_file.gdxarray(parameter='VREL', date_time=plot_time, level=500)[0]
u500 = u500 * units('m/s')
v500 = v500 * units('m/s')

wind = xr.merge([u500, v500])

print(wind)
print(wind['urel'])
print(wind['vrel'])

barbs = BarbPlot()
barbs.data = wind
barbs.time = plot_time
barbs.field = ['urel', 'vrel']
barbs.earth_relative = False
barbs.skip = (3, 3)
barbs.plot_units = 'knot'

cp = ContourPlot()
cp.data = ht500
cp.time = plot_time
cp.contours = list(range(460, 700, 6))
cp.linecolor = 'green'
cp.linestyle = 'solid'
cp.clabels = True
cp.plot_units = 'dam'

panel = MapPanel()
panel.area = [-120, -74, 22, 55]
panel.projection = 'lcc'
panel.layers = ['states', 'coastline', 'borders']
panel.title = f'500-mb Heights and Winds at {plot_time}'
panel.plots = [barbs, cp]

pc = PanelContainer()
pc.size = (12, 9)
pc.panels = [panel]
pc.show()

The output:

Dimensions:    (time: 1, pres: 1, x: 93, y: 65)
Coordinates:
  * time       (time) datetime64[ns] 2022-09-21T12:00:00
  * pres       (pres) int64 500
  * x          (x) float32 -4.226e+06 -4.145e+06 ... 3.17e+06 3.251e+06
  * y          (y) float32 2.035e+06 2.117e+06 2.198e+06 ... 7.155e+06 7.237e+06
    metpy_crs  object Projection: lambert_conformal_conic
Data variables:
    urel       (time, pres, y, x) float32 <Quantity([[[[ 0.9146438  0.9246438...
    vrel       (time, pres, y, x) float32 <Quantity([[[[ 2.710384    1.430383...
<xarray.DataArray 'urel' (time: 1, pres: 1, y: 65, x: 93)>
<Quantity([[[[ 0.9146438  0.9246438  0.8946438 ... -5.325356  -6.275356
    -6.875356 ]
   [ 1.5546438  1.7946438  1.5446438 ... -4.995356  -5.735356
    -6.455356 ]
   [ 1.1646438  1.3846438  1.2646438 ... -5.305356  -5.165356
    -5.765356 ]
   ...
   [22.764643  23.924643  24.304644  ... 19.364643  19.784643
    20.164644 ]
   [15.884644  21.864643  24.194643  ... 20.214643  21.474644
    21.594643 ]
   [11.184644  14.564644  19.484644  ... 18.704643  18.954643
    18.634644 ]]]], 'meter / second')>
Coordinates:
  * time       (time) datetime64[ns] 2022-09-21T12:00:00
  * pres       (pres) int64 500
  * x          (x) float32 -4.226e+06 -4.145e+06 ... 3.17e+06 3.251e+06
  * y          (y) float32 2.035e+06 2.117e+06 2.198e+06 ... 7.155e+06 7.237e+06
    metpy_crs  object Projection: lambert_conformal_conic
<xarray.DataArray 'vrel' (time: 1, pres: 1, y: 65, x: 93)>
<Quantity([[[[ 2.710384    1.4303839   1.140384   ... -2.3096159  -2.879616
    -2.939616  ]
   [ 2.630384    1.870384    1.710384   ... -1.499616   -2.439616
    -2.5496159 ]
   [ 2.390384    1.8303839   2.330384   ... -0.529616   -1.939616
    -2.079616  ]
   ...
   [ 6.6203837  10.700384   14.860384   ...  2.610384    0.46038398
    -1.399616  ]
   [ 0.95038396  7.9303837  14.220384   ...  1.740384   -0.409616
    -2.239616  ]
   [-3.0796158   2.220384   11.870383   ...  1.020384   -0.779616
    -2.239616  ]]]], 'meter / second')>
Coordinates:
  * time       (time) datetime64[ns] 2022-09-21T12:00:00
  * pres       (pres) int64 500
  * x          (x) float32 -4.226e+06 -4.145e+06 ... 3.17e+06 3.251e+06
  * y          (y) float32 2.035e+06 2.117e+06 2.198e+06 ... 7.155e+06 7.237e+06
    metpy_crs  object Projection: lambert_conformal_conic
Traceback (most recent call last):
  File "/chariton/decker/test/test.py", line 53, in <module>
    pc.show()
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 622, in show
    self.draw()
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 609, in draw
    panel.draw()
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 866, in draw
    p.draw()
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 1553, in draw
    self._build()
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 1577, in _build
    x_like, y_like, u, v = self.plotdata
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/metpy/plots/declarative.py", line 1538, in plotdata
    if 'degree' in x.units:
  File "/home/decker/local/miniconda3/envs/met433/lib/python3.9/site-packages/xarray/core/common.py", line 256, in __getattr__
    raise AttributeError(
AttributeError: 'DataArray' object has no attribute 'units'

This is with MetPy 1.3.1 and xarray 2022.6.0

[sgdecker]

Ahh, I knew I'd seen this before: #2118
That has the workaround.