Representation of polarimetry data in WCS

I've been writing polarimetric FITS files for some time now with no regards
for how others do things, but I'm now revising lots of old code and thought
I might take the time to properly standardize it. But first I need to
actually be sure I understand what the standard is!

WCS Paper I defines a 'STOKES' axis convention for storing polarimetric
data, which uses a series of integers to denote the various Stokes parameters
and/or other forms of polarimetry data. WCS Paper III defines the -TAB
convention for representing nonlinear coordinates via table lookup. It
seems like it should be fairly straightforward to combine these two forms to
easily handle arbitrary arangements of polarimetry data.

Paper III says that the CTYPEn keyword for a -TAB axis should follow the 4-3
format, using the 4-letter axis type. So then the proper value to use is
'STOK-TAB', correct?

Next question. Table 7 of WCS Paper I presents the mapping between integer
codes and polarimetry measurements such as Stokes Q or right-circular
polarization. But there are a few possible polarimetry measurements which
are left out of this table, in particular polarized intensity
(P = sqrt(Q^2+U^2+V^2). Is there any convention on an integer code to
denote that with?

Thanks,

- Marshall