RC2 advection calculation does not get coordinate data from DataArray

[kgoebber]

Try to include:

• Code Sample, a copy-pastable example if possible
• A "Minimal, Complete and Verifiable Example" will make it much easier for maintainers to help you

# Your code here
from datetime import datetime

import metpy.calc as mpcalc
from metpy.units import units
import xarray as xr

date = datetime(2020, 10, 26, 12)

ds = xr.open_dataset('https://thredds.ucar.edu/thredds/dodsC/grib/NCEP/GFS/Global_0p5deg/Best').metpy.parse_cf()

data_slice_dict = dict(time=date, vertical=850*units.hPa, lat=slice(55, 20), lon=slice(360-140, 360-60))

# First pull wind data for plotting barbs and computing advection
uwnd_850 = ds['u-component_of_wind_isobaric'].metpy.sel(data_slice_dict)
vwnd_850 = ds['v-component_of_wind_isobaric'].metpy.sel(data_slice_dict)
tmpc = (ds.Temperature_isobaric.metpy.sel(data_slice_dict)).metpy.convert_units('degC')
smooth_tmpc = mpcalc.smooth_gaussian(tmpc, 5)

t_adv_850 = mpcalc.advection(smooth_tmpc, uwnd_850, vwnd_850)

Output:

 ---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-8-5722b3ee1007> in <module>
      3 
      4 # Compute the temperature advection at 850 hPa
----> 5 t_adv_850 = mpcalc.advection(smooth_tmpc, uwnd_850, vwnd_850)

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/xarray.py in wrapper(*args, **kwargs)
   1158 
   1159             # Evaluate inner calculation
-> 1160             result = func(*bound_args.args, **bound_args.kwargs)
   1161 
   1162             # Wrap output based on match and match_unit

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/calc/kinematics.py in advection(scalar, u, v, w, dx, dy, dz, x_dim, y_dim, vertical_dim)
    292             (u, dx, x_dim),
    293             (v, dy, y_dim),
--> 294             (w, dz, vertical_dim)
    295         )
    296         if wind is not None

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/calc/kinematics.py in <genexpr>(.0)
    294             (w, dz, vertical_dim)
    295         )
--> 296         if wind is not None
    297     )
    298 

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/calc/tools.py in wrapper(f, **kwargs)
    894         if 'x' in kwargs or 'delta' in kwargs:
    895             # Use the usual DataArray to pint.Quantity preprocessing wrapper
--> 896             return preprocess_and_wrap()(func)(f, **kwargs)
    897         elif isinstance(f, xr.DataArray):
    898             # Get axis argument, defaulting to first dimension

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/xarray.py in wrapper(*args, **kwargs)
   1158 
   1159             # Evaluate inner calculation
-> 1160             result = func(*bound_args.args, **bound_args.kwargs)
   1161 
   1162             # Wrap output based on match and match_unit

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/calc/tools.py in first_derivative(f, axis, x, delta)
    970 
    971     """
--> 972     n, axis, delta = _process_deriv_args(f, axis, x, delta)
    973     take = make_take(n, axis)
    974 

~/miniconda3/envs/new_main/lib/python3.7/site-packages/metpy/calc/tools.py in _process_deriv_args(f, axis, x, delta)
   1273         delta = np.diff(x, axis=axis)
   1274     else:
-> 1275         raise ValueError('Must specify either "x" or "delta" for value positions.')
   1276 
   1277     return n, axis, delta

ValueError: Must specify either "x" or "delta" for value positions.

• Problem description When testing MetPy calculation functions from MetPy 1.0rc2, the advection calculation does not appear to properly use the DataArray coordinate information to fill in dx, dy values. According to the documentation, if you feed a scalar in as a DataArray, it should be able to fill in the dx, dy, dz values as needed.

• Expected output Should be able to feed in a DataArray for scalar (such as the smooth_tmpc in the example above) and not have to manually fill in the dx, dy, or dz values for computing advection.

[jthielen]

@kgoebber I guess this is what I get for poor testing in #1490...somehow the grid arg decorator got missed on the new advection. Also, this uncovered a separate bug in the grid arg decorator when there is a vertical scalar non-dim coord, so that was a good catch too. Does this fix work for you https://github.com/jthielen/MetPy/commit/fe46d137f56aeb493b8fbf54156f115b69e7e255?

[kgoebber]

That does the trick! Working as intended now. Thanks!