Source DM: Call For Scientific Requirements

During the spring 2018 Interop in Victoria, we had a hands-on session focused on the tools available to work with VO data models. The goal this session was to annotate data from different origins in order to make it inter-operable to eachother. The main concern expressed during this session was not related to the tools but to the lack of models for sources.

After some discussions, it appeared that having a global model for all missions and all archives would lead to a very complex model not usable in practice. The reason is that such a global model would have to merge all quantities of all user-cases it is supposed to cover.

It seemed to us more realistic to design a basic model, namely a core-model, with a reasonable set of quantities (ra, dec, ....) useful for all missions and to make it expendable with more specific components (e.g. parallax). So that a model aware tool could compare sources from different origins on the basis of the model components really instantiated. The data annotation process would also be simpler since only the core-model and some components of interest should be tagged.

The purpose of this page is to gather requirements from people involved in different missions and archives with the goal of proposing a draft in 2019. It is to be noted that Gaia people at ESA have already presented their own model proposal which is for us a good preliminary example of what is intended.

Core-model Components

• ra : right ascension
• dec : declination
• name: source name
• flux:photometric measurement or magnitude in several bands ,

LSST Specific Components

--

EUCLID Specific Components

--

CFIS Specific Components

--

GAIA Specific Components

As described in https://wiki.ivoa.net/twiki/bin/view/IVOA/SourceDataModel, the basic mechanism of Gaia Sources can be described as follows:

- Detections are done. In this case, the basic source (detection) data model can be used for positions, identifiers and fluxes/magnitudes. Also, using spectrometric techniques, radial velocity or redshift can be derived. Also, as known, fluxes in one or more bands are also associated to the detections (e.g. to create and spectrum)

- Some properties (e.g. proper motions and parallaxes) require a cross-identification of source detections. This is the case of different main catalogues of astronomy (e.g. Gaia) where different transient observations of different detections are consolidated in a single instance adding extra information. Also, metadata can be combined from different catalogues. This kind of objects are called crossmatched sources in the present DM.

For this crossmatched objects as output of different detections, the extended data model is needed. Some of this properties are:

- parallax (it is, in a naive way, the inverse of the distance but catalogues use to provide parallax as there are some possible errors involved in the calculation of the distance). See, e.g. http://www.mpia.de/~calj/parallax.pdf

- proper motions (pm_ra and pm_dec) are also an output of the crossmatch of different detections what provides the displacement of the sources in the sky

- Extra flux/magnitudes measurements of the source at different times can produce a source time series. Depending on the data model to be used (spectral data model or time series data model), the connection between detections can be done only to the spectral data model for detections and crossmatched sources or to the time series data model (if any) for the crossmatched sources

CDS Specific Components

--

Exoplanet Specific Components

Exoplanets domain looks at sources at, at least, two levels: stellar components & extra-solar planets. But also at the combination of the two.

The two are of course connected, thus the basic ra,dec&name would do as a primary identification of a system and its components (stellar_id + planet_letter).

Specific details are needed to properly characterize the stellar host (spectral type and physical parameters, activity indexes, ...) including the host multiplicity: is ita binary or not.

For the planet (as a source) caution has to be taken to separate candidates from confirmation and other levels. This implies that "status changes" apply, not only in a planet existing, but also in its classification (is it a Jovian or a Brown Dwarf?).

If the system itself is considered a source, than its decomposition should be taken into account, but that's probably a matter for the exoplanets data model iself (involving orbital details and such).

Time evolution plays also a role in detection, where radial velocity time series derived from high resolution spectroscopy play a role in detection and confirmation of candidates.

Chandra Archive

--

Diffuse and/or complex-morphology sources

Despite the fact that one can use some baricentric solution to roughly identify the position of a source, this is sometimes not enough. And the scenario gets more complex in multi-wavelength observations.

A subdivision that can be done is on

• compact sources
• diffuse sources

The former can be defined by ra,dec and ellipsoids, though it can be difficult associate/cross-match sources that appear single in one observational band and multiple at a different band.

The latter can be described or analytically or through a tesselletion solution, but surely ellipsoidal-error boundaries won't work in this case.

[The above description comes from morphological structures within the Milky-Way, but I've been told there exist other examples. Within the MW distance estimation and related parameters are useful]