New recipe to outline, or dedicated method for, attaching orography

A fairly common operation for users of cf, and one we know we will need for TWINE-VISION project purposes, is to read in a dataset defining orography and attach it to some other field to define the orography context for the information encapsulated by the field.

As clarified by @davidhassell today, this can be achieved with a three-step process involving reading in the separate orography, converting it from a read-in field to a domain ancillary, and finally attaching it to the other field. (One minor complication is that the X and Y coordinates need to be identified so they are assigned in the right order during the conversion to the domain ancillary, but that is about all that could go wrong.)

Given that users will often want to perform this, it would be nice to make it simpler for them than this, either by:

• having a dedicated method which will do the three steps in one to make the attachment, called something natural like attach_orogaphy or similar; or
• adding a new recipe to our documentation to outline how the attachment can be done, with descriptions of the steps.

It's not yet clear to me which is best, so I will leave that as an open question/decision for future.

Start by reviewing the existing three-step workflow described in the issue: read the orography, convert it to a domain ancillary, and attach it to the target field. Resolve whether this should be a dedicated method or a documentation recipe, including how X and Y coordinates are identified. Done means users have a clear, supported way to attach a separate orography dataset.