Fix bug when calculating UGRID cell areas when non-spatial coordinates span the discrete axis

[davidhassell]

When calculating UGRID cell areas with non-spatial coordinates span that span the discrete axis, an IndexError sometimes occurs:

IndexError: Invalid indices [0, 0, 0] for array with shape (13824, 4)

>>> print (f)
Field: air_pressure_at_mean_sea_level (ncvar%pmsl)
--------------------------------------------------
Data            : air_pressure_at_mean_sea_level(time(2), ncdim%ensemble(3), ncdim%nMesh2d_face(13824)) Pa
Cell methods    : time: mean (interval: 6 h)
Dimension coords: ncvar%ensemble(ncdim%ensemble(3)) = [0.0, 1.0, 2.0]
Auxiliary coords: longitude(ncdim%nMesh2d_face(13824)) = [70.3125, ..., 112.86804962158203] degrees_east
                : latitude(ncdim%nMesh2d_face(13824)) = [0.8827333450317383, ..., -47.359893798828125] degrees_north
                : time(time(2)) = [1978-09-01 12:00:00, 1978-09-02 12:00:00] 360_day
Topologies      : cell:face(ncdim%nMesh2d_face(13824), 4) = [[0, ..., 13755]]
>>> f.cell_area(great_circle=True)
Traceback
    ...
IndexError: Invalid indices [0, 0, 0] for array with shape (13824, 4)

Whether or not this happens depends on the arbitrary of the auxiliary coordinates stored inside the field (specifically within the embedded Constructs instance).

>>> cf.environment(paths=False)
Platform: Linux-5.15.0-94-generic-x86_64-with-glibc2.35
HDF5 library: 1.14.2
netcdf library: 4.9.2
udunits2 library: /home/david/miniconda3/lib/libudunits2.so.0
esmpy/ESMF: 8.4.2
Python: 3.11.4
dask: 2023.12.1
netCDF4: 1.6.4
psutil: 5.9.5
packaging: 23.0
numpy: 1.25.2
scipy: 1.11.3
matplotlib: 3.8.0
cftime: 1.6.2
cfunits: 3.3.6
cfplot: 3.3.0
cfdm: 1.11.0.0
cf: 3.16.0