Help with working with a complex dataset

[Matt Thompson]

This might be a problem where the answer is "Can't do it" but I thought I'd try and ask the gurus here. Namely, I work with a weather model that used to output pretty much exclusively in lat-lon output a la:

In [33]: print cube[0]

surface_pressure / (Pa)             (time: 1; latitude: 49; longitude: 96)

     Dimension coordinates:

          time                           x            -              -

          latitude                       -            x              -

          longitude                      -            -              x

That's the sort of thing I'm used to dealing with. However, internally, we actually use a cubed-sphere grid which is, as it sounds, a cube that is mapped to a sphere. Recently, there's been an effort to output our data essentially on a "native" grid to avoid interpolating our output to a latlon grid. (A plotting routine might do something, but the hope is to stay native as long as possible.)

This is something that the CF Conventions don't really understand (we'll probably propose something for it), but for now we've decided to represent it as:

In [34]: print cube_new[0]

surface_pressure / (Pa)             (time: 1; cubed-sphere face: 6; latitude: 48; longitude: 48)

     Dimension coordinates:

          time                           x                     -            -              -

          cubed-sphere face              -                     x            -              -

          latitude                       -                     -            x              -

          longitude                      -                     -            -              x

     Auxiliary coordinates:

          latitude                       -                     x            x              x

          longitude                      -                     x            x              x

Here you see that our variable is now sort of split into 6 faces (one for each of the cube). There is still a lat and an lon associated with it, though internally it is x,y (as seen with ncdump):

float SLP(time, nf, y, x) ;

SLP:long_name = "sea_level_pressure" ;

SLP:units = "Pa" ;

SLP:_FillValue = 1.e+15f ;

SLP:missing_value = 1.e+15f ;

SLP:fmissing_value = 1.e+15f ;

SLP:scale_factor = 1.f ;

SLP:add_offset = 0.f ;

SLP:standard_name = "sea_level_pressure" ;

SLP:vmin = -1.e+15f ;

SLP:vmax = 1.e+15f ;

SLP:valid_range = -1.e+15f, 1.e+15f ;

SLP:coordinates = "lons lats" ;

SLP:grid_mapping = "cubed_sphere" ;

So, I was wondering, is there a way to "collapse" or "squish" the 6 faces in a way that iris' quickplot or something similar can view it as 2D as it expects?