Incorrect documentation of "theta" CSUB observation type

[RichardWinston]

Describe the bug
In mf6io.pdf, the documentation of the "theta" observation type says its ID must be "icsubno" but in ex-gwf-csub-p04, a theta observation is included on line 47 in which the ID is "02_09_10" which is not an integer.

[langevin-usgs]

It looks like the following bit of code allows for icsubno or boundname. @jdhughes-usgs, does the observation table need to be updated for theta and also for the delay-* observations so that that they can also be provided as boundname?

    ! -- get icsubno number or boundary name
    if (obsrv%ObsTypeId == 'CSUB' .or. &
        obsrv%ObsTypeId == 'INELASTIC-CSUB' .or. &
        obsrv%ObsTypeId == 'ELASTIC-CSUB' .or. &
        obsrv%ObsTypeId == 'SK' .or. &
        obsrv%ObsTypeId == 'SKE' .or. &
        obsrv%ObsTypeId == 'THETA' .or. &
        obsrv%ObsTypeId == 'THICKNESS' .or. &
        obsrv%ObsTypeId == 'INTERBED-COMPACTION' .or. &
        obsrv%ObsTypeId == 'INELASTIC-COMPACTION' .or. &
        obsrv%ObsTypeId == 'ELASTIC-COMPACTION' .or. &
        obsrv%ObsTypeId == 'DELAY-HEAD' .or. &
        obsrv%ObsTypeId == 'DELAY-GSTRESS' .or. &
        obsrv%ObsTypeId == 'DELAY-ESTRESS' .or. &
        obsrv%ObsTypeId == 'DELAY-PRECONSTRESS' .or. &
        obsrv%ObsTypeId == 'DELAY-COMPACTION' .or. &
        obsrv%ObsTypeId == 'DELAY-THICKNESS' .or. &
        obsrv%ObsTypeId == 'DELAY-THETA' .or. &
        obsrv%ObsTypeId == 'DELAY-FLOWTOP' .or. &
        obsrv%ObsTypeId == 'DELAY-FLOWBOT') then
      call extract_idnum_or_bndname(string, icol, istart, istop, nn1, bndname)