Propose solar specific terms to the IVOA UCD List

[Cadair]

The UCD controlled vocab list is used by astropy to populate the world_axis_physical_types property of the WCS API.

There is a process to propose modifications to this list which involves:

1. Write a Vocabulary Enhancement Proposal.
2. Send VEP to the UCD Science Committee
3. Wait?

Writing a VEP

Here is a summary of what I can discern about writing a VEP from trawling the sprawling IVOA wiki.

• Documentation on writing a VEP: https://ivoa.net/documents/Vocabularies/20190905/WD-Vocabularies-2.0-20190905.html#tth_sEc4.2
• Template for UCD VEPs: https://wiki.ivoa.net/internal/IVOA/VEPsForUCD/UCD-VEP-template.txt
• A page about solar data: https://wiki.ivoa.net/twiki/bin/view/IVOA/UcdSolarDiscussion
• List of VEPs: https://wiki.ivoa.net/twiki/bin/view/IVOA/VEPsForUCD

What we would want in a VEP

I think one coherent VEP we should write would be pos. type words for all of the common solar physics physical coordinate systems (i.e. all the ones in the Thompson WCS paper).

Maybe other things?

[Cadair]

@ayshih I am happy to do the admin on this, but would like your input on what we should include.

[ayshih]

Sure!

From poking around, I see that pos.heliocentric and pos.heliographic have been pending/stalled since 2018 as part of their Request for Modification (RFM) process (https://wiki.ivoa.net/twiki/bin/view/IVOA/UCDList_1-3_RFM). I imagine a VEP would be needed to get some motion on those.

[Cadair]

pos.heliocentric is already on the list.

S | pos.heliocentric                  |  Heliocentric position coordinate (solar system bodies)

It's interesting to note that astropy doesn't use pos.heliocentric though.

We would need something for:

• heliographic stonyhurst
• heliographic carrington
• helioprojective

What other frames should we propose at the same time?

[ayshih]

It's interesting to note that astropy doesn't use pos.heliocentric though.

I believe that pos.heliocentric is solely a descriptor of the origin of the coordinate system, so there's no way to use it directly for coordinate components.

I don't understand how unique these descriptors are intended to be. For example, pos.eq.ra/pos.eq.dec are for equatorial coordinates, but presumably do not inherently distinguish between the many flavors of equatorial coordinates.

Even though Stonyhurst and Carrington heliographic coordinates are both heliographic, I think in their parlance both would already be (partially) covered as:

• Stonyhurst heliographic -> pos.heliocentric + pos.bodycentric + pos.bodyrc.*
• Carrington heliographic -> pos.heliocentric + pos.bodygraphic + pos.bodyrc.*

Unfortunately, pos.bodyrc only has alt and not r, while pos.spherical only has colat and not lat...

[Cadair]

I just noticed that astropy also has this: "DSUN": "custom:pos.distance.sunToObserver" in it's FITS mapping, so maybe also that?

I don't understand how unique these descriptors are intended to be.

Me either. I think there's probably a disconnect between what the UCD was written for and what astropy are trying to use it for.

I think in their parlance both would already be (partially) covered as

re-reading the doc, I am starting to think that astropy is "holding this wrong".