On Fri 2005-06-24T14:53:46 +0200, Andrea Preite Martinez hath writ:
> < Q | obs.atmos.refractAngle |Added
I am a bit mystified by the utility of this quantity.
In actual practice it is usually impossible to distinguish the amount of refraction from the amount of unmodelled telescope flexure. It is a rare situation when it is actually possible to make a measurement of the refraction angle. Any error is simply accommodated by the ever-requisite fine pointing adjustment of the telescope.
Under what circumstances is refractAngle actually used? What drives any requirements for it to be accurate or precise?
On the other hand, the differential refraction, which is to say both the change in the amount of refraction over a field of view and the atmospheric dispersion due to refraction are quantities which are often calculated and used.
The differential refraction and the atmospheric dispersion are needed for designing and manufacturing focal plane constructs for multi object spectrographs.
The dispersion enters directly into the decision about the position angle at which a slit spectrograph should be oriented.
A map of the differential refraction is necessary for multi-slit spectrographs and for fiber-fed spectrographs.
In my experience these two are needed far more often, and to greater precision, than the absolute amount of refraction.
> S | pos.earthop | Earth orientation parameters
> Q | pos.earthop.nutation | Earth nutation
These specifications seem incomplete, but I'm not sure what they are supposed to be communicating.
As a starter for discussion, does nutation refer to IAU 2000A model, IAU 2000B model, or the models resulting from the FK5-based revision performed between 1976/1984, or the nutation based on the older Newcomb models of precession, or what?
Is polar motion and UT1 relevant? Do they use the IAU 2000 models, or one of the older forms used in previous versions of the IERS conventions, or the even older FK5 forms, or what?
> Q | pos.eq.ha | Hour-angle
> Q | pos.az | Position in alt-azimutal frame
> Q | pos.az.alt | Alt-azimutal altitude
> Q | pos.az.azi | Alt-azimutal azimut
> Q | pos.az.zd | Alt-azimutal zenith distance
If I understand the work of the general relativists who have been contributing to various IAU initiatives over the past two decades then all of the above quantities do not have precisely-defined meanings. There is no mathematical framework for a rigorous mapping of celestial sphere to observation angles which is valid at the microarcsecond level.
The meaning of hour angle has been called into question at a precision much less tight by the IAU 2000 redefinition of UT1 and the introduction of the CIO/CEO- and TIO/TEO- based expressions. There is an active IAU committee trying to work out nomenclature for the "classical" vs. the "CIO-based" versions of these sorts of quantities.
Should the definitions of these admit that they may not be meaningful at the milliarcsecond level and give references to the draft documents which explain why?
On Tue 2005-06-28T15:16:26 +0200, Pierre Didelon hath writ:
> A.3.1 pos.earth
> Q | pos.earth.lon | Longitude on Earth
> exists but not latitude!
> Q | pos.earth.altitude | Altitude on Earth
> exists but not distance which give the distance to the center
>
> A.3.2 lon/lat/alt/dist
In the context of FITS WCS Paper III we decided that it was not a good idea to admit the concepts of latitude, longitude, and height into the standard.
(Note, by the way that I strongly prefer to use the term "height" rather than "altitude" or "elevation". The geodetic community also uses "height" at least in part to avoid confusion with the other two, which may be angles.)
Unless there is also a UCD word something like pos.earth.geodatum it is not possible to associate a precise meaning with .lon, .lat, and .alt
As we point out in FITS WCS Paper III, there are something like 1000 different geodetic datums in use, and some of their positions differ from each other by kilometers. Rather than burden FITS with the need to recognize even a small subset of those datums we decided to omit angular body coordinates and stick to Cartesian coordinates only.
True, even Cartesian coordinates technically require the specification of a datum, but in current usage all the Cartesian terrestrial datums differ by only about 0.1 m.
I recognize that there is a very strong tradition for the use of angular body coordinates, but if precise meaning is required they are not a good idea.
-- Steve Allen <sla-at-ucolick.org> WGS-84 (GPS) UCO/Lick Observatory Natural Sciences II, Room 165 Lat +36.99858 University of California Voice: +1 831 459 3046 Lng -122.06014 Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 mReceived on 2005-06-29Z00:13:36