mpcalc.advection dim_order in example, and missing values

Question

on the metpy webpage, they example provided for the 850hPa temperature advection is as follows:

adv = mpcalc.advection(temp_850 * units.kelvin, [u_wind_850, v_wind_850],
                   (dx, dy), dim_order='yx') * units('K/sec')

But based on the documentation, which states:

The order of the dimensions of the arrays must match the order in which
the wind components are given.  For example, if the winds are given [u, v],
then the scalar and wind arrays must be indexed as x,y (which puts x as the
rows, not columns).

It seems to me the dim_order in the example should be xy and not yx? as we give u then v, and dx then dy, rather than the other way around?

I also found out that even if my input temperature field has missing values at some grid points, metpy somehow still returns advection values at these points, rather than a missing value. How is this possible?

Answer 1

About the missing values, i found the issue. When applying the smoothing as per the example, this fills in where we have missing values. The advection outputs prior to smoothing have missing values where the input temperature and winds are missing. It's the smoothing which seems to interpolate where we have missing data.

But i am still confused about the dim_order though